College and Research Libraries The Lean Reference Collection: Improving Functionality through Selection and Weeding Christopher W. Nolan Librarians expend significant amounts of money, time, and space on their reference collections, yet a surprising number state that th~ir collections are full of materials which receive little use. Additionally, a large proportion of academic libraries do not have collection development poli- cies for this important asset. The author discusses this problem and develops a series of guide- lines for placing sources in reference, focusing especially on the suitability of the items for true reference functions and the expected frequency of use. ibrarians implicitly assign the reference collection a high value, yet the professional lit- erature contains little discus- sion of its management. 1 Expenditures are the most obvious proof of this value. Librarians spend a significant percent- age of library materials budgets on refer- ence books, serials, and, increasingly, automated products. Just in terms of size, the number of titles in academic li- brary reference collections range from a median of 5,000 titles in small college li- braries to almost 30,000 titles in ARL libraries-and many of these libraries in- dicate that only space limitations keep these numbers from growing. 2 Space de- voted to a reference collection provides another tangible measure of its value. Most academic libraries assign the refer- ence department and collection to a highly visible location, one that students and faculty will find accessible-what Mary and Victor Biggs call "highest- priority library space.' ' 3 Perhaps less tangible, but more impor- tant, is the usefulness of this collection for the provision of reference service. The reference collection is, along with the catalog, frequently the first (and sometimes only) source consulted while helping users. If the quality of reference service hinges on the resources mar- shaled for support, this close-at-hand collection should be critical. Many patrons also place significant value on the reference collection. Anec- dotal evidence suggests that patrons of- ten perceive reference items to be more authoritative than those items not in ref- erence. This logic is not difficult to fol- low. Reference books do not circulate, and often must even be requested at a reference desk. Librarians, asked for help answering a query, turn to some sources in reference and deliver the de- sired answer. Bibliographic instruction librarians demonstrate the use of in- dexes and provide lists of other suitable reference items. Current statistical sources are kept in reference, with ear- lier, outdated editions going to the stacks. All of these factors suggest to the library user that those people with ex- pertise, the librarians, significantly value reference books. THE PROBLEM: POORLY MANAGED COLLECTIONS This value should have resulted in Christopher W. Nolan is Reference Services Librarian and Assistant Professor at Trinity University, San Antonio, Texas 78212. 80 most libraries devising comprehensive policies for managing this resource. However, the Biggs' survey in 1985 found that only 7% of undergraduate in- stitution libraries, 17% of masters-level libraries, and 59% of ARL libraries had written reference collection develop- ment policies. Even fewer libraries had policies for managing ready reference collections. 4 Eugene A. Engeldinger found that less than 12% of academic li- braries had a written policy concerning . weeding of the reference collection. 5 Many reference librarians think their ref- erence collections are too large to be used effectively, and that it is difficult to keep track of what is in the collection and what should be updated or weeded. Does the absence of these policies mean that reference collections are not being managed effectively? It could be argued that other subject collection poli- cies and procedures cover reference ma- terials adequately. However, the Biggs' survey strongly suggested that a large number of reference collections are just too large to be used efficiently by librari- ans. Librarians at all types of academic li- braries estimated that over 25% of the items in their reference collections had not been used in the last five years, and that half the collection was not used in the last year. 6 Although these were rough estimates, they indicate that many reference librarians think their ref- erence collections are too large to be used effectively, and that it is difficult to keep track of what is in the collection and what should be updated or weeded. In addition to collection size, the qual- ity of information contained therein should also be considered. If patrons and librarians frequently regard refer- ence materials as more authoritative or valuable for answering their questions, and the collection contains seriously Lean Reference Collection 81 outdated materials, librarians may be re- lying on inaccurate sources. 7 Gail Sch- lachter's short survey of reference col- lections and policies reported that, although most of the libraries had out- dated editions of particular materials, those with reference collection develop- ment policies were more likely to pro- vide recent (and more accurate) editions than those libraries without policies. 8 REASONS FOR THE PROBLEM Thus, many libraries, through failure to manage carefully, have created large, unwieldy reference collections which contain outdated or unused sources. Four primary factors contribute to this deficiency. First, new reference tools have proliferated over the past few years. Librarians may have been more concerned with acquiring these new sources than with determining how their coverage compares with already- owned sources or with weeding titles that are no longer necessary. Second, we have not operated with a clear definition of a reference source. Re- sponses to Marcia Bates' question "What is a reference book?" were that the definition was obvious, but also dif- ficult to pin down-often from the same respondent. 9 Merely using the basic ALA Glossary definition-that a refer- ence book is a book used mainly for look- ing up definite pieces of information, rather than for continuous reading-is not sufficiently helpful because this defi- nition is relative. 10 What one reader may use only for referral-say, a commentary on the New Testament used to find in- formation on a particular passage- another reader may read cover to cover. Likewise, librarians evaluating a new ti- tle for inclusion in the reference collec- tion may debate whether a collection of essays with lengthy bibliographies will be used more for referral, or consecutive reading. This quandary apparently led Bill Katz to remove a very relative defini- tion of a reference book from later edi- tions of his standard reference text- book.11 Third, it is unclear what sorts of refer- ence books need to be in the reference 82 College & Research Libraries collection. Frequently, one librarian's "essential for reference" source is con- sidered obscure or trivial by colleagues. Of course, this difference of opinion can occur even when a well-defined collec- tion development policy is in force and when interpretation varies. But it is much more likely when oral tradition, rather than written policy, guides the decision. Fourth, the nature of the reference col- lection hinders its effective manage- ment. Reference sources can be classi- fied both as reference items and as items within subject disciplines. In many li- braries it is unclear whether a particular reference source will be ordered by the subject bibliographer or by a reference bibliographer. It is also unclear whose responsibility it is to determine the loca- tion of the item (reference or stacks) and who weeds and evaluates the reference collection. In such situations, strong co- ordination of reference collection func- tions is necessary. Yet the compilations of collection policies in the literature show that many libraries have not clari- fied this role, allowing the reference col- lection to be neglected. 12 STEPS TO A SOLUTION Given the need for more attention to reference collections, how do we man- age these issues more effectively? We will hardly be able (or want) to stern the flow of new reference sources in both traditional and electronic formats. More careful selection can avoid purchase of sources which duplicate existing items, and better book reviewing can aid care- ful selection. 13 Creating a reference col- lection development policy would be a positive step toward better manage- ment. However, most of the policies in the aforementioned collections give little rationale for the entire enterprise. In- stead, policies immediately launch into discussions of types of reference sources. 14 Libraries need to consider systemati- cally what they want in their reference collections, why they want them there, and how they are placed there. This pro- cess requires five steps: first, defining January 1991 more carefully what is meant by "refer- ence source," because this is the build- ing block of our collection; second, de- lineating the purposes of the reference collection; third, deciding just what types of reference sources belong in ref- erence to achieve the best level of refer- ence functionality; fourth, determining what makes some reference sources in- appropriate for the reference collection; and fifth, clarifying administrative re- sponsibility for the collection. Each of these steps is elaborated on below. 1. What is a Reference Source? The most common definition of a ref- erence source that appears in the pub- lished collection policies follows that al- ready cited from the ALA Glossary: a book designed to be consulted for defi- nite items of information or a book whose use is restricted to the library building. 15 The latter is, as Bates says, an administrative definition; an item is "reference" because we put it in the ref- erence collection. 16 Such administrative definitions are not particularly helpful in determining collection criteria. More helpful is the first definition. A book is a reference book because it has particular features-not defined in the Glossary- that encourage a certain form of use, i.e., consultation instead of consecutive reading. Some have argued that a reference book is not defined by intrinsic features, but that any item a patron uses for con- sultation in order to find discrete bits of information can be considered a refer- ence book. 17 Certainly librarians know that users will employ almost any source in the library to find the facts they wish to locate. But a definition this relative says little about why we put particular books in reference and others in the stacks. Following this logic, we might as well categorize reference books as tools that hammer in nails or prop open doors, for some users will do these things with books. Instead, we recog- nize intuitively that some sources work much better for reference tasks than oth- ers. That is, a reference book is not nor- mally read consecutively, but is con- suited, and a stacks book is more likely to be read in a continuous manner. Something about reference sources dis- tinguishes them from nonreference items. Bates argues that the key feature of a reference book is that it consists largely of files-that is, it contains records, or in- formation "individuals"-which are or- dered according to some principle. 18 These records can be further broken down into fields, or units, of data. For example, both an almanac and a bibliog- raphy arrange records into particular or- ders offering various access points. The arrangement of data into a file structure enhances the reader's ability to find a certain piece of information more quickly than would browsing through texts of continuous exposition. This definition fits with the intuitive rationalizations often heard for putting an item in reference. For example, most . librarians would classify an annotated bibliography of 200 pages with an intro- ductory essay of about thirty pages as a reference book. But most would send a 200-page essay followed by a thirty-page annotated bibliography to the stacks. Why? Wouldn't both work well for find- ing a number of key references on a topic? Yes, but the former seems most suited for reference and the latter for cir- culation, precisely because the first is primarily bibliography and the latter is primarily text. That is, the first consists predominantly of files and the latter of continuous exposition. Librarians ex- pect users to do lengthy, continuous reading only from the latter. - There are exceptions for placing books in reference, though Bates' empirical study demonstrates that these are few. Some are placed in reference for admin- istrative reasons, most notably theft- prevention. But her data demonstrated that most of the exceptions could be cat- egorized as "authoritative texts": U.S. or state law reporters, the Bible, classic histories and treatises. 19 Placing these volumes in reference appears to reflect the view that these titles are so impor- tant that users would be greatly incon- venienced by their absence from the li- Lean Reference Collection 83 brary. Also, in most cases, it is not expected that users will want to read them cover to cover. Instead, users will look up a particular Bible passage or U.S. court decision, or use the index in a monumental history to locate a key date or bit of biographical data. Thus, the ref- erence librarian sees these sources as be- ing used for consultation more than reading and can justify their location in reference. Bates' study is excellent in deriving a definition of the intrinsic nature of these reference books. However, two points need further discussion. First, she states that 90% or more of all reference books could be accurately assigned to refer- ence collections-even by a clerk- assuming that the person making the decision understands how to judge a book's file structure. This statement as- sumes that all reference-format books belong in reference collections. But this obviously conflicts with the practices (and the few policies) of most libraries. 20 Certain reference books routinely are placed in the circulating collection: su- perseded and outdated editions, highly specialized bibliographies, texts in lan- guages rarely encountered by the li- brary's clientele, etc. Second, Bates focuses on reference books. Her discussions do not explicitly include other media of information used by reference librarians and patrons. Yet a growing number of libraries use multi- ple sources, most notably machine- readable files, for solving reference que- ries. Online databases and CD-ROM indexes, even online catalogs, serve as reference sources on a daily basis at many desks. Like reference books, data- bases are built with a logical file struc- ture, composed of ordered records ar- ranged in such a way as to facilitate consultation for rapid retrieval of bits of information rather than continuous reading. These online sources thus fit the definition of a reference source and should be evaluated alongside other ref- erence sources. However, different methods of bud- geting for automated sources and, fre- quently, different ways of staffing for 84 College & Research Libraries their use sometimes exclude consider- ation of these sources from the normal reference collection management pro- cess. Sheila Intner calls attention to two of the many problems which can occur when these different types of reference sources are not considered together. First, the important issue of whether an automated database can serve the users' needs more effectively or cheaply than printed sources is ignored or avoided. The question now is which of these reference-format sources belong in the reference collection and which should be sent to the stacks or not be pur- chased. Second, the automated sources often be- come physically separated from there- mainder of the reference collection. This separation frequently occurs because of wiring problems or the convenience of sufficient space for the terminals. Thus, users attracted to the more flashy auto- mated sources may come into less con- tact with reference books-and refer- ence librarians. 21 This arrangement risks either misleading users into believing that all of the important tools are com- puterized, or at least keeping them from learning about the variety of other sources that libraries can bring to bear on their needs. Given that some agreement on what constitutes a reference source exists, the question now is which of these reference-format sources belong in the reference collection and which should be sent to the stacks or not be purchased. The crucial factor is how a potential source will be used. Does that use war- rant adding it to the reference collection? 2. Uses of the Reference Collection Patrons may use the reference collec- tion independently, or they may be re- ferred to particular sources by a refer- January 1991 ence librarian. In either case, patrons use the collection in place of the remain- der of the library's collection because they think beginning in reference seems to be more convenient or efficient than wading through the stacks. This follows logically enough from the nature of the sources housed in reference-items de- signed for quick consultations. Many authors of reference textbooks classify reference sources into two basic categories according to the type of an- swer they provide the user: those that give the information outright, and those that tell where the desired infor- mation may be found. 22 The former are often called II fact books, II for they are most often compilations of facts in de- fined areas of interest. Of course, they may be in formats other than print, such as a CD-ROM encyclopedia. Fact books, using the reference file format, permit the user to find these discrete bits of data more efficiently than would be possible in other sources in the circu- lating collection. The second type of reference source is what we might call a ''pointer,'' for it di- rects the user to another source which contains the information sought. These sources are often called ''bibliographic'' tools, for they most frequently contain bibliographic references to items on a given topic, by a particular author, or housed in a certain collection. Periodical indexes (paper or electronic), author bib- liographies, and the library's catalog are examples of this kind of source. Nor- mally, users of bibliographic sources have at least one extra step in their search for information than do fact book users; they must look up the citation, then locate the referenced item itself. In spite of the difference in search pro- cedures, users of both kinds of reference sources are following similar methods. They are using reference sources as sur- rogates for other information sources. Fact sources summarize and abstract data originally published in other mate- rials. Bibliographic tools also contain surrogates; for example, a citation which appears under a subject heading in a pe- riodical index is standing for the original source, alerting the reader to the origi- nal's attributes. Surrogates are not in all cases neces- sary, for a user could find the desired in- formation by browsing the stacks and never using reference sources. Many us- ers do, in fact, search in such a way. But growing collections become increasingly difficult to use when searching for a par- ticular bit of information. Most people probably do not need a catalog for their home libraries; they can find what they want much more informally. But a col- lection of hundreds of thousands of vol- umes cannot be effectively approached the same way. Thus, as collections in li- braries grew geometrically during the last century, a profusion of reference sources have also been created to serve as guides to these large collections. The library catalog is one tool which has served almost solely as a guide to the local collection. 23 However, other refer- ence sources have always included ref- erences to materials not owned by the lo- cal library. A reference collection alerts the user to the wider information uni- verse, not just what can be found on lo- cation. UCLA's collection development policy succinctly defines this role by in- dicating that the reference collection is the "key" to holdings of the local branch, the university's other libraries, and other libraries in this and other countries. 24 The level at which the local reference collection supports access to this wider world of information will vary, however, with the basic philoso- phy of reference service and with the budget. Unlike at UCLA, for example, a small college library staff may determine that undergraduates will do just fine by using only the items owned by that li- brary. Thus, they may acquire many fewer reference sources, avoiding those which list materials held outside their building. But whatever the collection philosophy of the library, the reference collection must still meet its role of being an efficient guide to the larger information universe of which it is a surrogate. Convenience and ease of use by users-patrons and li- brarians alike-are frequently cited as ob- Lean Reference Collection 85 jectives of the reference collection. 25 If the collection becomes too complex and un- wieldy, users will take longer to find what they need. In the worst cases, they will not use the collection, will stick with only a few known sources, or will be de- feated in their search for information. In fact, several studies indicate that users operate on the principle of ''least effort''; that is, they will do as little work as neces- sary to find information and will often prefer information with less pertinence or authority if it is more easily available than ''better'' information. 26 Consequently, li- brarians must balance comprehensive- ness of coverage with ease of use and avoidance of excessive complexity. 3. Characteristics of Sources Belonging in Reference Books and other media sources chosen for the reference collection should be evaluated in the same manner as other acquisitions, taking into account such factors as general quality, reputation of author or publisher, and suitability for users. But the previous discussion has indicated that reference sources have a particular character and are put in refer- ence to facilitate the rapid finding of facts or references. Based on this find- ing, the following characteristics should be met by titles to be added to this collec- tion. Librarians estimate that more than half their reference collections are not used in any one-year period and almost one- third are not used in five years. A. Reference format. As discussed above, sources predominantly struc- tured into files allow easier and quicker consultation. Those with mainly contin- uous text usually do not belong in the reference collection unless a case can be made for their convenience of use. B. Frequently used. Librarians esti- 86 College & Research Libraries mate that more than half their reference collections are not used in any one-year period and almost one-third are not used in five years. What, then, are these materials doing in reference? Surely items used less than once per year do not need to occupy space in a collection se- lected precisely to provide quick and convenient access. The reference stacks themselves are a file, an organized set of volumes which lead to facts or citations. Cluttering this file with rarely or never used sources merely dilutes the effec- tiveness of the remaining useful sources. 27 Richard Dougherty writes that ARL libraries need to start weeding the dead wood out of their circulating stacks so that users may more easily find the materials they are likely to use. 28 This is much more important for a reference collection. Even someone who has pro- vided lists of many different reference ti- tles, Constance Winchell, reminds us that the ''most important element in the equipment of such a department is an adequate and live collection of reference books ." 29 Many of the author's colleagues find frequency of use to be a debatable crite- rion. They argue that a reference collec- tion would not be complete without cer- tain sources, even those that fail to receive even occasional use. Yet most li- brarians often follow this frequent-use guideline. For instance, a good dictio- nary of Texas slang would probably be put into reference at a library in Texas, regardless of the size of the collection. But a small Midwestern college library would probably be less likely to do so, or perhaps would not even purchase such a volume. The determining factor is the difference in amount of expected use; the Midwestern school is probably less likely to see this item used. Implicitly, li- brarians do, in fact, use this criterion on occasion. What is needed is a more con- sistent and explicit application of it, both when a source is selected and when it is reviewed for weeding. 30 Another argument contends that pro- viding reference sources on subjects not well-represented in the general collec- tion is cheaper and easier than trying to January 1991 enlarge the circulating collection in those areas. 31 Certainly the availability of items through interlibrary loan gives merit to purchasing such secondary sources. But this does not support plac- ing them in the reference collection. On the contrary, if the subject is so periph- . eral that only a rare patron uses the sec- ondary source and the collection is weak in that subject, then it would be more beneficial for that patron to have lengthier access to the source, which cir- culation would aid rather than hinder. This guideline assumes that librarians can accurately judge or measure how much use the materials in reference re- ceive. Certainly, it would be preferable to gather firm data on the use made of each item in the collection, as is some- times possible for circulating items. A barcoded reference collection could pre- sumably allow each item to be scanned before reshelving, thus creating the data for this analysis. More probable is that li- braries which have done little previous analysis will need to do a detailed, man- ual use study or to have the reference li- brarians examine the collection and con- fer about perceived use patterns. Many writers have discussed the problems associated with attempting to gauge the use of library materials. 32 For example, scanning items before reshelv- ing may underrepresent true use be- cause many users will place the books back on the shelves themselves. How- ever, unlike the circulating collection, the appropriateness of materials in the reference collection is frequently moni- tored by reference librarians who use the collection and refer patrons to it. If use- ful sources are removed from the refer- ence collection, staff will receive prompt feedback about the mistakes, whether from their own searching or from patron comments. Consequently, even though some disagreement can be expected among the reference staff about per- ceived use, a surprising degree of con- sensus about those sources which have proved useless will be possible. C. Authoritative. As mentioned above, patrons often see items in the reference collection as more authoritative than other sources. Further, this collection is often the sole source for librarians at- tempting to help their users. Thus, sources located in reference must be worthy of this reliance. In fact, this seems so obvious that some librarians wonder why it should even be necessary to mention it. 33 Yet this criterion is often ignored. For example, The Gourman Re- port, a source which has received nega- tive re~Jiews for its unreliability, is rou- tinely added to reference collections and referred to by librarians. 34 No doubt many would respond that no other source fills its niche. Besides, its name recognition lends it a sort of cultural au- thority. But, if the critiques are valid, wouldn't a more prudent course be to leave the niche unfilled and explain to users the lack of dependable ratings ser- vices for colleges? If users will frequently choose an inferior, but more available, source of information over a more de- pendable source which is harder to ac- cess, selectors must be more active in their reviewing of potential reference sources and more choosy about select- ing titles. D. Current. Again, currency seems to be an obvious characteristic, one which countless policies and articles mention. However, Schlachter noted the fre- quency with which outdated editions re- side in reference collections, and Engeldinger reported on the rarity of weeding in most collections. Of course, date of publication alone does not deter- mine a volume's currency; its informa- tion may be as current as is needed. But the many handbooks and directories present in any collection mislead users when their data are several years old. Users may assume that the most current information is in those outdated sources, when dozens of other sources in books, articles, or online databases may be more recent and accurate. Further, past volumes of periodical in- dexes or serial bibliographies are usually kept in reference. However, many of these volumes do little more than gather dust. Few users (especially students) will search more than the most recent few years of an index. Past years of in- Lean Reference Collection 87 dexes which provide coverage of con- temporary reactions or styles, such as the Readers' Guide and New York Times In- dex, still remain quite valuable. But twenty-year-old volumes of Biological Abstracts, for example, could be moved to the stacks and perhaps be designated '' noncirculating. ''35 E. Provides unique coverage. Sources should offer a distinctive contribution to locating information within their subject scope. If a source is superficial or dupli- cates other sources already in reference, there is little reason to add it to refer- ence . We may all be guilty of purchasing the newest subject dictionary or statisti- cal handbook for reference even if it du- plicates others already in the collection. Of course, sometimes a particular sub- ject area receives so much use that it be- comes necessary to provide either more copies of the source or alternative sources. But the collection need not be cluttered (nor the budget encumbered) by purchasing sources which add noth- ing new to what is already there. An ex- ception here would be sources which, by their nature, contain a fair amount of bias or a particular slant. For instance, Bible dictionaries written from evangeli- cal Protestant, Catholic, and liberal Prot- estant perspectives may vary considera- bly in their definitions for certain concepts. In this case, proper coverage in the reference collection may require representative dictionaries from several major perspectives . 4. Types of Sources that do not Belong in Reference Obviously not meeting the character- istics listed above would provide rea- sons for placing a source into the circu- lating collection instead of reference, i.e., not in reference format, or rarely used, unreliable, or redundant. How- ever, librarians, citing II oral tradition, 11 frequently mention a few other justifica- tions for placing sources in reference. Some of these arguments are critiqued below . A. Requires instruction by librarian. This criterion seems to imply that because a reference source is difficult to use, it 88 College & Research Libraries should be near staff members who can explain its use to patrons. But if the source is very difficult to use and is thus rarely consulted, it need not take up space in the reference collection;' it does not serve well the purpose of more effi- ciently aiding users' research. However, if the source is essential and frequently consulted, it should be added to refer- ence anyway. It can also be argued that the stacks are full of sources which are difficult to use, but patrons are on their own with them. If this criterion means that sources frequently mentioned in in- structional sessions (and thus likely to be used) should be located in reference, then their usefulness and authority, not their difficulty of use, justify placing them in reference. B. Protection from theft or mutilation. Occasionally a source is put in reference because it either has been stolen or muti- lated in the past or that likelihood is sus- pected. Placement in reference presum- ably offers added protection. However, a reference collection created to enable efficient research and quick fact-finding does not function as well when it must dilute its file structure to become a safe haven for expensive materials. In almost all academic libraries, materials which need greater supervision than normal can be placed on reserve circulation. Theft-prone items are better placed on reserve (or in special collections rooms) than in reference. Many libraries locate heavily-used (and thereby theft-prone) reference items in a "ready reference" collection. This approach is valid when the items al- ready fit the reference criteria discussed earlier, but it is a waste of space and li- brarians' time (for retrieval, holding ID cards, etc.) when the sources are only there to prevent theft. C. Consistency of location. All things be- ing equal, most of us would prefer to put like sources in like locations. When we begin to purchase a monographic series of bibliographies, for instance, we often automatically add each new volume to the reference collection. However, this policy can easily lead to a bloated collec- tion. Many of these volumes may con- January 1991 cern subjects rarely pursued by the li- brary's clientele or may be less valuable than others in the same series. Each ad- ditional item in a series should be con- sidered individually according to the ac- cepted set of criteria. An exception to this would be a series in which one vol- ume refers to other volumes in the same set, such as the Dictionary of Literary Biog- raphy. D. Fills a niche. Reference librarians usually applaud a new publication which covers an area previously ne- glected. Impressed as we are by well- made reference tools, we tend to put these items into the reference collection because "we don't have anything else on that subject." However, most sub- jects will be covered in a general refer- ence collection by at least the broader sources, such as encyclopedias and guides to the literature. The new source should be evaluated by the other crite- ria, especially expected frequency of use. A source filling a small niche, how- ever nicely done and unique, should be sent to the circulating stacks if very little demand is expected. E. "Classic" source. Finally, there are those sources which one or more librari- ans (or sometimes patrons) believe must be in any quality reference collection. Some of these tools have earned this sta- tus by years of productive service to li- brarians, and they may indeed be essen- tial acquisitions. Others exhibit superior writing or editorial design. But each li- brary's collection serves a different set of user needs, and a source much used at one institution may have little use at an- other. If it is rarely needed, it probably should be located in the stacks collec- tion. These titles should be purchased for the general collection if they fit into the library's collection policies. If they become more valuable as a curriculum changes or subject interests shift, they can at that time be moved to reference. 5. Administrative Responsibility for the Reference Collection A perusal of reference collection de- velopment policies shows that various systems of selecting and supervising the collection are practiced. Rebecca Kroll lists a few of the people that usually se- lect the reference sources for the collec- tion, ranging from the subject bibliogra- pher, who does not work in reference, to the reference subject specialist. She also lists the people who most frequently function as the managers of this process, including heads of reference, coordina- tors for reference collection develop- ment, and committees of reference li- brarians. 36 However, both the noted lack of col- lection development policies for refer- ence and the comments of many col- leagues indicate that oversight for the reference collection is often lacking or poorly defined. A major reason for this confusion is that reference sources may be classified both as subject materials and as reference-format sources. Be- cause most academic libraries divide re- sponsibility for materials selection by discipline, overlaying further responsi- bilities by type of material can be diffi- cult . Many problems result from a lack of careful management. First, the assignment of selection for a discipline to just one individual bibliog- rapher allows the idiosyncrasies of that selector to color the makeup of the col- lection. Imbalanced selection, especially excessive additions to areas not calling for such build-up, is possible without oversight. Second, sources which do not fit neatly into a division of responsibility by subject disciplines may be missed by selectors who assume another librarian will order them. For example, interdisci- plinary items and general items (such as almanacs and biographical sources) of- ten defy easy categorization by subject. Third, selection may occur in the ab- sence of feedback concerning the use made of the collection by patrons and staff. This situation is especially likely when librarians who may spend little or no time with the library's clientele, such as subject bibliographers or heads of ref- erence, are responsible for adding to and weeding out the reference materials. The first two problems both stem from the lack of someone overseeing the en- tire reference selection and deselection Lean Reference Collection 89 process. The most obvious solution would be the appointment of an individ- ual or committee to ensure that the selec- tors properly address the entire scope of the reference collection and that the col- lection is balanced within the framework of a reference collection development policy's goals. This person or committee must have the authority to allocate funds directly from a reference account and also to question perceived errors of selection or weeding by the various ref- erence selectors . This position need not have the power to prevent the other se- lectors from ordering items they would like to add to reference, but it should have the ability to refer the discussion of adding these materials to an assigned point of resolution (e.g., the head of ref- erence). Whether a committee or a designated li- brarian has the authority for these deci- sions is not as important as is the explicit designation of someone to take charge of the process. Libraries cannot continue to allow a variety of individuals to make haphazard decisions about what the ref- erence collection should be. The third problem exists when the persons making decisions about refer- ence materials have little knowledge of the uses made of the collection. In aca- demic libraries with several staff mem- bers serving at the reference desk, it is impossible for any single librarian to have a complete picture of the use made of the collection. Thus, it is important to set up a formal mechanism for discus- sing appropriate titles to add or weed. The best scenario would include regular reference staff meetings to consider indi- vidual reference sources for addition or deselection and to voice staff experi- ences in working with patrons who use those types of materials. These meetings would not only provide better feedback about selecting good reference sources, 90 College & Research Libraries but would also serve as a form of contin- uing education. If staff size makes this an unwieldy process, a committee of several librarians with varied subject ex- pertise might be the best alternative. Whether a committee or a designated librarian has the authority for these deci- sions is not as important as explicitly designating someone to take charge of the process. Libraries cannot continue to allow a variety of individuals to make haphazard decisions about what the ref- erence collection should be. The refer- ence collection must be actively man- aged to meet the goals and objectives determined by the reference depart- ment. CONCLUSION Reference collections serve as subsets of, or surrogates for, the larger world of information and thereby permit users to search for the information they desire in a more efficient and convenient manner. This situation requires reference to be a highly functional collection of appropri- ately chosen sources, including as many as necessary to meet normal user needs while avoiding excessive complexity. January 1991 However, the reviewed literature dem- onstrates that librarians admit to the existence of cluttered, overgrown collec- tions which are all too infrequently guided by well-reasoned collection de- velopment policies. Consequently, it is imperative that those librarians responsible for selection and management of reference collec- tions do two things: · first, apply more rigorous thought to the criteria for deter- mining what gets located in those collec- tions; and second, review the collection on a regular, systematic, and aggressive basis in order to determine which sources are appropriate and which merely occupy space and dilute the use- ful materials. It is possible that the vast majority of reference collections would be significantly diminished in size if these proposed guidelines were actively followedY This, however, will not di- minish the effectiveness of those collec- tions. Instead, the leaner, trimmer col- lection will become more functional; users will find useful sources more eas- ily; and librarians can reclaim valuable space which can be put to other uses. REFERENCES AND NOTES 1. Lynn C. Hattendorf, "The Art of Reference Collection Development," RQ 29:219-20 (Winter 1989). 2. Mary Biggs and Victor Biggs, ''Reference Collection Development in Academic Libraries,'' RQ 27:69-70 (Fall1987). 3. Ibid., p.74. 4. Ibid., p.70-71. 5. Eugene A. Engeldinger, "Weeding of Academic Library Reference Collections: A Survey of Current Practice," RQ 26:367 (Spring 1986). 6. Biggs and Biggs, "Reference Collection Development," p.74-75. 7. Rebecca Kroll, "The Place of Reference Collection Development in the Organizational Struc- ture of the Library," RQ 25:99 (Fall1985). 8. Gail Schlachter, "Obsolescence, Weeding, and Bibliographic Love Canals," RQ 28:7-8 (Fall 1988). 9. Marcia J. Bates, ''What Is a Reference Book? A Theoretical and Empirical Analysis,'' RQ 26:40 (Fall1986). 10. Heartsill Young, ed., The ALA Glossary of Library and Infonnation Science (Chicago: American Library Assn., 1983), p.188. 11. William A. Katz, Introduction to Reference Work, 2v. (New York: McGraw-Hill, 1969), 1:12, and also 5th ed., 1987. 12. For examples, see Bill Katz, ed., Reference and Online Services Handbook: Guidelines, Policies, and Procedures for Libraries (New York: Neal-Schuman, 1986), V.2, and Paula D. Watson, ed., Refer- Lean Reference Collection 91 ence Policy and Administrative Documents (Chicago: Reference and Adult Services Division, American Library Assn., 1985), p.562-781. 13. Gail Schlachter, "Reviewing the Reviewers," RQ 27:468-70 (Summer 1988). 14. Interestingly, this process mirrors the structure of reference courses in many library schools, in which most of the courses are spent discussing and using particular reference sources, but little time is devoted to analyzing the reference interview, reference management, and other theo- retical issues. 15. Young, ALA Glossary, p.188. 16. Bates, "What Is a Reference Book?" p.38. 17. For example, see Donald Edward Davinson, Reference Service (London: Clive Bingley, 1980), p.12-13. 18. Bates, "What Is a Reference Book?" p.41-43. 19. Ibid., p.51-53. 20. Biggs and Biggs, "Reference Collection Development," p.72. 21. Sheila S. Intner, "A Question of Medium," Technicalities 7:13-14 (Oct. 1987). 22. For example, see Archie G. Rugh, "Toward a Science of Reference Work: Basic Concepts," RQ 14:293-99 (Summer 1975). 23. This is rapidly changing because many library online catalogs now contain bibliographic refer- ences to information, such as commercial periodical databases, not held in the local library. 24. UCLA's policy is reprinted in part in Watson, Reference Policy, p.583. 25. Cf. A. R. Spofford, "Works of Reference for Libraries," in Public Libraries in the United States of America: Their History, Condition and Management (Washington, D.C.: GPO, 1976), p.687-88, and Engeldinger, "Weeding," p.371. 26. See Herbert Poole, Theories of the Middle Range (Norwood, N.J.: Ablex, 1985), p.86-92; Esther G. Bierbaum, "A Paradigm for the '90s," American Libraries 21:18-19 Oan. 1990); and Winifred Sewell and Sandra Teitelbaum, "Observations of End-User Searching Behavior over Eleven Years," Journal of the American Society for Information Science 37:239-40 Ouly 1986). 27. Cf. Duke University's policy reprinted in Watson, Reference Policy, p.610. 28. Richard M. Dougherty, ''Ridding Collections of Deadwood,'' Journal of Academic Librarianship 15:3 (Mar. 1989). 29. Constance M. Winchell, Guide to Reference Books, 8th ed. (Chicago: American Library Assn., 1967), p.xiv (emphasis added). 30. The Georgia Institute of Technology policy, emphasizing this criterion, is reprinted in Katz, Reference and Online Services Handbook, 2:117-18. 31. Ruth C. Hale, "Academic Library Reference Collection Policy Statements: An Informed Con- sideration,'' in Katz, Reference and Online Services Handbook, 2:27. 32. See Daniel R. Arrigona and Eleanor Mathews, "A Use Study of an Academic Library Reference Collection," RQ 28:71-81 (Fall 1988). Classic discussions of use studies include Robert N. Broadus, "Use Studies of Library Collections," Library Resources & Technical Services 24:317-24 (Fall 1980); Allen Kent and others, Use of Library Materials: The University of Pittsburgh Study (New York: Dekker, 1979); and Paul Metz, The Landscape of Literatures: Use of Subject Collections in a University Library (Chicago: American Library Assn., 1983). 33. Fred Batt, "The Detailed Reference Collection Development Policy: Is it Worth the Effort?" in Bill Katz and Ruth A. Fraley, eds., Evaluation of Reference Services, The Reference Librarian, no. 11 (New York: Haworth, 1984), p.319. 34. DavidS. Webster, "Jack Gourman's Rankings of Colleges and Universities: A Guide for the Perplexed," RQ 25:323-31 (Spring 1986). 35 . This procedure is described in the Texas A&M policy included in Katz, Reference and Online Services Handbook, 2:214. 36. Kroll, ''The Place of Reference Collection Development,'' p. 96-99. 37. Cf. the major reduction in the size of a reference collection following a careful review reported in Paul T. Adaliar, Jr., and llene F. Rockman, "Title-by-Title Review in Reference Collection Development," Reference Services Review 12:85-88 (Winter 1984). NEW FROM PENN§YL v ANKA Middle i\ges Series -------- A Tainted Mantle Hercules and the Classical Tradition at the Carolingian Court Lawrence Nees Mar. 1991. 392 pp, 3 color & 88 b/w illus. Cloth, 8216 .. 7, $39.95 From Ad Hoc to Routine A Case Study in Medieval Bureaucracy Ellen E. Kittell i\pr. 1991. 256 pp, 9 illus. Cloth, 3079 .. 5, $35.95 (tent.) Literature Macropolitics of Nineteenth .. Century Literature Nationalism, Exoticism, Imperialism Edited by Jonathan A rae & Harriet Ritvo New Cultural Studies Series. Feb. 1991. 320 pp. Cloth, 8208 .. 6, $34.95 Lyric Contingencies Emily Dickinson and Wallace Stevens Margaret Dickie Feb. 1991. 208 pp. Cloth, 3077 .. 9, $23.95 American Designs The Late Novels of James and Faulkner Jeanne Campbell Reesman i\pr. 1991. 232 pp. Cloth, 8253 .. 1, $25.95 (tent.) Social Sciences Feedback Thought in Social Science and System Theory George P. Richardson Mar. 1991. 380 pp. Cloth, 3053 .. 1, $39.95; paper, 1332 .. 7, $19.95 (tent.) i\nthropology/Folklore Ethics and the Profession of Anthropology Dialogue for a New Era Edited by Carolyn Flue hr .. Lobban Feb. 1991. 256pp. Cloth, 8157 .. 8, $31.95 Rhetorics and Politics in Afghan Traditional Storytelling Margaret A. Mills Publication of the i\merican Folklore Society. Mar. 1991. 288 pp. Cloth, 8199 .. 3, $27.95 (tent.) Law Planning for Serfdom Legal Economic Discourse and Downtown Development Robin Paul Malloy May 1991. 232 pp. Cloth, 3055 .. 8, $27.95 (tent.) International Law and Pollution Edited by Daniel Barstow Magraw i\pr. 1991. 416 pp. Cloth, 3053 .. 1, $37.95 (tent.) History of Science The Literary Structure of Scientific Argument Edited by Peter Dear Feb. 1991. 208 pp, 6 illus. Cloth, 8185 .. 3,$28.95 The Japanese and Western Science Masao Watanabe Translated by Otto Theodor Benfey i\pr. 1991. 128 pp, 45 illus. Cloth, 8252 .. 3, $28.95 (tent.) University of Pennsylvania Press P.O. Box 4836, Hampden Station, Baltimore, MD 21211 Toll .. free: (800) 445 .. 9880