----, I I I I~· DEPARTMENT OF THE ARMY PAMPHLET 325-10 , en ... ~ C'j ij Standards of STAllS ICAL PRESENTATION HEADQUARTERS, DEPARTMENT OF THE ARMY APRIL 1966 PREFACE This pamphlet establishes standards and provides guidance for the presentation of statistical material in the Department of the Army. It supersedes the manual , STANDARDS OF PRESENTATION, published by Headquarters, Army Service Forces, in May 1946, which was issued as a guide to all major commands by the Chief of the Program Review and Analysis Division, Office of the Comptroller of the Army, in a memorandum dated 21 April 1950, and it also supersedes the Graphic Presentation chapter, Standards of Statistical Presentation, March , 1955 , prepared by the Office of the Comptroller of the Army. Purpose and Scope of Pamphlet The pamphl et is primarily concerned with statistical presentation for published and typewritten reports , but it also discusses problems peculiar to oral statistical presentations . It covers planning, design , production, and format, with major emphasis on design and production . The list of references includes some textbooks that cover the analysis of statistical data as well as the presentation of statistics . It is designed to serve as instructional material for persons not experienced in planning and prepar ing statistical reports and presentations and as a reference handbook for those who have had such experience . It contains sugges tions for typists and draftsmen as well as for statistical analysts . The publication includes information on how to design and construct tables and charts and how to use text in statistical reports . In addition, it contains chapters on production techniques, page and report format , and the preparation of displays for oral presentation . The chapter on charts co::1tains many examples of chart types and comments on appropriate uses of the various types . The standards prescribed are in general agreement with those set forth in AMERICAN STANDARDS : TIME SERIES CHARTS , a manual of Design and Construction, published by the Ameri an Society of Mechanical Engineers, New York , New York . Application of Standards The standards in this manual are designed p~imarily for use in formal published reports , but they are equally effective for informal or special reports throughout the Department of the Army . Sometimes it is not practical to follow the prescribed standards in all respects . The un expected "rush" report must often be produced with whatever skills and facilities are available , in the time allowed . Established reports may also present such a variety of problems that complete uniformity of trea~ment is not always possible . The standards established in this manual should be regarded as flexible enough to permit exceptional treatment in exceptional cases . Many apparently exceptional repor~s, however , are not as unusual as they may seem: they can often be made to conform with the established standards by careful consideration of alternative presentation techniques . When the standards cannot be carried out to the letter, the principles behind them should be f~llowed as closely as possijle . Pam 325-10 PAMPHLET HEADQUARTERS DEPARTMENT OF THE ARMY NO. 325-10 Washington, D.C., 4 April 1966 STANDARDS OF STATISTICAL PRESENTATION Page Chapter l . PRESENTATION l Qualities of Go~d · P;e~e~t~tio~ : : : l Major Problems in Presentation •.. 2 Major Procedures in Presenting Data , 2 The Three Forms of Presentation 4 2 . TEXT PRESENTATION , 5 How to Write Text 5 Numbers in Text •.. 6 A~rangement of Text . 7 Chapter Headings 7 Justifying •••.. 9 3 . TABULAR PRESENTATION ll Basic Principles . . ll Tabular Forms .... 12 Arranging Figures for Emphasis 12 Common Faults in Statistical Tables . l3 Basic Tabular Arrangements . . . . • 13 Position on Page .. . .... . 13 Terms Used in Tabular Presentation 14 Pages Required . . . . 15 Inter nal Structure . . 15 Component Parts of a Table 19 The Title 20 The Stub .. 22 The Boxhead . 28 The Field .. 31 Tabular Notes . 34 Tabular Rulings . . 38 Where to Put Rulings 38 Weight of Rulings . . 38 Blue Pencil Rulings . . ... 38 Other Ruling Considerations . 38 Space Requirements . . . . . 39 Table Width and Length . 39 Unusual Spacing Problems 40 Handling Extreme Cases 40 Preprinted Tabular Forms 42 4. GRAPHIC PRESENTATION 44 General ..... 44 Chart Components . 45 Chart Types . . . . 46 Curve Charts . 46 Surface Charts 5•J Column Charts . 52 Bar Charts .. 54 Combination Types . 58 Frequency Charts . 60 Graphic Tables . . 61 Graphic Types Not Recommended . 62 Pie Charts . . . . . . . 62 Area Charts . . . . . . . 62 Three-dimensional Charts 63 Perspective Charts ... 63 Perspective Plus )-dimensional Treatment 64 Pictographs . . . . . . . . . . . . . . . . 64 Maps ................. . 64 Typical Examples of Poor Presentation .. 65 Planning and Designing Charts . . . . . . . . 66 Development of Charts from Tabular Data . 67 Chart Sketches . . . . . . . . . . . . . 70 Data Sheets for Charts . . . . . . . . . 70 ii Chapter 4. GRAPHIC PRESENTATION -Continued Page Planning and Designing Chart s -Continued Desi gning Bar Charts • • . . 71 Insert Char ts . . . . . . . 72 Subdivided Surface Charts 73 Status-Change-Status Charts • 74 Changes in Content of Data 74 Logarithmic Charts 75 Chart Style . . . 76 Chart Construction 78 Grid Proportions 78 Chart Size .. 79 Scales ..... 80 Scale Rulings . . 84 Time-Scale Designations and Rulings . 86 Selecti~g Scales for Bar Charts . 88 Breaking Bars and Columns .. . 89 Spacing Bars (or Columns) .. . 90 Connected Columns ..••... 91 Overlapping Columns (or Bars) . 91 Goals and Projections .• 92 Shading .......... •. 94 Selecting Curve Patterns . . . 97 Preprinted Parts in Chart Construction 98 Chart Captions . . . . . . . . . 99 Labels •............ 100 Arrows and Other Label Pointers . 101 Chart Lettering ........ . 102 Line Weight Standards and Lettering Pens 104 Chart Specification Sheets 105 5 . FORMAT .... 108 Report Forma:, 1:)8 Page Format . llO Binding Methods 115 6 . PRODUCTION TECHNIQUES l17 Organizing the Material l17 Preparing Copy . 117 Processing Copy 121 Assembling Copy 121 Proofreading 122 7 . DISPLAY PRESENTATION . 126 Display Methods . . . 127 Choosing Between Projections and Posters 127 Projection Displays . 128 Poster Iisplays .. 130 Lettering for Displays.. 130 Shading Displays . . . . 132 Color in DisFlays . . . . 133 Hints for Oral Presentations . 134 List of References 135 Index . . . . . . . 137 iii CHAPTER l PRESENTATION The presentation of statistical data requires careful planning and the thoughtful application of principles and procedures cover ing the design, preparation, and reproductioL of text , tables, and charts . Presentation is important because it communicates the results of the whole chain of statistical operations, including the collection, compilation, processing and analysis of data. The effort required for presentation is usuall y small in comparison to the effort required for the earlier operations . What appears on a few pieces cf paper is often the only useful product of thousands of dollars and man-hours spent in the various steps necessary to assemble the basic data . The value of this effort may be lost if the results are not presented effectively. Some material that is poorly presented can be understood if the viewer spends enough time and effort in studying it; but sometimes an entire statistical program has failed tecause the findings were not properly presented and management did not have the time , patience , or knowledge to interpret them correctly. Too often, statistical reports cannot be interpreted cor rectly unless they are explained orally by the persons responsible for their preparation. Good presentation greatly increases the probability that the statistical product will be understood clearly and accurately and will be used properly as a basis for making decisions . Before data are collected and compiled, decisions should be made as to the purposes they are intended to serve , the audience they are to reach, and the form in which they will ultimately be presented . Before the physical design of reported material is begun, decisions should be made as to how the data are going to be produced , reproduced , and distributed. Decisions as to which data to include, which to emphasize most strongly, which to present in detail and which in summary form, which to present in charts and which in tables, frequently require the use of analysis as well as presentation techniques . These decisions are influenced by the nature of the data, the scope , format, and purposes of the report, and the av~ilability of time , equipment, and personnel. Analysis and presentation may be regarded as separate problems, but in practice they can not be entirely separated . Analysis includes some presentation and presentation includes some analysis. Whenever possible, analyses should be presented and presentations should be designed to highlight the most significant conclusions and facts developed through the analysis process . An understanding of the techniques of presentation enables the statistician to prepare a more understandable and convincing analysis; but perfection in format , typography, and the general appearance of a report cannot substitute for sound analysis . QUALITIES OF GOOD PRESENTATION Many of the principles of good presentation are so obviously matters of simple common sense that it seems almost unnecessary to call attention to them, but their application sometimes is not a simple matter. To be fully effective , statistical material ~~st be presented in a for m that can be clearl y understood and readily used at the level which it is intended to serve . Statistical information for management should be presented in terms of management problems and interests . Management is generally not interested in statistical techniques or terminology. Terms such as "skewness" produce only impatient mystification in the mind of the average administrator, yet he may be much interested in the problems or the conditions that are indicated by a skewed distribution. On the other hand , ma~erial intended only for trained technicians should be designed to meet the needs and experience of that group . A particular style and method of presentation is effective only insofar as i~ performs a specific mission for a specific audience . It is the failure to recognize this point that produces one kind of poor presentation, and helps to create impatience with statistics and statistical reports . Statisticians have a tendency to make their presentations more technical and more involved than they should be when writing for non-technical audiences . This is partly a matter of selfindulgence and partly a failure to recognize the nature of the audiences and their limitations and needs . Simplicity is a matter not only of form but also of c~ncept. The pr esentation may be physically simple , yet not clear in meaning because the concept is obscure or involved . Of course , it is possible to present statistical information too simply, and in doing so to run the risk of offending the reader; but it is better to offend him in this respect than to present him with material he cannot understand at all or can fathom only at the cost of considerable time and effort. On the other hand, simplicity and clarity are different things ; and it is impossible to present anything too clearly. Providing the reader with all the information available on a subject is seldom good presentation. Presentation as well as analysis is a selective process and many of the results of analysis must often be sacrificed in order to focus the reader 1 s attention on the more salient facts and conclusions . The question to be asked before presenting any statistical material should be "What do we want to snow?" rather than "What information do we have that can be shown?" Having determined the central idea to be derived from the data, the prese~tation should be built around it. This principle applies whether a single chart or an entire report is being presented . PRESENTATION MAJOR PROBLEMS IN PRESENTATION Most of the problems of statistical presentation can be covered by two questions : 1. What i s the best way t o present the information? 2. How nearly can this be done with available people, equipment, and time? To answer these questions the statistical analyst should know how the information ought to be presented, and how t o get it presented . For the first step he must know something about how statistical information can be presented, what forms are best for what purposes, and when to use which. For the second step he must know something about what can and cannot be done with a typewriter, what skills and equipment are needed t o prepare a chart, and what kinds of copies can be obtained from available reproduction processes. It is important also that the analyst knows how t o write readable and informative text and t o de sign clear and useful charts and tables. He must also know how t o get these things produced and how to put them t ogether into a presentation that helps the r eader t o find out what he needs t o know as easily, as quicldy, and as completely as possible. TYPICAL ERRORS IN REPORT PRESENTATION Understandable t o the Writer but not the Reader This error may occur because of a number of faults : poor organization, foggy writing, use of technical jargon and unfamiliar abbr eviations , poorly planned or designed illustrations or tables, or incomplete explanation of what the data represent. The most frequent fault occurs because the writer forgets that the reader does not know as much about the subject as he does . Not Aimed at the Reader ' s Needs The information carried may be accurate, meaningful, and clear; yet not pointed t oward the r eader ' s inter est . Different r eaders have different r easons for wanting t o know about the same subject. Investors, employees, and customers have different interests in the mat erial contained in a company's annual report. Doctors and patients have different interests in a r eport on public medical programs. A report on a new product interests the sales manager and the plant manager for different reasons . Not Easy t o Look At It is a mistake t o think that the only important thing about a report is the accuracy and significance of its content . The reader should be presented with something that is easy to look at. This means proper or gani zation of the main components of the report, attractive layout of the pages, careful distribution of emphasis, and legible lettering. Many other design features should be given careful consideration . Not in a Convenient Form A report should be easy to handle, easy to use , and easy to store . Physical details such as size, shape , thickness, margins, and method of binding are important fact ors in good pr e sen tation . So are index tabs, covers, and col or ed separator sheets in some circumstances . No mat ter how valuable a r eport is or how clear and easy it is to understand, it is poorly presented i f it is awkward t o u se . Contains Too Much Material Some r eports have t oo many pages and some have t oo much material on the page . Even if the r eader is eager t o get the information , he will resent having t o work his way through numerous unessential pages or through badly overcrowded pages t o find what he is looking for . Presentation is a selective process : in order t o get the reader ' s attention and focus it on the most significant facts, it is often necessary to omit many interesting sidelights, related data, and supporting documents that are informative, and possibly useful, but not actually needed. MAJOR PROCEDURES IN PRESENTING DATA The process of presenting data in statistical reports gener ally consists of eight basic steps or procedures. Some of these procedures may be fixed by the established policies or r equirements or by decisions of higher authority. In any case, it is important t o remember that each of these procedures is a distinct operation -and although some are not done as separate steps but are combined with others -each one must be thought about individually. The statistical analyst should keep all of them in mind because each one affects the others t o some extent . 1 . Define the Objectives Supposedly no one undertakes the presentation of data unless he knows exactly what he is trying to do. Too often the objective is not defined specifically and a large amount of material is put on paper with only a general notion of the purpose it is to serve . It is important for both the analyst and the reader t o have the purpose of a report stated clearly. A carefully defined objective at the start may also prevent a l ot of revision. It is desirable for the analyst t o write out on a piece of paper exactly what is intended t o be communicated to a specific audience . This helps t o clarify the objective. If MEASURING PROGRESS is the primary purpose, care must be taken t o be sure that the data are presented in such a way that they actually do measure progress. The best general method of measuring progr ess is t o relate r esults or actual performance t o some program, standard or goal. Often, however, t he nature of the subject is such that no absolute goal can be established. For some subjects, such as health, attainment of the theoretical goal (in this case the absence of all illness) is practically impossible . MAJOR PROCEDURES IN PRESENTING DATA Whenever a goal or objective is either lacking or unattainable, other yardsticks of progress must be substituted . Three alternative methods of showing progress are: (a) comparison of the same subject at two or more periods of time; (b) comparison of similar or comparable subjects at the same time period; and (c) comparison of similar subjects at two or more time periods . Some Army statistical presentations serve groups with different interests. Such presentations may serve more than one purpose or reach more than one type of audience . If so, it may be necessary to make compromises that will make it more difficult for any one audience to use the material effectively or efficiently. 2 . Choose the Medium of Presentation Decide what physical means are to be used to transmit the information. It might be a published 8" x 10. 5" report; it might be an oral presentation with display charts or film pro jections; or it might even be a self-contained, sound , moving picture . (See AR 108-6, Mcti on Picture Production. ) This step is not necessarily step number 2 . It may turn out to be step number 3 or step number 4, or at some point it may be decided to use more than one medium . The physical means of presentation, however, often controls many of the other steps of presentation; and for that reason it should be determined as early as possible . The reproduced written report is the best method of communicating when the audience is large and dispersed and when the data are likely to be needed by many offices as reference material in the future . Examples are reports that summarize the strength and health of the Army. Oral briefings with poster display charts or film projections are frequently preferred by high-ranking officers who have little time to read . Occasionally, a long, written report prepared at the request of higher authority is condensed to a short, oral presentation to facilitate coordination with other staff agencies and approval by senior officers . 3 . Select the Material Many people who work with statist ics do not realize that selecting the material is a part of presentation. However, in the larger sense, presentation is an editorial job. The analyst must determine what information he would like to give the reader, and how much volume he can expect the reader to examine. He must also determine what specific information will best convey t he essentials of the story, and how much detailed data must be included to support or supplement the major conclusions . In recurring reports the selection of mate rial should be changed as events unfold . If a crisis develops in a specific area more emphasi s should be given to the men, money, and material s committed to that area. 4. Construct the Story The next step i s to work out the sequence of the mater ial and to decide what emphasis each item shoul d recei ve . Actually, this step and the preceding one are seldom performed separately, but are usually a joint operation. Many different impressions can be obtained from any given compilation of data, depending upon what order and emphasis are used and what interpretation is suggested . Sometimes the nature of the material will fix the sequence and suggest the emphasis . 5. Select the Form of Presentation This step also is seldom performed separately. Usually decisions as to what to show in text, tabular, or graphic form begin to develop early in the organization and analysis of the data . In handling mass data, it is often difficult to know just what form of presentation to use until summary tabulations and comparisons have been made , and these usually cannot be made until the analysis is nearing completion. Some types of data almost always suggest a particular form of presentation, but it is poor presentation to use a particular for m just because it happens to be tne easiest thing to do . It is important to remember that text, tables, and charts each have specific advantages and disadvantages, and that it makes a differ ence whether the data are presented in one form or another. Each form performs certain functions that the others cannot do as well . Some of the factors to be considered when choosing between these forms are discussed in more detail at the end of this chapter. 6 . Produce the Presentation The preparation of reproduced statistical reports, surveys, reviews, studies, or analyses -whether in the form of single sheets, bound books, or large displays -is a special type of publishing activity; and it is important to recognize it as such . After decisions have been made as to what to do and how to do it, the manufacturing problem must be faced . In the publishing business the product is made better and easier by setting up production aids such as typographic and style standards, standard page layout forms and specifications , and by otherwise organizing to do the work systematically and efficiently. Statistical reports and similar products can be produced in this same way, especially when this material is to be produced in large quantities, or in small quantities at frequent intervals . Most Army statistical presentation is restricted by the limitations of the typewriter. This is a poor excuse, however, for inferior presentation. This manual attempts to show how excellent results can be obtained by using only a typewriter, and even better results by combining the typewriter with other available means of putting information on paper . There is a great difference between getting the most out of a typewriter and getting less than the most . PRESENTATION 7 . Reproduce the Presentation. Most presentations are r eproduced in order to obtain more than one copy . Repr oduction problems take two f or ms : using the wr ong process , and using the r i ght process but using it poorly. An understanding of all available reproduction methods helps to avoid both. 8 . Transmit the Final Product . Transmittal is not usually a problem, or only a very small pr oblem . It is a step , however that should be considered carefully. Even in t~ansmitting a repor t through the mail, there is a problem of whether or not to send a letter of transmittal and if so, what to say in it. When transmittal of the inform~tion is a direct, f ace-to-face process , as in a br iefing or conference , the pr oblems are quite different, and frequently they are more complex. No matter how the information is transmitted, this is a separate step of presentation and one that should be considered as an individual problem. THE THREE FORMS OF PRESENTATION The statistical analyst can choose from three forms of presentation: l . Text 2 . Tables 3. Charts Each has its own advantages and its limitations . Reports that a r e all text, all tables , or all char ts ar e useful for limited purposes but for the reasons given below they frequently do not make the best use .of the material . All Text. This type of presentation is not suitable for most statistical reports because only a r elatively small number of figures can be included in text without confusing the r eader. The very nature of statist~cs leads to numerical comparisons, and , except for the simplest cases, these can be made better in char ts or tables than in text. All text reports may be satisfactory when the statistics are extremely few or f ragmentary, or when the data are not good enough to show in tabular or graphic form . All Tables . Although tabular material is the backbone of statistical r eporting, all-tabular reports ar e seldom readable . All tables and no charts or text make a dull report . Unless the reader knows exactly what to look for and wher e to find it, important comparisons and other facts ar e often missed, and much effort is wasted examining unimportant details . All-tabular r epor ts lack the connective and interpretive qualities of good text, and the interest and forcefulness of charts . 1/ 1/ In reports prepared mainly for reference that is, those intended to give detailed information about a long list of items -a large volume of tabular material is necessary. Reports of this type , however, should begin with a summary or analysis of the tabular data. All Charts . Because graphic presentation is attr active and forceful , ther e is a temptati on to pr epar e statistical reports containing nothing but char ts . Such reports are effective for two purposes : to provide a collection of "high spots" calling the reader 1 s attention to a few impor tant points, or to provide a set of reference char ts which, except for the current details, are already familiar to the reader . For reports serving any other purpose -in other words , in most cases -all-graphic presentation is not satisfactory because it lacks a smooth flow of thought , and because it leaves too many questions unanswered . Some of these shortcomings can be ove r come by using a "nar rative" styl e of chart presentation (in which the title is expanded into a statement concer ning the chart) . See "Chart Style ," page 76 . A graphic presentation given at a briefing or conference is not all graphic -but is accompanied by oral explanation. Text + Tables + Char ts . Reports generally are mor e usefUl if two for ms of presentation are used rather than one ; but to be fully effective, a combiLat ion of al l the thr ee f orms is usually needed . In general, use text to describe and explain the statistics, use charts for emphasis , and tables to reinforce both text and charts . Determining the proper balance is a matter of judgment : it depends partly on the purpose of the repor t and partly on the nature of the material . One general principle should always be followed : text, tables, and charts should be mutually s upporting instead of i ndependent . Charts or Tables . It is often a waste of space and of the reader 1 s attention, to show identical material in both graphic and tabular form, but it is highly desirable to have supporting tabular data or specific references to these data for all published charts . One way to solve this problem is to include both a tabl e and a chart but to give each a di f ferent arrangement. For example , a chart showing percentages might be accompanied by a table containing the absolute amounts for the same information; or the chart might show totals and the table a detailed breakdown of the totals . In thi s way , chart and table support one another but show information that is dif ferent enough to warrant the use of both. Choosing Between Text, Char ts, and Tables . It is unreasonable to ask the r eader to read in text what he can plainly see in the charts and tables . The text should give key figures , but it should be used to emphasize and expl ain what the cha~ts and tables mean . The following three chapters, "Text Presentation, 11 "Tabular Presentation, " and "Graphic Presentation" explain in more detail how and when to use each form and how to get the most out of them . The "Display Presentation11 chapter points out some of the special problems of planning, designing, and using statistical material to be viewed by a number of persons at more than the normal r eading distance . CHAPTER 2 TEXT PRESENTATION In most statistical publications the reader must be told a great many things tha~ cannot be shown in tables and charts. He musJ often be told why, when, what , and how. Sometimes he must be told the meaning of the charts and tables shown him. In an or al report, all this is done by the speaker. In a written report, it must be done by text. Text should be used : l. To introduce . 2. To interpret or explain. 3. To emphasize. 4. To connect. 5. To summarize . If none of these needs to be done , text is not needed . The aim in the preparation of text shoul d be t o present clearly, accurately, and briefl y the information needed to understand and interpret the st atistical dat a in t he charts and t ables . The text should summarize , emphasize , supplement, or amplify -but never duplicate in detail -the graphic and tabular material. HOW TO WRITE TEXT Written words differ from spoken words in several ways . Spoken words can be supplemented with gestures, variations in the expression on the speaker 's face , and changes in the t one and inflection of his voice. The speaker can watch t o see if the listene r seems t o understand what he is saying. Written words do not have any of these aids. Punctuation, spacing, changes in the size and style of type , help some; but it is harder to tell someone what you mean in written words than it is in spoken words . Useful suggestions for the composition of text are given in DA Pamphlet No . l -10, "Improve Your Writing. 11 This pamphlet also lists (in Appendix I) a number of reference books concerning better writing that are available in libraries and bookstor es. To write good text it is importa.J.t to use short, familiar , specific words and t o avoid needless words . It is also important t o avoid unnecessary repetition. Coherent, unified and well organized sentences and paragraphs are basic requirements for presenting material in a forceful, emphatic manner. In stati stical text the objective is clear expression of thought without a colloquial or slangy tone . Do not sacrifice meaning for brevity; if the longer word is needed for precision then do not hesitate t o use it. Use the active voice generally, the passive voice less often: this will result in shorter and stronger sentences . Emphasis Emphasis may be achieved in several ways . Among them are the following : l . Allocat e space t o the vari ous subjects and to the various items in a subject accor ding t o their importance . 2. Put the most important sentence (the topic sentence) in an emphatic spot in the paragraph -usually the first or last sentence . 3. Use subject headings (chapter, maj or subdivision, minor subdivision and minor subdivision headings) t o call attention t o the impor tant t opics covered. 4. Supportthe t ext with charts and t ables. 5. Use ALL CAPITALS or underscoring to emphasize key words, phrases, or clauses, but use them sparingl y . 6 . List important facts numerically. MECHANICS OF STYLE In writing text, questions frequently arise concerning punct uation, abbreviations, spelling, hyphenation, and various other mechanical aspects of style, since in English there are often several acceptable forms . Although these alternatives do exist, it is advisable t o follow one st yle or system consistently within a report or series of reports . References The STYLE MANUAL of the U. S. Government Printing Office pr ovides a wealth of guidance mat erial on the accepted rules of capitalization, spelling, compound words , punctuation, abbreviations and other standards of form and st~le . Its supulement, "Word Division," gives rules for and example s of line-end wordbreaking . The DICTIONP_RY OF UNITED STATES ARMY TERMS, Army Regulations 320-5, contains t erms of particular interest t o the Army. The DICTIONARY OF UNITED STATES MILITARY TERMS FOR JOINT USAGE (Joint Chief of staff Publication l) contains terms of specifi c significance t o t wo or more military departments . AR 320-50, AUTHORIZED ABBREVIATIONS AND BREVITY CODE , contains a compr ehensive list of abbreviations. Most good dictionaries also provide general guidance . Punctuation Proper punctuation is a "must" in text writing. Otherwise , sentences may not make sense or may have meani~gs not intended by the writer. Two controlling pr inciples are t o : 1. Use punctuation marks to clarify the text and to make reading easier and faster . Omit punctuation marks whi ch do not accomplish these purposes . 2 . Use the correct punctuation marks . Each type of mark (comma, apostrophe , colon, semicolon, period, etc . ) has specific uses . Incorrect punctuation is more confusing than too · little, or none at all . TEXT PRESENTATION Abbreviations Abbrevi~tions often perform a real service to the reader by cutting down his reading time . If used to excess, however, they make text look clumsy and undignified . If abbreviations are incorrectly used, or are not easily understood, they are confusing and make reading more difficult. It is generally best not to abbreviate; but if done, these rules should be followed : l. Abbreviate caFefully and sparingly. 2 . Abbreviate only if it is believed that most of the readers will be familiar with the abbreviations . It is now common practice to use capitalized initials for the names of government agencies and other organized bodies when the meaning is clear (FBI, TVA, NATO) , and f or well-known contractions of military terms (DOD, MOS , AWOL) . The reader should be t old the meanings of all but the most obvious abbreviations . For example, a report intended solely for Army personnel need not define ROTC , RA , or TAG ; however, if the r eport r eaches non-Army readers, definitions should be furnished : Reserve Officers Training Corps , Regular Army, The Adjutant General . One way to tell the reader is to write the meaning (or the abbreviation) in parenthesis the first time an abbreviation is used . Another way is to furnish the definitions in an introduction or in a separate glossary immediat ely following the introduction. Sometimes the two methods may be combined . The following standards apply to signs and symbols : Can be Used : Sign or Symbol In Text In Charts and Tables # (number) No No ¢ (cents) No No $ (dollars) Yes Yes & (and) No Yes w %(percent) Q/ No Yes w w Avoid when possible. Use only when space is exceedingly tight. Q/ Always spell as one word. NUMBERS IN TEXT When and How t o Use Figures In articles, stories, and news accounts, the rule is to spell out most numbers that take no more than four or five words . In statistical reports, the rule generally is that all numbers ten and above are put in figures because there are so many figures and because they are usually the focus of attention. This not only saves a large amount of space but also makes the numbers easier to locate and compare . Use figures for : 1 . Numbers that are ten and above . . 2 . Money amounts : (5 cents , $0 .05 ; $5 ' 5 dollars ; $5 , 000, 000, $5 million; 5 million dollars; $64. 81) . 3 . Dates (conventional: December 7, 1941 ; military: 7 December 1941). 4 . Time of day (conventional : 8 a .m., 3 :15p.m.; military: 0800 hours, 1515 hours) . 5 . Decimals (2 . 8, 0 . 25 , 9 . 05) . 6 . Serial or ~eference numbers (Chapter 8, page 12, paragraph 2, column 6 , line 58, Figure 5, Table 10, Chart 1) . 7. Measurements (8 inches, 2 feet , 7 tons, 5 miles, 1 percent, 2 x 41 s , 2 hours 4 seconds, 10 acres , 3 gall ons, 6 pounds). 8 . Unit modifiers (5 -day week, a 5 percent increase, 1/2-inch pipe) . 9 . Numbered United States Army military unit designations, (other than the numerals for U. S . armies, which are spelled out) and Roman numerals for Corps (3rd Armored Division, First Army, II Corps) . 10 . A group or series of numbers when any is 10 or more (9 majors and 16 captains reported; 8 trucks and 10 drivers were needed) . Spell out: 1 . Mast numbers less than ten within a sentence (these three compounds, a six-month forecast, nine changes) . 2 . Numbers that begi~ a sentence . Generally it is better to reword the sentence . 3 . Numbers of United States armies (Third United States Army) . 4 . Numbers used formally as in titles (Federal Statistical Directory, Nineteenth Edition, the Thirteen Original States) ; unusually dignified subjects , or names of centuries (twen tieth century) . Other rules to follow ar e : 1. Use co~as in numbers with more than three digits (8,296, $35 ,701). This rule does not apply to dates , serial numbers, page numbers, military time, or digits to the right of a decimal point. 11 011 2 . In decimal numbers , put a before the decimal point if there is no unit (0 .16) . 3. Do not use a decimal point and ciphers to express a money amount consisting only of dollar s unless the amount is in a series including at least one figure in dollars and cents . (He had a $1 , 000 exemption; His expenses were $5 . 25 :or food and $8 . 00 for rent; but, His ex penses were $5 for food and $8 for rent . ) . NUMBERS IN TEXT 4. Do not repeat a spelled out number in figures (five-trucks, not five (5) trucks). 5. Avoid sentence structure which puts figures immediately before or after a numerical segment of a date (Wrong: more than 60,000 1965 models had defects; Right: more than 60,000 of the 1965 models had defects; Wrong: on 31 March 1945, 7,260 items were in the inventory; Right : on 31 March 1945, the inventory t otaled 7,260 i terns). Simplifying Figures Figures are much easier for the reader to think about if they are simplified. The more figures there are the more important it is to simplify them. Three ways of doing this are shown below: l. Simplify individual figures by rounding . 2 . Simplify groups of figures by averaging . 3 . Simplify the relationship between figures by using percentages and other ratios . When and how to use each method depends on the problem. Sometimes all three can be used effectively. Simplifying by Rounding . Rounding is a method of getting rid of needless detail. The purpose of rounding is to focus the actention on an approximate magnitude . A figure of $40,000 is easier to grasp and remember than $39,992.54; and 95 percent is generally easier t o comprehend than 94.89 percent. Rounding also makes relationships easier to see . Comparing 39,892.54 with 19,946 . 27 is much simpler ifwe know that the first number is about twice the second . Rounding the figures to 40, 000 and 20,000 makes this clear at a glance. By-Products of Rounding. Rounding is useful for other reasons besides making figures simpler. Sometimes it helps to make the figures realistic . Unrounded figures may seem more accurate or exact than they are . Many figures are computed or estimated rather than obtained by counting or measuring: they appear to be exact only because they have been carried to several digits in computing . For example, if you divide 100 peopl e by three, you get an answer of 33.33+ people; but this does not mean that each third has 33 1/3 people. Many other computed figures are j~st as illogical as this one when carried out to too many digits. Rounding should not be done until all computations are completed. Rounding is almost always preferablewhen the fi~u~e is an estimate. An estimate is at best only an approximate calculation. An unrounded estimate seldom is as accurate as it seems. Large rounded figures are easier to think about if some of the zeroes are trimmed off. For example, the figure 10,000, 000 is better for thinking purposes if written 10 million, and 8,900,000,000 is much easier to handle if written 8. 9 billion. ARRANGEMENT OF TEXT Much of the clarity of report presentation depends on arranging the parts in a logical order and identifyi ng them so clearly that the organization of the material may be readily ap parent to the reader. The planning of headings and subheadings often reveals weaknesses in organization or content such as omissions, faulty sequences , and repetition or overlapping . Headings and Subheadings Headings indicate wh8re different subjects begin. They introduce the various divisions and subdivisions of the report . Headings on each level of subordination are treated alike and they differ in appearance from those on other levels . In typewritten reports , these similarities and differences are obtained by using all capital letters, capital and lower case letters, underscoring, and indentation, and by leaving blank lines above or below the heading. A number of acceptable systems of headings may be followed. The standard adopted in this pamphlet is simple and yet provides for enough degrees of subordination to meet the requirements of most reports . It is described below in the order of succeeding levels of subordination. CHAPTER HEADINGS These are capitals, larger and much heavier than the Elite capitals provided on standard typewriters, of the same type face as the running heads, centered without underscoring. Chapter headings should appear at the top of the page . They indicate the principal divisions or components of a publication . SECTION HEADINGS These are al so capitals somewhat larger than typewritten Elite capitals, placed flush with the left margin without underscoring . Leave one blank line below each heading and two above it, unless it is at the top of a page . They indicate major subdivisions of chapters. MAJOR SUBDIVISION HEADINGS These are standard Elite typewriter capitals, placed flush with the left margin of the column, without underscoring . Leave one blank line between the heading and the following paragraph, and two blank lines between the heading and text above. These headings indicate the major subdivisions within the subject matter covered by section headings . Short sections do not always need major subdivision headings . Minor Subdivision Heajings These are initial capitals for all words except articles, prepositions, and conjunctions, placed flush wich the left margin and underscored.* Do not leave a blank line between this heading and the following paragraph -begin the text on the next line . Leave only one blank line between the heading and any text above it. Jf All underscoring should be done with drafting pen and ink . TEXT PRESENTATION Minor Subdivision Component Headings . These are init ial capitals, indented five spaces f r om left margin of the column (t en spaces in onecolumn text) , and underscored. This heading is followed by a per iod, two bl ank spaces, and the f i rst words of the text (on t he same line as the heading if t here is r oom for them) . Leave one blank line bet ween the heading and the text above . KEY LEAD WORDS may be put in all capitals in the first sentence of a paragraph to give these words emphasi s . All capit als lat er in t he paragraph are also used as ATTENTION-GErTERS. Another alternative met hod of achieving a similar result is to underscore like this . Both all capitals and underscoring, when used with discretion add t o the variety and attractiveness of a report. The number of different levels of subordination needed depends on the length and complexity of the report. Short simple reports may need only t wo or three levels of subordination. It is desirable to separate consecutive headings by at least one paragraph of text . The words selected for headings should be carefully chosen and reviewed to make certain that they are concise, yet clear and consistent. Body of the Text The body of t he text is single spaced, when typed for reproduction. Preliminary manuscript is usually double spaced or typed in one of the t wo columns of typing layout sheets tomakeediting easier . Rules for paragraphs : l. Separate paragraphs by one blank line. (This may be omitted i f space is t i ght) . 2. Indent the first lines of paragraphs : five spaces (st art in the sixth typewriterspace). in two-column t ext, and ten spaces (start in the eleventh space) in one-column t e xt. 3 . Do not number paragraphs, but do indent and number listings, such as thi s , which are a part of a paragraph. The numbers are indent ed five spaces in two-column t ext and ten spaces in one -column text. Leaving a blank line between the items in the list ing is optional, depending on the number of items, their length, and the availability of space on the page . In listings containing long i t ems it is better to allow the blank line if possible. 4. When the text is continued on a second column or page , carry it over t o that column or page and leave at the bottom of the first column or page not less than two lines of a paragraph. Sometimes a little stretching or squeezing is needed t o make a column of text come out to just the right length without r ewrit ing. If it is a little t oo short, put an extra blank line between paragraphs or above major or minor subdivision headings. If it is a little too long for the page , first try t o continue the last paragraph ( or some of it) on the next column. Then , if any paragraph contains a numbered listing of items or if the column contains any major subdivision heading, consider deleting the blank lines between the items or reducing the blank spaces above the heading to one line . The appearance of a short listing with no item l onger than two or three lines may sometimes be improved by typing the second and third lines of each item flush with t he first (nor mally in the lOth typewriter space) . This procedure may also be followed in listings with longer items to f i ll surplus space when fitting text t o a page . Text footnotes are placed at the bottom of t he page . In two-column text, they appear beneath the column t o which they apply. They are separated from the rest of the text by a line 0 . 01 inch thick, drawn with a ruling pen and ink. This line is 20 typewriter spaces wide in one-column presentations and 10 wide in two-column arrangements. A blank typewriter line is l eft above the ruling, but none below: begin the first text f ootnote in the first line below the ruling. The standard reference symbol if there is only one t ext footnote on the page is t he asterisk (*) . Two or more text f ootnotes on the same page are identified by numbers in series, which are indicated as reference symbols by a diagonal and an underscore , 11. Other commonly used footnoting systems for for mal reports number the text footnotes consecutively from the beginning t o the end of each chapter or of the entire report or publication. In the body of the text put the reference symbol immediately after its point of r eference, with no blank space intervening if the symbol is an asterisk, but with a one-space separation if the symbol is a number . For example : • • • obligations* and expenditures• • • •• •procurement actions 11 for the year •• • The reference symbol should be placed after punctuation marks : - •• • of September enlistments .lf They were ••• •••peak workloads : 11 The estimate was ••• But not after a dash : • • • vehicles -combat only 1/-in the ••• Or a closing parenthesis if it refers only to the material inside the parenthesis: ••• index (based on 1965 pr i ces*) r ose ••• Some other important rules. t o rememberabout text footnotes are : l . Single space all footnotes : no extra space between lines of the same note or between notes . 2. start each footnote with the reference symbol at the left margin of the page in one column text and at the left margin of t he appropriate column in two-column text. 3. Begin the first word after a symbol in the fourth space from the left margin if the reference symbol is an asterisk and in the fifth space if it is a number . ARRANGEMENT OF TEXT EXAMPLES of text footnotes are given below: * On a page of two-column text, place a single footnote at the bottom of the colQ~ in which the reference occurs . * In one-column text the ruled line would be two times as wide (20 typewriter spaces) and the footnote would be a full page wide . Place more than one footnote in series at the bottom of a column . ·11 The first word of a footnote starts in the fifth space if the symbol is a number and in the f ourth space if it is an asterisk. 61 Use number s if each column on a two-column page has a single footnote : put footnote 11 at the bottom of the first column and footnote 61 at the bottom of the second column . JUSTIFYING When text is set up in printer's type , all t he lines are made to come out even at the right. This leaves an even margin on the right side of the page , j ust as on the left. When text is typed on a typewriter , however , the lL~es seldom come out evenly, leaving an irregular margin at the right. This is one of the things that gives typewritten reports an unprofessional look. Some of the unevenness can be avoided if the typist breaks words that would extend beyond the standard length of line (instead of carrying the whole word t o the next line). All of the lines can be made exactly the same length by spreading the short lines, and squeezing the long lines. Thi s process of fitting text lines t o a standard length line is called "Justifying". Many typing jobs should not be justified. For formal reproduced reports, justifying is often worth the extra effort; but for most typing, it usually is not. Short, informal reports and statistical text typed in l etter or memorandum form should not be justified. In many rush j obs there simply is not time t o do it, even if it should be done . Justifying on a Typewriter Spreading lines is not as easy for the typist as it is for the typesetter. The typesetter presses a button and the extra space is spread evenly throughout the line; the typist usually must insert the extra space in units of a whole typewriter space , which means that the typewritten line will appear to have breaks in it. In using a standard office typewriter , there is no way t o avoid these space breaks : the objective is t o make them as inconspicuous as possible . * * Some typewriters have devices for spreading the extra space evenly throughout the l ine. Some others have a special space bar that lets the typist insert half spaces instead of full spaces . Either of these methods is better than inserting full spaces; but until such a device becomes standard equipment on typewriters, space must be inserted in whole units. The two general rules for inserting extra spaces in a line are : (1) spread them as evenly as possible throughout the line and (2) place them where t hey 1-1i ll be least noticed; i.e., between sentences, following commas, or where a natural pause in the sentence occurs. If possible, place them where they will be spanned by a long word in the lines above and/or below . It is inadvisable t o leave more than one extra space between words ; and seldom feasible to distribute more than four or five extra space s t o a line of two-column text, or more than five or six extr a spaces to a line of text if it is one· column wide . Taking out space to squeeze a long line is more difficult than inserting spaces . The only easy places t o eliminate a whole space are after a period or a col on . A good typist, however, can usually squeeze two half-spaces out of a line, especi ally if it has a period or a :::olon in it. As a general rule, however, if the last word is only one letter too long, it should not be broken or carried t o the next line . If the line is more than one letter too l ong , do not squeeze it. Frequently the best solution t o a bad space problem is t o change one or more words or to rewrite a sentence . 1'hree steps in Justifying Three steps may be taken t o make typewritten lines come out evenly : l . The typist first fills out all short lines with asterisks t o show how many spaces must be buried in the line to make it come out evenly. The prot~uding letter (or punctuation mark) at the ends of lines that are only one space t oo long is all that is needed to show that those lines must be squeezed one space . 2. The author (or the typist) next marks this trial typing with "hash" marks t o show where the spaces should be insert ed (or taken out) . 3 . Finally, the typist r etypes the text in final form, inserting or squeezing spaces where marked . Paragraph Typed Without Justifying "We , the people of the United States, in order to form a more perfect union, establish justice, insure domestic tranquility, provide for the common defense, promote the general welfare , and secure the blessings of liberty t o ourselves and our posterity, do ordain and establish this Const~tution for the United States of America." Step 1 : Typed for Justifying "We , the people of the United States, in'f** order t o form a more perfect union , establish**'< justice, insure domestic tranquility, providelf*lf for the common defense, promote the general welfare , and secure tte blessings of liberty t o**** ourselves and our posterity, do ordain and es-** tablish this Constitution for the United Stateslf of America." TEXT PRESENTATION "We, the _people of the United states, /in*** order/to form/a more perfect union, /establish*** j ustice, /insure domestic/tranquility, /provide*** for the common defense, promote the general wel fare , /and secure/the blessings/of/liberty to**** ourselves/and our posterity, /do ordain and es-** tablish/this Constitution for t he United states~f of America. " step 3 : After Justifying "We, the people of the United states, in order to form a more perfect union, establish just ice, insure domestic tranquility, providefor the common defense, promote the general wel fare , and secure the blessings of liberty to ourselves and our posterity, do ordain and establish t his Constitution for the United States of America. " JustifYing is Easier Than it Looks Making typed lines come out even at the right takes more work, but it need not take much more . If everyone who has to work on the text knows at the start that it is going to be justi fied, each step in the preparation is made with this end in mind. The typist ' s work should be made as easy as possible by providing her with typing layout sheets (paper printed with guide lines that match the typewriter spacing) . (See "Typing Layout Sheets11 , page117) • W.ord Division It is frequent ly desirable to break the last words of certain lines in order to make the lines more nearly even. When lines are to be just ified, breaking words is even more important . If the typist does not break a word at the end of a line, but carries it over to the next line, she will have to add more spaces between words to bring the line out to the full width of the page or column. The Word Division supplement to the GPO style Manual referred to on the first page of this chapter was prepared primarily for printers but is intended to give guidance to typists as well. It gives extensive detailed rules for breaking words and is a quick-reference aid for finding correct word breaks. It lists words broken into syllables but does not define them. The 1962 edition measures 3" x 5" , contains 190 pagesand costs 50 cents . Some General Rules for Breaking Words are given below: l . Break words only between syllables . If in doubt, consult the GPO Word Division Booklet or a dictionary. 2. Do not break words of one syllable , regardless_of their length: straight, through, range, etc. 3. Do not break any word -short or long that would leave a single letter standing alone : eo ulectric pinion niforrnity 4. Do not break the last word in a column or page . 5. Do not break numbers with less than six digits. 6. Avoid breaking words of five or less letters : in-to, da-ta, on-ly, etc . 7. Avoid the division of the final word of the last full line of a paragraph. 8. Avoid word breaks at the ends of more than two consecutive lines . CHAPTER 3 TABULAR PRESENTATION The presentation of data in tabul~r form or in tables is a basic statistical device . A statistical table is a systematic arra~gement of numerical data i n columns and rows . Its purpose is to present quantitative facts clearly, concisely, and effectively. It should facilitate an understanding of the logical relationships among the numbers presented . Tables are used in the compilatioP of raw data, in the summarization and analys~s processes , and in the presentation of statist~cs in final form . This chapter deals mainly w~th tables designed for reproduced typewritten reports prepared for Army use . A good table is the product of hard work. It is not just a package of figures put into neat compartments and ruled to make it pretty. It contains carefully selected data, put together with thought and ingenuity to serve E specific purpose . Tables are the backbone of mos t statistical reports . They provide the basic sub~tance and the foundations upon which conclusior.s of the report are based . WHY USE TABLES Tables are used in Army reports for the following reasons : 1. Clarity -they present many items of data in an orderly and organized way. 2. Comprehension -they make it possible to compare many figures qu ickly. 3 . Expl icitness -they provide actual numbers whi ch document data presented in any text and charts • 4. Economy -they save space -and words . 5. Convenience -they offer easy and rapid ~ccess to desired items of data. The disadvantages of tables include the following : 1 . Uninviting -tables generally look like difficult reading matter that could be ignored . 2 . Undramati c -significant relationships are often hard to find and comprehend . 3 . Specific -tabular data , because of their formal appearance, tend to make crude estimates and projections seem more precise than they actuall7 are . Much of the art of tabular presentation lies in the ability to maXliDlZe the advantages and minimize the disadvantages of this method . Sometimes other methods of presentation are better: 1 . Charts are often more effec~ive for showing simple relationships, trends and general magnitudes . 2 . When only a few key figures are to be presented it i s often desirable to show them only in the text. This ~s especially true if the figures lad: comparability and need explanation. BASIC PRINCIPLES The underlying principles of tabular presentation should be understood in order to design effective tables . Each table is more or less "custom made . " The following statements i ndicate some of the essential considerations to be oorne in mind : Determine the need for tabular presentation and its appropriateness on the basis of number of figures to be presented, compl exity of relationships between the figures , the requiremen~s for precise data, and the use of alternative methods of presentation. Defi ne the purpose of each tabl e in specific terms . This wil l ai d in establishi ng the scope of the material to be presented and in determining how many tables are requir ed . Each table must have a clear and definite purpose . Pl an the content of each table i n accordance with its specific purpose . This is the key to sel ection of t he i tems , thei r l ogi cal classificati on , and determination of t he proper size of groupings . Determine the extent to which rati os , percentages, averages or othe r measures wil l sharpen the focus of the tabl e on signi ficant facts and relationships . Examine critically, every element planned for inclusi on tJ be sure that it contributes to the accomplishment of the prime purpose of the table . Desi gn the table so that i t will be easy to read and understand, and pleasing to the eye . 1 . Make all titl es, units of measure , headings , stubs and notes concise and unambiguous . 2 . Di stribute categories hori zontall y and ver t i cally to produce properl y bal anced overall physical proportions . 3 . Use optimum spacing, ruli ng and indenti on for quick and certai n recognition of major and minor groups . 4 . Carefully plan the sequences and arrangements of i tems and total s to present a rational progressi on and to emphasize significant relationships . Time sequence data sho·.lld be from top to bottom or from left to right . 5. Place totals (which are generally of pri mary interest) at the top and to the left where they lvill be seen first . (See "Placement of Totals" , page 28 and accompanying examples . ) Monitor production of the table to insure that : 1 . The planned design has been followed . 2. Percentages, averages , rati os and any other computed data are accurate . 3 . Figures have been verified by machine check and other components carefully proofread . 4. Appropriate type face and clean, sharp unbroken type have been used . 5 . The table is properly balanced on the page . 6 . The ruling and spacing contribute to the ease of reading and reference . TABULAR PRESENTATION TABULAR FORMS When designing tables careful consideration should be given to the purpose that the table is intended to serve, despite the fact that distinctions between purposes are not always clear . Three distinct types are available to meet different objectives : 1. Text tabulati ons (or text tables) . 2 . Special purpose (summary or analytical) tables . 3. General purpos e (or reference) tables . Text Tabulations are small, simple, informal tables inserted in the text . They are merely listings of data without titles. They are actually part of the text and they must be placed where their subject is discussed , immediately after a colon. They are usually but not always indented from both margins as shown in examples on this and the following page . The purpose of these tables is to highlight relationships among a relatively small number of figures. They are intended to be read from beginning to end and should be limited to two or three columns. They provide as much clarity as other tabl es without as much interruption of the text . Text tabulations are more appr opriate for statements that are primarily narrative than for reports containing a considerable amount of statistical detail . Special Purpose (Summary or Analytical) Tables show selected data arranged to focus attention on important comparisons and relationships. They are meant to be read from start to finish . They are usually shown with text, and frequently are placed in the summary. If they are simple enough , they can be used as inserts on charts or on maps. Gener al Purpose (Reference) Tables give comprehensive coverage of the subject covered by a report. In tables of this type the data are arranged so that any needed figure is as easy t o find as possible . They are not meant for complete reading. They are "figure files" that the user refers to in about the same way he would look up a word in the dictionary. Reference tables usually appear at the back of a report . They are backup materi al to be used if detail is needed to supplement data summarized in analytical tables or charts . The distinctions between special purpose and general purpose tables are not precise: a special purpose table in one report might be a reference table in another , depending upon how the reader is expected to use it. The distinctions usually are not applicable to reports having only a few pages . Both special and general purpose tables are formal tables . A formal table is as complete and self-explanatory as possible. It usually does not depend upon the text -but has its own title and other explanatory information needed t o make its content clearly understood. ARRANGING FIGURES FOR EMPHASIS The location of figures within a table determines to a considerable extent the emphasis they will receive in the mind of the reader . In cesigning a table , therefore , it is important t o consider carefully which comparisons are of the greatest significance and then arrange the figures so that these comparisons receive the most emphasis . In tabular presentation, as elsewhere , the closer together things are the easier it is t o compare or contrast them. The most effective way t o bring two figures close together f or visual comparison is t o place one above the other. In this position the comparison may be made at a single glance . This is illustrated by the f ollowing example of a text tabulation: Location July Fort A • • 30, 250 Fort B 22,368 Placing the two figures side by side perrrits a somewhat less effective comparison: Month Fort A Fort B July • • 30,250 22,368 The presentation of two i terns of informati on for each fort necessitates additional choices. Cne way of doing this is shown below: Location and Month Number Fort A July . . . . . . . . . 30,250 August . . . . . . . . 27,942 Fort B July . . . . ... . . . 22,368 August . . . . . . . . 25,715 This makes it necessary for the eyes to skip a figure in order to compare figures for the same month . But this arrangement may be used to em phasize comparisons of what happened from July to August at Fort A with what happened at Fort B. This could be done more directly , however , by ~dding figures for actual or percent changes . In another arrangement shown below the eyes must move even farther to make the comparis on: Camp A Camp B Item July August July August :Number 30, 250 27,942 22 ,368 25,715 The same general principles can be applied t o a series of figures . A list of f igures in a column is easier t o scan and t o compare than the same figures listed acr oss the table on the same ~ine . In the vertical arrangement it is much easier t o l ocate the largest and smallest figures, partly because any differences in the numbers of digits may be seen at a glance . It is also easier t o l ook at a co lumn of f irst digits ARRANGING FIGURES FOR EMPHASIS and quickly see the differences in the general size of the numbers . An example of a single column is shown below: Year Number Jan •• • 109,246 Feb • • • 140,621 Mar 89,73 2 Apr • • • • • • • 6, 509 May • • • • • • • • • 25 , 824 This quick scanning is harder to do when figures are on the same line : Item Jan Feb Mar Apr May Number 109,246 140,621 89, 732 6, 509 25 , 824 The location of figures must often be influenced by other considerations. For example, the stub items or column headings sometimes must be put in a standard order that does not bring the most significant figures close together. Also , some figures may require comparison with so many others that some of them must be too far away for rapid comparison . It is important t o remember two basic principles : l . Numbers close together are easiest to compare , 2. Vertical comparisons can be made more rapidly and effectively than horizontal comparisons. COMMON FAULTS IN STATISTICAL TABLES A table that is not properly designed is sometimes worse than no table at all. A check list of common faults , any one of which ca~ spoil a table that is well designed otherwise , appears below: l . Wr ong emphasis -a summary table not arranged t o focus attention on the most significant comparisons; or a reference table not designed for ease of locating individual figures . 2. Poor internal construction -lack of clear, logical, orderly relationships . This includes the lack of leaders and columns so wide that it is hard to associate figures and stub entries. 3 . Not complete and self-sufficient leaves unanswered important questions such as scope , content, time, etc . 4, Cluttered with extraneous data-figures required t o compute t otals, percentages or ratios, but not necessary for the primary purpose of the table . 5. Lettering hard t o read -too small or poorly typed or reproduced. 6 . Badly fitted t o the space available t oo cramped , t oo spread, or poorly proportioned, 7. Vague, ambiguous or incomplete column headings or stub nomenclature -failure t o describe the data clearly. BASIC TABULAR ARRANGEMENTS In tabular presentation simplicity of design is just as important for quick, clear comIunication as it is in graphic present ation, All t oo often , however , it is necessary to use various more or less complex tabular arrangements t o show desired detail and relationships among :::n;:mbers. The designer of tables should give consideration t o three basic aspects of tabular arr angement: l . Position on page: upright or broadside , 2 , Pages required : single page or multipage , 3. Internal structure : treatment of stub and boxhead . Alternative choices concerning each type of Jabular arrangement are discussed and illus"Lrated below. The terms used in the discussion of tabular presentation to designate the component parts of a table are illustrated on the following page , vith the exception of field spanners which are illustrated and discussed in detail on pages 14 a~d 19-37. POSITION ON PAGE Upright Table The usual way to present a table is with the typing across the narrow dimension of the page in the same manner as the text. All of the tabular examples in this pamphlet are of this type . Broadside Table All too frequently, wide tables are typed across the broad side of the page -at right ~gles t o the text and t o the tables on other pages in the report. The report must then be turned sideways t o read the tables, making it di:ficult for the reader to go back and forth f rom analysis to reference , A broadside table can usually be redesigned to an upright type . Often its width can be reduced by reversing the position of the material in the stub and the boxhead , or otherwise changing its internal structure . The table might also be divided and shown upright on two facing pages as a double -page spread table , which is described below . Another method of avoiding a broadside table (not involving redesign) is to force the table into the upright position by reducing it in reproduction more than the rest of the report . Ttis decreases t he legibility of the table , and iLcreases the technical difficulties of planning and preparing the pages, as well as reproducing the final copy. When a broadside table is used, it should be arranged so that its top is on the left side of the page , TABULAR PRESENTATION TERMS USED IN DISCUSSING TABULAR PRESENTATION ~Subtitle TENTH ARMY FY 19X AL FUNDING PROGRAM AND Stub OBLIGATIONS, Y I NSTALLATION As of 31 December 19Xl Stub Headnote"""" Heading ~ (Thousands of dollars),-------' ,_________ _ , ' Installation f!/ 107,785 55,018 51.0 95,843 48,797 50.9 Fort Allen 20,612 10,265 49.8 6,519 Line ~~~~ Gates Montgomery • 10,942 44,939 5,526 23,413 50.5 52.1 2,702 ll,l23 Fort Schuyler • • • • • • 5,148 2,502 48.6 1,320 Fort Sullivan • 4,765 2,425 50.9 1,405 Leaders Camp Clark 2,375 1,131 47.6 719 525 Camp p Greene Putnam Stark Ward • • c/ 967 -4,406 1,125 564 493 2,040 623 379 51.0 46.3 55.4 67.2 253 1,321 265 :1851 ·_ _ J 221 1,045 237 r}_/ CLASS II ll '942 6,221 52.1 Buchanan Army Depot • 661 364 55.1 161 36 Dearborn Arsenal • • • • 1,534 822 53.6 375 3 Reference Funston Army Terminal 1,221 600 49.1 354 267 Symbol Hull Missile Plant 2,242 1,258 56.1 502 482 Pierce General Hospital • 1,262 557 44.1 417 288 after Proving Ground 1,196 636 53.2 296 264 Ta or Armory • 1,715 885 51.6 475 355 General Warr Ordnance Works • • 1,291 611 47.3 361 319 820 488 59.5 184 148 ~---O-th_e_r___ 1_----------------~--------~------~------~------~----~ NOTE: This table presents for the first time the status of funds for '1 Reference each installation for all appropriations for which funds are allocated 1 ~toTenth Army. i ~/ Data t or each installation include funds for subinstallations. : £/ Funds increased $11 million from 30 November 19Xl. 1 1 Source £/ An increase of $700,000 has been requested of higher headquarters , for activatien of two !lnfant!ry battalions. )d/ Less than $500. Installation to be inactivated. 1if Includes three District Engineer offices and Wheeler Army Depot. : / 1SOURCE: Status of Allotments and Operating Accounts, RCS ARAC0-14. Column Heading Cell Ruling .Footnotes TABLE: The term used to designate the entire tabular presentation. BODY: The 11 tabular 11 part of the table; excludes the title, headnote, and footnotes. FIELD: The area within the body of the table in which (excludes the Stub and Boxhead ) . the figures are entered BASIC . TABULAR ARRANGEMENTS PAGES REQUIRED The space occupied by a table may vary from less than one page to many pages. Tables requiring one page or less are called single-page tables . Those taking up more room are termed multipage tables . Single-Page Table Sometimes a table is a little too l ong , or a little too wide, or both, to fit in the space available on a page . This problem usually may be solved by changing the internal structure of the table or a careful pruning to make it small er . For example, blank space may be eliminated or abbreviations may be used to make the table fit. (See "Unusual Spacing Problems," page 40 . ) Multipage Table When the boxhead of a table fits across a single page, but the stub is much too long, the stub can be continued on one or more additional pages . The title and column headings are re peated on each page of an upright table and on the left-hand pages of a broadside table. Titles for multipage tables are discussed on page 22 . Double-Page Spread Table When an upright table is much too wide for the page width, it can often be placed on two facing pages with about half of the column headings on each page . If there is enough room, repeat the entire stub on the right side of the right-hand page . Otherwise , use line numbers in lieu of the second stub. If the stub captions are numerous, continue the stub on one or more sets of facing pages and repeat the title and column headings on each set. The title should not be centered across the two facing pages, but should appear separately on each page. Double-page spread tables should be avoided as a general rule but they are always preferable to extra-sized "fold-in" tables, which are not recommended under any circumstances . Fold-ins are expensiv~ they tear easily, and they are exceedingly hard to read with speed and accuracy. INTERNAL STRUCTURE Most tables fit into one of three primary patterns as determined by the organization of the stub and the boxhead: 1 . Single stub and single boxhead . 2. Divided stub and repeated boxhead. 3. Repeated stub and divided boxhead. Two secondary forms using field spanners are derived from these basic types, as is another form consisting of two or more subtables. These six primary and secondary types are discussed below and are illustrated on the pages that follow . The primary considerations in selecting the appropriate type of table are the data relation 214-853 0 -66 -2 ships which the table is intended to display and emphasize . The available space on a page and the space r equirements of a table are often major factors in selecting the type of table to be used . The table may be long and narrow and the available space short and wide, or vice versa. Sometimes the table may be too wide and too long to fit the page in either direction. Usually, these conflicts can be resolved by using a structural type which can be adjusted to fit the space. For example , a long, narrow table can be shortened to fit the available height on a page by dividing the stub and repeating the boxhead to produce a shorter, broader table . Similarly, a short, broad table can be narrowed to fit the available width on a page by dividing the boxhead and repeating the stub to produce a longer, narrower table. Single Stub and Single Boxhead Table This tabular pattern has a single stub under a single stub heading at the left of the table and a single boxhead across the top of the table. The stub applies across the entire width of the field and the boxhead down the entire depth of the field . The table may have both a stub and boxhead which are simple (no repeating segments), as in Table lA on page 16 . But it may have a compound stub (repeated line captions under different group captions), a compound boxhead (.repeatedcolumn headings under different spanner headings), or both as in Table lB. The single stub and single boxhead table is the most commonly used and the easiest to prepare and understand. Characteristics of data and inadequate space on the page for the table are the major reasons for selecting some other type . Before changing from this basic pattern, determine whether the table can be placed in the available space by reversing the positions of the stub and the boxhead. For exampl e , a table with more stub captions than column headings can be turned around and made shorter and wider; or a table too wide for the width of the page but with a short stub can be made longer and narrower . Divided Stub and Repeated Boxhead Table In this type of table, the stub is divided into two or more parallel columns with the stub heading repeated above the second and any subsequent columns, and the boxhead across the top of the table repeated for each division of the field preceding the next stub division, and each column heading applies down the entire depth of each field . Either or both the divided stub and the repeated boxhead may be simple or compound . Each is simple in Table 2A, and compound in Table 2B, on the adjacent page. A single stub and single boxhead table is a candidate for conversion to the divided stub and repeated boxhead structural typ~ if it meets TABULAR PRESENTATION SINGLE STUB AND SINGLE BOXHEAD TABLE SINGLE STUB AND SINGLE BOXHEAD TABLE With Simple Stub and Boxhead @ With Compound St ub and Boxhead Reen-May June Enlist-Indue-In5tal Month Total list ments tions l ation ments Received Adopted Received Adopted TOTAL . 158,.364 5.3,205 .32, 862 72,297 TOTAL 1,.392 109 1,488 19.3 July . . . 10, 096 .3,428 1,954 4, -4 CIVILIAN 894 75 925 1.31 Augu5t . . 9,.324 .3,170 1,772 4,.382 Camp X 8i 1.3 75 8 September . Camp Y Fort A October . . Fort B November . Fort C December . Other January • . MILITARY February . Camp X March . . . Camp Y Fort A April . . . Fort B May. . . . Fort C June . . . Other DIVIDED STUB A..'JD REPEATED BOXHEAD TABLE@ With Simple Stub and Boxhead Justi-Justi- Type of Complaint Total Type of Compl aint Tot al fied fied TOTAL . . . . . . . . . . . 1,065 354 Gr ade . . . . . . . . . . . . 21 7 Indebtedness . . . . . . . . . . 9 4 Adminis tration . . . . . . . . . . 115 60 Insurance . . . . . . . . . . . 4 0 All egations -Report of Conditions 158 35 Leave and Pa ss . . . . . . . . . 30 9 Allotments . . . . . . . . . . . . Marriage . . . . . . . . . . . . Army Re serve Matters . . . . . . . Me ss and Food . . . . . . . . . Boards . . . . . . . . . . . . . . Pay and Al lowance s . . . . . . . Denial of Permission t o See IG . . Post Exchange . . . . . . . . . Disciplinary Matters . . . . . . . Pr operty . . . . . . . . . . . . Discr·imination . . . . . . . . . Quarter s . . . . . . . . . . . . Duties . . . . . . . . . . . . . . Trans:fers . . . . . . . . . . . Ef fici ency and Charact er Ratings . School s . . . . . . . . . . . . Enlistments . . . . . . . . . . . Service s and Supply . . . . . . Enlistment Promises . . . . . . . All Other . . . . . . . . . . . DIVIDED STUB AND REPEATED BOXHEAD TABLE @ With Compound Stub and Boxhead May June May June Grade in which Grade in whi ch Serving Serving RA Other RA Other RA Other RA Other MALE . . . . 19 ,895 49,.372 20, 846 40 , 851 FEMALE . 1,547 .3,467 1,486 2,722 General Officer • .301 1.3 298 12 Gener al Officer 0 0 0 0 Col onel • . . . . 2,904 247 2,564 204 Col onel • . . . 3 1 .3 0 Lt. Col onel • . . 5,590 2,518 5,565 2 ,124 Lt . Col onel • . .34 4 34 2 Major • . . . . . Major • . . . . Captain • . . . . Captain • . . . lst Lieutenant . lst Lieutenant 2nd Lieutenant . 2nd Lieutenant BASIC TABULAR ARRANGEMENTS REPEAT ED STUB AND DIVIDED BOXHEAD TABLE @ Wit h Compound Stu b and Boxhead United St a t esTot al Tot al Depart mental Service End of Mo nth Total Veteran~! Total Vet erans Total Veterans Empl oy -Per-Empl oy-Per-Empl oy-Per - Number Number Number ees cent ee s cent ee s cent l9Xl March . . . . . . . . . . 601 , 450 228,61 7 38. 0 471, 745 217,485 46.1 1 6 ,773 5, 081 30.3 June . . . . . . . . . . 570,29 5 220 ,111 38 . 6 451, 870 209 , 218 46 .3 16,472 4 , 971 30 .2 September • . . . . . . . 552,829 213,793 38 . 7 432 , 988 203, 268 46 . 9 1 6 , 085 4 , 90 7 30 .5 De cember . . . . . . . . l 9X2 March • . . . . . . . . . J une . . . . . . . . . . Sep t ember . . . . . . . . December . . . . . . . . United States (Conclu ded ) Outside United St ates Field Servi ce End of Month Total Veterans Total Veterans Employees Number Percent Employe es Number Per cent l 9Xl ~rch . . . . . . . . . 454,972 21 2 , 404 46 .7 129 , 70 5 11,132 8. 6 J une . . . . . . . . . . 435 , 398 204 , 247 46 . 9 118 , 425 1 0 , 893 9 . 2 September . . . . . . . 416,903 198, 361 47 . 6 11 9 , 841 1 0 ,525 8 . 8 December . . . . . . . . l 9X2 -"-"March • . . . . . . . . . June . . . . . . . . . . September . . . . . . . December . . . . . . . . REPEATED STUB AND DIVIDED BOXHEAD TABLE REPEAT ED STUB AND SI NGLE BOXHEAD TABLE @ With Simple Stub and Boxhead fi:\ WITH FIELD SPANNERS ~Wi th Simpl e Stub and Compound Boxhead Number of Units State Strength June JulyAllotted Activated Accident Percent Percent Cat egor y Number of Number of TOTAL . . 718 696 77 , 688 Total Total Ar izon:a-. . 32 32 4,.192 Cal ifornia . 338 329 37,231 Acc i de nt s Idaho . . . Mo ntana . . . Nevada . . . TOTAL . . 2, 509 100 . 0 1 , 897 100 . 0 Oregon • . . . Utah . . . . Arrrry Aircraft 6 0 . 2 3 0 . 2 Washington • . Motor Vehi cle _ ,,322 52 . 7 55. 0 h2M Arrrry . . . ::.. , 048 41 . 8 812 42 . 8 Non-Arrrry Advisor Personnel Arrrry Fire . . Strength State Ot her . . . Of fi-En-per Total cers list ed Advisor Injuries TarAL . . 414 183 231 188 Arizona . . 19 10 9 221 TOTAL • . 1,594 100 . 0 1,1.37 100 . 0 Cal ifornia • . 193 85 108 193 Idaho . . . . Arrrry Aircraft 0 o. o 0 0 . 0 Montana . . . Motor Vehicl e ____4lQ 25.7 _2E 29 . 2 Nev ada . . . . Arrrry . . 51 .3. 2 40 .3.5 Oregon • . . Non-Army . Utah • . . . . Arrrry Fire . . Washington . . Ot her . . . TABULAR PRESENTATION REPEATED STUB AND REPEATED EOXHEAD TABLE WITH FIELD SPANNERS With Simple Stub and Boxhead Army Depot FY l9Xl FY l9X2 FY l9X3 FY l9X4 FY l9Xl FY l 9X2 FY l9X3 FY l 9X4 Rece:.ved Total Processed TOTAL . . . . . . . 516 ,773 649,926 441,122 311,842 508,569 657 , 535 438,240 311,515 Atlanta • Columbus Letterkenny . . . . . . . . . . . 29 ,361 69 ,945 44,411 99 ,545 24 ,516 59 , 431 20 , 818 36,079 28 , 586 69 ,540 44 , 723 99 ,769 24,379 59,440 21 , 255 35, 090 Memphis . . . Red River . . . . . . Sharpe . . . . . Schenectady . . Utah . . . . . . . Processed in Les s Than Four Days On Hand at End of Fiscal Year -TOTAL . 457 , 930 569 , 074 398,704 270,322 9, 913 2,304 5,186 5,513 Atlanta . . . . . . 24 ,561 33,987 19,708 20 ,129 775 463 600 163Columbus . . . . . 68 , 180 97,746 56 , 860 31,275 428 204 195 1,184 Letterkenny . . . . . . Memphis . . . . . . Red River . . . . . . . Sharpe . . . . . . . Schenectady . . . . Utah . . . . . . smGLE BOXHEAD SUBTABLES MUTIPLE BOXHEAD SUBTABLES With Compound Stu b and Simple Boxhead With Simple Stub and Boxhead @ UNITED STATES @ ARMY MILITARY PERSONNEL Component and Type Comm Warrant 30 June of Personnel Total Male Female Total Officers Officers Enlisted TOTAL . 617 , 402 608 , 140 9 , 262 l9Xl 817 , 430 79 , 943 8 , 448 729,039 l9X2 829,678 81 , 126 8 , 637 739,915 Regular Army . . 277,831 270 1495 7,33 6 l9X3 873 ,345 82 , 607 9 ,274 781,464 Commiss i oned Officers 14,015 12,999 1,016 l9X4 Warrant Officers . Enlisted Personnel . l9X5Other Com:eonents l9X6 Commissioned Officers l9X7Warrant Officers . . l9X8 Enlisted Pers onnel . P.RMY CIVILIAN PERSONNEL OUTSIDE UNITED STATES 30 June Total Salaried Wage Board TOTAL .•. 362 , 863 359,597 3,.266 l9Xl 272 ,477 144,412 128,065 Regular Army 272 , 807 271 , 044 l9X2 291 ,508 151,584. 139 , 924Commiss i oned Off i cers . 8 , 839 8,328 l 9X3 323, 018 164 ,739 15 8 ,279 Warrant Officers l9X4 Enlisted Personnel Other Component~ l9X5Commissioned Officers l 9X6 Warrant Officers l9X7 Enlisted Personnel l9X8 BASIC TABULAR ARRANGEMENTS certain conditions . These include the following: l. Too long for the available space and there is enough width to divide the stub and repeat the boxhead. 2. Awkward in appearance because of a com bination of many stub captions and few column headings . Tables with a narrow stub and one or two data columns will often fit in a half page width even after dividing the stub into two columns . Those with more column headings or long headings or stub captions generally require a full page . Repeated Stub and Divided Boxhead Table This arrangement puts the data on two or more decks or levels. The entire stub, including the stub heading, is repeated on each deck, and the boxhead is divided with a portion on each deck. The stub applies across the width of the field, but the po~tion of the ~oxhead on each deck affects only the entrie~ on that level. The bottom horizontal ruling of each level is drawn across the entire width of the table, and becomes the upper horizontal ruling of the succeeding deck.• Both the stub and the boxhead may be simple (Table 3B on page 17), or compound (Tabl e 3A), or one may be simple and the other compound . Tables that ara too wide for the available space may be rearranged into longer and narrower tables of this general type. For exampl e , short tables that are more than a full page wide may often be converted into longer tables that will fit the width of the page. A double-page spread table or two or more smaller, separate tables is often preferable to this complex type . Repeated Stub and Single Boxhead Table The data for this table form are also shown on two or more decks . The stub except the stub heading -is repeated on each level. The boxhead is common to all levels and appears only once at the top of the table . The stub and the boxhead may be simple or compound ; as illustrated in Table 4 on page 17, the stub is simple and the boxhead compound . Above each deck is a field spanner qualifying the columns of figures on that level. The first one is placed just below the boxhead . Horizontal rulings drawn the full width of the field inclose each field spanner. Tables that are short and wide may be converted to those with a repeated stub and a single boxhead . The same result can be obtained by changing field spanners to stub group captions . This avoids cluttering the field of the table with captions . Field spanners are practical, however, for sharp emphasis on the major groupings, or when the number of subcategories of data is so large that the stub otherwise would be complex and hard to follow. A further discussion of field spanners appears on page 31 and additional examples are on page 33. Repeated Stub and Repeated Boxhead Table This arrangement is similar to the preced ing type except that the boxhead is repeated across the top and each level has the same num ber of field spanners (usually two or three) as the boxhead has repetitions (Table 5 on page 18) . Its use is recommended when there is insufficient room for: l . A longer stub column by conversion to a single stub and single b~xhead table (i. e . by shifting the field spanners to the stub and repeating the line captions under each in a single elongated stub column) . 2 . Additional stub ~olumns by conversion to a divided stub and repeated boxhead table (i. e . , shifting the field spanners to the stub, repeating the l~ne captions under each, and dividing the stub into two or more columns) . 3 . Repeating the column headings of the boxhead under each field spanner by conversion to a repeated stub and divided boxhead table (i . e ., changing the field spanners to spanner headings and repeating the column headings under each). Subtables The basic and secondary tabular patterns are sometimes changed into a hybrid table form with the data grouped into two or more separate subtables under a common title . Stub group captions, boxhead spanner headings, or field spanners are converted to subtitles which are placed on the second li~e above the top horizontal ruling of the subtables they describe . A table-wide horizontal ruling is drawn be neath each subtable . This ruling is separated from the subtitle below it by one or two blank typewriter lines . The standard is one line for half-page-wide tables and two lines for fullpage-wide tables (this may be reduced to one if the space is "tight"). The boxhead is shown only at the top of the first subtable when it is applicable to all of the subtables (Table 6A on page 18) . Otherwise , each subtable has its own boxhead (Table 6B) . The stub is repeated for each subtable, but the s t ub heading is repeated only if each subtable has a separate boxhead. Subtables focus attention on major groupings of related data. Do not use them when such emphasis is not needed . COMPONENT PARTS OF A TABLE The terms used in this chapter to designate the component parts of a table are illustrated on page 14. The main parts of a table are : the title, stub, boxhead, field, and notes. These are discussed in some detail in the following sections . Line numbers and column numbers are not shown in the illustration, since they are used onl y in large detailed tables or when references to specific entries are anticipated. The principal components of tables are discussed and illustrated on the pages that follow. TABULAR PRESENTATION THE TITLE Evo~;· formal table should have a title stating briefly the nature of the data presented. Thi s should be a concise, explicit statement covering the following : l . Subject -the title should indicate the basic content of the table . Examples : Military Strength Civilian Empl oyees Procurement Deliveries Depot Receipts 2 . Coverage -it should indicate whether the data cover the whole subject (e .g., Strength of the Army) or only a portiJn of it (Fifth Army Strength) . It is to be assumed that an entire universe is covered unless some kind of a limitation is indicated . 3 . Cross-classification -the title should indicate how the data are cl assified and cross-classified. Only major crossclassifications should be indicated in the titl e , with detailed classificat i ons indicated in the stub and in the boxhead . The cross-classification usually follows the subject and coverage segments of the title, and is preceded by the word "By . " The cross-classification is usually expressed in the s~ngular rather than the plural form . Examples : By Branch and Grade By Type of Vehicle By Month and Type of Personnel 4 . Time -if applicable, the title should indicate the date(s) or time period(s) to which the data refer. This may be a specific date(s) or time period (s) , selected dates, or inclusive period(s) of time . Examples : 30 June l9X6 30 June and 30 September l9X6 30 June l9X4 and l 9X5 Fiscal Year l9X7 Calendar Year l9X2 Selected Dates, l9X5 , l9X6, ~nd l9X7 June l9X4 -July l9X6 The table title thus serves two functions : it indicates the contents of the table anc at the same time it serves as a reference guide . A single title may be used when a short, simple statement is sufficient to describe the data in the table or the title may consist of a IDEin title and a subtitle, to simplify an otherwise long, compl ex title or to indicate variations in content of a series of related tables . Wording of Titles Titles should be brief, but neither clarity nor accuracy should be sacrificed for brevity. The wording of titles should be in topical form (telegraphic style) rather than sentence form . Verbs are usually omitted as well as articles or other parts of speech not needed to under stand the title . Some other ways of obtaining brevity are to make the terminology concise and to use subti~les, headnotes (page 34), and foot notes (page 35) . The contents of the table should never be ~ess than tnat described by the title, but the table can contain more detail than the title ~ndicates . Generally, a short, less complete ~itle which may be read easily is preferred to a l ong , complete title which may not be read thor oughly because of length or complexity. It is often desirable to rely on footnotes or expl ana t ory notes for complete statements of content and coverage. Abbreviations should be used sparingly in ~itles and only when they are standard and widely accepted . Terminology should be standard and consistent among tabl es within the same report, and among related reports . Phrases like "Table Showing," "Number of," and "Distribution of" should be omitted from titles when the meaning is clear without them. Single or main titles which requir e two or (rarely) more lines, should be broken at logical uoints in order to make them easier to r ead and Understand . The division of words should be avoided . Emphasis in Titles Emphasis should be placed on the principal feature that distinguishes a table from other tables . Sometimes this can be done by placing the most important word or phrase first . When tables concerning entirely different subjects appear in the same volume , the subject should be emphasized . In related tables, the emphasis should be placed on the distinguishing feature, whether it is the main subject or not . Subtitles Subtitles are used to amplify the main title ·,rhen necessary. They are particularly useful to indicate the time to which the data refer and to jelineate differences in the content of related tables by indicating such things as the command, area, or population group covered . Dates in Title s Dates may be omitted from titles if they are the same for all of the tables in a report or section of a report, and the time is clearly identified on the cover of the report or on the title page or running head of the section. If the dates vary among the tables of a report or a section of the report, the appropriate date(s) should be included in the title of each table . When dates are used in titles, they usuallyshould be placed after that portion of the title to which they apply. If they apply to the en tire table, they should be placed at the end of the title or in a subtitle . When different dates apply to different portions of the title, each segment of the title should be followed by the applicable date(s) . Specify exact dates in the table title or subtitle only where it is impractical to show them elsewhere, or where the statistics might o~herwise be misinterpreted . The precise time specification can often be indicated logically COMPONENT PARTS OF A TABLE in the stub heading or in a spanner heading of the boxhead . For example, the shorter title "STRENGTH OF THE ARMY : 1910 -1925" is preferred to "STRENGTH OF THE ARMY : 30 June , 1910 -1925" when the month and day can be placed in the stub heading . However , use "STRENGTH OF THE ARMY : 30 June l9X4" instead of "STRENGTH OF THE ARMY : l9X4," or "STRENGTH OF THE ARMY : June l9X4, 11 since it would be awkward to show the precise "as of" date elsewhere and misleading unless the exact date is clearly indicated . When a table presents calendar year data, the time specification should general ly consist of the year designation(s) only. When fiscal year data are indicated, the year designation(s) should be preceded by the words "Fiscal Year . " Do not use "Fiscal Year l9X4" if "30 June l9X4" is meant . Unless the day and month are indicated in tables pertaining to numbers of personnel for example , doubt arises as to whether the data cover manyears or fiscal year-end strength . In reports where both fiscal and calendar year data appear, it is important to describe in figures in order to avoid confusion. When time series data cover a continuous period, the inclusive dates should be shown in the title, e . g . , April l9X4 -December l9X6 (but if the data are for the end of each month , this fact can be shown in the stub or spanner heading, as appropriate) . When time series data omit some of the intervening regular periods or dates , or cover irregular dates or periods, the dates or periods may be preceded by a phrase such as "For Selected Periods , " "For Selected Months , " "For Specified Years," "For Specified Dates," or "For Selected Years, 1900 -1980" ) . Use of such terms "During" and "Between" with dates should generally be avoided unless there is some special reason for including them. Percentages, Rates , and Ratios in Titles Frequently statistical tables are comprised of or contain percentages, rates, and ratios to facilitate analysis of data . When a table consists solely of such measures , that fact should be indicated in the title . When a table contains both absolute and relative numbers, the content of the title depends on the purpose of the table : l . If the primary reason for the table is to highlight the relative figures , and the absolute numbers are of secondary importance, limit the title to a description of the relatives . 2 . If the absolute numbers are of primary significance, and the relatives are of secondary interest, omit reference to the relatives . 3. If they are of equal importance for the use and interpretation of the data, indicate in the title that both the absolute and relative numbers are contained in the table . Percentages the title should make clear whether each sub-group is expressed as a per ent of the total population group covered by the table, or whether only one or a few of the subgroups have been selected to be expressed as a percent of the total population. Example: OFFICERS AS A PERCENT OF TOTAL MILITARY STRENGTH Rates -some of the ~escriptive informat ion about rates may be included in a title or subtitle . Example : ADMISSION RATES TO HOSPITALS AND QUARTERS Admissi ons per 1 , 000 Average Strength per Year The complete information which should be covered somewhere in the report is specified in Army Regulations 325-10, Standar d Computation of Rates , Par agraph 8 . Someti mes this i nformation is so vol uminous that it should be present ed in headnotes , footnotes , boxheads , or a combinati on of these . Ratios -usually the title should omit the phrase "Ratio of.'' Instead, it should cl early indicate which group is being expressed in relation to the other. Examples : NUMBER OF TRAINEES PER TRAINER NUMBER OF CIV:LIAN EMPLOYEES PER SUPERVISOR In a~dition, as indicated above for rates , the title may indicate the size of any standard unit of popul ation adopted . Placement and Capitalization of Titles Typewritten titles should be placed at the top· of the table, and centered across the width of the table . Main titles containing two lines should be presented in an inverted pyramid form . Subtitles should generally be centered directly ~~der the main title. Normally one blank line should be left between the title and the upper horizontal ruling of the boxhead . Main titles should be typed in ALL CAPITAL LETTERS . Subtitles should be typed in I nitial Capitals only. Punctuation in Ti tles The following punctuation rules should be followed in table titles to obtain clarity: Comma -use to separ~te three or more subjects, coverage indicators , cross-classifications, or dates, with tte last in each series preceded by "and. " Also use them to separate the subject and coverage segments from the crossclassification segment when they appear on the same line . In multiline titles, do not use a comma at the end of a line when the following line begins with a cross-classification or when it is ~ subtitle . Colon -use to precede the time indicator when it appears at the end of the title . Do not use a colon when time is indicated as a subtitle . Semi -colon use between series of dates when dates within one series are separated by commas . Hyphen -use to separate dates when the entire period shown by the dates is covered in the table, or when data for each date or time period covered by the ~ime indicator are shown in the table . Use the word "and" in lieu of a hyphen to separate two dates or time periods, and three non-consecutive dates or time periods . TABULAR PRESENTATION Examples are given below: 7 December 1941 -31 August 1945 Fiscal Years 1965 -1975 April and May 1942 April, June, and November 1917 Titles for Multipage Tables When an upright table covers more than one page , each page should contain the full title, and any accompanying headnote if applicable . The indication "(Continued)" should be placed at the end of the title on all but the first and last pages; the last page should have "(Concluded) " at the end of the title . When a broadside table takes more than one page , the title, modified as explained above with 11 (Continued)" and 11 (Concluded)" , should appear on the first page, on each succeeding left-hand page, and the last page . Table Numbers in Titles Table numbers are not an integral part of a formal table and therefore are not a "must" for statistical tables . This is particularly true for reports containing only one table or a very few tables . When a report contains many tables, however , a numbering system is generally used to facilitate references and identifications in the body of the report and in a list of tables. When numbers are used in titles, the standard practice i s to precede the number with the word 11 Table11 and place them to the l eft of the title as follows: TABLE 8 -NEW OBLIGATIONAL AUTHORITY FOR MILITARY FUNCTIONS Several methods may be followed when making assignments of table numbers or letters in a statistical report . Either numerals , or letters or some combination of both of these may be used . A system that is good for one report may not be good for another . The choice depends upon the size and the nature of the report, the number of tables, ar:d the complexity of the subject matter . Approved methods are: 1 . Use Arabic numerals throughout the re~ port, numbering each formal table in sequence, starting with the first summary or anal ytical table and ending with the last reference table . 2 . Use capital letters for summary or analytical tables and Arabic numerals for reference tables, when these two table types are cl early distinguished. 3 . Use capital letter prefixes with Arabic numerals where a large report contains several chapters or sections on different subjects . The single capital l etter prefixes may consist of a letter, or two l etters keyed to the subject matters . Examples : A-1, A-2 •• , B-1 •.••, etc.; MP-1 for the first table on military personnel, CP-1 for the first on Civilian Personnel and PD-l for the first on Procurement Deliveries. Roman numerals are not recommenied since they are easily misread and it is difficult to refer to them. THE STUB The list of captions at the l eft side of a table is called the Stub. These captions de scribe each row of figures . Where only a part of the stub is listed on one page and it is continued on the next page, the continuation is also placed on the left of that page . A table reqUlrlng two facing pages to show all data columns applicable to the stub should have a duplicate stub on the right side of the right-hand page . Line numbers may be used in lieu of the second stub if space precludes this listing on the right-hand page . This type of table is called a 11 double-llage spread. 11 Every stub should have a heading. The stub heading is used to describe, define, or amplify stub captions . Care should be used in determining this heading. Examples·: 30 June, Month-end; Calendar Year, Fiscal Year; Destination, Origin; Grade in which Serving, Permanent Grade ; Category and Location; Category anj Status; Army and Army Area . Designing a Stub A poorly designed stub is hard to understand and use . Extra care spent on the design of the stub sometimes makes the difference be tween a good table and a poor one. To design a good stub, carefully organize the items selected to be included in it and show this organization clearly. Organization means putting the parts in a useful order and grouping them in a logical way to emphasize the kinds of comparisons which are appropriate for the data, and its expected use . Organizing the Stub The first step in arranging a stub is to decide on the order of the items. The order is important because it helps the reader to make significant comparisons and it also helps him to find individual figures quickly. Order is not necessarily the same thing as classification. Data classified geographically such as command strength, by state need not be in geographical order: the order could be alphabetical, in order of size, etc. Sometimes the kind of classification fixes the order -as in data classified by time of occurrence -but not always. Useful Arrangements of Items in the Stub The following list explains some of the reasons why one order may be better than another. The first five are mainly arrangements for convenience; and the second five are primarily arr~ngements for analysis . (See illustrations page 25 . ) 1. ALPHABETICAL -in order of the names of the items (dictionary order, telephone lists, etc . ) . 2. NUMERICAL -the order in which the stub items are numbered (program numbers, unit numbers , Armies , etc. ) . COMPONENT PARTS OF A TABLE (The numerical order as well as the alphabetical order are frequently used and development , pre-production, curement, maintenance, etc . ) . pro when it is important to locate indivi dual items quickly. These arrangements are particularly useful for long stubs . This is good for showing how the status at a starting point changed to reach 3 . QUANTITATIVE -the order of the size of the items (l/4-ton, 3/4-ton, l l/2-ton, etc.; under 6 years of age, 7-8, 9-10 , etc .; 40-rnm, 57-mm, 75-mm, etc . ) . the status at an ending point . It is also useful for showing how many items have reached each stage of progress towards completion. The order of smallest to largest item (or largest to smallest) . This is often a customary or convenient way to list things. 9. IMPORTANCE -The order of rank or grade (top third in performance , middle third, bottom third; Col, Lt Col, Maj , Capt , lst Lt, 2nd Lt ; etc . ) . 4· GEOGRAPHICAL -the order fixed by geographical location (Maine, New Hamp This is useful for comparing rank, seniority, results, quality or performance. shire, Vermont, etc .; Continental U.S . Outside Continental U.S.; etc . ) . SometLLes , putting a list in one order puts it in another too: order of size This can be used simply because it is customary and so helps the reader to find items quickly. I~ is very usef~l to show how things differ from one location to another . This sequence is best when the location names can be put could be the order of importance and also the customary order ; order of location could also be the progressive order. When this happens , the question of which order to use is much easier to decide. in an order that suggests their actual Two or three different kinds of order geographic arrangement . Within major geographical areas , however , individual states or countries can be listed alphabetically f or convenience . may be used in the same stub . For example, the main order of stub items in a strength table might be geographical, with each geographical unit divided to 5. CUSTOMARY -the order that has come to show a major unit arrangement and each of these units divided to show a :::ustom be accepted (salaried and wage commissioned officers, warrant cers , enlisted personnel; etc . ) . board; offi ary (officer and enlisted) listing. Organizing a Corrplex Stub Many stubs are not just a straight list of This is generally used for short lists of items only. Like the four preced items . Organizing the stub is seldom that simple and easy. ing arrangements , it is useful when it is important not to take a new point of view . Some stub lists have two or three main item groups with many items under each; conversely, some have many main groups with two or three 6. CHRONOLOGICAL -the order of the dates at which the items occurred (Apr, May, items under each. Usually, these are simple patterns that are fairly easy to work out. Jun; lst quarter, 2nd quarter, 3rd quarter; etc . ) . This may be used to show growth , trend, or change . It is one of the most valuable analytical arrangements . The difficult ones are the stubs that have an irregular pattern: For example, a long list of major groups ; some with a few minor groups , some with many , and some with none . When a list like this has two or three levels of captions , t he problem becomes more difficult. 7. FREQUENCY -the order of how many there are of each item. This is not the same arrangement as shown in (3) above which There is no simple way of designing plex stub; but these rules will help: a com depicts order of size of items . Size measures how big the items are ; frequency measures how many, or how much there is of each item. 1 . Make the major divisions clear: let the reader see at a glance what the main component s are . To illustrate : stub listing of vehicles in the order 1/4-ton, 3/4ton, and 1 l /2ton would be by order of size of 2 . Set off major items : whether single items or groups o: items, those at the first level should stand out clearly. items . However, if we had 200 of the 1/4ton, 50 of the 3/4ton, and 500 of the 1 1/2-ton vehicles and wished to list them in a frequency arrangement, we would list the 1 l/2-ton, 1/4 ton, 3. Group related things : items that are related, even though some may be major , and some minor, should be set apart from other items . and 3/4ton vehicles in that order. 4. Make subordinate levels look subordi 8. PROGRESSIVE -the order of development of growth (feasibility study, research nate: each step of subordination should be easy to see and to identify. 23 TABULAR PRESENTATION 5. Split long lists for easy reading : a list of items all at the same level of importance is hard to read unless split into groups. 6 . Use clear , accurate wording : · make it clear to the reader exactly what each item means (what is included in it and what is not) . This applies especially to t otals and sub-totals. Some of the standard stub arrangements on page 25 illustrate these rules. Other tables in this chapter show possible solutions to several other stub problems. A special "double stub11 arrangement can be used when the stub contains both captions and units of measurement (see 11 Units of Measurement in the Boxhead," page 31) . This arrangement is of value when all or most of the lines require separate and distinct units of measurement . The two parts of the stub are usually separated by a vertical ruling as in the example below. Uem Unit ;ran Feb Mar Acid, Nitric . 1-Lb. Bottle Boots, Jungle Pair . . . . Concertina Wire, Barbed Roll . . . . Crossties . . Thous. Pieces Cup , Canteen . Thousand Lwnber . . . . Thous . Bd. Ft. Repellent Insecticide Gallon . . . Soap , Soft . . 25-Lb. Drum • Making the Organization Clear In a standard typewritten table four things may be done to make the organization of the stub clear : capitalize, indent, space, and underline. Other things such as boldface type, larger or smaller type sizes, and italics are not found on standard typewriters, and therefore relationships must be established without them. Using Capital Letters in the Stub Three degrees of capitalization can be made with a typewriter: ALL CAPITALS: EVERY LEI'TER A CAPITAL LETTER Initial Capitals : Every word except Prepositions , Articles, ·and Conj unctions start with a Capital Letter . This is also called Capitals and Lower Case . Normal Text Capitalization: Use no capitals except for the first letter of the first word and of all proper nouns. ALL CAPITALS should be used only for totals, and for captions of major groups. Initial Capitals should be used for captions of minor groups or for items under maj or groups when there are no minor groups. They may also be used for subtotals in short, simple tabbles, or where each major group has only a few items under it. Normal Text Capitalization should be used only for items under minor groups. The use of lower case only is generally limited to an explanation in parentheses that accompanies a caption. (This is rare : a footnote is often better) . The illustrations on this page and on other pages in this section show how to apply these rules. In addition they show how the rules may be modified t o meet special problems without diminishing the clarity of the table. Indention Use as much indention as needed to set off each step of subordination -but no mor e . Generally, two typewriter spaces for each step is enough . The standards of indention which are est ablished in this pamphlet are: TOTALS -indented four spaces start in the fifth space. MAJOR ITEMS -no indention (all the way out to the left edge of the table) . SUBORDINATE ITEMS -Indented two spaces start in the third space (each step of subordination is indented two more spaces) . Note : If a caption runs to two lines, the second line is normally indented two spaces more than the first ; four spaces if followed by subordinate items which are indented two spaces more than the first line. Sometimes the stub classification is so unusual that this system will.not work. For example, the line captions may be longer than the group captions, in which case the order of indention should be reversed to save space . The illustration on page 31 shows an example of this problem and how to handle it. Vertical Spacing Additional space is the standard device for setting o~~ the main components o~ the stub. The normal spacing (up and down) is : ONE BLANK LINE (double spacing) between : totals, and the items they summarize; two major items; or a major item and its component parts. NO BLANK LINE (single spacing) between : the items in a group, or a list of items not grouped under major headings. A long list of items, however, may be split by using 11 space11 breaks (one blank line) between each 3, 4, 5, or sometimes more items, depending on the nwnber of items, and the amount of space available. Changing the Normal Stub Spacing Normal spacing should be used in all tables when space is not a limiting factor; when space is tight, or too plentiful, normal spacing must be shrunk or stretched. Where possible, try to use the same spacing scheme for a series of tabl es with the same stub or similar stubs. COMPONENT PARTS OF A TABLE ~This stub lists offi ~cer grades in the customary order of highest first (also the order of importance) . Since there are no group headings , no indention is used except for the total. No spacebreaks (blank lines) are needed between each group of three items in a short list like this, but may be used if desired. ~A typical arrangement \.V for showing time captions in the stub. It follows the standard order of earliest date first . Note that yearly data are separated by blank lines , monthly data typed "solid'' ; also that monthly captions are indented two steps in from the outside edge . f?\ An illustration of \.V items put in an analytical arrangement (like things grouped together) but listed for ~onvenience (alphabetical order within each group) • The blank line between groups is the standard method of separation and should be used when there is plenty of room . (.;'\ Part of a long stub ~in which items first are classified by location then by nature of the violation. Second classification is shown in order of frequency. "Space breaks" could be used in the list of violations but are not actually needed in such a short stub. fn\ Two stubs having the ~same captions -but arranged differently. The first is an analytical arrangement (in order of the quantity af each item) ; the second is a convenience or customary arrangement (in order of the size of the item) . The first form is better for short analytical tables; the second , f or long J,ists in which ease of locating an item is most important . TOTAL General • Colonel •• Lt Colonel Major • • • Captain ••• lst Lieutenant 2nd Lieutenant Warrant Officer FISCAL YE.AR 19Xl FISCAL YEAR 19X2 FISCAL YEAR 19X3 1st Quarter • July August September ~ TOTAL • • Fort Allen . . . . Fort Gates . . . . Fort Montgomery Fort Schuyler • Camp Clark Camp Greene • . Camp Putnam • Others ~ VIOLATIONS TENTH ARMY Drunk • • Disorderly Pass Offenses • • Off Limits Vehicle Offenses Uniform Offenses Others TOTAL Trucks • • • _222. 2 1/2-Ton • 430 1/4-Ton • • 315 3/4-Ton • 220 TOTAL Trucks • • • 965 1/4-Ton • • 3153/4-Ton • • 220 2 1/2-Ton • 430 i'":i\ T~i~ short . stub is ~similar to ~he one at the left excep~ ~hat of ficer s have been grouped under grade headings . The grouping is emphasized by retaining the no~mal space between groups , but substituting an underline for the space normally used to separate the group head ings and the items under them. fi\ A simple a~angement \:J for ~howing status of supply. Note that all of t he items are given equal emphasis and that there is a blank line after each one . Changes during the period receive the same emphasis as the status at its beginning and end . fC\ Part of a long stub ~contains ~wo types of classifications (by fort and by type of personnel) . In this arraP~enent the first compariscn is between military and civilian at each fort. Note use of capitals and two-space indention for the two subtotals under the grand total. (;\ This stub stows three ~levels of ~ubordination. No~e tte ~ndention for each level aLd the use of underlining. When i tern:; in the stub are underlined all corresponding entries in the field should also be underlined. {.;;'\ An example c·f a fair ~ly complex ~tub that has been shortened as much as possible . Underlines are used to em~hasize the total and the subtotals (see page 28) . Note that "Army of the U. S." and "Retired Reserve " are not underlined, but that "National Guard," as a minor group .is underlined. TOTAL • General •• Field Grade Colonel • Lt Col onel Major • • •• Company Grade • Captain ••• On Hand 1 April Received in April Shipped in April On Hand 30 April TOTAL • • MIL;ITAB.Y CIVILIAN Fort Able • Military Civili an Fort Baker Military Civilian TOTAL AREA COVERED . Warehouse . Occupied Vacant Non-storage Shed Occupied Vacant Non-storage ~ TOTAL . . . Ac tiveArmy . . Regular Army Army Reserve National Guard Army of the U. S. ,ivilian Components Army Reserve Ready Reserve . Standby Reserve Retired Reserve National Guard Civilian Employees Salaried Wage Board TABULAR PRESENTATION Lengthened Two steps §) Lengthened One Step ---------------------,· /~ I TOTAL • 13, 456 TOTAL MAJOR GROUP • ll, 797 MAJOR GROUP Minor Group 10, 306 Minor Group Item • 1 Item Item Item 23 Item • Item • Item •• 456 Item • Item 7, 890 Item Item • Item 12 Item • Item 345 Minor Group Item 678 Item • Item 901 Item • Item • Item • Minor Group 1, 391 Item 23 Minor Group Item 456 MAJOR GROUP Item •• 789 Item 123 MAJOR GROUP Item • Minor Group 100 Item •• MAJOR GROUP 45 MAJOR GROUP MAJOR GROUP 1 , 588 Item • 678 Item • 910 MAJOR GROUP 26 \ §) See footnote on next page. ADJUSTING THE LENGTH OF A STUB TO FIT THE SPACE 13, 456 ll,797 10,306 1 23 456 7, 890 12 345 678 901 1, 391 23 456 789 123 100 45 1,588 678 910 26 I I NORMAL SPACING TOTAL •• MAJOR GROUP Minor Group Item •• Item Item • • • • • Item • Item Item • Item Item • Minor Group Item Item • Item Item Minor Group MAJOR GROUP MAJOR GROUP Item ••• Item ••• MAJOR GROUP 13 , 456 ll, 797 10, 306 1 23 456 7, 890 12 345 678 901 1,391 23 456 789 123 100 45 1, 588 678 910 26 \ COMPONENT PARTS OF A TABLE Shortened One Step Shortened Two Steps Shortened Three Steps TOTAL . 13 , 456 TOTAL . . 13, 456 TOTAL .. 13, 456 MAJOR GROUP 11, 79lMAJOR GROUP 11, 797 MAJOR GROUP ll, 797 Minor Group 10, 30 Minor Group 10, 306 Item 1Minor Group 10, 306 Item • . . . . 1 Item • . . .. 23 I t em -----"1 Item • . . . . 23 It em 456 I t em 23 Item • . . . . 456 It em 7, 890Item • 456 Item • . . . . 7, 890 Item 12I t em • 7, 890 Item • ... . 12 It em 345 Item • 345 I t em • 678 Item • .... 12 Item 678 I t em • 901 Item 345 Item 901 Minor Group 1, 391I t em • 678 Minor Group Item • . . . . 23 I t em 901 1, 3~~ Item • . . . . Item 456 Item • . . 456 Item 789Minor Group 1, 391 Item • . . 789 Item • . . . . 123 I t em • . ~ Item • . . 123 Minor Grou:e 100Item • 456 Miu.or G):'OUE 100 MAJOR GROUP 45 Item • 789 MAJOR GROUP 1, 58~ Item • ... . 123 MAJOR GROUP 45 Item •• . 7 Item • . . 910 MAJOR GROUP 1, 588 MAJOR GROUP 26Minor Group 100 Item 678 Item • . 910 MAJOR GROUP 45 MAJOR GROUP 26 MAJOR GROUP Item l, g~~ Item • . 910 MAJOR GROUP 26 WHERE TO Ptrr BLANK LINFS IN THE STUB Lengthened Stub Short ened Stub Normal Two One Spacing v ne Two OThree Steps §) St ep St ep Steps Steps 1 1 1 Above first stu b caption and below last 1 1 1 2 2 1 Below TOTAL Q/ . . . .. . .. .. . . 1 1 0 2 2 1 Between MAJOR groups . . . .. . . . . l l 0 2 2 1 Between MAJOR and minor groups . . . . 1 0 0 2 2 l Bet ween minor groups . . . . . . . .. 1 0 0 l 1 1 Between group headings and their items 0 0 0 1 1 1 Between every four items s/ . . . . . . 1 0 0 1 0 0 Between every item .. .. . . . .. 0 0 0 §) The stub can be made still l onger by doubling all normal spac ing (putting two blank lines where the nor mal spacing calls for one , and one where it calls for none) ; but t his third step of stretch ing makes t he t able look so empty that it should be used only as a last r esort. About the only reason for this much stretching is t o fill out a short page in the middle of a l ong seri es of full pages . £I Above , if total is at the bottom. £I Every three in short lists ; five or more in l ong lists . TABULAR PRESENTATION Sometimes a table may be shrunk or stretched for emphasis. Within reason, the size of a table should be tailored to give it appropriate emphasis . The illustrations on pages 26 and 27 show normal spacing and some procedures for shrinking or stretching the stub. These changes, however, do not alter the basic system of emphasis and grouping. Underlines Underlining should be used in typewritten tabl es much the same way that bol dface type is used in printed tables , to give emphasis to subtotals and totals . It is not necessary to underscore all subtotals and totals, particularly in small tabl es with adequate space between lines and for indentations or in unusual ly compl ex tables with many subtotals . Each tabl e should be examined to determine whether or not underlining will make it easier to use and understand . When captions are underlined, the figures in the field should al so be underlined. Wording in the Stub Use captions that are both descriptive and brief. This means to use mostl y nouns, and to omit as many articles , prepositions and conjunctions as possibl e without impairing clarity. "Total" means the total of all the subordinate items shown in the tabl e . "Total " used as an adjective (total strength, total man hours , etc . ) means the total of all there are-whether i ts components are shown in the table or not. Often ambiguities in the content of total s can be eliminated by substi tuting "all" for "total " , such as "all appropriations" or "all programs . " The use of "aggregate" in lieu of "total" generally should be avoided . Abbr eviations may be used in the stub when space is limited, but only those which may be l ocated in authoritative military or civilian sources ; such as Hq Co; Inf Div, Mech; Mil Sq Ft etc . Punctuation in the Stub Use no more punctuation than needed to make the meaning clear. Some of the rules of punctuation used in text are not followed in stub capt i ons . For'example : Periods a r e not used at the ends of captions , and may be omitted from abbreviations if the meaning is clear wi thout them . Periods may also be omitted from initials used as shortened names for countries , governmental agencies, and other organized bodies and for some other well establ ished abbreviations such as USSR, USAR, NATO, POW , WAC , and FY . Do not use apostrophes to show missing letters in words abbreviated by contracting . Use of Leaders Leaders (rows of periods in every other typing space) are used to connect each caption with the first data column . They are lined up vertical ly, working back from the second space before the first vertical ruling, to a point in the second or third space, as appropriate , after the last letter of each caption. In an informal table with no vertical rulings, at l east two blank spaces should separate the leaders from the widest figure in the first data column. No l eaders should be used in the field of the table. Do not use leaders (l) in narrow stubs, (2) in wide stubs when the captions end uniformly near the first vertical r uling, (3) in the duplicate stub at the far side of the right hand page of a table on facing pages, or (4) for captions that do not have any field entries opposite them. Always use leaders in wide stubs wi th short captions or those of irregular length. Line Numbers Line numbers are seldom necessary in singl e page tables . They may be used, however , at the left side of the stub when convenience of reference makes them desirable . When line numbering is required, all stub captions which identify rows of figures should be numbered consecutively. The numbers are separated from the stub by a vertical ruling. In the page at the right side of a table on faci ng pages, the ruling is between the last column of figures and the column of line numbers . If there is not enough room for the second stub of a table on facing pages , use line numbers instead, on the right side of the second page and match these with line numbers at the left side of the stub on the first page . The use of line numbers is illustrated in the table under "Lettered Columns ," on page 31. Placement of Totals Total lines are placed above the subordin ate numbers of which they are the aggregates . Exceptions may occur in tables dealing with f i nancial data, when conformance with accounting practice is required, but not otherwise . Totals generally are considered to be the most important figures in tables . Placing them at the top helps the reader to see them first. THE BOXHEAD Th~s term refers to all the column headings across the top of the table . These headings indicate Lhe kinds of information entered in the stub and field below. The boxhead is separated from the title by a single horizontal ruling which runs the full width of the table, with one blank line between the title (or the headnote, if any,) and the ruling. It is separated from the first caption in the stub in like manner. Contents of the Boxhead The tabular f orm is based on the relation ship of data in two categories, one listed in the stub and the other placed in the boxhead, The category consisting of the fewer items or shorter captions is usually placed in the box head , since standard page dimensi ons provide for COMPONENT PARTS OF A TABLE less horizontal space than vertical. Exceptions Spacing in the Boxhead t o thi s generalization may occur where : l . Other material presented on makes it necessary to have a is wi der than i t is high. the page table that 2 . The data require a comparisons to be emphasized stub that is short and a wide boxhead . Arrangement of Col umns The sequence of columns should be in some orderly arrangement, as noted for i tems in the stub (see pages 22-4 ) . When the same headings a r e shown in the boxheads of two or more tables , the arrangement and wording should be uniform. A Total column, if used , is pl aced to the l eft of the col umns of which i t is an aggr egate . This appl ies both to columns which contai n subtotal s of other col umns , and to the column which contains the total of all other col umns or subtotal s (and thus is the first data column to the right of the stub) . Exception: in tabl es dealing with financial data, when conformance with accounting practice is required , the total columns may be pl aced on the right of the columns of which they are aggregates . Derived figures , such as averages , ratios , percentages , changes , and deviations, should usually be placed in columns to the right of the basic figures . Grouping coordinate column headings under a common heading which describes all the columns beneath it, aids in reading and interpreting tables and shortens the column headings by avoidi ng unnecessary repetition of words . These group headings are referred to as spanner headings in this manual and they are similar to the group captions in the stub. Spanner headings may also be used to group and describe two or more lower or subordinate spanner headings . The use of spanner headings is i l lus trated on page 33 . Wording in the Boxhead Column or spanner headings should always read horizontally. Single words may be hyphenated and placed on two or three lines, as necessary. Headi ngs are typed with initial capital s , except for articl es, prepositions, and conjunctions , which are not capital ized. Brevity is necessary because of space limitations . Standard or otherwise easily understood abbreviations may be used to save space . Use of abbreviations in one column heading does not usually require their use in other column headings of a table . If they pertain to related things , however , such as the states of the United States, the standard is to either abbreviate all or none of them ; but not a combination of both. Punctuation in the Boxhead The general rules on the punctuation in the stub are also applicable to the boxhead. (See page 28 ) • In addition, if a word normally fol lowed by a comma ends a line of a column heading having two or more lines, the comma may be omitted if clarity is not reduced . The f i rst exampl e below shows normal vertical and hori zontal spacing of column headings . The second shows minimum vertical, and the thir d maximum r ecommended verti cal spaci ng . Army Command Strength ) )CD By Sex By Component Total Regular Male Female Other Army \(2) Army Command Str ength ) -By Sex By Component ( Total Male I Female R~ar I Other I )0 Army Command Strength Ey Sex By Component Total Regular Male Female Other Army Vertical -Each colurrn heading should be centered in its box . It should not be lined up at either the top or bottom wi th other headings w:Uch differ in \'ertical size . In an informal table , however , the bottom line of each column heading is underscored, in which case the bottom lines of all column headings are lined up . The heading 8f the box wi th the most typed lines should gene~lly be centered in a vertical height one typewriter line larger than the num ber of typed lines in the column heading; that is, leave one -half blank line between the column heading and each of the horizontal rulings , as in the first of the above examples . The other column headings are then centered vertically (up and down) in their respective boxes . In some cases, the half-line blank space allowance may be deleted or stretched to one line . Headings of more than one line should be single spaced . Horizontal -The widest line in each column heading should be centered in its box, with a minimum of one blank typewriter space on each side . The other ~ines are also centered, Where there is an extra blank space, or only one blank space, on any line, put it to the right of the typing . Exceptio~s : put it on the left (1) in the first (or onl y) line of the last column or spanner heading, (2) in short spanner headings where it will produce better centering of the heading 8Ver a·vertical ruling, and (3) in lines ending with a hyphen. Generally if two or more lines are required in the heading, tney should be in an inverted pyramid or pyra~d form, depending on the length of the words in each lin;, , logical breaks in wording, and necessary hyphenation. TABULAR PRESENTATION BOXHEAD CONSTRUCTION Preferred Practice Centering: Headings should be centered top to bottom in each box • Three Wide-column Total Line Ratio HeadingHeading NOTE : Centered headings are the recommended standard for formal tables. This arrangementf ollows standard publishing practice. (See the Government Printing Office Style Manual, page 180 . ) For informal unruled tables, align the headings at the bottom of the deepest heading; center them left to right over the figures in the field below. Number of Lines : Headings should be put on as fev1 lines as possible • • • Too fast Assuming for Conditions Improper Turn Ri ght of Way Reading Position : All column headings should read horizontally Field Grade Company Grade Prog Auth Act Prog Auth I Act Field Grade Company Grade Pro-Auilhor-Pro-Author - Actual Actual gram ized gram ized Box Heights : All boxes on the same l evel should be the same height (determined by deepest box ) : Active Duty Total Military All Other Personnel June July June July June I July Location of Total s : "Total" column should be at the left of t he columns it totals • Military Civilian Total Pet of Pet of Number Number Total Total Faulty Practice • • • do not align them at the bottom. Three Line Wide-column Total Heading Heading Ratio . . • or at the left, Three Wide-columnTotal Line Heading Ratio Heading . . • or at both bottom and left• Three Line Wide-column Total Heading Heading Ratio • • • do not put them on as many lines as possible . Too Assuming Fast Impr oper Right for Turn of Conditions Way . . • avoid vertical readings ~ abbreviating or by splitting wor ds . Field Grade Company Grade '"d '"d Q) Q) t;l t;l @ rl rl ·rl @ ·rl ;., ~ f..< 0 0 0 fib fib ~ ..c ""' () 0 ..d () H +" ~ 1j -< «1 ~ &:: £1 • • • avoid this kind of construction. Active DutyTotal M:i.litary All Other Personnel June July June July June July • • • avoid putting t otals at the right (but see "Arrangement of Columns" , page 29) . Military Civilian Total Pet of Pet of Number Number Total Total COMPONENT PARTS OF A TABLE Units· of Measurement in the Boxhead Units of measurement that describe t he con tent of the columnar data in the field (e . g ., pounds, dollars, percent, etc . ) , and notations which-indicate that the figures have been abbre viated ("ODDs" or "Thousands") are usually plac ed in the column or spanner heading if they re fer to a single column or group of columns. They are not used in the field, except in field span ners. They are seldom used in the stub, but are properly placed there when the measurement unit changes from line t o line . If a measurement unit or r ounding note is applicable t o all or most of the data in the table or .is too long or clumsy to include in the column headings, preferred placement is in a headnote immediately above the boxhead. Put units of measurement after or below the capt ions in column headings and inclose them in parentheses. Make them brief; use clear abbreviations where necessary. Capitalize the initial letter of the first word and use l ower case for all other words; f or example : "Deliveries (Millions of dollars), " ".Area (sq. ft. ), 11 11 Amnrunition (Thous . rd. ), 11 etc. Sometimes a unit of measurement comprises an entire column heading caption . If it does, the parentheses are omitted, and initial capitals are used on all words, except the prepositions , articles, and conjunctions; f or example : "Percent of Total , n "Number per 1, 000 Average St rength , 11 etc. Lettered Columns Columns seldom require lettering. Tables having many columns, however, and especially those occupying facing pages, occasionally ·have lettered columns for ease of reference , Column Number Cost Army..Area and Installation of of ( Vehicles Operation A B c 1 TOTAL . . . . 2 FIRST ARMY AREA 3 Boston Army Base • 4 Buffalo Recruiting Main station • . 5 Fort Devens . . . \ 16 SECOND ARMY AREA 17 Fort .Campbell 18 Carlisle Barracks L- numbers generally should not be used i :-. Army reports. Column letters should always run from left to right beginning with the stub, and be l ocated directly below the boxhead. They are separated from the boxhead and the field and stub by single horizontal rulings . The table above illustrates the use of column letters and line-numbers , THE FIELD That part of a table which contains the numerical data is called the FIELD . This is the part below the boxhead, and to the right of the stub. The field is set off from the rest of the table by three rulings : one at the top (to separate it from the boxhead)~ one at the left (to separate it from the stub} , and one at the bottom. Component Parts of the Field l . Cell -the space normally oc cupied by one figure entry in the fiel d . A cell without a figure entry should contain a word entry or a symbol to explain omission of the figure (see 11 11 Every Cell Must Have an Entry, bel ow) . 2 . Line -~ horizontal row of cells opposite a stub caption. 3 . Column -a vertical row of cells under a column headi~g. Field Spanners These are spanner he~dings that cross the table within the field as shown in Tables lB , 2 and 3 on page 33. A single boxhead, therefore, applies to all l evels of the table . An identi cal stub may be repeated with each level, or the stub captions may be different for each field spanner. F~eld spanners may be used for the following purposes: l. To reduce the width and increase the depth of a tab~e with an unusually wide boxhead and a shallow stub (see Tables lA and lB, page 33). 2. To avoid repeating the boxhead with each repetition of the stub (see Tables lB and 2, page 33) . 3. To reduce the number of times an identical svub must be repeated (see Table 3, page 33). 4. To permi~ placing long major group captions in the fie:d when space limitations prevent their placement in the stub (see Table 2, page 33). 5. To permit combining several short related tables with a common boxhead into a single compact table (such as a time series percentage distribution of enlisted personnel by length of service, age, education, and marital status) . As a general rule, field spanners should be avoided whenever uossible because they break across the column; and separate the column headings from the figures described . Sometimes these spanners can be shifted to the stub, as in Table lC, page 33 . This can be done by using abbreviations, or short terms with qualifying footnotes . If a stub can be made slightly wider, or an extra line used for each caption, the abbreviation might be avoided . 31 TABULAR PRESENTATION Uni ts of_ ~easure Units of Measure should not be placed in the fiel d . The standard place f or them is ins i de the boxhead as a parent het ical addition to a column heading or as a separ ate column heading, or in a headnote above the boxhe'ad . Occasionally a field spanner may also include or be a unit of measurement (see Table IB, page 33). Every Cel l Must Have an Entry An empty cel l , where a column and row meet , causes uncertaint y . Wer e no data reported? Was the figure received too l ate to be incl uded? Is there some reason why a figure is not called for in thi s space? Was the figure left out by mis take? Some of these questions may answer themselves ; but in most cases they do not . Even if the r eader could find the answer by studying the tabl e , it i s better to save him the trouble . The standard is to put something in every cell, and to use the fol lowing customary symbol s wher e appropr iate: gj Q/ etc . (Ref er ence Letter) means look for a f ootnote begi nning with this reference symbol at t he bottom of the tabl e . Always use a series of these when there is more than one reference note . (§/ can be used when there is only one. ) They should be used in al phabetical order -across and down the tabl e (see: "Wher e to Place Reference Symbol s ", page 35) . (Reference letters are used in tables and comparable ref er ence numbers are used in text to minimize errors arising f rom fail ure to r ealize that the symbol is a symbol. For example: 1Q/ in an otherwise blank cell could be read , or copied erroneously, as the quantity 10 . ) * (asterisk) means l ook for a reference note beginning with this symbol. It can be used when there is only one ref erence note . 0 (zero) is a number , and means a quantity of none ; (as: Percent completed, 0; Production during month, 0) . It does not mean an absence of data. (double dash) means that no number is applicable . It should be used only when the r eason is immediately obvious . For example , in a percent distribution col~ where only one subtotal is di stributed percentagewise , the group caption l i ne would indicate 100 . 0 and each line thereunder the appropriate percentage . All otherlines would cont ain the double dash. (dash) should be us ed in the fiel d only to mean "minus 11 or negative ; (as : -641) . + (pl us sign) should likewis e be used in the f iel d f or "plus11 or positive f i gures ~hen they are shown w~tp minus figures (as: Net Change ; Aug . ·;62 , Sep . 15) . The plus and mi nus signs may also be endorsed in parent heses to minimize the possibility of misinterpretation in small tables with ade quat e room . Symbol s such as E, for estimated; P , for preliminary; and NA , for not avail abl e may be used in exceptional situations . For example , they might be used for tables produced daily or weeY~y, with limited space, for sophisticated, EXAMPLES OF· ENTRIES IN THE FIELD STUB FIELD 0 BOXHEAD Q ~ ,. v TOT.AL . . . . -~ GROUP CAPT ION . . . . . Line Caption . . . . . . Line Caption . . . . . . Line Caption . . . . . . GROUP CAPTION Line Caption Line Caption Li ne Caption . . . . . . . . . . . . . . . . . . E 75 ,045 $14,075 + 150.1 p 10 , 755 100, 0 --gj 330 J 745 E 40,643 10,020 + 25 . 2 6, 350 59. 0 --550 165 E 10, 603 5,005 -b":b 6, 025 49. 7 --} --w 190 20,040 2,505 -7. 3 325 y 0 10, 000 2, 510 + 39.1 0 9. 3 --360 165 34 , 402 4,055 + 124. 9 p 4, 405 41 . 0 100. 0 195 165 10 , 291 1,030 + 159. 2 p 2, 205 20. 5 50 .1 25 40 24,111 2, 020 0. 0 1 , 200 11.2 27 . 2 70 12 7, 870 1,005 -34. 3 1,000 9. 3 22 . 7 100 113 ~ Footnote relating ~o figure shown. Q/ Footnote r elating to the omission of a figure . £1 Less than 0. 05 percent . E Designates an estimated figure . P Designates a preliminary figure . Designates a cell f or which data are not applicable, and the reason is clearly understood without a reference symbol and reference note . 0 Designates a zero quantity, not-the absence of data 32 COMPONENT PARTS OF A TABLE USE OF FIELD SPANNERS A short, wide table with repeated column headings under two or more spanner headings • • • • • • • • Average Number Number per 1,000 Average Strength Fiscal Year Total Confined AWOL Sick Total Confined AWOL Sick l9X2 . . . l9X3 . . . @ l9X4 . . . • • • when rearranged to make a repeated-stub But Field SpannerE can often be avoided by puttable, may require Field Spanners like this: ting headings in the stub, like this: Fiscal Fiscal Total Confined AWOL Sick Total Confined AWOL Sick Year Year Average Number NUMBER l9X2 l9X2 l9X3 @ l9X3 l9X4 l9X4 ® RATE fY' Number per 1,000 Average Strength l9X2 l9X2 l9X3 l9X3 l9X4 l 9X4 fY' Number per l,OCO average strength. Long, major group captions in a narrow stub may Stub repetitior-iE reduced by Field Spanners: be converted to Field Spanners : Baaic Unit Basic Unit Component Trained Trained Trained Trained Missile Total Jun Jul Aug System First Army Second Army Pay and Allowances TOTAL TOTAL . . Regular Army Missile A . . Army Reserve Missile B . Natl Guard Missile C . . 0 Army of U.S. Missile D . . Missile E . . '::'hird Army Fourth Army Missile F . . Missile G . . TOTAL Operations and Maintenance Regular Army CD Army Reserve TOTAL . . Natl Guard Army of U.S . Missile A . . Missile B . . Fifth Army Sixth Army Missile C . . Missile D . . Missile E . . TOTAL Missile F . . Missile G . . Regular Army Army Reserve Natl Guard Research and Development Army of U.S . l 214-853 0 -66 -3 TABULAR PRESENTATION regular readers. The meaning of these symbols should be indicated either in a footnote, or in a glossary of terms and abbreviations in the f r ont part of a report. Generall y , however, reference symbols (such as~' Q/, Q/, etc. ) are preferable . Zeroes and Decimals If a column of figures contains any decimal fractions , insert a cipher to the left of the decimal ~oint for figures l ess than one (including zero) ; and for whole numbers (numbers without fractions) , insert the decimal point and ci pher (or ciphers) to the right of the numbers . Exceptions may be made when space is exceedingly t ight or when the number~ have a speci al signi ficance such as those in this manual concerning line weights which are the same as certain pen numbers divided by 1,000. All of the figures in the column usual ly must have the same number of ciphers or digitsafter the decimal point. If the column of figures consis~s entirely of whole numbers , do not put decimal points or ciphers to the right of the numbers . Examples are shown below: Item A • Item B • Item C Item D 0 . 0 • 0 . 8 3 . 0 4. 9 Item A Item B Item C Item D 0. 00 0 . 81 3 . 00 4.98 Item A Item B Item C Item D 0 l 3 5 Dollar and Percent Si,ns A dollar sign ( ) is put before the first figure in a column of the field when all figures in that column refer to money. This sign may be omitted if a unit of measurement in the headnote , boxhead , or field spanner indicates that all of the figures in the table represent dollar values: (Millions of dollars, Thousands of dollars) . A percent sign (%) should not follow figures . The fact that the numbers r epresent percentages should be indicated in the title, boxhead, or stub. Usually it is better to use the wor d 11 percent11 and avoid the word 11 percentage11 in making t his explanation. Commas in Figures Commas should be used to set off each three digits in a whole number . If space is extremely limited, it is permissible to omit the commas, but this is seldom desirable and should be avoided, particularly if any of the figures ex ceed more t han four digits . It is often better to abbreviate the figures or redesign the table rather than to omit the commas . Pl acement of Figures in the Column The l argest figure i n each column should be centered within the vertical rulings . It should have a minimum of one blank typewriter space on each side of the figure ; the remaining figures ar e then lined up with the right-hand side of this figure. When t he number of bl ank spaces is uneven, put the extra space to the left of the figure . When the caption in the stub takes two or more lines, pl ace the figures on the same line used by the last line of the caption. TABULAR NOTES A note is a special explanation that the reader needs to help understand the table or a part of it. This mode of explanation should be used when the information provided does not logically belong in the title or in one of the column captions, or when it is too long to be included in such captions . Some of the notes that appear on tables may not be needed , but omission of needed notes is a more common fault with the result that the reader is given insufficient information about the content and coverage of figures to make proper use of ~hem ~n the decision-making process. Careful consideration should be given to the needs of the reader . Often he may be given sufficient information through precise wording of the t itl e (or subtitl e) , or of the headi ngs of the columns which eliminates the need for a note . A note should always be used, however, when i t will take a technical explanati on out of the title. Kinds of Notes Tabular notes may be grouped into three categories based on the location of the note . Notes at the top of the table above the top rul ing are referred to as headnotes. These notes are needed occasionally. Those that go inside the table , usually in the field, are referred to as body notes : they are rarely used and only in unusual situatio~s. Notes that appear at the bottom of the table below the last rule , includ ing general notes, and source notes as well as the more customary reference notes, a r e referred to as footnotes . Footnotes are used quite often. Headnotes A headnote generall y applies to the whole table and is so important that the reader must read it before he uses the tabl e . A headnote usually is a continuation of the titl e , explai ning all or part of it as illustrated below: (Based on a 5% sampl e of male commissioned Officers) (Oversea inventories based on incomplete reports) A headnote is not a subtitle that is a part of the title. A headnote is added information about the title. The most frequent use of a headnote is to show units of measurement, like : (Millions of dollars) (Thousands of short tons) The one exception to the rule that a headnote should apply to the entire table is when it is used to indicate a unit of measurement that applies to only part of the data. Thi s use of a headnote is appropriate when the statement would be too long or too awkward to include in a column heading, or when it might apply to many but not all of the columns . Examples of such excep tions follow: (Rates per thousand average strength) (Tonnage in thousands of short tons) (Strength in thousands) COMPONENT PARTS OF A TABLE Do not use a headnote to show the unit of a single column: This puts too much emphasis on one small part of the table . Do not use them unl ess they are necessary. placement. Put a headnote in the second line above the top ruling of the boxhead and in the second line below the title, or subtitle . This leaves one blank line below and above it. If space is tight, omit ~of the blank lines. If the note explains the title, omit the blank line above; if it is a unit of measurement, omit the line below . Examples : TITLE OF TABLE (Normal location of headnote) TITLE OF TABLE (Headnot e about t i tle here, if space is tight) =:J LL TITLE OF TABLE (Unit of measurement here if space is tight) A one -line headnote is always centered· left to right. The second line of a two-line headnote is centered under the first line . ~· Headnotes are typed in lower case letters (no capitals except the first letter of the first word and the first letter of proper nouns), and they are inclosed in parentheses . Body Notes The only reason for setting a note inside the table is to give it strong emphasis when t he reader is using the data. Few notes about one small part of a table are this important. The standard practice is to put a reference where the note applies, and make the note a footnote . The space available for a note in the body of a table is generally so restricted that the question of whether to put the note there or at the bottom of the table usually answers itself. But once in a while there is both the need and the space . Some examples are shown below: Production to start in July. No contract for this item. Discontinued l Oct 1945 . Cancelled. Footnotes Most tabular notes are footnotes , and these may be classified in four types : 1 . Specific reference notes -these refer to a specific part of the table and they start with a ~eference symbol (~, etc . ) . An asterisk(*) may be used if there is Jnly one note . 2. Special notes -these refer the reader to another page . Example: "See f oatnotes at end of table . " 3. General notes these relate to the whole table and start with the word "NOTE . " 4 . Source notes -these explain the source of the data and start with the word "SOURC~:. " When a source explanation is given :or only ~part of the table , use a specific reference note . A reference note explaining all sources may also be used . Most footnotes are speci:ic reference notes . General notes are seldom needed ; when they are , they often are used as a headnote . Source notes serve two purposes: They refer the user to the source o: the data, and they give credit to ~he originating organizatio~ or person. The source reference may be useful to i~dicate the tectnical content of the data and tells the user where additional detail may be obtained. Source notes are only one means of indicating sources of data: where one or more sources are used for a series of rel ated tables in a report, they may be indicated in the Introduction, Explanatory Notes , or Technical Appendix portion of the report, rather than be repeated on each table . Where to Place Reference Symbols In the title, the stub, or the boxhead put the reference Sj~bol after the word or group of w0rds to which it relates. In the field, put it before the number to which it relates, but if it relates to a word entry in the field, put it after the word . In a cell without a number or word entry, put it at the right side of the column (under the last two digits -as shown in table below) . Place the reference symbol before the first word of the footnote itself ~t the bottom of the table. A reference symbol should be close to the item to which it relates; for example : Copper ~ •• Q/ 2,964 It would be .better to put reference symbols half a space above, like this : copper ~ • )2/ 2 , 964 but this is hard to do unless the typewriter has a special ratchet :or half-spacing. Moreover, TABULAR PRESENTATION there is not room to raise the reference marks unless the lines are double-spaced. This generally wastes too much space : double-spaced tab_es are almost twice as l ong as singl e-spaced table& Leave one blank space between reference symbols and points of reference where space permits . Reference symbols generall y should go from top to bottom of the table and from left to right. When used opposite f i gures of about ~he same length they may be lined up vertically. The sequence of tabular r eference symbols is shown in the table at the bottom of this page. Where and How to Type a Footnote Footnotes of the table are al ways put at the end -not at the bottom of the page (unless the tabl e end is at the bottom) . The standard arrangement for single-page tables is to: l . Put the first l ine of a footnote in the third line below the l ast typed line in the body of the t able ; this allows one blank line above and one below the ruling at the bottom of the table . If space is tight, put it in the second line . 2. Put notes one under the other (each one starting a new l ine) whenever space is avai l able , Otherwise put a few short notes one after another on the same line . 3 . Start each r eference symbol (or the word 11 NOTE" or "SOURCE: ") at the left edge of the table . 4 . Start the first word after a lettered symbol in the fifth space from the left edge . Leave two blank spaces after the symbol . 5 . Start the second line of lettered notes at the left edge of the table . 6 . Single-space all footnotes , Leave no extra space between any lines . 7. Type in lower case l etters except first letters of first words and proper nouns . Footnotes on "Continued11 Tables , On multipage tables all footnotes should appear on the last page of the tabl e (they may appear on the last two pages of a table comprised of a series of repeated-stub, facing pages) . Al l pages except the last should carry two special footnotes : "(Continued on next page) 11 and "(See footnotes at end of table) , " or 11 (See footnotes on page000) , 11 Only the first notes appear on the next to the last page of a table consisting of a series of repeated-stub, facing pages , with footnotes divided between the last two pages . The first special not e is pl aced in the first l ine bel ow the bott om horizont al ruling, and the second special note is placed in the third l i ne below this r uling, One bl ank line is left between the first speci al note and other types of footnote . Order of Footnotes When more than one kind of footnote is used in a tabl e , pl ace them i n this order: 1 . Special notes . 2. General notes . 3. Notes st arting with reference symbol s . 4 . Source notes . A TABLE ABOUT TABULAR NOTES (When used , a headnote goes here . See page 35 . ) Where When ExamplesKind 'of Note Placed Needed on Page- HEADNOTE • . . Above About Title Above Occasionally 00 Unit/Measure Above Occasi onally oo,oo,oo BODY NOTE . . Inside Seldom oo,oo FOOTNOTE . . Bel ow Special e./ . lst Varies 00 General . . 2nd Seldom 00 Reference !?/ 3rd Often oo,oo,oo Source £1 . 4th Varies gj 00 (Continued on nAxt page) NOTE : A general note not important enough to be a headnote is put in a footnote like this . e./ Special instructions like 11 (Continued on next page)" are put first, as shown above . Q/ Reference notes start with a reference sym bol and are put in alphabetical order. Q/ A source note about part of the table is treated as a reference note . gj See text, page 000 . SOURCE: Handled like ~his if it applies to the entire table. The kinds of note used in a single table, a series of related tables , or a report, should pe kept to a minimum. SEQUENCE OF REFERENCE SYMBOLS eJ Spanner Heading gj Spanner Heading ~ Stub Heading Q/ Column Heading £1 Spanner Heading g/ Column Column Heading k/ Heading 1/ Column Heading h/ Column Column Heading U , Heading jj Column Heading U TOTAL 98 , 652 12,052 !!l/10' 976 30,829 20,696 12, 965 11,134 Line Caption Line Caption gJ Line Caption 20 , 256 8 , 705 6 , 410 3 , 069 2, 135 790 gJ 6 , 843 1 , 871 1 , 393 5, 260 l , 895 1 , 632 2,123 829 rJ 1,926 613 1,645 1, 035 1, 362 950 COMPONENT PARTS OF A TABLE Examples of Footnote Location Some of the standard procedures for locating footnotes at the ends of tables, which are discussed on the preceding page, are shown in the illustrations appearing below. The examples pertain to bot h single-page and mult ipage t ables prepared for typewritten SINGLE-PAGE TABLES I t em 67,423 116. 2 28,423 21 standard placement of a footnote . One blank line above and one below the bottom ruling. Item • . • . .• 67,423 116. 2 28,423 a Place footnote here only if space is tight. Item • • • . •• 67,423 ll6. 2 28,423 s/ Reference notes should be listed one under another in alphabetical order . Q/ This makes each reference easier t o find . £I This repeats the order in which the refer ence symbols appear in the table . Item • • • • • • 67,423 116 . 2 28,423 21 This one-after-another method is best when space is t ight. Q/ Less than 0 . 1 percent. Q/ See definitions page 00 . gj Not available . §) Completed 31 Jul. f) Less than $500. s/ All footnotes start at the left edge of the table. The first word in a lettered reference note begins in the fifth space, or in the fourth space if an asterisk is used. Q/ Not es are single spaced. Item •••• 67,423 ll6.J:: Army stati st ical reports. The styl e manual of t he U. S. Government Printing Office should be consulted when preparing tables that are intended t8 be printed by that agency. Note especially the s~andard indentions of the footnot es fron the l eft-hand margins of the pages. 11GONTINUED" TABLES Continuous stub ~ables : (All ?ages except last) Last line caption (Continued on next page) See foot~otes at e~d of table . (Last page of table) Last line caption sf Usually a note about title or column heading. h/ Usually a note about stub or fiel d item. Repeated-s~ub (double -page-spread) tables with footnotes on last two pages : (All pages except last two) Last line caption (Cont inued on next page) See footnotes at end of table . (Next to last page) Last line caption (Continued on next page) 21 This example assumes five specific notes, t otaling six lines. The first goes here . Q/ The second goes here . (Last page of table) Last line capti8n NOTE : The second line of a footnote is placed (Continued on next page) at the left edge of the table . s/ The third n8te goes here . 21 Footnotes are typed in l ower case l etters. i/ The fourth here. They are the only tabular elements that end with if The fifth here . a period. TABULAR PRESENTATION TABULAR RULINGS Rulings are used to help explain the organ ization of the table . When a table is extremely crowded -or extremely empty -rulings also help the eye to locate figures accurately. Hany tables are overruled -that is cut upinto many small pieces -which makes the table appear cluttered . Others are underruled or have no rulings at all. This makes them unnecessarily hard t o comprehend . WHERE TO PUT RULINGS Tabular rulings should be uniform from one table to the next, but rigid standardization de feats the primary purpose, which is to help the reader understand and read each table . When an unusual table needs a special ruling treatment use it: but be sure that the table really is un usual and not just one that should be redesign ed . Otherwise, follow these standards: USE A VERTICAL RULING between all columns of data (including column headings) and between the first column and the stub. Do not use outside vertical rulings (at the left edge of the stub or the right of the last column ) . USE A HORIZONTAL RULING above and below the boxhead and below the last line of data; also, between column headings and the spanner headings above them. When column letters are used, putthem below the bottom ruling of the boxhead and draw another horizontal ruling beneath the column numbers. Do not use horizontal rulings across the stub and field : instead, make best use of capitalization, underlining, indention, and spacingin the stub to separate and clarify successive lines or groups of data. Usually the boxhead rulings, and the one at the bottom of the table, are the only horizontal rulings needed . Excessive rulings, and rulings too heavily drawn, dominate the presentation; tend to obscure the figures; and are time-consuming to draw. WEIGHT OF RULINGS Two weights of ruling are usually enough . One should be clearly heavier than the other: lines that are just a little different are not recommended. Relatively sm~ll gradations of line weight may be useful if the tabl e is t o be used without reduction, but if it is to be reduced in reproduction, small differences cannot be seen. The following weights are recommended : 1. Heavy --. 02" (About 1 1/2 points*) 2. Light --. 01" (About 3/4 point) The line vleight scale shown on page 102 of the Graphic Presentation Chapter shows these and other standard ruling weights. * A point is a printer's measure= 1/72 inch. After the final typing, draw all tabular rulings, including underlining of stub captions and field entries, with a ruling pen and black drawing ink. In a simple table, only two lines need to be heavy: the top and the bottom horizontal rulings . In a large complex reference table, and in other special circumstances, heavy vertical rulings may sometimes be necessary. Heavy rulings,however, are most effective when used sparingly. BLUE PENCIL RULINGS Even when a table is typed on typing layout p~~er (printed with light blue lines that match typewriter spacing), the draftsman should usual ly be shown exactly where rulings are to be drawn . (Quite often the rulings should not be made on the printed blue lines, but between them. ) Each table should be ruled faintly in non photographing light blue pencil before giving it to the draftsman for inking. (Blue pencil lines that are heavy sometimes interfere with the even deposit of black ink. ) If the tables are repetitive and if the draftsman is thoroughly familiar with a standard pattern established for ruling such tables, the blue pencil rulings may be omitted . They may also be omitted if the draftsman is sufficiently familiar with both standard principles for rul ing statistical tables and the meaning of the tabular data presented . A professional table de signer can generally speed up production and ac curacy if he rules the tables in pencil rather than depending on the draftsman to make the rul ings after studying relationships between the various entries in the table . If there is insufficient time to draw in the complete sat of blue guide lines, put a light blue pencil mark to show where to placeeach ruling. OTHER RULING CONSIDERATIONS The above standards are intended primarilyfor tables prepared for reports that are to be reproduced by the photo-offset process . The standards for placement of rulings are also applicable to tables reproduced by other methods, but when necessary a single weight produced by stylus or special pencil may be used for all rulings. In tables for which all copies are produced through use of typewriter carbons, all horizontal rulings, including lines separating spanner headings from the column headings beneath them, are usually made with the typewriter. Vertical rulings may be omitted if spanner and column headings are spaced far enough apart to avoid confusion. If the table is a text tabulation, the only rulings necessary are those for the underlining of the bottom of each column heading, and, if appropriate, stub captions and field entries denoting sub-totals. SPACE REQUIREMENTS SPACE REQUIREMENTS In designing tables, adjustments often must be made to resolve conflicts between the needs for content and the availability of space. Frequently, a number of trial layouts are necessary in making these adjustments . TABLE WIDTH AND LENGTH Table WIDTH is determined by two factors: 1 . Horizontal space required for the stub and column headings in the boxhead. 2. Horizontal space required for the longest caption in the stub and the longest figure in each column in the field. Table LENGTH is determined by the vertical space occupied by the title and subtitle, headnotes, stub and column headings, stub captions, and footnotes . Generally, considerations of width are more important than length in planning the layout of a table. Long tables can be presented effectively on several pages, if necessary. Tables, too wide to present on one page, however, result in less effective presentation (i.e., the boxhead could be extended across two facing pages; the table could be arranged in two or more decks having the same stub, but with different column headings; or the material could be shown in two or more separate tables). The consecutive steps to follow in planning the width of a table and the arrangement of the boxhead are described below . The procedures will vary somewhat from table to table, depending on the specific problems encountered. Typing each step will be helpful, but is not necessary if the spacing is planned carefully and if counted accurately. The example which follows requires fitting a table to a width of 48 typing spaces (the width of this column of text) . The original layout of this table is illustrated at the bottom of this page in Figure lA. The first step is to count the number of characters and spaces needed for each column heading and for the longest stub caption and the widest field entry in each data column. This count should include an allowance for the standard one blank space on either side of the widest line in each column heading, the widest field 7 13 /3 /0 Obligations Obligation Month Payments Incurred Adjustments lA /O /1,/ • /~ /(p entry in each column and on the right side of the longest stub caption. This count may be recorded for convenience in pencil in each column heading and on the first line of the stub and field, as illustrated in Figure lA. The space count for the boxhead is then totaled, as is the count for the stub and field . In this trial layout, the boxhead requires 57 spaces and the stub and field entri es 72 spaces . The second step is ~o reduce the width of the boxhead to the required 48 spaces . In this table, a spanner heading is used for the first two data column headings, the word "Adjustments" is divided into two lines, and the phrase "End of Month" is re-phrased "Month-end"; in addition the blank spaces to the left of the stub heading and to the right of the last column heading are eliminated . This reduces the boxhead width to 45 spaces . The third step is to abbreviate the stub, and narrow the data columns by eliminating the dollar signs and, when appropriate , rounding the figures to thous~ds, using a headnote to indicate the unit of measurement . This reduces the combined width of stub and field to 45 spaces . The fourth step is to compare the revised space requirement for each column heading with that fer the stub and the field entries for each column, and reconcile any differences. That is, use the largest of the t~o space requirements for each column. After taking all four of the above steps, the material shmm in Figure lA could be presented in the more compact single-column table illustrated in Figure lB. (Thousands of do11ars1 Unpaid Obligations Obliga- Payments ~h Incurred Adjust- tions, 18 ments Month-end Jul 973,299 -199,269 1,016,454 5, 229 , 252 Aug 686,433 -40 , 770 944,225 4 , 930 , 690 Sep 700 , 249 + 18,119 934, 201 4,714,857Oct 762 , 938 -64,121 673 , 533 4 , 740 , 141 Nov 795,767 -106,087 1,019,016 4 , 410 , 805Dec 845,843 -42,611 759,314 4 , 454,723 The procedures and the principles described above also apply to fitting a table to a fullpage width (100 spaces in manual) . /"f. Unpa1d 7 + 13 + 13 + 10 + 14 = 57 Obliga"Lions End of Mon-:.h 16 10 + 14 + 16 + 16 + 16 =72 July $973' 298'770 -$199,268,904 $1,016,453,592 $5,229,252,349 August 686,432,912 -40,769,646 944,224,904 4,930,690,711September 700,248,981 + 18,118,822 934,201,147 4,714,857,367October 762,937,817 -64,120,825 673,532,620 4,740,141,739November 795,767,248 -106,087,495 1,019,015,658 4,410,805,834 December 845,843,181 -42,611,875 759,313,901 4,454,723,239 TABULAR PRESENTATION The f ollowing table summarizes these four steps and shows how to condense the ·requirement to the standard 48 spaces : Number of Spaces Col Col Col Col Step and Procedure Total Stub l 2 .3 4 l. Or iginal l ayout: Column headings 57 7 1.3 1.3 10 14 10 14 16 16 16 stub and field 72 Reduce width of 2 . column headings 45 6 10 9 10 10 3. Shorten stub and round figures 45 4 9 ll ll 10 4 . Use highest figure 6 10 ll ll 10in steps 2 and 3 48 UNUSUAL SPACING PROBLEMS Most tables when first designed will not fit the space exactly. A little spreading or squeezing, however, if carefully done, usually will bring them to the right size . Even though typewriter spacing is fixed , ther e are many ways to adjust the size of the tables. Adjusting a 11 Tight11 Table The usual trouble in tabular layout is too much material for too littl e space . The reasons for excess width usually include: too many columns , many-digit figures , lengthy column heacings, lengthy . st~b captions; and those for excess depth include too many stub items, too many lines or leveis in the column headings . If there is not enough space left to right: l . Look for columns of figures that could take more 11 r ounding . 11 2 . Look for the wide column headings that could be broken into more lines. 3 . Look for words in column headings that could be abbreviated. 4 . Look for related column headings that could be put in fewer spaces by placing duplicating words in spanner headings. 5 . Look for long stub capti ons that can be abbreviated or can be put on more than one line . If there is not enough space up and down : l . Remove the blank lines between group captions and their components . 2 . Remove the blank lines above group captions. 3 . Remove the bl ank l ine below the total. 4 . Look for column headings that can be put on fewer lines by abbreviating or by making the column wider. 5. Delete the standard blank line allowance in column headings (see page 29). 6 . Put two or more short footnotes on the same line (see page 37) . 7 . Omit the blank line bel ow ·or above a headnote (see page 35) . 8. Omit the blank line between the l ower horizontal ruling and the first footnote (see page 37) . Adjusting a 11 Loose 11 Table Sometimes the material is too sparse for the space . The reasons for surplus width include very short stub captions , very few columns of figures , and one or two-digit figures ; and surplus depth usually results from very few items in the stub. If there is too much space left to right : l . Add extra space to the width of the total columns . 2, Add extra space in the other columns. 3. Add extra space at the right side of the stub. If there is too much space up and down: l. Add blank lines between major breaks of the stub. 2 . Add a blank line or two in the boxhead . 3 . Add a blank line below secondary captions in the stub. 4 . Use more frequent l ine breaks in a long st> ries of items (for example , group by 2' s instead of 3 1 s , 3 ' s instead of 5 ' s , etc . ) . HANDLING EXTREME CASES Sometimes spreading, squeezing and rearranging is not enough . Some tables will not fit the space without drastic redesign. These troublesome tabl es generally are : l. Long and narrow (long stub, narrow boxhead) . 2 . Short and wide (short stub, wide boxhead) . 3 . Oversize (much too large in either or both directions) . Extremely l ong or short tabl es are discussed on page 41 , under 11 0dd-Shaped Tables . 11 Oversized Tables Some oversized tables are so large that the usual procedures cannot be used for fitting them to the standard page . This problem usually can be solved by 11 breaking11 the table or by reducing it. SPACE REQUIREMENTS Odd-Shaped Tables Tables that are too long and narrow or too short and wlde are awkward ; especially when t hey must be combined with other material . The illustrations on this page show two typical oddshaped tables and how they can be rearranged to fit standard page dimensions . Column ColumnStub Heading Heading Item 123,456 432 , 567 Item 16, 478 42 , 365 Item 9 , 543 9, 100 Item 5, 75 987 Item 320 , 47 76,532 Item 99 ,650 132 , 005 Item 57 , 453 516 ./--, Long , Narrow Table: Item 4, 359 3 , 255 ~few columns ; many stub items . Item 65,438 110 , 542 Item 8, 447 9, 789 Item 113, 662 46 , 347 Item 192 5 , 639 Item 430 89 ,035 Item 3, 534 400 ,164 Item 12, 345 5, 987 Item 478 13 , 705 Table lA could be spread to fit the standard width (half-page or full-page wide), but this would give it a 11 padded11 look. The location of the stub and column classifications could be switched but that would merely reverse the original problem: the table , instead of being too long and narrow, would be too short and wide . It would be even shorter and wider than Table 2A below : Stub Column Heading Column Heading Column Head . Column Head . Item 30,338 58, 567 123 185 Item 13, 864 40 ,345 567 296 Item 9,123 23 , 313 948 538 Stub Colwnn Heading Colwnn Heading Colwnn Head. Colwnn Head. Item Item 30,338 13,864 58, 567 40,345 123 567 185 296 Item 9 , 123 23 , 313 948 538 ---.J----.....--...L.--.....(.2-8}--'---,---'-.,--- Stub Column Head . Column Head . Column Head . Col. Head . Col. Head . Item Item 323 98 343 , 984 23,023 843,967 1 , 023 97 54 13 34 Item 14 34 , 931 19,324 68 15 The best solution is to convert Tabl e lA to a divided-stub table such as lB . I n this arrangement , the boxhead is repeated in each section of the table and a heavy r ule is placed midway from each side: Column Column Col umn Column Stub Stub Headi ng Heading Heading Heading Item 123, 456 432 , 567 Item 65 , 438 110, 542 Item 16, 478 42 , 365 Item 8, 447 9, 789 Item 9,543 9,100 Item 113 , 662 46 , 347 Item 5, 75 987 Item 192 5, 639 18 Item 320, 475 76 , 532 Item 430 89 , 035 Item 99 , 650 132 , 005 Item 3 , 534 400,164 Item 57 , 453 516 Item 12, 345 5,987 Item 4 , 359 3 , 255 Item 478 13, 705 The divided-stub arrangement is especially good for a table that has only one column of figures . In a long list of that sor t , the stub can somet imes be spl it into three or even four sections . A shor t table such as 2A could be widened to fill the page i-Tidth , but then its shape would be even more awkward . Reversing the stub and column headings wo·.lld make it almost as long and narrow as Table lA . The tabl e might be modified by putting the column headings in 11 decks11 or level s and repeating the stub on each level as in 2B . This arrangement is somewhat better than a wide table with many col umn headings and only two or three rows of figures . Sometimes the tabl e may be so wide that three levels of column headings are needed . SHORT , WIDE TABLE: A. many columns , few L/ stub items . Column Column Column Col. Col. Head . Head . Head . Head . Head . 323 343 , 984 843,967 97 13 98 23 , 023 1 , 023 54 34 14 34 , 931 =.9 , 324 68 15 The column headings on each level of this repeated-stub table relate to the data on that level on=.y . This differs from a repeated-stub table using field spanners (page 33) . Although these examples show how to redesign small summary tables , the same princip~es can be appl ied to larger tables . For example, a long table approximately half a page wide could be rearranged to make a shorter table a full page wide . A short table a page and a half (or more) wide could also be rearranged to make a longer table a full page ~ide. This avoids showing the table on more than one page . TABULAR PRESENTATION A third method (for wi de tabl es) is to use a folded page . This is not r ecommended . Fold-in pages are difficult to read, they are likely to tear and they present special printing and binding problems. Breaking Oversized Tables . Tables that fit the page one way but are much too big the other way often should be broken. For example , a table that is much too wide for one page may be broken down the middle and shown on two facing pages . These tables should r epeat the stub on the second page if possible . When the table is about the right width but runs much longer than a page, it can be continued on the following page . This is frequently done if the table would run to more than one and one -half pages . Large reference tables, such as those that show procurement or inventory data for several hundred items , are often continued for many pages . Sometimes these large tables are both long and wide. Some of these must be spread across facing pages, and continued for several pairs of pages . But even that space is not always enough . Some 11 continuedu tables are so wide that they must be reduced in reproduction to fit even the double-page width. Reducing Oversize Tables . Many oversized tables are difficult to 11 break. 11 For ex.amp::.e, an inch or two too long, is not long enough to carry over to the next page . A table that is a little too wide is seldom worth spreading to fill the facing pages . In fact, most oversize tables fall into this classification -they will not fit on one page, yet are not worth more than one . The only solution -after all methods of squeezing and rearranging have been considered and found unsatisfactory -is to reduce these tables in reproduction . The overall page dimensions of most Army statistical reports reproduced from typewritten 811 x 10 l/211 or 911 copy are x 13". The accom- STANDARD SIZES FOR FULL PAGE OVERSIZE TABLES (Before Reduction) Final Page Size : Final Page Size : 811 911 1311 X 10. 5" g/ X 91 Percent Reduction Width Length Width Length (Spaces) (Lines) (Spaces) (Lines) 10 87 5L, 93 68 16 93 58 100 73 20 98 61 105 77 25 104 65 112 82 30 lll 69 120 88 36 122 76 132 96 a/ Type page after reduction is approximately b. 511 X 8.1" . 91 Type page after reduction is approximately 711 X 10 .211 • panying table shows the width and length (in elite typewriter spacing) of the oversize originals for each of these final page sizes . These dimensions are for full page tables (oversize in both directions) . For tables that are oversize in only one dimension, it is important to remember that reducing any dimension reduces the other in the same pr oportion. A table that is too wide but not too long, reduces to less than a page long . This usually is not a problem if other material is available to put in the remaining space . A table that is too long but not too wide , reduces to less than a page wide. This is undesirable unless the page can be reduced to exactly half a page wide . The standard procedure for correcting this problem is to spread the table width so that it will reduce to exactly the page width. These standard page sizes provide for the tables that range from a little oversize to approximately a third oversize . It is desirable to reduce the table as little as possible . If there are several oversized tables that would be about the same size , however, make them all exactly the same . This simplifies the reproduction process , and also provides a more professional looking result than different reduct ions for each table. Titles of oversized tables generally should not be typed on the same sheet as the table . Instead, the title should be typed on a separate sheet of paper (if the table is the only material on the page) or on the sheet with the accompanying text (if the table is to appear on a page containing other material) . This keeps the Litles of all tables in a report the same size after reproduction. PREPRINTED TABULAR FORMS The standards of tabular presentation stated in this chapter were designed primarily for use in the preparation of pages that are prepared entirely by the typewriter (and, when needed, the ruling pen). These standards may be applied to a considerable extent, however, to printed or partially preprinted tables . In any case, the title, column headings and stubs should be prepared as typewritten copy in such a way that the material fits the page reasonably well before it is set in type or preprinted by some other process . Consideration should be given to the size and face of the type to be used and to page dimensions and the amount of reduction, if any. Preprinted titles, column headings and stubs are suitable for a large number of similar tables, or tables that are prepared in the same manner at recurring intervals . The preprinted material may be produced by the use of letterpress, if volume is sufficient PREPRINTED TABULAR FORMS to make it economical, or it can be done from typewritten masters by the offset process, or by specialized equipment utilizing punched paper tape . (It is also possible to preprint chart titles and grids to speed up final preparation of the report.) The figures may be entered on preprinted forms by typewriter, electrical accounting machine, or by electronic computer. When sufficient effort is spent, it is possible to program the computer so that totals and sub-totals appear above and to t he left of the i terns they summarize • CHAPTER 4 GRAPHIC PRESENTATION GENERAL Graphics have thes e f our important uses: 1 . To emphasize the main numer ical facts; 2 . To uncover facts that might be overlooked in text or t abular presentat ion; J. To summarize a cumbersome mass of tabular dat a ; 4. To add variety to text or tabular material, thus making the report more interesting to l ook at and easier to read and under stand. In statistical presentat ion, the first two uses are the chief r easons for using graphics and the third and fourth are usually by-product s. The fourth use, by itself, is s eldom sufficient reason for using graphics. To most people who work with statist i cs, graphic present ation means charts. This is because charts are much more widely useful (for presenting statistics) than maps or diagrams. For this reason the discussion about graphics in this manual is primarily about charts . Maps are discussed briefly (on page 64), and diagrams not at alJ.. KINDS OF GRAPHICS ••• used in statistical pr e sentation CHARTS: (amount r elationships) show HCM MUCH MAPS: (geographical r elationships) show WHERE DIAGRAMS: ( s chematic r elationships ) show HCM For showing statistical data, charts are usually the best, maps occasionally (see page 64), dia grams seldom. To use a chart effectively, three basic steps should be f ollowed : 1. Sel ect the one specific type that will be most appropriate; 2 . Design it to focus attention on the specif ic point the chart is intended to emphasize; ). Construct it in such a way that it is easy to r ead and understand. These three steps seem obvious . As principles , they are quite simple ; but it is not always easy to .put t hese principles into practice. The first part of this chapter discusses step 1; the PROS AND CONS OF GRAPHIC PRESENTATION Advantages 1 . QUICK : Shows the main f eature of the data at a glance . 2 . FORCEFUL : Carries much more emphasis than text or tables. J . CONVINVING: Proves the point instead of mer ely stating it. ·4. COMPACT: Packs a l ot of information in a small space . 5 . INTERESTING : Easier to look at than text or tables . Disadvantages 1 . TECHNICAL: Some r eaders not familiar with st atistical graphics. 2 . DEMANDING : Takes special "know how" to design effective graphics. ). COSTLY: Takes mor e time; also skilled help. 4 . NOT ALWAYS USABLE: Some data are not suitable for graphic presentati on . 5 . NOT PRECISE : Cannoc be quoted like text or :.ables . next part ("Planning and Designing Chart s" ) discusses step 2; the third ( "Chart Const ruction" ) discusses step J . Sel ecting the Right Kind of Chart In order to select the right type of chart , the designer must know exactly what he i s t rying to bring out of the data. Regardless of the general subject or of the general ki nd of comparison planned, a decision must be made as to just what specific point is being presented. Haziness her e may l ead to the s el ection of an entir ely wrong t ype of chart f or the purpose, or to the choice of a t ype that is unnecessarily ineffective. This problem is subjective, and can be solved cnly by the person r esponsible f or the chart . The draft sman, however exper t , cannot be expected to know enough about the purpose of the presentation, the characteristics of the data, or t he Lature of the audience, to decide this questioL himself . The designer may be cer tain about the speci f ic point he wants to get acr oss, but he cannot go far without knowing what kind of charts are available to choos e from. He needs to know what each of the basic chart t ypes can and cannot do, and how to modify each to adapt i t to the speci fic data to be emphasized . More can be done to fit the chart to the nat ure of the data than may be r eal~zed when the designer first encounters t he problem. Caution: Watch f or two pit falls in select ing charts do not limit them to one or two f avorite kinds, and do not make them so compli cated and obscure that no one else can underst and them. CHART COMPONENTS also that the amount scale runs horizont ally CHART COMPONENTS and that there is no time scale. Thi s page shows what the different parts of a chart are and tells what each part is called . A f ew additional terms are used i n other types of chart s; but these are explained later, The top illustration is a t ypical line or in the discussion about thes e types. 11 curve11 chart. Many charts shown in the next section on 11 Chart Types 11 f ollow this general arrangement. Column charts, which belong to this The terms used f or both curve and column same f amily, have a similar framework but make charts are those which were officially approved use of vertical bars instead of curves. by the Amer~can Standards Association in its American Standard 11 Time-Series Charts11 (ASA The lower illustration is a typical bar 215 . 2 -1938, r eaffirmed, 1947) . A r evi sion of chart . Although this is an entirely different this st.andard (ASA Yl5.2 -1960) published by and the Amer ican Society of Mechanical Engineers, kind of comparison, most of its components their names are the same as f or curve charts . 29 West. 39th St . , New York 18, N.Y., generally But note that a bar chart has a dist inguishing f ollows this t.erminology. Neither of these st anf eatur e : a list of i tems down the side. Not e dards covers bar charts. r Subtttle TOTAL ARMY CIVILIAN PERSONNEL !Y~ U. S. CITIZENS AND NON-U. S. CITIZENS THOUSAND Curves ( 400 Amount Scale Caption Vertical 300 Rulings AMOUNT 1 OuorterlSCALE ~ Ticks Monthly Ticks Amount Scale Numerals Time I Designations Horizontal \. Rulings J S DMJ S D MJ S D ',1 FY 19Xl FY 19X2 FY 19X3 / ---~---./ '------- ------- Bose SY PAID FR01>4 II FUNDS TIME SCALE Zero Line QJ PREVIOUSLY' CALLED UNGRADED CALLED GRADED /£./PREVIOUSLY Reference Symbols NET OBLIGATIONS. INCURRED FOR Bar MILITARY FUNCTIONS Outline Labels ~~T~MA 0~~""~~~~~~~~~~~~~~ 10 Caption I I ITEM 8 I STUB ~ YITEM c ~; ITEM D Item \. Shading GRAPHIC PRESENTATION CHART TYPES The following pages show typical examples of the four main kinds of charts used in s~atistical presentation. These are : curve charts, sur face charts, column charts, and bar charts. CURVE CHARTS In statistical work, a "curve" is any line that connects a series of plotted points. Cur ves can be used in many kinds of charts . Mos t of the curve charts used in the review and analysis of Army programs are time series charts. Such charts show quantities measured at specif ied times, such as procurement or strength by month. Frequency distribution charts are used EXAMPLES OF 11 CURVES 11 CD ------ STRAIGHT -LINE _ ----- --------._.. SIMPLE SLOPE-CURVE CHART CAPTION 80 6~--------------+-~~--~------~~ 2~--------------~--------------~ SIMPLE STEP-CURVE CHART CAPTION 5 0 4 r- 3 Fl I L--. 1--1 2 0 I I J F M A M J J A S 0 N D occasionally in Army analysis to show data in specif ied categor ies at a particular time, such as of ficers by age group on a given day (see page 60) . Time series are discussed below. Curve charts are the most widely used form of chart presentation: they can be used for the simplest comparison or for the most complex; they are extremely flexible and can be adapted to a wide range of uses; they are compact, easily carrying much more data than any other type of chart; they are easier to draw using the same data -than other types . Curve charts should generally be used when many points are plotted, several series compared, or emphasis is on movement rather than on actual amounts . (I2 As shown in Chart 1, curves can follow almost any pattern from a straight line to a zigzag. In most statistical presentation, they have a pattern somewhere between these extremes . A SIMPLE CURVE CHART shows amount changes in a single series of data. The curve can be eit her the slope type (2) or the step type (3), ·depending upon the nature of the data and the purpose of the chart . 0 A SLOPE CURVE is made by drawing a line directly from each plotted point to the next. Slope curves suggest that changes f rom point to poi nt are gradual or continuous, and therefore are usually the best way to show data that have a "carry over" from one time to the next. This type of time series is sometimes called "as of" dat a because it measures how things stand at points of time. Typical examples : month-end 1nventories, strength, unliquidated obligations. However, when fewer than 4 or 5 time points are plotted, a column chart is usually better. A slope curve can be used also for "during" dat a (measurement for periods of time, such as production, personnel separations, expenditures, etc.) although other forms of presentation are oft en better for this kind of information. (See next paragraph and "Column Charts," page 52 . ) (i) A STEP CURVE is made by drawing a norizonti:( line through each point and connecting the ends of these lines by vertical l ines . It is a good form of presentation for showing averages, or other measures that apply over periods of time . It is especially good for data that change abruptly at ir.regular in"Lervals, such as allocations of f unds, personnel ceilings, etc . Step curves are better than slope curves for showing "period" data, and are oft en used for this purpose, especially when the time series is a long one . (See also "Column Charts," page 52, and "Connected Columns," page 91.) CHART TYPES ~ A MULTIPLE SLOPE-CURVE CHART brings togeth er two or more related curves . The curves may be interdependent, such as a total and its com ponents; or independent, such as two or more totals; or nondependent, such as an actual re sult compared with a forecast or estimate. But whatever the purpose, this type of chart is bet ter if it does not compare too many things at once. A chart with four curves is not necessar ily twice as useful as one with two curves; in fact, it may be less useful, especially if the curves criss-cross or run together. Where many curves are to be shown, or several curves have the same magnitude, it is better to use two or more charts with the same scales rather than a single chart (see "Uniform Scales, " page 81) . ~ A MULTIPLE STEP-CURVE CHART serves the same general purposes as the multiple slope-curve type, but its use is much more limited. The reason for this is that step curves that cross are hard to f ollow, especially if they cross back and forth several times . Two or more step curves can be shown on the same chart, however, if they do not overlap or i f they cross cleanly, as in illustration 5. (See also "Combination Types, " pages 58 and 59. ) ® A CUMULATIVE CURVE CHART presents a running total. It should be used where the primary interest is in the cumulative picture rather than in the amounts for each period. Each point on a cumulative curve shows the aggregate to date : the total for the current period plus all earlier periods. It is therefore a good f orm of presentation for comparing results "so far this year" with those for the same period last year, or with an annual objective or goal such as the obligation or expenditure of funds . The process of adding the results for each successive period tends to smooth a cumulative slope curve. It is used mostly for data in which the aggregate can stay the same or in cr ease but can never decrease. If the nature of the subject does not obviously require cumula tive data, be sure to label the chart "Cumula tive from , "or "Cumulative Data" as shown in illustration 6. The cumulative idea can be used for step as well as slope curves . The usual limitations of s tep curves apply here too . A CUMULATIVE DEVIATION CHART, like any cumulative chart, shows the "total to date" at any point on the curve. This type, however, measures cumulative differences ~r deviations (for ex·ample, net gain or loss in strength, or cumulative deviation from budget or allowance). This kind of cumulative curve, therefore, can go down as well as up (when the net change i s a decrease instead of an increase) . See also how this idea can be used in combination with a column chart ("Combination Types, " page 59) . MULTIPLE SLOPE -CURVE CHART CAPTION 5 CD 4 .· ... ........ CURVE A ~ .. ....· 3 "V' / .... A........ .. 2 . ... CURVE B : \_, ...··...· ~ I 0 I 1 J s D M J s D MULTIPLE STEP-CURVE CHART CAPTION IOor---------------.-------------,-----, CD r-..--.r-,__ : CURVE B 25 ~~.--~~--;----------r--------~~-;--t==-_, _... D M J s D M CUMULATIVE CURVE CHART CAPTION100 .-------------,----------r (CU,.ULATIVE DATA) OBJECTIVE o~_.--~~--~--~~--~--~~--~~--~ D J F M A M J J A S 0 N D CUMULATIVE DEVIATION CHART CAPTION 20 .--------------------.-------------------, CUMULATIVE UNDER SCHEDULE OJ FM AM J J ASOND GRAPHIC PRESENTATION In addition to the kinds of curve charts shownonpages 46 and 47, several others are very useful for presenting statistics. Some of these charts are shown on this page and on the page that follows. These all have one thing in common: they require a special kind of scale or special scale treatment. Most of these special-scale charts vary widely in purpose and use. Some of them are useful primarily as analytical "work" charts and should be presented sparingly to readers not REPEATED TIME-SCALE CHART CAPTION 5 CD .... ......... ,_ YEAR A "',.,.." - o~~~--~~--~~--~~~~~~~ OJ F MAMJ J A S 0 N D MULT I PLE TIME -SCALE CHART CAPTION 400.-------.--------.------~~----~ 0 01..-------~------~------l....--------J 1941 1942 1943 1944 1945 1950 1951 1952 1953 1954 MULTIPLE AMOUNT-SCALE CHART CAPT ION A CAPTION B 5 100 CD 4 l--------------t----------------1f-------l 80 3 60 2 40 CURVE B 20 0 0 D M J s D M accustomed to them . Others are seldom appropriate, but extremely valuable when they are. Some of these examples are not recommended f or general use in statistical reporting but are included in this manual to assist the analyst in discovering basic relationships between figures. Taken together, this group of charts demonstrates how important it is and how possible it is to select the one particular type of chart that will focus on the key facts that should be conveyed to the reader. {1) A REPEATED TIME-SC.A.LE CHART is a method of ~wing a l ong curve as a series ofshort curves . For example, a curve measuring monthly data over a period of several years can be shown as several one-year pieces plotted on a one-year time scale . As illustrated in Chart 1, this method permits easy comparison of comparable time points . It is especially useful : (l) for comparing monthly data that have a seasonal pattern (generally high at certain times of the year, low at others), for example: sickness, AWOL rates, civilian personnel separations; (2) for comparing results "so far this year" with results for the same period last year on some earlier year (cumulative data) . 0 A MULTIPLE TIME-SCALE CHART foll ows · the same general scheme as a repeated time-scale chart, but compares non-repeating time periods . As illustrated in Chart 2, this kind of present ation is used to compare results or conditions during two similar historical periods . Although an extremely valuable chart when used properly, this type is much more difficult to design than a repeated time-scale chart. The main problem is in deciding how to match the periods . For example, in comparing data for WW I and WW I I, the time scales could begin at 1917 and 1941, 1914 and 1940, or 1914 and 1939 . Depending on the subject, any of these could be used . ~ A MULTIPLE AMOUNT-SCALE CHART brings together for cl ose comparison two or more curves that are measured in different units or are measured in the same unit but are so far apart in size that they would be difficul t to compar e . Preferabl y not more than two amount scal es should be used . If the aim is to compare change or growth, the zero line of both scales should coincide (either on or off the chart), and the scal e intervals should be so selected that both curves meet . (This purpose can often be served better by converting both series to a common base, thus requiring only one scale. For example : index numbers, or percent of average for the period . ) If the purpose of the chart is to compare the pattern of two or more curves -without regard for their size -the scales need not coincide at zero . Caution : many readers do not understand the characteristics of this type of chart and will misread it (see also page 65 for a misuse of multiple scales ) . CHART TYPES SUPPLEMENTARY AMOUNT-SCALE CHART ~ A SUPPLEMENTARY AMOUNT-SCALE CHART provi d es THOUSA D t wo ki nds of measurement on a single chart . 100 CD Fir s t , it measure s the variations in a s er ies of dat a, just l i ke any simple curve chart . Ther, in a ddition, it measures t he size of this s er ies i n r elat i on to another s erie s . For exampl e, a curve measuring officer st r engt h could be plotted again s t a ser i es of supplemen t ary curves that me a sure 6 , 9 , 12 , and 15 p er cent of total milit ary str ength (Chart 4) . Of course this t ype of chart can be used f or other ki nds of dat a : f or ex ample , s tock on hand r el ated to all owances, beds occupied r e lated to beds available , or stor ag e space oc cupied r el ated t o total usabl e space . J ~ AN INDEX-SCALE CHART shows data that have been conver ted i n t o per centages of a base val ue . INDEX-SCALE CHART For example , produ ct i on could be s hown ea ch month as a per c entage of the l eve l of January INDEX 1965 . Thus , February mig ht have a val ue of 200 UNE 19Xl : 100 102 . 3, Mar ch 105 . 7, etc . The principal use of i ndex chart s i s f or s howi ng "composite" data 150 ./"--such as price indexes. Howe ver, they c an be CURVE A CURVE B used also f or comparing t wo or mor e s erie s of 0 -"" da t a t hat are measured in differen t kinds of 100 .........-:: - unit s (such as wor kload and s t r eng th) or in di f ::;:::::..-' f er en t size units (such a s a total and on e of its component s ) . 50 NOTE : Excep t when working with s tandard econ omi c i ndexes such as t hos e f or prices, wage s , etc . 0 I I I I I I I I I t he kind of comparison shown on an i nd ex chart D J D J D J D may be clearer i f pre s ented a s s i mple per cen t ag e 19Xl 19X2 19X3 dif f er ences . For example, "Per centage incr e ase i n unit cos t s s i nce 1965 " shows ex actly the same picture a s "Ind ex of Un it Cos t s -1965: 100", ex LOGARITHMIC AMOUNT-SCALE CHART cept that the scale is di vided t o r ead 0 , 25, 50 CAPTION instead of 100 , 125, 150 . 10 CURVE A ./--/~ (D A LOGARITHMIC AMOUNT-SCALE CHART shows tne 8 6 / ./r-.....""" ............. ../' / r elativ e change f r om on e t ime to the n ex t . On this ki nd of chart a constant incr eas e of (say) / / ~......'l 10 percent each month would show a s a s t r aight 4 v ......... v line (slant ing up) : on an arithmetic-scal e / / chart, a const ant 10 per cent incr eas e would be a //VCURVE B curving line , slanting up bu t a t a s teeper and s teeper angle . 2 ~ Chart s l i ke I llust rat i on 6 are good f or comparing r ates of change . They can show whether in produc t i on are keeping pace wi t h change s 19Xl 19X2 19X3 19X4 19X5 19X6 change s i n cos t , whether changes in a small or ganizat i on are r e l ativ ely mor e or l ess than change s in a large organization , etc . ( See also MULTIPLE-LOG AMOUNT-SCALE CHART page 75 . ) CAPTION A CAPTION B 10 100 C2) A MULTIPLE -LOG AMOUNT -SCALE CHART shows the 8 ........... eo same type of r elat i onship as chart s wi th only on e ............. of thes e scales , but permi ts a comparison of 6 60 ....... CURVE A dat a t hat are me asured i n differ en t unit s ( or that are measured in the same unit but are much 4 ............... ....--40 ....... ~--"""" differ en t in size) . ~............ ----v / ~ --............... ....... NOTE : This t ype of chart g enerally shows r.el a s hows CURVE B 20 t i ve change f r om on e time to t he n ex t . I t 2 ........ how the r elative (per centage) changes i n on e s eries compar e wi t h the r el ative changes i n an other . I t s hould be used when that i s the pur CD "' 10 pos e of the chart , and not mer ely as an easy way to bring t wo curves c l os e together . 19Xl 19X2 19X3 19X4 19X5 19X6 49 GRAPHIC PRESENTATION SURFACE CHARTS area chart, each amount is shown as an area (theproduct of two dimensions); in a surface chart Surface charts are shaded curve charts. each plotted point represents a linear distanceSome s erve the same general purpose as similar (a measurement in one dimension). Shading thecurve charts; some perform duties that curve space below the curve connecting these point s charts cannot do at all. Some are extremely produces a shaded area; but not an area chart. simple, some not so simple . Several types of (For a discussion of area charts and why they surface chart are among the mos t useful charts seldom are suitable for s tatistical presentation, for showing time series data; but they are also see page 62, "Graphic Types Not Recommended.") among the easiest to use improperly. Broken scales . should never be used on surThe first thing to note about surface facfr charts because they distort the picture and charts is that they are not area charts : in an can giv e a false impression of the data. SIMPLE SLOPE -SURFACE CHART (I) A SIMPLE SURFACE CHART is merely a singlecurve chart with the space between the curve andthe base line shaded to form a surface . Thist ype of chart emphasizes the size of the totalamounts rather than the differences or changesin amounts, and is often used to make a simplegrowth curve look more impressive. Chart 1 is asimple slope curve that has been converted intoa surface chart. o~~~~~~~~~~~~--~~--~~ D J FMAM J J AS 0 N D SIMPLE STEP-SURFACE CHART (!) Chart 2 shows the same t reatment applied to a step curve (see page 46 for use) . This form CAPTION of s tep chart is similar to the connected column 500 0 chart and can be used in place of it when the time series is a long one (see "Connected Col400 umns," page 91). 300 200 ~ A SUBDIVIDED SLOPE-SURFACE CHART shows howthe component parts of a time series combine to 100 make the total. This type, which is also knownas a layer chart or st rata chart, is widely used0 in s t a tistical presentation.J F M A M J J A s 0 N D Only the bottom layer of a subdivided surface chart is measured directly from a fixedSUBDI VIDEO SLOPE-SURFACE CHART base. All other layers are measured from a CAPTION changing base, and their size can be read only 5oor----------------r--------------~ indirectly. For this reason, it is usually bestto put the most important component at the bot tom . However , the size and shape of the bottom CD 400~=-------------~--~--~~7.+~+-4 layer have such a great effect on the other layers, that it may not always be best to put themost important component first : if this or anyother layer fluct uates sharply, it should be puton top. This problem is discussed further onpage 73 . 100 Although subdivided surface charts can bevery effective, the weaknesses of this type ofpresentation call for caution in using it (see0 "Development of Charts from Tabular Data," pageD J F M A M J J A S 0 N D 68, for suggested alternatives). 50 CHART TYPES C±) A SUBDIVIDED STEP-SURFACE CHART serves the same general purpose as the subdivided slope type, but is a better form for data that move up or down sharply. When there are only a few time plottings, step presentation is effective for showing "during" data, as against "as of" data. It is especially good for showing such measures as a monthly average during the year, daily average for the month, etc. Like all step charts, this one is an intermediate type between slope curves and columns and has some of the advantages of each. ~ A 100 PERCENT SURFACE CHART is a subdivided surface in which the layers are shown as percentages of the total. This type of chart shows the change in the relative size of the components, over a period of time. Like all percentage relationships, this one can be misleading if the absolute amounts the percentages are based on are not fairly stable. For example, if 100 percent represents a sharply rising total, a decreasing percentage may actually represent an increasing amount . In such cases, there is a special advant age in picturing the absolute amounts in an accompanying chart or t able. This type of chart can be presented in step as well as in slope f orm. One of these types is about as good as the other, but the step type is better if the layers change abrupt ly in either direction or size. CD A BAND CHART compares two closely related curves and emphasizes the difference between them. It is useful for showing the spread between such things as cumulative expenditures and obligations, cost of an item before and after addition of equipment, or a tot al excluding and including a certain category. Note that in each case one measure must always be greater than the other. This type of chart is one form of a range chart. The range-column chart on page 53 is another . Both serve the same general purpose of showing differences between two series as well as their absolute magnitudes. (J) A DEVIATION-CURVE CHART compares gains and losses and emphasizes the difference between the two . It is a useful type of chart for showing inward and outward movement such as receipts and shipments, income and expenses, personnel accessions and separations, etc. Like the band chart, this chart shows two curves with t he space between them shaded to show how much one exceeds the other. But in this case the curves can cross : so the difference between them can have two meanings: a net gain, or a net loss . Contrasting shadings are used to set off the net gains from the net losses. Other ways to show this kind of comparison are illustrated on page 53, "Column Charts" and page 59, "Combination Types" (Chart 4). 214-853 0 -66 -4 SUBDIVIDED STEP-SURFACE CHART OJ FMAMJ J ASOND 100 PERCENT SURFACE CHART eo 60 40 20 0 0 J F M AM J J AS 0 N D J F M BAND CHART MILLI ON $4 CD DEVIATION-CURVE CHART CAPTI ON 4 0~-&--~--~~--~--~~--~--~~--~ J M J s D GRAPHIC PRESENTATION COLUMN CHARTS In column charts amounts are represented by the height of vertical bars or "columns." Column charts are generally used to compare data for a given item at different times . Horizontal bar chart (discussed on pages 54-57) are generally used to compare data for different items at the same time. Column charts are more suitable for data on activities or events that occur during a period of time (s uch as procurement during April) than for figures indicating status on a given date (such as inventory or strength on 30 April). Figures indicating status may be shown as column charts but usually they can be presented better SIMPLE COLUMN CHART CAPTION 4 0 J F M A M J J A S 0 N 0 GROUPED COLUMN CHART CAPTION 4 0 3 2 JAN FEB MAR APR MAY JUN SUBDIVIDED COLUMN CHART CAPTION 4 3 2 J F M A M J J A s 0 N D by curve or surface charts, which are discus s ed on pages 46-51 . A column chart makes a stronger picture than a curve showing the same data and so is especi ally good for showing growth or development in a striking manner . Columns are usually better than a curve when there are only three or four time pcints to be plotted, or when the series fluctuates very sharply. They are not the best form for comparing several series of data, for showing a long series of data with a great many plottings, or for showing numerous components of totals . (See also : "Planning and Designing Charts," pages 73 and 74, and "Chart Construction," page 91.) A SIMPLE COLUMN CHART consists of a series of vertical bars, each extending from the base line to a plotted point . This type of column chart and all types developed from it -must have a zero line . Since the entire height of the column is needed to show the size of the amount the column represents, the top end of the col umn should never be broken except for very occasional "freak" values . (See also : "Broken Columns," page 89 . ) (I) 0 A GROUPED COLUMN CHART is useful for comparing two (seldom more) series of data . This type is most effective when used for series ·that differ in level or in trend . To set off each group of columns from the next, the spacing be t\veen groups should be at l east as wide as a column (see page 90) . The two sets of col umns should have shadings that contrast sharpl y in tone . Sometimes this type of chart is drawn with one column in each pair overlapping the other; but tjis treatment should be used only when needed to save space and when the front set of columns is always shorter than the back set. (See "Overlapping Columns, " page 91 . ) (!) A SUBDIVIDED COLUMN CHART shows the component parts of a series of totals . This type of chart -also known as a segmented column chart serves the same purpose as the subdivided surface chart (pages 50-51) but is a better choice when the plotted values fluctuate sharply from one period to the next . There is a tendency to show too many components in this type of chart. When more than three or four segments are used, individual pieces are difficult to identify and to compare . I f a l arge number of component parts must be shown, the subdivided surface chart is usually better. (See al so : "Subdivided Surface Charts" page 73 . ) As a general practice, do not connect segments in adjoining columns by diagonal lines; such lines usually do more to confuse than to cl arify. CHART TYPES ~ A DEVIATION-COLUMN CHART shows differences between two series. Such charts generally present negative values plotted below the zero line as well as positive values plotted above it. In a deviation chart a colurrm can extend either above or below the base, but not both. This type is good f or showing how much results varied from an estimate or requirement . I t is especially useful f or measuring net gains and losses, and all similar plus-or-minus differences. Chart 4 shows a typical example . ~ The significance of net r esults is of ten clearer when they are compared directly with the gross figures from which they are derived. For example, in Chart 5 colurrms extending up from a base line might represent such data as receipts and those extending downward might represent shipments. The difference between each pair of colurrms is the "net", which appears as an overlapping deviation colurrm, above the zero line if receipts exceed shipments, below if shipments exceed receipts . This type of comparison can be shown also by a colurrm and curve combination (see Chart 5, page 59). (!) A FLOATING COLUMN CHART is a special variation of the subdivided colurrm type . The tot al length of the colurrm r epresents the tot al of two main components, one of which is plotted above the base line and the other below (often considered to carry an unfavorable connotation) . For example, the total length of the colurrm might r epresent the total number of contracts completed during a month, the portions above the base line indicating the number completed on or ahead of schedule and the portions below the base line indicating those completed behind schedule. This type of chart differs from t he deviat ion type in that each colurrm may extend both above and below the base. In effect, floating colurrms are subdivided colurrms with the dividing line between segments used as the base. The over-all lengths of the colurrms are subordinated t o the comparisons of the segments . (I) A RANGE-COLUMN CHART emphasizes the differences between two sets of values. In this type of presentation the highest and lowest values during each period are connected by a colurrm measuring the high-low range. Supplement ary inf ormation, such as the average value, can be r eadily included by means of cr oss-lines or other indicators. This type of chart is useful f or showing the spread in such data as personnel strength, inventories, prices, etc . As with all charts measuring r ange (see the surface chart and bar chart versions of this kind of comparison, pages 51 and 56), this type can be used eff ectively to show a supplementary range such as high and low tolerance limits, upper and lower levels of efficient operation, or other top and bottom "bench" marks . These are usually put in as light dash or dot lines across the entire chart . DEV IATION-COLUMN CHART CAPTION 100 --------------t 7: POSIT IVE 50 -----------i; ---· 0 ~m-~~~~~~~-~~~~~~~~~ :;:~: ~~~~~ ::::: :::: ===== Im;J :::: ~ ~~~~-~~~~~ w 11111 ·50 -l:J-------tl-1-1 NEGATIVE -100 ::::: ~ -----------1~1-------------------------- J F M A M J J A s 0 N D GROSS AND NET DEVIATION-COLUMN CHART CAPTION 40 FLOATING COLUMN CHART CAPTION 150 -----------~------ ® FAVO~BLE ,.,., 100 --..,,.,..-----t: -:::: ~'""' r-·-: .. 7: ... S:: :::: 50 -:::: -::::: .. :::::r-:: >-:::+ ::::: J F M A M J J A 5 0 N RANGE-COLUMN CHART CAPTION 5oor--------.----.-----.-------. 300 J F M A M J J A S 0 N D GRAPHIC PRESENTATION BAR CHARTS In a bar chart, amounts are represented by the length of horizontal bars . Bar charts differ from curve, surface, and column charts in that they have only one scale. This is an amount scale, and it measures across the chart; the vertical dimension is used for listing the items measured. The horizontal direction of the bars presents a direct contrast to the vertical arrangement of columns, and thus shows the reader at a glance that he is not looking at a time series chart. All bar charts (except one type, shown on page 56) need a zero line or other base line. The bars should not be broken except to avoid a long "freak" bar (see page 89) . SIMPLE BAR CHART CAPTION CAPT IONO 2 3 4 5 BAR-AND-SYMBOL CHART CAPTION CAPTIONO 2 4 6 8 10 ITEM C Wfi W'A 0 ITEM 8 ~~ Vff///~ ITEM E = =t -~ ITEM G :v~ ----i II ITEM A ~ --~EARLI ER Whl ITEM D MONTH MONTH ITEM F 0: . I I I I SUBDIVIDED BAR CHART CAPTI ON CAP TIO NO 5 10 15 20 25 ITEM A ITEM B ITEM E Bar charts may be used effectively to show how several items differ from each other in one or two characteristics, or to show how several items differ from each other in the distribution of their components. The order of the items in a bar chart is very important. Many kinds of i terns can be listed in several different ways, each giving a different emphasis to the data (see pages 22-25 of the Tabular Presentation chapter and page 71 of this chapter). The thickness of the bars has no measurement value and so should be uniform throughout the chart. A further discussion of bar chart treatment appears on pages 71 and 72. (!) A SH1PLE BAR CHART is mer~ly a series of horizontal bars drawn to the r1ght of a common base line. Each item may be plotted according to its absolute value, or may be expressed as a percentage of an appropriate total, goal, average, or other standard; for example : actual as a percent of scheduled deliveries. A single shading should be used for all coordinate bars, but a bar used to show a diff erent category, such as a total or average, may be set off from the other bars by a different shading or by additional space between bars. 0 A BAR-AND-SYMBOL CHART is a simple bar comparison with a secondary comparison added. The additional data, such as programed amounts, forecasts, averages, plans, or other standards, greatly increase the value of the chart. The symbols can be used also to show data for some earlier time : for exampl e, the preceding month or year, or the best previous performance. Thus this type of chart can often be used to summarize several time series charts. Adding two sets of secondary data is seldom satisfactory~ecause the symbols usually intermingle. However, if one set always represents smaller amounts than the other, two such supplementary comparisons may be effective. Different symbols may be used in this type of chart to emphasize different comparisons : to emphasize the bar comparison, use verticalcrosslines (as in Chart 2) or some other subordinate device such as a hollow diamond; to emphasize the comparison of each bar and its symbol, use black di~onds, filled circles, or some other prominent device. (The cross-line is the best for gener~l use: it is easy to make, and marks the pl ot ted value more precisely than f atter symbols. See also page 71 f or further discus sion and illustrations. ) ~ .A SUBDIVIDED BAR CHART shows each bar divided into its component parts. This type clearly shows how much effect each component has on the size of the total. Usually the largest or most important component is put next to the zero line. As in subdivided column or surface charts, only the component that starts from the base can be measured directly from the scale. CHART TYPES 0 A 100 PERCENT BAR CHART shows each bar divided according to the relative size of its components, regardless of the absolute size of the total value of the bar. It emphasizes the proportionate part of the total contributed by each component. This form of subdivided bar has the advant age of having two base lines: zero and 100 percent; and so provides direct comparison of the components at either end of the chart. Like other percentage comparisons it should be used ~ith caution when there is a wide difference in the absolute amounts on which it is based. 0 A GROUPED BAR CHART permits the comparison of a number of items in two respects at the same time. Although the bar-and-symbol treatment can be used to present this kind of information, grouped bars are better, when space permits, because they suggest the nature of the comparison better. Groups of more than two or three bars are seldom desirable because the eye i s confronted with too many comparisons at once . In designing this chart, the category of major interest is usually put first and given the darker shading . For spacing, the standard is to make the space between groups no less than the thickness of a bar. (An easy way to lay out this chart is to make the space between pairs the same as each bar. ) See also "Overlapping Bars, " page 91. (!) A PAIRED BAR CHART is another way LO compare a number of items in two respects. I t is pr ef erable when different unit s or scales must be used for each category. Instead of being grouped, the bars are put opposite each other, one set extending to the left of the item names, one to the right . In this chart, the bars at the left might represent cost per unit and those at the right the number of units produced; or number of motor vehicle accidents per 100,000 miles and number of miles operated. Bars extending to the left (like columns extending below the zero line) tend to suggest unfavorable results or conditions . If Lhe na ture of the data permits, the chart should be designed to take advantage of this. (J) A DEVIATION-BAR CHART also presents bars extending to the left and right; but in this case each item has only one bar, and it goes to the lef t or right. This type of chart is useful for comparing dif fer ences between actual results and a program or st andard, especially when these differences are so small that they would be hard to compare on a chart showing the totals. Chart 7 might show any kind of "more-thanless-than" data, such as amount over or under allowance , percent over or under estimate, etc . Like any deviation chart, this type is effective also for showing other net measurements such as net gains and net losses, or net ins and net outs . 100 PERCENT BAR CHART CD PERCENT CAPTION 20 40 60 100 ITEM A ITEM 8 ITEM c ITEM D ITEM E ITEM F ITEM G GROUPED BAR CHART CAPTION O 5 CAPTION 10 15 20 I I I ITEM A t::::;:;;:;~::::~:t/::::::::::::;;::~::~:t~~~~~::::t::::::;;:::;::~:~:~~t~::::·::}:t:::::~ MEN I I I WOMEN ·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:::3 ITEM 8 . .'.' -:-::::::::::::::::::<<:-:-.... ::<::::::::::::::::::::::::::::::::::::::::<::::::::::::::::;] 1 T :::::::::~:::::::::::::::;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:::::::t ITEM C ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::<::::::::::::::::: 1 1 1 !•!·!·!·!·!·!·!·!·!·!•!•!•!·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!·!· ITEM D 1 1 :·:·:;::·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:~ITEM E ::::::::::::::::::: : :::::::::: : :~l -T I I I PAIRED BAR CHART CAPTION B 2 3 4 I ITEM D F~=~=~= ITEM A p.w;.;.;.F~=.:.:.:.t:.:.:.:.:J E:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:; ITEM C ...-..:.:.;..;.;~~=.:;:;.:,:;.;..;;;;~ ::::;:::;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:; ITEM E I I ~~~~~~~ f::::::::::;:;:;:;:;:;:;:;:;:;:;:;:;:; ITEM G J.:.:.:.:.:.:w:;~~~ DEVIATION-BAR CHART CAPTION CAPTIONCAPTION _ 3 0 4 "i -, ,,,,,,, '' t ,,,,,,,,, 't,,,,,,,,,,,]l ITEM A ITEM B If:::::::::::::::::::::::::;.·:: :::::::::::::::::::: ( FAVORABLE) ITEM C ~:::::::::::::::::::::::::::::::::::::::::::::::::::::::: ITEM D (UNFAV ORiBLE) :;:;:;:::::;:::;:::;:::::::;:::;:;:::;:;:;::::::j ITEM E ~:::::::::::::::::::: ITEM F ITEM G GRAPHIC PRESENTATION SLIDING BAR CHART CAPTION CAPTION 40 20 ITEM A ITEM 8 ITEM c ITEM D ITEM E ITEM F ITEM G RANGE-BAR CHART CAPTION CAPTION 0 20 40 ITEM A ITEM 8 ITEM C ITEM D ITEM E ITEM F ITEM G RANGE-BAR CHART CAPTION CAPTION 10 20 30 40 I A I 8 c D ® E F CHANGE-BAR CHART CAPTION CAPTION 0 10 20 30 ITEM A ITEM 8 ITEM c LAST MONTH ITEM D ITEM E ITEM F ® (!) A SLIDING BAR CHART is a special sort of subdivided bar chart. The length of the bar represents the total of two main components, one extending to the left and one to the right of the base line. For example, personnel might be divided into military on one side of the base and civilian on the other; or the quantity of an item in storage divided into ready for issue and not ready. The main advantage of this sliding treat ment is that both of the components can be measured directly from a common base. This chart, therefore, is primarily a comparison of part s, with the comparison of totals s econdary : a conventional subdivided bar is primarily a comparison of totals. Sliding bars can be used to show either ab solute or percentage values . The two segmentsof the bars can be shaded differently and even subdivided further if appropriate. (!) A RANGE-BAR CHART shows the range or spreadbetween low and high amounts rather than justthe size of single amounts. In effect, a rangebar is the r esult of placing one simple bar on top of another: omitting the part where the shor ter overlaps the longer leaves a range bar. Range bars do not start from a common base, and so cannot be compared directly; instead, they show the size of each range in relation to the two amounts i t s eparates . This sort of presentation is useful when interest is in the amounts at each end of the range as well as in the dif:erence between them. By adding a cross-bar (or some similar de 50 60 vice) to a simple range chart, this type of chart can show a comparison of averages, plusthe range of values behind these averages . Such a chart often tells a lot more than one showingonly the averages. @ Sometimes the value of a range chart can be increased further by adding still another set of H IGH MO. measures, such as goals or tolerance limits. Ill ustration 10 shows one version of this added f eature . I t would be possible under certain special conditions to go a step further and s ubdivide these range bars, but this might well make the chart too cluttered to be understood easily. ® CHANGE-BAR CHART . By adding a direction indicator, such as an arrowhead, the basic rangebar form can be used to show change instead of 40 50 simple range. The addition of the pointer not only tells which end of the bar marks the curr ent r esult , but also tells whether the changeis an increase or a decrease. Unf avorable changes (whether decreases or increases) are usually shown as black bars to make them stand out clearly from favorable changes. Thi s type of presentation is not necessarily limited to showing change from one time to another. The bars might show such inf ormation as original and r evised programs for various items, or performance before and af ter adoption of new methods or procedures . CHART TYPES ~ Sometimes a third set of values can be add~to this chart with profit. For example, the addi tion of an objective or some other bench mark might help put the current performance into better perspective . Illustration 12 shows how this idea can be used, even when the objective is a range instead of a single value and is diff erent for each item. It is occasionally useful to add still another set of values, such as the best previous performance or an estimate for some future period ; but this may make the chart cluttered and confusing unless these dat a f all into a clear pattern and do not scatter all over the chart . A small circle -about one-half the width of the bars -is usually the best symbol for such data. Like every other type, t his chart has its limitations; but i t is an eff ective means of showing which way things have been moving, how far they have moved, and where they now stand. ~ A PROGRESSIVE OR STEP-BY-STEP BAR CHART is an unusual f orm of subdivided bar chart. It shows the segments of a bar as a series of steps one below the ot her . This treatment calls attention to one step at a time and emphasizes the number of steps involved. Putting the pieces one under t he other in t his way also permits easier and fuller l abeling, which is especially helpful when the labels are long descriptive captions. The step-by-step method can be used either to build up a total or to break it down. For exampl e, it could show how several sources of fund s add up to the total available, or how the total available has been used. The adding and subtracting steps can be combined to show a complete cycle of operation or performance. This type of chart is especially useful for taking a close look at the status at the start of a period, the steps that affected the status during the period, and the status at the end of the period. For example, this twoway chart might be used to explain changes in the quantity in stor age of an item of equipment. (BD The pieces do not have to be sorted to show first the inward steps, then the outward. When appropriate, the steps can progress in actual order of occurr ence . For example, this type of chart could be used effectivel y to show a succession of program changes such as: increase, decrease, decrease, increas e , etc . NOTE: Step-by-step presentation uses much more space than conventional subdivided bars: it should not be used unless the nature of the data and the importance of the subject warrant it. Make sure also that t he pieces are really progr essive steps, or at least have a logical sequence. Do not use this type of chart just to add novelty to a simple subdivided bar; and especially do not use two progressive bar charts when one chart comparing two subdivided bars CHANGE-BAR CHART CAPTION 0 40 CAPTION 80 120 160 200 ITEM A ® ITEM 8 ITEM c ITEM D ITEM E ITEM F PROGRESSIVE OR STEP-BY-STEP BAR CHART MILLIONS OF DOLLARS CAPT ION 0 10 20 30 40 AUTHOR I ZED FOR FY 19X6 EXPENDITURES MATERIAL SHIPPED FROM U.S. ~ MATERIAL ISSUED FROM LOCAL DEPOTS I HANDLING COSTS TO DATE ... ..... . 0 I LOCAL PROCUREMENT ..... I OTHER NEARBY h PROCUREMENT ..... LJ I I AVAILABLE 31 DEC. 19XS TWO-WAY PROGRESSIVE BAR CHART THOUSANDS OF POUNDS CAPTION 0 2 4 6 8 ON HAND 1ST OF MONTH FROM SOURCE A FROM SOURCE B TOTAL AVAILABLE DURING THE MONTH TO USE C • I TO USE D ON HAND END OF MONTH will tell the story just as well . GRAPHIC PRESENTATION COMBINATION TYPES By combining basic chart types a chart can often be devised that will fit the data better than any of the basic forms . The illustrations on these two pages are typical examples of how combination types can be developed . They show how these special forms can help to bring out more clearly the particular comparison the analyst wants to show. Of course, these are only some of the many variations possible: other data and other needs will suggest other ways to combine the basic types . As mentioned earlier SLOPE AND STEP-CURVE CHART MI L LION $8~------------~--------------~ FUNDS AVAILABLE CD o~~~~~~~~~~~~~~~~~ J S D M J S D M J FY 19X1 FY 19X2 SURFACE AND CURVE CHART CAPT I ON 100 ,----------,--------y-------, LABEL A 0 JUN DEC JUN FY 19X1 FY 19X2 SUBDIVIDED SURFACE AND CURVE CHART THOUS 0 D M J S D M J S D M J S D 19X1 19X2 19X3 in this chapter, however, avoid charts that are so complex and so obscure that they discourage, rather than invite the reader's interest. Co~bination types should serve a definite purpose other than novelty: do not use them just to be different or just for variety. Do not use them with the idea that combination types are always better than the basic types : they are not . Also, be consistent : always use the same kind of combination for the same kind of comparison. CD SLOPE AND STEP-CURVE CHART . Slope curves and step curves can of~en be combined to make a more effective chart than either kind alone . The example at the left shows a typical use of t hi s sort of combination. Notice that the step curves are used for data that change abruptly from one level to another; and that the slope curves are used for data that change more or less continuously. CAUTION : Do not use the two types of curve together unless the data require it. A step curve used just for variety will usually make the chart confusing instead of more revealing . 0 SURFACE AND CURVE CHART . There are many uses for a chart combining a surface and a curve. For example, a curve representing a standard or requirement can be combined with a simple surface chart showing actual performance or results . As illustrated at the left, this is a simple but effective combination chart . Shaded surfaces should not be used if the two curves shown are close together or if more than two curves would run through the same shaded area . (3) A SUBDIVIDED SURFACE AND CURVE CHART is a useful combination when it works . Chart 3 shows a good example of how effective this sort of combination can be. This one chart tells the story much more clearly and completely than two separate charts one dividing occupancy, the second comparing total authorized, available, and occupied. But this type of combination must be handled carefully when a curve runs through the surfaces : when this happens, be sure to make the curve heavy enough (and the curves outlining the surfaces light enough) to set it off clearly from the background. If more than one curve runs through the shaded surfaces, or if there are several bands of shading, this type of chart may become confusing unless the curves follow a pattern quite different from the shaded bands . Even then, extra care is necessary to arrange the labels so that there is no question of which labels Hhat. CHART TYPES A COLUMN AND STEP-CURVE combination is sometimes a more logical way to show two series than two curves. This form is especially suitable for comparing actual and program data, or for showing two different kinds of things. For example, columns could be used to show act ual numbers of students in training while a contrasting curve might be used to show the number programed. When a step curve is combined with connected columns (see page 91), it is important to make it heavy enough to look like a curve: otherwise it will seem to be only a separator line that divides the columns into segments. (I) A special column and curve combination which has proved very useful is illustrated at the right. In this type, columns are used to measure results for each period against a standard or goal (shown as a straight line or step curve). A slope curve showing the cumulative difference between results and goal is then superimposed. This type of chart shows at a glance the results for each period and the total effect of these results t o date. It is especially good for comparing actual expenditures and a budget or program. (!) Another useful type of chart combines columns showing current results with a curve showing monthly averages for several earlier years. The dashed lines show the average range of the highs and lows in the earlier period. If appropriate, they could show the extreme high-low range instead of the average range; in which case the lines would be irregular instead of straight. This is a good method of showing how much of the change from one month to the next is merely the usual seasonal change, and at the same time co~paring current performance with the level of earlier years. This type of chart is useful for quality control work. (I) SPLIT-SCALE CHART. This is actually a twoin-one chart rather than a combination type; b t since it usually makes use of two types it is included here. The main purpose of such charts is to provide both a close-up of the current period and a quick l ook at the past (or at t he future -with the columns placed at the right of the curves, as illustrated in Chart 3, page 84). In Chart 7 this type has been used to show current data, cumulated monthly, for comparison with the preceding year and with the total for earlier years. The same kind of combination could be used to show quarterly averages over a number of years as compared with monthly da~a for the current year, etc . (See page 87.) This two-in-one type can be used also for non-cumulated data: for example, the current part could show actual monthly production (or strength); the other part, monthly averages for preceding (or succeeding) years. Such a chart, however, can be used only for comparing levels, not trends . Since the time scale for the early (or future) years is very much compressed, it makes trends seem much sharper than on the stretched-out monthly scale. COLUMN AND STEP-CURVE CHART CAPTION 5 J FMAMJ J ASOND COLUMN AND SLOPE-CURVE CHART PERFORMANCE 5 J F MAMJJ ASOND COLUMN,SLOPE-CURVE AND RANGE CHART RAT~E~------------------------------------~ 10 J FMAMJ J ASO NO CAPTION 6 SPLIT-SCALE CHART 4 2 19Xl 19X3 19XS D M J s D GRAPHIC PRESENTATION FREQUENCY CHARTS Most !trrrry charts are time series charts that present chronological changes in data, such as month-end strength for a five-year period. Frequency charts are needed to show data grouped by size or magnitude. The numbers of military personnel, for example, may be classified in terms of age groups, height, weight, etc . In these examples, t he numbers of personnel is called the "frequencies" and the differences between the upper and lower limits of each group as the "class intervals." Simple Frequency Charts Each type of frequency distribution chart has its own uses. The slope frequency curve (Chart 1) is usually best for continuous data where differences in the magnitudes within the class intervals are infinitesimally small (such as age and length of service). It is also best when the number of class intervals is large so that the curve is relatively smooth. The histogram, step, and column forms are good for small numbers of groups or for discrete data (measured in indivisible units) such as number of dependents per family, or pay grades. Slope curves are generally best for comparing two or more distributions in the same chart. Separated columns are usually best when one or more horizontal-scale units are not included in the distribution and must be shown as an open space. Most frequency charts show data by groups or class intervals instead of individual units, as illustrated in Charts 1, 2 , and 4. When the number of individual units is large, . grouping them simplifies the chart and clarifies itsmeaning. (See page 83 for the number and size of groups and methods of labeling and plotting them .) Comparing Frequency Distribution When comparing two or more distributions in the same chart, a scale of absolute frequencies should be used if the purpose is to show simply how many are in each category. If the purpose is to compare the proportion in each category, plot the percentages. Several frequency distributions can be compared by converting them into subdivided bars (see pages 54-55). This method uses a series of graduated shadings -properly labeled -to indicate the class intervals or individual measurement units. Four or five subdivisions are about as many as can be shown satisfactorily, and so this method sacrifices some detail; but the result is a compact summary chart that is easier to grasp than a series of frequency charts or a 3eries of overlapping curves on a single chart. Cumulative Frequency Charts These are useful for showing how many (or what percentage) of the total frequencies were less than any specified quantity (Chart 4A) or more than any specified quantity (Chart 4B) . This type of chart should be used only when it can be clearly expl ained in the title. (See also page 57, Chart 13. ) SIMPLE FREQUENCY DISTRIBUTIONS FREQUENCY 60 FREQUENCY CURVE 17 18 19 20 21 22 23 24 25 26 CATEGORY FREQUENCY 60 0 CONNECTED COLUMNS 20 2 3 4 5 6 7 8 9 10 CATEGORY FREQUENCY 40 SEPARATED COLUMNS 2 3 4 5 6 CATEGORY CUMULATIVE FREQUENCY CURVES FREQUENCY 160 MORE THAN @ O~...::~:.......L......t.....J.._._...L..................., $I 2 3 4 5 6 7 8 9 10 $1 2 3 4 5 6 7 8 9 10 CATEGORY CATEGORY CHART TYPES GRAPHIC TABLES Graphic tables, usually referred to as charts, sometimes may be used to portray statistical data effectively. In this type of table, the entries are sorted into a few main groups, and a suitable graphic device used to represent each group. This grouping procedure is a form of summarization, and the effectiveness of this type of (!) A STATUS CHART highlights good or bad conditions. It can be used for showing either the quality or quantity of performance or status, production, or inventory. For example, this type of chart could be used to show the status of supply at oversea bases for different types of supply, in which case the key might be expressed in quantities such as "less than 50 percent of allowance," "50 to 99 percent of allowance" and "100 percent of allowance or more . " This type of presentation is the most widely useful of all graphic tabulations, and can be adapted to many kinds of tabular data. (!) A PROGRESS CHART is a graphic check list us-ed to record progress, or lack of progress, through a series of processes, steps, or operations: for example, the progress of a suppl y item from depot to oversea command. In using this type of presentation, it is often desirable to show the date of completion of the last completed step or operation. This may be done by putting the date in the last "completed" box for each item. Symbols instead of shadings can be used for charts of this type and are usually better when there are several degrees of progress to be recorded. The "square dial" symbols shown below offer a simple and effective method of showing as many as eight steps: G) A RANK CHART compares the status or performance of a number of items in several different respects. In this type of comparison, individual amounts are not shown graphically : instead they are represented in a general way by showing which of a few broad amount-ranges each item falls into. It is usually best to limit these groups to three or four, especially if there are only a few items . Numbers can be inserted in the blocks to show numerical rank; or, if they can be shown in one or two digits, the actual amounts represented can be inserted. (A good way to do this is to show not total amounts but the difference between each amount and an average or goal .) The series of columns across the chart can also be used for different periods of time, instead of different categories, as in the progress chart above. presentation depends largely on the aptness of the summarizing. This kind of presentation combines the "picture" value of graphics with the main feature of a tabular arrangement -cross classification -and so provides some of the advantages of both forms. The three examples on this page are particularly good for summarizing a number of more detailed charts or tables . STATUS CHART (!) I'V'I NOT KEY : mJ POOR []FAIR 0 GOOD IC>I APPLICABLE CATEGORY CAPTION A 8 c D E F G H J K L ITEM A ITEM 8 ITEM c ITEM D ITEM E ITEM F ITEM G (!) PROGRESS CHART KEY , mcoMPLETED o IN PROCEss oNOT sTARTED CAPTION PROCESS OR OPERATION I 2 3 4 5 6 7 8 9 10 II 12 1TE M A t:;:::;:;:~>;!;l:l:l::~=::::::::::< 0 r<) C\1 0 r<) a:: w w z (5 z w ..J QUARTERMASTER ~ ~ SIGNAL ~ ENGINEERS § MEDICAL ~ Chart E shows for each technical service the TOTAL NUMBER of vehicles disabled four days or more, and also shows how many were delayed in repair by lack of spare parts and by other causes . The main comparison here is between TOTALS; if a comparison of only the number of disabilities due to lack of parts is wanted, only that portion of the bars should be shown (see Chart G below) . However, putting the "lack of parts" segments next to the base line permits this secondary comparison (See page 72 for comments about the type of insert chart used in Chart E) . PERCENT OF VEHICLES DISABLED FOUR DAYS OR MORE DUE TO LACK OF PARTS® SEPTEMBER AND OCTOBER PERCENT SERVICE 0 25 50 75 100 ENGINEERS MEDICAL SIGNAL TRANSPORTATION CHEMICAL QUARTERMASTER ORDNANCE Chart G emphasizes the comparison of the "lack of parts" percentages shown at the left of Chart F . Although the "other causes" segment is not shown in this chart, it can readily be seen as the difference between the bars and the 100 percent line . As an added feature, this chart also shows the same measurement for an earlier period . Thus it permits a three way comparison: of current results, of earlier results, and (indirectly) of the change from one time to the other. 214-853 0 -66 -5 time . The third provides a secondary comparison that measures one ~eriod against another . The fourth summarizes results during the entire time span ocvered. Note that all of these charts are bar charts selected from types discussed on pages 54-57 . PERCENT DISTRIBUTION OF VEHICLES DISABLED FOUR DAYS OR MORE-OCTOBER 19Xl CD PERCENT OF TOTAL SERVI CE 0 25 50 75 100 ORDNANCE TRANSPORTATION CHEMICAL QUA.RTE R MASTER :::::;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:; r:.:::e::.:t.::=x.~ SIGNAL ENGINEERS MEDICAL Chart F shows ~he PROPORTIONATE PART of de lays in repair which resul ted from lack of spare parts and from other causes . Since this type of presentation measures the relative importance of each component, it often reveals facts not readily apparent in a chart showing absolute values . However, a percentage distribution such as this may have little meaning if there is a wide range in che size of the absolute values . The significance of percentage rel ationships may be clarified by showing the absol ute data as a column of figures in the stub (see page 71, "Designing Bar Charts" ), or in a separate chart or table . PERCENT OF VEHICLES DISABLED FOUR DAYS OR MORE DUE TO LACK OF PARTS RANGE SERVICE FOR 0 PERIOD 25 MAY-OCT PERCENT 50 19X1 75 100 ENGINEERS MEDICAL SIGNAL TRANSPORTATION CHEMICAL LOW MONTH QUARTERMASTER ORDNANCE Chart H carries the scory presented in Chart G a step further . It shows an over-all picture for the entire period covered, instead of just the current period. Emphasis is on the range in the "lack of parts" percentage during the period, without identifying specific months . The averages ' show the general level for the entire period and indicate indirectly how the monthly values are distributed within the range. Note that Transportation, with a slightly higher average than Signal, has a much narrower range . GRAPHIC PRESENTATION CHART SKETCHES Sketches are a vital part of chart preparation. They are valuable not only f or trying out different types of chart but also f or three other important uses : 1. Testing the data: t o see if they show the sort of r elationships expected as well as to see if certain r elationships will need special explanation. * 2 . Testing the design: to work out probl ems of scaling, spacing, labeling, and other "bugs" that do not appear until the sketch is started. 3. Instructing the draftsman: to provide a model of the kind of chart wanted, plus any special instructions or specifications needed in ~rawin~ the chart. Without a sketch (or earlier chart of the same data) all of these things have to be done the hard way, and often with the usual by-products of all trial-and-error methods : mistakes, r evisions, unsatisfactory final r esults. Sketches are useful also f or getting approvals and clearances that are needed in advance of publication. Sometimes they can even be used to do the work of finished, inked charts. Illustration B on this page shows how satisfactory a careful sketch can be. Drawn with a medium soft pencil (2-H), with lettering typed or penciled in (free hand), a good pencil sketch can be done in less than half the time it takes to do the same chart in ink. Even longhand l ettering is satisfactory if done neatly. For many informal presentations, and especially f or hurry-up, 2 or 3 -copy jobs, sketch * It is important t o r emember that charts serve a double purpose : they are useful not only for demonstrating facts t o the reader, but also f or revealing f acts to the writer . Often, an analysis written before the data have been chart ed will have to be rewritten because the char t shows a different picture than expected or r eveals additional facts which were not apparent. This is one good reason why charts should be r oughed out in sketch form early in the devel opment of the r eport. 33 __________] 23 ~ ROUGH S~TCH to make a quick pictur e of the data and present a general idea of how t he chart will l ook. charts have many advantages . Excellent copies can be made (by offset or other photographic processes) if the originals are drawn with the understanding that they are to be reproduced. Caution: Do not draw sketches on ordinary commercial graph paper if they are to be r eproduced. Most graph papers are printed with green ink which generally reproduces black. This results in a coarse screen of backgr ound l ines that make the chart hard to r ead. If the lines are printed with a light non-photographic blue ink (a f ew commer cial papers come this way) they will disappear in r eproduction processes. DATA SHEETS FOR CHARTS It is important t o furnish chart data to the draftsman in a way that he will not misinterpret it. Prepare special data sheets designed f or plotting purposes. On these data sheets it is desirable to : 1 . Include only the plotting data, or at least clearly mark which columns or rows are to be plott ed. 2 . Identify the data with the same captions that will be used on the chart. If this is not f easible, mark the columns of data A, B, C, etc . and then mark the chart sketch to correspond. 3. List the data in the same order they will take on the chart. For example, data for three curves -total, group B, and group A should be listed on the data sheet in that order. 4. Cumulate data for subdivided charts. Instead of showing the amounts f or components A, B, and C, show the figures for A, then f or the total of A plus B, then f or the total of all three. 5. Round figures for ease in plotting. A three-digit figure is about as close as you can plot o~ a small chart; a four-digit figure, on a larg~ chart. Sometimes the need f or speed prevents the preparation of proper data sheets, but poorly prepared plot ting data sheets increase. the length of time r equired to prepare the charts and they are likely to result in errors that also add to the delay. ARMY PROCUREMENT DELIVERIES Billions of Dollars 0 10 20 Item A Item E Item C It em D Item E . Item F Item G ® CAREFUL SKETCH with all details of spacing and captioning worked out . Used for approvals or even as the final chart. PLANNING AND DESIGNING CHARTS DESIGNING BAR CHARTS Bar comparisons have several special advantages that should not be neglected. RANKING . The items in a bar chart can be listed in various ways . (See pages 22-24 of the Tabular Presentation chapter.) Probably the most useful is in order of the magnitude of the data. The standard arrangement is from best to worst. Exception : When the aim is to emphasize unfavorable conditions, items are ranked from worst to best . RANKING TO EMPHASIZE : GOOD RESULTS POOR RESULTS AWOL RATE AWOL RATE CD 15 20 15 20 When grouped bars are used, ranking is based on the most important set of bars . When one set represents current data, one earlier data, the current bars are normally shown first and shaded the strongest. In bar-and-symbol charts, ranking is normally based on the data for the bars . (See Chart 3.) GRAPHIC SYMBOLS . Adding supplementary data by means of graphic symbols is especially useful in bar comparisons . These symbols are effective for spotting key values such as schedules, programs, or other standards or goals . Symbols are useful also for showing data for some earlier period -say, last month or last year. Used in this way, a bar-and-symbol chart can summarize several time series by showing what changes have taken place from past to present . The standard symbol (a heavy "slash-line" that crosses the bar and extends on either side) does not work well for connected bars . Instead, use diamonds or circles . Diamonds locate the plotted points more clearly than circles; circles are easier to draw. SYMBOLS FOR GRO~PED BARS ······.... .. .:.:.:.:.:.::. :.:.:.::: 0 ,.....,....:.;.;·· ··.:.;.;..., ..:.;.;.:.;.;.:.;.;······ ----0 ·::;:::::::::::::::;::::::::::::::::::::::::::::•1 t :::::::::: ::::::::::::::::::::::::::::l-----e • Caution: Filled-in diamonds or circl es are so black that they dominate the chart . This may be desirable when the symbols represent goals or standards and so should be prominent; it is undesirable when they represent an earlier date. A third type of symbol a wedge, or half diarr.ond -is usef~ for special cases . Its main use is as a bench mark . This symbol is most effective when it cuts into the bar; least effective when it floats beyond it. Chart 3 shows one way to use this device; Chart 4 another. Still another use: to measure progress to date, against subdivided bars showing progressive steps of construction, production, or readiness . ACTUAL r:L JUN EST .LJ JAN EST . TOTAL RQMTS. STUB DATA. Sometimes there is an advantage in addi~g s pplementary or supporting data in the stub of a bar chart. These figures usually are a different kind of measurement than shown by the bars (for example, dollars vs. percent, tonnage vs . days of supply, workload vs . completed). Adding these supplementary data increases the value of the chart cons:.derably, because it helps to round-out t he picture and make the bar comparison more meaningful. SUPPLEMENTARY DATA IN THE STUB 0/c OF S.:.: ~: ·. : ·. _ .·.· .·.·.·.·. ·.·.·.·.·.·. ·.·.·. .11 15 ITEM 2 ~·.·.·.·.·.·: .. ITEM 3 :-:-:-:-:-:-: ITEM 4 ·.·.·. 10 ~--~--------~--r---~~--------~ o~~~--~~--~~~--~~~~~~ D J F M A M J J A S 0 N D "BLOWUP" INSERT CAPTIONCAPTION O 10 20 30 40 I T EM A JU=U==:.:Mi====t:===~=~= I T EM 8 :;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;: IT EM C F="==.:.:.:.r~'-"' ITEM D 2 ITEM E ITEM F ITEM G IITOTAL" INSERT MILLIONS OF DOLLARS CAPTION 0 I 2 3 4 5 MIL. PERS. AVAILABLE APPO RT IONED OBLIGATED MAINT. a OPER. AVAILABLE APPORTIONED OBLIGATED PROCUR. a PROD. AVAILABLE APPOR TI ONED OBLIGATED 0 5 10 15 TOTAL~ OTHER ~~~~T .:.:·:·:·:;:;;:;:;;::;::;;;:;:;:;:;:;::::::::: AVAILABLE APPORTIONED OBLIG . ;.·:·:·:· I OBL IG ATED PLANNING AND DESIGNING CHARTS SUBDIVIDED SURFACE CHARTS TITLE This type of chart is used many times when it should not be . I t is excellent for showing how the component parts of a total change in importance over a span of time, but it shows only a "broad brush" picture . Unless the changes are pronounced, this form of presentation hides them instead of revealing them. Also, it is not an easy chart to read . Only the total and the bottom layer can be read directly from the amount scale: the other components must be read indirectly, by subtracting the amount of one curve from another to find the difference. Sometimes exact measurement is not important; but when it is, this is not the best type of chart to use . (See Chart B, page 68) . Besides being hard to read, subdivided surface charts can sometimes show a misleading picture . This is most likely to happen when the data take a sudden change in direction, or when one of the components is very irregular (large one month, small the next, then large again, then small, etc.) Sharp Trends I f you show the components of a sharply rising series, watch out for a bad optical illusion. In Chart 1, the top layer seems to be getting smaller toward the end of the year: as you can clearly see in Chart 2, this component actually is larger in the last three months of the year than in any other. The trouble in Chart 1 is that the eye finds it easiest to measure across this layer at right angles to the direction of the slope instead of vertically. Even when you know that such a chart is misleading it is difficult to make your eye see it correctly. A recommended solution is to change to a curve type of chart, or to a column chart, if appropriate, as in Chart 2 . Of course when you change to either of these types, the top component becomes a series of disconnected segments instead of a continuous layer; but this is a smaller fault than the distortion that the layer treatment produces . "Irregular" Components There is another illusion that occurs in surface charts, and this one also is caused by the way the data behave . If the values in one THOUS 5 CD 4 r-------_,---------+---------r-------1~ 3 Q~,W;W:..;,w,;,.:.;..;.;.!:.:.:.;.;:.;w;.:..:..:w;.;w;.;.~=;,;.;,.;.,;w;.;.:..;,w,;,.;w;w~w;w;w;w;=~ D J F M A M J J A S 0 N D TITLE THO US 5 0 4 3 2 DJFMAMJJASOND of the components fluctuate sharply in other words, i: one of the layers is very uneven -it makes any layer put on top of it seem uneven too . In Chart 3, layer C makes A and B follow such an irregular path that they seem much more uneven than they are . Chart 4 shows a better arrangement of these components , Chart 5 a still better one . Of course, it is not always possible to shift the order of the components to f it the be havior of the data; sometimes one order is either so fixed or so logical that any other seems wrong . But when one component is extremely errat~c, try to find a logical arrangement that will put this layer at or near the top of the chart . GRAPHIC PRESENTATION STATUS-CHANGE-STATUS CHARTS There is often a need to show both inventory data and production data for the same subject . By presenting these together, it is possible to show how things stood at the beginning of each period, what activity took place during each period, and the resulting status at the end of the period . A combination of thi s kind provi des a rounded story that can be much more informative than eicher series alone . But if you attempt to combine the two kinds of measurement into one continuous time series, you may get some strange looking charts . Chart lA shows the most acceptable way to interweave status and change data . This form follows the same general method used in constructing a progressive bar chart (see page 57) . I t presents a f airly clear picture if the reader understands that each column r epresents t he end of one month and the begi nning of the next . Chart lB is less satisfactory because i t looks like a seri es of divided double bars, which it is not . Chart 2A i s generally a better way to show the ins and outs and their net effect on the status . By separating the production dat a from the inventory data, it presents a clear easy-tounderstand picture . If there is a large volume of ins and outs but a small volume on hand, as there might be at a Reception Station, the inventory curve will cross the columns, producing a simple combination type such as the one shown on page 59 . If there is a large volume on hand but a very small volume of ins and outs, the two series may be too far apart for easy comparison . One way to remedy this : put the ins and outs on a separate grid directly above the curve chart . Chart 2B shows a method to use if the ins and outs combi ne to make an important total, such as additions to and shipments from storage (total \vork volume) . Although this method uses more space, it has a compensating advantage : the outs extend downward and so appear to pull down the on hand curve when they exceed the ins . CHANGES I N THE CONTENT OF THE DATA Sometimes the composition or coverage of the data changes suddenly (or the effect of a gradual change is finally adjusted for) . A typical example of this would be a change in the definition of the geographic area under the control of an oversea command , or a change in the meaning of the term "deliveries . " When these changes are important enough to change the level of the data, they should be clearly indicated to prevent the reader from misreading the chart . For example, without some warning signal, a sharp drop in a production curve will be misinterpreted to mean a decrease in production instead of oerely a change in definition that excludes iteos formerly included . COMBINING STATUS AND CHANGE DATA @ OUT JUL AUG SEP I 31 I 31 I 31 JUL AUG SEP .. . fib JUL AUG SEP JUL AUG SEP HOW TO SHOW A CHANGE IN THE DATA 0 1----o. -A I * - I I \, / 8 * - I I l _l l I I I I 19Xl 19X2 19X3 * NOTE ABOUT CHANGE IN COMPOSITION OR COVERAGE. Curve A on Chart 3 shows the standard method of showing a basic data change : a light dash l ine connecting tb~ old and new series . Unless stated in the labels, the change should be marked with an asterisk and explained in a footnote . Curve B shows the method to use if the change in level is very small. If the data are available, there is an advantage in including in t he footnote the value for both the old and new series at the point of the change-over . This makes it clear that the break is not an omission of data . PLANNING AND DESIGNING CHARTS LOOARITHMIC CHARTS Charts having logarithmic amount scales are useful for comparing ratios instead of amounts . This type of chart, sometimes called a ratio chart, is widely misunderstood and misused . It should not be used unless the writer and the typical reader understand its characteristics . On an arithmetic scale, equal distances represent equal amounts; thus the distance from 1 to 2 is the same as the distance from 2 to 3, 3 to 4, and so on . On a logarithmic scale, equal distances represent equal ratios . For example , the distance from 1 to 2 is the same as the distance from 2 to 4, 4 to 8, 8 to 16, and so on in each case a ratio of 1 to 2 . This property of a logarithmic scale is valuable for two reasons : (1) on this scale a series pf constant percentage changes appears as a straight line; (2) equal relative changes always take the same angle of change whether oc curring at the top or bottom of the chart . These are really two ways of saying the same thing, but they lead to two different uses of this type of chart . The first use is in the mathematical projection of trends . This is a technical use appropriate only for certain kinds of data in a special kind of time series analysis . It is not useful for most statistical reporting . The second use is in comparing relative changes (of a single series of data at different times, or of two or more series at the same time) . This kind of presentation is sometimes appropriate for recurring Army reports. If, for example, the relative change in workload has been more or less than in personnel, or if deliveries of item X have increased at a faster or slower pace than Y, a logarithmic chart will show this very clearly. Chart lB demonstrates these relationships. Charts that have a logarithmic amount scale at the side and an arithmetic scale at the bottom are called semi-log charts . They may have one "cycle" which has amount scale captions ranging from 1 to 10 (and multiples thereof) or two or three cycles with ranges of 1 to 100 or 1 to 1,000. ARITHMETIC LOGARITHMIC 10 10 T CURVE X .J' 8 I __...... ..J'- 8 v 6 CURV/ -v v 6 v 4 4 v cu~v CURVE~ 2 v 2 I.. / v @ @ ~ v 0 ARITHMETIC vs LOGARITHMIC SCALES 0 2 3 4 5 6 7 8 9 10 ARITHMETIC: Amounts equal; ratios unequal. 2 3 4 5 6 7 8 9 10 I I LOGARITHMIC: Ratios equal; amounts unequal. The following illustration shows another valuable use of semi-log charts . Note that small a~ount changes at a low level may be more important relative to their level than large changes at a high level . This kind of perspective can be gained only from a chart like 2B. It is especially helpful in evaluating data that increase from a low to a high level . ARITHMETIC LOGARITHMIC 20 20@ 10 1-------+-----l 8 1------~~----l 6 1-----~~-----l OL-------...J Charts that have logarithmic scales on the bottom as well as the side are useful for studying certain production quantity-cost relationships, such as those for aircraft and missile items . These are called learning or progress curves . The virtues of logarithmic amount scales can be important; but they are gained at a cost . Since most people expect a level twice as high to be twice as great, they can be badly misled by logarithmic charts if they do not understand that they do not show that kind of comparison. Logarithmic charts cannot be used to show zeroes or minus figures . They are not as easy to construct as arithmetic charts . Also, log-scale charts look technical and discourage some readers from trying to understand them . For these reasons they are not recommended for Nidespread use . They are, however, excellent for the specialized purposes discussed above when it is fairly certain that the user understands their peculiarities and significance . GRAPHIC PRESENTATION CHART STYLE Styl e -the manner of presentation -i s an important factor in chart design. It is not something a chart may or may not have: all charts have style of one kind or another. However, the same data can be presented i n many styl es : plain or fancy, bold or deli cate, conventional or unusual, geometric or pictorial. Style should al ways serve a purpose other than novelty. Introducing style flourishes to attract attenti on or to disguise the fac~ that the chart is a chart will distract the reader rather than help him get the i dea it contains . The right styl e is the one that best fits these two conditi ons : 1. The purpose of the chart and the kind of audience i t is designed for . 2. The means of presentati on (either report or di spl ay, etc . ) . Of the two, the first i s the more important in setting the basic style; the second usually requires merely an adjustment of the basi c style to fit the physical means of presentation. Style and Audi ence The importance of style should not be overemphasized; often a chart styled for one audience will serve very well for others. Nevertheless, style can make a difference. The wrong style can make a chart seem inappropri ate and out of place: an i nept style can obscure the purpose of the chart, or even discourage the audi ence from reading i t . The accompanying examples illustrate some of the main chart styles. Many variations of each of these are possi ble, including styles that make use of color, photographs and other art work, and numerous other speci al effects; however, these s i x are typi cal and show the gen eral range that can be applied to most statisti cal charts. G) "Statistical" Style. The basic chart style . Useful for a wi de range of audiences if the concept involved in the chart is a s i mple one. Permits f iner accuracy and detail than most other styles, but should not be used to give ap proximati ons the appearance of exactness . It i s t he best style for audiences accustomed to sta tisti cs . The style used in thi s manual i s a s i mplified version of this form . 0 "Narrative" Style. A modifi cati on of the "statistical" style. Tells the s t ory in the title, the graphi c part of the chart servi ng to demonstrate the statement in the title. An effective style for audiences who must deal in amounts but are not accustomed to the "stati sti cal approach" and so need to have the meaning of the chart interpreted or summarized in words. Wri ting the text takes time . G) "Poster" Style. Useful for trai ning aids, educati onal displays, briefing char~s, and s i milar uses. This style sacrifices some accu racy and detail for boldness and should be used for audiences who are not concerned with mi nor variations or intricacies of the data . 76 0 "Decorative" Style. Appropri ate for use in informational booklets, etc . having other types of illustrations and art work. For audiences who receive charts better if they are "non-statistical" in appearance . (D "Three Dimensional" Style . A vari ety of the decorati ve style . For the same type of audience, but usually more striking . It i s another method of avoiding the "statistical" look. Not recommended because the third di mension often distorts the picture. When the third dimension is not used for decorati on but i s designed to measure something, the chart becomes more, instead of less, complicated . (See also page 63) . ® "Pictorial" Style. The farthest removed from the "statistical" style and useful for audi ences at the other end of the knowledge scal e . This is the extreme method of di sguising a chart . The main weakness in straight pi ctori al t reatment i s i ts loss in accuracy. But i t i s also costly in time and money . Two compromise ~ethods, although not as stri king, are more reliable: (1) addition of pictori al materi al to "statistical" style charts, and (2) conversi on of the chart to "pi ctograph" style (see page 64) . Most of these styles are useful at one time or another, but only the f irst two are recom mended for stati stical reports . The so-called stati stical style, as illustrated throughout thi s manual, i s the standard for Army reports and all s i mil ar working documents . Style and Means of Presentati on The follow1ng summar1zes style adjustments needed for the mai n methods of presenting charts : Statisti cal Reports. Includes stati sti cal studi es and analyses, operating and other management reports , financial, economic and similar technical publications. 1. Use a style that emphasizes accuracy and completeness rather than popular appeal. 2 . Use a style that can be standardi zed, and will permit rapi d production. 3. Use a style that can be reproduced effectively by the less expensive reproduction methods, avoiding color, half-tone, and other costly features . Printed Publications. Includes i nformational booklets, training manuals, and s i milar publi cations that are not pri mar ily statistical. 1. Adjust the style to match the general style of the publi cati on . Make it bolder or more restrained, larger or smaller, simpler or more elaborate, as needed. 2 . The lettering may be handled in a style similar to that used in the rest of the publicat i on . Use the same kind of type and follow the specificati on of the publication as to size and placement, particularly of titles . Exhi bits and Displays . Includes lecture, briefing and conference charts . 1. Simplify theseby eli minating unnecessary detail. 2 . Make all lettering big and bold and read horizontally. L------------------------------------ PLANNING AND DESIGNING CHARTS PRODUCTION OF MISSILE M 104A2 THOUS 60 ------------------------------------ CD ACTUAL r __ J I ---r--.r----_J 20 J F M A M J J A s MISSILE PRODUCTION UP-STILL BELOW SCHEDULE Retooling delays are still slowing production of the redesigned Missile Ml04A2 ; October i s expected to be up 5,500; 8,000 off schedule. THOUS 60 -0-2------------ r--J I r----.J ---r- 20 J F M A M J J A s MISSILE PRODUCTION THOUS 60_(D_3______________________ 40 --------------------~ J A S J J A THOUS 60 - MISSILE PRODUCTION J F M A M J J A s MISSILE PRODUCTION 50,000 24,000 JAN JUL SEP NOTE : The data plotted in these charts are purely fictitious . GRAPHIC PRESENTATION CHART CONSTRUCTION GRID PROPORTIONS This section of the manual tells hbw to lay out, space, and draw charts . Some of this information is mainly of interest to the person who draws the final chart . But some of it, especially the parts about size, proportion, and scale selection, is about things that the analyst must consider also . In fact, these are construction matters that the draftsman is seldom in a position to decide alone . The Draftsman ' s Part in Graphic Presentation The draftsman has a key part in chart construction. Although no amount of drafting skill can make a poor chart good, poor drafting can ruin a good chart . Drafting skill is more than the ability to draw a clean line and to make and space lettering properly. It is the ability to understand what the analyst is trying to achieve graphically, and to produce this result . This means that the draftsman can do a first class job only if he knows what the chart is about, and, to some extent, why and how it is going to be used. Unless the analyst provides this background information, no draftsman can be expected to do his best work . With a working knowledge of what the chart means and what purpose it serves, an experienced draftsman will not only carry out the analyst's ideas more effectively but will suggest ways to make them even more effective Why Construction Standards are Valuable The last part of this section shows drafting specifications for lettering, line weights, curve design, and shading patterns . These standards are useful in three ways : l. They help to get work done more smoothly by providing a standard way of doing it. 2 . They help to produce better charts, by using the best graphic practices .* 3 . They help to make the charts easier for the reader to understand and use, by following accepted conventions of chart construction, and by presenting similar information in a uniform way. Therefore, the standards should be followed as a regular procedure. However, they should be used with judgment . The specifications recommended in this manual are intended to serve as a help, not a hindrance . There will be times when they cannot be followed, or when the chart requires some change in them to meet a special problem. But these standards should work well in most cases . As a general practice they should be followed closely for all Army statistical reports that are to be reporduced . Even when charts are not presented in printed publications or when they are made in less finished for~, the principles behind these standard specifications should be followed if possible . * Practices approved by the American Standards Association have been used as a guide. The shape of a chart has an important effect on the picture it presents . A short, wide chart :1attens a curve and makes the changes in it seem to occur gradually. A tall, narrow chart exaggerates the movement in a curve and makes the changes seem to occur abruptly. Charts B and C illustrate this clearly. The shape of the chart affects all time series charts -curve, column, surface in the same way. For a horizontal bar chart, these extreme shapes have the opposite effect : a wide chart exaggerates the differences in the lengths of the bars, a narrow chart reduces them. 4® HOW THE SHAPE OF THE CHART AFFECTS THE PICTURE : In most cases, a shape somewhere between these extremes works best . The "normal" grid proportion for the average time series chart is about 2 high by 3 wide . Most of the time series charts in this manual are drawn in that proportion or close to it. This same general shape works well also for the average bar comparison (see page 90 for further discussion) . The regular exceptions to this 2 to 3 grid proportion are : (a) a long time series with a great many plottings, which require a wider than normal chart, (b) a long series of bars, which require a taller than normal chart. As shown on pages 85 -86, the standard procedure for a long time series is to make the chart the full width of the page . For a large number of bars, there is no standard chart height, the size varying with the number of bars included (see page 90). While a 2 to 3 proportion is considered the norm, this is not to be construed as an inflexible rule that must be followed in all but the most extreme charts . The shape can be adjusted where r.ecessary . :f every chart becomes a special case, however, all of the advantages of standard sizes are lost . It is generally best to make most charts approximate the standard proportion and sizes. One reason for this is because it helps to keep less important charts from getting more space -and more attention -than the more significant charts . It also helps to simplify the planning and preparation of the report . CHART CONSTRUCTION CHART SIZE The size of the chart i s important for three reasons: it has an effect on how the chart l ooks, on how much emphasi s it gets, and on how i t f i ts on the page. Many chartmakers make the mistake of letting the space control t he chart: if there is plenty of space available, i t becomes a big chart; i f there is very little space, a small chart. Instead, they should choose a size t hat will: l. r- 1 I I Provide enough space for all parts of the chart to f i t in comfortably (but no more) . Whether they call for this upright I shape, or the l ong hori zontal s hape shown I across the bottom of the page, few charts I need to be bigger than a quarter page . I I I f+-------Half-page wide* Half page high* I f---______1_______ I I I I Many charts need to be no more than an I eight h of a page. In smaller reports (letI ter size or "pocket" s i ze) a quarter page is I about as small as a chart should be unless it is extremely s imple. I I I I I Quarter page I high* I I I I I ,- -------_ _j L---__l_ I Charts that are a 2. Fit the chart to the layout dimens i ons used for the report. 3 . Give the chart the emphasi s it deserves (in relation to other material on the page and to other charts on other pages). Sometimes l and 2 call for a different size than 3 . However, it is almost always possible to ~eep the chart legible, make it fit the page format , and still have it about the right size to match i ts importance . NOTE : Complex charts wi th a lot of detail take much more space than s impler charts. Before de signing a compl ex chart, make certain that i t i s wor t h the added emphas i s that i ts size gives i t . Often a simpler, summary type chart will be just as effective. Sometimes a chart l ooks complex because it is poorl y designed . Redesign and simplification of ~it les and capti ons will often change an apparently complex chart to one that is simple . Standard Sizes As shown on pages 105-107, this manual re commends two standard wi dths for charts : t hose that are a full page wide and a half page wide. The half-page wi dth i s the same as one column of text or a one -column table when a two-column format is used; the f ull-page width is the full wi dth of the "copy area" (2 columns plus the white space between). NOTE: ''Wi dth" means width of the entire chart including scales and scale capti ons . The speci f i cati on pages 105-107 show exactly how big to make several typical kinds of charts. Many of the other charts in t his manual illustrate how s i ze can be varied to fit the data in t he chart and to give the chart the emphasi s it is meant to get. For exampl e, pages 80-82 show how small charts can be if they are simpl e enough. It i s generally better not to use nonstandard widths for single charts . However , charts may be paired or grouped to total either of t he standard wi dths, (see pages 59, 62, and 80-82) . ----------------t--i full page wi de (as wide as this outline) are usually needed when a chart I shows plottings for a long period of time, or when i~ is made up of two or three small chartsI grouped under one mai n title. This height (quarter page) can be increased (to a third or even aI half page) i f the chart is important enough to get tha~ much space, or i f the nature of the data 1 demands i t . (See pages 85, 105 and 107 for examples of f ull-width charts . ) I I ~-----------------------------------Full-page wide* -----------------------------------4----~ Quarter page high* I L_ __ l __j * These dimensions would of course be proporti onaly smaller in smaller reports. GRAPHIC PRESENTATION SCALES How the chart looks to the reader depends as much on the scales as on the data. The four charts below demonstrate this clearly: they all present the same data; yet, because they have different amount scales, each presents a different picture . Only one can be the best . IQO I 0 1---d:::===::::l Although the principles of scale selection are the same for time series charts, frequency charts, and bar charts, the problems of applying these principles are not quite the same, and so they are discussed separately. Scales for Time-Series Charts Amount Scale Range is important because it has a direct effect on the height and movement of the curve . Usually the amount scale should begin at zero and extend upward (and downward if necessary) to include the largest and smallest values in the data . Excessive range should be avoided because it depresses the curve, minimizes differences, and wastes space . Breaking the Amount Scale (omitting the zero end or an interior section of the scale) is not recommended as a standard practice . I t is sometimes useful for analyti cal "work" charts, to give the analyst a close-up of the changes ; but it is not reliable for presentation charts because it is likely to mislead the reader. EXCEPTION: Occasionally it is necessary to show small but important differences that would be lost if the amount scale extended to zero. One way to show t his would be to measure the differences or changes, and picture just that ; but sometimes there is an added value in ret aining the original scale of measurement . For instance, in a comparison of actual military strength with program, the differences or changes are not only important in themselves but also in relation to the totals . For comparisons of that kind -where small differences between large amounts must be enlarged for close inspection -you can omit an interior section of the salce, using only enough scale range to include the curves .* BUT REMEMBER: a chart drawn in thi s way i s a compromise; it throws the over-all picture out of perspec tive to provide a close-up of the detail. NOTE: When an interior secti on of the scale is omitted from a chart, you must give the reader fair warning that the scale is not complete . As shown in Chart 3, the standard method is to show the zero end of the scale separated from the rest of the chart by a double "torn" edge . * Applies only to curve charts : zero or an interior section of the scale should never be omitted in column or surface charts . There is no simple rule for scale selec~ion. The principle to foll ow is to try to choose scales that give a true picture of the data considering the nature of the sub,iect and the purpose of the chart. The following four ~ages 0ell how to do this, and also explain how to divide scal es, and how to arrange and label them. CD UPPER LIMIT OF SCALE RANGE 8 10 .-----------., Normal practice is • • but when most to include highest of curve is near plotting, but no top, range may be 0 more ...... . increased. 0 L...--------' @ LOWER LIMIT OF SCALE RANGE 10 10 .--------. • • because the chart presents a misleading picture if zero is not included. Standard practice is to extend the scale to zero. o.__________, @ BREAKING THE AMOUNT SCALE 250 ACTUAL REVISED PROGRAM ' ' ' /,... ....... 30 JUNE 'Xl'' I --..... __/ PROGRAM ~-- 230 ~----------------~----------------1 CHART CONSTRUCTION Breaking Curves . Another method of breaking the amount scale is to stop it short of the highest curve value and "break" off the top of the curve . This is permissible if the high point is not significant or if it is the result of some special condition that makes it not comparable with the remainder of the data . Chart 1 shows the same series scaled (A) to include the highest value and (B) to exclude it. Method A is correct if the high point is part of the story; B is correct if the high point is a freak point that should be disregarded . Cautions : Do not break off peaks merely to increase the movement of the remainder of the curve (use a taller chart). Do not break off peaks that run just a little above the top grid line (unless they would run into the title or some other lettering). When a curve is broken off, the remaining side-pieces should be slanted as though the value were actually plotted. The amount value of each plotting that is "off the chart" should be shown in small numerals directly above the top grid line . Uniform Scales . Placing comparable or related charts together invites comparisons between them . Comparable scales should be used for charts that are to be compared . If charts are grouped merely for convenience, and a comparison among them is either not possible or not significant each chart may be given an independent scale . However, readers unconsciously compare similar charts placed close together and this should be kept in mind when selecting scales . Amount Scale Intervals . The scale should be divided into as many parts as the reader needs to read the chart accurately. Statistical charts, however, are meant . to be read approximately not exactly and a few scale divisions are usually enough . (See "Scale Rul ings," pages 84-85 . ) When relatively exact reading is desired, the amount scale should be divided into units of 2 or 4 (or these multiplied by 10, 100 , etc.) . The reason f or this is that 2 or 4 can be easily divided in half by eye, then divided again, for fairly accurate r eading . When less exact reading is needed, units of 1, 5, or 25 are better because they are most familiar, especially for measuring money or other financial data. When measuring percentages, units of 10 or 25 are often useful because they provide a 50 percent line (often an important bench mark in i\ercentage comparison) . ' ) , \ Do not use awkward units such as 7, 11, 13, 17, 19, etc . , even though these may provide just the number of horizontal rulings you wish . They are too difficult to use in reading the chart, and produce unfamiliar multiples such as 14, 22, 39, 51, etc . Units such as 6, 12, 15, or 30, however, may be used when appropriate . For example, it would be logical to measure "time on octive duty" in units of 6 or 12 (months), or "rations" in units of 15 or 30 (days of supply) . CD BREAKING CURVES 7 .9 8 0 3 ® 6 4 2 l----/ J ~ I I 0 @ UNIFORM SCALE FOR RELATED CHARTS 20 20 Comparable charts shown together should have the same amOWJt scale. 0 ~------------~ @ UNIFORM SCALE-SPACE SAVING METHOD 20 12 This method keeps a uniform scal e rut adjusts the scale ranges to fit the. data. 0'-------------- @) DIVIDING AMOUNT SCALES 20 20.-------------~ 1615 12 10 8 f.nsi l y -d i v i ded 5 Fami lia r i nter va l s in ter va ls a r e 4 a r e hes t f o r r e.'v!-best for r eadi n'! t::e sca l e . intJ the c11r ve , 0 0 GRAPHIC PRESENTATION Location of Amount Scales is standardized as follows: put the scale at the left side of half-page -wide charts, and at both left and right of full-page-wide charts . EXCEPTIONS : ( l ) When two charts are used side by s i de under a s i ngle heading, put the scale for the l eft chart at its left, the scale for the right chart at i ts right. (2) When there are three or four charts side by s i de under a single heading do thi s : (a) if all the scales are the same, show them only at the left and right ends of the row ; (b) if the scal es are di fferent, show one at the l eft of each chart . Amount-Scale Captions should indicate the unit of measurement and should be used to avoid large numbers in the amount scale. They are placed above the amount scales in one or more hori~ontal lines. Captions should be kept short . Do thi s by: l . Abbrevi ating : Use MIL SQ FT instead of MILLIONS OF SQUARE FEET; THOUS MEAS TONS or THOUSAND MEAS TONS instead of THOUSANDS OF MEASUREMENT TONS, etc . Be sure to use abbreviati ons that cannot be mis understood by the reader. 2 . Shortening: Use RATE, instead of INCIDENCE PER 1000 STRENGTH; TONS instead of CAPACITY IN TONS ; etc . Use either THOUS ROUNDS or THOUS ) whichever fits the ROUNDS) space best. Also use MILLION) instead of (MILLIONS OF DOLLARS $100 ) (100. Another good way to shorten long cpati ons is to put some detail into the subtitle or labels, or simply move it inside the chart. I nstead of INDEX : l9X9 100, put only INDEX above t he scale, l9X9 100 i nside the grid . Or, instead of THOUSANDS OF BARRELS PER YEAR, put THOUSAND above the scale, and use BARRELS PER YEAR as a label ins i de the chart or include it in the t itle . Time-Scale Range should normally begin wi th the earliest pl otting peri od and end ·wi th the latest . EXCEPTIONS : (l) When the data do not start at a major time divisi on the standard is to begin the chart earlier than the first plotted point. (2) When the chart i s to be brought up to date from time to time, the standard is to draw it >·lith room for several future plottings . Time Scale Divisions. Normally there i s a scale division (but not necessarily a time l a bel) for each plotted point . (The illustrati ons on pages 86-87 show how this standard method of dividing and l abeling is applied to typi cal time scal es.) EXCEPTION : When data are measured at unequal time intervals (for example, l Jan, 15 Mar, 7 Apr, 20 May), the divisions of the time scale should be equally spaced but the data should be plotted "irregularly" (see Chart 2) . PLOTT I NGS AT STANDARD PLOTTING BOTH PERIOD AND POINT DATA IRREGULAR INTERVALS CAPTION CAPTION 3 3 CD 2 / --v ~ (Label and rot on the lines) (Label the spaces -.,. I 0 I 0 ... J F M A M J J F M A M J J F M A M J TREATMENT WHEN SOME SPLIT SCALE TO SHOW CHANGE IN PLOTTING INTERVAL DATA ARE OMITTED CHANGE IN TIME INTERVAL CAPTION CAPTION 30 15 5 ,____------J 0 1.-....J........._._....J 1234J FMAMJ JFMAMJJASOND 1ST 2ND 3RD 0 N LAST YEAR THIS YEAR CHART CONSTRUCTION When data for both a point in time and a period of time are shown on the same chart, plot the data as indicated in Chart 3, page 82 . When data for some of the time units are omitted (not available, not reported, or not comparable) the time scale should include space for them, and the reason for the omission should be explained in a footnote (see Chart 4, page 82). When the t i me unit changes, as: annually to monthly, or quarterly to monthly, space the intervals proportionately. That is, make each year as wi de as twelve months, each quarter as wide as three months (see Chart 5, page 82). Otherwise, trends will be badly distorted and changes will seem more sudden and severe in the early part of the chart than in the later part . EXCEPTION : Time intervals need not be spaced proportionately i f the chart is separated into two grids : one for (say) quarterly data, one for monthly (see Chart 6, page 82). The Standard Location of Time Scales is underneath the grid . All lettering should read horizontally. * When time scale labels are so close that they become crowded, abbreviate them or omit some or both (see page 86 for examples of recommended standards). When two or more rows of charts grouped under the same title have the same time scale, do not repeat the scale on each chart . Instead, place each row of charts close to the one above, and show the time scale for the bottom row only. Ti me-Scale Captions are not always needed : but when they are, they should normally be centered below the scale or the parts of it to which they relate . When it is i mportant to call the reader ' s attention to them they can be moved up inside the grid (usually in the upper left corner). Captions like END OF MONTH, AVERAGE FOR MONTH, can be omitted if their meaning is * See examples on specification pages (105-107). carried in the chart title . This can also be done with captions such as QUARTERLY or ANNUAL AV~~GES but the explanation must be given somewhere on the chart. (See exampl es on page 87) . In all other respects, time-scale capti ons should be handled the same as amount-scale captions, except that abbreviati on is not as i mportant because there is usually plenty of space for spelling out. The standard is to include the caption FISCAL YEAR in fiscal time scales covering more than one f i scal year . It should be spelled out when there i s plenty of space; abbreviated FY when space is tight . (See page 86 for examples . ) Scales for Frequency Charts Three important considerations when planning frequency distribution charts are : the number of groups, the size of the groups, and the method of labeling the groups. The number of groups is important in bringing out the pattern of the distribution . Too few groups hide the pattern; too many groups break it up. As a general r ule do not use fewer than five groups or more than twenty. Ten to fifteen is the best range for most di stributions . Although not a rigid rule, it is generally best to use groups of equal size . Frequently data divided into unequal groups are easily misunderstood, even when drawn properly. (See Charts lB and lC below ) . EXCEPTION : When the data naturally divide into unequal groups, the chart is usuall y more meaningful i f drawn that way. (Example : educational levels.) The hori zontal-scale labels should explain the size of the groups clearly. "Overlapping" labels such as 0-10, 10-15, 15-20, etc . do not tel l which group includes the repeated f i gure. For continuous dat a, the preferred method is : less than 10, 10 t o 19 .9, 20 to 29.9, etc. For discrete data, the preferred method is : less than 10, 10 to 19, 20 to 29, etc. EQUAL VS UNEQUAL S IZE GROUPS FREQUENCY 120 EQUAL-UNEQUAL-UNEQUAL @ -CORRECT 21 @ -INCORRECT .!Y ® r-CORRECT £1 eo r t- r- r r- - 40 - . - r- - ;- 0 ~10 9 I I L 2 H 21 l 21n15 : ····:. 7 -"t----, o 1-2 3 4 5 6 1 0-1 2 3-4 5-7.9 o 1 2 3 4 5 0 .9 1.9 2 .9 3 .9 4 .9 5. 9 6 .9 7 .9 0 .9 1. 9 2.9 3 .9 4 .9 0 .9 1.9 2 .9 3 .9 4 .9 7 .9 SIZE 5f An accurate picture of the frequency in each size group . Q1 Inaccurate because not comparable. Total frequency for the enlarged group (5-7 .9) should be divided by 3 to show the average frequency per 10-unit group. s/ The dashed line shows the standard method of showing the 5 -7 .9 SIZE SIZE enlarged group . This preserves the general pattern of the distribution. The dotted outline shows an acceptable method more suitable for analvtical and administrative use: retaining the ~otal frequency but compressi ng the group to the standard wi dth. GRAPHIC PRESENTATION TOO MANY RULINGS SCALE RULINGS Vertical and horizontal r ulings or scale lines should be usedto g~ide the eye in locating and reading points on the chart. Do not use too few or too many of them. Some people like to show a grid ruling for every division of the scale . This is fine for mathematical or engineering charts (or even for some statistical "work" charts) but not for statistical presentation . Closely spaced rulings are not needed on charts intended primarily to show relationships. When close measurement is needed, the values should be shown in an accompanying table . Too many rulings clutter the chart and make it harder instead of easier to read. On the other hand, a complete lack of ins i de rulings is not good either . Although a chart is primarily a picture, the reader should be able to gauge the approx i mate amounts represented by the graphic part of the chart . This i s hard to do without rulings to guide the eye. Also, the picture may be misleading if it does not have enough rulings. Chart 2 shows how an optical illusion can mislead the eye: the difference between the two curves is exactly the same in March and in September . If horizontal lines are drawn through these plotted points, it is evi dent that this is true . Most curve and surface charts should have at least 2 or 3 verti cal rulings and 3 or 4 hori zontal ones. Extremely small charts, however, do not need inside rulings. Few charts need more than 8 or 10 vertical or horizontal rulings. There is no easy answer to the question "How many rulings is the ' right ' number?" Charts that are to be read precisely need more rulings than those meant to give only a general picture; tall charts need more rulings than short charts; wide charts need more than narrow ones . The illustrations on the next page show how to apply these princi ples to charts having standard widths. Special Ruling Problems : Column and bar charts usually need only one set of rulings. Column charts need no verti cal ruli ngs because the columns themselves serve the same purpose . EXCEPTION : When a column chart is pai red with a curve chart, as in illustrati on 3, end rulings are often used for uniformi ty. Bar charts have only one scale and do not need a second set of rulings. Rulings should not cross columns or bars, especially when they are s ubd i vi ded. As shown in Chart 4, rulings that cross look like division lines and are confusing. Averages, ceilings, author.izations, programs, or other data that appear as a straight line should be treated as curves (drawn through bars or columns). 30 25 0 r- ...........!\ / I I'\ 20 ~~ 15 \ Jv "'.... 10 ' I ' 5 0 D J FMAMJ J A S 0 N D NOT ENO UGH RULINGS 30 0 2 0 10 0 ~_.__.__.__._~--~--~~--.__.__._~ D M J s D END RULINGS FOR COLUMN CHART 400 PROGRAM 0~ ~CTUAL 300 -1- AUTHORIZED r---"" 200 100 I I 0 JJASONDJ FMAMJ 30 JUN FISCAL YEAR 19X1 'Xl 'X2 DO NOT RULE THROUGH COLUMNS OR BARS 3--------f 3 PREFERRED J F M A M J J F M A M J CHART CONSTRUCTION Charts with several curves usually need more HORIZONTAL rulings than those with one curve. Rules for Curve and Surface Chart Scale Lines l. The bigger the chart (on the same size page) the more rulings it should have . 2 . Use some inside rulings -both horizontal and verti cal -except for extremely small charts . J . ~ule all divisions of the AMOUNT scale for which scale numerals are shown . 4. Rule only the key divisions of the TI ME scale, following a consistent scheme . (See also the next two pages . ) .-----------------.----------------,6 A tall grid like this needs more HORIZONTAL rulings than a short grid like the one below . IJ should never have fewer than 2 or 3 i nside horizontal rulings . An irregular curve or one that has a wide range needs fewer HORIZONTAL rulings than a smooth, or fairly flat curve . - 50 '-../ A short grid like this should never have more than 3 or 4 inside HORIZONTAL rulings . 40 1-o. ---....__ 30 " "'--- ~ 20 10 I I I I I I I I I 0 A wide grid like this needs more VERTICAL rulings than narrow grids like those above . Usually it will have more anyway, because wide grids are generally used for longer periods of time; but no matter what the time scale, wide grids should have more vertical rulings than narrow grids . GRAPHIC PRESENTATION TIME-SCALE DESIGNATIONS AND RULINGS Year labels not used unless are the chart covers more than one year: for a year or less, Time-scale designations and captions should the year is usually put in the chart title . always read horizontally. Label every month in NOTE: do not repeat years in the title if shown t ime ranges up to one year; every third month in ~n the time scale . ranges of 15, 18 or 24 months. Cal endar Year Data Fi scal Year Data I First 6 mooths ~First 6 mcnths 0 I I ol JAN FEB MAR APR MAY JUN JUL AUG 5EP OCT NOV DEC I Las t 6 months JLast ths 6~mcn 0. I :=:J JUL AUG 5EP OCT NOV DEC JAN FEB MAR APR MAY JUN Use rulings only for major time divi si ons ; [ Firs; 9 ~ths= J • "t icks" for the i n-between months . 0 J F M A M J J A 5 Note that year-end lines are heavi er than ot her monthly lines when there are more than J1 ~ear::moot~s ~ three vertical rulings . Thi s hol ds for both I ~~ cal endar year data and f i scal year data . =~ D J F M A M J J A 5 0 N D' .... J':::h: I :~ I .::~ l J~· ~~~ I : ~I :~ I D M J s D M J J s D M J s D 19X1 19X2 19X1 19X2 by ~ths Note that year-are not heavy when end lines I I : I I I : 1 I they are the only rulings . D M J s D M J s D 19X1 19X2 D J D J D J J D J J D J D D 19X1 19X2 19X3 FY 'Xl FY 'X2 FY 'X3 Four years by months is about the limit for a chart that is half a page wide. A longer pe 0 ~~:~~.'~, ~::.t::::'I ' ' I' 'I ' ' I ' : I ' ' I''=· I ' ' I ' ,] • riod co ld be covered if data were for longer D D D D D intervals (quarterly or annual data). 19X1 19X2 19X3 19X4 Note that full-wi dth charts take more scale labels than half-page charts covering the same peri od . 1-years by ~:s ~ [, ~ I , ] : t o~ I 1 I I 1 I I I I I D M J s D M J s D M J s D M J s D ·= ·= 19X1 19X2 19X3 19X4 0 [s:~·~~:z.::t~·] ' , I ~ ~~~ :. I [ L I I I ' I ' I ' D J D J D J D J D J D 19X1 19X2 19X3 19X4 19X5 0 ~~~~~~~~~ .t.:::r. ,,1..........[...,,...1u. , .. ,.1....: ,..1:.::,1.~........ I, ... , .. 1 19X1 'X2 ' X3 'X4 'XS 'X6 'X7 'XS 'X9 19YO CHART CONSTRUCTION Time scales for quarterly and annual data are labeled in the same general way as monthly data. The number of labels increases as the number of time divisions increases, but when the scale becomes crowded some labels may be omitted . Quarterly Data .* In time scales up to three years long, each quarter should be labeled; but for scales four or more years long, quarterly labels should be omitted . NOTE : The standard method is to use number labels (1, 2, 3, 4, without QTR) and use the caption QUARTERLY DATA in the upper part of the grid . When both monthly and quarterly data are shown on the same chart, label the scale with monthly designations and include the words QUARTERLY and MONTHLY in the appropriate portions of the grid, as illustrated below : 4 2 3· 4 2 3 4 YEAR A YEAR 8 CIJ:,[:LJ YEAR A \ YEAR 8 YEAR C YEAR 'o YEAR E D M J S Q M J S D M J S D YEAR A YEAR 8 YEAR C Annual Data.* For time scales up to five years long, label every year, using four digits . For scales that are six to ten years long, label every year, but only the first and last with four digits, omitting the 19 in all others . For scales over ten years long, label every fifth year (see illustration) . For very long time scales (50 or more years) label only every lOth year . 19XO 'Xl 'X2 'X3 ·xs 19X6 19XO 19X5 19YO 19Y5 Split-Scale Charts . These should f ollow the system used for continuous time scales, omitting some labels as the scale range increases. The illustrations below show how to handle special cases . 3 QTR 4 QTR D J D J D 19X1 19X2 19X3 MONTHLY AVERAGE FOR THE YEAR 19X6 'XS 'YO 19Y2 19Y3 19Y4 19Y5 General Procedures . The following standard methods are recommended for general use : TIME SCALE CAPTIONS : When needed, put caption in the grid or below the time scale but not both. EXCEPTION : Omit it entirely if included in the title . (See page 83 . ) FISCAL YEAR DATA : For one fiscal year by months or quart ers, spell out (as, FISCAL YEAR 19X6) in the chart title . For more than one year, use FY in front of year labels in the time scale . For annual fiscal data, use the caption FISCAL YEARS under the yearly labels. UNEVEN TIME RANGE : In "odd" time-scale ranges (such as : 2i years by months, 23 years by years) always label the first and last time divisions , regardless of how many other divisions are labeled . = Use time caption either in the grid or below scale : but not both. D M J S D M J s D 19X1 19X2 • END OF MONTH * Recommended for half-page-wide charts: for full-page-wide charts use the labeling method recom mended for half-page charts having half the time scale range . For example, a full page chart covering eight years by months is handled the same as a half page chart covering four years . 214-853 0 -66 -6 GRAPHIC PRESENTATION SELECTING SCALES FOR BAR CHARTS Problems of scale selecti on for horizontal bar charts are much the same as those for t i me seri es column charts (see pages 82-83) . Because bar charts have no t ime scal e the scaling rel ates only to the amount scal e . As in all graphi c presentation, the shape of the chart has an i mportant effect on the scale. In a bar comparison, a tall, narrow shape minimizes amount differences; a l ow, wide shape emphasi zes them (a di rectly opposite effect than these shapes woul d have on a column chart) . Some bar charts may present scale problems because of poor layout or awkward data . The f i rst probl em i s unnecessary and can be avoided; and the second i s unavoidable, but i t can be partly overcome . Poor bar chart l ayout may result from too much space in the stub, or careless scaling. The two charts below show t ypi cal exampl es of t hese faults and how t hey can be avoided . The standard for spaci ng bar charts i s to allot at l east two thirds of the chart ' s wi dth to the bars, to provi de space for effective scaling . If necessary, abbrevi ate stub capti ons or put them on two lines. The standard procedure for scaling a bar chart i s to l et the l ongest bar come out to the f ull wi dth of the chart . If t his l ongest bar extends beyond half a scale i nterval, include the THE WRONG WAY TO SCALE A BAR CHART THOUSANDTYPE OF EQUIPMENT 0 2 4 6 LOCOMOTIVES LOCOMOTIVE CRANES FREIGHT CARRIERS SPECIAL EQUIPMENT THIS WAY MAKES BETTER USE OF THE SPACE THOUSANDTYPE OF EQUIPMEN T 0 2 4 LOCOMOTIVES LOCOMOTIVE CRANES PASSENGER CARRIERS FREIGHT CARR IERS SPECIAL EQUIPMENT scale r uling for the next scal e divi si on; other'"i se, l et the bar extend beyond the precedi ng scale di vi si on . EXCEPT=ON : In a seri es of comparabl e charts, the scale range should always be uniform, regardlesscf t he length of the bars. A1vkwar d data may result from too much or too li~tle range between hi gh and l ow amounts . Two standard methods of overcoming thi s probl em of extreme range are to : use an insert chart to enlarge the smallest bar s (see page 72) or expand the scale by "breaking" extremely l ong bars t hat t he reader ei ther expects to be much l onger t han the rest, or that are not entirel y comparable and therefore need not be shown completely. (See di scussi on on next page . ) The probl em of small differences i n amounts is harder to solve . As illustrated in the charts below, one way to show up small but important differences i s to measure the differences directly and to compare t hose instead of the tot als . These solut i ons al: ai m at the same result : bringing out differences that otherwise would not show up cl earl y . Caut i on : it is not always desirabl e t o magni fy soall di fferences. Somet imes the whol e poi nt of the comparison i s that t he di fferences are small . IMPORTANCE OF THE POINT OF VIEW PERCENT COMPLETED PERCENT NOT COMPLETED 0 50 100 0 10 20 A :::::::::::;: :::::::::::::::::::::::;::::::::::::::::::j A::;:::::::::::: ::: ::: :::::;:;:;:;:;:;:::; :::;:;:;:;:::;:] I 8 ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ::1 8 ::::::::::::::::J I c ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::1 C ::::::;: :: :;::::::::::::: A I 0 :::::::::::::::::::::::::::::::::::::::::::::::::: :::::::1 ... ::::: ::: ::::::::::::::::::1 I I MEASURING DIRECTLY VS. INDIRECTLY % OF SCHEDULE %UNDER OR OVER SCHEDULE 0 50 100 10 5 0 5 10 A 8 c ...... c 0 Location of Scal es. In all bar charts , the scale i s shown at the top of the chart directl y below the t itle. When the chart i s extremel y tall anj there are more than 3 or 4 scal e rulings , repeating the scal e at the bottom of the chart may make i t easi er to read . Scale Numerals. These are centered above the scale rulings. Speci al care should be taken to shorten numerals to prevent them from running together (see page 82) . Scale Capti ons . The standard procedure i s to use a scal e capti on for all bar charts. The capti on should be centered above the scale numeral s . No capti on is needed for a scal e that i s repeated at the bottom of a chart. CHART CONSTRUCTION BREAKING BARS AND COLUMNS Usually the only good reason for breaking a bar or column is that it is a "freak" and not comparable with other bars. Typical examples : a bar showing a high accident rate because of one acci dent in a small i nstallation; or a bar showing a large amount of unobligated funds resulting from ·Ghe cancellation of a major contract. Before breaking bars or columns, be sure that the important point of the chart will not be lost in the process : the fact that one item is much larger than the others may be the main fact that the reader should know . Broken Bars If a long bar should not be broken, these three things help: (l) Reexamine the design to see if the best possible use has been made of the space . (For example, condensing a wide stub adds more space to the length of the scale.) (2) Reassay the value of the chart : it may be important enough to increase its size, for example, from a half-page to a full-page wide . (3) See if the chart can be laid out to take a small i nsert chart that repeats the smallest bars on an expanded scale (see " 1Blowup 1 Insert Charts," page 72) . NOTE : Never scale a chart so that half of the bars will be broken. Do not include a "TOTAL" that is the sum of all the other bars and so will have to be broken (this does not apply to totals that are ratios or averages.) Never break bars by cutting off the base end of ·the scale, or by cutting out an inside piece of the scale . Broken Columns These are used in the same way as bars, and for the same reasons . Typical examples of extremely long columns that could be broken: high lost-time record because of a flood or strike; temporary rise in unit cost because of model change over. One other problem occurs more often in column than in bar charts: negative values. The standard way to trea~ negative columns is this: stretch the scale range to include them if the amount of the negative value is important; break them if only the fact that they are negative is i mportant and the amount is unimportant . How to Break Bars (or Columns) When the end of a bar must be "broken off" it should be done like this : l . Break the bar beyond the last grid line. 2. Use a bold, simple break. 3 . Do not show the square "end" of the bar. 4. Show the value of the bar in small numerals, just above or below the break . Breaking Subdivided Bars When a bar breaks in the last segment, it should be handled just like any other broken bar. When it breaks before the last segment, follow the method shown in the illustration opposite . When an unusually long "freak" bar compresses the rest of the chart too much, expand the scale to fit the rest of the bars and break the long one (see text) . Break beyond last ruling. Use a simple, bold break keep breaks uniform. Show amount above or below end of the bar : not beyond or inside Broken subdivided bars are not always effective, but can sometimes be used. 0 5 10 15 20 0 2 3 4 Breaking Bar -and-Symbol Charts beyond the end of the scale, is to put an arrowThe standard method of showing a circle, head at the end of the dashed connecting line diamond, cross line, or other symbol that extends and then show the figure above or below it. GRAPHIC PRESENTATION SPACING BARS (OR COLUMNS*) SPACING PATTERNS There is no singl e system for good bar spacing . The reason for thi s i s that several thi ngs have an effect on how thick t he bars seem : the s i ze and shape of the chart, the number of bars, the l ength of the bars, and even the shading used in them . For exampl e , bars t hat are the right thickness in a chart 4 inches wi de would be too l ong and t hin in a chart 8 inches wi de . Sati sfactory bars can be designed, however, by follm-Ting thi s basi c principle : the shorter or closer the bars , the thinner theY should be; the longer or farther apart the bars, the thicker they should be . The spacing patterns at the right show how thi s works out for half-page-wide bar charts. These same spacings could be used for column charts up to a third of a page high; but for wi der bar charts or taller column charts the bar s or columns should be made a little thi cker in relation to the space between them. A good way to do this i s to use the next higher ratio (for example 3 to 2 instead of 4 to 3) than the one recommended in the pattern chart . A s impler system i s to reduce the number of spaci ng patterns to, say, three : one for only a few bars, ~nother for a moderate number of bars, and a thi rd for a l arger number. This method is easi er for the draftsman to carry out and works fai rly well in most cases, but requires some ad justment for charts that are much larger or smaller t han the "normal" s i ze. Many of the charts in thi s manual were designed by this t hree-way spacing system, as follows: For a standard half-page-wide bar chart half as tall as the pattern chart -(see page 78 for discussion of grid s i ze) : 2 or 3 bars : bar -2 units of space, space between bars -3 units of space, space above f irst bar and below last bar -1~ units . 4 to 8 bars : bar-3, space-2 , end spaces-2. 9 or 10 bars : bar-2 , space-1, end spaces-1. For more than 10 or 11 bars, use a taller chart; for these same numbers of bars in a much taller chart , use a ratio that gives a smaller proportion of the space to the bars and a larger proportion to the spaces between bars . When in doubt, remember the basi c pri nciple stated in the first paragraph on thi s page and refer to the exampl es in the pattern chart . A still s i mpler method of spacing i s to choose the desired bar thickness and then use one standard spacing ratio for all bar charts, simpl y increasing or decreasing the height of the chart to f it the number of bars needed. Thi s i s a much easi er method and has been found generally satisfactory for the average bar chart, us ing a spacing ratio of 3 (bar) to 2 (space) . There are several reasons, however, why t his one-rati o system cannot always be us ed : (1 ) i t * The material on this page discusses spacing as rel ated to bars, butthe same principles appl y to columns . Two special probl ems that apply more often to columns are explained under "Overlapping Columns " and "Connected Columns " on page 91 . A B C D E F G H RAT IO: BAR 2 3 5 6 4 3 5 2 SPACE* 3 4 6 5 3 2 3 ··r · · ...'1'·""•r"• • · • , .. • • • · ·r • • • • • • ,. • • • • • ·r · · ..· · ,.· · · ..·r .. : t [I [ [ [ [ ~ ~ 1 1 I t ~ (This I: t [ [ ( layout [ ~ ~ is twi ce I: [ : [~ I t~ as high : [I [ I [ F t~ as the I I I [ I f I ( "normal" I I E I [ t" [ t (2 high r: 1 I~ [ I F~ t 3 wide) ~ [ I F [ c t chart. [ I :I~~~~~~ To make : I t ~ ~ ~ F E ~ it taller ~ I [ F t I [ I or shorter I [ I F [ ' simply ajd t or cut off I lc ~ : c ~ F ~ ~ I [ I ( ( ' bars to I I 1 [ 1 [ 1 ( fit the : [ I I [I~ t assigned I I ~ [ I F~ t space . t [ [I I I ~ F ~ t I I [ [ [ ( ~ t ,,I, ,,•••• J ••••••• l ...... J.......1 •••••• J•••••••L•••••• ••••••••I,,, *For tLe spaces above and below the bars, use from 1/2 to the same as the space between bars. may require more (or l ess) space t han i s availabl e , (2) it may give the chart more (or l ess) emphas i s than i t deserves, (3) no one spacing rati o l ooks equally well under all conditi ons. CAUTION: For half-page-wide charts to be reduced about the same as in this manual (16 percent) do not draw bars any thi cker than about 1/2 inch or t hinner than about 1/8 i nch . Exceptions : (1) connected bars or columns (see page 91 ) can be as t hin as 1/10 inch; (2 ) since a column chart i s spaced al ong the l ong dimens i on of a standard half-page -wi de chart~nd therefore must f it a larger space, the columns should be thi cker t han the same number of bars spaced along the short di mens i on . For exampl e, in a 2-column chart the columns could be as thi ck as 3/ 4 inch. Spacing Grouped Bars (or Columns) The aim i s the same as for ungrouped bars : to avoid bars that are extremel y thi ck or thi n . The standard spacing ratio i s 1-1-1 . .. . That i s, the space between groups is the same as the thickness of each bar . This applies to 3-bar groups as well as to the more usual 2-bar type . (See pages 45, 55 , 65, 71 and 72 for exampl es of grouped bar charts . ) EXCEPTIONS : When t he 1-1-1 rati o would produce thinner bars than in column H, above, the space between groups may be reduced (to no less than ~ the bar thi ckness)**. When the standard rati o would produce bars thicker than in column B above , the space may be increased to 1~ t he bar thi ckness . However, it is better to change the height of the gr i d to retain the 1-1-1 rati o with bars about as thick as in columns C-G . ** Overlapping bars are a better solution when workable . See page 91. CHART CONSTRUCTION CONNECTED COLUMNS Sometimes so many columns must be fitted into a chart that the only good solution is to leave out the spaces between them. The resulting chart is about half way between a regular column chart and a shaded step-curve chart . This form should be used whenever separated columns would be so narrow that they would be difficult to draw.* In a standard chart, half a page wide, this point is reached at about 20 columns . Two years by months usually should be drawn as connected columns. Three years by months should be a step curve or slope curve . (See pages 46 and 50 . ) OVERLAPPING COLUMNS (OR BARS) Sometimes double columns can be overlapped to good effect . The usual purpose of overlappi ng is to save space and at the same time to avoid very narrow columns . This is a good design when it is appropriate, a poor one when it is not . It should be used when each ofthe front set of columns is shorter than the back set. When a bar in the back column is shorter, it appears to be a narrow sliver at the side of the front column rather than a whole column (Chart 4) . In spacing this type of chart, the standard method is to overlap by 1/2 the column widt h. This holds each pair together and also makes t he spacing easier to lay out . When there are quite a few columns it may be necessary to overlap by 2/3 the column width to provide a clear separation between each pair. Usually it is best to separate pairs by no less than 1/2 the column width . Overlapping is not recommended for subdivided columns unless they are extremely simple . If the columns are divided into more than two parts, the selection of matching shadings becomes difficult . (This is one of the places where color would be effective : background columns of, say, blue, could be shaded wi th the same series of patterns as the overlapping black columns . ) Bars can be overlapped in the same way as columns, but there is seldom as much need for it. The height of a bar chart is more flexible than the width of a column chart, and usually can be increased to avoid the need for overlapping. Exception : I f a long series of double bars takes up much more space than the chart warrants, overlapping can be used to reduce the height of the chart, if one set of bars stays shorter than the other . * Tryi ng to draw a series of narrow columns exactly the same width is trouble enough : trying to keep thin white spaces between them exact ly equal takes more time and effort than it is worth . Besides, closely-spaced columns do not look separated anyway : they look like connected columns with a white line instead of a black line between them. · Connected columns are more economical also because they can be shaded much more easily. CONNECTED COLUMNS : : :::: : : : : : :: : : ::: J F M A M JFMAMJJ J M J s f'2\ SPACING FOR ONE YEAR BY MONTHS \!:.) 2 2 1-2.J 2.~-~·~ 30 -?:::.::._:::·:!--------------_--2,3 1 ~~~~ 17-: !':': r.: .. :::: ,. ·.. .. 10 .. :::: ~--Hf .__...... :: llll :: llll :: @~ :: l:l :: :_ .··:l:·:l·::;' : ::: o ~~~1~i~~~~~~~~~mu~~~~~~-:~m~~~~i ·: · :: :: J FMAMJJ A SOND SPACING FOR 6 MONTHS BY MONTHS J F M A M J CD DON'T OVERLAP SERIES THAT CROSS 30 20 10 J F M A M J GRAPHIC PRESENTATION HOWTO INDICATE GOALS AND PROJECTIONS .CTUALIPROJEC~D Many types of charts may be used to show THRU SEP goals or projections. Some of these present no special problem (such as Chart 4, page 59 and Chart 3, page 84), but many do . Two standard methods of showing such data are the "ghost" method and the "signpost" method . Charts l and 2 show typical examples of each. Note that in both methods, appropriate labels are used to indicate the type of gDal or projection used . Additional labeling to specify dates or other identifying information should be used when needed . Curve Charts The ghost method uses a different kind of line or shading for the projection or goal . The standard is a broken line, because this suggests ACTUAL (ORECAST the conditional nature of the data . Sometimes, • > especially in large display charts, an unbroken line of lighter tone or a different color can be used in the same way . The signpost method relies on a pointer to tell the reader where the actual data end . It has one big advantage: it permits the use of broken lines for other purposes . Dot or dash lines are often needed for setting off one curve from another, especially when they cross, or when several curves are shown on the same chart . Sometimes broken lines are needed for IF curves in the past instead of the future : for showing what the actual results would have been IF such and such a condition had been met; or what the effect would be IF the data included A 7G) B certain adjustments . For example, an unbroken '0'. .: '::::~ = 10'.•:·:. ::::: o/: curve measuring "patient l oad" might be paired ~ .;.; ==~= ~· \vith a broken curve showing the ef fect of chang·::m... .: .;:;; .:::.: r,::. ·· ~.·.:~.··:.:.·.· ':-' ;.·~=..~=.. . . .... ing the stay of certain cl asses of patient . =~=~ :::: :::: ~.~;:.j: :;:;: •••• ·,:. .. ..~:.:••• ::.:;::.=: Broken lines are useful also for showing aver~=~=~ :::: ;;:: ·::: . . ages, standards, or other summary or benchmark ;:;:; ~=~= ~~ll ..: ..:.::.. :;:: .. ::::: .:~:~ •!•. !·!· data . ::.::.: .... :::. :::: ::':::..::.::':::::... ·:::: ~=~: ::;: Shaded Charts ~=~=~ .:::: ·:·. ·.:..·.:. ~;;;; ... .. . . The same considerations applyto projections on surface charts : the ghost method is a good llll ~:;·:·::1 )llll 11111 .. l1l1 :;:~.:';:. one when it works; but mechanical problems are :::: ::::: :.:• .... I . . ····· •.•. =.:=:·..=::=:·. I often such an obstacle that the signpost method •••• • ::::: .: ::::==:.==:: ::::::::: !•!•! •• •••• .. .. I is a better choice . The more components there : ::::: :::: :::: .. I •···· .. :::· :.::.: .... .. are (and therefore the more different shading ::::: ::::: :::: .... :::: :::: I I I I patterns) the more difficult it is to find matching dark and light shadings for each pair of actual and estimated· pieces . r- B r .,..,. Chart 3 illustrates two methods of showing projected data in a column chart. Both are ef·.·. f '0'. fecti ve : A is better for suggesting estimated I. ~ s t-f- Jl:: f or preliminary data, B for genuine projections . .::: ~· ~. .•.. Chart 4A shows sti ll another way to suggest the \l\l •!•! ~;;; ~ i~;~ \\\\ difference between actual and projected . In this ...• :::: :::: !•!• ii!i method there is no need for lighter shadings, a ;;;; ;~;~ ~;;; t r oublesome problem when the columns are subdi•!•! 1111 ;;;; :::: •... ...• vided . 4B is a treatment similar to 3B . I t works ~;;· ~~~~ ... I wel l only when the trend in each layer of divi ::: llll ... sions is fairly steady. (See al so "Split-Scale Ill! :::: I ····· Combination, 11 page 59. ) ~;;;\l\l\ \1\l ::::: .... !Ill ;~;~ Of course the signpost method used in :::: ~;;;; ·:·: Charts l and 2 can be used for column charts \1\11 l~ll ::::· :::: ~~~~ CHART CONSTRUCTION also. In fact this treatment is recommended for column charts when they are presented with curve or surface charts that use it. Chart 5 shows two ways to add projected data on a bar chart . Method A is better because its meaning is clearer; but it can be used only when the projected data are all larger than the current amounts . When some of the future data are smaller, method B must be used . The groupedbar arrangement used in Chart 6A is the standard way to show projected data for subdivided bars. It works equally well for simple bars (6B), but uses twice as much space as methods 5A or B. Successive Projections A series of forecasts made at different times may present special difficulties. This problem occurs most often in time series charts, but can appear in bar charts also (see page 71) . The standard method of showing successive forecasts on a curve chart is to change the curve design for each forecast, as in Chart 7 on this page . It is important to choose patterns that will suggest which forecast was made first, which second, etc . (especially if they cross) . EXCEPTION : If the projected curves are so well separated that there is no chance of confusing them, the same pattern can be used for all . Labeling is often the hardest part of showing several projections . Be sure to: 1 . Keep labels as simple as possible . 2 . Try to arrange them in the order in which the estimates were made . 3. Group any related labels (see Chart 7) . If direct labeling does more to confuse the chart than to clarify it, use a key. But be sure to use a distinctive pattern for each projection. Actual-Estimated Arrows There are several methods of labeling actual and future parts of the chart, but the t woway pointer used in this manual* is about as clear and foolproof as any . It shows at once that some of the data are not actual results, it shows where the actual data end, and it is flexible enough to meet most needs . The illustration opposite shows a useful assortment of past-and-future labels . The style at the left is considered preferable (the abbreviated forms on the middle are useful when space is tight). The alternative method shown at the right, with lettering inside the pointers, somewhat simpler but less flexible in use . is Two hints for using this device: 1. When the latest 11 actual11 point is so close to either end of the chart that spelledout labels won't fit, use the abbreviated form . 2. When there are only one or two months of actual data, omit the 11 actual11 half of the arrow and its label; when there are only one or two months of projected data, omit the 11 future 11 half. * See pages 92 and 106. CD _ ~: :: ~ :: : . ~---1 A 8 ~--::::_:::~:::::::: : : :::::::::: : -~:::::::: . -~-.:~ J · :.:· .:.:::-:. ·:.:::··.::::::: ::::: /: .·.·..·.··.·.···.·.· ·.·.·.·.·. .-.·.·.·. --J 1 ·:::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ::: =:J .......... .... . ... --, ..........·.·.·.··.·.·.·.·.·.·.·.·.·. __ J ·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.·.· .· .·.·.·. --, :-:-:-:-:-:-:-:-:-:-:-::-:-:-:-:-:-:-:-:-:-:-:-:-::-: --J A ® 8 \ :·:·:·:·:-:-:·:·:·:·:·:·:·:·:·:·:·.·:·:· \· ~-~-~-~-~--~::.-=--_j ; ::::::::::::::::::::::::::::::::::::::::: r-------_-: ~:: ~ ·:::::::::::::::::::::::::::::~:::::::1 \-----J .. ........ .. .. . .·.·.·.·.·:·: ·:·:·:·:·:·:·:·:·:·:·:·· ,_------_J 30 CUM. FROM I JUL Xl INCL. 0 20 31 AUG X2 ESTIMATE 10 19Xl 19X2 l9X3 19X4 19X5 19X6 FISCAL YEAR ACTUAL PROGRAMED rT.,PROG) (ACTUALJ PROGRAMED) ;cr., PROJ> (ACTUAL IPROJECTED> rT.,F'CAS> (ACTUAL IFORECAST > tT.,EST. > (AcruAL lESTIMATED) ;cT., PLAN> (ACTUAL IPLANNED > tT.,SCHE> (ACTUAL ISCHEDULED) GRAPHIC PRESENTATION HOW TO SELECT AND USE SHADING SHADING Shad i ngs look best when all are about the same "coarseness " . Very coarse or unusual patterns are hard to look at and distort the bar . Even-spaced patterns are better for .M charts than ~ i rregular patterns . White and/or black can be used for very small segments . They help to get better contrast be-~~~~~~~~~ tween other shadi ngs . (See ~~~~~~~ al so examples on pages 52-53 . The standard order of tones is dark to M light . (See ~ text for ex cepti ons). The patterns should be placed at a 45° angle. Verti cal or horizontal M designs dis-~ tort the area shaded. ~ ~ Vertical shading on bars or horizontal shadings on columns have another disadvantage: in small segments they are confusing because they look like two or three smaller di vi s i ons. The kind of shading used and the way i t i s usedfu exceedingly i mportant. I n fact, an otherwi se gocd chart may be ruined by poor shading . Selecti~g Shading Patterns Shading patterns should not be so coarse that they di sturb the eye or distort the areas they shade, nor so fine that they fill i n when reproduced . I t i s sometimes hard to stay between these extremes ; but i t ~be done much more often than i t generally i s . If patterns are too coarse, they are too noticeable. Shadi ng should serve its purpose qui etly and effi ci ently wi thout calli ng itself to the reader 's attenti on . Patterns that are to be reduced consi derably in reproducti on should be coarse enough to stand the reduction, but peppermint-stick or checkerboard patterns are seldom justi fied . Di fferent patterns in any one chart (or in a seri es of charts presented together) s hould be about the same coarseness . I f one pattern i s coarse and the next fine, the coarse one seems closer to the eye and it breaks up the unity of the chart . Thi s sort of contrast has about the same effect -and the same l ack of logi c -as outli ning one bar with a thin l i ne and the next wi th a thick line. Shadi ng -~rangement The safest scheme is to put the strongest tone next to the base l i ne , the weakest the f ar thest away . Someti mes thi s arrangement can be changed to good advantage, but i t i s the best general acheme to follow. Sometimes -especially when several shading patterns are used -there i s a temptation to alternate the tones~ f i rst light then dark then light then dark . However, thi s should be avoi ded as much as possi ble because i t breaks up the unity of the chart and makes i t hard to look at . Shadi ng :or Sense Sometimes the logical order of the data does not put the most i mportant part close to the base l i ne . For example, i n comparing how the cost of two items is divided into materi als, manufacturing, and di stributi on, the main interest might be in manufacturing cost . Shadi ng t hese components i n the standard way would put the emphas i s in the wrong place . Sometimes the nature of the data leaves no choi ce of how to shade the chart. For exampl e, a chart divided to show the number of cl oudy days and sunny days would look more sensi bl e i f a darker shading were used for "cl oudy, " a light er shading for "sunny . " I n the same way, data di vi ded into such categories as white and nonwhite, blacktop and concrete, would make more sense if shaded according to the nature of the subject. CHART CONSTRUCTION These exampl es point up the rule t hat the choice of shadings s hould always be t hought through, and ~ot l eft to whim or habit. Al though the answer i s seldom as obvious as in these illustrations , most charts can usually be worked out so that the shading helps to tell the story instead of hieing it. Using Black and White Shadings Although black and white are not recommended f or standard use, sometimes both can be us ed to very good advantage. Black is easy to add to the chart, white requires no work at all; and since the chart designer often needs more tone steps than he can get wi th shading patterns, these two extremes are frequentl y useful. White is usually a poor tone (on white paper) because it does not set off the shaded part from the background . I t i s never effective for large areas; but it i s sometimes good for small ones if the outline of the bar or column is heavy enough to set it off from the background. In fact, for very small segments , white i s often better t han a shading pattern becaus e it is l ess likely to f ill in and be lost in reproduction. Black i s good for small segments but poor for large ones . Black i s so strong t hat i t can very easily dominate the chart if i t is not used wi th restraint . However, as shown in t he accompanying i~lustrations, it can hel p to s implify many shading problemsand to make the chart easi er to understand. Shading Problems Double or triple bars usually should have the stronges~ shadi ng in the top bar because this usual ly shows the most recent or most important data . If a double bar cons i sts of an actual result and a program or forecast, use the stronger shading for the accompli shed fact . Double or triple columns should follow the same shading principles used for grouped bars. Overl apping bars or columns s hould al ways have t he stronger shading: the overl apped data should be lighter (see page 91). When subdivided bars, columns , or surfaces change from actual data to forecast , program, etc ., avoid two sets of shadings unless they match in pattern and gradation (see pages 92-93). When shading is added to a time series surface chart to bring i t up to date, remove the shading back to t he nearest verti cal grid l ine , then apply a new area wide enough to include the current data . SUGGESTED SHADINGS 1 Shading Charts wi t h mostly long bars need a lighter tone t han t hose with mostly short ones . 2 Shadings The same rule applies to parts of bars also. (Use black or white onl y for very small divisions.) 3 Shadings Note that different combinations are used for different di stributions of parts within the total . (See text . ) 4 Shadings The same principles apply when shading f our segments . 5 Shadings About the limit (unless color can be used) . Note: For the pattern numbers and other information about these shadings, see page 96 . GRAPHIC PRESENTATION Shading Patterns for Different Reductions Many shading patterns, printed on transparent f ilm with a dry adhesive backing, are available commerci ally . The four series of shadings illustrated below should meet most chart needs . The patterns listed in the left column are for charts requiring only a little reduction. This list includes t he shadings used for most of the illustrati ons in this manual . The second and third lists are for charts planned to take a moderate to heavy reduction. The fourth series i s for large di splay charts that are planned for reducti on to letter size . SHADINGS FOR DIFFERENT REDUCTIONS ~ For Reductions of: Tone Up to 1/3 1/3 -1/2 1/2 -2/3 2/3 -4/5 Q/ Q/ Q/ ......... ... . ............ ........ .... • • • . 1 :::::::::::::::::::::::: • • ............ . • ................... ...... • .• gJ 83 3 84 gJ e!7 J . 93 ////////. &fff}~~~f}l}~ ////////. / //// 2 ////////. :~~~~~ff~~~~~~~ ////////. /// / / ///~ gJ 501 T-B 502 000 gj gJ 269 3 • • • ~~~• • • • F-6 258 36 266 79 268 83 92 4 ~~(XX> 22 305 62 306 gJ 407 ·~~~.J 5 •••••••••• •......•.. 205 56 95 97 208 6 --~~ 304 25-R 307 27-R 209 7 1111111111 111111111 -- 29-R 106 ~ 8 7-R gj 7-R gj 7-R gj 7-R 5f Figures under each sample are manufacturers ' pattern numbers . Numbers at the left are ZIP-A TONE numbers, those at the r i ght are CRAF-TONE numbers . (A thi rd brand -ADD-A-TINT -i s avail able in the same patterns as ZIP-A-TONE . ) Pat terns i ncluding the letter R are printed in red (whi ch reproduces black) . The red i s transpar ent, and so enables the draftsman to see the bar or surface outline the shading must be cut to fit . The ori ginal samples have been reduced 16 percent in this manual . 2/ These patterns will stand up under a greater reduction than the top limits suggested, but only when reproduced with extreme care on a good quality paper. ~ Not available in ZIP-A-TONE. ~ Not available in CRAF -TONE. 96 Notice that in each seri es all the patterns have about the same coarseness . Two reasons for thi s are that : (l) a shading that is coarser than tLat for the remainder of the chart i s hard tc look at and distorts the chart, and (2) a shading that is too fine for the reducti on used will look "muddy" after reproduction. NOTE : Eight patterns are included in each seri es to provide a wide selecti on. This does not mean that all eight should be used on the same chart; in fact some of these patterns are merely alternatives that produce about the same effect . In many charts, more contrast is needed than i s provi ded by these successi ve steps . When no more than four shadings are required, good contrast is provi ded by using alternate patterns . Some of the patterns in the fourth series are not quite as coarse as they shoul d be for a 4/5 1 s reduction. However, they are the coarsest patterns availabl e . (The secti on on "Shading Di splays" in the Display Presentation chapter suggests some other methods for making the l arge di splays . ) Sometimes i t is necessary to use a pattern from the next coarser series, to provide a clear di sti nction between several patterns on the same chart;but this should be done only when unavoidable . Mi xi ng up coarse and f ine patterns "breaks up" the chart (see "Selecting Shading Patterns," page 94) . Whenever there is any questi on about the quality of reproduction, choose a shading series that wi ll be sure to reproduce clearly. It i s better to use patterns that l ook a little coarse, but are easy to di stinguish, than to use f iner patterns that fill in. However, coarse patterns are undesirable on a small chart or oL any chart that is viewed from close up. Do not use patterns that are coarser than are actually needed to show a difference between one pattern and another. SUBSTITUTE SHADINGS ~ For Reductions of : Tone Up to 1/3 1/3 -1/2 l/2 -2/3 2/3 -4/5 Zip Craf Zip Craf Zip Craf Zip Craf 1 87 75 4 85 33 81 83A - 2 - - - - - 814 97 208 3 01 255 - 265 - 207 61 97 4 85 61 X 302 - - - - 5 9 33 54 261 29 94 - 269 6 - 301 - - - - 20-R - 7 - 103 - - NH-1 - NH-5 109 8 An..v sjtisfa1ctor~ methld of Fettinr blalk. 5I These are simil ar to the suggested patterns shown in the table at the left, and may be substituted with reasonably good results . Pattern numbers at the left are ZIP-A-TONE; those at the right, CRAF-TONE. CHART CONSTRUCTION SELECTING CURVE PATTERNS Curve desi gn, like shading, should be carefully thought out . A thoughtful choi ce of curve pattern and weight can make the chart easi er to understand and can help bring out the facts the reader should learn. It i s not necessary to use a different design for each curve in a chart . The only times when different patterns are needed are: (a) when curves cross ; (b) when the nature of the data demands i t. In most charts each curve does not cross other curves and one pattern can be used for more than one curve wi thout any chance of confusion (Chart l). When curves do cross, i t i s not al ways necessary to use different curve patterns . Varying the weight will often work j ust as well (Chart 2) . Sometimes the nature of the data suggests the desi gn of the curve; sometimes i t practically di ctates it. Broken-line curves, for exampl e, are appropriate for showing estimates, projections, averages (when compared with the unaveraged data), and schedules or programs. Either dot or dash lines maybe used for these purposes . (See pqges 92-93 . ) It i s desi rable to avoid a different pattern for every curve because : l. Only two or three patterns are easy to draw. Except for the continuous or soli d-line pattern -all patterns are time-consuming and hard to draw uniformly .* 2. Changes in pattern automatically change t he emphasis; and there i s no simple way to control these differences in emphasis . In selecting curve patterns, make important curves prominent and less important ones less prominent. When the emphasis on all the curves i s t he same , make dot or dash curves slightly heavier than soli d curves so that they will appear to be the same weight . The following recommended patterns will meet most needs : Solid line -for a single curve and for any other purpose i t can be made to serve . Dash line for showi ng projection of a solid line curve, averages, upper limits (such as ceili ngs or quotas), lower l i mits (such as mi nimum requi rements), and in general for all "ghost" curves (such as curves showi ng how the data would l ook i f certain adjustments had not been made, or how i t mi ght look i f certain changes were made . ) Also, a dash line curve i s * "Prepared" curves (preprinted on a transparent adhesive film) are avai lable in a wide range of dot patterns and some dash patterns . They provi de more uniform results than most draftsmen can draw and are excellent when they can be us ed. They are best on large simple charts; for a small chart or for many close plottings this method i s extremely tedious . ..... . . .... ....·· ··. ... ········ a good second choi ce when a solid line curve cannot be used . Dot curve -for all the uses listed above for dash curves, especially when a dash pattern must be used for one of the main curves on t he chart. Dot curves are good for showing revi sed estimates or for projections (when original estimates are shown by a dash-line curve) . Do~ dash curve -sometimes necessary when a fourth pattern i s needed . Easily confused with a dash curve unless carefully drawn. Ball and l i ne curve for data plotted at irregular interval s or when i t is i mportant to emphasize the plotted points. (For example, when data are plotted in t he spaces between vertical rulings.) This works best when the plottings are not too cl ose together. Circl e and line curve as a secondary curve to accompany the ball and line pattern. Thi s also can be used for estimates , projecti ons , etc ., by making connecting lines dashed instead of solid . STANDARD PATTERNS The basi c three: Solid } Dash will meet most Dot charting needs. Dot Dash } Addi t i onal pat-Ball &Line ~erns for spec• • • c· l &1 . 1al problems . lrc e lne See text. These patterns, in varying wei ghts when necessary, are the recommended standards for statisti cal charts . GRAPHIC PRESENTATION PREPRINTED PARTS IN CHART CONSTRUCTION USEFUL "STICK-ON" PARTS Preprinted materials are not entirely effective for making charts because they can seldom be rigidly standardized . Such materials should be used, however, when they will speed up the work and improve the appearance of charts . I t is often possible to preprint component parts such as labels, titles, time scales, or any other elements that are used repeatedly. For example , chart or page headings (such as Training, Military Personnel, Health, Safety, etc.) used month after month in a recurring report, might well be prepared in quantity in advance and then used as required . When the same layout can be used for several charts, it is sometimes profitable to reproduce copies of the entire chart framework in advance (including lettering and any other component parts that are common to all charts) . Even plotted data can be preprinted : for example, the average for a group of installations could be included on reproduced skeleton charts that would each be filled in with the data for one installation. In addition to "tailor-made" parts, several commercial materials are useful as prepared parts for chart construction. Some of the most useful of these are "type" and "tone" sheets of thin transparent film with a pressure-adhesive backing . y "Type" sheets (for example, ADD-A-TYPE, ARTYPE, CRAF-TYPE) contain alphabets, nu~bers and letters in series, arrows, asterisks, and other useful graphic devices . The illustration at the right shows some of the items available on these sheets . The two most useful for ordina r y charting are the long arrows (used as label pointers) and the broken lines (used for curves and other chart lines) . This cut-and-press-on method is not always practical for chart curves . A curve with a great many close plottings would require so many small patches that it would be more trouble to put on than an ordinary pen-and-ink curve . Also there are times when the chart needs a different weight curve than these preprinted lines provide . However, they are worth remembering : often they can help to cut production time; and like all preprinted parts, can help the draftsman to turn out a better looking chart . In addition, they help to get uniformity throughout a series of charts drawn by several draftsmen. The lines in the lower part of the illustration were cut from shading "tone" sheets (ZIP-A-TONE, CRAF-TONE, ADD~A-TINT). 2/ Some of these lines fit charting needs very well . However, the chore of slicing one line from a shading pattern and putting this on the chart is difficult unless the pattern is coarse enough to provide a fairly wide strip. 1/ This is a dry wax adhesive . Sheets must be bur nished down to adhere evenly and firmly. ~ Examples of shading patterns are shown on pages 94-96 . ARRq.JS AND LINES Sheet No. } 3ooo IV 1009 £1 A355 g./ } 2001 tJ} 1005 Q/ A379 g./ 2o73 IV } 3007 tJ} 1010 Q/ A361 g./ } 3020 tJ} 1012 £1 .....•............•••••••••..••.....•.•. A373 gj -----------------}3009 tJ} l Oll Q/ A363 gj LINES FROM SHADING SHEETS Pattern No. T-B gj ---------------------}501 §/ 629 £/ } 79 w 259 §/ 7l3f./ } 54 w 268 §/ 837 £/ } 97 w 208 §/ 850 £/ 1 27-R s!/ ~ 209 §/ ~ 835 £/ a/ ARTYPE (Artype, Inc . ) b/ CRAF-TYPE (Craftint Mfg . Co.)c/ ADD-A-TYPE (Arthur H. Brown & Co . ) d/ ZIP-A-TONE (Para-tone, Inc.) e/ CRAF-TONE (Craftint Mfg . Co . ) £/ ADD-A-TINT (Arthur H. Brown &Co . ) The two-way arrows shown at the bottom of page 93 are excellent subjects f or preprinting. They are such a nuisance to draw that a ready supply of them is a time-saver as well as a convenience . These devices can easily be made by offset reproduction . They can then be pasted on with rubber cement or two -sided adhesive tape . Security labels can also be reproduced and stocked for future use. These are also obtainabl e through regular supply channels, in three sizes, on transparent film sheets (ARTYPE No . 3031) . CHART CONSTRUCTION CHART CAPTIONS Chart captions are titles, subtitles, general notes, and any other words that apply to the chart as a whole . Most captions -especially titles tend to be too long . They should be long enough to be clear and accurate; but clarity and accuracy are not necessarily gained by using more words . Good captions choosing the right words properly. are and the result of arranging them Titles -----The main title should give the reader a quick, clear understanding of the content of the chart . The title (and in fact every caption on the chart -but especially the title) should be written as much as possible from the reader's point of view . The words should be as simple as possible and the concepts behind these converted into familiar ideas. Location. A title is normally centered above the chart grid, but individual titles may be placed inside the grids when several r elated charts are covered by a single main title. Length. One-line titles are usually best; two lines are sometimes needed, but never three . One way to keep main titles short is to put secondary information in a subtitle . For example, the title : OFFICER, ENLISTED AND CIVILIAN OPERATING PERSONNEL IN WASHINGTON, D.C. METROPOLITAN AREA can be rearranged: OFFICER, Et-jLISTED AND CIVILIAN OPERATING PERSONNEL In ~ash+ngton, D. C. Metropolitan Area Many titles are too long because they contain information that belongs somewhere else in the chart . The title above is a good example of this . If "officer," "enlisted" and "civilian" appear on the cliart as labels, the title should be cut to eliminate these words . Avoid the poor practices of merely repeating labels in the title or of cataloguing every detail of the chart in a subtitle . This is an example of the second fault: ARMY CIVILIAN PERSONNEL Actual Strength, Ceilings, Projections 19Xl1X5 by Months When long titles must be broken into two lines, break them for sense: at a logical break in their meaning . 1For example : I Use : DECORATIONS AND AWARDS TO U. S . ARMY PERSONNEL Or : DECORATIONS AND AWARDS TO U. S. ARMY PERSONNEL But not : DECORATIONS AND AWARDS TO U. S. ARMY PERSONNEL If a title qan be broken in two places, use the one that will avoid two exactly equal lines or one very long and one very short line. Abbreviations should be avoided in titles as a general rule . EXCEPTION : When a term is repeated in a series of chart titles, abbreviate not only to keep the titles short, but also to help the reader see the difference between titles more quickly. Subtitles A subtitle is a secondary caption that explains or qualifies the main title . Subtitles may be used for such items as date (see page86), index base period, and limits of coverage . They are usually placed beneath the title, but may be put within the grid . General Notes A general note applies to the chart as a whole. If it should be seen before the chart is read, it is put at the top of the chart (between the title and the grid); if it is very long, or not important enough to need first attention, it is put at the bottom of the chart, aligned at the left side . A general note is generally preceded by the word "NOTE:" . Reference NQtes These are special explanations about a portion of the chart or data . In general they should be used only when the reader needs them to understand the chart properly. A reference note is normally put at the bottom of the chart directly below a general note if there is one . It is sometimes possible to put a short reference note directly at the point i t refers to (for example, a dip in production might be labeled with the word "strike"). Reference notes are put one under the other in order of their references in the chart . When they are very short, or when space is limited, several notes may be put one after the other. Every reference note starts with a reference mark matching the mark shown at the point of reference . When only one mark is needed the asterisk (*) is the standard . When more than one reference mark is needed, use lower case letters set off by an underscore and slant (a/, b/, c/, etc . ) . --- For an example of how reference notes are handled, see the upper chart on page 45 . Source Notes These are capti ons that tell either the source of the data or the source of the chart or both. Charts in reports do not need to note the chart source except on charts prepared by a different office than the one responsible for the report . The source of the data should be noted whenever the reader needs~know which of several likely sources the data came from or when the publisher of the report does not assume responsibility for the data . Source notes appear below all other notes (if any) at the left side of the chart, preceded by the word "SOURCE ." GRAPHIC PRESENTATION HOW TO USE LABELS LABELS When there is only one curve or one set of simple bars or columns, the title and other chart captions usually provide all the information needed . But when the chart carries more than this, more identification is needed . Such a chart can be clearly understood only if the graphic parts of it are labeled carefully and completely. Direct labeling is almost always preferable to a key or legend because it provides the quickest and most accurate understanding of a chart . Location. A label should be close to the thing it labels but not quite touching it. A good rule of thumb is to place the label about 1/2 to 3/4 the letter height away. Do not scatter the labels all over the chart: try to arrange them in an orderly way. Do not put labels directly opposite the ends of bars : place them either above or below . Do not put labels on a slant . Do not zig-zag them to follow the shape of the curve . NOTE: Sometimes several label arrangements should be tried in order to find the best one. For all but the simplest charts, make labels as separate pieces that can be moved . Do not trim labels too close : leave a margin of white space around the outside (about 1/2 the height of the letters -no less than 1/20") . "Pointers." If the label cannot be put near enough to show clearly what it relates to, it can be placed well away and connected by a "pointer." The standard pointer is a straight line with a small arrowhead (see page 98). Keys . Legends and keys formerly were widely used in statistical charts but they generally should be avoided except for series of charts that use exactly the same labels . Unless keys can be memorized easily, they waste the reader ' s time and they are likely to cause confusion. If used at all, i t is best to put the graphic samples first, then the explanations . POOR BETTER POO~ BETTER ALL CAUS ES BATTLE I NJURIES BATT LE AND WOUNDS DISEASE INJURIES /8 WOUNDS / DI SEASE ,..It' ,...!., Place labels to read in order: avoid having arrows point in too many different directions . S MAL L ARMS gill ~ LI GHT 8 MED AR TI LLERY 1.1 Keep labels all inside or all outside the parts they label: otherwise they cut up the chart . U. S. Relate all labels to a single bar, and if possible arrange them in the proper order. Leave a margin of white space around labels : Avoid putting labels inside bars: put them above do not put them too close to rulings. or below; if too crowded, overlap slightly. CHART CONSTRUCTION ARROWS AND OTHER LABEL POINTERS If labels are placed carefully and put cl ose to the t hi ng t hey label, there i s us ually no questi on about which labe l relates to what . But sometimes the most careful pl acement still l eaves the reader in doubt . In such cases some sort of pointing devi ce is necessary. Although it may seem like a minor matter, the kind of poi nt ing device used and t he way i t is put to use can have an important effect on t he clarity of the chart. The purpose of a pointer i s s i mply to point to something . Thi s can be done with a fai r l y l i ght, simple line with an ar rowhead. Fancy or overwei ght devi ces are more trouble to make and di str act the reader ' s attention from the thing they point to . Hi nts that will help to place poi nti ng de v i ces effectivel y follow : l . Try to put pointers where they will not cross a curve or some other part of the char t to whi ch t hey relate . 2 . Use strai ght lines. Curved or bent lines add one more unnecessary flourish and take l onger to draw. 3. Make the l i nes l ong enough to see . If t he space i s just barely bi g enough to take an arrowhead, a pointer i s not needed. 4. Draw t he l i ne at a sharp angle (45 ° i s good), preferably in the di recti on opposi te t he curve. Vertical or horizontal pointers do not stand out from the grid l i nes . 5 . When more than one pointer is needed, draw them at the same angle and , i f possible, in the same di rection (see page 100) . 6 . Do not make the line thicker than the l ettering to which i t relates -thi nner is better. 7. Do not make arrowheads too bi g : a l ong one i s better than a s hort, fat one . * NOTE : Several manufacturers make arrows on transparent press-an f ilm. The ones used in this manual are ARTYPE, Design# 2073 (see page 98) . * However, attention arrows (the short, squat kindthat are all head and no body) are excellent for marki ng an important spot on the chart, such as the l ocation of V-J Day or D Day, or some level such as a year-end objecti ve . TYPE C ADJUSTED POINTING DEVICES FOR LABELS ARROW LINE BENT ARROW-LINE GOOD ARROW UNDERLINE FAIR BOX AND ARROW CUR VED ARROW LINE ARCHI TECTURAL ARROW POOR ARROW HEAD TOO SMALL ARROW HEAD TOO LARGE WHEN TO USE ARROWS POOR PREFERRED AVAILABLE FUNDS EX PENDITURES Use arrows when curves are crowded or when labels are too l ong to fit close to curves , but do not use more than you actually need . OUTSIDE U.S. CONT. U.S. -/ " \ \ CONT. U.S. \ \ \ \ ' ' Arrows are especi ally useful in l abeling subSometi mes another arrangement of the l abel s will divided columns or "interlocked" curves . remove ~he need for any pointers . GRAPHIC PRESENTATION CHART LETTERING The aim in lettering statistical charts is to use lettering that is easy to read yet does not overpower the rest of the chart . When the charts are used in statistical reports, extra care must be taken to avoid the extremes of lettering that is too big or bold for the text mat ter or too small or weak to be legible.* The lettering used throughout this manual is the recommended standard for report charts . It is a single stroke "gothic" · (block) letter made with "mechanical" lettering guides . The standard lettering sizes made with these guides and the appropriate pens are listed in the following table . LETTERING GUIDES AND RECOMMENDED PEN SI ZES LEROY WRI CO 2/ Letter Height !:/ Pen Line Pen Line (Capitals) Number Weight Number Weight .080" 00 . 013" 13 . 013" . 100 0 . 017 17 .017 .120 0 . 017 17 .017 .140 1 . 021 21 .021 .175 2 .026 26 .026 . 200 2 . 026 26 . 026 .240 3 . 035 35 .035 . 290 4 . 043 43 . 043 . 350 4 . 043 43 . 043 .425 5 .055 55 . 055 . 500 6 . 067 67 .067 NOTE : Letters larger than . 500 (~")are not needed for report charts . For .table of larger guide sizes, see Display Presentation chapter . !:/ Guides are referred to as number 80, 100, 120 , 140, etc. EJ Scriber type; also made in stencil type . St andard Sizes The table at the upper right tells what size guide to use for each lettered part of the chart . Note that except for two items the sizes are one step larger for full-page -wide charts than for half-page-wide charts . These sizes are recommended for report charts to be reproduced same size or to be reduced no more than 20 percent in reproduction (this manual was reduced 16 percent) . As a rough rule of thumb, increase these sizes one step (on the original chart) for each additional 15 percent reduction. For example, in a half-page -wide chart drawn for a 35 percent reduction, make the title size 140 instead of 120; for a 50 percent reduction, make the title 175 . * When making displays such as large wall charts, easel charts, or charts for projection, the major problem is to get the lettering big enough . For displays, the lettering should be larger and heavier -in relation to the rest of the chart than for charts in printed reports (see Display Presentation chapter). NOTE : The sizes specified in this tabl e were selected for charts to be used wi th elite typewritten text or tabul ar data . However, they are equally eff ective for charts that appear on report size sheets without text or tables . LETTERING SI ZES FOR STATISTICAL REPORT CHARTS Chart Width Chart Component Half Full Page Page!:/ Chart Number E/ .080 " . 080" Main Title .120 .140 Subtitle :::/ .100 .120 Amount Scale Capticns .3/ . 080 .100 Numerals .100 .120 Time Scale Months or years only e/ .100 .120 Months wi th years ~-. 080 .100 Years with months .100 .120 Time captions .3/ . 080 . 100 Labels !) . . 080 .100 Bar Charts Stub headings . 080 . 100 Stub items :::/ .100 . 120 Footnote .080 .080 a/ For charts one -third of a page or more high . b/ If used . S/ Use one size smaller ~ captions are long and cannot te abbreviated . d/ Make one size larger if used as subtitle . e/ Use same size for quarterly designations . I/ Use one size larger for short labels such as WW II, V-J DAY . Letter Count for Chart Titles and Notes The tabulation below gives the average numbe.r of CAPI TAL letters that will fit in one line without crowdi ng . This is useful for estimating how many words will fit on a line, and for helping to judge whether long titles and other captions need to be shortened . Average Number of Letters per :Guide Chart Wi dth rij Size Inch Half Page Full Page 80 12+ 50 104 100 10 40 83 120 8+ 33 68 140 7+ 29 60 a/ Based on width of chart originals for this manual : 4" and 8 . 3". A space between words should be counted as one letter; but ·if the letter count is high, word spacing can be squeezed slightly. Captions with several letter I ' s or figure l ' s will take less space than indicated; those with a large number of M' s and W's will take more . CHART CONSTRUCTION Spacing Lettering Good l etterlllg requires good spacing . Poorly spaced letteri ng not only looks carel ess but i s hard to read . Lettering guides are a great help i n drawing even, uniform letters; but they are no help in spacing. Spacing i s a matter of eye judgment, not mechanical measurement . Letter Spacing The nearest approach to a rule for spacing l etters i s to make the AREA between them roughly equal. Equal LINEAR di stances will not produce good l ettering. WHITE SPACE APPROXIMATELY EQUAL-CORRECT SPACING LINEAR DISTANCE EQUAL -INCORRECT SPACING u u UL.J u In spacing l etters, t he amount of space inside the l etters should be considered . For example, two O's must be spaced cl oser together t han two N' s . A safe rule for making lettering readabl e is to use about l /4 to l/3 the letter height as t he normal space between letters : l . Put thi s normal space between all parallel-s i ded combinat i ons (AV, HE, NE, WA, MP, etc . ) ; 2. Put round l etters (like 0 , C. Q and G) and overlapping combinations (like LY and TA) a littl e cl oser together than parallel-s ided combinat i ons ; 3 . Give the l etter I a little more than normal space, because it has no white space within i t . Word Spacing The space between words should be about 3/4 t he l etter height (the average letter wi dth) . The space between sentences shoul d be half again to twi ce as much as between words . Line Spacing The space between lines should never be l ess than 3/4 the letter height : more i s generall y desirabl e . Never use less space between lines than between wor ds , otherwise the lettering will l ook like a mass of words instead of a seri es of lines. The table at the right shows how much space to use between lines of l ettering (or between l ettering and grid lines) . Thi s spacing i s illustrated on pages 105 -107 . The second column of the table shows the relative distance to use between lines, regardless of t he s i ze of the chart or the l ettering s i zes used . The following four columns show in inches these relati ve distances as applied to standard-wi dth charts us ing the l ettering s i zes recommended in the tabl e on the right side of page 102 . Wherever feasibl e , these measurements have been rounded to permi t maximum use of the l Oths and 20ths divisions of a triangular scal e . RECOMMENDED SPACING BETWEEN LINES Use t his amount of space from bottom of one line to : Bottom of In Top of next E) next !2/ planning these Rela-Inches Inches lettered t i ve parts : Dis-Half Full Half Full tance Page Page Page Page rJ Wi de Wi de Wi de Wide ALL CHARTS Chart No ., Ti t l e* l .13 .14 .25 .28 2-Line ':'i t l e 3/4 .08 .ll . 20 .25 Title, Subti tle* l .10 .13 .20 .25 2-Line Subtitle 2-Line Label 3/4 3/4 .08 .07 .08 .08 .18 .15 .20 .18 TIME SERIES CHARTS Title*, Gri d g/: 1-Line Scale Cap. 2 .25 .30 --- 2-Line Scale Cap. 3-l/3 .40 .45 ----Subtitle*, Grid Q/: 1-Line Scale Cap . 2-l/2 .25 .30 --- 2-Line Scal e Cap . 4 . 40 .45 --- 2-Line Scale Cap. 3/4 .07 .08 .15 .18 Scal e Capti on, Grid l .08 .10 ----Numerals, Gri d l .10 .10 ---- Time labels, Grid: Months !ij' w/years l .07 .10 .15 .20 Years w/months !ij' 2-l/2 .25 .33 .35 .45 Mos . or Yrs . Onl y l .10 .13 .20 .25 Time capti ons*, Time Labels l .07 .10 .15 .20 Footnote*, Ti me Designations l-l/2 .12 .15 .20 .25 2-Line Footnote 3/4 .07 .08 .15 .18 BAR CHARTS Title or Subti tle, Scale capti on* l -l/2 .12 .15 .20 .25 Scale Caption*, Numerals l .08 .10 .18 .22 Numerals, Rulings l /2 .05 .05 ----Stub heading i/ ----------Stub i tems, Grid l .10 .10 ----Footnote, Rulings l -l/2 .12 .15 .20 .25 * See footnote rJ. £I Excluding hei ght of l etter . Q/ Including he ight of l etter in second line . rJ Stated in terms of letter height of t he i tem in the stub of this tabl e (the i tem mar ked with an aste~isk when l ettering s i zes differ); e .g., make space between chart No . and t itle the same as the height of t i tle letters . gj For column charts, betvreen t itle or subti tle and top of columns when extending above grid. !d Or quarters. £I Cent er oppos i te scal e caption and numerals . GRAPHIC PRESENTATION LINE WEIGHT STANDARDS STANJARD LINE WEIGHTS FOR REPORT CHARTS The tabl e at the r i ght speci fies line Line Pen wei ghts for each graphic part of a chart. Chart Type and Component Weight Size These standards are ill ustrated i n the chart 111 Q/ speci f i cati on s heets shown on the foll owing pages . CURVE AND SURFACE CHARTS Fairly compl ex charts, such as those showi ng several series or estimates, or both, may need several line wei ghts . The tabl e l i sts recommended wei ghts for curve or surface charts that are ei ther a full-page or a hal f -page wide and wei ghts recommended for column or bar charts . Many charts, however , have so few components that t hey do not r equire all of t hese wei ghts . Many s i mpl e line or surface charts need only three weights : one for the base line , one for the scal e lines , and one for t he curve line . These line wei ght standards are for charts to be reduced up to 20 per cent i n reproducti on . * I f the reducti on i s to be mor e than 20 percent, i ncrease the line weights as fol lows : Line Weights Line Weights for a Reduction of:for a 0-20% 30 % LIJ% 50% Reduction . 010 11 . . . --.01211 --.014" --. 017" . 014 . . . --.017 --.020 -.024 .017 . . --.020 . 024 .029 .021 . . -.025 -. 029 -. 035 . 024 . . . -.029 . 034 -. 040 - . 028 . . . . 034 -. 039 . 047 - - . 030 . . . 036 . 042 -. 050 . 035 . . . .042 .049 . 059 --. . 043 . . . --. 060 --. 072 - . 05 2 When charts are to be drawn for a heavy reducti on (of 50 percent or l ess of t he ori gi nal s i ze) avo i d us ing line wei ghts that vary just a l ittl e because after reduction there will not be enough di fference to tell them apart . Lettering Pens for Rul i ng Leroy and Wri co l etteri ng pens can often be used for l i ne work s uch as curves, bar and col umn out lines, and scale rulings . Many draftsmen prefer them to ruling pens, especi all y when drawi ng heavy lines . They put a f l atter f i lm of i nk on the paper, and they provide f i xed l i ne wei ghts that avoi d the tri a l s and adjustments need ed to match l ines wi th a r ul i ng pen . The tabl e on the l eft of page 102 shows t he nominal l ine wei ght -i n thousands of an i nch of both Leroy and Wri co l etteri ng pens. The line -wei ght standards were selected with these pens i n mi nd . Note , for example, that Leroy pen number 0 makes the l i ne weight ( .017") recommended for year-end verti cal r ulings; Wrico pen number 35 makes the line ( .035 11 ) recommended for s i ngl e curves ; etc . * These speci f i cati ons assume that accompanying text or tabular material will be reduced the same amount as the charts . Half Page Width (about 411 ) Base (Zero ) Line , •.• . 02411 Scale-lines and Ticks . 010 000 Year-end Verticals . 017 0 100% Lines • .017 0 Curves: Simple Curve Chart . 03 5 3 Multiple Curve Charts: -Primary or t otal curve . 030 Secondary curve -solid-. 024 Secondary curve -dash -. 030 Secondary curve ·-dot 3 Surface Charts: Single curve . 024 Total or primary curve . 024 Secondary curve . • • .017 0 Coordinate curves •• . 017 0 Full Pat}e Width (about 811 ) Base (Zero) Line ••• . 030 Scale-lines and Ticks • .010 000 Year-end Verticals • • • • .024 100% Lines • • • • • .024 Curves: Simple Curve Chart . 043 4 Multiple Curve Charts: -Primary or t otal curve . 035 3 . se-condary curve -solid . 030 Secondary curve -dash -. 035 3 Secondary curve -dot 4 Surface Charts: Single curve . 030 Total or primary curve . 030 Secondary curve • . 024 Coordinate curves • . 024 COLUMN AND BAR CHARTS { Number o f Columns or Bars: Base Line . 035 3 l-3 { Outline . 028 Base Line . 030 4-6 Out line . 024 { Base Line . 028 7-9 { Outline . 021 1 Base Line . 024 10-12 Outline . 017 0 { Base Line . 021 113 or more Outline . 014 00 Scale-lines • . 010 000 100% Lines £1 Column or Bar Divisions , gj Cross-li ne Symbols §/ §) The sampl e line welgn-cs as r eproaucea apyrox imate the actual wei ghts speci f i ed . Small vari ati ons m~y have ari sen in reproducti on. Q/ Leroy pen numbers . Wri co pens are availabl e in same size but use a two-number system t hat shows pen s i ze : e . g ., pen No . 35 =.035". ~ Make these t he same wei ght as the out lines . Q/ Make these 2/3 t he weight of the outlines . ~ Make these t he same wei ght as the base line . CHART CONSTRUCTION SPECIFICATIONS FOR CURVE AND SURFACE CHARTS Chart no. Title Subtitle 100-0 ~~20-0 I _CHART 12 25" ·-_ _l _____ .OVERSEA COMMAN TRENGTH f r Scale lines •20" L _________ EXCLUDING EN ROUTE PERSONNEL_ _ .010" or 000 pen / ~H~USANO_t_ .25" Scale caption . .08" 80-00 -.to" high .0511 high Single curve .035"or 3 pen I " r --_j) L.....I......._...L...J......l.....L....L.....L..J.....L-..L.....&-.J......l.....L....L....J-.J.-'--L...J-.J......~...r::~_ -I .15 L ---_J s 0 M J s 0 M J I 35" / FISCAL YEAR 19Xl FISCAL YEAR 19X2 ___ .J' ~~ Main title 80-00 100-0 ~40 / -I Chart title r----_STATUS OF FINANCIAL PROPERTY ACCOUNTS~ •30" ~ ___STOCK ON HAND _l_RECEIPTS AND ISSUES MILLION .25" THOUSAND $4 ~------------~--------------. .-----~------~--------------.$800 CURVE A 0 0 19Xl 19X2 CIV I LIAN STRENGTH-FY l9X2_f_ Total curve NUMBER .25" Primary curve 1500 .030" Secondary curve .02411 CROSS -REFERENCES : Curves, lines -pages 98, 104 . 0 Months Shad i ng -pages 94-96. .2o"i /l1oo-o Letteri ng -pages 102-103. '--~ J A s 0 N D J F M A M 214-8 53 0 -66 -7 105 GRAPHIC PRESENTATION SPECIFICATIONS FOR COLUMN CHARTS Chart no. Title Subtitle / 80-00 -~ 100-0 CHAR;,~--/ NOTE: ·2~r:___ _ACTIVE ARMY. TRENGTH These illustrations have .20 L_ _ _______ _30 JUNE_ I_ been reduced 16 percent . Lettering designations and line .2511 MILLION_!_ weights refer to Leroy template and pen numbers. 1.5 .o8" CROSS-REFERENCES: Column spacing -page 90. 1.0 Shading -pages 94-96 . Lettering -pages 102-103. Line weights -page 104. Bose •25" .035 Scale caption ./~~ f 80-00 EXCL. 600 NOTE: Scale numerals/ Cross-line symbol s 100-0 and step-curves should be heavier than the column outlines (same • weight as base line is good rule of thumb). Connecting lines from columns to cross-line symbols should be same weight as scale lines. Main title Chart title Subtitle / 140 -I I20-0 /00-0 "'_COMBAT VEHICLE AND PROGRAMMED PROCUREMENT .2C(_ _L _ _ _ _ _ _ _ _ _ _ _ _ ____CALENDAR YEAR 19X1 Actual-Estimated arrows · 25"~--SELF-PROPELLED WEAPONS_!_ TANKS/ (Seepage93) .2511 NUMBER / NUMBER DELI.EAIES,PROG>AMMED t 1:: MED~~:"'~uvt"~l'~oor; ~~ ;r ::: :·:·: ····· ····· ::::: :::::. ····· ~~~~~ ::::: ~~~~~ ~~m~ 111111 I ::::: 50 ~~;~~ ~~~~~ ······ 200 II ~~~;;;~~;; :;::: 25 100 ~l~l~ ·•·•·· ::::: :::::: ···•·• ······ I :::::· ~~~~~ ~m~~ :::::. ::::: ;;m .... :::::: ~~~~~ 0 0 J F M A M J J A s 0 N D JFMAMJJASOND CHART CONSTRUCTION SPECIFICATIONS FOR BAR CHARTS Chart no. Title Scale caption Scale numbers 100-0 ~ /f20-0 ~ CHART 4 ______ -/-____ /-______ ; ,.,~ NOTE: FY 19Xl MILITARY APPROPRIATIONS_ .J:~5 These illustrations have 11 20 11 ., PERCENT __________::...--, been reduced 16 percent. Let DEPT. 0 20 40 60 80 11 100-=::::-= :::::.::.1~=:::J .05" tering designations and weights refer to Leroy line tem plate and pen numbers. ARMY Labels 80-00 CROSS-REFERENCES: Bar spacing -page 90 . 100% line and Shading -pages 94-96 . NAVY Bar outline Lettering -pages 102-103. .028" Line weights -page 104 . Bar outline .021" Subtitle Bose .035 svc 500. Stub caption 80-00 DELIVERIES Not e t hat bar s pac ing Stub items I REQUIREMENTS and line weights ( ex c ept /00-0 f or sca l e lines) var y wi t h the numb er of bar s · ( see pages 90 and 104). Subtitle 120-0 Scale lines F Zipatone .010"or 000pen PERCENT OF FUNDS OBLIGATED Bar outline 017 r-~~5~'f_-_~l;c~C-YEAR 19Xl AND EARLIER APPROPRIATIONs/· " Symbol •25" L_ .Q~TJ:G.9~'L _____ 0 25 5~ /?~ / •024" 100 1 TANK AUTOMOTIVE :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::f:::::::::::::::::::::::::::l AM M UN ITI 0 N ;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:::1:::;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;1 CLOTHING AND EQUIPAGE ELECTRONICS EQUIPMENT SUBSISTENCE CHAPTER 5 FORMAT Format is defined as the shape , size, style and general make-up of publications, reports, pages , tables or charts . These physical characteristics are subject to wide variation. When a series of publications is designed to serve the same general purpose, however, the variation should be kept to a minimum. Uniformity of format contributes greatly to the convenience, efficiency, and usefulness of the series and prevents distracting the reader with superficial details. REPORT FORMAT Elements of a Report The elements of a report are the physical parts that go into its make-up 1/. Since a report may be a single piece of paper, a bulky bound volume, or, more often, something in between it is not possible to establish a standard as to what elements should be included. The only element a report must have is a body . Other elements may be added before and after the body if they are required. The simplest form of a report is all body -more elaborate reports require additional elements. A complete, formal report could contain all of the following : l. Cover. 2 . Title or security page. 3. Distribution list. 4. Letter of transmittal, foreword, or preface . 5 . Contents page. 6 . Body. a . Introduction b. Summary c. Detailed review and analysis d . Reference material 7. Index. Because of the importance of the body of the report it is considered first in the paragraphs that follow. The Body of a Report Too often the presentation of a report consists of binding together a collection of basic charts and tabl es without explanation other than a transmitting paragraph . Reports of this type sometimes are misunderstood or misinterpreted and contribute to many erroneous decisions. Recurring reports, issued frequently for limited distribution, may require only a little explanatory information. This is particularly true when all individuals receiving the report are thoroughly familiar with its content and characteristics. l/ This discussion of report format is concerned with analytical statistical reports designed to assist top management in making decisions . These are "end-product" reports which generally contain a small portion of the available information selected for its significance . These reports usually are published . Guidance concerning report forms and directives for the collection of raw data is given in DA Pamphlet No . 335-2 and AR 335 -10. Observance of the following principles generally will assist in presenting the information in su ch a manner that the r eader can read it easily, grasp it quickly, and understand it accurately. First, an INTRODUCTION, with or without a heading, is generall y desirable at the beginning of the body of the report. It may contain information about: l. The nature of the subject. 2 . The procedure and method followed . 3. The sources of information. 4. The characteristics and limitations of the data. 5. The meaning of special terms. It is not essenti al that these items be included in an introduction that is a part of the body of a report. It is important that careful consideration be given to each of these items and that explanations be availabl e to the reader somewhere. Decisions as to the content of the expl anatory material and its presentation must be based on the nature of the material and the needs and t echnical subject-matter knowledge of expected users. If, for example, special terms are voluminous , they may be explained under a heading such as Definitions or Glossary. If the sources are many and varied they may be explained under a heading such as Explanatory Notes or in an Appendix. In many cases these special terms and sources can be explained most effectively in footnotes. Second , the body of the report should contain a SUMMARY of the main findings or results. Depending upon the size and scope of the report, this ma~r be a sentence , a paragraph, a page, or a dozen pages including summary charts and tables . Even when a report covers unrelated subj ects, it is often desirable to use a brief summary, giving the key facts about each subject in a digest (either narrative or tabular) at the beginning of the report. Greater detail may be shown in a summary which may precede each of the individual subjects . Third , a DETAILED REVIEW AND ANALYSIS or other material supporting the summar y should be arranged in a l ogical sequence . To do this, the text, charts, and tabl es should be presented in judicious combination using each to accomplish the purpose for which it is best suited. They should be placed in the order in which they are to be read. Tables should not be grouped arbitraril y at the back of the report for reference. Fourth, actual REFERENCE MATERIAL which is not essential to an understanding of the report may appear as an appendix. Other Parts of a Report The COVER has two basic purposes : to identify the report and to protect the material inside . Sometimes it serves additional purposes , such as a combination cover and contents page, or a combination cover and foreword . REPORT FORMAT Identity is provided by a ~itle, a date, and the name of the issuing agency, and sometimes an appropriate symbol, insignia or design. Protection is provided by a security classification at least as high as the highest classified material in the report, and by a paper stock of sufficient weight t o conceal and physically protect the contents over the period of expected use and retention of the report. The title should be brief yet accurately descriptive:--As in all titles, t he cover title sometimes can be shortened effectively by reducing some of the description t o a subtitle. Long subtitles, however, are almost as undesirable as long titles. A good rule of t humb is t o use one line for titles, and no more than t wo lines for subtitles. The nature of the data determines whether the dat.e is an "as of" date (e.g., 7 December 1941; 30 June 1945), or a period of time (e.g., June 1945; First Quarter, Fiscal Year 1945; Fiscal Year 1945; or 7 December 194~ -31 August1946). Sometimes, the time element of the material has no particular significance, and the date on the cover can be the date the report was issued or completed. Sometimes both the "as of" date or period and the date of publication are needed. At least one of these dates should be used. The name of the issuing agency is usuallythat of the higher organizational element to which the actual preparing agency belongs. For example, if the preparing agency is "Office, Director, Review and Analysis, Office, Comptroller of the Army," the issuing agency shown on the cover may be "Office, Comptroller of the Army." A TITLE PAGE which may also serve as the Security Page generally follows· the cover. This page is the first right-hand p~ge inside the cover. It includes the title of the report; its Reports Control $ymbol, if any; the security paragraph, if the report contains classified material; and the name of the preparing agency. A DISTRIBUTION LIST is ofter-, but not always, included in a report after the title page. In a bound volume it is reproduced on the reverse of the title page. It indicates the recipients of the report, and sometimes the number of copies distributed t o each of them. The words "DISTRIBUTION LIST" are centered above the listing. The list may be typed in either one or in two columns as described below for the contents page. This page may contain instructions on how to obtain additional copies or be put on regular distribution for future issues. The best place for such information, including the complete name and address of the office concerned, is below the distribution list, or in its absence, at the bottom of the contents or the title page. A LETTER OF TRANSMITTAL, FOREWORD, or PREFACE may be included, particularly in a publication superseding another one or in the first issue of a recurring report. It may be used to explain the purpose, scope , and application of 214-653 0 -66 -6 the publication. It is placed ahead of the contents page. Some reports have more than one of these three in the sequence shown at the beginning of this paragraph. The CONTENTS page of a report usually is placed on the next right-hand page . This page (or pages) shows the main organization of the report and indicates the number of the page on which each subject begins. It shows the location of major subdivisions, as well. When relatively few items are listed, the contents page may be presented in a single colunn centered left to right on the page . If a single column does not provide sufficient space to list all the subjects covered, the contents may be presented in two columns utilizing the full width of the type page, with three or four spaces between columns depending on whether the full width is an odd or even number of typing spaces. On the contents page the major divisions of the report are indicated by. headings typed in all capitals and aligned at the left margin of the typing width selected. The subordinate divisions and the subject items are typed in initial caps, and indented four spaces,* for each degree of further subordination. Leaders, consisting of periods typed in every other space and lined up vertically, are used to extend the captions to within two spaces of the widest page numbers. A minimum of one blank line is left between major divisions: extra blank lines may be inserted if needed to balance the material on the page, or t o equalize the length of columns when the two-column arrangement is used. Frequently the TYPING AREA required by the contents page fills less than a page . If this occurs, a one -column arrangement should be centered horizontally t o occupy not less than 70 percent of the width of the type page . Vertically, about one-third of the blank space should be left above the page heading (to the top of the typing area), and about two-thirds of the blank space left below the last typed line (to the bottom of the typing area). The same procedure should be used for the distribution list and the foreword, or the equivalent. The INDEX is an important part of many reports. It is an alphabetical, cross-referenced, subject listing at the end of the report with page references for each subject. The sole purpose of the Index is t o make al l of the material in the report easy to find. The Contents page usually serves this purpose adequately in short reports; but in long, involved reports covering many subjects, a comprehensive Index will more than pay for the trouble of preparing it. Physical Characteristics of a Report Report Size. Most Army Statistical reports are reproduced directly from typewritten copy in 811 final sizes of x 10. 511 or 9" x 13". The larger size is .convenient for use and filing, and it contains a much larger typing area than the 8" x 10. 5" page. * For one-column pages. Indent two spaces for two-column pages. FORMAT The larger typing area allows incr eased flexibility in the arrangement of material and reduces the number of pages to be processed. In a detailed review and analysis of a series of subjects, it facilitates getting a maximum of text, tabular, and graphic presentation for each subject on a single page or pair of pages, without excessive reduction of material in reproduction . This makes it simpler t o find material in the finished report. The larger size also helps to solve one of the most perplexing problems encountered in preparing many statistical reports: how to get a maximum amount of tabular data in a mininrum of space. The Adjutant General reproduces DA publications in size 8" x 10.5" unless the proponent agency justifies the use of another size because of content and usage. DA administrative, supply, and equipment publications (most of which are printed rather than reproduced from typewritten copy) are published in the 8" x 10.5" size. This size can be handled more rapidly, especially for mass distribution to many geographic areas. It will also permit the use of automated counting and packaging equipment in AG publication centers. Moreover, the publications can be printed on economical, highly automated presses. The 811 x 10.5" size is good t o use i f all or most of the tables in a report fit readily into its smaller typing area. Otherwise, it is easier to prepare the report in the 9" x 13" page size. The use of larger page sizes than these two is not recommended. Larger sizes are difficult for the reader to use, and present filing, handling, and mailing problems. Use of larger sizes can usually be avoided by adequate planning for a standard page size. The standard sizes generally result in more professional-appearing and readable reports which have better space utilization throughout. Two-side and One-side Repr oduction, Reports reproduced on one side of the paper can be prepared and reproduced faster than those reproduced on two sides of the paper. This practice is wasteful of materials, however, and often makes a long report forbidding in appearance and inconvenient to use because of its bulk. Two-side reproduction, therefore, is preferred generally for all reports of more than a few pages. A good rule to follow is to use twoside reproduction when a report consists of 10 sheets or more in addition to the preliminary pages and the back cover (if any). It is highly desirable to present closely related material in such a way that reference may be made t o and from related parts without turning pages back and forth, especially in reports showing tabular and graphic information as well as text. Reproduction on two sides of the paper gives twice the area for reference without turning pages and is usually justified on this consideration alone. PAGE FORMAT Elements of a Page The physical elements of a typical page in a formal Army statistical report are the running head , page number, classification label, type page (the "typing area" in which text, tables and charts or other descriptive or identifying material are placed), and the margins. Theillustrations on pages lll and 112 show the usual pl acement of all elements in the standard page ·format, and also give the standard dimensions common to each page of a 9" x 13" and an 8" x 10 . 5" report when reproduced. RUNNING HEADS usually appear at the t op of each page of a report. They are prominent headl i nes which identify the general subject matter on the page, There is no standard arrangement that is best for all situations , but for any indi vidual report a suitable method ordinarily is suggested by the composition of the material in i t . In.thi s manual the chapter title appears on the left hand pages and the major sections of the chapter on the right hand pages. Stymie Medium type -the style used in this manual because of its similarity to elite typewriting -i s recommended for running heads . If i t is not readily available, Bodoni Bold type or any sinilar lettering is an accept able substitute. Proof sheets of one of these compositions may be cut and pasted at the top center of the typing layout sheet. (See page 98.) If commerci ally prepared type is not available, the runni ng head may be lettered directly on the sheet wi th lettering pen and template or some other lettering device. Horizontal rulings are frequently drawn above and below the running heads (as in the illustrations on pages lll and 112) . The top ruling, however, may be omitted, The f i rst pages of the distribution lists, forewords or equivalents, and contents pages do not carry running heads. Instead, the applicable identification is centered above the first typed l i ne in the same style and size of letters as the running heads on other pages of the report, but without the horizontal rulings (or ruling) which customarily set off a running head from the rest of the page . SUcceeding pages of these report elements cont ain running heads with the words "DISTRIBUTION LIST, 11 "FOREWORD, 11 "PREFACE~ or "CONTENTS, 11 as appropriate. Similarly, no running head is used on the :FIRST NUMBERED PAGE of a report. The title of t he report is placed on the page in the same t ype style and size as the running heads. If the report begins with an introduction or a S2ffiffiary, one of these terms may be substit uted for the report title . I: there is no introduction or summarv, the title of the first chapter or maj or subdivision of the report is sometimes used . CLASSIFICATION LABELS, identifying the material by "Top Secret," "Secret," "Confidential" or other security classification, are attached to each typing layout sheet that contains any classified material, after the final typing on t 1.0" RUNNING HEAD LINES C!l .4" r ---------, FORMAT I ~~---------------------------------------4-------------- L--____ _ ...J CLASSIFICATION (LEFT HAND PAGE),--- (RUNNING HEAD: STYMIE MEDIUM TYPE OR BODONI BOLD TYPE),--- 0. 511 I I I I I /10.2" TYPE PAGE I 1311 ~ \ TRIM SIZE I I I I 9 11 -----r--------1--------.---1.---1-------+---------,-- 1 I I 1 I I I I I This page presents the format spec ifications for a l eft hand 911 page of a x 13 11 report after reproduction, For a right hand page, the positi on of both the page number and the classification labels would be reversed. I I I I I L---,0. 4 " @_____________i __ PAGE NUMBER _ ..J I I IL ___ _ I I _j 0 . 4':---------, --...:.L L ------_: (LEFT HAND PAGE) (a~~smB~I8~) lll -~ 0 . 5" RUNNING HEAD LINES .,: 0.4" FORMAT r--------, ------------------~----~-------------------------------4~ t (jRUNNING HEAD : STYMIE MED!UM TYPE L ________j 0 . 5" OR BODONI BOLD TYPE) CLASSIFICATION ~---- I _illiGHT HAND PAGE) . --.·.1 ~, I 1 I 8. 111 ~ TYPE PAGE \ ...---1.011 ~M---------+-----6. 5"!-+----------+-----------l.io0.5" / 10.5'' TRIM SIZE I . 011 ~----_J_______ ~-------L-~----------+---------L-- 8 This page pr esent s t he for mat speci fications for a right hand 811 10. 511 page of an x r eport after r epr oduction. For a l eft hand page , the posit ion of both t he page number and the cl assification labels would be reversed. I ____ j I_ _ _j L---- r-------, 0 . 411 I I L ______ .J 00 ----PAGE-NUMBER CLASSIFICATION ' RIGHT HAND PAGE)(RIGHT HAND PAGE) 112 PAGE FORMAT the white paper is completed. Each page carries the highest security cl assification of any mate r ial on the page. Pages with no material that i s classified do not require an;r classification labels . Two labels, prepared in advance from proofs of condensed Gothic type or the equivalent and incl osed in ruled boxes, are a~tached to each cl assified page. One is centered over the lower l ine of the running head and aligned with the outside margin. The other is placed below the type page and aligned with the inside margin. The bottom ruling of the box should be about four-tenths of an inch below the type page after reproduction. See the illustrations on pages lll and 112 . PAGE NUMBERS in the same type face as the running heads are placed at the outside l ower corner of each page directly opposite the classification l abel and the same distance below the type page . No number appears on the ti~le or security page, but Roman numerals (beg~nning with II) general ly appear on the other preliminary pages, such as the distribution list, foreword or equivalent, contents page , etc . Page "one" in Arabic numeral s is usually the first page of the body of a permanently bound report. It is always a right-hand page, as are all other odd numbered pages . All even numbered pages are always on left-hand pages . Page numbers should be assigned to blank pages, but these numbers should not be reproduced . This insures that odd numbers will continue to be on the right-hand pages and even numbers on the left. Final copy is normally made on oversized originals which are reduced in reproduction. The TYPE SPECIFICATIONS below show for various reductions the type size (in points) to use for running heads and page numbers : Running Heads Page Numbers Percent Reduction Stymie Bodoni Stymie Bodoniof Page Medium Bold Medium Bol d With Double Ruled Running Heads 10 . 18 18 12 14 16 . . 18 18 12 14 20 18 13 12 14 25 . 24 24 18 18 30 24 24 18 18 36 . 24 24 18 18 With Single Ruled Eunning Heads 10 . 12 14 l:? 14 16 . . . 12 14 12 14 20 12 14 12 14 25 . . 18 18 18 18 30 . 18 18 18 18 36 18 18 18 18 The fol lowing line weights (in i nches) are appl icabl e to the running head r ulings and the rulings of the box inclosing the classification label. Percent ReducRunning Head Cl assification tion of Page ~ulings Box Rulings 10 . 017 . 010 16 .017 .010 20 • . 017 . 010 25 , 020 .012 30 . 020 . 012 36 . 024 . 014 NOTE: These type sizes and line weights apply to both 8" x 10 . 5'' and 9" x 13" reports . Both TEXT and TABLES are generally typed on 12-pitch typewriters (that is, 12 characters to the inch) equipped with standard Elite letters (the kind used fc" these pages) . Text for published reports is usually typed six lines per inch when single spaced, except as noted under the TEXT PRESENTATION chapter of this manual . One-col umn and Two-column Pages Material placed in the type page area of a page may be presented in either a one -column or two-column arrangement . The two-column format , used in this manual, is generally preferred because the text is easier to read, and more text can usually be placed on ~he page . The two col~nn arrangement permits greater flexibility in alternate use of tables and charts that are a half-page-wide and those a full-page-wide . A long line of type is harder to read than a short one . The best length for easy reading is a line 39 to 52 characters long. In typewritten reports, however, a column of typing 39 characters wide is so narrow that the typist has diffic~ty in making the lines come out even at the right. A column should be at least 45 characters wide to provide sufficient flexibility to permit the typist to avoid or to adjust short lines. Two-column text is preferred if a page is sufficientl y wide for a minimum of 93 typewriter spaces (each column 45 characters wide , with at least three spaces between columns) . Satisfactory text also may be typed in columns that are 43 characters wide with four spaces between the columns . For pages containing a typing area width of less than 93 spaces, the one-column text usually should be used . Layout and Organization The layout process between the submission of manuscript and the final copy is considerabl y more difficult for tables and charts than it is for the text. The requirements of presentation and of content often conflict, making necessary numerous compromises that must be worked out be fore final copy is prepared. In designing a page on which a table or a chart is presented with text, the table and chart should be planned first since ~hey are much less flexible than text. FORMAT The general requirements of the whole page must be kept in mind , however , while designing the table or chart in order to attain the proper emphasis . Many statistical reports have a "sameness" from page to page which may make an inherently i~teresting report dull and uninteresting . It takes two forms : l . Lack of variety in the typ6S of charts used. 2 . Close similarity in the arrangement of the text , charts, and tables . The repeated use of one or two standard types of charts in a report seldom is desirable . Many types from which to choose are shown in the chapter on Graphic Presentation. If a report contains many charts, several types are generally needed to make the best comparisons . The pages may be diversified by making some of the tables and charts a full-page -wide and some of them a half-page -wide . Variations also may be made in the amount of space allotted to each subject, the amount of text and types, number , and sizes of charts and tables . The amount of time available for layout and organization of reports necessaril y affects the quality of the reports . If it.is essential to meet an extremely short deadl ine, it is obvious that something must be sacrificed. Sometimes time can be saved by preparing and reproducing semi-finished charts and tables (complete without data and footnotes) before the raw data are received. This greatly reduces the period from receipt of the data to the compl etion of the final copy f or reproduction. A logical arrangement should be followed in presenting statistics if they are to be clear and easily understood. Correct us e of the tool s -charts , tables , and text -wil l hel p to insure an orderly sequence of data and the completion of each subject before the introduction of another . Good planning almost always results in putting near each other the things that belong together . Otherwise, the report becomes disjointed and the reader is apt to lose interest. For example , it is improper to put a chart on page nine when it belongs with the material on page three , simply because there was more extra space on page nine. When a page contains closely rel ated text , charts , and tables , it is general ly preferabl e to connect the charts and tables by placing some of the text between them. This gives. the presentation a balanced look and leads the reader nat,rrally and smoothly from one chart or table to the next . In assigning space on a page, allowances must be made for standard separations between text, tables, and charts . The established separation is two typewriter lines . When space is "tight", this may be cut to one ; or , when space is more "loose", increased to three lines . The standard separation for a text table is one blank line from the text which precedes and follows it. When one-column text is used in a report, it may be continued beside half-page-wide charts and tables . The "runaround" text should be separated from the charts and tables by three or four blank typewriter spaces , depending upon whether the layout sheet has an odd or even total number of spaces . This problem is not often met in a two-column presentation. If a table comes at the bottom of a page of text, the tabular and 'text footnotes should be clearly differentiated. The method of doing this varies , depending on whether the table is halfpage-wide or full-page-wide and whether the text is one-column or two-column. The following arrangements show the successive order of text, table and footnotes which usually works best for each particular situation: l . Half-page-wide table and 2-column text Text Two blank lines Table Tabular footnotes, no wider than table Blank line Separator ruling (10 spaces wide) Text footnotes If the other text column also has footnotes, put them at the bottom of that column following the standard blank l ine and 10-space separator ruling. 2. Full-page-wide table and 2-column text - Text Blank line Separator ruling (or rulings -one un der each text column -if both text columns have footnotes) 10 spaces wide Text footnotes (under each text column if necessary) Two blank lines Table Tabular footnotes 3. Half-page-wide table and 1-column text with runaround text beside the table - Text Two blank lines Table Tabular footnotes , no wider than table One blank line Separator ruling, 20 spaces wide Text footnotes , full-page-wide 4. Full-page-wide table and l-column text - Text Two blank lines Table Tabular footnotes One blank line Separator ruling, 20 spaces wide Text footnotes Narrow and Broadside Pages . Whenever possible, material should be arranged so that it will read across the narrow or short length of the page . PAGE FORMAT This upright arrangement should be used for charts and tables . Material sho~d be presented to r ead broadside or the long way of the page only in extreme and rare cases . Some effective method of avoiding this arrangement can almost always be found . Particular effort should be made to avoi d the use of broadside material iL an upright report where the broadside material is at right angles to the text. This makes it unnecessaril y difficult to read . Fold-in Pages . A folded page bound with a report is extremely undesirable . It is onl y rarely that this expedient is the best compromise in presenting material in reports. Some material that cannot be presented effectively on one page may be redesigned as -a "double-page spread," arranged on two facing pages . Other alternative arrangements are discussed in some detail in the chapter TABULAR PRESENTATION. (See pages 39 and 40 . ) BINDING METHODS When a report has more than one page , the pages should be fastened togetLer . Fastening and binding methods range from a single staple in the upper left corner to sewn and glued bindings used in bookmaking . For mo3t typewritten statistical material , binding ~ethods between these two extremes are better. Binding methods can be divided into three categories : l. Stitching or sewing . 2 . Mechanical binding. 3. Loose-leaf binding. Stitching T~is group includes both stitching and sewing with binding thread, corner stapling, side wire stitching, and saddle (wire) stitching. Sewing is generally used only when the presentation is to be a bound book , and so is not described here . Corner Stapling -this is the quickest and easiest binding method . It is good for a few sheets of an informal presentaticn. It is used frequently for Army statistical reports. Pages stapled in this way tear loose easily, so this method should not be used for papers that will be handled a lot. Pages printed on two sides may be printed "tumble" style (the back upside-down), but this can be a nuisance . Side Stitching -in this binding method , wire stitches or s~aples are driven through the sheets in the binding margin. This method is quick and easy, but it has two disadvantages : (l) Side bound publications are hard to open and will not stay open unless the pages are folded back and heavily creased . (When there are a lot of pages even this does not work very well) . (2) Stitching in the side margin uses about a half inch of the margin, and the thicker the report the bigger the margin must be . The copy area on the page , therefore , must be narrower, which re-sults in less copy per page . Saddl e Stitching -in this method , folded sheets ·are opened in the middle and placed over the '' saddle" of the stapling machine . Two or more staples are then driven through at the back fold and clinched on the inside . Saddle stitching is not good for thick volumes . Also , it requires that printing be done on sheets that are b..,ro pages wide (which means that the first page is pri nted on one half, and the last page on the other half of the same sheet; and so forth). Saddle stitching has one big advantage : the pages lie relatively flat when the publication is opened . B~cause of the location of the stitching, less binding margin is needed and more of the page width ;::an be used for "copy . " Mechanical Bindings Mechanical binding methods were developed to combine the advantages of side stitching and saddle stitching. The best known are the spiral binding (a coiled wire threaded through a large number of small holes) and the plastic binding (a cylindrical plastic "comb" with the teeth inserted through a seri es of slots) . These and other patented bindings come in several sizes . They generally require special equipment not available in many offices . The main advantages of these types of binding are that they permit the sheet to lie flat, and need only a narrow binding margin . Both stitching and mechanical bindings are .relatively permanent: that is, bound sheets cannot be removed or new sheets added without unbinding. Loose-Leaf Binding These binding methods include brass fasteners and similar binding devices as well as ring binders . Sheets for all loose-leaf binders must be punched with two or more holes . Fasteners -these are strips of soft metal that are put through the punched holes and then bent to hold the sheets together. Some fasteners are loose and some are attached to covers . The simple brass fasteners (twopivoted prongs joined at the base in a f l ange) are the cheapest. ------------SIDE STITCHI G -------------SADDLE STITCHING -OPEN BACK COVER SINGLE FOLD COVER DOUBLE FOLD COVER SINGLE FOLD COVER 2 PIECES l PIECE l PIECE l PIECE FORMAT BINDING METHODS -WIHEN TO USE EACH A few sheets -informal presenta CORNER tion -when time is short. Semi STAPLING permanent -best with thin paper lowest cost. A few sheets , to a report half an inch thick (special equipment willSIDE stitch books up to 1 1/2 inches STITCHING thick) -semi-permanent a l ow cost method . Formal presentations such as annual reports, printed booklets, etc . - SADDLE quick method but requires careful STITCHING planning of pages -opens flat a low cost method . When sheets are to be removed (or added) later -a few sheets to half FASTENERS an inch thick or more medium cost. When sheets are to be removed (or added) and must open flat a few RING sheets to an inch or more (depend - BINDERS ing on diameter of rings) -best loose-leaf type -not recommended for many copies highest cost. MECHANICAL BINDING When sheets will not be removed (or (Spiral, added) but must open flat -Plastic, high cost . etc.) Fasteners are available in several lengths : short, for a few sheets; and longer, for .many sheets . The accompanying illustrations show several methods of fastening sheets in -and to cover stock with fa$teners of this kind . The strip fasteners (see illustration) also come in different lengths, separate or attached to binders . They are used in the same way as the brass fasteners , but they are a little fa~ter and easier to handle . These are best used as a semi -permanent binding. Rigid metal posts or screws -such as those used in post binders are widely used for accounting records, but they are too clumsy for most statistical presentations . They have the same disadvantages as fasteners , and also the disadvantages of ring binders. Fasteners are a medium-cost binding method . They should be used for a few copies only. Their principal advantage is the loose-leaf feature sheets can be add ed or removed easily. MECHANICAL BINDINGS SPIRAL PLASTIC FASTENERS BRASS STRIP C=R===Head® =ound=~ ~ Flat Head(D ~ COVERS WITH FASTENERS SINGLE FOLD COVER SINGLE FOLD COVER Full Fastened Half Fastened-Back OPEN BACK COVER DOUBLE FLANGE COVER ---~- - ;; ----- • Flanged I The principal disadvantage is that they are unwieldy: the pages do not open flat, and the more pages there are, the more trouble this is. Ring Binders these are available in any size from vest pocket to large display size , with various numbers of rings , · and in many kinds of covers , from cardboard to leather. Two or three rings is the standard for letter size, three for larger pages . When more than three are used, a punching probl em arises . Two (711 hole punches center to center) , and threehole punches (4 1/411 center to center) , are the standard, and the only ones that are widely and cheaply available . The diameters of the rings generally range from 1/211 to 1 1/211 in letter size binders . Ring binders are heavy , bulky, expensi ve , and hard to label . They also require considerable file space and should be used only when interchange of pages and flat opening are of prime importance . CHAPTER 6 PRODUCTION TECHNIQUES The preparation of a report is a difficult job at best : it can be a nightmare unless the contents are planned carefully, developed systematically, and reviewed methodically. The application of sound production techniques will not guarantee a good report, but failure t o apply t hem will result almost invariably in a poor report with undetected errors, laGk of balance, incomplete coverage, and unsatisfact ory presentation of the material. ORGANIZING THE MATERIAL A report should be carefully planned before the work of presenting it is started. The first step in planning the report should be t o define both i ts purpose and scope, so as t o provide a criterion for judging the pertinence of all subsequent work. As soon as the nature and extent of the dat a and their likely implications are apparent, the organization of the report should be planned . As text, tables, and charts are deve l oped , this preliminary organization should be revised as necessary, the overall structure being kept in mind while the details are devel oped . When t his met hod is f ollowed, the parts should grow into a unified whole, r ather than being forced into a pattern after they have been completed independently of each other . Preparing an Outline A good way t o start is t o prepare a tentative list of contents and assign space for each item in accordance with its importance . This will force summarization of more voluminous data on the relatively unimportant subjects and the expansion of and perhaps more thor ough analysis of t he more important subjects . In preparing the list of contents , try t o gr oup related items t ogether and arrange the groups, as well as the mat eri al within each gr oup , in a logical or der . This pr ocess should be continued t o the or gani zation Df each of the items, so that a working outline results , which needs only t o be expanded in devel Dping the complete report. Making the Report Dummy Af'ter the content of the report i s fairly well crystallized, but befor e any of the parts are prepared in final form, the physical fea tures of the report should be considered. At first, a tentative report dummy should be pre pared, showing as many pages, including blank pages as the t otal number planned for thereport. The material t o be presented on each page should then be indicated in pencil. This helps to determine the size and shape of the t ables and charts which must be pl aced on the same page with text or with each other and , particularly, aids in arranging related material in an adj acent position or on facing pages . Several adj ustments of the material t o be presented may be necessary t o make it conform t o the relatively inflexible conditions imposed by page and r eport f ormat . These limitations should constantly be kept in mind in preparing any final pages . Production Schedule With the dummy as a blueprint, the production of the report should be planned step by step and checked periodically against the schedule. Each page and each insert (chart or table) should be listed separately. The steps to be planned and checked separately will pr obably include those listed bel ow . Often two of these steps can be combined f or control purposes. 1. Tables designed. 2. Text written. .3 . Charts sketched. 4. Trial typing made . 5. Final typing made . 6 . Tables ruled and text underlined. 7 . Charts drawn . 8. Pages assembled, numbered, and classification labels affixed . 9. Final check and review made . PREPARING COPY In developing a report for publication a distinct function must be performed after the manuscript is complet ed but befor e final copy is prepared . This function may be thougpt of as page make-up or l ayout, or page design. It is not the function of the draftsman or the typist and should not be left t o either, but should be handl ed by the person preparing the report or by a specialist in this field. Typing Layout Sheets The procedures for designing pages and preparing typed and drafted material are greatly facilitated by the use of typing l ayout sheets. These are sheets of paper ruled t o correspond to typewrit er spacing and drawn in pr oportion to the st andard format dimensions of the report. TRIAL TYPlliG of text and tabular material may be done directly on typing layout sheets f r om manuscript, without any preliminary typing on other paper. Decisions as t o final page arrangement are more easily reached because the limitations of the f ormat are kept constantl y before t he person preparing the page . Furthermore , instructions t o the typist f or changes or adjust ments are mor e easily given and under st ood when specifi ed on the l ayout sheet than when sho'tffi on ordinary paper . The principal purposes of such trial typing, other than t o provide as clean copy as possible for final typing, are : (1) t o determine the space requirements for the text and t ables ; (2) t o check and r evise, if necessary, the tent ative page arrangement; and (3 ) to serve as a basis for justifying the lines. CHART SKEI'CHES may also be made on typing layout sheets (see "Using Typing Layout Sheets for Charts," bel ow) . PRODUCTION TECHNIQUES Specifications of standard Typing Layout Sheets. Standard blank forms are used in Headquarters, Department of the Army for reductions of 16, 30, and 36 percent f or 9" x 13 " reports. These sheets after reproduction have about the same format specifications as those shown on page lll. The sheets can also be used for other reductions for 9" x 13" reports and t o prepare pages for 8" x 10.5" reports which will reduce to about the same format specifications shown on page 112 (see "Using Typing Layout Sheet s for 8" x 10 . 5" Reports and for Other Reductions" , below) . An additional layout sheet for 8" x 10.5" pages is available . It is marked for reductions of 10, 15, and 20 percent. The dimensions of the 9" x 13" standard layout sheet for a 16 percent r eduction are specified in the illustration on the facing page. Note that: 1 . The spacing of the vertical rulings is the same as the spacing of letters typed on an elite typewriter (12 per inch) and the spacing of the horizontal rulings is the same as the spacing of single-spaced lines on any standard typewriter (6 per inch). Thus, each small rectangle defines the space in which a character may be typed. 2 . The ruled area corr esponds to the type page . The running head lines are indicated at the t op of the page , and the l ocation of the page number (in the l ower outside corner) i s also shown . Typing layout sheets may be used for either one-column (full page wide) or t wo-column text, charts, and tables (half page wide) . The full width of the type page is used for one-column presentations. For t wo-column presentations , the type page i s divided into t wo columns of equal width separated by three or four blank typewriter spaces (three i f the type page has an odd number of spaces; four if it has an even number). The accompanying illustration contains two columns of 48 spaces each, separated by four blank spaces. The sheets are printed on an inexpensive green or yellow paper for trial layouts and on a better grade of white paper for final (camerar eady) typing. The lines are printed in a light blue ink which do not reproduce when the mat erial on them is phot ographed with most standard film. Using Typing Layout Sheets for 8" x 10 . 5" Reports and for Other Reductions . Each of the three standard typing layout sheets for 9" x 13" reports can be used t o prepare pages of the s~ reduction for 8'1 x 10 . 5" reports by applying the type page (typing area) dimensions in t ypewriter spaces and lines indicated in the second and third columns of the table on this page (forreductions of 16, 30, or 36 percent as appr opriate). The page numbers are usually placed at the outside margin three or four typewriter lines below the last line of the type page , as revised -three in page reductions of 16 percent and four in page reductions of 30 and 36 percent. The running head rulings in black ink on the final copy should be as wide as, but no wider than the revised type page . When the typing layout sheet on the next page is used for an 8" x 10.5" report it would utilize 93 spaces and 61 lines. The type page would end on line 58 and the page number would appear on line 61 . In both 8" x 10 . 5" and 9" x 13" r eports use the 16 percent reduction typing layout sheet for page reductions of 10 and 20 percent and the 30 percent reduction typing layout sheet for page reductions of 25 percent. The type page specif ications for those reductions shown in the table on page 113 are applicable and the page numbers are uniformly placed on the third line below the revised t ype page . The standard dimensions of the 16 percent reduction typing layout sheets are not suitable for a 20 percent reduction t o the 9" x 13 " size . This r eduction r equires using five additional spaces and four additional lines with the type page ending on line 77 and the page number appearing on line 80 . The running head lines must then be extended t o the full width of the type page . Adjusting ·the Length of the Type Page for Deletion of Running Heads . Running heads are sometimes omitted, especi ally in short r eports. In such cases, the type page should be lengthened by utilizing the space normally occupied by the running head . The number of typewriter lines by which its length can be increased depends upon whether or not classification labels are required. The foll owing table shows the length in lines of the type page with and without classification labels but without the running head : Report Size : Report Size : 811 911 10 . 511 fY 13" fY PerGent X X Reduction of Page With With Without Without Labels Labels Labels Labels )2/ )2/ 10 . . . . 56 60 70 74 16 . . . . 60 64 75 79 20 . . . . 63 67 79 83 25 . . . . 68 72 85 89 30 . . . . 72 76 91 95 36 . . . . 80 84 100 104 See pages 111 and 112 for length of type page when running heads are used . Align one classification label with the outside margin of the type page and place the top ruling of the box inclosing the label four lines above the type page . Place the other label in its cust omary position at the inside l ower margin directly opposite the page number and an equal distance below the type page . (See page 111 . ) i 1. 2" 1 411 EUTE TYPING LAYOUT SHEET 0~ t ,----------------------, 1----------------~------l 1 I I I I I I I I I I I I I I I 11 / 12 . 2 I I I I I I I TYPE PAGE I I I I I . I I I I I ..--1. 2" 8.3" 1. 2"-+- I I I I I I I I I I 15 . 5"~ I I I PAGE SIZE I I I I I I \ I I I I I This model typing layout sheet is de I signed for a r eduction of 16 per cent . I A complete working copy would have a grid of nonphotographic light blue l i nes I with each cel l j ust l ar ge enough for a single elite typewriter l etter or figure . I The spacing of these lines is indica I ted by the length of the da shes surr ound ing each of the two columns on the page . I The specifications given apply to a I right hand page before reducti on . (See page 112 for format specifi cations after I r epr oduction. ) I I I I I I I I I I I I I I I L_____ _ ________________ I I________ _ r-------., I I I I 0. 5" I I L--_____j l _______ CLASSIFICATION PAGE NUMBER 119 -"- PRODUCTION TECHNIQUES Using Typing Layout Sheets for Charts. Although primarily intended for the typing of text and tables, typing layout sheets are also con venient to prepare chart sketches . If the chart as finally drawn is to be given the same reduction as the other material appearing on the page with it, the rulings on the layout sheet make it easy to determine the number of typewriter lines and spaces needed for the chart and the number remalnlng on the type page for other information after the chart has been designed . The final chart is drawn to the same overall size as the sketch, either directly on the layout sheet containing final copy or on a separate sheet which is then affixed in its proper place on the lay out sheet. A complex chart which needs more room for effective drafting of necessary detail may be prepared for a larger reduction than the other material on the same page. It is wise in such cases to make the chart sketch and the final drawing equal in size . The space requirements for the chart are computed by the method described below for oversize tables. Typing layout sheets may also be used in the design of charts for transparencies, slides, and posters as well as charts for printed reports . Chart sketches may be reproduced and distributed for comment and coordination with other offices before final decisions as to content, lettering, spacing, coloring, etc. This is sometimes referred to as the "precharting" technique. Oversize Tables A larger reduction for a table which will occupy a full page when reproduced will require more than the normal number of typewriter spaces and lines. The entire page is then given a larger reduction -as determined by the number of additional spaces and lines needed to type the table -than the other pages in the report. For example , if the normal reduction of the page s in an 8" x 10.5" report is 16 percent, an oversize table that fits into lll spaces and 69 line s would take a r eduction of 30 percent. When an oversize table takes up only a part of a larger type page, the excess space may be used for text or charts . Since all of the material then receives the same reduction, the text on the reproduced page will not be as legible as that on other pages reproduced at a smaller reduction. This is a serious defect in page reductions of more than 20 percent and can be avoided in the photo offset process or reproduction . The photo offset process allows oversized tables to be prepared separately for larger reductions than the material on the rest of the page . Separate negatives made at the two different reductions are stripped together and the image is transferred to a zinc plate for the final reproduction on the finished page. A table prepared in this manner must be carefully planned to insure that it will be red~ced in proper proportion to fit the available space . For example , a full-page-wide table 100 typewriter spaces (8 .3") wide by 18 typewriter lines (3") long for a 9" x 13" report will be 7 inches wide by 2. 5 inches long after being reduced 16 percent. The original dimensions of a table planned for the same space after a reduction of 30 percent would be . lO inches (7 ; 0 . 70) by 3 . 6 inches (2.5 ~ 0.70) or 120 spaces (l0xl2) by 22 lines (3 . 6 x 6) . In either case, if the other information on the page with the table is reduced 16 percent, it should be typed in an area 100 spaces (8.3") wide by 55 lines (9 .2") long -including 2 blank lines to separate this material from the table. The type page after reproduction will then be l0. 2 . inches long (2 . 5 inches for the table plus 7 . 7 lnches for the other material) . Prepared Materials Many expedients may be developed that are helpful in preparing final copy ·for reproduction. Some of these , such as the reproduction in quantity of repeating elements, depend on the fore thought and ingenuity of the personnel producing the report copy. Others , like the use of pre pared shading sheets for applying tones and patterns to charts, involve prepared materials that are available commercially. It is important to take advantage of all such expedients in order that reports may be produced quickly and economically and that their appearance may be of a quality commensurate with their content. MASTER FORMS, containing all repeating ele ments (such as running heads, titles, column headings , and rulings), may be prepared and re produced in appropriate quantities when a number of tables or charts of a similar nature are pre sented in a uniform manner . This procedure pro vides partly finished originals which are then filled in to completion thus saving a great deal of time and assuring uniform presentation. Most master forms are made the same size (with the same format specifications) as the typing layout sheets used in the report. If necessary, they may be made larger to accommo date the material. If typewritten material is entered on the forms, however, the extra reduc tion is likely to result in copies that are dif ficult to read. For example, master forms for oversize tables should not be designed for a larger reduction than 36 percent. Master forms may be designed for a series of oversize tables or charts which occupy less than a full page when reproduced at a larger reduction than the other material on the same page . The size of the forms is determined by the space occupied by the tables or charts after reproduction and the percent of reduction. EXAMPLES : The original charting area dimensions of a master form for a series of charts for a 9" x 13" report each chart to be 7 inches wide (full-page width) by 4.5 inches long (less than full-page length) , are as follows before reductions of 25 and 30 percent: l . 25 percent. 9 . 3 inches wide (7 • 0 . 75) by 6 in~hes long (4. 5 ; 0.75). 2 . 30 percent. 10 inches wide (7 -0 .70) by 6 . 4 inches long (4. 5 ; 0 .70) . PREPARING COPY When ordering master forms , the reproduction office should be cautioned that high quality work is necessary in order that copies may be suitable for further reproduction. PROCESSING COPY Report processing is divided into two basic steps . In the first, all the component parts of the report are typed or sketched; then proofed, edited , and dummied to form compl~te pages . In the second step , these pages are translated into final "camera ready" copy by retyping the text and tabular material (preferably ~n an electric typewri ter equipped with a carbon paper ribbon); by preparing the charts as finished inked drawings ; by assembling and marking all copy for reproduction; and by a final check for accuracy and completeness of the material. Procedural Guides to Good Copy The preparation of the page dummy which completes the first step , puts the report in semi -finished form . Using as a guide the report dummy , the text and tables are cut and pasted in the places assigned to them on each page , and space for each chart and each oversize table is marked off. Last-minute corrections and changes are often necessary at thi s time . They may be minor, such as changing a word or two ; or extensive, such as major changes in text, tab~es, or charts to remedy faulty layout, spacing, or other aspects of design. Major alterations may necessitate changes in arrangement of material on other pages of the report . One source of trouble stems from the ease with which a chart can be adjusted to a given space . The result is a tendency to prepare text and tables first and to squeeze the charts t o fit the remaining space , giving them less emphasis than they should have . Decisions concerning allocation of space to text, tables , and charts affect both arrangement and relative emphasis , and should be made when the report dummy is prepared (subject to review and revision as the production cf the report progresses) . Before final copy preparation is started, all material should be marked to indicate exactly how it is to appear in final form, so that typists and draftsmen can convert it verbatim into finished "reproduction originals . " This means that all information necessary for converting the chart sketches into completed inked drawings and retyping the text and tables should be provided, as indicated below . l . For chart construction the following are essential: accurate sketches and separate data sheets for each chart. The tabular material that is submitted for inclusion in the repor t frequently requires further processing for plotting purposes . For columns or bars that are segmented , for example , the values of the segments should be cumulated from the base line, so that the draftsman can plot each segment from that line rather than from the precedingsegment. Key number s should be ent ered to indicate the page on which each chart is to appear. 2. The f i nal typing of text and tables should be keyed to the page dummy which shows the page sequence and exac~ position ofeach part of the material. 3 . Corrections and changes which can be indicated clearly and easily followed, such as minor revisions in text and tables and simple adjustments in spacing, may be shown in pencil on the preliminary copy . Major revisions usually involve retyping some of the text or tables or redrawing some of the chart sketches prior to preparation of final copy . ASSEMBLING COPY The report dummy previousl y discussed, revised, if necessary, to reflect the final decisions as to paging, allocation of space , and arrange~ent on the page , serves as a basis for assembl ing copy for r eproducti on . Some one person should be responsible for thi s impor tant procedure . At this stage of repor t production, all the material i s seen together in f inished for m for the first time , and it is at this point that unanticipated problems of continuity, page arrangement, and other physical detai ls may come to light. The general steps to be followed are as follows : ASSEMBLING THE MATERIAL BY PAGES and checking it l.fith the dummy. Changes in unsatisfactory pages can be made at this stage . INDICATING PAGE NUMBERS on each page and on each piece of copy to be inserted on the page . Pages should also be marked to indicate where each inser t goes, and the inserts al so should be marked to indicate their position on the page . CHECKING OF COPY FOR SIZE to make sure that all material will fit properly on the page . All inserts should be checked to see that proper instructions are given for reproduction, especially when they require unusual reduction or other special treatment . CHECKI NG THE ENTIRE REPORT to make cer tain that final copy has been prepared for all pages, including all special pages (the cover, security page , contents page , etc . ) that are used . Reviewing ~inished Copy Always proofread all material prepared in any stage of the production of a report. Errors and omissions can occur whenever text and tables are retyped or chart sketches redrawn . Unless immediately discovered and corrected, they will probably be repeated in the next stage and may not be detected . A final review of the entire report should be made after all related material has been prepared and assembled for reproduction. Although no rules can be definitely established to i nsure that an adequate review will be made , since this is dependent largely on a thorough knowledge of the subject matter, neverthel ess, certain specific checks are highly desirable . PRODUCTION TECHNIQUES CHECKS FOR COMPLETENESS. Is all information listed in the "Contents" actually included in the report, and vice-versa? Are page numbers and running heads affixed to applicable pages? Are all chart and table titles, and necessary labels, and all other details included? (In the various stages of processing not only such details, but even whole sentences may occasionally be omitted . ) CHECKS FOR ACCURACY within text, tables, and charts . For example, is all spelling correct? Do percentages in the tables add to 100 and are chart values correctly plotted? CHECKS FOR CONSISTENCY among text, tables, and charts . Do figures relating to the same subject matter agree? Does material in onepart of the report seem to contradict or invalidate material in another part? CHECKS FOR UNIFORMITY of treatment. Is all text appropriately typed and underlined? Are all tables properly typed and ruled? Are similar charts or tables treated similarly as to their titles and other details? Preparing a Final Dummy After the final copy has been assembled and reviewed for accuracy and completeness, and the last-minute changes have been made, the report is then ready to be sent to the reproduction unit. A dummy should accompany the material to indicate the order of the parts, the pagination (with a separate page in the dummy for each numbered, unnumbered, and blank page in the report), and the location of each insert chart or table. The dummy shows where each component of the report is to be placed and it is a guide to proper assembly of the printed r eport. PROOFREADING Checking text, tables, and charts for errors is one of the most tedious but more important chores in statistical presentation. Even a small harmless error is bad , because it raises doubts in the reader 1 s mind about the entire job. Some of the most serious mistakes have survived several earnest proofreadings . A few samples of errors that might occur are shown in the paragraphs below: "Construction in March decreased almost 150 percent. " (One decimal place too many: should have been "15 percent . ") "It is the opinion of this Committee that the proposed schedule can be met." (Should have been "cannot. ") "Both manufacturers have stated that they won 1 t convert their plants before the end of the year. " (Stenographer misread her shorthand: should have been "want to" . ) One reason why mistakes like these get by is that many reports must be done too quickly. Everything about these jobs is pushed through with such haste that many errors are made and some are not caught. Some things can be hurried, but proofreading is not one of them. The main reason why errors slip through is that most of us see what we expect to see . We reaj words, and even groups of words , by scanning them quickly. We expect certain combinations of l etters and words , and we can see them whether they are there or not. The more familiar we are with the subject written about, the more we see what we expect; t herefore, a writer is usually the worst proofreader of his own writing. Some writers have trained themselves to do this job well, but they are the exceptions . Catching Errorfi There is no foolproof method of finding mistakes, but a good way to find most of them is to use a system -make proofreading and checking a nethodical procedure : 1 . Split the proofreading job. Most of us lack the training and skill to do this job all alone -especially if there is a lot of it. The best way is to divide the work : have two persons "read back" the text and tables; and have others machine-check the tables, review charts , and look for particular kinds of errors . 2 . Check for one thing at a time . It is easier to find mistakes if you look for one kind fi~st ar.d then for another. This may take longer but it is safer. 3. Have one check made by a "stranger . " Someone who does not plan or write the material may pick up mistakes that the writer may miss . 4. Make sure each proofreader knows what he is supposed to check . Unless each knows what pages to check and what to look for, some things may be checked twice; some not at all. 5. Make sure that everyone initials what he has checked . This shows which pages have and ha7e not been proofread. These five rules for checking may seem like too many . They may be in some cases . Like all rules , they must be applied with judgment: for a 3hort report a simpler one -man system may work well; for a long report only two or three people may be available for the proofreading job. When t here is a large amount of checking to be done, however, and sufficient personnel to do it, this method is effective . Readi~ Back e best way to check typed (or drafted) material is to read it back with someone else . One person reads aloud from the original copy (long hand or preliminary typing); the other scans the proof (the next preliminary typing or the final typing, depending on the processing stage) to see that it agrees with the reader1 s words . Two people working as a team are likely to catch errors they would miss if checking separately. This method lets each person fix his atte~tion on one sheet of paper instead of reading ba~k and forth between two . Also, some errors are more noticeable when words and figures are spoken ~han when they are read silently. PROOFREADING Marking Copy for Changes All Typing (or printing) mst be proofread t o det ect errors -those made by the author as well as the typist. These errors can be in spelling, punctuation, omission of words , capitalization, alignment, spacing, or any of the other details of putting typewritten words or figures on paper. Marking errors or changes clearly and corr ectly is an important part of proofreading and editing. Poorly marked-up copy leads t o more errors, more work and more time lost. Always use light blue pencil on final copy, because blue does not photogr aph (with most film). Proofreaders ' Marks. These are symbols indi cating specific types of errors or changes . They are easy for a t ypist t o learn and follow. The most important proofreadErs' marks are explained in the list on the next page . They are s t andard marks used in printing and publishing (except for a few changes to fit typewriting problems that differ from typesetting problems) . Using Proofreaders ' Marks . There are t wo systematic methods of correcting proof -by putting the corrections on or between the lines of type directly at the point of error, and by marking the errors in both the text and the margin . The first usually works well in double-spaced copy, but seldom is practicabl e in single-spaced copy because of insufficient r oom bet ween the lines for corrections. The second is applicable under either condition, and is especially suitable in single-spaced copy because it requires a minimum of niarking in the text. The examples below show the application of proofmarking procedures under each method . When corrections are made at the point of error only, standard marks may be modified wher e necessary in order t o improve cl arity (as shown in the first example) • EXAMPLE OF COPY CORRECTED ON OR BETWEEN THE LINES : _..~Example 1 W This method may be foilowed when th~ypewriter ~ [lines are.., spaced.ene~ apar~ ~therCwise, the ~emplifi~method\should be used in marking ~rors. 9{ Regardless of the method ~'Q) impor tions) should be ~theline to •ffiiefi • , .1'1-, hides they re:ffer t o. The typlst,r copy.:z:uol der c=.:J ~ ). AA 9. the copy and co~ections below~· EXAMPLE 1 (As Corrected) This method may be foll~wed when the typewriter lines are double-spaced. Otherwise, the method exemplified below should be used in marking errors. Regardless of the method used , one important rule always should be observed. This rule is : al l changes (or corrections) should be on or above the line they refer to. The typist's copyholder hides the copy and corrections below . EXAMPLES OF COPY CORRECTED IN THE TEXT AND IN THE MARGIN : EXAMPLE@ 2 tf-. / J.. In this method" t he errors teat are .i./A/# indRated in the text are expJijl§lned b;y;:the t/i./0/~ proofreadersv marks ,0'n the margin11 as con~ trasted ~the method sho>l!l in ~ample c 1 above , it has a geU:eral rather than a ]}"1/W JP..mited us[ age . ~en there is more than · f CD mark on a line of t ext, the marks and ~ co~ections in the margin are entered in .uA~ pFel'liso]y the same or der in which the -~--~ erro~ rrv:r'11 . The marginal marks are separatedAj.J.Yfc. t.e_ach.ftrormtifte. other by ¢lash lines. EXAMPLE 2 (As Corrected) In this method, the errors indicated in the text are explained by the proofreaders' marks in the margin . As contr asted with the method shown in Example 1 above , it has a general rather than a lLmited usage . When there is more than one mark on a line of t ext, the marks and corrections in the margin are entered in exactly the same or der in which the errors occur. The marginal marks are separated from each other by slash lines. EXAMPLE 3 Both margins may be use~for proo([readers 1 margine when a page contains a one"column~~/ arrange:ment of text. Each mark should appear in l!.lhat margin which is,rlos;(st to the ujtrf-v-' coiAesponding error. This procedure ~good iB oitt1ations ~·l-J:E?f9 there areJhU many errors t o be corrected. It affords t wice as much marginal space for correc tions! tt times , even this is scarcely enough . PRODUCTION TECHNIQUES This Proofreaders 1 Mark 1\ / # C/) = ][ 0 Yi STANDARD PROOFREADERS 1 Means : Take out letter(s), word(s), or punctuation mark(s) indicated Delete and close up space Close up space Insert letter(s) or word(s) as indicated Change word(s) as indicated Change letter(s) as indicated Insert typewriter space(s) or line(s) where indicated Insert period where ~ndicated Insert comma where indicated Insert apostrophe where indicated Insert quotation marks where indicated Insert parentheses where indicated Insert hyphen where indicated Use lower-case f or letter(s) indicated Use initial capitals for word(s) circled Transpose or transfer letters or words indicated Replace with a capital the letter(s) indicated Let it stand as it was . Disregard changes above the dots Move to right ( J ) or left ( [ ) as indicated Move up (rl) or down (LJ ) to the point indicated Spell out the number(s) or abbre viation(s) marked with a circle Start a new paragraph as indicated Should not be a separate paragraph Query . Word, spelling, etc . in doubt MARKS Mark in the Mark the Marcin Typing Like This : He marked the proo¢f. He marked the ~proof. He marked; the proof. SomeEone marked the proof. He maZ:ked the proof . He m~ed the proof. He" proof . He marked~ proof. He m¢rked the proof. :;.-He marked the..proof . Standard marks were used . 0 He marked the proof~ YesAhe marked the proof . He used proofreadersvmarks . v v fl/fl He said, Mark the proof . (j) He marked 15" or more" pages . I:/ The copy is self" explanatory. He ~arked the proof. €arking proof~ He ~roof)the\matke~. He ~arKed\ proof. He marked the paftof. He 'Iiiarked the proor(lway). ~8 marked the proof. ~ ~He ~fl.ESJ~ the proof. ] He] marked the proof . ll -He marked the proof~ marked the ~proofreading . IThe r eader marked !RaPlEefr~ ~e proof was read by He § of-marked) the copy . PROOFREADING Proofreaders should always be on the alert to avoid two possible pitfalls : 1. Words that sound the same but are not spelled the same . 2. Details that cannot be read, such as alignment, spacing, and other layout features . The English language has many homonyms words that have the same sound but different meanings , such as : coarse/course, principle/ principal, capital/capitol, etc. The meaning of the sentence usually shows which spelling is the right one; but proofreaders are concentrating on words , not thoughts. Typists and other clerical proofreaders should not be expected t o think through all of the sentences. This is especially true when the volume of checking is large and when it has t o be done quickly. When reading proof with ~~other person, either spell out homonyms or em?hasize the key letter. For example , "principal with an s.", or "capitol with an .2", etc. When reading proper names , make it clear whenever the spelling is not the most common one : "Thomson -without a :Q", "Andersen with an e" , etc. Spacing and other non-readable features of the copy are harder t o check by the reading-back method . The reader can note a new paragraph, a word in all caps., or an underline easily. It is harder t o note and explain proper indention, spacing between lines, or centering of captions. It is generally better t o have these checked by one person who knows how they should be done and can make separate checks for these design details. If all typing and drafting is done on typing layout sheets, these layout details are easier t o read back and check. Safety Rules for Reading Back . Always read fr_om the original draft to the proof. This will increase the probability of finding errors. FOR THE READER: Read in a clear, audible voice . Call out every punctuation mark, capital, underline, paragraph, and all other typographic or style features of the original copy. FOR THE; LIST~R: Take nothing for granted; remember that the burden of proof is on you . Foll ow the reader 1 s voice closely; do not read ahead of it. Be sure to listen to every word . Mention anything that appears on your proof that the reader does not read. Machine C.ht?c.king. Besides reading back the words and figures , all totals and calculations in a table should be checked on adding or calculating machines as an added insurance against mistakes. Derived figures, such as averages, percentages, and rates, should be recalculated. Totals should be checked by adding their subordinate figures . If a total is divided into subtotals-, each subtotal should equal the sum of its subordinate figures, and the t otal should equal the sum of the subtotal s . When the figures are additive both horizontally (across the rows) and vertically (dovm the columns) , they should be verified both ways . I~ is best t o machine check a table in each stage of preparation from worksheet through all intermediate typings to final copy, though this is not always possible due to quantity of work and time limitat ions. At minimum, a table should be machine checked from the original worksheet (prefer ably by someone other than the person who made the computations) , and from the final copy. CHAPTER 7 DISPLAY PRESENTATION In this manual, large charts that are intended to be seen beyond the normal reading distance are called display charts. This general term is used to cover briefing charts, wall charts, easel charts, flip charts, conference charts, and lecture charts. Whatever they are callBd, and however they are shown -as large drauings, blow·ups of small drawings, or as pr ojections on a screen -they provide an image large enough t o be seen by a group of people. REFERENCE CHECK LIST Before starting to prepare displays, become familiar with t hese t opics : Graphic Presentation Page Chart Types • • • • 46 Planning and Designing Charts • 66 Shading •••.. • 94 Tabular Presentation Basic Tabular Arrangements 13 Component Parts of a ~able 19 Space ~equirements 39 Display Presentation is a Specialty The first thing t o know about display presentation is that it is not easy t o do well. The suggestions in this chapter will help, but display presentation is an art in itself which is acquired by experience. A good display chart cannot be made by merely enlarging a report chart, because the requirements are not the same . Displays should be bolder and simpler because the reader cannot bring a display up nearer t o his eyes for close examination and he usually is not able t o study it long. Selectivity -the First Key t o Effective Display In all statistical presentation there is a tendency t o show t oo much. This is bad enough in a written report but it is worse in an oral report. On a printed page , the reader can skip some of the material, choosing the things that he wants t o study. In an oral presentation the reader must l ook at everything. Most readers are able t o absorb and retain only a limited number of visual messages . If they are presented with too much material at one sitting, they must ignore some of it. Therefore, the fir st rule of display presentation is t o show only the highlights. Make a picture of only the principal things the audience should learn from the presentation. Explain the "why" and the "how" orally. Most viewers are able to grasp only one idea at a time. Usually one major idea in a display is enough. Some secondary information can be presented but the typical viewer will get the basic idea quicker and more certainly if the presentation is focused on only one thing. Most viewers are able to follow only a limited amount of detail. In presenting an idea, include only material that is needed . Information added as a matter of interest detracts from the basic idea even though directly related to it. Usually the viewer is given only a limited time to study each display, therefore it must be understood quickly. It is important to keep the presentation as simple as possible. This means simplifying terms and concepts so that they may be grasped at first reading. It also means r educing all elements, such as titles, labels and scales to their simplest form . Size -the Second Key t o Effective Display A display is almost useless if the audience cannot read it. Since the viewer cannot examine it closely, the designer must be especially care~ ful t o make everything large and bold enough to be legible at a glance. This means that the entire display should be easily readabl e by the farthest person in the audience . This usually requires making lettering and lines much larger and heavier than seems necessary to the person designing the displays. DISPLAY PRESENTATION 1 Displays For Highlights Only 2 One Idea Per Display 3 Needed Material Only 4 Keep It Simple 5 Make It legible ~ Good display presentation. Poor ~ RULES FOR USING DISPLAY PR~ENTATION 1. Use display material only for the highlights or key points in the presentation. 2 . Show only one idea or concept on each individual display. 3 . Include only material that is actually needed to understand the display. 4 . Keep the words and concepts as simple as possible . 5. Follow standards for getting at least minimum legibility. One or more of these rules must be broken occasionally, but if they are not followed, the display will always be less effective. The effectiveness of display presentation is directly related to how cl osely these few but important rules are followed . DISPLAY METHODS DISPLAY METHODS Display presentation can be divided into t wo basic methods : displays projected on screens and those drawn, painted, or otherwise prepared as posters. Both basic methods may be subdivided into two maj or types : Projection Displays : 1. Transparent projections : using a transparent film or slide (usually made by a diazo process or by phot ographing the original material) . 2. Opaque projections: using any small opaque original. Poster Displays : 1. Big original displays : drawn the s:::.ze they are t o be shown . 2. Blowup displays : drawn small and then "blown up" t o an enlarged phot ographic print. Changes in facilities, equipment, supplies , and techniques have been introduced rapidly in recent years. It is desirable, t herefore, f or personnel responsible f or statisti cal presentations t o obtain the latest informat i on on capabiliti es, characteristics and lioitations from l ocal personnel responsible f or the preparation and projection of displays and t o inquire about new developments from time t o time. This manual does not discuss speci alized display methods such as magnet boards (the metal boards t o which display material i s fastened by small magnets) or flannel boards (velvet covered boards t o which display material -with a sandpaper or "flock" backing-is held by friction) . These are in general not suitable for statistical present ation . Ordinary chalk boards such as blackboards or green boards are not discussed separately because they are generally used t o present broad concepts, outlines, and other nonstatistical material . Moving-picture projection is not discussed because it is a large subject almost entirely out of the field of statistical presentation . CHOOSING BETWEEN PROJECTIONS AND POSTERS Many factors must be considered before deciding which of the display methods to use and no general rule will hold in all cases. The choice of method will depend partly on the nature of the presentation, partly on the occasion, and partly on the size and characteristics of the audience. The kind of equipment anu faci lities available may greatly influence the choi ce of method . In recent years various types of projection display have been used with incr easing f r equency f or ora_ briefings. This growth has r esulted in large part from improvements in equipment used in preparing material f or such present ations. BR=EF SUMMARY OF PROJECTION METHODS Type of Projector and Maj or Transparency Size Advantages Slide Lightweight, small. ~X2" Slide s inexpensive. Provides slides of more 2 3/4" X 2 3/ 4 " workable size. Widely available stand3 l/4" X 4" ar d size in many organizations. Overhead !1" May be operated by 711 X 7", speaker, who faces audi 1011 8" X or ence. Uses transparen 1011 1011 X cies large enough t o write on . No slides or transparen~ector Type) cies needed. Up t o 10 11 x 10" Q/ §) Ot her sizes include : 4 1/2" x 4 1/2", 6 1/211 h/ Size of material t o be projected. Maj or Lighting Disadvantages Conditions Requires high degree of Semi-darkened skill. r oom needed for good results. Equipment f or making and Same as f or 211 snuwing not as widely avail-X 2". able as the other slide sizes. Limited t o horizontal dis- Same as f or 211 plays (i: full sl ide area X 211 • is used) . Slides are heavy and bullw (in quantity) • Transparencies are bulky Satisfactory with (in quantit y) • aver age r oom lighting; better with lighting dimmed . Image not as clear as Darkened r oom others because of light needed f or bright l oss. screen image . x 6 1/2", 7 1/2" x 10" and 13 3/4" x 18 1/4" . ~14 · 853 0 -66 -9 DISPLAY PRESENTATION Projections can be shown effectively either to large or small groups when appropriate equipment, facilitie.s , and space are available . Films and slides are generally easier to handle , transport, and store than posters. These factor s are especially important when the material has a security classification. Poster-type displays are often better when frequent reference to the material is desired . Poster displays are also useful for public exhibits that will be shown continuously for days or weeks. They may be hung on walls singly or in groups . In fact , an entire series of charts can be hung around a room (often called a "chart room") or placed on sliding display panels, so that it is easy to make comparisons from one chart to another. Charts that can be rolled up are used extensively for oral presentations that are repeated many times to small groups at different locations and under unpredictable conditions. A speaker1 s notes may be written on poster displays in light blue pencil. These can either cover supplemental information, such as numerical values of high and low points, or an outline of the entire presentation. Overlays may be used in certain poster displays . Sometimes speakers make their own display charts _and illustrations as they present their oral analyses . This may be done by using large blank pads on easels and either crayons or wicktype ink markers. "Bleeding" onto the page below may be avoided by placing a sheet of acetate above it. PROJECTION DISPLAYS Three different methods may be used to make projections and it is important to know t he type for which the displays are being made . An original designed for one type of projection may not be good for another. For example, overlays may be used in addition to the basic chart for transparencies. Opaque curves, bars , etc., are frequently placed on transparent plastic sheets that are hinged to the basic illustration. This technique can seldom be used with slides. Basic slide or transparency sizes are discussed below. Originals intended for one size may not be nearly as good for the others. Each size has its merits and its weaknesses. Some of these are related to how ·large or how small the slide is, some to the type of projector used . Slide Projectors . Most Army install ations have facilities for projecting three si zes of slide: 2" x 2", 2 3/4" x 2 3/4" , and 3 l/4" x 4" . A projector designed primarily for a large size can be used to project a smaller size by using interchangeable slide ~arriers for the smaller size . Projectors designed primarily for the small size can often be used for larger slides if appropriate changes are made in the lenses as well as the slide carriers. The larger slides and overhead projectors are frequently used for large audiences (over 100) . The smaller slides and opaque projectors may also be used if the room and the screen, lenses, and other equipment are designed to produce legible images. The 2" x 2" slide requires comparatively little space when being carried or stored. Since it is square, it works equally well in the projector for showing either a horizontal or vertical display. But this slide is so small that it requires the highest grade of photographic skill, especially if paster-shadows or other marks must be taken out on the negative . It is almost impossible to hand-color sli.des · this size; but they are the least costly when using color photography since they can be made from standard 35 millimeter color film. The larger 2 3/4" x 2 3/4" slide is easier to work with than the 2"·x 2" and the 2 l/4" x 2 l/4" image that it uses is easily taken with a reflex type camera, (i.e . , it is easier to ge t a sharp focus on the ground glass of a reflex camera than it is through the common typRs of viewfinders on other cameras) . The standard mask for this slide is square, but masks can be made to fit both horizontal and vertical displays . The 3 l/4" x 4" size is the standard in many technical fields . Some libraries have accumulated slides of this size over many years. The good and bad points of this size are midway bo3tween those of the smaller and larger sizes . The main advantage is that this is a long-established size for which equipment is widely available. The main disadvantage is its horizontal shape, which limits displays to this shape . (A vertical shape can be used, but it wastes much of the projection area. ) Overhead Projectors. These are generally used to project large film transparencies that are placed on top of the machine . (Overlays can be used to develop a topic one step at a time . ) Light is reflected through the transparencies to an overhead mirror which deflects it to the screen. Standard sizes for these transparencies are 7" x 7", 8" x 10" , or 1011 x 10" . The most frequently used size is the one usually referred to as the 8" x 10", although the usable image surface generally measures about 7" x 9". A great variety of overhead projectors has been developed during recent years . Most of them are designed for use with small groups but some are equipped with lenses having long focal lengths. These may be used successfully in auditoriums seating several hundred people, if the screen is sufficiently large . Most of these projectors are portable, but many are permanently installed in rooms designed for briefings or for giving special instructions. Some are located in projection booths. Some may be placed beside or in front of the speaker and they project the image to a reflecting screen behind the speaker at the front of the room . This is known as a front projection. It permits the speaker to face the audience instead of the screen. He can draw or write on the film. If desired, the speaker 1s notes may be written on the film mounts . CHOO SING BETWEEN PROJECTIONS AND POSTERS Some speakers do not like to operate a projector. If a projector operator is used it is necessary to establish a system of communication from speaker to operator and it is embarrassing if the system goes wrong . Other projectors are located at the front of the room behind the speaker and they project the images toward the audience t o the back side of a translucent screen. This is known as a rear projection and it is less common than the other reflecting type . If overlays are to be used for both front and rear projection they should be hinged at the edge of the frame. The advantages of overhead projectors in clude the following: a large image at a short screen distance ; the speaker faces the audience instead of the screen; the traLsparencies are large enough for the speaker to read easily; the speaker can draw or write on the film during the projection; the images are bright enough to read in semi-dark or lighted rooms; and overlays can be used to develop a topic a step at a time. The disadvantages are that the transparencies are relatively bulky if used in large numbers . This can result in a troublesome problem if they have a security classification and must be transported by air. The portable projectors usually are carried on a cart, which is no problem unless the projector must be moved up or down stairs or from one building to another . Opaque Projectors . These project reflected images of any non-transparent material that is printed, typed, drawn or written. The larger models take material up to 1011 x 1011 • The advantages of opaque projectors are that they do not require slides; can project any material that is on hand; and they are especially good when no copies of the original material are required . The disadvantages are that: the images are not as clear or bright (with the same size lamp) as transparent projections because of light loss they must be viewed in rooms tha~ are at least semi-darkened; the heat of the lamps can scorch originals ; wax adhesives tend to melt, stain, and fall from the copy after it is removed from the projector; and they cannot be used for rear screen projection. Two or more projectors are provided in some briefing rooms . One projector may be used f or a basic image on one screen and a second may be used for a supplemental image on a second screen. Two or more projectors may also be used to show first a basic image and then a supplemental one on the same screen simultaneously. Both uses of more than one projector produce the same end effect as an overlay, but with less change-over time and without the confusion necessitated by bringing the overlay transparency into position. A change in the content of the basic image, however, requires changes in each of the transparencies on which it appears. When more than one projector is used, each projector usually requires a separate operator. Production Hints for Projection Displ ays. Useful information concerning the preparation and projection of transparencies may be obtained from manuals and catalogs prepared by the manufacturers and some distributors of projection supplies and equipment . Positive slides (3 1/4" x 4" or larger) can be brought up to date (by extending curves, adding simple bars, etc . ) with a special ink. They cannot, however , be revised by removing any of the existing image . Negat ive slides cannot be brought up to date , but can be altered by painting· out parts to be removed . Changes of this type can be made more easily on the larger sizes (7" x 7", and 10" x 10") . Cautions When Using Projection 1 . Before preparing slides it is desirable to find out how large and what shape the screen will be . If the screen is small, try to make the slide material simpler and larger to compensate. I f the screen shape is wider than it is high, try to avoid narrow charts because they will use only about half the screen area. 2 . If the audience is spread out , use a matte-finish screen, not a beaded one . The beaded surface is brighter for the people directly in front of it, but not so good for those at the sides . 3 . If the projector points up at the screen at a sharp angle, tilt the screen forward, if possibl e, to place it at right angles to the projector . Otherwise the image on tte screen will have a "keystone'' distortion: that is, the top will be much wider than the bottom. 1/ 4. It is important to know what room lighting will be used. In a room just light enough for reading or making notes, the projection lamp must be stronger than in a room that is darkened . 5. If the screen image is pale or if it lacks contrast, try to move the projector closer . Usually a smaller, bright image is better than a large washed-out one . If the machine will take a brighter lamp, that is a better sol ution. ~ 6 . Check all materials and equipment beforehand under conditions which resemble those expected during actual presentation as closely as possible. l/ Some manufacturers of portable screens supply devices to correct this effect. Screens that are fastened to the wall or hung from the ceiling can usually be tilted by pulling the bottom edge back or moving the top edge forward . Some are permanently installed at the angle that is the most satisfactory. ~ It is important to check manufacturer' s instructions before using a larger lamp: additional heat may damage lenses if the projector is not designed for it. DISPLAY PRESENTATION POSTER DISPLAYS Posters are usually made on large sheets of illustration board . This has a good drawing surface and takes all kinds of ink, crayon and paint (except oils) . Its erasing qualities are good . Illustration board comes in two standard 3011 4011 8011 sizes : x and 4011 x 6011 • A 6011 x size is made , but is not widely available . 4011 The 3011 x size is good for fairly close viewing (10 or 12 feet) . It can be used at greater distances only if the chart is so simple that the lettering can be made large . The 4011 x 6011 size is best for general use, but is not big enough for large audiences . 8011 The 6011 x size is best for groups of 100 or more , but it is so big that it is awkward to work on and to handle . If an image larger 6011 than 4011 x is needed , projection of some type is usually better. The step-at-a-time technique can be used in poster displays by hiding part of the chart with one or more sheets of white paper (lightly attached with tape or rubber cement at the top corners) and then uncovering it at the desired moment . Production Hints for Poster Displ ays . Get striking 11 reverse11 (white on black) effects by making photostat negatives of titl es, or other key parts of the display and attaching them to the original . Another method is to fasten a strip of black paper to a chart, then attach white gummed l etters . This is a good way to get emphasis without using color. Photographic blowups are useful: (1) when available personnel lack experience in making poster-type originals , but can make good small originals. ( 2) when existing small size material is to be used for display. (3) when more than one copy is needed for display . Blowups are usually made with pr ofessional photographic enlargers . They can be made by proj ecting a negative film to a sheet of sensitized paper tacked to the wall, but this is an emergency make-shift method . Even if the proper enlarger is available, the large size sheets of photographic paper are not always available when needed . Production Hint for Photographic Blowups . Wet photographic blowups before mounting so that they will dry flat and without wrinkles . Mount a sheet of Kraft paper on the back of the cardboard to prevent the blowup from curling as it dries . For quick-mounting a blowup, make it 3 or 4 inches oversize each way; then fold this excess ove r the back of a piece of beaverboard and staple . The r esult is not as flat as it would be if the material were pasted, but it is usable . Flop-Overs . Either big originals or blowups can be made as rigid displays or as 11 flop-overs11 • FloJ?-overs1 sometimes called 11 flip sheets11 , or 11 fll p char-cs11 , are flexible sheets fastened at the top with staples, cords , screw posts , or rings, or between two strips of half-round wood . After each chart is shown and discussed, the speaker or an assistant turns it up , over, and down behind t he others . Flop-overs are not as bulky as illustration boar d , and so are often used when a set of displays must be carried about from place to place . They have three disadvantages : If much larger 3011 4011 than x , they are hard to lift over; they are likely to tear; also, since the sheets are fastened ~n a fixed order, they are hard to use except in this order. They are satisfactory for fixed sequences of ideas (as in training aids) but are of limited use for the presentation of statistical material. The tearing of flop-overs can be prevented by backing the paper with a sheet of mounting linen or binding the edges with a strip of transparent or white scotch tape . Fl op-overs can be made or mounted on stift sheets if hinged at the top with flexible strips such as linen binding tape . LETTERING FOR DISPLAYS Display charts are more likely to have lettering that is hard to read than other charts . The reasons for this are that many designers of displays tend to: 1 . Underestimate the sizes of lettering needed . 2 . Space words and l etters too close together. 3 . Request more degrees of subordination in the size of letters than can be shown successfully with available type sizes . 4. Include more information than can be shown l egibly. A large variety of TYPE FACES is available for use in display charts . Legibility depends primarily on size and design of the type . The size of type is generally measured in 11 points11 • A point is equivalent to about 1/72 of an inch. Point size can be approximated by measuring the distance between the top of a lower case letter with an ascender (such as an h) to the bottom of another lower case letter with a descender (such as a y) . The large variations in relative amounts of space assigned to the ascenders and descenders greatly affect the distances at which different type faces having the same number of points may be read . Wide , black l etters with short ascende rs and descenders and sufficient space between the letters can be r ead much farther than narrow, thi n letters with long ascenders and descenders S··t close together . LETTERING FOR DISPLAYS RULE-OF-THUMB FOR SIZES AND DISTANCES If you know the distance from the display to the farthest viewer, divide this distance (FEET) by 32 to get the smallest legible lettering size (INCHES) . Example : distance 16 ft. 7 32 = .5 inches = lettering l/2 inch high on the poster or screen. If you know the lettering size, multiply this (INCHES) by 32 to get the farthest distance (FEET) at which it will be legible . Example : Lettering . 35" high, multiplied by 32 = legible at ll ft. For individual type faces, however, point sizes are significant. They are useful for establishing limits on the smallest letters to be used . A limit of 14 points, for example , is generally used for 8" x 10" transparencies for projection. It is much easier to establish standards for Wrico and Leroy lettering ~~ides than for type faces. The standards presented below should be regarded as guidelines that should be modified to meet the needs of specific conditions, including size of the screen or poster, its distance from the farthest viewer, type and capacity of the projector, the shape of the room and the amount of light in it. If the originals are made small, and e~larged for presentation (by using blowup prints or screen projection) the l ettering methods for making report charts may be used. The following table lists minimum lettering sizes recommended for these small-size originals : SIZES OF SLIDE ORIGINALS AND LETTERING SIZES ~ When Viewing Conditions Are : If Smallest Above Below Lettering Average c;) Average Q/ Average g/ Size Is: (Capitals) Use this size original §/ u 6 1/2" 5 l/4" 4 1/4" , .100" 7 l /2 6 1/4 5 .120 9 7 1/2 6 .140 ll 1/4 9 3/8 7 l/2 .175 12 3/4 10 1/2 8 1/2 . 200 15 12 1/2 10 . 240 18 l/2 15 1/2 12 1/2 .290 22 1/2 18 3/4 15 .350 27 22 1/2 18 . 425 32 1/2 27 21 1/2 . 500 ~ This also applies to material used in opaque projectors or to originals used for blowup charts . ~ } D~stance from { ~ ~/2 x} Im~gc:e~:ze gj v1ewer t o screen x 9 ~ Longest dimension (not counting margins) . t/ Standard Gothic letters or equal. Before selecting the size and face of the smallest l etter in a display it is important for several people to look at samples of different sizes and faces under the most adverse viewing conditions that are anticipated. Typewriting is a satisfactory method of l ettering, especiall y for text and tabular displays, if the original is small enough . A good rule of thumb for material prepared on a standard typewriter is to use an original no larger than 1/2 a letter-size sheet. The originals may be larger if the less common machines with l arger type are used . When used in display work, typing should be clean, but heavy. The light, sharp typing needed for good report reproduction is generally too weak for display presentation~ Headliners also may be used, when available, for both the large and small l etter sizes . The exact specifications may be obtained from re tailers' catalogs. Type stamping (with printers ' type set in a large metal type holder a~d used like a rubber stamp) is a good method for small displays , especially when certain words or numbers are repeated ; but the equipment needed is not widely available. For big original poster-type displays , the choice is between three l ettering methods : large size l ettering guides , paste-on l etters, or freehand lettering. Tracing stencils (letters ihked and filled in free-hand) are a good last resort when other methods are not available . PASTE-ON LETTERS 'are available either wi t h or wi t hout a gummed back. These l etters take more t ime than lettering guides but they provide large, bold lettering and do not require waiting time for work to dry. LETTERS PRINTED ON CARD STOCK generally may be fit ted (face down) into a composing stick, fastened in a strip with a piece of double adhesive t ape which is then turned over and pressed into position on the display. Thi s is a fairly quick lettering method but usually is not suitable for displays to be shown in original, for each letter card shows a shadow that gives the display a patched-up appearance . LETTERS PRINTED ON TRANSPARENT ADHESIVEBACK FILM are cut from the sheet and placed directly on the display, one at a time. They are applied easily , but t his is a fairly slow method . These printed l etters are thinner than cardboard letters and so cast no shadow; but they do reflect light that can be objectionable on a display vie1o~ed in original. WRICO AND LEROY LETTERING GUIDES may be used for display lettering. The table at the top of the following page lists recommended minimum sizes of poster let ters (or screen images) for various distances. These sizes are for the regular "Gothic" letter style . The l argest Wrico guide is 2 1/2" high and the largest Leroy guide is 2" high . . DISPLAY PRESENTATION The key to legibility is the size of the material in the display, not the overall size of the poster . A smaller display should use the same size letter ing and line weights as a larger display unless it is to be seen at a closer range . MINIMUM LETTERING SIZES FOR POSTER DISPLAYS ~ LEROY Q/ Distance -WRICO g/ Display to Farthest Guide Letter Pen Guide Letter Pen Viewer No . Height gj No . Height gj (Feet) §/ 350 . 35011 4 lO -ll 35 . 35011 3 ------ll-12 40 . 400 A 425 . 425 5 12-15 ----- 500 . 500 6 15-18 50 .500 A ------18-21 60 . 600 B 700 . 700 9 21-23 -------. ----23-25 75 . 750 c 1000 1 . 000 10 30-35 100 1 . 000 c ------35-40 125 1 . 250 D 1350 1.350 12 40 -45 ----- ------45 -50 150 1 . 500 D ------55-60 185 1.850 E 2000 2.000 14 60-70 -----------70-85 250 2 . 500 E NOTE : These standards apply also to other lettering methods , such as headliners , gummed letters, or free-hand lettering of equal size and the same style . ~ Displays to be viewed in original size or images on a screen. Q/ Scriber type guides . gj Stencil type guides (manufacturer also makes scriber type guides up to . 625 11 ) . gj Numbered sizes are tubul ar type; lettered sizes are 11 brush11 type pens . §/ Recommended maximum for good viewing conditions . If lighting, seating, etc ., are poorer than average, decrease these distances by l/3 or increase lettering size by l/2. Style of Lettering . The best practice is to use the simplest and most legible style of letters. Standard lettering guides make 11 block11 letters with all strokes the same thickness and with no serifs. This style is recommended for most display work because it is easy to read and easy to draw . Th~ particular alphabet used in standard Ler oy and Wrico lettering guides is "engineering Gothic". All the capital letters except the 11 P are about the same width: (about 3/4 the letter height) . Displ ays that are all lettering are usually easier to read in a lettering style that uses both thick and thin strokes and seri fs . Distinctive styles are sometimes effective for main titles (if the title is short and the style is legible) . CONDENSED LETTERS (less than half as wide as high) are not recommended for standard use because they are less legible than normal width letters . Larger condensed lettering, however , can be more legible than smaller non-condensed lettering that takes the same space from left to right . For this reason long titles that cannot be shortened should have condensed letters. Caution: If condensed letters are made too bold (thick stroke) they seem to "fill in11 at a distance and are hard to read . CAPITAL LETTERS are the standard for display charts and most tables . Long notes or subtitles, however, (especially those that take two or more lines) are usually more legible in lower case . Text displays are l ettered in all capital s if they are very brief. If the text is more than 25 words long, use capitals for the title and lower case for sentences and paragraphs . SHADING DISPLAYS The principles to follow in shading displ ays are the same as for any other graphic presentation. Small originals (to be used for blowups or projection) can be shaded with the same materials that are used for report charts . Shadings for large original displays must be considerably coarser. Patterns must be big and bold on the large charts to make them easy to distinguish at a distance , and especially if the large charts are to be greatly reduced in order to make small hand-out copies or record co~ies. The shading patterns shown in the last column of page 96 will be satisfactory for charts no larger than 4011 x 6011 • These may be reduced i n order to bring them down to letter-size copies . These patterns may be supplemented by open (white) and solid (black or a dark color) . RULES FOR LEGIBLE DISPLAY LETTERING l . Use a simple letter style . Block (uniform line weight with no serifs) is usually best. 2. Use standard letter proportions (letters averaging about 3/4 a s wide as high) . As a standard practice, condensed lettex-'s should be avoided . 3 . Make lettering contrast sharpl y wi th the background . Use black lettering on white ; wij~te lettering on black. Grey l ettering must be much l arger than bl ack or white to be equally legible. (Follow the same "contrast" principl e when using coloJ;'. ) 4; Be sure letters are large enough . (See tables of minimum lettering sizes . ) Few objections will occur if lettering is too large -many, if it is too small. 5 . Space letters properly. Keep white space (not distance) between letters approximately equal, and large enough to prevent them from "running together" when viewed . SHADING DISPLAYS As explained previously, black and white should be reserved for shading only the smallest areas . Colored shading is available (sometimes on special order) in several coarser patterns, and is effective for display presentation. It is important to use a different pattern f or each color if black and white copies are to be made . * If no copies are needed, co~o r may be substituted for shading; but this is even more difficult than shading, especially when several different areas or items must be identified. Another shading method, especially useful for charts larger than 4011 x 60 11 , is to 11 shade11 areas with strips of cellophane tape . In this method , the part to be shaded may be colored first (with colored paper, colored film, paint, or some other pigment) . Colors that will reproduce white are then striped with narrow strips of dark cellophane tape ; colors that will reproduce black are striped with white tape. Tapes in various colors also may be used. Tapes are available in widths l/32 of an inch and up. Tape that is l/16" or l/8" wide is easier to handle than l/32" and reproduces much better in small black and white copies. The stripes are usually about an inch or more apart. Crosshatching can be made by adding a second set of stripes at right angles to the first. All of the shading lines should be at a 45 degree angle . This line-at-a-time method of shading is effective for two reasons : it permits the use of color, but helps t o remove the need for many different colors on the same chart. For example, blue could be used for three different areas : one solid, one crisscrossed with white hatching, and one crossed with diagonal white lines . However, this shading method is fairly slow and painstaking, and should not be used if standard shading sheets are available . COLOR IN DISPLAYS Display charts are well suited to the use of color . Color costs relatively little if displays are made in one copy only, and color can help tell the story quickly and clearly. Two basic rules for using color in displays are : l . Make the color do some work. Do not use it without purpose . 2 . Use it in such a way that relationships are clear if displays are reproduced in black and white . Unless the first rule is f ollowed, poor displays can easily result. Applied only t o make a more vivid chart, color can do more harm than good by placing the emphasis in the wrong place or by making the display harder t o understand, rather than easier. * Shading f i lm usually comes with the glossy surface face up. This produces a bad glare that is undesirable on display charts to be viewed in the original. Some sheets are printed with the matte (dull) side up. These are much better for display purposes. Some typical ways to make color work effectively are to : l. Use light t ones of subdued colors (such as blue or green) to deemphasize secondary or background data such as earlier estimates, etc . Use br ight colors (such as red or orange) to emphasize primary important relationships, key items, etc . 2. Try t o standardize on certain colors for specific purposes of meanings in a series of simil ar or related charts . For example , r ed is often used for program and bright blue for program accomplishment . The rule t o apply color in a way that will make understandable black and white copies does not always apply, but it is gener ally desirable to code colored curves or areas by using a dif~ ferent patt ern for each color (including black). Tapes a_~d f i lms in display sizes are available commercially in many patterns and colors. See also "Selecting Curve Patterns" and "Shading" ; in t he graphic presentation chapter . When using color in displays, it i s important not to use too much of it. CO~ORING METHODS can be summarized only briefly in this manual. They can be divided into three basic groups : l . Pencils, crayons, pastels . 2. Inks and paints. 3 . Colored papers, adhesive tapes rod films. The first group seldom provides the best method . Pencil or crayon color is hard t o put on evenly, and pastels are messy. "Bui lders" crayons that are one-half inch thick, however , are useful for making some types of quick, emergency displays . Inks . For poster displays, opaque or pigment inks give clearer, more even col or than the transparent inks. For coloring glass slides or films, however, inks must be as transparent as possible. Paints. Water colors (opaque, transparent and tempera) have been used for display presentation for many years. A newer substance , however, acrylic-vinyl copolymer , has been gaining in popularity and has been replacing rubber-base paints. · Oil paints (either artists 1 or decorators 1 ) are not recommended. Color ed Papers . These are good for large areas, because they provide flat, even colors. They are not self-adhesive, so must be pasted on (usually with rubber cement) . Several grades are available . For most display uses, sheets that are prin~ed or coated with color are better than those color-dyed during manufacture . They are usually available in larger sheets and in a better variety of colors . Color ed paper s , however , stain easily and must be handled carefully when posting. DISPLAY PRESENTATION Self Sticking Tapes . These are coated on the underside with a 11 pressure11 adhesive . They are excel lent f or making col ored l ines or bars , and even for shading large areas. Tapes are often used in making transparency masters or originals as well as for posters . The narrower tapes can be used as rules . Caution should be exercis ed in using narrow colored tapes that are to be photographed, because they tend to darken to the extent that colors cannot be distinguished. Tapes are available in polyester film and in cellophane , vinyl, and plastic coated cloth. The relatively new polyester film is more tough and more resistant to moisture , chemicals , and to temperature changes than cellophane or vinyl tapes . They are also more stable and do not shri nk , even when attached for long periods of time . Most cellophane tapes have a shiny finish that may ref lect a troublesome glare, especially when used to cover large areas . Glare can be eliminated by using tapes with a matte finish . Adhesive Film Sheets . These come in two principal · types: those having a wax backing that adheres when the film is pressed down with a burnisher , or a tacky adhesive (non-melting) backing that adheres with light pressure . Many matte finished sheets which do not reflect light are available . Some tapes have color only on the surface . These are gener ally used when it is desired to rub some or part of the color off for art work separation. Tapes with col or not l imited to the surface do not need to be handl ed as carefully as the others . The polyester film discussed in the preceding section on tapes i s also available in sheets . These film sheets have the same characteristics as the tapes and they are widely used in making transparencies for overhead projecti on . HINTS FOR ORAL PRESENTATIONS 1 . Determine when displays will be visible. Indicate exactly in text when the displays and any overl ays are to be shown the audience and when they are to be withdrawn. They should be on long enough to be read , but not so l ong as to cause loss of interest. 2 . Sel ect times when the speaker will be silent. The audience should have a chance to study new and complex displays . 3. Determine what lights will be used . It is al so important to determine when to use each light. 4. Determine who will do the pointing. If it is to be someone other than the speaker , instruct him thoroughly so that there will be no communication failures . 5. Sel ect the best pointing method . Many different devices , including el ectrical pointers , are available to call attention to highlights . 6. Determine when the pointing is to be done . Many presentations are l.Illpali'ed by too much or too little pointi ng . 7. Determine what equipment will be used . This means what specific type· and size of pr oject or, screen, easel, and l ights will be employed , etc . 8. Learn how to operate this equipment . Some lecterns contain remote control and other electrical equipment for dimming lights , changi ng slides , etc . These must be investigated thoroughly before facing an audi ence . 9. Determine exactly wher e equipment will be pl aced . This includes the l ectern, if any , and the screen and projector or the easel. It is al so important to insure that the display is high enough so that the whol e audience can see all of it. 10. ~ehearse all aspects of the presentat ion. Obtain suggestions from associates and eliminate deficiencies . 11. Inspect all equipment shortl y bef ore speaking. Make certain that everything is in proper position and in operating condition. Make plans in advance for any necessary repairs of defective microphones or other devices . 12 . Do all these things before the audience appears . LIST OF REFERENCES The American Society of Mechanical Engineers American Standards Time Series Charts (ASA Yl5 . 2 -1960) , 1960 . Illustrations for Publication and Projection (ASA Yl5 . l -1959) , 1959 . Arkin, Herbert and Colton, Raymond R. Statistical Methods as Applied to Economics , Business, Psychology, Education and Biology, 4th Ed ., Barnes and Noble, New York , N. Y. , 1959 . Croxton, Frederic E. and Cowden, Dudley J.Applied General Statistics, 2nd Ed . , Prentice Hall, Inc . , 1955 . Haemer , Kenneth W. Making the Most of Charts . An ABC of Graphic Presentation. American Telephone and TelegraphCo . , New York , N. Y., 1960.Making Your Meaning Clear. Six Steps In Effective Communication. American Telephone and Telegraph Co ., New York, N. Y. , 1959 . Hanson, Kermit 0.Managerial Statistics, Prentice Hall, Inc ., 1955 . Marting, Elizabeth, EditorReports to Top Management, American Management Association, New York , N.Y. ) 1953 . Mills, Frederic C.Statistical Methods , 3rd Ed ., Henry Holt and Co ., New York , N. Y., 1955 . Neter, John and Wasserman , WilliamFundamental Statistics for Business and Economics , 2nd Ed ., Allyn and Bacon , Boston, Mass ., 1961 Schmid, CalvinHandbook of Graphic Presentation, Ronald Press, New York , N. Y., 1954. The University of ChicagoA Ma11ual of Style, The University of Chicago Press, Chicago, Ill. , 1949. United States Government Bureau of the CensusBureau of the Census Manual of Tabular Presentation, prepared by Bruce L. Jenkinson, 1949 . Department of the Army AR 320-5, Dictionary of Uniced States Army Terms . AR 320-50, Military Terms, Abbreviations, and Symbols -Authorized Abbreviations and Brevity Codes , 1964 . AR 325-10, Computation of Rates. DA Pamphlet No . l -10, Improve Your Writing. DA Pamphlet No . 325-5, Statistics. Federal Statistical Standards, 1965 . Joint Chiefs of Staff Dictionary of United States Military Terms for Joint Usage , 1964. United States Government Printing Office Style Manual, 1959. Word Division Supplement to Government Printing Office Style Manual , 6th Ed ., 1962 . 135 INDEX A Abbreviations : Boxhead 29, 30, 31 Charts 66, 82, 83, 99 Text.. 6 Titles . 20 Accuracy. 11 , 121, 122 , 125 Advantages of: Char ts 4, 44 Tables 4, 11 Text . ....... . . 4, 5 Amount scale for char ts: Breaking curves . . . . 81 Breaking the amount scale . . 50, 65 , 80 Captions (see also Captions) . 82, 88 Int erval s or divisions . 81 Location of 82; 88 Range ..... 80 Uniform scales 81 Analytical : Tables 12 Arrangement of stub 25 Arabic numerals For tables 22 Area char ts . . 62-63 A:ea vs . linear measurement 62-63 Arithmetic vs . logarithmic scales 75 Arrangements of items in stub: Alphabeti cal 22 Anal ytical 25 Blank lines . 27 Capital letters 24 Chronological 23 Customar y 23 Einphasis 23 , 25 Frequency 23 Geographical 23 , 25 Importance 23 , 25 Indention .. 24, 25 Items in stub 22, 28 Lengthening 26 Location 23 , 25 Numerical . 22 Progressive 23, 25 Quantitative . 23 Shortening 27 Spacing .. 24 Status 25 Subordinate 23 ; 24; 25 Arrangements, basic tabular: Divided stub and repeated boxhead table 15, 16 Repeated stub and divided boxhead table 17, 19 Repeated stub and single boxhead table. 17, 19 Repeated stub and repeated boxhead table. 18, 19 Single stub and single boxhead table . 15, 16 Subtabl es .... . 19 Arrows : Actual-estimated . . 93 Arrows and other label pointers 98 , 100-101 Pr epri nted . . . . . . . . . .. . .. 98 B Band char t ...... . . 51 , 73 , 105 Bar charts : Bar-and-symbol char t . . ..... 54 Breaking bar s . . 54, 65 , 88, 89 Change-bar chart .. 56-57 Components of . . . . 45 Designi ng for char ts ..7l Deviation-bar char ts .. 55 Graphic symbols, use of . .. 7l Grid pr opor tions . . . 78, 90 Gr ouped bar chart . . 55 Lettering sizes . . . . 102 Line weight staqdards . 102, 104 Need for zero or base line 52, 54 One hundr ed per cent bar chart . 55 Overlapping bar s . . . . . . . . 91 Paired bar chart . . . . . . . . 55 Progressive or step-by-step bar chart . 57 Range-bar chart . . 56 Ranking of i terns . . :i4, 7l Scale captions . . 88 Scale location . . 88 Scale numerals . . 88 Scal es for . . . . 88 Simple bar chart . . 54Sliding bar chart . 56 Spacing bars . . . 90-91 Spacing between lines . 103 Spacing letters .. . 103 Specification sheet . 105-107 Subdivided bar char t . . ... . 54 Supplementary data in the stub . . . 7l Page Bar charts -Continued Two-way progr essive bar chart 57 Uniform t~ickness of bars 54 Bar spacing . . . 90 Bar thickness . . 54, 71 , 90 Binding methods : Loose lea:' 115-116 Mechanical binding 115-116 Stitching or sewing 115 Blank lines, rules :'or: F::Jotnotes . 8, 9, 114 Paragraphs . 8 Rulings 114 Stub . 27 Tables 114 Text 114 Blowups , photographic 130 Board , illustra~ion 130 Boxhead: Abbr evi ation •. 31 Arrangement of co~umns. 29 , 30 Capitalization 31 Columns .. . 29, 30, 31 Construction 30 Contents 28 Location 14, 28 Punctuation 29 Spacing .. 29,' 30 Units of measurement . 31 W;,rding . . . 29 , 30 Breaking words .. .. . 5, 10 Breaking amount sca~es . . 50, 65 , 80 Breaking bars . . 54, 65 , 88, 89 Breaking columns . . 52, 65 , 89 Breaking curves 80 Broadside pages . 13 ' 114-115 c Capitalization: Boxhead 31 Stub 24 Titles 22 Captions : Amount-scale . . 45 , 82 , 88 Chart . . . . .. 99 Time-scale . . 83 ' 86 , 87 Change-bar chart . 56 , 57 Chart captions (see ~l;o -C~pti~n;) ( General notes . 99 Reference notes . 99 Source n::>tes -99 Subtitles . 99 Titles . 99 Charts . 44, 107 Chart components : Bar charts 45 Column charts . 45 Curve charts 45 Char t construction: Arrows and other label pointers 101 Breaking bars and columns . 89 Char t captions . 99 Char t lettering . 102-103 Char t s i ze. . . . 79 Connected colunns Curve patterns : 46 ,' ~~ Dr aftsman 1 s role i n . . . 78 Goal s and pr oj ections . 92, 93 Grid proportions . 78 Importance of construction standards . 78 Labels . . . . . . . . . . . 100 Line weight standards . . . 102, 104 Overlapping columne or bars . 91 Plotting. . . . . . 82 Preprinted par ts . 98 Processing copy 121 Scale rulings . . 84-87 Scales 80-83 Selecting soales for bar charts . . 88 Shading . . . . . . . . . 92-96 Spacing bars and columns . . . . . . 90 Specification sneets . . . . . 105-107 Time-scale designations and rulings 86-87 Chart content and design , keep simple ... 66 Chart data sheets . . . . . . . . . . . . . . . . . . . . . 121 Chart development . (See Development of char ts from tabular data. ) Chart labels ,. . . . . . . . . . . . 93 , 98, 100-101 Chart lettering: Letter count for char t titles and notes . 102 , 103 Letter spacing . . . . .. 103 Lettering guides and recommended pen sizes 102 Lettering sizes for charts 102 Line spacing . . . . . . . 103W::>rd spacing . . . . . . . 103 137 INDEX Page Chart selectipn, problems of Chart shape . . Chart size Chart sketches Chart specificati~n~ Char t style : "Decorative " style Examples . . . . . 11 Nacrative" style "Pictorial" style "Poster " style "Statisti cal" style Style for exhibits and displays Style for printed publications Style for statistical reports . "Three-dimensional" style ... Chart subtitles . . . . . . . . . Chart titles . (See Titles, chart . ) Chart types Char t width . Charts, poor Checking: Copy Machine Ciphers . Classification labels, security Color display . . . Column charts : Breaking columns . Components of . . Connected columns Deviation-column chart . Floating column chart . Grid proportions Gross and net deviation-colwnn chart . Grouped colwnn char t · . Lettering sizes to use . Line weight standards . . . . Overlapping columns (or bars) Range-column chart Simple column chart Spacing between lines Spacing columns . . . Specification sheet . Subdivided column chart Column and slope-c urve chart Column and step-curve chart . Column, slope-curve and range chart Columns : Arrangement of Boxhead ... Figures, placement in Lettered .... Number on page Overlapping . . Thickness . Combination type charts: Column and slope-curve chart . Column and step-curve chart . Column, slope-curve and range chart Slope and step-curve chart . Split-scale chart . . . . . . . . . Subdivided surface and curve chart Surface and curve chart Commas , in figures Components of: Charts Tables Field . Computer-printed tables (a~d · t~t~l;) : Contents page Copy : Assembling Checking Marking for change . Preparing . Processing Reviewing , finished Cover of report . . . Cubes .... . .. . Cubic vs . linear measurement . Cumulative curve chart Cumulative deviation chart . Cumulative frequency chart . Curve charts : Components of . . . . . . Cumulative curve chart Cumulative deviation chart Curve patterns, selection of Grid proportions Index-scale chart . . . Lettering sizes to use Line weight standards Logarithmic chart .. Multiple amount-scale chart 44 78 79 . . 70 105-107 76 :76, 77 .76, 77 : 76 , 77 . 76 , 77 . 76 , 77 . 76 , 77 76 . . . . 76 . 63 , 64 , 76 , 77 . . 99 46-61 79 65 121 11t 125 3, , 34 . 110, 111, 112 , 113 . . 133 -134 . 52 , 65 , 89 45 91 . 53 . 53 . 78 . 55 . 52 ... 10?. . 102 , 104 53 , 91 . . 53 .. 5?. . . 103 . 90, 91 . 106 52, 73 . 59 . 59 . 59 . 29 . 29 . 34 . 31 . 113 , 114 . 91 . 91 . 59 . 59 . 59 . . . . . 58 . 59, 84, 87 . 58 . 58 . 34 . 45 14, 19-38 . . 31 . .. 42-43 108, 109 , 110 121 121 123-125 117 121 121 108-109 . 63 . 63 . 47 :t6 . 45 . 47 . 47 . 97 7B . 49 102 : io2, 104 .. 49 , 75 . . .. 48 Curve charts -Continued Multiple-log amount-scale chart Multiple slope-curve chart Multiple step-curve chart Multiple time-scale chart Repeated time -scale chart Simple curve chart Simple slope-curve chart Simple step-curve chart Spacing between lines . . Specification sheet . . . Supplementary amount -scale chart Curv~ patternr., selection of Curves , kinds of . . .... . D Dash · · · · · Data changes in charts, how to show Data cheets for char ts Dates : In report titles In text . Decimals "Decorat~ve" chart style Development of charts from tabular data: Deciding how to chart . . Deciding what to chart Decidi~g whether to chart Examples Definitbns Designing oharts : Changes in content of the data Char t sketches Chart-style ..... Check list for Data sheets . . ... Designing bar charts Development of char ts from tabular data Insert char ts . . . . . . . . . Keep c~arts simple in content and design Planni~g charts . . . . . . . . Presenting charts to the reader .. . Status-changA-status charts . . . . . Subdivided surface charts . . . . . . Typical examples of poor presentation Designing e_ffective tables Deviation-bar chart .. Deviation-column chart Deviation-curve chart Diagrams . . . . . . Display presentation : Chart style for . Poster ..... . Production hints Projection Reference check list Rules . . . Selec-:.ivity . . Size ..... . Distribution list, report Dollar sign . . . . Double -page spread : Basic arrangement Stub Double stub . Dummy, :-eport E Emphasis , by allocation of space on page Emphasis in charts Emphasis in tables: In tbe stub ... Placing figures close together Vertical and horizontal comparisons Emphasis in text: Headings : How achieved Key lead words Underscore . . . Exhibits , chart style for F Faults , common : In report presentation . In statistical tables Field, of a table : Commas in figures Compcnents Decin:als . Dollar and percent signs Every cell must have an entry Field spanners . . . . . . . . . 49 . 47 47 48 48 46 46 46 103 105 49 97 97 32 74 70 109 6 31. 76 , 77 67 67 . 67 67 -69 108 74 70 76 , 77 66 70 71 67-69 . 72 66 66 66 74 73 65 11 55 53 51 44 76 , 77 127 , 130-134 129 127 ' 128-129 126 126 126 126 109 , 110 . 34 13, 15, 114 . 22 . 24 . 117 , 122 . . 121 . 44, 67 . 22 -25 12-13 , 15 12-13 7 -8 . 5 . 8 . 8 . 76 . 2 13 34 31 34 34 32 31, 32 INDEX Field, of a table -Continued Lucation ... ... . Placement of figures in the column Reference symbols . . Units of measurement Zeroes Figures : Arabic for table numbers Commas in . . . . . . . . . . In text. (See Numbers in text. ) In titles . . ... Placement in column . Film sheets . Film display, letters on Floating column chart . . Flip charts or Flop-aver charts Fold-in pages . . . . . . . . . . Footnotes in tables (see also Notes in charts, and Notes , tabular) . . . . . . . . . . Footnotes in text (see also Notes , tab~lar) : Examples , text Indention . Location Punctuation Reference s~~ls Ruling for Spacing for Systems . . Foreword Forms, master Frequency charts : Comparing frequency distributions Cumulative frequency chart Histogram . . . ... . Scales for . . . .. . Simple frequency charts G Glossary . . . . . . . . . . . . . . . . . . . . G"als, objectives , perfor mance, plans , requirements 58,59 , 92 , 93 Graphic labels . . . . . . . . . 64 Graphic presentation, pros and cons of 4, 44 Graphic symbols, use in bar charts 71 14 34 36 32 34 22 34 22 34 .134 . 131 . 53 130 .15, 42 , 115 Footnotes in text . 32 , 35-37 8-9 9 , 109 8, 114 8 8 8 8 8 108 , 109 , 110 120 60 60 60 83 60 .108 Graphic tables : Progress chart Rank chart Status chart Graphic types not recommended : Area charts Examples of poor presentation Maps Perspective charts ..... Perspective plus )-dimensional Pictographs Pie charts Three-dimensional charts Graphics : Kinds of Selecting right kind of chart Uses of . 61 61 61 62-63 . . 65 44-64 63-64 treatment 64 64 62 63; 64; ?6; 77 44 44 44 Grid proportions . . 78 , 91 Grid rulings (see also Scale rulings and Time-scale designations and rulings) . . . . . . . . . . 44-45' 84-87 Gross and net deviation-column chart 53 Grouped bar chart . . . 55 Grouped column chart . 52 Guide , lettering . 102 H Headings and subheadinga in text : Chapter . . . . . 7 Hajor subdivision 7 Minor subdivision 7 Minor subdivision component 8 Reference symbols in 36 Running heads 110, 111 , 112, 113 Section . 7 Headliners 131 Headnotes 34-35 Histogram . .. 60 Indention : Footnotes 109 Listings of items 8 Paragraphs 8 Stub 24-27 Index report 108 , 109, 135 Index-scale chart . 49 Inks 1)) Insert charts . 72 Introduction, report 108 Issuing agency , name of ... . J Justifying typewritten text (for an K Key lead words . . . . Keys , use of in char~s L Labels and label pointers (chart) . Labels , graphics (pictographs ) Layer chart . . . . . . . . . . Layout and organizat~on, page . Semifini shed pages . Variety . ... . . . Layout , page , examples Layout sheet, typing . Leaders, stut of table Legends , use of, in charts Lettering: Boxhead Display . Chart (see ~l;o .Cha;t · l~tte;i~gj Guides Pens .. Sizes Spacing Letter of transmittal Line numbers . . . Line weights: Footnotes , text General standards for charts Tabular rul ings . . . . . Linear vs. area measJ.ITement . Linear vs . cubic measurement Listing of i t ems Location: Figures in column . Figures in tablA . Footnotes, tabular Footnotes, text Logarithmic chart .. M Haps . . ... . . Master forms Measure , units of: Boxhead Field . .. .. . Mul tipage t ables, titles . Multiple amount -scale chart . Multiple-log amount -s·,ale chart ·(s~e· ais~ Loga~ith;.i~ ~h~rts j 49 Multiple slope-curve ''hart Multiple step-curve chart Multiple time-scal e chart . N "Narrative" char t style .. .. . .... . Notes , explanatory, contents and coverage . Notes , tabular (see also Footnotes in text, and Notes i n charts) : Body notes 11 Continued11 tables Headnotes Kinds of Location Footnote~ General . Reference symbols Sequence. Special . Sour ce Notes in charts (s~e -ais~ Notes General notes Reference notes Source notes Numbers : Line Numbers , page Numbers , table, in titles Numbers in text: Othe r rules Rounding Simplifying What to show in numerals What to show in ••ords . Numerals , Roman , for tables Page :-Tu9 even right margin) . .. 9-10 8 . 100 . 93 ' 98 ' 100-101 . . 64 . . . . 50 .. 113-115 42 -43 , 114 . .. 114 . . . . . . . . . 111 , 112 lll, 112 , 117 , 118 , 119 , 120 . 28 . 100 . 29 . 130-133 . . 102-103 102 , 131 , 132 . 102-103 . 130-132 ... 103 108, 109 28 , 31 . . . . . 8 102 , 104-107 . . 38 . 62-63 63 8 . 34 . 12-13 . 35-37 8-9 49 , 75 . 44, 64 . 42, 120 31 32 22 48 47 47 48 . 76 , 77 108 Footnote in tables 35 , 36 36, 37 34, 35 , 36 34 32 , 35 , 36 , 37 tabular and Footnotes 35 , 36, 37 35 , 99 35-36 36 . . 35 .. 35 in text) : 99 99 99 28 , 31 111 , 112 , 119 , 121 . 22 6 -7 7 7 6 6 22 INDEX 0 One hundred percent bar chart 55 One hundred percent surface chart 51 Order of i terns in : Bar charts 54 Stubs 22-28 Outline report 117 p Page : Broadside 13 , 114-115 Elements llO Fold-in 15 , 42 , 115 Format llO Number lll' 112' 113 Reduction 42 , lll, 112, 118, 119 , 120 Size 109-110, lll, 112 , 119 Specifications lll, 112 Paints . . . . . 133 Paired bar chart 55 Paragraphs : Carryover to next column 8 Indention 8 Numbering 8 Rules for 8 P~ns, lettering 102, 104-107 Percentages . . . . 21 Percent sign . . 34 Perspective charts . 63 -64 Perspective plus 3-dimensional treatment in charts . 64 Photo offset 120 Pictographs . . . . 64 "Pictorial" chart style 76-77 Pictorial symbols 64 Pie charts .. 62 Plotting charts .. 82 Pointers , label 93 , 98 , 100-101 Pointing 134 Poor presentation, typical examples of .. . .. 65 "Poster" chart style ... 76 , 77 Posters , big original and blowups 127 ' 130-134 Pr echarting technique ..120 Prepared materials 42, 98 Preprinted forms 42-43 Preprinted parts in chart constructi on .. 98 Printed publications, char t style for 76 Preface 108·, io9," llO Presentation': Equipment .. 134 Good quality l Major problems 2-4 Oral, hints . . . . . . . . 134 Procedures (eight basic steps) 2-4 Three forms (text, tables and ch~rts) 4 Production hints : Oral presentations 134 Poster displays • . 130 Projection displays 127 , 128-129 Schedule ll7 Techniques 117 Program, objective , performance, plan, requir ement 58, 59, 92 , 93 Progress chart 61 , 92 , 93 Progressive or step-by-step bar charts 57 Projection: Methods 127 Qpaque • 127 ' 129 Overhead 127 , 128-129 Slide .. . . 127 128 Transparent 127 , 128-129 Projections and goals, how to indicate 92-93 Projectors .127 , 128, 129 Proofreaders' marks 123 , 124 Proofreading 122-125 Punctuation : Boxhead 29 Footnotes . . a Importance of 5 Rules for 5, 6 Stub . 28 Q Quality control . ...... . . ............ . 59 R Range -bar chart . . . . . . . . . . ..... ... . 57 Range chart. (See Band chart, Change-bar chart, Column , slope curve and Range chart . ) Range-column chart 53 Rank chart 61 Rates . . . . . . 21 Ratios . . . . . . .. 21 Reduction, page : ll9, 120 Page Reduction , page -Continued Page numbers . ll3 Runnin5 heads ll3 Reference : Materi~l . 108 Notes . 32 , 99 Numbers and letters 32 , 35-36 Symbols 7 -8, 35-36 Tables ... 12 Repeated time scale charts ... 48 Report checki ng . . 11, 121 , 125 Reproduction, two-side and one -side no Review and analysis, detailed, in a large report 108 Rounding figures 7 Rulings , for charts .. 84-87 Rulings , tabular: Blue ~encil 38 Helping t he draftsman 38 Other ruli ng considerations 38 8 Text ..... .. . Wei ght of ru.Lings 38 Where to put rulings 38 Runarour.d , text . . ll4 Running head . . . . . llO, 111 , 112 , 113 , 118, ll9 s Scale r ulings . . . . . 84.$7 Scales : Amount-scale capo1ons 82 ft~ount-scale intervals 81 Amount-scale range . . 80 Break~ng the amount scale 50 , 65 , 80 , 81 Breaking curves 81 For bar charts . . . . 88-89 For f~equency charts . 83 For kme-series charts 80-87 Location of amount scales 82 , 84-85, 88 Location of time scales 83 ' 84 ' 86-87 Time -scale captions 83 , 86 , 87 Time -scale divisions 82 , 83 Time-scale range 82 Uniform scales 81 Screen: Distance 129 , 131 Image 129 Size .. 129 , 131 Securit;: Labels 110, lll, 112, 113 Page · · 108 Shading , selection and use of 94, 95 , 96 , 132-133 Signs ~~d symbols: . 6 As t erisk . 32 Footn:>tes 32 Minus 32 Plus .. . 32 Reference symbols 35-37 Simple bar chart . . 54 Simple column chart 52 Simple curve chart . 46 Simnle frPquency chart 60 Simple slope-curve chart 46 Simple slope-surface chart 50 Simple step-curve chart 46 Simple step-surface chart . 50 Simpli£ying £igures . . . . . 7 Size of report . . 109-llO, lll , 112 , 119 Sketches . . . · · · 70 , ll7 Sliding bar chart 56 Slope ~~d step-curve chart 58 Slope-curve charts . 46 , 47 Slope-surface charts 50 Source notes . . 35 , 99 Spacing: Between report elements ll4 Boxhead.... 29 Paragraphs . . . . . . . . 8 Pr oblems in tables . . . 39-42 Single or double column .. 8 Spacing bars (or co1umns) 90, 91 Spacing letters, words, and lines on charts . 103 , 105 -107 Stub . . . ...•. · · · · · 24 , 26 , 27 .. . . Table width and length . . . . 39-42 WordE and letters for display . . . .. 130 Spanners , field ....... . 14, 17, 18, 19, 29 , 31, 33 Special purpose tables . . . . . . . . . . 12 Specification, chart construction 105 , 106 , 107 Split-scale chart 59, 84, 87 Standard sizes of charts . . 79 "Statistical" chart style 76-77 Status char t . . . . . . . 61 Status-change-status charts . 74 Step-b;r-step bar chart 57 Step-ct:.rVe charts . . 46 Step-s~face charts . . 50, 51 INDEX Page Strata chart . . ---so Stub: Arr angement of items in.spacing 22 , 24, 25 , 26, 27 Capital letters . 24 Double designing . . 22 , 24 Heading . . 22 I ndention . 24 Leader s 28 Lengtheni-ng" 26 Line numbers 28 Major i terns 24 Organizing . 22 , 23 Punctuation . . . 28 Shortening . . . . . . 27 Spacing, normal . 24, 26, 27 Spacing, vertical . . . 24 Subordinate i terns . . . 24 Totals, placement . 24 , 28 Underlines . . . . . . 28 Wording ...... .. . .. 28 Style (see also Char t style) 76-77 Styl e , mechanics of: Abbreviations . 6 Numbers in text 6 Punctuation 5 References . 5 Word Division" : . . 5, 10 Subdivided bar chart . 54 Subdivided column chart 52 Subdi vi ded slope-surface chart . 50 Subdivided step-sur face chart . . . . 51 Subdivided surface charts (see also Surface charts ) : Arrangement of "ir regular" components . 73 Imprope r use of . . . . . . . . . . 73 Subdivided surface and curve chart . . 58 Subtables . 18 , 19 Subtitles, chart . . 99 Summar y : Section of a repor t 108 Tables . . . 12 Text . . . . . . . . . 5 Supplementar y amount-scale chart 49 Sur face char ts : Arranging i r regul ar layers in subdivided surface char ts 73 Band char t 51 , 73 ' 105 Broken scales never used 50 Deviation-cur ve char t . . 51 Differ f r om an area chart 50 Grid proportions . . . . . . . . 7B Improper use of st;bdivided :.;urretce c :Jo.1· t s 73 Lettering sizes to use ... . . . 102, 105 Line weight standards . . . . . . . 102, 104 One hund r ed percent surface chart . 51 Simple s l ope-surface chart . . 50 Simple step-surface chart . 50 Si mple surface char t . . 50 Spacing between lines . . 103 Specification sheet ... 105 Subdivided slope-surface chart . . 50 Subdivided step-surface chart . 51 Surface a nd curve char t . 58 Symbols , gr aphic . 71 Symbols, pictorial . . . . 64 Symbols , "square dial" . . 61 Symbols , tabular .... 32 , 34-36 T Table : Boxhead . 28-31 Br oadside .. 13 , 114-115 Components of 19-22 Double-page spread .. 15 Field ..... 31-34 Formal . 12 General purpose . . 12 Informal . . .. 29 Multipage . . . . 15 , 22 Number s in titles .. 22 Odd shape . 40-41 Oversized . . 40 Pr oduction .. 11 Position on page 11 Single page . 15 Special purpose . . 12 Structure . . ... 15-19 Stub (see also Stub, table) 22 -28 Tabular notes . 34 Tabular rulings . . . .. 38 Text .... .... . .. 12 Title (see also Title, table) 20-22 Width and length . . . 13 ' 15 ' 39-42 Tables , types of: Single stub and single boxhead . . 15 , 16 Divided stub and repeated boxhead . 15 , 16 Tables , types of -Continued Repeated stub and C.ivided boxhead . 17, 19 Repeated stub and single boxhead • 17 , 19 Repeated stub and repeated boxhead 18, 19 Subtables ............ . 19 Wi th field spanners .. • ..... 17 ' 18 Tables , advantages and di sadvantages of . 4, 11 Tabl es , ove r sized 40-42 , 120 Tabular: Notes . 34 Rulings .... . 38 Tapes , adhesive . . 134 Tapes , double adhesive . 131 Terms used : In graphic presentation . . 45 ' 105-107 In tabular presentation . . 14 Text : Advantages and disadvantages of 4, 5 Arrangement ....... . 7, 114 Purposes .. .. ..... . . .. 5 Tabulations of tex~ tables . . . . 12 Three-dimensional charts . . . . . . . . 63, 64, 76 , 77 Three-dimensional plus perspective treatment, in charts . . . 64 Time scale : Captions (see al so Captions) 82-83 , 86-87 Divisions . . .. . ..... . 82 82 Range ...... .. . . Standard location of . . . . . . . . . . . . . . . . . . 83 Time-scal e de signations and rulings (see al so Scal e r ulings,1 : Annual dat a 87 Cal endar year data 86 Fiscal year data . 86 Quarterly data .. 87 Spl it-scale charts 87 Title : Page ... . 108 , 109 Repor t .. . 109 Title , t able : Capitalization 21 Double-page spread 15 Dates in titles 20 Emphasis in titles 20 "15, 22 Multipage tabl es . Percentages , rates and ratios in titles 21 Pl acement and capitalization of titles 21 Punctuation i n titles 21 Single page .. .... . . . 15 20 Subtitles . .. .. . . . . 22 Table number s in titles Titles for multipage tables . 22 Wording of titles . 20 Titles , chart: Abbreviations in 99 o/) Length ... Letter coQ~t per l i ne 102 o/) Location . . Totals : Computer printed 43 In stub 24 , 28 Location .... 25-28, 30 Transpar enc i es , projection 127, 128-129 Trim size . . . . . . . . . . . . . . . . . . . . 111 , 112 Two-way progressive bar chart (see also Progressive or step-by step bar chart) . . ..... . . . 57 Type (see also Chart , lettering) : Face and s~ze specification 110-113 Faces, display 130 Faces , tex-: 110, 113 Page . . · · · · · 110, 111, 112 , 119 Sizes for display ..... 131 Typing: Area . 110, 111, 112 , 119 Layout sheet 111, 112 , 117, 118, 119, 120 u Underscore a~d underline (see also Ruling) : In stub · . 24, 25-28 In text . . . . . . . . . · .. .. ..... 8 w Width : Broadside pages . . . . . . . . . . . . . . . . l3, 114, 115 Double-page spread . . . . . . . . . . . . . . 13 , 15, 22 , 114 Fold-in pages . . . . . 15, 42 , 115 Half-page and full -page tables 114 Narrow pages . 114 Word divisio~ . . .... 5, 10 z Zero 32, 34 Zero line , omission of 54, 80 141 By Order of the Secretary of the Army: HAROLD K. JOHNSON, General, United States Army, Official: Chief of Staff. J. C. LAMBERT, Major General, United States Army, The Adjutant General. Distribution: Active Army: To be distributed in accordance with DA Form 12-9 requirements for Administration E. NG: None. USAR: None. U. S. GOVERNMENT PRINTING OFFICE : 1966 0 -214-853