lib-MOCS-KMC364-20131012113204 194 Journal of Library Automation Vol. 14/3 September 1981 today's large academic libraries struggle, there is, nonetheless, room for criticism of library priorities. This study must be viewed as only a first step (largely tentative and exploratory) in relating automation with service attitudes. It suggests that online systems may be asso- ciated with managers more positive in their view of the management role and more pos- itive in their attitudes toward users than batch- and manual-system managers. Fur- ther research would be useful at this point to compare levels of automation (manual, batch, and online) with circulation-staff service attitudes or those of patrons using the systems. REFERENCES l. Laurence Miller, "Changing Patterns of Cir- culation Services in University Libraries" (Ph.D. dissertation, Florida State University, 1971), p.iii. 2. Ibid., p.149. 3. Robert Oram, "Circulation," in Allen Kent and Harold Lancour, eds., Encyclopedia of Library and Information Science, V.S (New York: Marcel Dekker, 1971), p.l. 4. William H. Scholz, "Computer-Based Circu- lation Systems- A Current Review and Eval- uation," Library Technolo gy Reports 13:237 (May 1977). 5. Robert Oram , " Circulation," p.2. 6. James Robert Martin , "Automation and the Service Environment of the Circulation Man- ager" (Ph.D. dissertation, Florida State Uni- versity, 1980), p.22. Statistics on Headings in the MARC File Sally H. McCALLUM and James L. GODWIN: Network Development Office, Library of Congress, Washington, D.C. In designing an automated system, it is important to understand the characteristics of the data that will reside in the system. Work is under way in the Network Devel- opment Office of the Library of Congress (LC) that focuses on the design require- ments of a nationwide authority file. In support of this work, statistics relating to headings that appear on the bibliographic records in the LC MARC II files were gath- ered. These statistics provide information on characteristics of headings and on the expected sizes and growth rates of various subsets of authority files. This information will assist in making decisions concerning the contents of authority files for different types of headings and the frequency of up- date required for the various file subsets. TheN ational Commission on Libraries and Information Science supported this work. Use of these statistics to assist in system design is largely system-dependent; how- ever, some general implications are given in the last section of this paper. In general , counts were made of the number of biblio- graphic records, headings that appear in those records, and distinct headings that ap- pear on the records. The statistics were bro- ken down by year, by type of heading, and by file. In this paper, distinct headings are those left in a file after removal of duplicates. Dis- tinctness will not be used to imply that a heading appears only once in a source bib- liographic file, although distinct headings may in fact have only a single occurrence. Thus, a file of records containing the dis- tinct headings from a set of bibliographic records is equivalent in size to a MARC au- thority file of the headings in those biblio- graphic records. METHODOLOGY These statistics were derived from four MARC II bibliographic record files main- tained internally at LC: books, serials, maps, and films. The files contain updated versions of all MARC records that have been distributed by LC on the books, se- rials, maps, and films tape:; frum 1969 through October 1979, and a few records that were then in the process of distribu- tion. The files do not contain CIP records. A total of l ,336,182 bibliographic records were processed, including 1,134,069 from the books file, 90,174 from the serials file, 60,758 from the maps file, and 51,176 from the films file. A file of special records, called access point (AP) records, was created that con- tains one record for the contents of each oc- currence of the following fields in the bib- liographic records: Type of Heading personal name corporate name conference name topical subject geographic subject uniform title heading Fields 100,700,400,800,600 110,710,410,810,610 111,711,411,811,611 130, 730, 650 651 830,630 Only the 6XX subject fields that contained LC subject headings (i.e., second indicator = 0) were selected asAP records. The main entry data string was substituted for the pronoun in the series (4XX) fields that con- tained pronouns. The AP records also con- tained information from the bibliographic records that assisted in making the counts, such as the date of entry of the record on the file, the identity of the type of bibliographic file, and the language of the bibliographic record. A third file was derived from the AP file that contained a normalized character string for each AP record heading. These normalized AP records were used to pro- duce the counts of distinct headings by clus- tering like data strings. Normalization in- cluded conversion of all characters to uppercase, and masking of diacritics, marks of punctuation, and other characters that do not determine the distinctness of a head- ing, but would interfere with machin~ de- termination of uniqueness. The subhelds included in the normalized string, hence used for all heading comparisons, are given below. Only use-dependent subfields, such as the relator subfield, and those that be- longed to title clusters in author/title head- ings were excluded. Examples of the AP file field contents and the normalized forms are: AP field contents: Chuang-tzu Chuang-Tzu [Blaeu,Joan] 1596-1673 Blaeu, Joan. 1596-1673 Blaeu,Joan, 1596-1673 Byron, George Gordon Noel Byron, Baron, 1788-1824 Byron, George Gordon Noel Byron, baron, 1788-1824 Byron, George Gordon Noel Byron, baron, 1788-1824 Byron, George Gordon Noel Byron, Baron, 1788.1824 Communications 195 normalized forms: CHUANG TZU BLAEU jOAN 1596 1673 BYRON GEORGE GORDON NOEL BYRON BARON 17881824 Distinct headings for this study were de- termined by comparing on the following subfields: Type of Heading personal name corporate name Sub fields a, b,c,d a, b, k, f, p, s, g conference name a, q, e topical subject a, b, x, y, z geographic subject a, b, x, y, z All occurrences of repeating subfields were included. The relator data of subfields were dropped from personal and corporate name headings as were the title subfields in author/title headings. A separate study will examine the occurrence of author/title headings. Approximately 8 percent of the name headings in the files carry title sub- fields: 6 percent are series and 2 percent are author/title subjects or added entries. Two types of distinct heading counts were generated for topical and geographic subject headings. One takes account only of main terms, the a and b subfields, excluding all subject subdivisions. The other com- pared the complete heading strings, includ- ing subject subdivisions. CHARACTERISTICS OF THE FILES The four bibliographic files from which the statistics were derived were begun in different years and are of unequal size. Ta- ble 1 presents the number of bibliographic records added to each of the MARC files by the year that the record was first entered into the file. The records added in the first months of 1979 have been eliminated from tables 1-3, thus the total number of records under consideration is 1,210,809. In the combined file, the records for books domi- nate the contributions from other forms of materials, representing 85 percent of the combined file records. After the addition of the films and serials records in 1972 and 1973 the total number of records added each year leveled off to around 115,000 but jumped to an average of slightly more than 150,000 records per year following the ad- 196 Journal of Library Automation Vol. 14/3 September 1981 Table 1. Number of Records Added to Each File by Year Year Entered Book Serial Map Film Total 1968 11,812 0 0 0 11,812 1969 43,874 0 1, 104 0 44,978 1970 86,004 0 3,467 0 89,978 1971 105,390 0 8,857 6,280 114 ,247 1972 73,437 0 4,665 6,280 84,382 1973 92,512 3,720 5,566 8,929 110,727 1974 99,004 10,682 6,246 8,457 124,389 1975 86,527 15,866 6,721 8,604 117,718 1976 120,106 19,098 6,876 5,432 151,512 1977 140,011 17,999 7,011 4 ,797 169,818 1978 169,044 12,643 5,584 4,464 191,735 Total 1,027,721 80,008 56,117 46,963 1,210,809 Table 2. Numbers of Headings and Distinct Name Headings Added to All Files by Year Number of Headin gs Number of Distinct Headin gs Year Personal Corporate Conference Personal Corporate Conference Entered Names Names Names Names Names. Names 1968 14,526 3,138 155 12,620 2,139 143 1969 53, 134 21,206 1,027 39,184 9,364 909 1970 104,365 42 ,798 2,175 63,037 14,286 1,769 1971 129,617 57,496 2,742 64,029 15,216 2,158 1972 91,040 45,768 1,942 41,246 9,891 1,402 1973 118,188 57,847 2,625 48,703 12,653 1,862 1974 127,588 73,303 2,972 51,623 17,129 1,983 1975 113,622 76,417 2,519 50,291 18,135 1,742 1;}76 154 ,7 18 88,207 3,454 73,182 23,120 2,306 1977 182,860 87,985 3,487 89,353 23,906 2,333 1978 218,535 97,042 4,192 99,780 24,280 2,831 Total 1,308, 193 651,207 27,290 633,048 170, 119 19,438 Table 3. Numbers of Subject Headings and Distinct Subject Headings Added to All Files by Year Number of Distinct Headings Number of Headings First Terms Only Full Headings Year Topical Geographic Topical Geographic Topical Geographic Entered Subjects Subjects Subjects Subjects Subjects Subjects 1968 10,615 1,857 4,390 489 7,775 1,512 1969 45,161 9,047 8,104 1,980 23,617 5,426 1970 89,304 21,054 8,170 4,263 34 ,526 10,179 1971 115,220 31,278 6,853 5,417 36,689 12,862 1972 92,247 20,760 4,236 2,597 26,201 7,074 1973 121 , 161 27,890 4,460 3,105 33,061 9,819 1974 137,843 31,814 4,524 3,553 39,262 11 ,4 13 1975 130,980 30,650 4,203 3,417 40,129 11 ,818 1976 168,840 39,886 5, 125 4,142 55,468 15,472 1977 185,331 44,973 5,718 4,194 59,529 16,676 1978 222,565 49,923 7,151 4,034 69,856 17,855 T otal 1,319,267 309,132 62,934 37,191 426, 113 120, 106 clition of major non-English roman alpha- bet language records in 1976. The increase is noticeable primarily in the books and se- rials files since the maps file had been add- ing those languages since 1969 and only a limited number of non-English-language audiovisual materials are cataloged. The unusually large number of records added to the books file in 1971 resulted from a special project to add retrospective titles to the file. The large increase in books records in 1978 was due to the CO MARC project in which retrospective LC records that had been con- verted to machine-readable form by other libraries were contributed to the LC MARC file. Approximately 12,000 COMARC rec- ords were added in 1977 and 28,000 in 1978. The fall in numbers of film records produced in 1976-1978 reflects a general fall in production of instructional films in the United States. Counts of items cataloged that are com- piled by LC processing services from cata- logers' statistics sheets show that LC cata- loged approximately 225,000 titles in 1978; thus, approximately 73 percent of LC cata- loging is currently going into machine- readable form. The principal exclusions are records for most nonroman material (only nonroman records for maps have been transliterated and added since 1969) and a few records for music, sound recordings, in- cunabula, and microforms. The portion be- ing put into machine-readable form should rise significantly as the romanized records for items in several nonroman alphabets are added in the next year. NAME HEADINGS Table 2 presents the number of occur- rences of name headings in the MARC bib- liographic files and the number of distinct name headings, both by type of heading and by year. The number of distinct head- ings that were new to the file in a year was determined by comparing the headings added in a given year against those added in all previous years. It is not surprising to find that 66 percent of name-heading occur- rences are personal names, 33 percent are corporate, and only 1.4 percent are confer- ence. The figures shift when considering the distinct names, where 77 percent are per- Communications 197 sonal and only 21 percent are corporate. Looking at ~he total figures in table 2, while 1 ,308,193 of the headings that ap- peared on the records were personal names, only 633,048 or 48 percent of these were distinct. Of the rest, 52 percent were dupli- cates of the distinct headings. Similarly, 26 percent of corporate names were distinct, with 74 percent being duplicates; and 71 percent of conference names were distinct, with only 29 percent being duplicates. In 1968, 87 percent and 68 percent of personal and corporate names, respec- tively, were distinct, i.e., 13 percent and 32 percent "had been used previously" when they appeared on a bibliographic record during the year. As the base file of names grows, the percentage of names appearing on new records but which "had been used previously" rises, to 60 percent and 77 per- cent in 1974. While the figures reported in table 2 indicate that the percentage of head- ings used that were repeats fell slightly again in 1977 (51 percent and 73 percent), this is probably due to the influx of new names with the addition of new languages in 1976-77. Additional statistics gathered on English-language items show the per- centage of repeating headings becoming steady after 1974. SUBJECT HEADINGS Statistics concerning distinct topical and geographical subject headings were col- lected for main terms, excluding subdivi- sions, and for full subject heading strings. Table 3 gives the numbers of headings and the numbers of distinct headings of each type found in the MARC file. Looking at the total figures, only 4.8 percent of topical first terms are distinct, the rest are dupli- cates. This indicates an average occurrence of 20.8 times for each first term. Slightly more, 12 percent, of the geographic first terms are distinct. When the full headings with topical, pe- riod, form, and geographic subdivisions are considered, the percentage of headings that are distinct rises to 32.3 percent for topical subjects and 38.8 percent for geographic subjects. Thus, 67.3 percent of topical and 61.2 percent of geographic are duplicates of existing headings. In the yearly figures, sub- 198 Journal of Library Automation Vol. 14/3 September 1981 ject headings show the same tendency as name headings in that the percentages of headings that appear on new records but which "had been previously used" rises as the stock of headings increases and then levels off. Subjects were also affected by the addition of other roman alphabet lan- guages in 1976-77 but not to a very large degree. For all access points, name headings and full string subject headings, name headings account for 55 percent of the headings that occur in the bibliographic records, with only 45 percent attributable to topical and geographical headings. It should be noted that 12 percent of the name headings that appear on the bibliographic records are names used as subjects. FREQUENCIES OF OCCURRENCE Counts were also made of the frequency with which name headings occurred in the bibliographic files. Table 4 summarizes the frequency data: 66 percent of distinct per- sonal names, 62 percent of distinct corpo- rate names, and 84 percent of distinct con- ference names occur only once in the files. The percent of corporate names with single occurrences is surprisingly close to that for personal; however, the percent of names having multiple occurrences falls more slowly for corporate than for personal names. While 5.47 percent of corporate names occur ten or more times, only 1.92 percent of personal names occur ten or more times. The figures for personal names roughly correspond to those obtained by William Potter from a sample taken from the main catalog at the University of Illinois at Urbana-Champaign. That study showed 63.5 percent of personal names occurred onlyonce. 1 The number of occurrences of different types of headings are compared in figure 1. The bars show the numbers of personal, corporate, conference, topical, and geo- graphic headings that appear in the biblio- graphic files. The shaded areas represent the number of headings that are distinct, thus the upper part of each bar represents additional occurrences of the headings from the shaded area. For personal, corporate, and conference headings a further distinc- tion is made between distinct headings that occur only once, the crosshatched area, and those that have multiple occurrences. Thus the multiple occurrences of corporate names may be seen to come from a small Table 4. Frequency of Occurrence of Name Headings in All Files Distinct Distinct Distinct Number of Personal Names Corporate Names Conference Names Occurrences Number Percent Number Percent Number Percent 1 456,328 65.65 116,250 62.02 18,02 1 83.90 2 119,68 1 17.22 30,185 16.10 2,049 9.54 3 46,247 6.65 11,563 6.17 587 2.73 4 23,951 3.45 6,814 3.64 289 1.35 5 13,820 1.99 4,109 2.19 163 .76 6 8,790 1.26 2,958 1.58 98 .46 7 5,827 .84 2,175 1.16 56 .26 8 4,056 .58 1,673 .89 48 .22 9 2,998 .43 1,395 .74 36 . 17 10 2,153 .31 10 ,037 .55 18 .08 11-13 4,116 .59 2,180 1.16 44 .20 14-20 3,748 .54 2,632 1.40 41 .19 21- 50 2,678 .39 2,901 1.55 23 .11 51-100 448 .06 936 .50 4 .02 101-200 149 .02 374 .20 2 .01 201-300 47 .01 109 .06 1 .00 301- 400 19 .00 46 .02 0 .00 401-500 11 .00 21 .01 0 .00 501- 1000 5 .00 53 .03 0 .00 1001 + 2 .00 18 .01 0 .00 Total 695,074 99.99 187,429 99.98 21,480 100.00 number of distinct corporate headings, as was indicated by the slow decrease of the multiple-heading occurrence rate (i.e., a small group of corporate names have a very large number of occurrences). FILE GROWTH As a bibliographic file grows and the stock of names and subjects that are con- tained in the associated authority file in- creases, the number of new-to-the-file 1400 1200 1000 "' <:> 800 z i5 : .. 0 a: w 600 ID ::;: "' z 400 200 1,444,726 Personal Names Corporate Names Communications 199 headings that are required for the new bib- liographic records would be expected to fall. Figure 2 illustrates that tendency and shows that there is a leveling off of the num- ber of new-to-the-file headings per new bibliographic record after the biblio- graphic file reaches a certain size. For ex- ample, after approximately 700,000 biblio- graphic records are in the file, for every additional 100 bibliographic records ap- proximately 298 name and subject headings 30,417 Conference Names 1.468,804 Topical Subjects Geographic Subjects D DISTINCT HEADINGS - DISTINCT HEADINGS THAT OCCUR -ONLY ONCE Fig. 1. Number of Headin gs by Type. 200 Journal of Library Automation Vol. 14/3 September 1981 will be assigned, and, of these, approxi- mately 53 will be new personal names, 14 new corporate names, 2 new conference names, 35 new topical subjects, and 10 new geographic subjects; the remaining 184 headings used will already be established in the authority file. Thus after a certain bib- liographic file size is reached, the growth of the authority file is approximately a linear function of the growth of the bibliographic file. IMPLICATIONS The reoccurrence frequency of headings in a bibliographic file is often cited as a fac- tor in designing bibliographic and authority-file configurations. Discussion 1.2 II I 0 0 .9 a: 0 u w a: .8 ~ ~ .7 z 5 :.\ " .6 ~ z 0 .5 a: w "' ~ . 4 z .3 centers on the necessity of carrying author- ity records for headings that occur only once in a bibliographic file . With reference to the name-heading data in table 4 and figure 1, carrying authority records only for headings that occur more than once could 'potentially reduce the size of the authority file from that indicated by the whole shaded area (including shaded and cross- hatched) to the plain shaded area, i.e., from 903,983 records to 310,123, a 66 percent decrease. Controlling multiple occurrences of a heading is, however, only one role of the authority record. More important perhaps is the control of cross-references connected with the heading. Preliminary work with a • PERSONA L NAMES ---- 9 TOP~CAL SU8JECTS ... CORPORATE NAMES 2~ ----------~----~---------- & GEOGRAPHI CA L SUBJ ECTS 'Y CON FERENCE N AMES » ~~~~r=~~~~~==~==~==~~~==~==~==~==~-- 100 200 300 400 500 600 700 800 900 1000 11 00 1200 1300 NUMBER OF BIBLIOGRAPHIC RECORDS Cthousands) Fig. 2 . N umber of N ew Headings p er R eco rd for All Files. random sample of personal names in the LC file indicates that less than 17 percent of personal names require cross-references. Thus the personal name headings that occur only once but would require authority rec- ords because of cross-references could be less than 17 percent. The frequency data combined with reference structure data could have a significant impact on design. Out of a total of 695,074 personal names in the authority files associated with the MARC bibliographic files examined here, 456, 328, or 66 percent, occur only once. Of these, fewer than 77,575 would be expected to have cross-references, thus the name- authority file for personal names could be reduced in size from 695,074 records to 316,321, a 55 percent decrease. If separate authority records are a system requirement, the occurrence figures might then be useful for defining configurations that employ machine-generated provisional records for single-occurrence headings that do not have reference structures or that simplify in other ways the treatment of these headings. These figures may also be useful in making deci- sions on the addition of retrospective au- thority records to the automated files. REFERENCE 1. William Gray Potter, "When Names Collide: Conflict in the Catalog and AACR2," Library Resources & Technical Services 24:7 (Winter 1980). RLIN and OCLC as Reference Tools Douglas JONES: University of Arizona, Tuc- son. The Central Reference Department (so- cial science, humanities, and fine arts) and the Science-Engineering Reference Depart- ment at the University of Arizona Library are currently evaluating the OCLC and RLIN systems as reference tools, to see if their use can significantly improve the ef- fectiveness and efficiency of providing ref- erence service. A significant number of the questions received by our librarians, and presumably by librarians elsewhere, in- Communications 201 volve incomplete or inaccurately cited ref- erences to monographs, conference pro- ceedings, government documents, techni- cal reports, and monographic serials. If by using a bibliographic utility a librarian can identify or verify an item not found in printed sources, then effectiveness has been improved. Once a complete and accurate description of the item is found, it is a rela- tively simple task to determine whether or not the library has the item, and if not, to request it through interlibrary loan. Additionally, if the efficiency of the li- brarian can be improved by reducing the amount of time required to verify or iden- tify a requested item, then the patron, the library, and, in our case, the taxpayer, have been better served. The promise of near- immediate response from a computer via an online interactive terminal system is clearly beguiling when compared to the relatively time-consuming searching required with printed sources, which frequently provide only a limited number of access points and often become available weeks, months, or even years after the items they list. We realize, of course, that the promise of instantaneous electronic information re- trieval is limited by a va):'iety of factors, and presently we view access to RLIN and OCLC as potentially powerful adjuncts to- not replacements for-printed refer- ence sources. Given that RLIN and OCLC have databases and software geared to known-item searches for catalog card pro- duction, our evaluation attempts to docu- ment their usefulness in reference service. A preliminary study conducted during the spring semester of 1980-81 indicated that approximately 50 percent of the ques- tionable citations requiring further biblio- graphic verification could be identified on OCLC or RLIN. The time required was typically five minutes or less. Successful verification using printed indexes to iden- tify the same items ranged from 20 percent in the Central Reference Department to 50 percent in Science-Engineering. Time re- quired per item averaged approximately fifteen minutes. Based on our findings, we plan a revised and more thorough test during the fall se- mester of 1981-82, which will include an assessment of the enhancements to the