College and Research Libraries Commentaries on Collection Bias Marcia Pankake, Karin Wittenborg, and Eric Carpenter The following commentaries are in response to the article, "Potential Collection Development Bias: Some Evidence on a Controversial Topic in California," by Dave Harmeyer on page 101 of this issue of College & Research Libraries. COLLECTION BIAS: ETERNAL VIGILANCE THE PRICE OF LIBERTY Dave Harmeyer's study exhibits a grave fault when it leaps from the num- bers of copies held to conclude that the librarians who built the collections are bi- ased. The only conclusion the study safely makes is that libraries in California hold a larger number of copies, editions, or titles of these particular pro-choice books than of these particular pro-life. books. What is the definition of a balanced · collection? Surely not equal numbers of books. By their very individualistic na- ture, books cannot be equated one for one. One comprehensive book may do the job of several smaller or less ambitious books; one side of a question may be represented by a thorough well-organized book and another side of the question by a half- dozen books presenting specific aspects, or by a subscription. Librarians protect ideas as much as books. Do these eight books represent the lit- erature of the pro-choice and pro-life movements qualitatively or quantita- tively? The author and his judges label the books, but offer no assurance that these books are particularly important, are the most necessary or the best books. Neither do they quantitatively represent the literature. Books in Print lists about 450 books on abortion; only one of these eight titles is listed under the more re- cently used specific headings for pro- choice and pro-life movements. Set aside the question of whether a few books reflect the dissemination of ideas. The author's belief that the librarians were biased and deliberately never ac- quired the books provides only one possi- ble explanation for the disparate numbers of holdings of these books. Many other factors influencing selection must be ex- amined before we accept this assertion. The author acknowledges that library policies affect selection. An examination of these policies would help more to un- derstand the library collections than does the distance between the judges' interviews. What are the libraries' poli- cies? What are their audiences, their needs, and demands? Circulation needs, for example, may require multiple cop- ies of one title, thus increasing the num- bers of "units" identified. Do the libraries serve vocal communities where patrons actively request pro-choice books and librarians buy modest numbers of pro- life books to represent that viewpoint? Selection and acquisitions procedures may influence collections more than do librarians' prejudices. Where were the books reviewed? Only four of these eight Marcia Pankake is Bibliographer at the Wilson Library, the University of Minnesota, Minneapolis, Minnesota 55455; Karin Wittenborg is University Librarian at the Alderman Library, University of Virginia, Charlottes- ville, Virginia, 22903; and Eric Carpenter is Collection Development Librarian at Oberlin College, Oberlin, Ohio44074. 113 114 College & Research Libraries books appear in Book Review Digest within two years of publication, suggest- ing that reviews for the other four did not appear in the common sources. Were the books reviewed while they were still in print? Were they publish- ed in editions large enough so that stock remained available after the re- views appeared? Were they reviewed positively, offering librarians reasons to select them, or negatively, offering rea- sons to pass them by? Or compared to other books on the topic which were se- lected instead? Did librarians reflect bi- ases from other sources? Were the books listed in nonreviewing trade sources such as the American Book Publishing Record? Did they move through the organized book trade, were they sold by distribu- tors? While libraries should not limit their collections exclusively to materi- als handled by jobbers, for very practi- cal reasons most materials in most libraries are obtained expeditiously through vendors. Local holdings and collections pro- vide only part of the information librar- ies supply. How do the librarians augment their collections with sources not owned, information available from remote loca- tions, information to which librarians supply access? Online bibliographic da- tabases do not portray how any library publicizes its holdings or directs patrons to electronic information or interlibrary loans. Identifying shortcomings of this study does not serve to defend librarians charged with bias, or excuse less than rigorous selection practices, nor justify librarians' complacency about impartial collections. The criticism is offered in- stead to encourage more careful investi- gation of the topic, which may suggest how librarians can improve their selec- tion practices. Rather than accepting rea- sons for why books do not appear in library holdings, we need to identify causes of weak selection practices and then counteract them. For example, if the book trade sources do not list certain publishers, librarians must solicit those publishers directly. Librarians must ex- amine their patterns of information for . March 1995 selection to add supplementary sources, sometimes identifying them through personal contact with readers. We should not assume that library col- lections are static. Good collection man- agers are always engaged in repair and renovation, often rethinking titles not in their collections. Such reevaluation does not refer to the reasons why a title was passed by, but instead reconsiders the title as a new purchase. How does its content relate to local needs and to the strengths or weaknesses of the library, which may differ from when the book first appeared? Books and the ideas they present can be vulnerable in libraries and society. Librarianship is a privileged occupation, for everyone who works with books and other library materials to make informa- tion available engages in a noble and necessary work in a democratic society. Library staff must vigorously promote the right to read and protect minority interests. All library staff must positively defend freedom of information by carry- ing out their responsibilities honorably, sometimes by acquiring and preserving materials of which they may personally disapprove. Acquisitions staff must or- der titles expeditiously, from suppliers likely to supply materials quickly, and then receive and move the material into the cataloging stream as fast as possible. Catalogers must catalog fully, accu- rately, and quickly. The book must be marked and shelved accurately. At every step along this chain of actions the book is vulnerable, to individuals on the li- brary staff or from outside the library who seek to repress it from the public. A huge web of trust maintained by every- one who works in the library makes in- tellectual freedom possible. When readers realize the variety and number of poten- tial hazards that lie between books and readers, they stand amazed that libraries have any controversial materials at alL- Marcia Pankake COLLECTION BIAS: WHAT'S RIGHT? Dave Harmeyer's article on collection development bias is provocative and opens some interesting avenues for fur- ther research. His literature review gives a useful overview of the classic and more recent articles on issues relating to cen- sorship and selection in libraries. While I am not convinced that the case has been made that selectors in academic and public libraries are introducing bias into their collections, the article serves a use- ful purpose by focusing attention on a potential problem and identifying some important questions that could benefit from exploration. Perhaps the most provocative ques- tion raised is both philosophical and practical. ALA's Library Bill of Rights states, "Libraries should provide materi- als and information presenting all points of view on current historical issues." What exactly does it mean to present all points of view? In the abstract, it may seem desirable to balance a position on one side with a position on another. In reality, it may be neither reasonable nor desirable. Does presenting all points of view require academic libraries to pur- chase an equal number of books on both sides of an argument? Does it mean that we must reflect certain perspectives in our collections even if the books are not scholarly? Harmeyer touches on some possible explanations for why pro-life books are not selected for academic libraries. He mentions that conservative books often are not widely reviewed and that both collection policies and budget constraints can present legitimate barriers to repre- senting all points of view on an issue. Indeed, both collection development policies and budget constraints are key to the building of academic library col- lections and merit further attention. A number of questions come to mind. Se- lection in academic libraries is closely linked to faculty research and teaching interests. Once selectors have deter- mined that the subject of a book is rele- vant, they usually assess whether the material is scholarly in nature. Table 2 indicates that of the sample eight books, three were "very pro-life," but none was moderately pro-life. In contrast, in the sample there were two "very pro-choice" and two moderately pro-choice titles. In Commentaries on Collection Bias 115 a note, Harmeyer mentions the imbal- ance and suggE7sts that it may result from the nature of the pro-life issue or simply that a more diversified sample of pro-life titles was needed. It would be worth exploring whether the extreme pro-life titles were not selected by academic li- braries because their polemical character was considered unscholarly. The influence of budgets on collec- tions building has been abundantly clear in the last few years with the rise in serials prices and the drop in mono- graphic acquisitions. It is a phenomenon that affects libraries differently. Har- meyer's study would have been strength- ened by more information about the size and the scope of the libraries in the sam- ple. Is there any correlation between bias and size of libraries? Are library collec- tions increasingly being influenced by demand and use? An interesting follow- up study to Harmeyer's might be to ex- amine use of pro-life and pro-choice titles and evaluate the degree to which current ·selection is based on the use of the existing collections. As "access" becomes a larger element of every academic library's operation and .as scarce dollars are used to pur- chase titles expected to be in high de- mand and high use, . it will be an interesting challenge to maintain a di- versity of viewpoints in the collection. Perhaps the greater challenge will be how we let users know about the uni- verse of materials available beyond the core collection. This study also made me curious about the possible influence of approval plans. I suspect that the religious-affili- ated institutions are smaller and may not make extensive use of approval plans. If that is the case, and if as Harmeyer sug- gests, the conservative books are not widely publicized or distributed via mainstream channels, the imbalance could be driven by the use of approval plans. The average selector in a large academic library will have a broad range of responsibilities and may not readily observe if the approval plan delivers a disproportionately high number of pro-choice books. A selector who is firm 116 College & Research Libraries ordering every title is likely to have a greater first-hand knowledge of the col- lection and can more readily seek bal- ance. Can the argument be made that the abortion debate is a more significant topic for religious-affiliated institutions than it · is for other libraries and thus attracts more attention? Again, it would be helpful to know more about the size and character of the libraries studied. The article piqued my curiosity about the size of the literatures on either side of the abortion debate. A quick search of the public-access version of the OCLC database surprised me. The subject heading "pro-life movement'; identified just over 600 records, while the pro- choice movement identified just under 200 records. Given a pro-life literature that may be nearly three times the size of the pro-choice selection, it may be that the academic and public libraries have many pro-life titles, but simply not the ones in Harmeyer's sample. That might bear further investigation as it relates to the issue of whether the academic librar- ies did not select the "very pro-life" titles because they were not scholarly and may have been viewed as polemic. A similar study with a larger and more diversified sample might be informative. Harmeyer's study deserves attention from academic libraries, and I hope it prompts discussion well beyond collec- tion development circles. Sweeping generalizations about balance and di- versity need to be interpreted for indi- vidual institutions and for selectors. Is bias creeping into the collections inad- vertently? What is our commitment to ensuring the representation of multiple viewpoints and how do we ensure that it is realized? What really is our obliga- tion? The larger issue is that academic libraries face a future in which they con- tinue to purchase a decreasing portion of an increasing publishing output, yet we need to teach our users about the larger universe of information and facilitate their use of it. While electronic initiatives show some promise in that regard, the overall environment is much more com- plex and difficult to navigate.-Karin Wittenborg March 1995 ADEQUATE REPRESENTATION AND NUMERIC EQUIVALENCY: HOW MUCH IS ENOUGH? Dave Harmeyer's article is timely and provocative. It questions an assumption we are loathe to challenge: some librari- ans are "consciously or unconsciously, biased in building collections on contro- versial social and political issues." Spe- cifically Harmeyer charges that public and academic librarians developing col- lections examined in California "con- sciously or subconsciously discriminate against a conservative social/political perspective, in this case, the pro-life point of view." Harmeyer's charge is based on a comparison of holdings in various academic and public libraries in California with a list of eight books rep- resenting the spectrum of opinion on the abortion controversy. Why were these particular books cho- sen for the sample? Is this sample mean- ingful? Four sampling criteria are listed: availability in academic and public li- braries, number of book reviews, cur- rency, and "the sense of being a classic." These criteria are weak and inadequate. A more meaningful sample would have included the most important, repre- sentative, widely cited and reviewed, and available books on abortion publish- - ed in a specific period (e.g., the past twenty years). A sample based on these criteria and selected by a panel of recog- nized experts on opposing sides of the abortion issue would have been more compelling. More serious than this methodological weakness is the assertion that adequate representation of opposing views in a li- brary collection requires numeric equiva- lency. Most librarians assume that "libraries should provide materials and information representing all points of view on current and historical issues."1 This is a basic prin- ciple in our . professional creed. Har- meyer challenges the adequacy of our response to this quintessential profes- sional responsibility. While his research was limited to California, Harmeyer im- plies that the conservative point of view on political and social issues is not ade- quately represented in academic and public library collections across the country. Indeed, Harmeyer calls for similar investigations on "other contem- porary social/political issues such as gay /lesbian/bisexual rights, feminism, fetal tissue experimentation, school based clinics, political correctness, or gays in the military." This charge of bias in collection development is a serious matter demanding careful consideration. The basic question at issue is not whether libraries should represent all points of view on controversial issues, but rather what constitutes adequate representation. Harmeyer tests two hy- potheses, both concerning "a significant difference between the number of repre- sentative pro-choice and pro-life books selected by California academic and public librarians." In Harmeyer's view numbers alone determine adequacy of representation of opposing viewpoints. His argument hangs on the proposition that a collection must have equal, or very nearly equal, numbers of titles advocat- ing opposing views on abortion to en- sure adequate representation of these views. A librarian whose collection does not have this equilibrium is guilty of bias in collection development. LeRoy Charles Merritt, in a work cited by Harmeyer, refutes this notion. He re- fers to "The Myth of Library Impartial- ity" by pointing out that "libraries cannot supply an equal number of titles on both or all sides of every political issue. They must follow the pattern of book publication and cannot wait for a title to appear on the "other" side before making a purchase. It is necessary, how- ever, that the authentic and important books on every political issue which meets the normal selection criteria be acquired as they are published. "2 Merritt is correct. It is impossible to ensure that an equivalent number of titles on oppos- ing sides of every political and social issue is acquired for a library collection. What then constitutes adequate repre- sentation of opposing viewpoints? Li- brarians must ensure that major and representative voices on opposing sides of controversial issues are acquired. Do- Commentaries on Collection Bias 117 ing this requires heroic effort in a poorly funded library, but it will provide ade- quate representation of opposing views. Excessive concern about numeric equiva- lency of texts advocating opposite view- points is not warranted. Harmeyer assumes that the absence of equal numbers of titles on opposing sides of the abortion issue is the result of bias by the selector. Is it possible that factors other than bias are responsible? Harmeyer concedes that "collection de- velopment policies and limited budgets can present legitimate barriers for li- brarians striving to select for a collection that does present all points of view on issues." Librarians working in seriously understaffed libraries do not have time to ensure that equivalent numbers of ti- tles are acquired on opposing sides of every issue. Collection development requires bal- ancing competing needs. Often the most pressing is patron demand. Patron de- mands may conflict with a librarian's commitment to provide adequate repre- sentation of opposing viewpoints on controversial issues such as abortion. Academic librarians on some campuses may face demand for more pro-choice than pro-life books, and a public librar- ian serving a conservative community may have difficulty justifying expendi- ture of public funds on books advocating freedom of choice on abortion. Inadequate selection tools also make it difficult to ensure adequate representation of opposing viewpoints on controversial issues. Harmeyer charges that "quality conservative religious books, critical of liberal social movements" are not ade- quately reviewed in mainstream review- ing sources. This is at least partially true. Conservative religious and social change publishers are not adequately covered by major review journals. Harmeyer's call for research on this subject is on target. Research is also needed to deter- mine whether books by conservative publishers are readily available through library book jobbers and how quickly these books go out of print. Though Harmeyer's insistence on nu- meric equivalency to ensure adequate 118 College & Research Libraries representation of conservative views is not justified, there is some truth to his charge that the conservative viewpoint is often underrepresented in libraries. Eric Moon's assertion in 1969 that "radi- cal right pressures on libraries have gained some strength from the general truth of their argument that library col- lections tend to favor liberal over conser- vative points of view" is still true. 3 How then should academic and public librar- ies respond to this challenge to provide more adequate representation of ·Con- servative vi~wpoints on controversial issues? The answer is obvious-by exam- March 1995 ining existing collections to ensure ade- quate representation of conservative views. This requires checking holdings to ensure that works by Jerry Falwell, Rush Limbaugh, Phyllis Schlafly, and Cal Thomas are included on library shelves with more liberal writers who oppose their views. Position papers and news organs from right-to-life organiza- tions should be acquired along with those from pro-choice advocacy groups. Such a response by selectors for aca- demic and public libraries will suffice; no undue concern for numeric equiva- lency is required.-Eric Carpenter REFERENCES 1. Library Bill of Rights (Chicago: ALA, 1980). 2. LeRoy Merritt, Book Selection and Intellectual Freedom (New York: Wilson, 1970), 16. 3. Eric Moon, Book Selection and Censorship in the Sixties (New York: Bowker, 1969), 7. IN FORTHCOMING ISSUES OF COLLEGE & RESEARCH LIBRARIES Cyrillic Transliteration and Its Users Alena L. Aissing Students' Perceived Effectiveness Using the University Library Cheryl McCarthy Government Information Systems: A Quantitative Evaluation John V. Richardson, Jr. and Rex B. Reyes