College and Research Libraries JOHN J. KNIGHTLY Library Collections and Academic Curricula: Quantitative Relationships In recent years several groups of academic libraries have surveyed their collections as part of their background studies for network de- velopment. This study of twenty-two ·Texas state-assisted senior col- lege and university library collections analyzes collection ''duplica- tion" or "overlap" with respect to common degree programs. The concept is examined that libraries acquire similar book collections in support of similar academic programs. In all nineteen subjects studied, collection overlap is found to be extensive. However, each library does make a unique contribution to potential statewide net- work development. IT IS AXIOMATIC that academic library collections bear a "close" relationship to curricula; however, the nature of the quantitative relationship is seldom de- scribed in other than formula terms. This article reports on a recent study of li- brary collections of the twenty-two state- assisted colleges and universities of Tex- as which provided data on common un- derlying components of different library collections with respect to common de- gree programs and levels of the degree programs. In recent years, particularly since 1966, several groups of academic libraries have surveyed their collections as part of their background studies for development of academic consortia.1 Desiring to mitigate some of the high costs associated with individualized or labor-intensive practices, these libraries have sought to identify the extent of duplication or overlap in their collec- John ]. Knightly is assistant professor, Graduate School of Library and Informa- tion St:?ience, the University of Tennessee, Knoxville. tions in order to plan for effective co- operative efforts. The study reported in this article, which compares subject col- lections to curricula, adds a useful di- mension to these previous studies of what has been termed "duplication," "multiplication," or "overlap" in the collections of academic libraries. PRoCEDURE Nineteen academic degree programs were chosen for study: astronomy, bi- ology, botany, chemistry, economics, French, geography, German, English, his- tory, management, mathematics, philoso- phy, physics, political science, psychol- ogy, sociology, Spanish, and zoology. Book titles were assigned to correspond- ing degree programs on the basis of Dewey classification numbers; 2 juveniles and paperbacks under one dollar were ex- cluded. The data gathering instrument consisted of stratified random samplings from the U.S. book titles published in a given year in each of the nineteen subject fields as recorded in the Ameri- can Book Publishing Record (BPR) an- / 295 296 I College & Research Libraries • July 1975 TABLE 1 HoLDINGS AND DUPLICATION OF U.S. BooKs BY AcADEMIC LEVEL Number Composite of Average Level Subjects Holdings Doctoral 14 61% Master's 15 38 Bachelor's 16 30 No Degree 9 26 nual cumulative edition. The total book · sampling, consisting of 845 titles, . was checked -against the public catalog of each of the twenty-two state-assisted senior colleges and universities of Tex- as. The data on book holdings were compared to academic degree programs of each school as defined and invento- ried by the state's college and university system coordinating board. As indicated in Table 1, library holdings and dupli- cation between libraries for each sub- ject were determined for: ( 1) schools with no degree program in that particu- lar subject, ( 2) schools with a program terminating at the bachelor's level in that subject, ( 3) schools with a program terminating at the master's level in that subject, and ( 4) schools with a program terminating at the doctoral level in that subject. Schools included in the study were: University of Texas, Austin Texas A & M University Texas Tech University North Texas State University University of Houston University of Texas, El Paso University of Texas, Arlington Lamar University Southwest Texas State University East Texas State University Texas A & I University Stephen F. Austin State University Sam Houston State University Angelo State University Texas Southern University Texas Woman's University Midwestern University · · West Tex~s State -University Range of Composite Average Average · Duplication Duplication 69--86% 76% 44-74 56 34-65 47 15-58 37 Pan American University Tarleton State College Sui Ross State University Prairie View A & M FINDINGS Standard Deviation 5.24 8.28 9.37 11.09 Collectively, the libraries held 783, or 92 percent, of the 845 titles in the total sample. These 783 titles were acquired, collectively, 6,982 times, showing an av- erage rate of acquisition of 8.9 per title. Ninety-one percent of all titles from the list which were acquired by any one library were duplicated one or more times elsewhere among the twenty-two libraries. The library whose holdings of the sample were most duplicated, pro- portionately, in other libraries was li- brary 12, with 64 percent of its total holdings held on the average by each of the other twenty-one libraries. Library 20, the state's largest library and the one with the largest holdings of the sample, was numerically the most duplicated in other libraries but proportionately the least duplicated, with 45.6 percent of its holdings of the sample held on the average (mean) by each of the other twenty-one libraries. Average (mean) duplication in all cases was extensive, ranging from 45.6 to 64 percent (Table 2) : In all nineteen subject areas· there was a .pronounced · tendency for holdings a~~ duplication to rise prog~essively from low levels among those schools with no degree program in the particu- lar. subject, to higher levels among those schools with bachelor's and master's lev-· el programs, and to the highest levels Quantitative Relationships I 297 TABLE 2 . HoLDINGS AND DuPLICATION-TEXAs AcADEMIC LmRARIEs Total sample: 845 Holdings of Sample Average Duplicated by Other Libraries 0 Library Number Percent Number Percent 1 278 . 32.8% 158.6 57.0% 2 342 40.4 182.7 53.4 3 396 46.8 205.9 52.0 4 225 26.6 128.0 56.8 5 472 55.8 231.3 49.0 6 158 18.6 92.9 58.8 7 101 11.9 46.4 46.0 8 279 33.0 151.3 54.2 9 345 40.8 199.4 57.8 10 295 34.9 164.7 55.8 11 137 16.2 72.5 52.9 12 138 16.3 88.4 64.0 13 296 35.0 162.3 54.8 14 538 63.6 247.3 45.9 15 256 30.2 139.6 54.5 16 478 56.5 240.2 50.2 17 236 27.9 132.5 56.1 18 448 53.0 220.6 49.2 19 398 47.1 208.0 52.2 20 556 65.7 253.9 45.6 21 399 47.2 200.9 50.3 22 211 24.9 121.8 57.7 mean 377 37.5 165.8 52.2 median 296 35.0 164.7 53.8 0 Average duplication= Average (mean) of the duplication between all pairs of libraries. among schools with doctoral level pro- grams. First of all, on the lowest level there were nine subjects (Tables 1 and 3) in which some of the twenty-two libraries held titles but did not support degree programs. Average duplication among the libraries at this "no degree" level ranged from 15 to 58 percent among the nine subjects, with composite average duplication of 37 percent. It was possi- ble to gather duplication data at the bachelor's level in sixteen subjects (Ta- bles 1 and 3). Average duplication among the libraries at this bachelor's level ranged from 34 to 65 percent among the sixteen subjects, with a com- posite average duplication of 47 per- cent. The extent of duplication continued to rise at the master's and doctoral lev- els. Tables 1 and 3 show duplication data at the master's level in fifteen sub- jects and at the doctoral level in four- teen subject areas. Average duplication ranged from 44 to 7 4 percent among the subjects at the master's level with a com- posite average duplication of 56 per- cent. At the doctoral level average dupli- cation ranged from 69 to 86 percent in the fourteen subjects, with a composite average duplication of 76 percent. An additional question asked was whether the extensive collections at the doctoral level included most of the titles held by the other libraries. The data indicated that an average of 69 percent of the titles in a subject field could be expect- ed in the largest collection at the doc- toral level; that 23 percent of the titles were in the remaining libraries; and that 8 percent of the sample were not acquired by any of the participants. CoNCLUSIONS The aggregate data on duplication 298 I College & Research Libraries • July 1975 TABLE 3 ExTENT OF HoLDINGs AND ExTENT OF DuPLICATION AccoRDING TO DEGREE PRoGRAM AND LEvEL oF DEGREE PROGRAM Mean Mean Size of Range of Holdings Duplicationt Book Number of Holdings Num- Per- Num- Per- Subject Sample Level Schools Low High her cent0 her cent English 143 Doctoral 5 61-93 79.4 55.5 60.5 76.1 Master's 14 15-70 48.7 34.0 33.3 68.4 Bachelor's 3 35-47 40.6 28.3 21.0 51.7 No degree 0 History 98 Doctoral 4 47-70 53.2 54.2 37.5 70.4 Master's 14 11-53 31.0 31.6 13.6 44.0 Bachelor's 4 10-52 23.1 23.5 8.3 36.0 No degree 0 German 28 Doctoral 1 15-15 15.0 53.5 Master's 4 8-16 13.2 47.1 9.8 74.4 Bachelor's 6 7-11 9.3 33.2 5.9 63.7 No degree 11 3-13 6.8 24.2 3.0 44.3 Spanish 31 Doctoral 1 15-15 15.0 48.3 Master's 11 5-18 11.2 36.1 6.1 55.1 Bachelor's 10 1-13 7.4 23.8 3.2 43.2 No degree 0 French 31 Doctoral 1 20-20 . 20.0 64.5 Master's 6 14-20 16.1 51.9 10.4 65.0 Bachelor's 13 2-18 9.5 30.6 3.8 40.6 No degree 2 7-10 8.5 27.4 3.0 35.2 Philosophy 34 Doctoral 1 26-26 26.0 76.4 Master's 1 21-21 21.0 61.7 Bachelor's 5 12-26 18.2 53.5 11.9 65.3 No degree 15 2-18 10.4 30.5 5.1 28.3 Psychology 33 Doctoral 3 20-24 22.3 67.5 17.1 76.6 Master's 9 10-24 17.3 52.4 11.4 66.4 Bachelor's 6 10-22 14.5 43.9 8.0 55.1 No degree 4 2-6 4.2 12.7 0.6 15.5 Economics 64 Doctoral 3 39-41 40.3 62.9 31.0 76.9 Master's 7 7-42 23.5 36.7 13.6 58.0 Bachelor's 11 7-33 17.9 27.9 8.7 48.6 No degree 1 14-14 14.0 21.8 Sociology 47 Doctoral 1 27- 27 27.0 57.4 Master's 11 7-33 18.7 39.7 10.6 56.9 Bachelor's 6 6-29 17.6 37.4 10.2 58.3 No degree 4 4- 29 14.0 29.7 5.3 38.0 Political Science 53 Doctoral 2 43-43 43.0 81.1 37.0 86.0 Master's 15 9-44 26.1 49.2 15.0 57.4 Bachelor's 5 11-41 20.2 38.1 9.1 45.0 No degree 0 Geography 31 Doctoral 2 16-20 18.0 58.0 14.0 77.7 Master's 2 8-11 9.5 30.6 5.0 52.6 Bachelor's 7 5-12 10.2 32.9 5.8 56.8 No degree 11 4-14 7.8 25.1 3.0 39.1 Astronomy 31 Doctoral 2 15-25 20.0 64.5 14.0 70.0 Master's 0 Bachelor's 1 10-10 10.0 32.2 No degree 19 3-15 8.2 26.4 2.8 34.5 Botany 30 Doctoral 4 16-22 19.0 63.3 16.0 84.2 Master's 1 . 7- 7 7.0 23.3 Bachelor's 1 8- 8 8.0 26.6 No degree 16 6-17 10.7 35.6 6.2 58.3 Quantitative Relationships I 299 TABLE 3-Continued Mean Mean S:ize of Range of Holdings Duplicationt Book Number of Holdings Num- Per- Num- Per- Subject Sample Level Schools Low High her cent0 her cent Biology 31 Doctoral 5 11-19 15.0 48.3 10.5 70.0 Master's 13 3-16 11.4 36.7 5.6 49.2 Bachelor's 4 5-16 9.2 29.6 3.1 34.4 No degree 0 Chemistry 32 Doctoral 5 13-24 10.6 61.2 13.8 70.4 Master's 13 1-25 12.8 40.0 6.2 48.4 Bachelor's 4 5-14 8.2 25.6 3.6 44.7 No degree 0 Physics 31 Doctoral 5 19-23 21.6 69.6 17.6 81.4 Master's 9 5-18 10.8 34.8 5.3 49.1 Bachelor's 7 3-21 8.1 26.1 3.0 38.2 No degree 1 0- 0 0.0 0.0 Zoology 31 Doctoral 4 10-22 16.7 53.8 11.6 69.8 Master's 1 8-8 8.0 25.8 Bachelor's 0 No degree 17 2-15 9.0 29.0 3.8 43.1 Mathematics 31 Doctoral 5 17-25 20.0 64.5 15.9 79.5 Master's 13 2-14 8.9 28.7 4.0 45.8 Bachelor's 4 4-8 5.5 17.7 2.3 42.4 No degree 0 Management 31 Doctoral 4 20-24 22.0 70.9 17.0 77.2 Master's 10 3-22 12.2 39.3 6.1 50.6 Bachelor's 8 2-12 7.1 22.9 2.8 39.7 No degree 0 0 Percentage of the sample held by each subset. t Average (mean) duplication between every pair of libraries within a subset. gathered in the course of this study tend to substantiate the findings of re- cent network studies in other regions of the country as cited above. As did these other recent studies, this Texas study re- vealed considerable overlap in the col- lections of academic libraries. Certain differences in the various studies suggest that comparison should be made with some caution but, with this qualification in mind, a comparative summary shows the following: The collection overlap at the libraries of the six New England state universities ranged from 38 to 67 percent for current imprints and aver- aged 46.9 percent. The average duplica- tion at the nine Colorado academic li- braries ranged from 23 to 44 percent and averaged 30.8 percent. Average du- plication at five Washington, D.C., uni- versity libraries ranged from 43 to 60 percent and averaged 48.8 percent. For the twenty-two Texas academic libraries, average duplication ranged from 45 to 64 percent and averaged 52.2 percent. The data concerning commonality of subject collections at different libraries call to mind the controversial proposal made by Massman and Olson in 1971 calling for common selection at the na- tional level of current materials for small academic libraries. 3 The data from this study of Texas libraries sug- gest the extent to which libraries already are selecting in common within common curriculum areas. It may be inferred from the data that curriculum is a ma- jor controlling influence on collections but that other variables must be consid- ered in order to account for remaining differences or uniqueness in the subject collections. 300 I College & Research Libraries • July 1975 Some of the variables deserving con- sideration are: 1. Suitability or equivalency of titles; that is, some libraries may be se- lecting different titles but with sim- ilar subject content. 2. Vagaries in the selection process; for example, Massman and 0 Is on note that librarians and faculty members may be short of time, overburdened with other duties, in- adequately acquainted with books, or opposed to hard work. Further- more, that even when conscien- tiously performed, the selection process may be undermined by the haphazardness of the manner in which books get into the reviewing journals. In this regard, it is inter- esting to speculate on the influence of an authoritative book selection guide on the extent of duplication. For example, of 845 book titles in the total sampling, 114 were identi- fied as having been reviewed in Choice. In individual subjects as well as in the total sample, Choice books were duplicated at nota- bly higher rates than non-Choice books. The Choice duplication as a percentage of non-Choice dupli- cation ranged from 128 to 185 per- cent in the individual subjects and averaged 143.5 percen;t of the total sample. 3: Budget; at the doctoral level, for example, libraries in this Texas study acquired 61 percent (mean) of all books classified in a subject and duplicated among themselves 76 percent of these acquisitions. Some of this high duplication could be due simply to each li- brary's acquiring a large propor- tion of a finite population of books. At other curriculum levels, when the institution may lack mon- ey to purchase all the basic books, the budget may be forcing some of the differences in the subject collections. This variable is some- what obviated in this Texas study since the budgets of fifteen of the twenty-two libraries were sufficient to add 20,000 volumes or more to their collections during the time period studied. 4. Another variable which may ac- count for differences in the subject collections is the difference in ma- turity of the collections. During any one year, depending on the time when the degree was begun, some libraries will be seeking to c'catch up" in collection develop- ment thus adding more and differ- ent titles than their counterpart li- brary whose collection is better de- veloped. 5. A fifth variable concerns differ- ences in teaching style and the ob- servation made by many that course content and demands on the library vary though course titles are similar. While recognizing that this variable may be operating, its merit may be challenged to some extent on grounds that each library should be acquiring the important works dealing with other aspects of the subject in addition to those desired by individual faculty mem- bers. The above variables are some of those which may account for differences in the subject collections. Duplication of titles within common subject areas is high, but the collections as presently constituted do have differences. Each li- brary makes to some extent a unique contribution to overall subject coverage. The findings of this study have cer- tain general implications for coopera- tive or interinstitutional library pro- grams. In the first place, the data indi- cated a potential problem area for con- sortia arrangements based upon the as- signment of specified subjects and speci- fied subject levels to individual libraries within the consortia. It was noted that until degree programs move higher than the bachelor's level, the data showed more differences than duplication in the collections. Even at the doctoral level, the most complete single collection in a subject lacked, on the average, 23 per- cent of the titles held by libraries other than the largest. Consequently, the con- centration of collecting responsibility at one library could have the unintention- al effect of reducing subject coverage. In order to avert such a development, provisions for periodic evaluation of collections, including routine sampling, would seem to be an advisable compo- nent of specialized purchasing agree- ments. Implications may be drawn, also, for certain assumptions underlying auto- matic ordering plans based upon institu- tional profiles. (For the one-year-period focused upon in the book sampling ( 1968), none of the libraries studied had in operation an automatic ordering plan. Several have begun such plans since that time.) It has been suggested that such plans place considerable reli- ance on curriculum as a controlling fac- tor on library collections. The data from this study indicated that curricu- lum tended to influence collections but did not, in .itself, adequately explain Quantitative Relationships I 301 the differences between library collec- tions. An implication may be drawn from the data that libraries with similar curricula would tend to acquire more books in common under an automatic ordering plan than they would without such a plan. For the present, libraries should be aware that such plans may al- ter the nature of their collections. For the future, the significance of such al- terations may be a subject of further research. The question of relationship between duplication and use is an important re- search area. It is suggested that library consortia might well undertake coopera- tive research on the question of whether materials now highly duplicated are the materials that have high rates of usage. Impelled by mounting pressures to de- velop new services, to maintain access to an expanding national and world lit- erature, and, at the same time, to stabi- lize costs somehow, academic libraries are necessarily seeking the benefits of cooperative collection development. Studies of the duplication and use of materials among cooperating libraries can provide some of the essential infor- mation required for effective network planning. REFERENCES 1. For example, duplication studies for New England, Washington, D.C. ; and Colorado libraries are reported in the following: Wil- liam R. Nugent, "Statistics of Collection Overlap at the Libraries of the Six New England State Universities," Library Re- sources & Technical Services 12:31-36 (Winter 1968); Ralph Halsted Parker, A Feasibility Study for a ]oint Computer Cen- ter for Five Washington, D.C., University Libraries, Final Report (Consortium of Uni- versities of Metropolitan Washington~ D.C., May 1968); and Lawrence E. Leonard, Joan M. Maier, and Richard M. Dougherty, Cen- tralized Book Processing: A Feasibility Study Based on Colorado Academic Libraries (Metuchen, N.J.: Scarecrow, 1969). 2. The classification num hers grouped under each academic department are thought to be those which would reasonably be expect- ed by most librarians. The McGrath studies, which have reported assignments of classifi- cation numbers to academic departments on a quite detailed basis, largely corroborated the assignments made in this investigation. William E. McGrath, "Determining and Al- locating Book Funds for Current Domestic Buying," College & Research Libraries 28: 269-72 (July 1967); William E. McGrath and Norma Durand, "Classifying Courses in the University Catalog," College & Research Libraries 30:533-39 (November 1969). 3. Virgil F. Massman and David R. Olson, "Book Selection: A National Plan for Small Academic Libraries," C allege & Research Libraries 32:271-79 (July 1971).