College and Research Libraries Literature Obsolescence-, Dispersion, and Collection Development George V. Hodowanec This study determines annual book obsolescence rates for individual instructional departments within a university. Analysis o sue factors as immediacy and intensity of peak usage, use dispersion, and the commonality of use have helped to develop an acquisition priority weight- ing (APW) formula. The function of the APW is to serve as a guide in the collection develop- ment process. The Circulation Commonality Table, the Dispersion Table, and the Immediacy/ Intensity Table were used to refine the well-known Pareto Principle and Trueswell' s 80/20 rule. This study identifies specifically which one-third of total resources receives two-thirds of total use, thus, the 113-2/3 rule was developed. everal factors may contribute to book obsolescence: the usage rate of a book from the high point of its circulation, subse- quent natural decline in circulation, and fi- nally its ultimate low pomt of circulation. Although some of these factors are per- haps too varied and too subjective to enu- merate and analyze, especially in aca- demic libraries, it is yet possible to observe certain general, repeated patterns in the circulation rate of books in all subject clas- sifications from the time of acguisition to the time of least circulation. 1 Likewise, . several factors may contribute to book dis- persion, 2 the use of books classified in one subject area by students majoring in an- other subject area-the use of math books by music majors, for example. Obsoles- cence and dispersion may be interrelated to some extent. Widely dispersed use may affect obsolescence rates in certain subject classifications. Widely dispersed use soon after acquisition may increase a book's rate of obsolescence because such use raises its peak usage rate. However, widely dis- persed use throughout a book's life may tend to stabilize its rate of obsolescence be- cause the instructional department whose curriculum that book supports is not solely responsible for its use. At the other extreme, however, narrowly dispersed and nondispersed library holdings may exhibit slower, more stable rates of obso- lescence because no ''outside' demand ar- tificially increases a nondispersed vol- ume's peak usage rate. Conversely, high • or low obsolescence rates may affect dis- persion by encouraging or discouraging widespread early or lifetime use of materi- als. Obsolescence and dispersion may both be influenced by the hard or soft (axio- matic or judgmental) nature of subject ar- eas, 3 by variations in instructional meth- odology, and by periodic curricular modifications at the college or university of which a given academic library is a part. Since obsolescence and dispersion are re- flections of user behavior and since hard or soft subject areas and differences in teaching methods or curricula can be ac- counted for, study of the statistical data that reflect user behavior and consider- ation of the types of and rationales for user behaviors thus reflected would enable ac- ademic libraries to prepare meaningful, practical collection development guide- George V. flodowanec is director of university libraries, University of Akron, Ohio. 421 422 College & Research Libraries lines based on local use patterns. Such guidelines would be an accurate response to the needs (as demonstrated by use) of an academic library's primary patrons, the students of the college or university. The goal of collection development guidelines developed from analyses of obsolescence data, dispersion data, and behavioral re- search will be to provide an academic col- lection which maintains adequate hold- ings to support the curriculum of a college or university while avoiding high obsoles- cence rates and low use rates, but sustain- ing well-dispersed holdings. LITERATURE REVIEW Obsolescence of materials in libraries has been recognized throughout the his- tory of libraries, but only since World War II has it been more and more frequently studied. In 1947, Gosnell sought to estab- lish mortality formulas for books akin to insurance company analyses of mortality rates among groups of people. Just as in- surance companies predict the number of people in given groups who will die after varying periods of time, although no one can predict just which people will die, Gosnell wanted to predict mortality rates for books in libraries. 4 In 1950, Gosnell ad- vocated ''systematic weeding'' of library collections, bringing considerations of ob- solescence into the realms of effective li- brary management. 5 In 1967, Hardin re- minded librarians that there is no finite limit to the rate of increase of a collection; therefore, microfilming, microcarding, and the like are not really solutions to the problem of rapidly growing collections. Hardin contended, ''either we must con- tinually diminish the rate of increase or we must introduce what may be termed a mortality factor and eliminate individuals whose procreation we have permitted. " 6 In 1970, Brookes analyzed the "obsoles- cence of special library periodicals," refin- ing the. techniques used to determine ob- solescence rates. 7 In 1973, Brookes presented graphic methods for plotting obsolescence in periodical literature and observed that ''at the present time there is no general agreement on how scatter should be defined or measured," al- though he added, "It seems likely that November 1983 scatter and obsolescence are related, but that both are determined by rate of growth-the faster the rate of growth, the less the scatter and the more rapid the ob- solescence."8 In 1974, Line and Sandison warned that "obsolescence tends to be a loaded term because it does not distin- guish decline in use from decline in value or recognize the possibility of increase in use and that current nonuse does not nec- essarily mean either obsolescence or lack of value. " 9 In 1975, Hodges noted that the most useful part of a book's existence is in the first three years after publication and that the acquisitions and cataloging proce- dures take too long a time out of a book's period of greatest demand. 10 The decade of the 1970s continued to see the publica- tion of studies addressing obsolescence and collection growth. An analysis of the entire range of previ- ously published studies reveals, however, that most obsolescence research deals with periodical literature, not with books. Moreover, identifying literature relevant to the relationship of obsolescence and collection development is hampered by the indexing of publications under head- ings not readily discernable as pertinent to the subject. Therefore, as an aid to future studies ·of obsolescence and collection de- velopment, a suggested reading list of published articles appears in appendix A. METHODOLOGY In order to provide direction for the pro- cess of data accumulation for this study, the decision was made to analyze one aca- demic library's annual circulation (Wil- liam Allen White Library, Emporia State University, 1980 calendar-year circula- tion), utilizing distinct academic instruc- tional departments as units of compari- son. These departments are relatively stable units that enable patterns of library use to be related to curricular programs. Departmental curriculum-supporting collections were identified through analy- sis of the content of each course listed in the university's course catalog and subse- quent comparison of that content with subject classifications in the Dewey deci- mal system, by which the university's li- brary materials are classified. Next, any conflicts between course titles and de- scriptions and the appropriate Dewey classifications were resolved. All calendar-year 1980 records of stu- dent circulation were examined, and the catalog numbers of volumes that had been circulated were grouped by Dewey divi- sional numbers (the second-level classifi- cations summarizing one hundred subject areas) corresponding to instructional de- partments. Data on books circulated in 1980 were then grouped to show the year of acquisition for each volume circulated, and this information was collated with each instructional department's previ- ously identified curriculum-supporting collection. This process was applied to volumes circulated in 1980 and acquired in any of the preceding twenty-two years. For the purpose of this study, acquisition records were accepted as indicators of a book's first availability to library patrons. The total number of volumes acquired in each Dewey division during each of the preceding twenty-two years was listed. Then, the 1980 figures on circulation of volumes in each Dewey division were ar- ranged to reflect the year of acquisition for each item acquired within the preceding twenty-two years. Finally, the number of volumes circulated in 1980 from each Dewey division's annual acquisition list was divided by the total number of vol- umes in each corresponding annual acqui- sition list. This process for the hypotheti- cal Dewey division XYZ would be illustrated thus: Dewey Division Number: XYZ Total1980 Circulation: 1,000 volumes Annual ratios of XYZ volumes circulated in 1980 (by year of acquisition) to the total XYZ volumes acquired (by same year of acquisition): 1979 1978 1977 1976 1975 1974 100/200 180/220 210/215 150/205 95/195 90/210 .50 .72 .98 .73 .49 .43 By grouping the circulated volumes ac- cording to their Dewey divisional classifi- cations and their years of acquisitions, in- formation on book use as a function of time for each library subject area of each instructional department's curriculum- supporting collection was generated. This information was used to plot obsolescence Literature Obsolescence 423 graphs for the individual instructional de- partment collections, for groups of these department collections comprising four broad disciplinary divisions (fine arts and humanities, life science, pure and applied sciences, and social and behavioral sci- ences), and for the overall library collec- tion. Circulation graphs were plotted by year of acquisition for every instructional de- partment; these graphs were then grouped according to their unique behav- iors of usage, and five representative graphs were selected to show standard and unique obsolescence patterns. The curves from these five graphs were super- imposed to illustrate diverse patterns of book use as a function of time. The super- imposed curves represented different use patterns observable within library circula- tion. The social studies/anthropology (SSA) curve most closely correlated with the li- brary average (LA) curve, with a mathe- matical coefficient of correlation of . 98. Similar use patterns were demonstrated in the curves of the biology, education, English, library science, mathematics, and physical sciences collections. The extent to which these curves paralleled the LA ~urve varied, but their coefficients of cor- relation (ranging from .80 for English to .98 for physical science) indicated a high degree of similarity in use patterns. The curve for the business (BUS) collection was considered a relatively close match to the LA curve and the aforementioned curriculum-supporting collection curves, with one exception. The BUS curve dis- plays an extremely high use rate for the first two years after books in this collection are acquired; then the intensity of use drops, and the remaining twenty years of use patterns show close correspondences to the LA curve. Because of the steep rate of use during the first two years of circula- tion, the BUS curve showed only a . 78 co- efficient of correlation to the LA curve. The home economics (HE) collection curve illustrated an erratic use pattern. Its coefficient of correlation (.83) was some- what lower than the . 98 coefficient for the SSA curve, but was within the high range of correlations to the LA curve. The art, 424 College & Research Libraries health and physical education, industrial education, psychology, and speech collec- tions revealed similar erratic use patterns, with coefficients of correlation to the LA curve ranging from .57 for health and physical education to .97 for industrial ed- ucation. All the curriculum-supporting collections in this group exhibit multiple circulations of books during the first two years following acquisition. The foreign languages (FL) collection re- vealed a unique use pattern. Its rate of cir- culation of books remained consistently below the LA rate, but in all other respects it closely paralleled the LA curve. Its .85 coefficient of correlation indicated a high correspondence to the LA curve. The use pattern of the fiction collection, studied separately, exhibited highly erratic use. Repeated attempts were made to ma- nipulate the plotted data into mathemati- cal, computer-generated expressions de- scribing book use as a function of time; however, all attempts (except linear repre- sentations of obsolescence rates) to gener- ate mathematical expressions of book use as a function of tiirie were unsuccessful. Therefore, the data reflecting book use over time were analyzed as representa- tions of user behavior patterns. Conclu- sions reached by such analyses were dis- cussed with the librarians responsible for collection development as well as with fac- ulty who are professionally responsible for directing student use of library hold- ings. These people's expertise was called upon to refine or refute preliminary ·con- clusions. All comments were considered valid, and an academic collection use analysis was included in this study so that the use of academic collections by declared majors specializing in an area could be compared to university-wide use of these same aca- demic collections. With such analysis and comparison, those responsible for making collection development decisions could avoid slighting collections showing heavy use by declared majors in a given subject area. Although past, present, and potential user behavior patterns were discussed with those professionally involved with the university and the academic library be- ing studied, across the board, no one was November 1983 able to offer new insights ~to the obsoles- cence pattern of books or into the factors that might affect the obsolescence se- quence. The most immediately compre- hensible graphic depiction of obsoles- cence seemed to be linear plotting of usage rate as a function of time, so linear repre- sentations of the book obsolescence se- quence were determined for the overall li- brary, the four broad disciplinary divisions, and the individual instructional departments (art, business, and mathe- matics, for example). Higher correlation coefficients between divisional and de- partmental obsolescence curves and the overall library obsolescence curve were found to exist with linear regression plot- ting than with exponential decay plotting. (See figure 1, column F.) Since no useful conclusions could be drawn from the "plateau" years, the years of steady low circulation of a book after it reaches its lowest point of circulation, and since these "plateau" years' circulation data lowered the coefficient of correlation of divisional and departmental curves with the overall library curve, preferability was given to the linear representations and to conclu- sions that could be drawn from them. While circulation is admittedly not an exact measure of book use, for the vast majority of library holdings, circulation is an adequate barometer of use, thus circu- lation data were accepted as indicators of use in all cases. Book use could be mea- sured by citations of library holdings in re- search works done at an academic library, but such a measurement of use would be equally as incomplete as circulation rec- ords. Reshelving records could be kept and used to measure book use, but such records would not include books inadver- tently removed from open stacks or books reshelved by patrons who ignored re- quests not to reshelve materials. It seems much more acceptable, then, to relate use to circulation, acknowledging the flaws in such a relationship, but achieving a mea- sure of the actual possession of a library holding by a library patron for whatever reason. ANALYSIS OF OBSOLESCENCE FINDINGS Decreased use over time is a normal, predictable, well-known characteristic of library holdings. However, knowing that obsolescence is characteristic of library materials is of little use; knowing the spe- cific rate of decreased book use, or book obsolescence, within well-defined subject groups would be more valuable because acquisitions requests and decisions are made in regard to such groupings. Column A of figure 1, the Obsolescence Analysis Matrix, reveals the annual obso- lescence rate for four major disciplinary division collections (fine arts and humani- ties; life science; pure and applied sci- ences; and social and behavioral sciences), for sixteen individual academic depart- ment collections comprising the four disci- plinary categories, and for the entire li- brary. In the process of calculating the individ- ual department collection obsolescence rates, books circulated in 1980 were grouped by the year of publication (acqui- sition) and by Dewey second summary di- visional subject areas that directly corre- lated to the content of courses offered in each instructional department. With each of these departments previously defined as a unit made of specifiC' informational subject areas (based on the collation of the course descriptions and the subject classi- fication headings in the second summary divisions of the Dewey list), a relatively stable subject-area profile of curriculum- supporting collections was revealed for each department. The annual obsolescence rate for all aca- demic department collections displayed a range of 6.23 percent, with a low of 2.27 percent in the foreign languages collection and a high of 8.50 percent in the business collection. The department collections varied within one standard deviation for almost two-thirds of these collections. The obsolescence rate of each individual instructional department collection indi- cates the rate at which books become less frequently used and, therefore, have less ~nformational value to library patrons . The obsolescence rates of the foreign lan- guages collection (2.27 percent per year) and business collection (8.50 percent per year) indicate that books in the former col- lection do not need to be replaced as quickly as books in the latter because Literature Obsolescence 425 books supporting the curriculum of the foreign languages department obsolesce at roughly one-quarter of the rate of pooks supporting the curriculum of the business department. Although such a generaliza- tion seems supportable by reference to ob- solescence rate patterns, most generaliza- tions based solely on the ranking of obsolescence rates are insufficiently repre- sentative of user behaviors to determine collection guidelines . Closer consideration of obsolescence rankings suggested that variations in the obsolescence rates of individual academic departments might result from several in- fluences. For example, 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 col- lection. Also, an instructional field that re- lies heavily on visual examples or descrip- tions of processes (art and architecture, for instance) to generate fresh approaches to its subject matter will tend to make its ac- quisitions obsolesce at a higher annual rate because users seem to obtain and re- view the volumes as soon as the books be- come available in the library, causing early multiple circulation patterns, then circu- late these volumes later at more-standard rates. Moreover, academic fields which are in a developmental or redevelopmen- tal state, undergoing refinements in the methodology and technology of their in- formational domain (e.g., industrial edu- cation, home economics, and computer science) tend, likewise, to demonstrate higher than average rates of obsolescence in their collections . Conversely, academic fields that rely upon revised versions or new editions of already existing materials (e.g., English) tend to amass collections with below average annual obsolescence rates. Finally, certain instructional fields and certain curriculum offerings are pri- marily textbook oriented. When these fields and courses do not exhibit rapid ex- pansion of theory, research, and publica- tion, their curriculum-supporting collec- tions tend to have below average annual use and obsolescence. There are other factors closely related to obsolescence that may be fruitfully con- A B c D E F G H RANKED BY RANKED BY RANKED BY RANKED BY RANKED BY RANKED BY RANKED BY RANKED BY OBSOLESCENCE YEAR OF RATE OF YEAR OF RATE OF COEFFICIENT NUMBER OF NUMBER OF RATE PEAK PEAK LOWEST LOWEST OF DEWEY DIV. DEWEY DIV. CIRCULATION CIRCULATION CIRCULATION CIRCULATION CORRELATION IN WITH NO. TO DEPARTMENTAL SIGNIFICANT LIBRARY AVG. COLLECTIONS DISPERSION (Percent of (Circulations (Circulations amiual decline) per 100 ilol.) per 100 vol. ) LIBRARY LIBRARY LIBRARY LIBRARY LIBRARY AVERAGE - 4. 64 AVERAGE - 2 AVERAGE - 69 AVERAGE - 13 AVERAGE - 18 - - - - - - - - - - - - ForL 2. 27 Art 1 Bus 141 Math 12 ForL 5 SocS .98 SocS 34 SocS 26 Eng 3.00 Bus 1 Psy 135 PhvS 12 Math 9 IndE • 97 ForL 14 ForL 14 SocS 3. 33 HomE 1 Art 126 Art 13 Mus 9 PhyS .97 LibS 10 LibS 9 Math 3. 82 LibS 1 HPE 121 ForL 13 LibS 10 Psy .96 Eng 6 IndE 4 LibS 4.15 Math 1 HomE 118 Mus 13 In dE 12 Educ .95 IndE 6 Art 3 PhyS 4.17 Psv 1 IndE 118 Bio 14 Eng 13 Mus • 95 PhyS 6 Eng 3 Mus 4.18 ForL 2 Sve 110 Educ 14 Psy 13 Art .90 Art 5 PhyS 3 Bio 4. 36 IndE 2 PhyS 68 Eng 14 SocS 15 LibS .90 Bio 3 Bio 2 Edu 4.45 Mus 2 Bio 66 HPE 14 Edu 17 Spe .89 Bus 3 Psy 1 Spe 6.42 SocS 2 Educ 66 LibS 14 Bio 18 Math .87 HPE 2 Bus 0 HomE 6.50 Sve 2 LibS 64 SocS 14 PhvS 18 ForL .85 HomE 2 Educ 0 Psy 7.00 Bio 3 Mus 55 s2e 14 Bus 22 Bio .83 Psy 2 HPE 0 In dE 7.07 Educ 3 SocS 55 Bus 15 HomE 27 HomE .83 Sl!e 2 HomE 0 Art 7.33 Eng 3 Math 51 IndE 15 Spe 33 Eng .80 Educ 1 Math 0 I HPE 7.45 HPE 3 En11. 46 HomE 15 Art 38 Bus • 78 Math 1 Mus 0 Bus 8.50 PhyS 3 ForL 30 Psy 19 HPE 39 HPE • 57 Mus 1 Spe 0 -------- ------~ -------- -------- ------- ------- ------- ------- FAH 4. 27 FAH 2 SBS 87 FAH 13 PAS 17 SBS • 97 - - - - - - - - LS 4. 36 PAS 2 PAS 74 LS 14 LS 18 PAS .96 - - - - - - - - PAS 4. 38 SBS 2 FAH 69 PAS 15 FAH 22 FAH .93 - - - - - - - - SBS 4.69 LS 3 LS 66 SBS 15 SBS 26 LS .83 - - - - - - - - KEY: Art--Art; Bio--Biology; Bus--Business; Educ--Education; Eng--English; FAH--Fine Arts & Humanities; ForL--Foreign Languages; HomE--Home Economics; HPE--Health & Physical Education; IndE--Industrial Education; LibS--Library · Science; LS--Life Science; Math--Mathematics; Mus--Music; PAS--Pure & Applied Sciences; PhyS--Physical Science; Pay--Psychology; SBS--Social & Behavioral Sciences; SocS--Social Studies/ Anthropology; Spe--Speech FIGURE 1 Obsolescence Analysis Matrix I J RANKED BY RANKED BY TOTAL NUMBER NUMBER OF OTBEll OF DEPTS. DEPTS. USING USINC EACH MOST DISPERSED DEPARTMENTAL DEWEY DIVSION COLLECTION - - - - - - - - SocS 14 Bus 12 HPE 14 Educ 12 Bus 13 Spe 12 Edu 13 HPE 11 Sve * SocS 11 Eng 10 Eng 7 Psy 8 Pay 7 Art 6 Art 7 HomE 5 HomE 3 LibS 5 LibS 2 PhvS 3 Bio 1 Bio 2 In dE 1 IndE 2 Math L Math 2 Mus 1 Mus 2 PhvS 1 ForL 1 ForL 0 ------- -------- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - *Core 000' s in LibS, Core 790' s in HPE sidered in conjunction with this analysis of obsolescence rates. Two of the most sig- nificant of such factors are the period of peak use and the corresponding magni- tude of peak use. The period of peak use reveals the ''immediacy'' of user need, and the magnitude of peak use reveals the ''intensity" of user need. Rankings of im- mediacy and intensity appear in columns B and C of figure 1. The immediacy factor indicates the ur- gency with which books are needed by li- brary patrons. Six academic instructional departments (shown in column B) exhibit peak circulation of their library materials one year after books in their curriculum- supporting collections have become avail- able in the library. Four of these same de- partments (shown in column C) also appear in the top-intensity group, exhibit- ing the highest ratios of circulation trans- actions per one hundred books acquired. These reflections of immediate and in- tense use by four academic departments support decisions to acquire materials for their collections as quickly as possible. These departments, and only these de- partments, demonstrate patron behaviors which recommend such quick acquisition of new publications. Instructional departments whose peak use of their curriculum-supporting collec- tion occurs in the second or third year after acquisition do not exhibit sufficient imme- diacy to necessitate urgent purchasing of their requests. Book purchases for these departments should be executed deliber- ately and carefully so that resources are not wasted and so that curricular pro- grams do not suffer. Most instructional department collections with delayed peak use periods also exhibit an intensity of use that is lower than collections with immedi- ate peak use periods. For example, column B of figure 1 shows that materials in the music department collection reach their peak usage period two years after ac- quisition, and column C shows that these materials' intensity of use during this pe- riod is only 55 circulations per hundred books acquired. Books in subject areas that reach peak immediacy after two or three years offer excellent possibilities for resource-sharing Literature Obsolescence 427 efforts among libraries. Items in categories with a moderate or low use intensity and a later immediacy factor give' a library plenty of time to determine whether inde- pendent acquisition or shared access is an appropriate response. Interlibrary loan records of items in low use and late imme- diacy categories will pinpoint the occa- sional heavily used item, and the library can acquire it. Other items in these catego- ries are perhaps best borrowed from li- braries at institutions where the curricular programs they support are especially em- phasized. Data in columns D and E represent the years that books within the various curriculum-supporting collections reach their lowest use and the rate of use in each curriculum-supporting collection during those years. The range of lowest rate of use varies from 5 to 39 hundred volumes and occurs twelve to seventeen years after books are acquired by the library. Although some studies have speculated that a certain low point in use justifies weeding via secondary storage or discard- ing, no sound decisions about weeding can be made unless each library considers the period of lowest use along with its own financial and physical resources, in- cluding operating funds, availability and accessibility of primary and secondary storage space, and storage costs in pri- mary and secondary storage locations. Each library should determine for itself the cost-effectiveness of storing and main- taining given volumes if their chances of circulating in a given year are only one in twenty. At some point, borrowing infre- quently requested volumes through inter- library loan will be more cost-effective for medium-sized libraries with limited bud- gets and space than holding those vol- umes will be. However, this point will vary for each library. Decisions about dis- carding, putting into secondary storage, or continuing to shelve certain volumes at the primary library facility can be sup- ported in part by the data in columns D and E. Differing circulation rates at the lowest points of use in each curriculum-support- ing collection suggest a variety of conclu- sions. Although the data appear contra- 428 College & Research Libraries dictory, curriculum-supporting collec- tions with higher annual rates of obsolescence also tend to be collections with higher rates of use at their lowest points of use. Conversely, the collections with lower annual rates of obsolescence tend to show lower rates of use at their lowest points . For example, the business, health and physical education, art, indus- trial education, home economics, and speech collections all display high rates of obsolescence and relatively high rates of circulation fourteen to seventeen years af- ter acquisition of books, for those collec- tions. However, the foreign languages and mathematics collections show low rates of obsolescence and relatively low rates of circulation twelve to thirteen years after volumes have been acquired. There- fore, one must not assume that high rates of obsolescence indicate collection hold- ings that may become very little used, nor can one assume that low rates of obsoles- cence indicate stable, high use. Six of the nine departmental collections in the lower half of column D, collections with relatively high circulation rates at the end of their obsolescence sequence, are also in the upper half of column I, which ranks the range of dispersion of each curriculum-supporting collection. Widely dispersed use of a curriculum-supporting collection apparently tends to elevate that collection's circulation rate at the lowest point of its obsolescence sequence. Column F shows how well the individ- ual collections' graphs of circulation pat- terns throughout the obsolescence se- quence matched the corresponding graph for overall library circulation. Circulation plotted as a function of time for each aca- demic instructional department's curriculum-supporting collection demon- strated many variations among depart- mental collections . Measured against overall library circulation and in terms of the coefficient of correlation,. however, de- partmental collections generally correlate highly. The social studies/ anthropology collection reveals a . 98 coefficient of corre- lation to the overall library graph of circu- lation during the obsolescence sequence. In this study, therefore, the graphic pat- terns of the social studies/ anthropology November 1983 department curve can be used to repre- sent the use patterns of overall library cir- culation. Columns G, H, I, and J present interre- lated data on the factor of dispersion. These columns reflect data showing book use as a function of the magnitude and range of the circulation of an academic de- partment's curriculum-supporting collec- tion by patrons not majoring in that de- partment's curricular programs. Dis- persion appears to be a very important fac- tor to consider in the priority ranking of books that have been requested for acqui- sition. Column G identifies the number of Dewey divisions that directly relate to the subject content of the courses offered by a given academic department. These Dewey divisions form the fundamental curriculum-supporting collection for a particular department, that is, the collec- tion of books in subject areas most closely related to the unique curricular program of that particular department . The social studies/ anthropology collection has the largest number of Dewey divisions that di- rectly relate to subject areas within its cur- ricular programs, while the music depart- ment has the smallest number of such Dewey divisions. While some variations of the boundaries of a particular academic department's cur- ricular domain are likely, the overall sub- ject profile of that particular department is likely to remain reasonably stable. Hence, column H identifies the number of Dewey divisions considered unique to a particu- lar department which are not significantly circulated by users associated with other instructional departments. For instance, it was found that twenty-six Dewey divi- sions are, for the most part, relied upon only by patrons associated with the social studies/anthropology department. How- ever, there are eight Dewey divisions (compare column Hand G) that support the social studies/anthropology curricu- lum but are utilized by students not asso- ciated with the social studies/anthropol- ogy department. At the other end of the scale, no Dewey divisions in the business department's collection are used exclu- sively by business majors. Columns H and G also show the extent to which individual academic-department collections with many Dewey subject divi- sions may maintain curricular individual- ity. Relatively exclusive use of the re- sources in those Dewey divisions associated with a particular instructional department shows the extent to w_hich certain collection development requests will need to be considered simply on the basis of departmental curriculum support. No significant dispersion of a given num- ber of Dewey divisions in a departmental curriculum-supporting collection sug- gests the need for curricular guidelines as part of the collection development review process. The larger the number of Dewey divisions unshared by other instructional departments, the more that acquisitions decisions must be based upon the require- ments of curricular programs within the particular instructional department. Column I identifies the total number of instructional departments using each in- dividual instructional department's curriculum-supporting collection. For in- stance, a total of fourteen teaching depart- ments use Dewey divisions that are de- fined as unique subject areas relating to the curricular programs of the social stud- ies/anthropology department. One analy- sis of this factor shows that books in the curriculum-supporting collections of cer- tain academic departments are used not only to support the informational needs of that department's own students but also of other students not associated with that particular department. Thus, while the collection development selection process should give priority consideration to books that are needed to support the cur- ricular programs of a particular instruc- tional department, the argument favoring the acquisition of a given book is strength- ened whenever it is used by patrons asso- ciated with other departments. Column J identifies the dispersion range of the single Dewey division in each aca- demic department collection with the highest range of dispersion. For instance, one category in the business department collection is also used by students associ- ated with twelve other academic depart- ments. While there are a number of indi- vidual Dewey divisions that are used by Literature Obsolescence 429 students associated with departments other than the department drawing curric- ular support from these subject classifica- tions, the dispersion thus reflected may be concentrated in only a few instructional departments. The figures in column J, however, provide information that will as- sist in assigning relative importance to ac- quisitions requests on the basis of general use. The opposite argument can then be made that if there are none or very few dis- persed Dewey divisions related to a given academic department, acquisitions deci- sions concerning its curriculum- supporting collections will be almost ex- clusively based on the importance of the library's support of that department's cur- ricular program. To determine university-wide book-use relationships and departmental needs for library support of curricular programs in which students may take an academic de- gree, circulation and collection-size data were further analyzed. Figure 2, Aca- demic Collection Use Analysis, presents the findings of this usage investigation. The number of volumes in each individual academic department's curriculum- supporting collection, already identified as the group of pertinent Dewey divi- sional classifications, was compared to the library's total holdings in curriculum- supporting collections so that the size of each departmental collection could be ranked as a percentage of the total size of all departmental collections. Likewise, a twelve-month total of circulations for each academic department's collection was compared to the library's total twelve- month circulation of volumes from curriculum-supporting collections in or- der to rank circulations from each individ- ual academic department's collection as percentages of the library's total circula- tions of curriculum-supporting volumes. The ratio of an academic department's percentage of total curriculum-supporting circulations to its percentage of total curriculum-supporting holdings provided a coefficient of usage, a mathematical quo- tient reflecting the relative university- wide patron demand for books from each academic department's collection. In a similar manner, the twelve-month 430 College & Research Libraries November 1983 Departmental Major's Use Low (0.00 - 0.60) Standard (0.61 - 1.80) ,..... 0 c: Business ("") .c bOI Speech "" ::Z::r-i ("") r-i .._, ,..... 0 ("") Education "d ~r-i l1l QJ "d I Industrial rJ) ~ ::::;, I"Clr-i .j..IOC) QJ C/) "d 0 "" .._, ~ I >. .j.J "" ,..... rJ) 0 ~ 00 Biology QJ 0 > "" ~ ) I ::::;, 0 ...:lO 0 0 .._, I Key For Collection Use Analysis Departmental Collection Art Biology Business Education English Foreign languages Health and physical education Home economics Industrial education Library science Mathematics Music Physical science Psychology Social studies/anthropology Speech Health and Physical Education Home Economics Psychology Library Science Education Foreign Languages Mathematics Music Physical Science University-Wide Coefficient of Usage 1.34 0 .63 1.77 0.84 0.79 0.34 2.20 1.33 1.01 0.93 0.71 0.74 0.58 2.70 0 .83 2.03 FIGURE2 Academic Collection Use Analysis High (1. 81 - 5. 00) Art Social Studies/ Anthropology English Departmental Majors ' Coefficient of Usage 2.55 0.56 0.51 0.44 4.72 1.75 1.73 1.61 0.49 1.31 0.84 1.14 0.81 0.75 4.15 0.45 totals of declared majors in each academic department were compared to the univer- sity's twelve-month total of declared ma- jors in order to rank the number of de- clared majors in each department as a percentage of the year's total number of declared majors. The ratio of each aca- demic department's percentage of total curriculum-supporting circulations to its percentage of total declared majors pro- vided a coefficient of departmental ma- jors' usage. By grouping departmental collections according to their low, standard, or high university-wide use and their low, stan- dard, or high declqred majors' use, the cross-comparison format of figure 2 was generated. Departmental collections ap- pearing midway ?P _the university-wide axis reflect soundly developed collections for which the relative use rates match the relative collection sizes; current acquisi- tions and weeding policies should proba- bly be continued in future collection de- velopment. Departmental collections appearing at the bottom of the university- wide use axis reflect overdeveloped collec- tions for which the relative use rates are less than the relative collection sizes; ac- quisitions should probably be more care- fully screened and judicious weeding of unused volumes should be applied in fu- ture collection development. Departmen- tal collections appearing at the top of the university-wide use axis reflect underde- veloped collections for which relative use rates are greater than relative collection sizes; judicious broadening of acquisitions policies might be considered for future collection development. On the departmental majors' use axis, collections appearing midway reflect stan- dard use by declared majors in the depart- ments listed. Collections appearing at the left end of the departmental majors' use axis indicate light use by declared majors; the curriculum and/or the teaching meth- ods for the departments listed are not "library-intensive." Collections appear- ing at the right end of the departmental majors' use axis indicate heavy use by de- clared majors; the curriculum and/or teaching methods for the departments listed are "library-intensive." Review of an academic collection use analysis table based on circulation and declared-major data for any given aca- demic library should indicate, to those re- sponsible for collection development, the relative importance of the library's sup- port of various departmental curricular programs in addition to the relative im- portance of the library's support of aca- · demic collections showing varying de- grees of dispersed use. The use-support relationships are dynamic, changing as university emphases change. Academic li- brarians must keep both campus-wide needs and degree-program needs in mind as collection development decisions are Literature Obsolescence 431 made. It is important to note, however, that figure 2 was compiled from 1982 cir- culation data. Yet its results reinf6rce the use findings from the 1980 circulation data presented in figure 1. Thus, replication of the study at a different time and for differ- ent purposes tended to support finds of low, standard, and high usage of groups of library holdings affiliated with aca- demic departments. CONCLUSIONS AND RECOMMENDATIONS Initial evaluation of all the above- mentioned data made it tempting to offer a variety of general policy directives. However, it is not the purpose of this study to determine the minimum size of core collections or to specify the number of titles that should be acquired in any spe- cific subject area. In both instances, such determinations require subjective judg- ments from librarians, taking into account the existing size of the resources in a par- ticular subject area, the level and depth of an instructional program, and the overall goals of a university. However, consider- ation of the number of titles to be pur- chased in a particular subject area should be based not only on use-related factors but also on the number of titles being pub- lished in that subject area. First, a study of circulation patterns sug- gests that "basic" or "core" or "essen- tial'' collections should be developed after review of a number of integrally related use factors rather than developed around arbitrarily specified numbers of volumes per student. Analysis of the use patterns in c.urriculum-supporting collections avoids the subjectivity with which titles are recommended in standard library guides for collection development. Stan- dard catalogs and lists of ''best books,'' al- though published regularly and used widely, cannot reflect individual academic libraries' local use patterns and corre- sponding patron needs as effectively as a faculty member's preference or a librari- an's knowledge of actual patron use and of strengths and weaknesses in a particu- lar collection. Circulation data also seemed to indicate 432 College & Research Libraries that certain books were being acquired in certain subject areas of the Dewey divi- sions in preparation for later academic use. Certain collections in the low-to- middle range of immediacy and intensity of use continued to show regular growth in acquisitions. This behavior seemed to indicate that some books in these collec- tions were acquired on the basis of per- ceived potential use, not on the basis of ac- tual curriculum-supporting use. This study was structured from the out- set to specify curriculum-supporting col- lections that reflect closed subject areas of various academic departments' speciali- zations. Use data, though, indicated sig- nificant dispersion of volumes found in several so-called closed academic collec- tions. In other words, use data seemed to reflect the interdisciplinary dispositions of academic library patrons. Initially, disper- sion appeared to be a balancing factor that prevented rapid rates of obsolescence in collections with wide circulation among patrons who did not show declared ma- jors in the departmental curricula that these collections support. It seems, how- ever, that if usage is widely dispersed early in the library life of certain compo- nents of a collection, then the early part of that collection's obsolescence sequence will vary greatly from the overall library pattern. If, on the other hand, usage is well dispersed throughout the entire li- brary life of certain components of a collec- tion, then that collection's obsolescence sequence will tend to correspond to the overalllib~ary pattern. Continued review of the particular dispersed components of a collection is necessary to verify this ob- servation. Data from figure 1, the Obsolescence Analysis Matrix, were evaluated to deter- mine whether or not they revealed any in- formation valuable to the forming of col- lection development guidelines. Column B, ''Year of Peak Circulation,'' showed the immediacy with which books in each academic collection circulated. Column C, ''Rate of Peak Circulation,'' showed, in terms of circulations per one hundred vol- umes, the intensity of circulations in each academic collection during the years iden- tified in column B. Since the probable im- November 1983 mediacy of need and the probable inten- sity of use are important to the making of collection development decisions, these data were combined in figure 3, the Imme- diacy/Intensity Table. Combined, the data reveal three distinct groups or communi- ties of use: first, high intensity (110-150 circulations per hundred, multiple circula- tions of volumes) and first-year immedi- acy; next, moderate to high intensity and first to second-year immediacy; finally, low to moderate intensity and first to third-year immediacy. Books in the de- partmental collections making up the high-intensity and first-year-immediacy group show the earliest and heaviest use; therefore, priority should be assigned to their acquisitions when faculty or librari- ans request them. Books in the other two groups show less immediate need; there- fore, other factors would be necessary to encourage early acquisitions of new books in the subject areas comprising these de- partmental collections. The dispersion data from columns G, H, I, and J of figure 1 were arranged to reflect the actual range of dispersion and rate of circulation of the dispersed Dewey divi- sions within each academic curriculum- supporting collection. Dewey divisions not assigned to curriculum-supporting collections were eliminated from this group so that a view of multidepartmental circulation of Dewey divisions with curriculum-supporting use could be framed. Such a view represents the aca- demic institution's direct influence on book use. First, only those Dewey divi- sions outside a given instructional depart- ment's assigned collection but circulated by that department's majors were listed. Next, the total circulation by declared ma- jors of books from collections outside each instructional department's curriculum- supporting collection was tabulated for each department. Finally, "significant" dispersion was accepted to begin when- ever an "outside-major" circulation rate of more than 1. 90 percent of the total circu- lation by each department's declared ma- jors occurred. The entire year's circulation figures for books in the Dewey divisions showing such significant dispersion of use could then be compared to the year's total circulation of books to determine the per- cent of total library circulation for which each range of narrowly to widely dis- persed subject divisions accounted. This comparison is reflected in figure 4, the Dispersion Table. Circulation of bo<;>ks in thirty-two of the one hundred Dewey di- visional classifications showed both cur- ricular support and significant dispersion. Books in Dewey divisions with signifi- cantly dispersed use by majors in one or more departments outside the depart- ment whose curriculum these Dewey divi- sions support accounted for slightly more than 66 percent of the total library circula- tion. In other words, two-thirds of total li- brary use was found to stem from circula- tion of books in one-third of the possible Dewey divisional classifications. To be certain that the circulation data, on which collection development recom- mendations would be made, reflected user preferences not solely dictated by curriculum-support requirements, a final tabulation was made of the year's total cir- culation grouped by patrons' declared major departments but without regard to the curricula of those departments. As presented in figure 5, the Circulation Commonality Table, this tabulation re- flects any circulation of books by declared majors from all the departments making up one of the four broad academic discipli- nary divisions of the university: fine arts and humanities; life sciences; pure and applied sciences; and social and behav- ioral sciences. The tabulation was ar- ranged to show whieh Dewey divisions recorded any circulation of books by ma- jors from all departments in four, three, two, or one of the broad academic disci- plines. As the Dispersion Table reflects a curriculum-generated core collection that accounts for 66 percent of book use, so the Circulation Commonality Table shows a user-generated core collection. This user- interest core collection accounted for twenty of the possible one hundred De- wey divisions and 53.5 percent of the year's total library circulation. Of the thirty-two dispersed curriculum-support- ing Dewey divisions appearing in the Dis- persion Table, eighteen also appear in the top twenty Dewey divisions reflecting Literature Obsolescence 433 user interests in the Circulation Common- ality Table. The remaining fourteen dis- persed Dewey divisions accounted for 15.5 percent of the year's total library cir- culation. The user-generated collection and the curriculum-supporting collection together accounted for 69 percent of the year's total library circulation and thirty- four of the one hundred Dewey divisional subject classifications. The well-known Pareto Principle, echoed in True swell's descriptions of the use of a library's total collection, seems to be. refined by this breakdown. Rather than 20 percent of the library's holdings satisfying 80 percent of user demand, roughly one-third of the li- brary's subject classifications satisfied two-thirds of user demand. Additionally, the use of precisely identified subject clas- sifications as the basic units of comparison in this study enables one to determine which one-third of the library holdings ac- count for two-thirds of user demand. Within these collections is the basis for sound use-based collection development decisions. Reference to any of the bibliographic utility data bases or to a CIP can yield the identity of a book's Dewey decimal divi- sional classification (or the corresponding Library of Congress classification). To be- gin the process of formulating collection development decisions, the subject-area classification for each requested book should first be identified by the Dewey di- visional classification. Then, reference should be made to the Immediacy/Inten- sity Table (figure 3) in order to determine the relative priority ranking of the re- quested books in terms of time and use. Next, the Dispersion Table (figure 4) and the Circulation Commonality Table (fig- ure 5) should be referred to in order to de- termine the relative priority ranking of the requested books in terms of curricular- generated and user-generated dispersed use. Combined, all this information can be put into a simple Acquisition Priority Weighting (APW) formula to aid in the collection development process. The basic formoftheAPWisAPW =[I+ I]+ [ + D] + [C]; it is a sum of ranking values as- signed to the position of a book's curriculum-supporting collection and Range of Highest Circulation Rate per One Hundred Volumes Acquired: 10 20 30 40 so 60 70 80 90 100 110 120 130 140 Art Mathe- Library Home Business rna tics Science Economics Psychology --------- --- ~- ------- DIVISION DIVISION DIVISION OF FINE OF PURE OF SOCIAL ARTS AND AND AP- AND BE- HUMAN- PLIED HAVIORAL Foreign Music TIES SCIENCES SCIENCES Industrial Languages --------- ------- --------- Education Social Studies/ Anthro- Speech pology. LIBRARY AVERAGE ---------- DIVISION OF LIFE SCIENCES English --------- Health and Biology Physical Education Education Physical Sciences 20 30 40 so 60 70 80 90 100 110 120 130 140 lS~ ~ 0 ·.c a:: =s u Oil ~ 0 0.. 0.. ;:I (/) 5 ;:I 8 "§ u ] 1::: ~ ~ 0.. ~ 0 .5 "' 1::: -~ "' ~ 0 ;a 1::: .9 ~ >. ~ ~ 0 !3 ~ >. ~ ::.<: ~ "' ~~d2 lfl ' ~g~ lfl ' ~R- ~ lfl .. g ;,:._::l "'Oil~ .;..;~ · u; <~~ Literature Obsolescence 435 Dewey divisional classification on the Im- mediacy/Intensity, Dispersion, and Circu- lation Commonality tables. The horizontal columns of the Immedi- acy/Intensity Table refer to the year in which peak usage occurred in a given aca- demic collection. The vertical columns identify circulations per hundred books within the same academic collections~ All Dewey divisional classifications in each academic department's collections are uniformly assigned a priority value based on the whole collection's immediacy and intensity ranking. First-, second-, and third-year peak usage are given rankings of 3, 2, or 1 respectively. Circulations per hundred are given rankings from 0 to 1 to reflect their circulation ratios; collections with peak circulations per hundred of 40, 60, and 110, for example, are given rank- ings of 0.4, 0.6, and 1.1 respectively. The sum of each department's year-of-peak- usage ranking and its rate-of-peak-usage ranking will comprise the Immediacy/In- tensity [I + I] factor in acquisitions prioriti- zation. For example, second year peak us- age at 50 circulations per hundred would be prioritized thus: 2 + 0.5 = 2.5. The next component of the APW for- mula is the Dispersion [1 + D] factor. Books in subject-area classifications with no dispersed use, in other words, with only major-department use, are assigned a dispersion factor of 1 to reflect one- department use. Books in subject-area classifications showing use by two or more departments are given priority rank- ing values by adding the major- department dispersion factor (1) to the product of values assigned to the ''Percent of Total Library Circulation" and "Range of Departments Using Dispersed Dewey Divisional Classifications" categories. A value of 1 is assigned to the ''High Circula- tion" category; 0.75, 0.50, and 0.25 re- spectively are assigned to the ''Moderate to High,'' ''Low to Moderate,'' and "Low" circulation categories, the hori- zontal columns of the Dispersion Table . The vertical columns of this table, reflect- ing ranges of dispersion, are given values of 1, 2, 3, and 4 for "Low," "Low to Mod- erate,'' ''Moderate to High,'' and ''High'' dispersion respectively. Thus, a book z z o...-.. HO E'-< • ::r:~oo (.!) I iH ~ ::r: u • ::r: (.!) r:>:\D H'-' u HZ ::r:o...-.. HO 0 E'-< • E-~ E'-< u . :~ r:>:H'-' l>lU Cl §! 0 l:zl H E'-< j ~SO' :::> Cl E'-< • ~ §! j 1 H :::>~ u 0 u . e'-:N >< H '-' ~ s u RANGE OF DEPARTMENTS USING DISPERSED DEWEY DIVISIONAL CLASSIFICATIONS: LOW DISPERSION (2-4) 780 (2/3.3%) I.,OW TO MODERATE DISPERSION (5-7) 330 (7/2.8%) 970 (6/2.7%) 320 (5/2.2%) MODERATE TO HIGH DISPERSION (8-10) 360 (9/2.7%) 150 (8/2.8%) 810 (8/2.6%) HIGH DISPERSION (11-14) Fiction (14/7.3%) 650 (13/5.7%) 370 (13/5.3%) 790 (14/3.4%) 300 (12/4.0%) 610 (11/3.6%) ~ ~--~~--------------------4---------------------~-------------------+--------------------~ ~ 390 (4/1. 8%) 530 (2/0. 9%) :;J ~....... 800 (4/1. 5%) 730 (2/0.8%) ~ ~ c: 000 (4/0.8%) 540 (2/0. 7%) E'-< :l: j ~ 940 (3/1. 7%) 290 (2/0.4%) ~ S B c: 34o (3/LO%) 58o (2/0.4%) ~ e 5lo (2/1.4%> l~ (2/0.4%) ~ u 640 (2/1.3%) 500 (2/0.3%) ~ 910 (2/1.0%) 670 (2/0.3%) ~ .ll_Q (2/0. 2%) 740 (6/1. 5%) 820 (5/1.9%) (Note: Dewey divisions with at least two department dispersion--major department and one other-- account for more than 66% of total library circulation. Dewey divisions with only major department use account~r less than 34% of total library cir~ulation.) ~ Vl 0'1 (') 0 = f't) (JQ f't) ~ ~ f't) Ill f't) e: n =- ~ 0: ~ ..... f't) Ill z 0 ~ ~ f't) tot ~ 1.0 00 Vl Literature Obsolescence 437 from a subject classification in ''Low to Moderate Circulation" and "Moderate to High Dispersion'' would be prioritized 1 + (.75 X 3) = 3.25. The third component of the APW for- mula is the Circulation Commonality [ C] factor. Dewey divisional classifications appearing in the categories of one, two, three, or four broad-academic-discipline circulation receive priority values of 1, 2, 3, and 4 respectively. Books classified in the 720, a Dewey division appearing in the two-academic-discipline circulation group on the Circulation Commonality Table, would receive a priority ranking factor of 2. The formula, then, for each requested book's APW (acquisition priority weight) is as follows: APW = [Value of Peak Usage Year+ Peak Number of Books Circulated per 100 Acquired/100] + [1 + (Value of Library Circulation Rank x Value of Dispersion Range Rank)] + [Value of Circulation Commonality Rank] Thus, books appearing in the Dewey di- visional subject classification 650 would be prioritized in the following manner: APW650 = [3 + 140/100) + [1 + (.75 x 4)) + [4) = 4.14 + 4 + 4 = 12.14 To determine the usefulness or applica- bility of the APW formula at libraries other than the William Allen White Library of Emporia State University, a Spearman Rank Order Correlation (rho) test was conducted to compare the use patterns of the Emporia State University, Bucknell University, Colorado State University, and University of Pittsburgh libraries. In collection sizes, these university libraries range from 400,000 to 600,000 volumes (Emporia State and Bucknell) to over 1,000,000 volumes (Colorado State) to over 2,000,000 volumes (Pittsburgh). Ac- cording to the Spearman Rank Order Cor- relation test, the use-pattern correlation between the Emporia State library and each of the other libraries is significant. (See appendix B for each individual corre- 438 College & Research Libraries Dewey Divisions Showing Any Circulation By Majors From All Departments In All Four Academic Disciplines November 1983 Dewey Divisions Showing Any Circulation By Majors From All Departments In· Three Academic Disciplines ~~~~L~~~~~~~~-_1--~~::~~~~~-~~~~~:~~~!~~~~~~~~~~~- -~~~~r-~~~~~~~~-_L __ ~::~~~~~-~~~~~:=~~!-~~~~~~=~~~-- 150 Psychology 390 Home Economics 300 Social Studies/ Anthropology 620 [No Curricular Program) 330 Business 640 Home Economics 360 Social Studies/ Anthropology 810 English 370 Education 970 Social Studies/ Anthropology 610 Health and Physical Education 650 Business 790 Health and Physical Education (Speech) Fiction English Dewey Div sions Showing Any Circulation By Majors From All Departments In Two Academic Disciplines . I ~~~~!-~~~~~~~~-+--~~::~=~~~-~~~~~:~~~~-~~!!~~!~~~ 290 f Socal Studies/ Anthrolopogy 630 t [No Curricular Program) 720 I Industrial Education 740 I Art 820 I English 910 f Social Studies/ Anthropology I I I I I I I I I I I I I I I I I I I I I Dewey Divisions Showing Any Circulation By Majors From All Departments In One Academic Discipline I -~~::~!-~~~~~~~~-+--~~~~~=~!~~-~~~~~~=~~!-~~!!~==~~~ 130 Psychology 380 Social Studies/ Anthropology 500 Physical Science 510 Mathematics 530 Physcial Science 540 Physical Science 510 Biology 580 Biology 590 Biology 660 [No Curricular Program) 690 Industrial Education 750 Art 780 Music 800 English 940 Social Studies/ Anthropology Fed Gov' t Doc [No Curricular Program) FIGURE 5 Circulation Commonality Table lation value.) This significant correlation supports the hypothesis that use patterns are generally applicable. Emporia State University's APW values for the one hun- dred Dewey divisions should be broadly consistent with other universities' APW values for similar subject classifications. (See appendix C for Dewey to LC conver- sion of subject classifications pertaining to academic departments.) Curricular differ- ences may slightly alter individual APW values, but within the high, middle, and low groups of APW values, the same sub- ject classifications should appear, accord- ing to the Spearman correlations. (See ap- pendix D for the APW values of the one hundred Dewey divisions for Emporia State University based on 1980 circulation data and curricular offerings.) Admittedly, there would be weaknesses in using any formula indiscriminately. However, the APW formula provides quantifiable rationale that would be useful in the often subjective process of making collection development decisions . The APW formula rests on the assumption that various types of use are identifiable and that use justifies collection develop- ment, with heavier use in certain areas justifying heavier acquisitions in those ar- eas. Used widely and adapted to the cur- ricular programs of an individual univer- sity, the APW formula can provide objective guidelines for the bulk of collec- tion development decisions, decisions that must fit the most useful books re- quested into limited acquisitions budgets. As a refinement of Pareto's Principle and Trueswell's 80/20 rule, the circulation commonality and dispersion analyses showed that over 50 percent of a year's to- tal library circulation is reflected in 20 per- cent of the Dewey divisional classifica- tions, and 69 percent of a year's total library circulation is reflected in 34 percent of the Dewey divisional classifications. Appendix E presents a possible method of applying such findings to the budgetary considerations necessary in collection de- velopment. Using the total of the APW values computed from William Allen White Library circulation records at Em- poria State University, one may ascertain a relative potential-use factor by dividing each APW value by the total of all APW values and multiplying the result by 100. The resulting percentage demonstrates the relative use an acquisition in a given Dewey divisional classification might rea- sonably be expected to ·have. Literature Obsolescence 439 If academic libraries carefully follow prioritized potential-use patterns in dis- pensing their book budgets-proportion- ately allotting the first half of available funds to those subject areas showing the first half of expected book use, allotting the first two-thirds of available funds to subject areas showing the first two-_thirds of expected book use, and so forth-then sound collections can be developed in support of local-use patterns and demon- strated patron needs. REFERENCES AND NOTES 1. The general obsolescence pattern consists of three elements: first, a high point of use sometime within the first four years after an item's acquisition; second, a decline in use over a period of years (roughly ten to fifteen years) until a low point of use is reached; and, third, a period of relatively steady, low use lasting, as found in this study, at least through an item's twenty-second year of shelf life. 2. "Dispersion," in this study, refers to the use of books that belong to a given academic depart- ment's curriculum-supporting collection by students not majoring in that department's curricu- lum . The term is distinct from "scatter," identified by Wilfred Ashworth ("The Information Ex- plosion," Library Association Record, 76:63-68, April197 4) as the "order of decreasing productivity of papers relevant to a given topic" across a large number of periodicals that publish articles per- taining to a given discipline . As Samuel Bradford's law points out, journals principally devoted to the discipline carry, as a small group, more articles on a relevant topic than all the journals in the larger group of publications with a less-restricted focus. Thus, "dispersion" is a term for a type of user behavior while ''scatter'' is a term for a factor of diminishing productivity in relation to infor- mation gathering for a given topic. 3. Discussion and application of the "hard/soft" distinction are found in William E. McGrath's "Re- lationship Between Hard/Soft, Pure/Applied and Life/Nonlife Disciplines and Subject Book Use in a University Library," Information Processi~ & Management, 14:17-28, 1978, and in George V. Hodowanec's "Analysis of Variables Which Help to Predict Book and Periodical Use," Library Acquisitions: Practice and Theory, 4:75-85, 1980. 4. Charles F. Gosnell, "Obsolete Library Books," The Scientific Monthly, 64:421-27 (May 1947). 5. Charles F. Gosnell, "Systematic Weeding," College & Research Libraries, 11:137-38 (Apr. 1950). 6. Gamet Hardin, "The Doctrine of Sufferance in the Library," College & Research Libraries, 8:120-26 (Apr . 1967). 7. B. C. Brookes, "Obsolescence of Special Library Periodicals: Sampling Errors and Utility Con- tours," Journal of the American Society for Information Science, 21:320-29 (Sept.-Oct. 1970). 8. B. C. Brookes, "Numerical Methods of Bibliographical Analysis," Library Trends, 22:18-43 (July 1973). 9. Maurice B. Line and Alexander Sandison," 'Obsolescence' and Changes in the Use of Literature with Time," Journal of Documentation, 30:283-350 (1974) . 10. Kenneth Hodges, "Chronological Order," Library Review, 25:57-62 (Summer 1975) . APPENDIX A: SUGGESTED READING LIST Brookes, B. C. "Numerical Methods of Bibliographic Analysis," Library Trends 22:18-43 (July 1973). ____ . ''Obsolescence of Special Library Periodicals: Sampling Errors and Utility Contours,'' Jour- nal of the American Society for Information -Science 21:320-29 (Sept.-Oct. 1970). 440 College & Research Libraries November 1983 Burlington, William Stone. "The Obsolescence of Engineering Books ." Masters thesis, Columbia . Univ., 1951. Gosnell, Charles F. "Obsolete Library Books," The Scientific Monthly 64:421-27 (May 1947). ____ . "Obsolescence of Books in College Libraries," Collection Ma1Ulgement 2:167-82 (Summer 1978) . ____ . "Systematic Weeding," College & Research Libraries11:137-38 (Apr. 1950). Hardin, Gamet. "The Doctrine of Sufferance in the Library," College & Research Libraries 8:120-26 (Apr. 1967). Hodges, Kenneth. "Chronological Order," Library Review 25:57-62 (Summer 1975) . . Line, Maurice B., and Sandison, Alexander. " 'Obsolescence' and Changes in the Use of Literature with Time," Journal of Documentation 30:283-350 (1974). Morse, Phillip, and Elston, Caroline. "A Probabilistic Model for Obsolescence," Operations Research 17:36-47 (1969). . Rouse, S. H., and Rouse, W. B. "Analysis of Monograph Obsolescence at Two Levels of an Interli- brary Loan Network," Information Processing and Management 15, no.5:219-25 (1979). Seymore, Carol A. "Weeding the Collection: A Review of Research on Identifying Obsolete Stock; Monographs," Libri 22, no .2:137-48 (1972) . APPENDIX B: SPEARMAN RANK ORDER CORRELATION A Spearman Rank Order Correlation (rho) test and at-test were conducted to determine the con- stancy of circulation patterns at four academic libraries. The curriculum-supporting Dewey divisional classifications were grouped according to the primary instructional departments they serve and then ranked according to the percentage of total library circulation that they accounted for. Comparisons were made between subject area classifications at Emporia State University and three other schools: Bucknell, Colorado State University, and the University of Pittsburgh. After the comparisons were made, the subject-area classifications were grouped according to the departmental curriculum- supporting alignment at Emporia State. Then, based on individual subject-area circulation figures from the other schools, rankings were made according to each departmental group's percentage of total library circulation. The rank order correlations between circulations of subject-area classifications aligned by individual instructional departments at Emporia State and three other schools are illus- trated below. The University of Pittsburgh's system of reporting circulation by subject areas made it necessary to use broader bases of comparison between Emporia State and Pittsburgh than the depart- mental bases otherwise compared . SPEARMAN RANK ORDER CORRELATION and T-TEST: Emporia State University and Bucknell University Emporia State University and Colorado State University Emporia State University and University of Pittsburgh rho = 0.645; t = 3.158 (significant at P ~ .01) rho = 0.688; t = 3.457 (significant at P ~ .01) rho = 0.845; t = 4.74 (significant at P~ .01) Therefore, all correlations show a probability of significant rank order correspondence at levels of 99 percent or more. Literature Obsolescence 441 APPENDIX C: DEWEY TO LC CONVERSION CHART DEPARTMENT DDC SUBJECT AREAS LC DEPARTMENT DDC SUBJECT AREAS LC Art 700 The Arts CJ Home 390 Customs, Etiquette, GR 730 Fine & decorative N Economics Folklore 740 Plastic arts, sculpture NB 640 Home Economics GT 750 Drawing NC TT950-TT999 760 Painting & paintings ND TX Graphic arts NE Printmaking & prints NK Industrial 600 Technology NA NX Education Buildings TT1-TT160 670 Manufactures Tl-T53 TT697-TT924 Civic & landscape art 680 Architecture T201-T995 Biology 570 Life Sciences GN49-GN296 690 Til 580 Physical Anthropology QH3Ql-QH67l 710 TS1-TS154 590 Biology QK 720 TS195-TS1982 Physiology QL TS222Q-TS2283 Microbiology QP TT161-TT267 Botanical Sciences QR TT387-TT695 Zoological Sciences Z48 610* Medical Sciences R Zll6-Z276 RB-RG R.J-RM Library 000 Generalities AE RS-RT Science Library & Information RV 010 Science AG RX 020 AI RZ 030 AM 040 AN Business 330 Economic s HB1-HB847 050 AP 340 Labor HB37ll-HB3840 060 AS 650 Financial HC 070 AY. Land HD 080 AZ Cooperatives HF5001-HF6191 090 CD921-CD4279 HG 'PN4699-PN565Q · Public finance K Q300-Q385 International economics KD Zl-Z39 Production KE Z278-Z8999 Macroeconomics KF I Mathe- 1 510 I Mathematics Law QH75-QH77 QA Management T55 .4 -T60 matics Z43-Z45 I Music 1780 1 Music Z49-Zl04 M ML . This subject area(s) supported no curricular program in MT this study. DEPARTMENT DDC SUBJECT AREAS LC DEPARTMENT DDC SUBJECT AREAS LC Education 370 Education L-LH Physical 500 Pure Sciences CE w Science Natural History LT 520 Microscopy GA Astronoaiy English 400 Language CN 530 Physics GB40Q-GB2998 410 Linguistics Pl-P86 Chemistry 420 Literature Pl01-P900 540 Geology GC 800 PE Meteorology 810 PN1-PN1551 550 General Hydrology Ql-Q299 820 PN160Q-PN1999 Paleontology PN3311-PN4500 560 Paleozoology QA PN6011-PN6790 QII-QE PR QH1-QH74 PS QH201-QH278 Zl05-Zll5 620* TA Fiction PZ TC TF-TG Foreign 430 Foreign languages P901-Pl081 TJ-TL Languages 440 and literature TN 450 PA 630* S-SK 460 TS212Q-TS2159 470 PB TT30Q-TT385 480 660* TP 490 PC no• TR 830 I Psychology I ~~~ I Psychology 840 PD BF 850 860 PF 870 Social 100 Philosophy and related B-BD 880 PG 890 Studies/ 110 disciplines BH-BJ PH Anthro- 120 Metaphysics PJ-PH pology 140 Cosmology PQ 160 Aesthetics PT 170 Ethics I 180 Logic Health & 610 Medical Sciences GV 190 Physical 790 Recreation QH 200- Religion BL-BX Education RA421-RA954 290 RA1001-RA1270 300- Social Sciences H-HA 320 . Books related to physical education programs only . Pol)u.lation Demography HB848-HB3700 * This subject area(a) supported no curricular prograa in this study . 442 College & Research Libraries November 1983 APPENDIX C, CONTINUED DEPARTMENT Social Studies/ Anthro- pology Speech DOC 350 360 380 901)... 990 Portion of 000 Portion of 790 SUBJECT AREAS Political Science Public Administration Military art & science Social problem services association Commerce, Communications, Trar;sportation History & Geography Coaa:ounication Performing Arts LC HE,HF1-HF2701 HJ-HX RA5-RA41Q RA960-RA99B T54-T55 J-JX U-UH V-VM c-cc CD1-CD920 CR-CT D-DX E-G GB1-GB399 GB5000-GB5030 GN700-GN875 Z4l-Z42 P87-P96 PN1560-PN1590 PN200D-PN3299 APPENDIX 15: ACQUISITION PRIORITY WEIGHTING (APW) This appendix provides APW (acquisition priority weighting) values for the one hundred Dewey second-level divisions based on the analyses of book circulation and curriculum offerings at Emporia State University in calendar-year 1980. These values can be used for relative comparisons to support subjective judgments in collection development decisions. Even though a Spearman Rank Order Cor- relation (rho) test shows significant statistical correlation between the rank ordering of Dewey divi- sional circulation at Emporia State and at three other universities, other academic libraries interested in using this formula may want to develop an immediacy/intensity table, a dispersion table, and a circulation commonality table based on the specific curricular structure and emphasis of their own universities . Individual academic library rankings of immediacy and intensity of use, dispersion of use, and commonality of use would not help verify the findings of the Emporia State University analy- sis but also would provide individual academic libraries with directly applicable data tailored to partic- ular universities. RANKED BY DEIIEY DIVISIONS 000 - 4.85 010 - 4.6 020 - 4.6 030 - 4 . 6 050 - 4.6 060- 4.6 070 - 4.6 080 - 4.6 090 - 4.6 100- 3.5 !10- 3.5 120 - 3. 5 130 - 6.3 140 - 3.5 150 - 10.8 160 - 3.5 170 - 3.5 180 - 3.5 190 - 3.5 200 - 3.5 210 - 3.5 220 - 3.5 230 - 3.5 240 - 3.5 250 - 3.5 RANKED BY APW FORIIULA 12.4 - 650 10.8 - 150 10 . 4 - 330 10.4 - F 9.6 - 370 9.5 - 300 9. 2 - 610 9.2 - 790 9.0 - 360 8.45 - 390 8.45 - 640 7. 7 - 740 7.5 - 970 7.1 - 620 6. 9 - 810 6.35 - 720 6.3 - 130 " 6.2 - 750 5. 75 - 510 5 . 75- 910 5.75- 290 5.65 - 340 5.45 - 730 5 .35 - 690 5.2 - 700 RANKED BY RANKED BY DEIIEY APW DIVISIONS FOR!IULA 260 - 3.5 5.2 - 760 270- 3.75 5.2 -770 280- 3.5 5.0 - 780 290 - 5. 7 5 4. 9 - 820 300 - 9.5 4.85 - 630 310 - 3. 5 4. 85 - 660 320 - 4. 5 4 . 85 - 000 330- 10.4 4.75- 380 340- 5.65 4.75- 940 350 - 3. 5 4. 6 - 010 360 - 9.0 4.6 -. 020 370- 9.6 4.6 - 030 380- 4.75 4.6 - 050 390- 8.45 4.6 - 060 400- 2.4 4.6 - 070 410 - 2.4 4.6 - 080 420- 2.4 4.6 - 090 430- 3.3 4.5 - 320 440 - 3. 3 4 . 35 - 6 70 450- 3.3 4.1 - 600 460- 3.3 4 .1 - 680 470 - 3.3 4.1 - 710 480 - 3.3 3 . 85 - 500 490 - 3. 3 3.85 - 530 500 - 3.85 3.85 - 540 RANKED BY RANKED BY DEIIEY APW DIVISIONS FORIIULA 510 - 5. 7 5 3. 85 - 580 520 - 2.6 3. 75 - 270 530- 3 .85 3.65 - 800 540 - 3. 85 3.6 - 570 550 - 2. 6 3. 6 - 590 560 - 2 . 6 3.5 - 100 570- 2.6 3.5 - 110 580- 3.85 3.5 - 120 590 - 3. 6 3. 5 - 140 600- 4.1 3.5 - 160 610 - 9.2 3.5 - 170 620 - 7 .1 3. 5 - 180 630- 4.85 3.5 - 190 640- 8.45 3.5 - 200 650 - 12. 4 3. 5 - 210 660- 4.85 3.5 - 220 670- . 4.35 3.5 - 230 680 - 4 .1 3. 5 - 240 690 - 4 .1 3. 5 - 250 700 - 5.2 3.5 - 260 710 - 4.1 3.5 - 280 720- 6.35 3.5 - 310 730 - 5.45 3.5 - 350 740- 7.7 3.5 - 900 750 - 6.2 3 . 5 - 920 RANKED BY DEWEY DIVISIONS 760- 7.6 770 - 7. 7 780 - 5.0 790- 9.2 800- 3.6 810 - 6. 9 820 - 4.9 830- 3.3 840 - 3. 3 850 - 3. 3 860- 3.3 870 - 3.3 880 - 3. 3 890 - 3. 3 900- 3.5 910 - 5. 75 920 - 3.5 930 - 3.5 940- 4.75 950 "- 3. 5 960 - 3.5 970 - 7.5 980 - 3.5 990 - 3.5 F - 10.4 RANKED BY APW FORIIULA 3.5 - 930 3.5 - 950 3. 5 - 960 3.5 - 980 3.5 - 990 3. 3 - 430 3.3 - 440 3. 3 - 450 3.3 - 460 3.3 - 470 3.3 - 480 3.3 - 490 3.3 - 830 3.3 - 840 3.3 - 850 3 . 3 - 860 3.3 - 870 3. 3 - 880 3.3 - 890 2.6 - 520 2.6 - 550 2.6 - 560 2.4 - 400 2.4 - 410 2.4 - 420 Literature Obsolescence 443 APPENDIX E: POTENTIAL USE OF PERCENTAGES (APW VA LUES TOTAL OF 475.35 100%) 650 2.61 760 1.09 580 .81 930 . 74 150 2. 27 770 1.09 270 .79 950 .74 330 2.19 780 1.05 800 .77 960 .74 F 2.19 820 1.03 570 . 76 980 • 74 370 2.02 630 1.02 590 . 76 990 .74 300 2.00 660 1.02 100 .74 430 .69 610 1.94 000 1.02 110 .74 440 .69 790 1. 94 380 1.00 120 .74 450 .69 360 1.89 940 1.00 140 .74 460 .69 390 1. 78 010 .97 160 . 74 470 .69 640 1.78 020 .97 170 .74 480 .69 740 1.62 030 . 97 180 . 74 490 .69 970 1.58 050 . 97 190 . 74 830 .69 620 1.49 060 . 97 200 .74 840 .69 810 1.45 070 . 97 210 . 74 850 .69 720 1.34 080 .97 220 . 74 860 .69 130 1.33 090 .97 230 .74 870 .69 750 1.30 320 .95 240 .74 880 .69 510 1.21 670 .92 250 .74 890 .69 910 1. 21 600 .86 260 .74 520 .55 290 1. 21 680 .86 280 .74 550 .55 340 1.19 710 .86 310 .74 560 .55 730 1.15 500 .81 350 .74 400 .so 690 1.13 530 .81 900 .74 410 .so 700 1.09 540 .81 920 . 74 420 .50