perrault.p65 316 College & Research Libraries July 1999 The Effects of High Median Age on Currency of Resources in Community College Library Collections Anna H. Perrault, Richard Madaus, Ann Armbrister, Jeannie Dixon, and Rhonda Smith In 1998, a comprehensive study was conducted of the monograph col­ lections of the twenty-eight public community colleges in Florida. This article reports the findings of that study with respect to median age and currency of resources. The rationale for the interpretation of the findings in the Florida Community College Collection Assessment Project is congruent with the philosophy that college collections should emphasize the instructional and curricular needs of students and, there­ fore, that the collections should emphasize current resources rather than retrospective depth. The findings on median age from the Florida Community College study serve as an example for the discussion of the implications of median age on currency of resources in college library monographic collections, especially in the professional, scien­ tific, and technical fields. The major recommendation is a Continual Update Collection Management Model for college collections. raditionally, median age and currency of collections have not been concerns in the re­ search library arena because, by definition, research libraries seek to build collections with retrospective depth. Under this collection-building philosophy, the humanities and histori­ cally oriented disciplines were the prominent collecting areas in academic libraries. Traditional principles of collec­ tion development began to undergo change in the 1980s when the effects of the escalation in serials pricing on mono­ graph purchasing were beginning to be apparent. The ARL statistics series re­ vealed a decline in purchasing power and in actual number of materials pur­ chased in the latter 1980s. 1 The catchphrase for academic libraries caught in the acquisitions budget crisis became “Just in time, not just in case.” The just-in-time philosophy would seem to favor fields in which current information is paramount. Indeed, research revealed that in ARL libraries, the science/technology Anna H. Perrault is Associate Professor in the School of Library and Information Science at the Univer­ sity of South Florida; e-mail: perrault@luna.cas.usf.edu. Richard Madaus is Director of the College Cen­ ter for Library Automation; e-mail: madausr@lincc.ccla.lib.fl.us. Ann Armbrister is Associate Director of the College Center for Library Automation; e-mail: armbria@lincc.ccla.lib.fl.us. Jeannie Dixon is Assis­ tant Director of Library Software Operations at the College Center for Library Automation; e-mail: jeannie@lincc.ccla.lib.fl.us. Rhonda Smith is Reference Librarian at the Florida Distance Learning Refer­ ence and Referral Center; e-mail: rhonda@dudley.lib.usf.edu. 316 mailto:rhonda@dudley.lib.usf.edu mailto:jeannie@lincc.ccla.lib.fl.us mailto:armbria@lincc.ccla.lib.fl.us mailto:madausr@lincc.ccla.lib.fl.us mailto:perrault@luna.cas.usf.edu The Effects of High Median Age 317 fields had gained in percentage share of monographic acquisitions compared to the humanities/arts and social sciences.2 Yet, little attention has been given to currency and obsolescence in monograph collections with respect to the effects of high median ages, particularly in science/technology materials and increasingly technologically focused curricula. This article reports the findings of a study of community college collections in Florida with respect to median age and cur­ rency of monographic collections. The in­ terpretation of the findings generalizes the effects of median age on currency of collec­ tions to the universe of college libraries. The Florida Community College Collection Assessment Study Florida is unique in that the public com­ munity colleges share one statewide net­ work, the Library Information Network for Community Colleges (LINCC), which is provided through the College Center for Library Automation (CCLA). From the LINCC aggregated database, standard­ ized data for all twenty-eight Florida com­ munity college library/learning resource center collections can be obtained. A com­ prehensive study of the monograph col­ lections of Florida community colleges was conducted by CCLA in 1998 through data extracted from the LINCC database. This comprehensive study was preceded by an earlier study which was reported at Library Research Seminar I, “Partners and Paradigms,” held in Tallahassee, Florida, November 1–2, 1996.3 In the past few years, most current information, reference and indexing services, and selected journal literature have been converted to electronic formats. The major finding of the 1996 study of the aggregated resources base of the Florida community college collections by imprint year was that the monographic resources of Florida community colleges are significantly out of date. In addition, it was found that in the 1990s, the num­ ber of older materials had increased in percentage share to that of current mate­ rials. The analysis by subject divisions revealed that outdated materials are prevalent in all major subject divisions, including the sciences and technology in which currency of materials is impera­ tive.4 The profile by age obtained for the aggregated resources base and three in­ dividual library/learning resources cen­ ters in 1996 suggested that more in-depth quantitative analysis was needed. The findings of the 1996 collection as­ sessment study were incorporated into a Program Review of the community col­ lege library/learning resource centers (L/ LRCs) in Florida commissioned by the Florida Division of Community Col­ leges.5 One consequence of the LINCC Collection Assessment and the Program Review was the preparation of a budget request for remedial funding to address the inadequacy of resources in the com­ munity college L/LRCs. The Learning Resources Standing Committee of the Division of Community Colleges also rec­ ommended that a comprehensive study be conducted of the collections of all twenty-eight Florida community colleges. The comprehensive study of Florida com­ munity college collections and the full report of the collection assessment project were completed in September 1998.6 The Problem As a background to the problem of median age in college collections, it is necessary to differentiate the collection development philosophy for community colleges and primarily undergraduate four-year aca­ demic libraries from that of research librar­ ies. The most recent statement of collection development philosophy for community college collections is in Wanda K. Johnston’s Administering the Community College Learning Resources Program.7 Community college collection devel­ opment is directly related to the col­ lege mission. It requires an under­ standing of the informational and instructional needs of its students, 318 College & Research Libraries July 1999 faculty, administrators, and broader college community. The primary pur­ pose of the college’s collection devel­ opment is to support the instructional program. Consequently, appropriate resources are selected to serve diverse student learning styles and abilities; to support transfer, vocational, devel­ opmental, and community interest courses; and to use alternative infor­ mation delivery systems. Faculty re­ search and esoteric requests usually are fulfilled through external sources via resource sharing agreements. Thus, participation in broader coop­ erative networks is necessary to ex­ pand the instructional resources available on campus and to support research needs.8 This statement is in the same vein as earlier statements on the mission of col­ lege libraries. At a conference in 1975, Evan Farber provided a description of the purpose of college libraries: … a college library is very different from a university library, not just in size but also in purpose. Moreover, the needs of college undergraduates have to be determined by different criteria than those used for univer­ sity students. A college library must have, first of all, a collection of cul­ tural and recreational materials that can expand students’ horizons; sec­ ond, a good basic collection that will meet most of their curricular needs; and third, a good reference collection that will serve as a key to the imme­ diate library, and to resources else­ where. Only after these three needs are met should we think about a col­ lection to fill the occasional research need. We should aim for a well-cho­ sen basic collection that meets the first two needs, plus enough ad­ vanced materials to meet most of the students’ research needs, and then depend on outside sources for the re­ mainder … College librarians should be thinking of “reference-centered” libraries, not “book-centered” (that is warehouse-type) libraries.9 Both these statements emphasize the de­ velopment of collections that support the instructional program and fit the curricu­ lar needs of the students. Although neither statement directly emphasizes the need for current materials, it is implied in that both statements stress instructional and refer­ ence services. In the past few years, most current information, reference and index­ ing services, and selected journal literature have been converted to electronic formats. Many states, including Florida, have elec­ tronic resources packages provided through statewide consortial arrange­ ments. Although the availability of elec­ tronic resources meets a portion of the cur­ rent information needs of students, the development and maintenance of mono­ graph collections remains an important aspect of the mix of resources being pro­ vided by the college library. The problem is twofold: the analysis of the collections to study the distribution of monographs by subject and age (pub­ lication date), and the interpretation of the findings with respect to median age and its effects on the currency of collections. Do the monograph collections of the com­ munity colleges contain materials in suf­ ficient numbers and level of currency to provide adequate resources and services for the students of those institutions? Research Design The research design for the study is a col­ lection analysis of the Florida community college collections, both individually and collectively, as an aggregated resources base. The twenty-eight public community colleges in Florida share a statewide da­ tabase, LINCC, which reflects the aggre­ gated resources of community college L/ LRC collections in Florida. Data extrac­ tion from the LINCC database for the monographic collection assessment took place in May—June 1998. Only biblio­ graphic records with a monographic tag were extracted. Five broad subject group­ ings and forty-seven individual disci­ The Effects of High Median Age 319 plines and fields were defined for the study. The study analyzed monographic records divided into five-year periods be­ ginning with 1970. All imprints prior to 1970 formed a grouping of pre-1970 titles. Analysis and interpretation of the data are by the aggregated database, LINCC, and each of the twenty-eight community col­ lege L/LRCs individually. The analysis of the study findings in the Florida Commu­ nity College Assessment Report concen­ trates on the distribution of publications by subject and age (publication date). The Florida community colleges can be regarded as representative of the majority of the community colleges in the United States. This study is unique in that it is the first statewide study of community colleges uti­ lizing data extracted from a cooperative database. Hence, the data are comparable for each of the twenty-eight community colleges. The distribution of collections by age is not a measure often calculated for any size and type of library collection. The data necessary to calculate median age have not been routinely gathered. Median age re­ quires that the distribution of titles by sub­ ject and imprint year be known. Few online catalog databases include this type of data report as a standard feature. Indeed, it is a difficult data set to obtain due to the pro­ gramming involved, even when the ven­ dor or database management is cooperative. There are no reported data from like stud­ ies with which to compare the findings, es­ pecially with regard to median age The Florida community colleges can be regarded as representative of the majority of the community colleges in the United States. The funding trends in higher educa­ tion for the past fifty years, on average, have been similar for institutions of higher edu­ cation. The trends in collecting found in the LINCC study can be assumed to be typical patterns for many college libraries. To set the findings of the Florida Community Col­ lege Collection Assessment into the broader arena of academic library collections, it is necessary to review the literature for re­ search germane to the context of the study. Review of the Research One approach to the problem of judging the appropriateness of median ages is to look at the rate at which materials become obsolescent by disciplines or fields of study. This review of research first sum­ marizes previous reviews on obsoles­ cence and weeding. Then the findings of those individual studies germane to the interpretation of the findings of the Florida Community College Collection Assessment Study are reviewed. Reviews on Obsolescence and Weeding The most active period of research in ob­ solescence was the 1960s and 1970s. It is not within the scope of this review to cover again all the studies in this body of litera­ ture, which has been the subject of several thorough review articles. The most com­ prehensive review of the literature of ob­ solescence was published by Maurice B. Line and Alexander Sandison in 1974.10 This is the definitive review of the research in obsolescence, being both analytical and critical. Line and Sandison define the con­ cept of obsolescence as “the decline over time in validity or utility of information.”11 Their review includes an appendix of “Studies giving date or age data for li­ brary uses or references citations.”12 A total of 170 studies are reviewed and listed in the appendix, which is in tabu­ lar format and divided into sections ac­ cording to type of study: • Use studies: Individual libraries • Use studies: Interlibrary loans • Reference/citation studies: —Synchronous (one source period) —Two or more source periods —Diachronous —Relative • Analyses of references For each study, the type of library, form of material, year of data collection, num­ ber of items studied, and intervals of data distribution are given. This listing of the various studies with brief descriptions of the type of data collected forms an inven­ 320 College & Research Libraries July 1999 tory of the research up to that point in time. Line and Sandison summarized the characteristics of the studies: In the first place, most of them are concerned with science and technol­ ogy (with medicine having more than its share); there are a few in so­ cial science, a few library studies spanning several subjects, and al­ most none devoted to the humani­ ties. Secondly, they nearly all deal with research literature or academic libraries. Thirdly, many are based on very small samples, and few are com­ parable with one another, because of the differences in, for example, age grouping. Fourthly, many, especially the synchronous studies, rely on age- biased data so that their conclusions are misleading or invalid.13 It is interesting to note that the studies divide into two main types: use studies, and references or citation studies. Many of the use studies are performed on the mono­ graphic literature because books circulate. The studies of references or citations are almost all studies of the journal literature. Very few studies have been conducted that either deal with median age in monograph collections or study obsolescence of mono­ graphs according to discipline. The section of the review article by Line and Sandison on “Use of Monographs” includes the find­ ings of only four studies. Those germane to the research reported in this paper are reviewed in the section on individual stud­ ies later in this review. From their analysis of the research on obsolescence, Line and Sandison suggest hypotheses for examination: Literature may decline in use faster when (a) it deals with data of ephemeral rel­ evance (b) it is in the form of a ‘report’, thesis, ‘advance communication’ or pre-print (c) it is in a rapidly advancing tech­ nology Literature may decline in use more slowly when (a) it is descriptive (b) it deals with concepts (c) it is critical14 These hypotheses are useful to formu­ late broad guidelines by category or type of material rather than obsolescence rates by discipline. In their conclusion, Line and Sandison comment that any sum­ mary of the present state of knowledge in the obsolescence of materials “would be misleading.” They state as hypotheses what some have stated previously as fact. They list twelve recommendations for further research.15 Line and Sandison simply conclude that the research find­ ings in obsolescence of library materials are insufficient to formulate obsolescence rates. They contend that the point of ob­ solescence studies should be to project the future use of materials because that is what librarians really need to know. In 1981, D. Kaye Gapen and Sigrid P. Milner updated the previous reviews of the research on obsolescence.16 Few stud­ ies had been published since the review by Line and Sandison. Gapen and Milner conclude that: Much basic research remains to be done on obsolescence. Researchers have taken the concept as proven, but in fact it still only a [sic] hypoth­ esis. The studies that have been done have concentrated heavily on scientific fields at the expense of the social sciences and the humanities, and journal articles at the expense of monographs. More should be done in the humanities, if only to determine whether obsolescence is a concept which cannot be usefully applied outside the sciences.17 Gapen and Milner observe that many studies have been “motivated by the need to withdraw something and have been in­ terested only in what should be discarded, not in an ideally objective model.”18 They are critical of the findings of most of the research in obsolescence: “Ideally for re­ mote storage or discarding, research on ob­ solescence has produced many mathemati­ http:sciences.17 http:obsolescence.16 http:research.15 http:invalid.13 The Effects of High Median Age 321 cal formulas, but unfortunately they have been neither simple nor universally appli­ cable.” They term only that research that has been “transmogrified into bibliofolklore,” such as journals can be dis­ carded after seven years, as simple, but “much of it is generally incorrect.”19 The authors propose a “problem-solving man­ agement model,” which they describe in general terms but do not actually de­ velop.20 Line updated the previous review of obsolescence studies in 1993.21 A consid­ erable number of studies had been per­ formed since the 1974 review, most of which were citation studies of journals or journal articles. Line’s summary of the lit­ erature is very similar to that of the 1974 review: “there has been a good deal more discussion of the matter, and additional light has been shed on the theory, but much research remains to be done, unwarranted statements continue to be made, and there has been little contribution to the practical applications of literature use decay.”22 Of relevance to the research reported in this article is the observation that because “ar­ ticles are intended to report research at the frontiers and therefore date more rapidly than books, which consolidate knowl­ edge,” “citations to journals show a faster decay rate than citations to books.”23 Line concludes by observing that the “virtual library” may ease the space pressure on li­ braries, but that if books and articles are to be accessed remotely, there “will have to be immense improvements in indexing.” He proposes that it is technically possible to make the title pages, contents, pages, and even indexes to books available online.24 A theoretical article by Dianne Rothenberg appears in the same issue of Library Trends in 1993 as the updated re­ view by Line.25 Rothenberg suggests that parameters of obsolescence—characteris­ tics of the knowledge base, characteris­ tics of publications studied, characteris­ tics of uses, characteristics of users, char­ acteristics of the setting, and time span— are all “co-occurring events that affect changing uses of print materials over time.”26 Rothenberg poses several re­ search questions for further study and observes that “studies of changing uses of print materials over time need more sophisticated analysis than conclusions reached through frequency counts.”27 The literature of obsolescence cannot be separated entirely from the literature of weeding because much of the research in obsolescence has been conducted for the purpose of developing criteria for weeding. Likewise, much of the research in obsolescence is concerned with use of materials because the literature of weed­ ing and obsolescence center on the factor of use. Indeed, many of the same studies are reviewed, whether the focus of the review is obsolescence or weeding. The literature of weeding has been thor­ oughly reviewed by Stanley J. Slote, who published the first edition of Weeding Library Collections in 1975 with the most recent edi­ tion in 1997.28 The bulk of the literature on weeding is concerned with procedures for weeding based on determining past and future potential use of materials. The classic studies in weeding were con­ ducted in the 1960s and 1970s. These stud­ ies sought to take advantage of the com­ putational powers of computers to conduct research on the use of library materials. The theoretical studies by Winston C. Lister, 29 Aridaman K. Jain,30 Herman H. Fussler and Julian L. Simon,31 Michael K. Buckland,32 Richard W. Trueswell,33 Allen Kent et al,34 and Slote35 all sought to de­ velop mathematical formulae or objective criteria for the selection of materials for storage or weeding. Trueswell’s research became the most prominent of the weed­ ing studies with his theory of weeding by “last circulation date.”36 Using the findings of earlier research, he sought to identify a core collection by assuming that a small proportion of the collection satisfies a large proportion of the circulation. He examined circulation records to determine the age of the materials circulating and the number of times these materials had previously cir­ culated. In addition, he defined core collec­ tion as “a percentage of the collection that should satisfy a given level of the user cir­ http:online.24 http:velop.20 322 College & Research Libraries July 1999 culation requirements” rather than a set of specific subject materials.37 The converse of this was to attempt to define what per­ centage of user needs is not satisfied by the core collection. From the circulation studies he had conducted, Trueswell then set out to develop a weeding procedure based on the core collection principle. This was the “last circulation date” theory. He emphatically stated that all books should be removed that had “not circulated dur­ ing the previous eight year period.”38 Trueswell had developed a quantitative approach that predicted the size of the core collection and that helped identity candi­ dates for weeding and the desired effect of the weeding on circulation. His proposal of weeding by “last circulation date” re­ ceived much attention and was incorpo­ rated into methods for weeding library collections by Slote and others. From this review of the literature on obsolescence and weeding, only a few studies were found that are directly ger­ mane to the consideration of age of mate­ rials, or median age, or obsolescence of monographs by field or discipline. Those studies are reviewed in the next section. Individual Studies The landmark study of the effects of age on use of monographic literature was that of Fussler and Simon, which has been cited in every literature review for obso­ lescence and weeding. Fussler and Simon investigated the patterns of book use in several universities. Looking at book use by publication period, they found that the “more a book was used at first, the faster it was likely to decline in use,” that “past use was the best predictor of future use,” and that decline with time “differed sub­ stantially between subjects.”39 Fussler and Simon tested objective cri­ teria they developed for weeding against the expert opinion of faculty in a number of disciplines, including chemistry, eco­ nomics, and literature in English and Teu­ tonic languages. There was almost total agreement between the objective criteria and the experts in the “cumulative disci­ plines” of chemistry and economics, but not in the humanities. This was seen as an indication that “a differential plan might well be the best policy” for weeding.40 One of the earliest studies of obsoles­ cence in monographic literature is by Charles F. Gosnell. His article “Obsoles­ cence of Books in College Libraries” was first published in 1944 and reprinted in Collection Management in 1978.41 Using three book lists of recommended college library acquisitions, Gosnell found that different subjects in the lists had varying obsolescence rates of from 1.5 to 31.3 years, with the overall averages in the 8.1­ to 9.6-year range. His interpretation of “obsolescence rate” was that the lower the rate, the higher the number of older ma­ terials in that subject area. Materials that become obsolescent quickly have a higher obsolescence rate. A low number, such as four or five, means the material obsolesces more quickly, in four to five years. Con­ versely, a higher number, such as twenty- one, means the material has a long pe­ riod of usefulness. After analyzing three standard selection lists, Gosnell then studied the collections of five college li­ braries and previous circulation studies in college libraries. He found the librar­ ies had lower obsolescence rates, that is, a higher proportion of older materials than the selection lists.42 Gosnell emphasized the validity of the formula he developed for calculating ob­ solescence rates rather than findings of obsolescence rates by certain subject dis­ ciplines. Indeed, in the table “Rates of Obsolescence in College Libraries,” there are only four subject categories listed with obsolescence rates for each of the five col­ lege libraries in the study. The rates vary considerably for the same subjects from one college to another.43 The highest over­ all obsolescence rate in libraries was 4.9 years. There are actually little data on ob­ solescence rates of literature by discipline in the Gosnell study. From his findings, Gosnell considers an average obsolescence rate for a college library to be five percent a year. Assum­ ing this rate, he calculates the half-life of the collection to be approximately four­ http:another.43 http:lists.42 http:weeding.40 http:materials.37 The Effects of High Median Age 323 teen years. “That is, half of the useful col­ lection will be in titles fourteen years old or less. The average life or life expectancy would be about twenty years old.” An­ other finding was that “Titles over thirty years old in many college libraries com­ prise at least half of the collection.”44 Line and Sandison are critical of Gosnell’s methodology, describing the data analyzed as having been drawn from “five well-weeded college libraries.” The find­ ings “can do little more than reveal the principles on which weeding was done.”45 Jain reviewed all previous work in us­ age studies in an attempt to develop a mathematical model for use in selecting books for storage.46 The relevance of his study is that he is one of the few research­ ers to prefer age as a criterion for weed­ ing rather than use. From his review and study he concluded: In spite of the recent tendencies to overemphasize usage histories, this study shows that age is a significant variable in studying use of mono­ graphs … while usage rates of indi­ vidual monographs have consider­ able variation even over a short pe­ riod of time, the usage rates of vari­ ous age groups do not show any sig­ nificant differences over time.47 The study that has the most relevant findings for the interpretation of median age and obsolescence rates by discipline was conducted by George V. Hodowanec. He sought to establish obsolescence rates of the monographic literature by discipline for use in academic libraries through a study to determine annual book obsoles­ cence rates at the University of Akron in the early 1980s.48 Hodowanec sought to develop a Priority Weighting Formula (APW) for book budget allocation. The for­ mula was based on determining annual book obsolescence rates for individual in­ structional departments within a univer­ sity. The obsolescence rates were calculated through defining periods of “peak use,” which reveals the “immediacy” of user need, and the magnitude of peak use, which reveals the “intensity” of user need. These factors were incorporated into an Obsolescence Analysis Matrix to calculate the percent of annual decline in use by four major disciplinary divisions and more spe­ cific disciplines. The study found a range in obsolescence rates by subject discipline from a low of 2.27 percent a year in foreign languages to a high of 8.50 percent in the business collection. For the major divisions, the percentage obsolescence rates per year were: • fine arts and humanities, 4.27 per­ cent; • life sciences, 4.36 percent; • pure and applied sciences, 4.38 per­ cent; • social and behavioral sciences, 4.69 percent. The overall average library obsoles­ cence rate was 4.64 percent a year.49 Thus, according to the findings of the Hodowanec study, close to five percent of a collection becomes obsolescent each year. Although the percentage rates of obso­ lescence per year in his study were derived from one case study on a medium-sized, four-year institution, Hodowanec’s findings can be used to suggest that an average with­ drawal rate of five percent a year is reason­ able for any collection in which currency of information is a priority. Indeed, the ACRL/ AECT Standards suggest a three to five per­ cent withdrawal rate a year: 6.4 Obsolete, worn-out, and inappro­ priate materials should be removed based on a policy statement. Commen­ tary. Deselection and weeding on a regular basis is indispensable to a useful collection and should be done systematically. A written policy should govern what should be re­ moved, what should be replaced, and what should be permanently re­ tained. Not only do obsolete and in­ appropriate materials occupy expen­ sive storage space, but they also de­ tract from other current materials containing important information. From three to five percent of the collec­ tion should be replaced annually. The http:1980s.48 http:storage.46 324 College & Research Libraries July 1999 condition of the collection should be reviewed regularly and needed re­ pairs should be made.50 From his findings Hodowanec formed several conclusions that are relevant to consideration of currency in library ma­ terials by discipline. These are similar to the hypotheses posed by Line and Sandison. He observed that: … substantial growth and expansion of theory, research, and publication in a particular instructional field (e.g., business) seem to result in a higher annual rate of book obsolescence for that field’s curriculum-supporting collection … Moreover, academic fields which are in a developmental or redevelopmental state, undergo­ ing refinements in methodology and technology of their informational domain (e.g., industrial education, home economics, and computer sci­ ence) tend likewise, to demonstrate higher than average rates of obsoles­ cence in their collections. Conversely, academic fields that rely upon re­ vised versions or new editions of al­ ready existing materials (e.g., En­ glish) tend to amass collections with below average annual obsolescence rates. Finally, certain instructional fields and certain curriculum offer­ ings are primarily textbook oriented. When these fields and courses do not exhibit rapid expansion of theory, research, and publication, their cur- riculum-supporting collections tend to have below average annual use and obsolescence.51 The findings in the study by Hodowanec are similar to those by Gosnell forty years earlier. Both sought to develop formulas for calculating obsolescence by discipline based on use, and both found an average overall library obsolescence rate to be near five percent a year. Although each used a different formula for calculat­ ing obsolescence by discipline, their find­ ings are in agreement, for the most part, in that the scholarly research disciplines such as history, literature, and math have lower obsolescence rates than the professional and applied fields, such as business and technical fields. This review of the research in obsoles­ cence and weeding has shown that very few studies have sought to establish obso­ lescence rates for monographic resources by fields or disciplines. The accepted pro­ cedures for weeding of collections have employed use or “last circulation date” as a primary criterion for deselection, rather than the age of the material. Although Line and Sandison define obsolescence in terms of both validity and use of information, the research in obsolescence and weeding has concentrated almost solely on use rather than validity of contents. The contents or quality of the information in the book re­ ceives less attention than the objective fac­ tor of use. The assumption is that low use equates with outdated content—the reason the materials are not being used. But the contents may or may not be outdated in­ formation or research. Obviously, a book cannot contain information on events or developments after its publication date, but an older edition may not necessarily con­ tain inaccurate information, just not up-to­ date information; and it can still be “use­ ful.” Use of materials may or may not be related to the currency or accuracy of the information contained in a given title. Many outdated and inaccurate texts are circulating from libraries of all sizes and types because they are what the user finds on the shelf. Indeed, one of the objectives of bibliographic instruction programs is to teach students the value of determining the currency and accuracy of information, rather than just checking out any book on an assigned topic. These points were sum­ marized by Line and Sandison: Knowledge is commonly recorded in documents and in studying changes in their validity or utility it has been usual to study what happens to docu­ ments, although the nature of the re­ lation between document use and in­ formation utility over time remains http:obsolescence.51 The Effects of High Median Age 325 obscure. Decline in document use can occur even though the information recorded is still valid and potentially useful; it is not possible to tell from decline in document use which type of ‘knowledge obsolescence’ is occur­ ring—changes in use do not neces­ sarily correspond to changes in value and validity. Studies of document use can therefore serve only as partial indicators of “knowledge obsoles­ cence.”52 Collection assessment is concerned with quality. With the exception of the studies by Gosnell and Hodowanec, the findings of research in obsolescence for mono­ graphic literature do not form a firm basis for considerations of median age in library collections. For the most part, the contem­ porary collection manager is left to fall back on what Gapen and Milner term “transmogrified bibliofolklore” in analyz­ ing age data for library collections and forming judgments about the proper ra­ tios of current to retrospective materials. The interpretations in the analysis of the age data by subject in the Florida Commu­ nity College Collection Assessment study are based on research findings reviewed here, augmented by “bibliofolklore.” Interpretation of Findings The rationale for the interpretation of the findings in the Florida Community Col­ lege Collection Assessment Project is con­ gruent with the philosophy that college collections should emphasize the instruc­ tional and curricular needs of students and, therefore, that the collections should emphasize current resources rather than retrospective depth. The interpretation concentrates on the findings for median age and the broader analysis of the dis­ tribution of monograph records in the LINCC aggregated resources base by sub­ ject and age. Median Age The median age calculation in the Florida Community College Collection Assess­ ment Study uses 1970 as a base year. All pre-1970 titles are counted as 1970 because individual title ages prior to that date are unavailable in the study data. The calcula­ tion is performed using the number of titles in the five-year time blocks defined for the study. Actual median age of the collections would be older if median age were calcu­ lated on an individual title and year basis. The findings with respect to median age of the collective monographic resources of the Florida community colleges are dis­ played in table 1 and figure 1. For the LINCC database, the overall median age is close to twenty-four years. Half of the titles were published before 1974 and half after 1974. Median age for the collective resources base ranges from a low of nineteen years for the sciences to a high of twenty-eight years in the gen­ eral category. All the median ages are in the 1970s. Whereas twenty-six years for a median age may not be cause for concern in the humanities, twenty-three years appears to be high for the social sciences, which include business, psychology, edu­ cation, political science, and law. Withdrawing older materials can lower the median age of a category. The sciences have a median age of nine­ teen years, but technology has a median age of twenty-one years. Although many older standard texts in the basic sciences may still be useful, the materials in the applied engi­ neering and technological fields should be current for those fields. Although the sci­ ences and technology have lower median ages than the humanities and social sci­ ences, the median ages for all the broad sub­ ject groupings appear to be high. The twenty-eight community colleges are grouped into three peer groups by size of institution. Figure 1 shows median age in the broad subject groupings for the LINCC aggregated database and the three peer groups. Although there are slight dif­ ferences in the median ages among the peer groups and LINCC, they are not pro­ nounced. Within the three peer groups, individual colleges have higher or lower median ages than the average for LINCC 326 College & Research Libraries July 1999 TABLE 1 LINCC Monographic Bibliographic Records by Median Age Subject Year Reached 50% Median Age in Years General 1970 27.76 Humanities 1971 26.44 Social Sciences 1975 22.97 Sciences 1978 19.19 Technology 1976 21.31 LINCC total 1974 23.59 and the group. For many institutions, the median age of materials corresponds to founding date and funding patterns. For the collective resources base, fully half the material is in titles published be­ fore 1974. It would seem that if material is to be considered “current,” median age should be closer to the mid-1980s in order for one-half of the collections to be less than fifteen years old. One effect of high me­ dian age is that it means the percentage of current materials is low. Withdrawing older materials can lower the median age of a category. Conversely, not weeding older materials raises the median age un­ less acquisitions of newer materials in­ crease to offset the “weight” of the older materials. This effect is readily apparent in the full data analysis by percentage distri­ bution of the collection according to sub­ jects and age in the next section. Percentage Distribution of Collections by Age Median age is one type of analysis for age of collections. A more detailed analysis was performed of the distribution of LINCC monographic bibliographic records ac­ cording to percentage of titles for each time period for the forty-seven subjects areas de- FIGURE 1 LINCC And Peer Groups Median Age of Records fined for the study. The results of this analysis are included in table 2. The high percentage shares of older materials and low numbers of current acquisitions that result in low percentages of current titles show clearly in this analysis. From table 2, areas that might contain large numbers of outdated materials can be identified. Table 3 summarizes the data from table 2 by broad subject grouping across the thirty- year time span of the study. It is easy to see why the median ages are in the 1970s by looking at the percentage distribution of records. Reading the “LINCC average” line in table 3 horizontally from left to right, it can be seen that for the LINCC database, 36 percent of the holdings are in pre-1970 imprints. This means that as we rapidly approach the year 2000, more than one- third of the collective resources are thirty years old or older. Advancing toward the present across the bottom of the table, it can be seen that each five-year period in the analysis comprises a smaller percent­ age share of the total for the entire data­ base. Although it may be cause for concern that more than one-third of the titles in the collective resources base are imprints thirty years old and older, 29 percent are in the 1970s, with another 11 percent before 1985. Thus, only 23 percent of the titles in the collective resources base are less than fif­ teen years old. Comparing the 1990–1994 time period with the 1970–1974 time inter­ val, it can be seen that in all categories the older time period has a much larger per­ centage of records than the current time frame. Even for the sciences, the percent­ age of current materials is lower than for the twenty-year-old materials. Although it is desirable to have retrospective strength in some scholarly disciplines, for commu­ nity colleges in which the curricula empha­ ses are on current applied and technologi­ cal fields, the collective resources show an alarming decline in numbers of new titles added in the past decade. Although retrospective depth is desir­ able in the humanities, with classical texts in all fields continuing to be read, new criti­ cal and reference apparatus are necessary The Effects of High Median Age 327 for current teaching in any field. Table 3 shows that more than 45 percent of hu­ manities materials in the Florida commu­ nity college collections are older than thirty years and that less than 10 percent of titles in the humanities are in 1990s imprints. The humanities rank last among the broad sub­ ject groupings in the percentage of current titles. The decline in acquisitions in the humanities may be reflecting changing pri­ orities. It is possible that humanities sub­ jects are no longer as high a priority in the instructional programs of the community colleges as they were twenty to thirty years ago and that enrollment may now be con­ centrating in the allied health and techni­ cal fields. The social sciences have one-third of titles in the pre-1970 time period. As with the humanities, the percentage of titles declines with each five-year interval mov­ ing toward the present. The number of 1990s imprints is nearly 14 percent of the total number of social sciences titles, a better proportion than in the humanities. Slightly more than a third of titles in the social sciences were published after 1980. In the LINCC database, the sciences seem to have fared the best over the years. Although 26.76 percent of the science col­ lections are in the pre-1970 time period, nearly 19 percent of the science titles are in 1990 imprints. The years in between 1970 and 1990 may be in need of exami­ nation as the percentage of materials in those time frames seems high. With 26 percent in 1970s imprints and another 28 percent in 1980s imprints, there is the possibility that the science collections are still retaining many books with outdated scientific information and should be ex­ amined for deselection. As with the other broad subject group­ ings, percentage share of total for the tech­ nological fields does decline moving for­ ward in time. Although the sciences have 19 percent of total share in 1990s imprints, the highest of the broad subject group­ ings, technology has a 12 percent share of total in 1990s imprints. Only the hu­ manities have a lower share of total (10%) in current materials. TABLE 2 Number of Monographic Bibliographic Records per Subject by Five Year Period Showing Records per Period as a Percentage of Total Records per Subject Pre 1970 1970-1974 1975-1979 1980-1984 Subject Records % of Subj. Records % of Subj. Records % of Subj. Records % of Subj. Total General 3,728 49.56% 694 9.23% 814 10.82% 622 8.27% Philosophy 17,967 55.28% 4,467 13.74% 3,292 10.13% 2,203 6.78% Religion 24,239 44.49% 7,834 14.38% 6,344 11.64% 5,087 9.34% Music 15,865 40.66% 6,552 16.79% 5,784 14.82% 4,210 10.79% Arts 30,957 40.55% 12,618 16.53% 10,524 13.79% 8,101 10.61% Architecture 4,848 38.44% 1,766 14.00% 2,031 16.10% 1,418 11.24% Lang. & Lit. except. 73,045 51.17% 25,470 17.84% 20,096 14.08% 9,361 6.56% Romance & Germanic Langs. 2,321 53.93% 513 11.92% 361 8.39% 358 8.32% Literary History & Collections 32,639 40.74% 12,439 15.52% 10,692 13.34% 8,876 11.08% Literature: Romance Langs. 17,373 58.81% 3,825 12.95% 2,569 8.70% 1,869 6.33% 1985-1989 1990-1994 1995-1998 T�T�� Subject Records % of Subj. Records % of Subj. Records % of Subj. Records Total General 2,018 6.21% 1,787 5.50% 764 2.35% 32,502 Philosophy 4,786 8.78% 4,200 7.71% 1,986 3.64% 54,486 Religion 3,240 8.30% 2,351 6.03% 1,005 2.58% 39,020 Music 7,028 9.21% 5,088 6.66% 2,013 2.64% 76,341 Arts 1,364 10.82% 856 6.79% 328 2.60% 12,612 Architecture 7,445 5.22% 5,407 3.79% 1,886 1.32% 142,740 Lang. & Lit. except. 360 8.36% 305 7.09% 86 2.00% 4,304 Romance & Germanic Langs. 7,450 9.30% 5,849 7.30% 2,178 2.72% 80,125 Literary History & Collections 7,450 9.30% 5,849 7.30% 2,178 2.72% 80,125 Literature: Romance Langs. 2,130 7.21% 1,413 4.78% 352 1.19% 29,539 328 C ollege & R esearch L ib raries Ju ly 1999 ��������������� ������������������������������������������������������� ������������������������������������������������������������������������������������������� �������� ��������� ��������� ��������� ������� ������� ���������� ������� ���������� ������� ���������� ������� ���������� �������������������� ������ ��������������������� ������ ����������� ����� ���������������� ������� ���������� ������ ������������������������ ����� ��������������� ������ ���� ������ �������� ������ ��������� ����� ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ����� ������� ����� ����� ������ ������ ����� ����� ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ����� ������ ����� ������ ����� ����� ������ ������ ����� ����� ����� ����� ������ ������ ������ ������ ������ ������ ������ ������ ����� ������ ����� ������ ����� ����� ������ ����� ����� ����� ����� ������ ������ ����� ������ ����� ����� ����� ����� ����� ������� ��������� ������� ���������� ��������� ������� ���������� ��������� ������� ���������� ����� ������� ������������������� �������������������� ����������� ���������������� ���������� ������������������������ ��������������� ���� �������� ��������� ������ ������ ����� ������ ����� ����� ������ ����� ����� ����� ����� ������ ������ ������ ����� ����� ����� ����� ����� ������ ����� ������ ��� ������ ����� ����� ����� ����� ����� ����� ����� ������ ����� ����� ����� ����� ����� ����� ����� ������ ����� ����� ��� ������ ����� ��� ����� ����� ����� ��� ����� ����� ����� ����� ����� ����� ����� ����� ����� ����� ������� ������� ����� ������� ������ ������ ������� ������� ������ ������ T h e E ffects of H igh M ed ian A ge 329 TABLE 2 (cont.) Number of Monographic Bibliographic Records per Subject by Five Year Period Showing Records per Period as a Percentage of Total Records per Subject Pre 1970 1970-1974 1975-1979 19�0-19�4 Subject Records % of Subj. Records % of Subj. Records % of Subj. Records % of Subj. Anthropology 7,170 33.66% 3,927 18.44% 3,476 16.32% 2,209 10.37% Recreation 7,738 25.48% 5,529 18.21% 6,565 21.62% 4,286 14.11% Social Sciences, General 7,738 5.30% 681 2.24% 647 2.13% 590 1.94% Business, Finance & Economics 1,608 23.84% 19,479 14.39% 21,043 15.55% 20,323 15.02% Sociology 32,259 19.59% 21,633 17.88% 21,383 17.67% 16,485 13.63% Political Science 23,707 43.86% 8,070 16.09% 6,113 12.19% 5,124 10.22% Law 21,994 21.16% 7,916 16.64% 7,996 16.81% 6,617 13.91% Education 10,064 27.94% 16,312 19.26% 12,262 14.48% 9,460 11.17% Military & Naval Science 23,659 28.92% 2,207 13.61% 2,054 12.66% 2,901 17.88% Bibliography & Library Science 4,691 23.90% 5,167 15.76% 5,420 16.53% 4,957 15.12% 1985-89 1990-94 1995-98 ����� Subject Records % of Subj. Records % of Subj. Records % of Subj. Records Anthropology 1,934 9.08% 1,857 8.72% 721 3.38% 21,301 Recreation 2,911 9.59% 2,289 7.54% 1,042 3.43% 30,367 Social Sciences, General 2,911 10.24% 351 7.83% 146 3.26% 4,482 Business, Finance & Economics 459 13.26% 16,550 12.23% 7,728 5.71% 135,339 Sociology 17,940 12.31% 14,412 11.91% 8,471 7.00% 120,985 Political Science 14,889 7.50% 3,557 7.09% 1,518 3.03% 50,146 Law 3,759 13.44% 5,896 12.39% 2,683 5.64% 47,569 Education 6,391 10.25% 9,136 10.79% 5,154 6.09% 84,675 Military & Navel Science 8,682 17.21% 1,135 7.00% 435 2.68% 16,221 Bibliography & Library Science 2,791 13.41% 3,536 10.78% 1,464 4.47% 32,787 330 C ollege & R esearch L ib raries Ju ly 1999 ��������������� ������������������������������������������������������� ������������������������������������������������������������������������������������������� �������� ��������� ��������� ��������� ������� ������� ���������� ������� ���������� ������� ���������� ������� ���������� ��������������������� ������� ������ ������� ������ ������� ������ ������� ������ ������������ ������� ����� ��� ����� ��� ����� ��� ����� ���������������� ����� ������ ����� ������ ����� ����� ����� ������ ����������� ����� ������ ����� ������ ����� ����� ����� ������ ����������������� ������ ������ ����� ������ ����� ������ ����� ������ ������������� ������ ������ ����� ������ ����� ������ ����� ����� ���������������������������������� ������ ������ ����� ������ ����� ������ ����� ������ �������� ����� ������ ����� ������ ������ ������ ������ ������ ��������������������������� ����� ����� ��� ����� ����� ������ ����� ������ ������� ��� ����� ��� ����� ����� ������ ����� ������ ������� ��������� ��������� ����� ������� ������� ���������� ������� ���������� ������� ���������� ������� ��������������������� ������ ������ ������ ����� ������ ����� ������� ������������ � ������ ����� ����� ��� ����� ������ ���������������� ��� ������ ����� ������ ����� ����� ������ ����������� ����� ����� ����� ����� ����� ����� ������ ����������������� ����� ����� ����� ����� ����� ����� ������ ������������� ����� ������ ����� ������ ����� ����� ������ ���������������������������������� ����� ������ ������ ������ ����� ����� ������ �������� ����� ������ ����� ������ ��� ������ ����� ��������������������������� ������ ����� ����� ������ ����� ������ ������ ������� ����� ����� �� ������ �� ������ ��� T h e E ffects of H igh M ed ian A ge 331 TABLE 2 (cont.) Number of Monographic Bibliographic Records per Subject by Five Year Period Showing Records per Period as a Percentage of Total Records per Subject Pre 1970 1970-1974 1975-1979 19�0-19�4 Subject Records % of Subj. Records % of Subj. Records % of Subj. Records % of Subj. Other Systems of Medicine 87 14.22% 70 11.44% 88 14.38% �7 1�.8�% Total Sciences 75,629 26.76% 36,821 13.03% 37,908 13.41% 39,911 14.12% Agriculture 7,164 29.99% 4,654 19.48% 4,486 18.78% 2,699 11.30% Technology: General 1,944 26.19% 1,070 14.42% 1,082 14.58% 1,064 14.34% Engineering: General, Civil, Construction 4,622 21.72% 4,067 19.11% 4,205 19.76% 3,253 15.29% Engineering: Mechanical 1,380 25.93% 676 12.70% 1,384 26.00% 1,069 20.08% Engineering: Electrical 3,977 24.49% 2,307 14.21% 2,526 15.56% 2,599 16.01% Engineering: Automotive, Aeronautical 3,683 28.38% 2,339 18.03% 2,467 19.01% 1,720 13.26% Mining, Metalurgy, Chem. Tech. 1,791 36.60% 1,029 21.03% 739 15.10% 673 13.75% Manufactures & Handicrafts 3,217 21.11% 3,541 23.23% 3,871 25.40% 2,075 13.62% 1985-1989 1990-1994 1995-1998 ����� Subject Records % of Subj. Records % of Subj. Records % of Subj. Records Other Systems of Medicine 82 13.40% 89 14.54% 99 16.18% 612 Total Sciences 39,160 13.85% 35,383 12.52% 17,828 6.31% 282,661 Agriculture 2,400 10.05% 1,736 7.27% 749 3.14% 23,888 Technology: General 971 13.08% 881 11.87% 407 5.48% 7,422 Engineering: General, Civil, Construction 2,364 11.11% 2,167 10.18% 596 2.80% 21,278 Engineering: Mechanical 456 8.57% 276 5.19% 81 1.52% 5,323 Engineering: Electrical 1,995 12.29% 1,598 9.84% 1,234 7.60% 16,236 Engineering: Automotive, Aeroautical 1,515 11.68% 921 7.10% 330 2.54% 12,976 Mining, Metalurgy, Chem. Tech. 348 7.11% 227 4.64% 86 1.76% 4,893 Manufactures & Handicrafts 1,388 9.11% 783 5.14% 361 2.37% 15,240 332 C ollege & R esearch L ib raries Ju ly 1999 E< :. � N u0 e.. � Z e . 0. 0 e0 c.u :. -. 0 c.u . u0 ' . .u. ... . -. .0 c0 ..e .u0 ' . - . . 0 ' -. u. e. . 0. E0 . - . u0 ' . .u. �." tQ tQ I t. tI I t 0QI 0. ': c:" 'l �" '0. II � 0" ': c: �" '0. II � 0" ': c: �" '0. II � 0" ': c: �" '0. II � 0" ': c: �0 �( �' 0�0 �·' ( N O .. No N'' N ' \ . NN' N 0 0. No oM' N o M . 0O' 0 l;Q " '.: 0Q0 !; !Q 0-0 MI M!� MM iI0 i- MM� M! 0.i i0 ..� i\ \MI i. I\� 0 � . \.0 !0 Q !0 !I� !i0 I. i iI IQ� M\0 M. . i! I.� M!! !M ! ii !M� i0- I i 0-0 i00 Q i 00. i00 I i 00- 0 � . ': c:" 'l <" '0. '' � 0\ ': c: <" '0. '' � 0\ ': c: <" '0. '' � 0\ ': c: <" '0. '' �0 �( �' 0�0 �·' ( . 0 O0 . No' . N No 0 o N' '.0 N O .' .' 0N -;Q " '.: Q !; .0 000 .Q0 0.� 0 0 . 00Q t� . 000 Q� . . .0 0 0 � . . 0. .Q0 00� .0. 0. t 000 � 00 .0 0. 0� Q .. t.t The Effects of High Median Age 333 Technology does not follow the same pattern as the sciences. One reason for the differences may be that the universe of monographic publications in these fields is smaller than in the basic sciences and allied health fields. The practical nature of the education and training in the technological fields makes them less library and information resources oriented. With 25 percent of total in pre-1970 imprints, this subject grouping may need even more con­ centrated deselection attention. More specific subject areas can be examined using table 2. Look­ ing at the figures for the health sci­ ences in table 2, it can be seen that while only 10 percent of total is in pre-1970 imprints, 26 percent con­ centrates in the 1970s, with 17 per­ cent in 1980–1984, and 19 percent of total in 1985–1989. The acquisi­ tions rate for the twenty-year pe­ riod 1970–1989 appears to have been adequate, but many of the ap­ plied materials from that time span would now be outdated practice, especially those before 1985, now almost fifteen years old. The general, physical, and life sciences all have more than 40 per­ cent of total in pre-1970 imprints, with mathematics at 30 percent and oceanography at 35 percent. The latter two subject areas display a more erratic acquisitions pattern over time, increasing share of total in some years and decreasing in others. In the 1990s, the positions of the subject areas in the sciences are somewhat reversed from earlier decades, with the health sciences fields occupying larger proportions of total and the basic sciences de­ clining in share of total. Computer science titles classify in several different call numbers— QA, TK, and in the Hs. It is espe­ cially alarming that the math (QA) and electrical engineering (TK) lines do not show a more current 334 College & Research Libraries July 1999 age or rate of acquisitions pattern. Me­ chanical engineering has the highest per­ centage of materials in the 1980–1998 time frame, but it still has more than 50 per­ cent of titles in the pre-1980 time frame. In fact, fully 62 percent of titles in tech­ nology are pre-1980s imprints. Business, education, law, and electri­ cal engineering have larger percentages of older imprints with decreasing percent­ ages of the total number of titles in more current years. Medicine and nursing have an opposite pattern with larger numbers of titles in more recent years and smaller percentages of titles in older materials. In fact, table 2 shows that the health sciences categories, the lines for medicine, thera­ peutics and pharmacology, nursing, and “other systems of medicine,” have the highest percentages of total in the LINCC database of all subjects in the 1990s. To reiterate, in the Florida community college collections, the sciences have fared better in acquisitions numbers overall than the humanities and social sciences. The main problem is in the number of older imprints that could contain out­ dated information, especially in the pro­ fessional, technical, and scientific fields. With acquisitions slowing in the past twenty-five years, there are probably materials that contain outdated informa­ tion. Many areas may need newer edi­ tions of basic, standard material. The next section further considers the problem of age of collections as it relates to the provision of high-quality resources. Hypothetical Median Age As previously stated, the rationale on which the analysis for the Florida commu­ nity colleges is based is that the collections be composed of materials that are intellec­ tually viable and contain accurate, current information. It is useful to ponder what a percentage share of total for the broad sub­ ject groupings would look like if a hypo­ thetical matrix were constructed to achieve a desirable balance of more recent books to older materials. Table 4 resembles the actual data from the Florida Community College study contained in table 3 but is constructed to take the differences in schol­ arship between the broad subject group­ ings into consideration. It is a hypotheti­ cal look at percentage share of total with an emphasis on current materials. Table 4 illustrates a seesaw pattern. The sci/tech fields are light on the older side and “heavier” on the current end. The scholarly disciplines are tipped toward the older end, although they have a higher percentage of share in recent materials than the patterns found in the Florida commu­ nity college study. The differences between the patterns of distribution by time period in the two tables is striking. Obviously, it is simple mathematics that the percentage share of total collections post-1980 must be at least 50 percent in order to achieve median ages in the 1980s. The hypothetical distribution in table 4 would result in median ages (indicated by bold type) of the 1970s for the humanities, in the mid-1980s for the social sciences, the late 1980s for sciences, and the early 1990s for technology. Although the percentage shares could be adjusted endlessly, the percentages in the table do allow for dif­ ferences in the production and use of knowledge in the broad groupings. The humanities do have the highest percent­ age in retrospective materials. The social sciences have one-fourth in older materi- TABLE 3 Percentage Shares by Time Period for Broad Subject Grou[ings in LINCC Pre '70 70-74 75-79 80-84 85-89 90-94 95-98 Total Hwnanities 45.39% 14.77% 11.73% 9.52% 8.94% 6.74% 2.90% 100.00% Social Sciences 33.37% 16.55% 14.29% 11.49% 10.28% 9.45% 4.54% 100.00% Sciences 26.76% 13.03% 13.41% 14.12% 13.85% 12.52% 6.31% 100.00% Technology 25.23% 18.22% 19.44% 14.57% 10.96% 8.07% 3.50% 100.00% LINCC (avg.) 36.35% 15.50% 13.54% 11.32% 10.33% 8.82% 4.13% 100.00% The Effects of High Median Age 335 TABLE 4 Hypothetical Share of Subject Groupings to Achieve Lower Median Ages Pre 1970 1970s 1980-84 1985-89 1990-94 1995-99 Total Humanities 37% 23% 12% 11% 9% 8% 100% Social Sciences 25% 16% 13% 13% 15% 18% 100% Sciences 19% 13% 10% 15% 19% 24% 100% Technology 15% 10% 11% 10% 25% 27% 100% (llis table does not contain actna! data.) als to allow for the historical nature of the scholarly disciplines in the social sciences. The percentages in the last two time peri­ ods may look shockingly large for the so­ cial sciences, sciences, and technology com­ pared to the humanities. But it must be borne in mind that these figures reflect an idealized distribution by percentage share within the four broad areas of knowledge and not annual acquisitions rates. As pre­ viously observed, the larger the percent­ age of older imprints in an area, the smaller the percentage of more current materials in a 100 percent calculation. This table of a hypothetical distribution of titles by five-year periods for the broad subject groupings is meant to be thought- provoking. It illustrates the differences between fields and disciplines with high obsolescence rates and those with materi­ als that remain viable for longer periods of time. This discussion has been for the purpose of raising the issue of median age and its effects on the quality of resources and impact on the services those resources provide. Although research libraries do not aim to lower the median age of collections, there are areas in professional, technical, and scientific fields that should have lower median ages than the scholarly retrospec­ tive disciplines, no matter the size of the library. Only the largest of research insti­ tutions can justify keeping outdated pro­ fessional, technical, and scientific informa­ tion for historical purposes. Recommendations The problem of high median age and in­ sufficient numbers of current acquisitions in the Florida community college collec­ tions was addressed in the LINCC Report with a proposed collection management model based on systematic additions and withdrawals—a Continual Update Model. This model is similar to the no- growth, “optimal collection size” theory posited at the conference in Chicago in 1975 by Trueswell and further elaborated on by Daniel Gore. The classic studies in both weeding and obsolescence had nearly all been reported by the mid-1970s. By this time, it was ap­ parent that the rapid collection building of the 1950s, 1960s, and early 1970s, and the continued increase in the number of pub­ lications, monographs, and journals, were causing both space and budgetary prob­ lems in academic libraries. A conference was held in Chicago in 1975 to address these problems. At that meeting, Trueswell posited the idea of a no-growth collection, which he defined as a static-size collection to which new additions would still be made.53 The static size would be main­ tained by a steady withdrawal rate bal­ anced with the same rate of new acquisi­ tions. This “optimal collection size” theory was explicated by Daniel Gore in a research paper added to the conference proceedings which gave the volume and the movement an identity—“Farewell to Alexandria: So­ lutions to Space, Growth, and Performance Problems of Libraries.”54 Gore attempted to answer the question, How large should a library be? His answer was that a collection should be large enough to produce the performance (sat­ isfaction) rate desired by the institution. He emphasized that an optimum performance rate could be achieved with smaller, but 336 College & Research Libraries July 1999 more judiciously selected, collections. His rationale was based on making the collec­ tion smaller to provide more and better service, which he termed “acceptable per­ formance rate,” which in turn was based on “availability rate” of materials the li­ brary owned. Gore’s argument was that a new building would never be needed if the number of volumes required to maintain any specified performance rate also re­ mained constant. Although the titles held by a library will change from year to year, the total number remains constant. The outflow rate of withdrawn materials will match the intake rate of new volumes. The withdrawals would be selected by the Trueswell criterion of weeding by last cir­ culation date. Gore based the explication for optimal collection size on a hypothetical collection of one million volumes. This theoretical model was based on the findings of pre­ vious research in obsolescence, use, and weeding. But it was not found to be prac­ tical. The no-growth concept, which would seem to have been a logical conse­ quence of the need to control growth and weed collections, was very controversial and not accepted in the research library arena. The literature of the library field does not contain reports of the adoption of the optimal collection size model. Continual Update Model The Continual Update Model incorpo­ rates the finding by Hodowanec that the overall obsolescence rate for college col­ lections is approximately five percent a year. The suggested model for collection management is to add five percent new titles a year at the same time withdraw­ ing five percent older materials. Although this model resembles the optimal collec­ tion size theory, it does not require the calculation of performance rates, avail­ ability rates, and so on. It is not based on a predetermined or optimal collection size but, rather, on a dynamic collection of continued additions and withdrawals within which median age by discipline and field can be monitored. The empha­ sis is on maintaining currency of re­ sources rather than a certain optimum collection size. A recommended collection profile, us­ ing data from the LINCC study to illus­ trate the Continual Update Model, is shown in figure 2. The 1990–1994 time period, which is the last complete data interval in the study, is used to calculate a projection for the LINCC database through the year 2004. The graph assumes five percent additions and withdrawals on a moving five-year scale. Thus, the last three bars in the graph are the same per­ centage of collection because the same acquisitions level is used for all three. The bars in the main graph in figure 2 show the distribution of the existing resources according to the time periods utilized in the study. The line (shaded area) shows what the distribution by imprint date would look like had the suggested model been in practice since 1970. Collections in which the Continual Up­ date Model is practiced would have an age profile almost directly the opposite of the present profile of the LINCC aggregated resources base. Over the span of nearly thirty years, the collective resources by imprint year would have had a substan­ tially different age composition had the suggested model been in place. A system­ atic process of additions and withdrawals would produce a profile similar to the shaded area in figure 2. As new materials are added and older, outdated materials withdrawn, the percentage of older mate­ rials diminishes in proportion to current imprints. The median age in this hypotheti­ cal collection would have been an overall eleven years instead of the actual LINCC median age of twenty-four years. If such a model were implemented, withdrawals would naturally be spread over the entire collection, although not evenly, but more heavily in some fields than others. The percentage of materials for specific disciplines needing replace­ ment each year could become a factor in a formula for book budget allocations, just as Gosnell and Hodowanec had suggested. The point of this model is not to save space, although that is one effect. Rather, The Effects of High Median Age 337 the point is to maintain collections at a cur­ rent level while at the same time retaining the intellectually viable titles in the collec­ tions and discarding superceded and just- plain-wrong information. Literary texts, basic science texts, and any other materi­ als still intellectually viable could remain or be replaced with newer editions that would be more attractive to users. Classics no longer in print could be retained and supplemented with newer critical or ex­ plicative apparatus. The net effect of such a policy is to keep collections current in areas in which currency is paramount and at the same time to maintain breadth and depth in areas in which historical material is desirable. Such collections impress us­ ers as fresh and up-to-date. They also im­ part to users the confidence that a college library is adequately supported by the in­ stitution and that its professionals are knowledgeable and capable of assisting them in the educational process. The model requires that new additions to the collection be sustained at the level of at least five percent per year. Although five percent may not seem high, calcu­ lated for many academic libraries on just the monograph budget, or by number of titles acquired annually, it is, in reality, a substantial figure. The adoption of such a model could be the impetus for a long- range funding plan for college collections because implementation requires the ad- If libraries are to continue to be buildings with physical materials as well as virtual services, the collec­ tions need to be vibrant and not give a warehouse impression. dition of, at a minimum, five percent new materials a year. Conclusions From the analysis in the Florida commu­ nity college study, it appears there is a rela­ tionship between median age of collections, level of funding, and founding date. Most of the community colleges were founded in the 1960s. Start-up funding was adequate or more than adequate for this time period. However, by the 1980s, higher education in general was beginning to experience re­ duced funding or at least reduced purchas­ ing power. In many institutions, the library/ learning resources centers were simply not supported as well as other components. For these and a variety of other reasons, the col­ lections of the community colleges in FIGURE 2 LINCC Age Profile of Collection: Projected versus Recommended 35% 30% 25% 20% 15% 10% 5% 0% P er ce n t o f C o lle ct io n Recommended Pre 1970 70-74 75-79 80-84 85-89 90-94 95-99* 00-04* Year of Publication 338 College & Research Libraries July 1999 Florida experienced declining growth be­ ginning in the late 1970s. As we move for­ ward in time, the majority of subject areas in the study decline in acquisitions num­ bers resulting in median collection ages in the 1970s. The profiles of the community colleges and the collective resources base reflect an uneven pattern of support for the collections, which has resulted in high me­ dian ages and serious deficiencies in cur­ rent materials. The findings of the 1998 LINCC col­ lection analysis have led to concern and recommendations for collection manage­ ment in college library collections. As we approach the end of this decade and cen­ tury, age of collections should be a mea­ sure for examination of the quality of ser­ vice provided by the resources. The ac­ quisitions buildup of the post World War Two years through the late 1970s has stocked academic libraries with materi­ als that now are twenty to fifty years old. Have many of these materials from the peak collection building years now be­ come outdated and too misleading to re­ tain? Are there high median ages due to the lack of current acquisitions to bring down the “weight” of the larger percent­ ages of older materials? Although community colleges and undergraduate libraries have the clearest mission to emphasize and maintain cur­ rent collections, with the rapid changes technology has brought to almost every professional, scientific, and technical field, currency of information is a much more paramount consideration now than it was thirty years ago when academic libraries focused on building retrospec­ tive collections to serve the future. If li­ braries are to continue to be buildings with physical materials as well as virtual services, the collections need to be vi­ brant and not give a warehouse impres­ sion. As we approach the new millen­ nium, do we need to revise collection management thinking? Notes 1. Association of Research Libraries, ARL Statistics, 1989/90 (Washington, D.C.: ARL 1991): 7. 2. Anna H. Perrault, “The Shrinking National Collection: A Study of the Effects of the Diver­ sion of Funds from Monographs to Serials on the Monograph Collections of Research Libraries,” Library Acquisitions Practice & Theory 18, no. 1 (1994): 3–22; “The Changing Print Resource Base of Academic Research Libraries in the United States,” JELIS (Journal of Education for Library and Information Science) 36 (fall 1995): 1–16. 3. Anna H. Perrault, John DePew, Richard Madaus, Ann Armbrister, and Jeannie Dixon, “An Assess­ ment of the Collective Resources Base of Florida Community College Library Collections: A Profile with Interpretive Analysis,” Resources Sharing and Information Networks 14, no. 1 (1999). In press. 4. Ibid. 5. W. Lee Hisle, “Program Review of the Florida Community College System Libraries/Learn­ ing Resources” (Tallahassee, Fla.: Florida Division of Community Colleges, Aug. 2, 1996), Report. 6. Anna H. Perrault, Richard Madaus, Ann Armbrister, Jeannie Dixon, and Rhonda Smith, The Florida Community College Collection Assessment: A Study of the Collections of the Twenty-Eight Florida Community College Collections (Tallahassee, Fla.: College Center for Library Automation, 1998). 7. Wanda K. Johnston, Administering the Community College Learning Resources Program (New York: G. K. Hall, 1994). 8. Ibid., 66. 9. Evan Farber, “Limiting Library Collection Growth,” in Farewell to Alexandria: Solutions to Space, Growth, and Performance Problems of Libraries, ed. Daniel Gore (Westport, Conn.: Green­ wood Pr., 1975), 39–40. 10. Maurice B. Line and Alexander Sandison, “‘Obsolescence’ and Changes in Literature over Time,” Journal of Documentation 30 (Sept. 1974): 283–349. 11. Ibid., 283. 12. Ibid., 323–49. 13. Ibid., 298. 14. Ibid., 317–18. 15. Ibid., 320–21. 16. D. Kaye Gapen and Sigrid P. Milner, “Obsolescence,” Library Trends 30 (summer 1981): 107–24. 17. Ibid., 116–17. The Effects of High Median Age 339 18. Ibid., 117. 19. Ibid., 107. 20. Ibid., 119–22. 21. Maurice B. Line, “Changes in the Use of Literature with Time—Obsolescence Revisited,” Library Trends 41 (spring 1993): 665–83. 22. Ibid., 665. 23. Ibid., 670. 24. Ibid., 679. 25. Dianne Rothenberg, “Changing Values in the Published Literature with Time,” Library Trends 41 (spring 1993): 684–99. 26. Ibid., 694. 27. Ibid., 697. 28. Stanley J. Slote, “Analysis and Review of the Literature of Weeding,” in Weeding Library Collections, 4th ed. (Englewood, Colo.: Libraries Unlimited, 1997), 48–79. 29. Winston C. Lister, “Least Cost Decision Rules for the Selection of Library Materials for Compact Storage” (Ph.D. diss., Purdue University, 1967). 30. Aridaman K. Jain, “A Statistical Study of Book Use Supplemented with a Bibliography of Library Use Studies” (Ph,D. diss., Purdue University, 1967); ———, “Sampling and Short-Period Usage in the Purdue Library,” College & Research Libraries 27 (May 1966): 211–18; ———, “Sam­ pling and Data Collection Methods for a Book-Use Study,” Library Quarterly 39 (1969): 345–52. 31. Herman H. Fussler and Julian L. Simon, Patterns in the Use of Books in Large Research Librar­ ies (Chicago: Univ. of Chicago Pr., 1969). 32. Michael K. Buckland, “Systems Analysis of a University Library,” Occasional Papers, no. 4 (Lancaster, Eng.: Univ. of Lancaster Library, 1970). 33. Richard W. Trueswell, “A Quantitative Measure of User Circulation Requirements and Its Possible Effect on Stack Thinning and Multiple Copy Determination,” American Documentation 16 (Jan. 1965): 20–25; ———, “Determining the Optimal Number of Volumes for a Library’s Core Collection,” Libri 16 (1966): 49–60; ———, Analysis of Library User Circulation Requirements (Amherst, Mass.: Univ. of Massachusetts, 1968); “User Circulation Satisfaction vs. Size of Holdings at Three Academic Libraries,” College & Research Libraries 30 (May 1969): 204–13; ———, “Growing Li­ braries: Who Needs Them? A Statistical Basis for the No-Growth Collection,” in Farewell to Alex­ andria: Solutions to Space, Growth, and Performance Problems of Libraries, ed. Daniel Gore (Westport, Conn.: Greenwood Pr., 1975): 72–104. (This last paper summarizes the previous research con­ ducted by Trueswell.) 34. Allen Kent et al, Use of Library Materials: The University of Pittsburgh Study (New York: Marcel Dekker, 1979). 35. Slote, “Analysis and Review of the Literature of Weeding.” 36. Trueswell, “Growing Libraries.” 37. Ibid., 93. 38. Ibid., 84–85. 39. Herman H. Fussler and Julian L. Simon, Patterns in the Use of Books in Large Research Librar­ ies (Chicago: Univ. of Chicago Pr., 1969): 144–47. 40. Ibid., 128. 41. Charles F. Gosnell, “Obsolescence of Books in College Libraries,” College & Research Librar­ ies 4 (Mar. 1944): 115–25; Collection Management 2 (summer 1978): 167–82. 42. ———, Collection Management, 176. 43. Ibid., 177. 44. Ibid., 181. 45. Line and Sandison, “’Obsolescence’ and Changes in Literature over Time,” 302. 46. Aridaman K. Jain, “A Statistical Study of Book Use Supplemented with a Bibliography of Library Use Studies” (Ph.D. diss., Purdue University, 1967). 47. Ibid., 125. 48. George V. Hodowanec, “Literature Obsolescence, Dispersion, and Collection Development,” College & Research Libraries 44 (Nov. 1983): 421–43. 49. Ibid., 425. 50. “Standards for Community, Junior, and Technical College Learning Resources Programs: The New Standards, Approved by ACRL and AECT,” College & Research Libraries News 51 (Sept. 1990): 757–76. 51. Hodowanec, “Literature Obsolescence, Dispersion, and Collection Development,” 425. 52. Line and Sandison, “’Obsolescence’ and Changes in Literature over Time,” 284. 53. Trueswell, “Growing Libraries,” 72–104. 54. Daniel Gore, ed. Farewell to Alexandria: Solutions to Space, Growth, and Performance Problems of Libraries (Westport, Conn.: Greenwood Pr., 1975): 164–80.