College and Research Libraries The Effectiveness of Book Selection Agents in a Small Academic Library Christopher Millson-Martula Small academic libraries typically rely more heavily upon classroom faculty as book selectors than do large academic libraries in which librarians and book jobbers are the principals in- volved. Given the various constraints that small academic libraries face, it is important for them to rationalize the entire collection development process and also to employ the most effec- tive agents as book selectors. This article describes a study of the relative effectiveness of class- room faculty and librarians as book selection agents in a small academic library. nflation, declining or stable en- rollments, revised institutional priorities, curricula, and other factors have resulted in a re- thinking of collection management for many academic libraries. In some institu- tions a specific staff member has been identified as the chief collection develop- ment officer with at least coordinating re- sponsibility for all aspects of collection management. In many other institutions approval plans and other methods have been adopted to .stretch as far as possible every dollar expended, and greater atten- tion has been given to collection evalua- tion, especially with regard to periodical subscriptions. The overall result for many academic libraries has been a greater, more intense focus on collection develop- ment and an elevation of collection man- agement as a specialty area to a level ap- proaching that of public services and technical services. If library literature is an accurate indicator, the majority of signifi- cant collection management activity is oc- curring in a large academic and research li- braries. However, as the study described in this article indicates, significant collec- tion management activity is also occurring in small academic libraries-collections with fewer than 200,000 volumes and an- nual materials budgets of less than $150,000. Regardless of the size of an academic li- brary, relatively little attention has been given to the question of who can most ef- fectively select materials for the collection. This lack of attention to the effectiveness of selection is especially surprising since a collection constitutes a large investment in dollars, time, and space, and any practical administrator should determine not only the rate of return on that investment but · also should investigate the quality of the investment decisions that are made. In small academic libraries, the two groups most often responsible for selection are classroom faculty and librarians. Does one group of selectors, either classroom fac- ulty or their librarian colleagues, generally make more effective collection decisions than the other? THE SETTING The library described in this study serves an urban commuter college located in Chicago. Approximately two thousand students are enrolled in both traditional liberal arts courses and career-related pro- grams in areas such as business, criminal Christopher Millson-Martula is an academic librarian residing in the Chicago area. 504 justice, education, and nursing. In addi- tion to offering courses at the undergradu- ate level, the college also offers some career-related programs at the master's level. The library collection consists of approx- imately seventy-five thousand volumes and four thousand volumes are added an- nually. Membership in two consortia, in addition to OCLC, provides direct access to almost ten million volumes. The materi- als budget for both books and periodical subscriptions is about $100,000, with allo- cations for book purchases in various sub- ject areas based on a mathematical for- mula in which use (both external and internal) and average cost per book are the principal elements. Although funds for li- brary acquisitions are allocated to the li- brary budget, the individual allocations for subject areas are jointly administered by the library and the appropriate aca- demic departments until April 1 of each year, after which time all funds not yet spent or encumbered are administered solely by the library. Responsibility for selection of materials is shared by the classroom faculty and the three reader services librarians. Each reader services librarian has two master's degrees and has liaison, library instruc- tion, and collection development respon- sibilities in one of three broad areas (hu- manities, natural sciences, and social sciences). Both librarians and classroom faculty initiate requests. When a librarian makes a selection before April1, the librar- ian forwards the order card to the appro- priate academic department head, who, in turn, may authorize the request and for- ward it to the library for purchase. Al- though the extent of librarian involvement in the selection process varies from one subject area to another, the number of re- quests initiated by a librarian rarely drops below 40 percent of the total requests in a given subject area. THE STUDY This was a conventional study of the use of the history section of the collection, in- cluding all countries and all time periods. Only the history section of the collection was chosen because it was one of but three / The Effectiveness of Book Selection 505 subject areas in which a librarian held a second master's degree. Consequently, selection by the librarian should have been done at a relatively high level of com- petence. The investigation was limited to books, including both monographs and serials, but not periodicals. Unlike a con- ventional use study, however, this inves- tigation did not assess the use of materials added to the collection but rather investi- gated the relationship between use and selection responsibility. Hypotheses The study was designed to test two hy- potheses: 1. History books selected by classroom faculty show greater circulation activity (a greater number of circulations per book) than those selected by their librarian col- leagues. 2. History books showing circulation activity (at least one circulation per book) have a higher level of activity during · the first three years of inclusion in the collec- tion than in later years. For many years, the history section of the collection has constituted a relatively little-used part of the collection, contribut- ing an average of no more than 4 percent of the total circulation activity. History fac- ulty members frequently recommend spe- cific book titles to students for research pa- pers and other library-related classroom assignments; therefore, it was assumed that history books selected by history fac- ulty would have a greater amount of circu- lation activity than those selected by li- brarians. The Kent study at the University of Pittsburgh and other studies in academic libraries revealed that books are most likely to show circulation activity within the first three years of their addition to the library collection. 1 After this period, use drops off considerably or ceases. Items showing no circulation activity during the first three years following acquisition are not likely to have any activity in following years. Based upon the investigator's expe- rience, there were no factors likely to make the local situation different from other academic situations. A review of the 506 College & Research Libraries literature follows the discussion of the study. Data Collection and Analysis The study focused on those 691 history books that were added to the collection during the period July 1, 1977, through June 30, 1980. Evaluation of their use cov- ered the period between July 1, 1977, and June 30, 1983. For each book, two types of data were collected: status of selector (classroom faculty or librarian) and amount of external use (as measured by recorded circulation transactions). The selector's status was easily deter- mined by noting the name of the requestor on the order card for each item ordered during the data collection period. The amount of external use was determined by two methods, one for uses before October · 1980 and one for uses after that date. Until October 1980, the library used a manual circulation system. To determine uses that occurred before October 1980, the date due slip in each of the books involved in the study was checked for circulation ac- tivity. When a book was not on the shelf at the time of the data collection, the stacks were rechecked two months later. If, at that time, the book was still unavailable for examination, it was dropped from the study. In October 1980, when the library joined the Library Computer System (LCS), an automated circulation system and resource-sharing network, the library discontinued its use of date due slips, rely- ing instead upon a date due card that was valid for one circulation only. To deter- mine uses that occurred after October 1980, the circulation activity for each item involved in the study was generated from LCS circulation records. The transition from a manual to an automated circulation system served a useful purpose in that it provided a clear line of demarcation that could be used in easily determining what effect, if any, the length of time an item had been in the collection had upon its cir- culation activity. The data were analyzed in order to de- termine the relationship between use and selector, especially the relative effective- ness of the two selector groups, and that between use and the length of time an November 1985 item had been in the collection. It was de- termined that the chi square test applied to a 2 x 2 contingency table at the .01level of significance could be used effectively to test the study's first hypothesis. Results During the three-year acquisition period covered by the study, 691 books were added to the collection and charged to the history fund. Librarians selected the great majority of books, a total of 523 books or 75.69 percent, that were added to the col- lection during the study; classroom fac- ulty selected 168 books or 24.31 percent of the total. Although the total number of books selected in each of the three years ranges from a high of 420 to a low of 87, the low can be considered typical for a year during which 4,000 books are added to the collection. The number of books added during the first two years of the study was significantly higher due to retrospective collection development based in large part upon the second edition of Books for College Libraries. It is not unreasonable to assume that in many academic libraries history materials receive high use due to the nature of the information transfer process in history and the widespread prevalence of course assignments requiring library use. Ac- cording to the data collected in this study, however, history materials constituted no more than 4 percent of total circulation ac- tivity in each of the years covered by the study. Of the 691 books added to the col- lection, only 420 books or 60.78 percent showed any circulation activity. The total number of circulation transactions was 1,176 or 1.70 circulations per book for all books added and 2.80 circulations per . book for all books having at least one cir- culation transaction. Tables 1 and 2 pro- vide more detailed information about cir- culation activity for books selected by each group of selectors. The data seem to indi- cate that in terms of circulation per book classroom faculty make more effective book selectors than their librarian col- leagues. In table 3 the results are pre- sented in a slightly different way. This ta- ble indicates equal effectiveness in selection activity on the part of classroom The Effectiveness of Book Selection 507 TABLE 1 CIRCULATION ACTIVITY FOR BOOKS SELECTED BY CLASSROOM FACULTY Circulations Circulations Books Book per Book for All ra;r:~~ Year Circulated Circulations Books Added Circulated 1977-78 54 183 2.23 3.39 1978-79 36 121 1.95 3.36 1979-80 14 29 1.21 2.07 Composite 104 333 1.98 3.20 TABLE2 CIRCULATION ACTIVITY FOR BOOKS SELECTED BY LffiRARIANS Circulations Circulations Books Book per Book for All ra;r:~ Year Circulated Circulations Books Added Circulated 1977-78 203 557 1.65 2.74 1978-79 75 217 1.78 2.89 1979-80 38 69 1.10 1.82 Composite 316 843 1.61 2.67 TABLE3 BOOKS CIRCULATED AS A PERCENTAGE OF BOOKS SELECTED Selector Classroom faculty Librarians Books Selected 168 523 faculty and librarians. Whether or not this would still be the case if retrospective ac- quisitions (those books selected from Books for College Libraries) were removed from the study remains to be seen. How- ever, the Pittsburgh study shows that cur- rent publications receive greater use than those purchased on a retrospective basis. 2 Unfortunately, retrospective acquisitions were not singled out during the study. Two other analyses were deemed ap- propriate for the study: (1) the percentage relationship between selectors and books selected, number of titles circulated, and total number of circulations; and (2) there- lationship between selector and single versus multiple circulation transactions. Table 4 indicates that while both groups of selectors may be equally effective in terms of circulation activity, classroom faculty make a greater contribution in terms of se- lecting books that have multiple circula- tion transactions. This is confirmed by ta- ble 5. These data also indicate that .librarians may be selecting those books Books Circulated 104 316 Books Circulated as a Perces~fe~t~~ Books 61.90 60.42 more likely to be used by .either nonhis- tory students or history students who do . not consult with their instructors for rele- vant sources. Although there are no rele- vant data available, perhaps one can as- sume that the selections of librarians added breadth to the collection while the selections of classroom faculty tended to provide the collection with a greater de- gree of depth in relatively few subject ar- eas. While these analyses of the data provide valuable insights into the relationships among categories of selectors, books se- lected, and books circulated, other analy- ses were needed to test the research hy- potheses. To test the first hypothesis, a 2 x 2 contingency table was created, and the chi square test was used. With one de- gree of freedom, a chi square value be- yond 6.635 is statistically significant at the .01 level. Since the resultant chi square value was .0000085, the first hypothesis was rejected. The second part of the study involyed 508 College & Research Libraries November 1985 TABLE4 RELATIONSHIP BETWEEN SELECTOR AND ACQUISIDONS AND CIRCULATION ACTIVITY IN PERCENTAGES Selector Classroom faculty Librarians Books Selected 24.31 75.69 Books Circulated 24.76 75.24 Book Circulations 28.32 71.68 TABLES SINGLE VERSUS MULTIPLE CIRCULATIONS BY SELECTOR Selector Classroom faculty Librarians Books Circulated Once Number Percentage of of Books Books CircUlated 30 125 28 .85 39.56 an analysis of circulation activity before and after the library's transition from a manual to an automated circulation sys- tem. While analyzing circulation before and after October 1980 may be considered primitive because it did not allow for a uni- form time period under each circulation system, the change in systems nonethe- less provides a demarcation line for deter- mining use in the recent and distant pasts. Under both circulation systems it is ex- pected that items not used in the first three years after their addition to the collection are not likely to be used in succeeding years. Approximately 38.8 percent of the books were not used at all. When one combines the books not used at all with those books with a decrease in use under the automated circulation system (71 per- cent for 1977-78 books, 68 .5 percent for 1978-79 books, and 52.9 percent for 1979-80 books), it seems that for this par- ticular collection the majority of books have little or no value to users after a rela- tively short period of time. Table 6 pro- vides more detailed information. While about 30 percent of the books showed decreased circulation activity un- Books Circulated More Than Once Number Percentage of of Books Books CircUlated 74 191 71.15 60.44 der the automated system, about 25 per- cent of the books experienced increased use; the 41.5 percent increase for 1979-80 books is probably artificially high since most of these books did not have the op- portunity for significant circulation activ- ity under the manual circulation system. While the number of total book circula- tions and of titles circulated decreased un- der the automated circulation system compared to the manual system, this was not consistently true for the first two years of the study (see tables 7 and 8 which con- tain data for books circulated under both circulation systems). Although the in- crease for 1979-80 books under the auto- mated circulation system is quite expected due to the brief period of time those books were available for circulation under the manual system, the same cannot be said for the increase demonstrated by the 1978-79 books. Thus, since the data indi- cate inconclusive results, the second hy- pothesis cannot be supported. COMPARATIVE DATA How do the results of this study com- pare to data gathered at similar institu- TABLE6 Year of Acquisition 1977-78 1978-79 1979-80 CHANGE IN CIRCULATION ACTIVITY FROM FIRST TO SECOND CIRCULATION SYSTEM AS A PERCENTAGE Increase Decrease Some Use 23.56 32.86 5.48 27.17 29.35 4.35 41.35 13.80 5.75 No Change No Use 38.10 39.13 39.10 The Effectiveness of Book Selection 509 TABLE 7 CIRCULATION ACTIVITY UNDER MANUAL SYSTEM FOR BOOKS CIRCULATED UNDER BOTH CIRCULATION SYSTEMS Year 1977-78 1978-79 1979-80 Book Circulations 427 157 25 Books Circulated 192 75 18 Circulations _per Book for Books Crrculated 2.22 2.09 1.39 TABLES CIRCULATION ACTIVITY UNDER AUTOMATED SYSTEM FOR BOOKS CIRCULATED UNDER BOTH CIRCULATION SYSTEMS Book Year Circulations 1977-78 315 1978-79 180 1979-80 72 tions? A review of the literature indicates that smaller academic libraries are not un- dertaking studies of the relationship of book use and book selector responsibility, or, if they are, the results are not being broadly disseminated to the profession. The author located only one similar study, and it was but one portion of a broader DePauw University study that partially replicated the 1978 University of Pittsburgh library materials study. 3 Larry Hardesty studied the circulation activity of slightly more than 1,700 books selected by both librarians and classroom faculty. Librarian selections constituted only 173 of the total. The results led Hardesty to conclude that ''librarians can play a useful role in the selection of books for the aca- demic library."4 He found that 31.2 per- cent of librarian-selected books were not used at all, while the comparable figure for books selected by classroom faculty was 34.2 percent. In addition, his results indi- cated that librarian-selected books consti- tuted a higher than expected portion of books with moderate or heavy use (more than. six circulations). Hardesty concludes that ''the main difference between the classroom instructor-selected books and librarian-selected books is that the librari- ans selected fewer gr:aduate level books in narrow specialties.''5 Although the Hardesty study seems to be the only reported study of book use and selection responsibil_ity in a small aca- Books Circulations per Book Circulated for Books Circulated 157 2.01 80 2.25 42 1.71 demic library, several such studies have been conducted in university libraries. A landmark study was conducted by Gayle Edward Evans at four public and private university libraries located in the Midwest and the Rocky Mountain region during the 1960s.6 His study involved an exami- nation of circulation · activity in relation to selection agent: librarian, classroom fac- ulty, or book jobber. Evans hypothesized that selection agents having the greatest number of contacts with the greatest num- ber of library users would select the high- est percentage of titles showing circula- tion activity. To test this hypothesis Evans studied the circulation of 6,891 titles (English-language current imprint mono- graphs) based on a random sample of at least 500 titles per type of agent per institu- tion. Analysis of circulation activity cov- ered the first twelve months of availability in the collection. The study results con- firmed part of Evans' hypothesis, namely, · that librarians tended to select the greatest number of materials that were used; fac- ulty and jobbers followed. However, Evans apparently did not feel sufficiently confident to attribute the librarians' per- formance to their extensive contact with li- brary users. At first, he thought that the differences between librarians and faculty arose from different selection philoso- phies. Later he stated that both types of agents appeared to be selecting from the same philosophical point of view. Unable 510 College & Research Libraries to make any- definitive statement, Evans concluded with an expression of need to determine the cost-benefit factors for each type of selection agent. Robbie Bingham carried out a similar study approximately ten years later at four university libraries in the South. 7 While Bingham also investigated the relation- ship between categories of selectors and use of selected items, she added two cate- gories of selection agents to Evans' three: faculty/jobber and librarian/jobber. Sam- pling 7,224 titles (450-500 items for each single category of selection agent and at least 100 items for each of the two combi- nation categories}, Bingham hypothe- sized the following descending order of · use by selection agent: (1) faculty/jobber; (2) librarian/jobber; (3) faculty; (4) librar- ian; and (5) jobber. Bingham found that the single categories of selection agents were more effective than the combination categories, with faculty selecting the greatest number of materials used. How- ever, for materials dealing with the hu- manities, librarians replaced faculty as the most effective selection agent. CONCLUSION It is not surprising that the results of the studies cited above are not uniform. Per- haps the factors exercising the greatest in- November 1985 - fluence are key institutional charact- eristics-curricula, library, faculty, and students-that are unique to each institu- tion. It is also not surprising that a great deal of work remains to be done before aca- demic libraries, especially smaller aca- demic libraries, can establish collection development processes that are appropri- ate for meeting users' demonstrated needs. In order to do so, considerable analysis needs to be done concerning se- lection agents. Factors such as educational background, relevant experience, philos- ophy of selection, and sources or methods used for selection (together with the re- lated costs involved) have a direct bearing on the effectiveness of collection develop- ment. Likewise, selections should be ana- lyzed in terms of those pre-acquisition in- dicators that Weeks claims predict book use: type of publisher, language of publi- cation, date of publication, and type of publication (single or multiple author, conference proceedings, serial reviews, and bibliographies).8 What can be said with a reasonable de- gree of certainty is that librarians play a significant role in book selection. The goal . remains, however, to enhance that role in a way that will make collections more rele- vant to curricular or research needs. REFERENCES 1. Allen Kent and others, Use of Library Materials: The University of Pittsburgh Study (New York: Dek- ker, 1979). 2. Ibid. 3. Larry Hardesty, "Use of Library Materials at a Small Liberal Arts College," Library Research 3:261-82 (Fall1981). 4. Ibid., p.275. 5. Ibid. 6. Gayle Edward Evans, "The Influence of Book Selection Agents Upon Book Collection Usage in Academic Libraries" (Ph.D. diss., Univ. of Illinois, 1969). 7. Robbie B. Bingham, "Collection Development in University Libraries: An Investigation of theRe- lationship Between Categories of Selectors and Usage of Selected Items" (Ph.D. diss., Rutgers Univ., 1979). 8. Kenneth Weeks, Determination of Pre-acquisition Predictors of Book Use (Berkeley, Calif.: Institute of Library Research, Univ. of California, 1973).