College and Research Libraries uniquely burdened is therefore simply with- out substance. In regard to describing which libraries fine and which do not, the most useful variable investigated was the nature of institutional support. State-supported institutions clearly tend toward punitive measures, while pri- vately supported institutions do not. STEPHEN TONEY Research Notes I 313 REFERENCES 1. ARL Library Index, 1978-79 (Report of the Task Force on ARL Membership Criteria, May 1980). . 2. United States Bureau of the Census, Statistical Abstract of the United States (Washington, D .C . : Govt. Print . Off., 1979), inside front cover. A Cost Database for Branch Library Resource Allocation and Performance Evaluation A major gap in the knowledge of the man- agement of the Smithsonian Institution Li- braries (SIL), as in most libraries, is in the precise allocation of expenditures in terms of library goals. Traditional accounting systems are primarily concerned with expenditures by fund and by type of item purchased, i.e. , ob- ject class or line item. However, management purposes (by which is meant planning and re- source allocation, as opposed to accounting purposes) are better served by knowing: (1) for what organizational goal an expenditure was made; and (2) what users were benefited by an expenditure. Management Control in Nonprofit Organi- zations by Anthony and Herzlinger offers a brief survey of accounting methods that illus- trates how accounting practices have re- flected the increasing importance being placed on the budget as a planning tool, in addition to the budget's traditional role as a request for funds. 1 In effect, planning at the top level consists of making decisions about how resources should be allocated to fulfill the goals of the organization, and modern budgeting methods result in a document that expresses those de- cisions. A budget resulting from one of these modern methods clearly reflects the priorities of the organization, and thus will probably Stephen Toney is manager for systems and plan- ning, Smithsonian Institution Libraries, Washing- ton, D.C. have resulted from an evaluation of priorities. The executive branch of the federal gov- ernment now practices zero-base budgeting, in which all programs are reevaluated each year in terms of the goals of the organization. However, although the Smithsonian designs its budgets using the zero-base method, the method is not reflected in its accounting sys- tems. That is, the accounting systems used by most recipients of federal funds, including the Smithsonian, have only rudimentary capabilities to assign to expenditures the pur- pose of the expenditure in terms of goals. Ac- counting reports show expenditures by object class, which are of little use in evaluating the degree of success in fulfilling the goals stated in the budget. ANALYSIS OF THE PROBLEM Conversations among the director, the budget officer, and the author established the need to tag each SIL expenditure according to fiscal year, library goal, object class, fund type, fund source, cost center, and benefit center (these items are defined below). The expenditures so tagged could then be manip- ulated to show total costs by any of these clas- sifications. The most essential of these classi- fications to the project's purposes were the library goal and the benefit center (i.e., the branch library). Tagging in this detail also en- abled cross-tabulations; for example, the ex- penditures by any branch for fulfilling any goal could be isolated. 374 I College & Research Libraries • july 1981 Fiscal year was included so that expendi- tures could be compared from year to year. Library goal was coded according to a goal classification established as part of SIL' s routine annual planning. Object classes of a high level of detail were judged not to be of interest. Expenditures were broadly coded for the classes (and sub- classes) of: personnel (direct and contracted); library materials (monographs, serials , stand- ing orders, and other); automated services and computer equipment; other equipment; supplies; training; travel; buildings and space; and other. Fund type showed whether the expendi- ture was from Smithsonian federal funds, Smithsonian trust funds, or non-Smithsonian funds. Fund source showed whether the expendi- ture came from the SIL budget, from the budget of the benefittor, or from "other source." (Some Smithsonian units donate ser- vices, rent, etc., to the branch libraries that serve them; these units are the benefi'ttors of the services of the branch, the branch being the benefit center. A gift to the library system as a whole would be coded "other source.") Cost center is a coded designation for the SIL department or other Smithsonian unit authorizing the expenditure. For example, Cataloging "authorizes" the costs of salaries, OCLC services, etc., for its work. Cataloging J!light also "authorize" the expenditure of non-SIL funds , as when a cataloging contract is paid for by the benefittor. Benefit center shows where services were received. Since the SIL organization reflects that of the Smithsonian, to say that the branch library serving the Museum of American His- tory is the benefit center is to say that the museum itself is the benefittor. By this con- cept, the cost of providing service to each component of the Smithsonian is made clear. The difference between a cost center and a benefit center may be illustrated this way. If the Cataloging Department buys a typewriter or purchases OCLC services or hires em- ployees, those expenses are assigned to Cataloging as the cost center-the unit incur- ring the cost. However, the ultimate bene- ficiaries of the typewriter are the SIL users. Thus, in order to determine resource alloca- tion among the user groups, it is useful to allocate proportionally the cost of the type- writer among them. For practical purposes, benefit centers are · the reader services branches of the SIL and certain independent libraries in the Smithsonian. The cost centers and benefit centers were coded in such a way that breakdowns on sev- eral levels were possible. Thus the tag SILR-MNH-ANTH is attached to costs for the Anthropology subbranch of the Museum of Natural History branch of SIL Reader Ser- vices division. METHODOLOGY Recent rental of a Lanier stand-alone word processor provided the basic method for stor- ing, manipulating, and printing the data. The Lanier can sort records of up to 256 characters on both fixed-length and variable-length fields, can extract records based on certain characteristics, and can print selected records or parts of records. The Lanier also has cer- tain arithmetic functions. For the purposes of the cost database (informally called the costgrid), sorting and column addition were the only special functions used. For simplicity, a fixed-length record format with fixed-length fields was chosen. Table 1 shows the fields in the record , their positions, their lengths, and their purposes. Gathering of.the Data Expenditures of the SIL in fiscal year 1979 (FY79) were gathered from Smithsonian ac- counting reports (for salaries and benefits), from acquisitions accounts (for library mate- rials), and from examination of all SIL pur- chase orders for FY79. From these sources, each SIL expenditure was tagged, but goals were not assigned to personnel costs at this stage . For expenditures represented by purchase orders, assigning the goal was easy-shelving supports the goal of physical care of the col- lections, a binding contract supports the bind- ing goal. We avoid goals like "support re- search," which is based on someone else's goals, preferring those like "provide refer- ence service" or "perform research," which are based on library goals. A single goal con- solidating the whole materials budget, such as "build the library collection," is much easier to work with than trying to determine how many of the materials support reference ser- vice, how many the curriculum, etc. Research Notes I 315 TABLE 1 CosTGRID RECORD FoR~tAT Field Order Field Name Field Length in Characte rs Purpose 1 Fiscal year 2 Goal 2 4 Fiscal vear of expenditure Classification number of library goal that expendi- ture supported 3 Object class 4 Object name 3 13 Type of item or service purchased with expenditure Name of item or service purchased with expendi- ture 5 Fund type and source 2 13 (See text) 6 Cost center Benefit center What SI organization or department within the SIL made the expenditure 7 13 7 2 What S~ org~nization or department within the SIL benefited from the expenditure 8 Cost The amount of expenditure 9 Source of cost A number that refers to a note of where cost in- formation was gathered from and what adjust- ments were made Personnel costs in technical services were also easy to assign to goals. However, the ac- tivities of Reader Services staff, at least at SIL, touch on nearly all of SIL' s goals. For example, the filing of catalog cards in the branches supports the goal of providing cataloging (at least the way we view it). There- fore, branch staff were asked to provide esti- mates of how they spent their time. This gave only a very rough idea of cost by goal, but time constraints prevented the thorough re- cording of activities that was needed. To identify expenditures for library ser- vices by Smithsonian units other than SIL (which had never been determined), a ques- tionnaire was sent to the head of each major unit. Only object class and fund type were asked. The year, fund source, and cost center were evident. The benefit center was always the cost center. Breakdown by goal was not attempted for non-SIL expenditures, since those goals govern only the SIL. Allocation oflndirect Costs In order to provide a true picture of the cost of serving each component of the Smith- sonian, technical services costs had to be ap- portioned to the branch libraries. This was done on the basis of the amount of work done in each technical service unit for each branch, rather than by a general formula. In order to be as accurate as possible, each technical service unit was analyzed to identify its m~jor activities; then the records for each activity were used to determine how much of the work was for each branch library. In some cases, however, allocation of technical ser- vices costs was hampered by lack of knowl- edge of exactly how time is spent. In general, lower-level tasks and automated activities were fairly easy to assign, but planning, for example, was harder. Generalizations were made when needed, small amounts of time were ignored, and analogies were used when possible. Supervisory and administrative time at all levels except the central administration was prorated into the various units. Costs of the central SIL administration could have been allocated to branch libraries. However, no meaningful way of doing so was discovered. Certain figures had to be adjusted. For example, since expenditures for OCLC ser- vices are for both the cataloging goal and the acquisitions goal (preorder searching), the line in the costgrid for OCLC expenditure (added during examination of the purchase orders) was struck and replaced by two lines, with the OCLC cost for each goal reflected in the line for that goal. Notes were kept of the adjustments made, and coded in the costgrid in the "source of cost" field (see table 1). To illustrate, the resulting two lines were: 79 IC2 3 COMPUTER FC SILT-AD SILT-ACQ 10460 7 79 lA 3 COMPUTER FC SILT-AD SILT-BSS 31381 7 Thus, SIL spent $10,460 in FY79 for goal 1C2 (acquisitions), for object class 3, com- puter services, of federal funds, from the cen- tral libraries budget, authorized by the Assis- tant Director for SIL Technical Services and 376 I College & Research Libraries • july 1981 benefiting the Acquisitions department. Likewise $31 ,381 benefited the Bibliographic Support Section in fulfilling goal1A (catalog- ing). (The number 7 at the end is the "source of cost" field, and refers to note 7.) This example illustrates another point- that there are several levels ofbeneficiary. As mentioned before, the costgrid allocates all expenditures except those for central admin- istration to a branch library. But in the exam- ple, it appears that a technical services unit is a benefit center. However, this was only a technique for grouping costs into the BSS unit, as a step preliminary to allocating BSS costs to the branch libraries. Thus it can be seen that the allocation pro- cess may have several cycles, but to skip one cycle results in loss of the audit trail and thus loss of information. Unfortunately, the above technique does result in rather more notes than were desired, in order to explain the two levels of cost allocation. Because of these two levels of allocation, technical services costs appeared twice in the costgrid if an adjustment had not been made. To illustrate: assume that SIL has only three branches, A, B, and C. (These will be coded SILR-A, SILR-B, and SILR-C. The R stands for Reader Services.) Also assume that A gets 50 percent of the effort of the BSS unit, B 30 percent, and C 20 percent. The costgrid then shows, for the OCLC costs discussed above: 79 1C2 3 COMPUTER FC SILT-AD SILT-ACQ 10460 7 79 lA 3 COMPUTER FC SILT-AD SILT-BSS 31381 7 79 1A 3 COMPUTER FC SILT-BSS SILR-A 15691 8 79 1A 3 COMPUTER FC SILT-BSS SILR-B 9414 8 79 1A 3 COMPUTER FC SILT-BSS SILR-C 6276 8 The dollars column now adds up to $73,222 for OCLC services, rather than the actual $41,841 spent. This is a result of the two phases of allocation. To correct this, a line was inserted into the costgrid containing a nega- tive amount equal to the amount that would be doubly entered-in this case -31,381. A final methodological point is that of data integrity. More than 500 lines of data were entered for FY79, and the possibilities of ac- cidental error were great, as with any numer- ical or coded data. Thus two quantities were carefully checked after each update of the grid. First, the number oflines was calculated by programming the Lanier to print a certain number of lines per page. Second, a total for expenditures was kept. Each new listing of the grid was compared against the previous, as adjusted with the new data input. A second technique for checking data in- tegrity, which should be employed for any data-processing project or for any large word-processing project, was to keep a backup diskette with a second copy of the costgrid. Mter each update to the grid was verified, the old backup was erased and a backup of the new master made. The disk- ettes were named COSTGRID-MASTER and COSTGRID-BACKUP, and the versions of the grid were named for the date of their cre- ation, e.g., COST800610. DISCUSSION AND CONCLUSIONS This paper has attempted to show how li- braries can improve their knowledge of re- source allocation. At the Smithsonian Institu- tion Libraries, we now have new knowledge of where the money goes. The figures on de- gree of support given to each branch we are quite confident of; the figures on allocation by library goal are only indicative, not precise. However, any management information project, even one as rudimentary as that de- scribed, is a developmental process. At SIL we now know the general outlines of the terri- tory and can work toward more precise knowledge each year. It is hoped that several improvements to the database will broaden its usefulness. First among ~hese is to develop the output side of the data. Expenditures are one way to quan- tify ~he inputs to the library system (the staff, the materials, etc.). What is needed to discuss library efficiency is a quantification of outputs (reference questions answered, circulations, etc.). If these were added to the database, routine evaluations of efficiency could be made. Of course, efficiency is only part of the ef- fort to measure performance. What is ulti- mately desired is a means of measuring li- brary effectiveness . . Then comparisons be- tween branches or between libraries cari be meaningful. One way to do this in the context of the costgrid is to modify the measures of quantitative branch output with coefficients of effectiveness for each branch activity. The coefficients might come from an annual study of the branch's effectiveness in performing each of its services. This study would need to focus on the needs of users in fulfilling their goals, a task notoriously difficult in the library context. Some ideas for pursuing this have been gathered from Lancaster's Measurement and Evaluation of Library Services, 2 and also from the papers presented at the preconfer- ence on library effectiveness of the 1980 ALA Annual Conference. 3 Two intermediate steps between the purely quantitative efficiency rating and the true effectiveness rating can be imagined. The first might be a quality measure used to adjust the raw quantitative measures of out- put. The supposition made here is that high- quality output is more effective that low- quality (that is, it better fulfills the library's goals). The quality measure has the advantage Research Notes I 377 over the true effectiveness measure in that the users' trt·e needs need not be studied, but only the library's products. The second possible intermediate step in approximating a true effectiveness rating is to employ what has been referred to as the "managerial rating model. "4 This model adds to the rating formula an assessment of the rel- ative importance of each of the library's prod- ucts (obviously based on the importance of the library goals) . The assessment is usually performed by the director; however, to re- flect the diverse user needs in each branch library, a ranking of the' product importance by each branch chief could be used to modify further the scores for his or her branch. It is perhaps easier to see the progression of DISCOVER THE WORLD of academic libraries with College & Research Libraries News News from the Field keeps you informed of ongoing research by academic libraries, as well as recent special collections acquired and grants awarded. People and Profiles gives you back- ground on the leading librarians in the profession. Job Listings will help you find the position that best suits your talents. Mini-reviews of recent professional literature allow you to stay up-to-date. A Calendar of Events will brief you on upcoming workshops, seminars, and conferences. You will receive 12 months (11 issues) of C&RL News for $10. Send orders to: C&RL News Editor, ACRUALA, 50 E. Huron St., Chicago, IL 60611. 378 I College & Research Libraries • july 1981 sophistication using an example. (The exam- ple comes from technical services, which of course can be rated in ways similar to branch libraries). The first, or purely quantitative, step would be a formulaic comparison between outputs and inputs, such as : n Ia ; x; ; = I S, = ~-c-- where S1 = final score; Xi = the quantities of various products completed (titles originally cataloged, copies added, etc.) ; a i = a coefficient that expresses the relative effort required for each product (so that one title cataloged is "worth," for scoring purposes, ten copies added (the actual coefficients must be derived from time studies) ; and c = the cost of the department during the time period under study. The next step would be to have the outputs of the department evaluated for quality in some way; original cataloging, for example, could be evaluated by recataloging a random sample and comparing the results with the original work. The formula to score this method might be: n Ia;x;q; ; ; I Sz =---c-- where qi =the quality scores (on a one-to-ten scale, a percentage scale, or some other ratio scale). The next step employs the managerial rating model, in which each product is ranked for importance, again on a ratio scale. The formula might then be: n Ia; X ; q; r; ; ; I c where r t = the ranking given by the library administration to the relative importance of each product. To reiterate, none of these formulas have addressed the true effectiveness question, namely whether the user has received the service he or she needs. This is a question almost impossible to answer , since users themselves often cannot express , indeed sometimes do not realize , their information needs . 5 Solving the problems of such assess- ment will be the work of years. REFERENCES 1. Robert N . Anthony and Regina E. Herzlinger, Management Control in Nonprofit Organiza- tions {Homewood, Ill.: Irwin, 1975). 2. F . W . Lancaster, The Measurement and Evalu- ation of Library Services (Washington , D.C . : Information Resources Press, 1977). 3. Library Effectiveness: A State of the Art (Chicago: American Library Assn. , Library Ad- ministration and Management Assn . , 1980). 4. For discussion of the history of this idea, see Terry Abraham , "Managerial Rating: A Library Effectiveness Model, " in Library Effectiveness: A State of the Art (Chicago: American Library Assn. , Library Administration and Management Assn., 1980). 5. See Robert S. Taylor, " Question-Negotiation and Information-Seeking in Libraries, " College & Research Libraries 29:178-94 (May 1968). NOW ... The Concise AACR2 MICHAEL GORMAN While the original AACR2, published in 1979, is a formi- dable work-covering as it does all the biblio- graphic situations cata- logers are likely to encounter- its very thoroughness limits its usefulness where the material to be cataloged is less com- plicated or where only the operative principles of the code must be grasped. The latter is the case in teaching, for example. Without disregarding the importance of properly cataloging the more diffi- cult material, catalogers wil1 recognize that most of their questions can be answered by a simpler version of the Rules. The Concise AACR2 was de- signed for that purpose. It emphasizes essential principles, dropping separate treatment by medium and the Jess frequently used rules. In many cases it simplifies the wording of the rules while often giving additional explanation. The Concise AACR2 should be consulted first when problems occur; it fol- lows the rule numbers of AACR2 so that if it cannot provide the answer the complete version may be used. 176 pages Paper ISBN 0-8389-0325-8 (1981) $6.50 Anglo-American Cataloguing Rules, 2nd ed. Cloth ISBN 0-8389-3210-X (1979) $15.00 Anglo-American Cataloguing Rules, 2nd ed. Paper ISBN 0-8389-3211-8 (1979) $12.00 Sample entries shown in typewriter .... type to guide users in the exact preparation of their own original copy. Order Department The Famil Edward S y of man • I C teichen • • created b Ma~/:1/ection of JOJ photo r y Add. ntry under title '8 apfts hy a·ariour ed entry llluler th h . . people) Wh e eadmgfor Steichen Y a duck? • from the M • Visual and edited by Rairx Brothers mo vierbal gems . chard J v es 1 Mom entry Iinder . ' . Anobile Added t1r,e <'tltr v under the lteadin The N g fo r Anoh1Z ew York times e American Library Association so East Huron St. , Chicago, IL 60611 early fills a need?' valuable addition to our fessionalliterature?' clearly written, vital renee to the field?' t's what they've been ng about the The professional review media and librarians and inforrnatio~ specialists at every level have joined in their praise of ALA World £1/t~lfclopcdin, which made its debut in Spring This favorable reception supports the editors' initial rnr•vi.rtirui that there was a need for such a work that, "in one co1nv1>niPnl vo.Jume seeks to explain fundamental ideas, record events and activities, and portray those personalities, living dead, who have shaped the field." The Em·ydopcdin is the result of international collabora spanning four years of planning and editorial work and enga almost 400 advisers and contributors from 145 countries world. It contains 624 pages, 700,000 words, 452 signed 300 illustrations both historical and contemporary, and 144 tical tables. AlphabeticaUy-arranged articles, averaging than a thousand words in length, have been edited for ,.,,..,,..,c,,,", , 1--------------------------------.---- readability, and comparab LA World Encyclopedia Library and Information Services coverage. There are two special feat of the work. A topical of· Contents gives an Wedgeworth, Editor major contribution to rative librarianship a publishing milestone the information field. are a few of many favorable and " ... clearly fills a need for a lne, -vcllume reference work that extensive and information related to libraries and information Services:• -COLLEGE & RESEARCH LIBRARIES " ... a valuable addition to our literature and [one that) undoubtedly be widely and intensively used." - JOURNAL OF LIBRARY HISTORY "I congratulate the editors for produced so truly magnum -ROBERT M. PIERSON, School of a.nd Information Science, Catholic University of America. of the organization of the ume and a fresh view of the ganization of knowledge in field; a precise and in Parallel Index in the margins of the text pages bring index references and related alphabetical articles and ready reference information. The ALA World EHcyclopcdia enhances reference collec serves as an invaluable resource for library school provides comprehensive coverage for practitioners and readers. 624 pages Cloth ISBN 0-8389-0305-3 (1980) $85.00 Order Department " ... a carefully-edited, well-organized, and handsomely presented volume:•- THE JOURNAL OF ACADEMIC LIBRARIANS HIP "The arrangement, featuring cross references alongside the main entries, makes it very handy to use. -ANTIQUARIAN BOOKMAN "What a massive piece of work! Congratulations."-ANN MORGAN CAMPBELL. Executive Director, The Society American ArchiVists " ... a particularly valuable to our professional col because of its auth•orJLtativ·enes~ and international scope. M. RICHARDS, Chief, Library D!Vision,lnternatlonal Commu.nlc,ll.tlc>rj Agency RISS FANG, Professor, School ofLibra.ry Simmons College. AMERICAN LIBRARY ASSOCIATION 50 East Huron Street, Chicago, Illinois 60611