College and Research Libraries The Selection Decision: Defining Criteria and Establishing Priorities John Rutledge and Luke Swindler This article discusses the specific selection criteria used in making collection development deci- sions. The criteria are grouped into six major categories, and within each category they are arranged in order cf relative importance. The proposed schematization provides librarians who have collection development responsibilities with a holistic and explicit model for arriving at a selection decision as well as a mechanism for assigning a specific priority to each selection. Use of such a model can help to rationalize selection decisions; it relates acquisitions effectively and convincingly to a library's fiscal environment; and it promotes cooperative collection develop- ment. · ~nl German proverb states, ''Who- l 'rJA~ ever has the choice, also has the '(i}'-~ ~; misery.'' Making choices is no ~ £ .... ~~ easy business, yet selecting ma- terials is one of the principal functions of collection development officers. Even in libraries that rely heavily on approval plans, selectors must review titles individ- ually to ensure an effective collection de- velopment program. Despite the central- ity of selection decision making to the collection development process, there are few tools that offer practical assistance for the performance of this intellectual task, particularly in academic settings. More- over, the guides found in the published literature are not truly comprehensive, and none provides a practical means of re- lating selection to a library's acquisitions budget. While a collection development officer may be called upon to select or re- ject hundreds of titles during the course of the working day and is generally profi- cient at making these choices, he or she could still benefit from a convenient tool that rationally organizes the factors con- tributing to an acquisitions decision. This would be particularly helpful if funds are insufficient for acquiring all appropriate ti- tles or if the library is seriously attempting to implement cooperative collection de- velopment programs. Selection officers typically receive little assistance from their own library when making decisions. Even the best and most widely known collection policies merely state what a library ideally would select in a world without financial constraints, while a few policies also indicate the exist- ing level of collecting. 1 A unique tool, the Bibliographer's Manual, prepared by the collection development staff of the Uni- versity of Texas at Austin Library, serves as a guide to the collection development system and selection procedures rather than to the decision-making process it- sel£.2 A more general guide used in many libraries is ALA's Guidelines for Collection John Rutledge is West European Bibliographer and Luke Swindler is Social Sciences Bibliographer at the Uni- versity of North Carolina, Chapel Hill, North Carolina 27514. 123 124 College & Research Libraries Development. 3 This otherwise useful hand- book covers the formulation of a collection development policy and the evaluation of the results of selection, but does not deal with the specific selection decision mak- ing that is central to the collection devel- opment process. EARLIER APPROACHES TO SELECTION The task of setting forth the criteria used to select materials for libraries offers an in- triguing challenge and has appealed to many other writers. During the past few years, as the field of collection develop- ment has matured, a number of texts have appeared. As one would expect, certain themes are common to all attempts- quality of the materials, relationship to the patrons, cost-to name only the most fre- quent. Usually, however, treatments of selection decision making are very gen- eral, discursive, and incomplete. More- over, the selection. criteria are often tied to the scope and organization of specific tools. An examination of major works by Ar- thur Curley and Dorothy Broderick, Rob- ert Broadus, and William Katz shows that they all develop some general principles, concentrating, as textbooks must, on broad issues rather than on the intricacies of the decision-making process. When specific selection criteria are discussed, they are treated independently; that is, the authors do not relate the various crite- ria to each other. The reader is left won- dering which criteria are the most impor- tant and when to apply them. Finally, all three texts focus on public libraries. In Building Library Collections Curley and Broderick discuss some of the principles of selection using a series of nine debate top- ics.4 Under each one they present a range of contrasting viewpoints. The discus- sions are rarely prescriptive; indeed, they are not intended to be. Rather, they point to the diversity of opinion on such issues as high culture versus popular materials, catholicity in collecting, and demand as the governing factor in selection. More- over, because the authors' primary inten- tion is to stimulate thought, their princi- ples do not serve as a guide to selection. In his textbook, Selecting Materials for Li- March 1987 braries, Broadus offers a wide-ranging ex- position of the issues and practical wis- dom about selecting books and most other types of library materials. 5 Although not grouped conveniently, the principles he posits are (1) user needs as primary; (2) re- lation to existing collection; (3) relation to other libraries; (4) the sources or pub- lishers; and (5) book-intrinsic criteria such as content, recency, veracity, reputation of the author or publisher, and format. While Broadus is a useful discussion of se- lection, the specific application of selec- tion criteria is not developed. 1 In Collection Development: The Selection of 1 Materials for Libraries, Katz advances a set of ten selection criteria. 6 These are (1) pur- pose, scope, and audience; (2) difficulty; (3) authority, honesty, and credibility of author and publisher; (4) subject matter; (5) comparison of a title to others in the collection; (6) timeliness; (7) format; (8) price; (9) curriculum support; and (10) de- mand. We believe that these criteria touch upon most of the issues. However, as with the other survey texts, the relationship of the criteria to each other remains undevel- oped. Jean Boyer Hamlin, in a contributed chapter in Robert Stueart and George Miller's handbook on collection develop- ment, develops a list of nine selection cri- teria.7 Her factors, paraphrased, include (1) pertinence to areas covered, (2) interest to users, (3) relationship to existing collec- tion, ( 4) cost, (5) patron objections and threat of theft, (6) probable quality, (7) ne- ,. cessity of continuing financial commit- ment, (8) duplication of existing material, and what might be called (9) "bibliothecal convenience" or ease of handling. A more tightly organized grouping would sim- plify the classification, since some criteria are much more narrowly focused than others. In addition, the criteria are neither well developed nor prioritized. To illus- trate practical decision making, Hamlin recasts the criteria into questions that should be asked when selecting books from a dealer's catalog. While these ques- tions can help to refine one's thinking about a particular title, they provide no way of evaluating the results of the exami- nation. Recent journal literature has provided further attempts to delineate selection cri- teria, sometimes more focused on the decision-making process itself. Hendrik Edelman develops a model of decision making based on the organization of li- brary materials by source and type of pub- lication.8 He suggests that to this universe of published knowledge one may apply certain historical-, linguistic-, and geographical-elimination factors, corre- lated to the collection level descriptors. A further distinction is made between selec- tion for short-term goals and selection for long-term goals. The main criteria, accord- ing to Edelman, are established by the col- lection development policy. To this policy one must bring to bear the virtues of ''bal- ance, reliability, and comprehensiveness, in that order. ''9 This article is very general and does not discuss specific criteria, nor how they affect individual selection deci- sions. The suggestions, while accurate in the main, do not yield a guide to microde- cision making; indeed, Edelman's main focus is macrodecision making. In an article on selection decision mak- ing for preservation purposes, Dan C. Ha- zen develops another distinction. 10 Pursu- ing the close relationship between preservation decisions and collection de- velopment decisions, Hazen adduces five criteria that pertain both to preservation and to new-title selection. These are (1) ac- ademic activity or user demand, (2) histor- ical precedent and tradition, (3) the vol- ume and cost of materials, (4) the availability of alternatives to purchase, and (5) discipline-specific models of access to information. These criteria apply best to preservation decisions; for acquisitions they are incomplete and lack specificity. Although most of these criteria are valid, no priority is assigned to them nor is a method of application suggested. John N. DePew presents an explicit model of the acquisitions process, consist- ing of a detailed flow chart with weighted inputs and a formula that results in a selec- tion decision. 11 Although the article does make one aware of the complexity of selec- tion, the criteria are inadequate and not well developed. In addition, they include considerations that should not be rele- vant, e.g., whether the requester will cause trouble or whether the title is a gift. The Selection Decision 125 Finally, the formula he presents is very ar- bitrary and too cumbersome. Ross Atkinson's recent article on what actually happens when collection devel- opment librarians select a title represents one of the few attempts to describe the se- lection process itself. 12 His article is useful for understanding some of the dynamics of selection microdecisions and how crite- ria relate to each other. Atkinson sees the selection process as the interaction of a se- lector with a bibliographic citation; the se- lector resolves the decision by using three contexts, the archival (what is already in the collection), the communal (the re- search needs and interests of the clien- tele), and the thematic (what has been or is being published on the subject). While this article is stimulating, it is primarily a theoretical treatment and not intended to be a practical guide to selecting. It is also too concentrated on the bibliographic cita- tion itself to serve as a comprehensive guide to selection decision making. Our review of the literature has not un- covered a practical and holistic model for microselection, the selection of materials on a title-by-title basis as is· done in li- braries every day. We see the need for the provision of a comprehensive and practi- cal model that has a high level of applica- bility to any selection decision. The model presented in this article not only delin- eates and defines the appropriate criteria, but also displays them visually to show the relationship of the criteria to each other. In addition, we provide a numerical rating system to allow the librarian to rank each title and thereby relate selection to available funds. EVOLUTION OF THE MODEL The model consists of the criteria dis- cussed below, coupled with a priority sys- tem, which grew out of the tasks and op- portunities faced by collection develop- ment librarians at the University of North Carolina at Chapel Hill (UNC-CH). Col- lection development for the main library system at UNC-CH was totally reorgan- ized during the mid-1970s, when selection became a library responsibility. At the same time the university and library ad- ministrations greatly increased funding for acquisitions. These two developments 126 College & Research Libraries created expanded collection development possibilities, making selectors acutely aware of the selection process and the ra- tionale for selection decision making. Since then, occasional budgetary reverses have given the selectors experience in the painful application of a triage system, in which all selections were classified as first, second, or third priority. With the reorganization of collection de- velopment, long-standing cooperative ac- quisitions programs with neighboring re- search libraries took a new lease on life. 13 This priority system has proven useful for cooperative collection development, par- ticularly in facing the practical necessity of deciding which specific titles should be ac- quired locally and which should be ac- quired by the library holding the primary collection commitment. Within the con- text of broad cooperative agreements, a common understanding and an explicit statement of selection criteria encourage collection development librarians to think in terms of priorities and thereby to for- mulate cooperative programs with a large measure of specificity. REDEFINING THE SELECTION CRITERIA If, as Atkinson says, selection is difficult to describe, 14 it is also difficult to prescribe. The selection criteria presented below are the results of years of wrestling with the intellectual issues involved in the selec- tion of library materials. We have tried to include all the relevant factors involved in selection decision-making. For the sake of clarity we have arranged the factors into six internally coherent and, insofar as pos- sible, mutually exclusive categories, thus avoiding redundancy. At the same time the arrangement indicates relationships between the criteria. Each of the criteria causes the selector to ask specific ques- tions about any given title; in answering the questions the selector brings objective information to bear on making the deci- sion. Subject Subject constitutes the first and most important selection factor. Since all books and other library materials are about something, both collection development March 1987 policies and ·staff are typically organized along subject lines. The selector initially discerns the subject of a work, e.g., France-History-Revolution, or Science Fiction, then evaluates the item in terms of the information or knowledge universe. At the same time, and perhaps more im- portantly, he or she attempts to relate the item under consideration to the programs at his or her institution. How well the item supports institutional objectives and pro- grams is the paramount consideration, but it is always seen in relation to the larger intellectual universe. Indeed, un- less one knows the subject of a work, a ra- tional selection decision cannot be made. Intellectual Content In actual practice, especially when selec- tion decisions are made from a biblio- graphic citation, it is sometimes difficult to arrive at an informed estimate of the intel- lectual content of a work. Nevertheless, the question must be addressed. In assess- ing intellectual content, one asks how it relates to what has gone before. Is the work a key title in its field, whether a great work of literature or a seminal study? How valuable have the author's past contribu- tions been? Is it "raw data" of the field such as statistical tables? How authorita- tive is the work and what is the nature of its contribution? Is it narrowly focused or a general essay? Or is the primary concern propagandistic? High intellectual content alone cannot determine the selection deci- sion; nor can trivial or polemical works be rejected automatically, since they some- times become the subject of research. Potential Use Having considered the work in terms of its subject relevance and its intrinsic intel- lectual integrity, the selector next reviews it in the light of his knowledge of the pa- trons' needs. Potential use is considered only after subject and intellectual content have been ascertained, in order to ensure that appropriate, quality materials are added. One must know what a work is about and something of the nature of its contents before one can predict level of use. What is the likelihood that the item in question will be used? What level of use justifies its acquisition? The selector ... l should know of urgent research interests and be able to predict probable interest based on knowledge of the course offer- ings, research programs, and circulation patterns. There is also the category of broad, general information, material to which educated men and women will want to have recourse. These works ' are desirable but must take second place. At the lower end of the range are works that, because of considerations such as style or prerequisite knowledge, are deemed less accessible. From the bibliographic citation the selector can usually determine whether one use in fifty years would be the maximum expected. Relation to the Collection This factor echoes the concerns dis- cussed under subject. Whereas subject re- lates an item to the information universe, here an item is scrutinized in terms of its relation to a specific library's collection. Typically the questions asked here are ones that will be posed by librarians, who, generally, are trained to look at the integ- rity of the library's collection, to fill in gaps in the collection, to establish balance and comprehensiveness in the collections, and to maintain cooperative programs with other libraries. Here there is an inherent tension: how to meet current user de- mands and yet continue to build upon his- toric collection strengths and specialties. Bibliographic Considerations To a certain extent you can judge a book by looking at its cover. Bibliographic con- siderations parallel those criteria found under intellectual content. The interre- lated issues of publisher and format fur- ther refine the selection. The reputation of the publisher or sponsoring agency and the type of publication or format of the work both play an important role in any selection decision. Obviously these fac- tors require some knowledge of the book trade. The highest priority within this cat- egory is assigned to the titles of distin- guished publishing houses that over the years have built up a reputation for excel- lence. At the other extreme, there is a lush undergrowth of "quasi-publications" such as working papers and research re- ports. Between the two poles is a wide The Selection Decision 127 range of specialized publishing as well as the output of the major trade publishers, each of which has a reputation for quality and subject specialization. Language Language is a criterion because it speaks to the issue of potential use, yet it is dis- tinct from it. Language also relates closely to the topic of the work. The major work- ing language of a given field deserves spe- cial consideration. Similarily, the second working language of a topic will have to receive a relatively high priority. In some cases the major language of the topic will not be English. If the major language of a field is Italian, the selector cannot exclude an item in Italian if the topic is central. Does the language of the item augment or detract from the capacity of the work to in- form? Travel guides in the vernacular, for example, may convey a fuller understand- ing of the country than English-language editions; on the other hand, a foreign- language book on computer science has little capacity to inform an English- speaking audience. Perhaps the foreign- language item helps to educate the poten- tial user about areas not well covered in the English-language press? Or, would the language of the item have to be consid- ered distant from its topic, e.g., Italian- language studies of Czech literature? Fi- nally, hoping to gore as few oxen as possible, we recognize that some lan- guages are less central to scholarly inquiry than others, although factors such as user interests can cause the item to receive an overall higher ranking than the language factor alone would indicate. Cost as a Nonfactor Although many writers include cost as a factor, price is irrelevant to making a selec- tion decision as distinct from a purchase decision. We agree with Atkinson that "the budget should be viewed not as a cri- terion for selection but rather as an influ- ence upon the relative extent to which se- lection criteria are acted upon. " 14 While high cost typically results in more care be- ing taken in making the selection decision, the priorities-those items that the library must have, should have, or could have- do not change in response to budgetary 128 College & Research Libraries limitations; they remain the same, whether money is available or not. Fur- ther, it should be noted that librarians generally select titles within specific budget lines, e.g., new subscriptions, au- diovisuals, expensive titles, current books. As a consequence, a costly micro- form collection or multivolume set does not compete against a current book or a new journal subscription but against other possible expensive purchases. One can therefore use the proposed model to de- termine the relative priorities among a group of expensive titles. Finally, just as high cost should not influence a selection decision, low cost or bargain prices also should be irrelevant. If a title is ranked at priority three, its ranking does not change in response to the offer of a discount. Only when two items of equal ranking are being considered for acquisition in times of fi- nancial adversity might price determine which is actually purchased. USING THE MODEL In choosing a chart or tabular form of presentation, we are attempting to pro- vide selectors with a succinct, rapidly scannable tool for guidance in title-by-title or microselection (see table 1). Although the model presented was developed in an academic library for the rationalization of book selection, it could be adapted with only slight modifications by any type of in- stitution and can be applied to selection of all types of library material. When deciding whether to acquire a ti- tle, a selector usually considers many fac- tors of varying importance within the con- text of the inclusions and exclusions of a collection development policy. The factors that we consider most relevant are grouped into six columns. Each column contains a discrete set of criteria, made specific by descriptive phrases. The fac- tors affecting selection are (1) subject, (2) intellectual content, (3) potential use, (4) relation to collection, (5) bibliographic considerations, and (6) language. More- over, because some selection factors are more significant than others, the colum- nar sets are presented from left to right in descending order of importance. Al- though each column represents a distinct and internally coherent set of criteria, the March 1987 second three columns echo respectively the first three. Using this model one evalu- ates each title horizontally in terms of the six selection factors and vertically in terms of its rank within each column. Within each column the criteria are listed in descending order of importance. We have divided the columns to create three basic levels of priority: (1) the library must have the item: the title is essential and is the first to be reviewed against available funding; (2) the library should have the item: the title is an important ad- dition to the collection, and users could reasonably expect to find it in the library; and (3) the library could acquire the title: al- though peripheral to the collection, the ti- tle is appropriate and there is a possibility that it will find a user. There always will be a subjective ele- ment to selection and evaluation: it is an art-not an exact science. Nevertheless, there may be some circumstances in which one wishes to weigh each title quantitatively for the sake of comparison. In such instances we propose a method of assigning a relative value to each criterion considered in the evaluation (see table 2). One can thereby derive a numeric rating for each selection. It seems simplest to set up the selection values so that the highest possible score totals 100 points. The cu- mulative score is the total of the values as- signed by the selector in each of the six columns. The first factor, subject, carries the high- est number of points: any item being eval- uated receives a score of 1 to 30 points for subject. Intellectual content ranks just un- der subject with a slightly smaller range of 25 points possible; similarly, other factors receive proportionately fewer points. If a title receives 67 to 100 points, it ranks as a first priority; if a title receives 34 to 66 points, it scores as a second priority; and if 1 to 33 points, it equals a third priority. The decision to assign a specific score of 16 points, rather than 25, for subject within that allocated range will be deter- mined by subjective judgments that can- not be eliminated from the selection pro- cess. Herein lies the art of selection. We also believe that some libraries will wish to change our weighting of the criteria. This can be simply done and would allow the 1 TABLE 1 SELECTION CRITERIA Subject Intellectual Content Potential Use Relation to Collection c?~~~::fe:!t~~s Lan~ase First Directly supports Key work in field Known research or Central to existing Distinreished Major language(s) of Priority pros_rams or Key author program interest collection publis er topic mstltutional Major critical study Patron request, Closely related Significant .English and second emphases Substantial new based on need Provides specialized sponsoring body working language(s) Major field of contribution to Probable need, based information about a srecialized publisher of topic ~cho!arship or learning on known interest central strength o high quality Major foreign mqurry Necessary to Major trade language accessible intellectual integrity publisher to users Second Ancillary to General essay General interest Develops existinftt Specialized publisher Treatment in foreign Priority programs Narrowly focused Title recommended collection strengt Published lan~age of topic not Specialized topic work by patron, without Historic collecting dissertation we covered in Minor field of Narrow intellectual specific need strength Popular publisher English ~cho!arship or perspective Immediate use Foreign lanrage unlikely treatment o mqurry Popular treatment local/national issues ~ =-Third Tangential to Raw or unedited Presents problems of Completes serial, Research report Foreire language ttl Priority programs material accessibility series, or set held Unpublished perip eral to topic or Ul ttl Mar~inal area of Marginal or Infrequent use Very specialized dissertation user ;" n scho arship or polemical work material Working papers Treatment in foreiBn -..... inquiry Trivial literature Assigned to lans_ua~e of matenal Q Pamphlets = Propaganda cooperative partner Ephemera availab e in English 0 Language accessible ttl to tiny minority of n ..... Ill likely users cs· = 1-1 N 1.0 130 College & Research Libraries March 1987 TABLE2 j J SELECTION VALUES Intellectual Relation to Subject Content Potential Use Collection ~~~~i':fer~~i~s Language (30 points) (25 points) (20 points) First 30 25 Priority 29 24 28 23 27 22 26 21 25 20 24 19 23 18 22 17 21 0 Second 20 16 Priority 19 15 18 14 17 13 16 12 15 11 14 10 13 9 12 0 11 0 Third 10 8 Priority 9 7 8 6 7 5 6 4 5 3 4 2 3 1 2 0 1 0 results to reflect local emphases more closely. In any case, such a point system could then be correlated with available funding to determine which selections are actually purchased. By no means do we wish to suggest that every title needs to be treated with such mathematical precision. A library could limit its use to certain classes of materials that require particular selection care (e.g., reference titles); special formats (e.g., au- diovisuals); expensive items (e.g., books costing more than $100); or critical budget lines (e.g., new subscriptions). Further- more, since the first three criteria are more heavily weighted to reflect their greater importance, it will not always be neces- sary to go through the entire grid in order to arrive at a title's relative priority, partic- ularly since the second three factors echo the first three and carry far less weight. 20 19 18 17 16 15 14 0 0 0 13 12 11 10 9 8 7 0 0 0 6 5 4 3 2 1 0 0 0 0 (12 points) (8 points) (5 points) 12 8 5 11 7 4 10 6 0 9 0 0 0 0 0 0 0 0 0 0 0 0 0 o · 0 0 0 0 0 0 8 5 3 7 4 2 6 3 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 2 1 3 1 0 2 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 IMPLICATIONS OF THE MODEL The proposed model will find its great- est utility in adding precision to and en- hancing the consistency of the individual selection decisions. Consequently, use of this model can improve the quality of a li- brary's collection development efforts. Moreover, because the schema presented .1 here is comprehensive, holistic, and ex- plicit, it can serve as a practical guide for training all selectors in a library and can - ( / aid in the rationalization of selection deci- · ~ sion making throughout a library system. If all selectors have a similar understand- ing of the criteria to be used and their in- · - terrelatedness, then unintended dispari- -~ ties in collecting levels can be minimized . . By extension, the model has applicabil- ' ity to cooperative collection development · programs. In our experience we have found that cooperation with other li- braries is most practicable when lower pri- ority materials are unper discussion. Be- cause no library is lil