College and Research Libraries GLORIA S. CLINE College & Research Libraries: Its First Forty Years College & Research Libraries began publication in December 1939. This study examines the changes that occurred in its publication and citation pat- terns during the forty years from 1939 through 1979. Data are generally de- scribed in terms of eight five-year periods, and the findings of this study are compared with the results of similar studies of various subject literatures. An overall trend toward greater adherence to the norms of scholarly publication in other disciplines was observed. CoLLEGE & RESEARCH LIBRARIES (C&RL) (7) seek to stimulate research and experimentation for the improvement of the service and publish the results ; and marked its fortieth year of continuous publi- cation in December 1979. Widely recognized as a leader in the field, C&RL has ranked for (8) many years among the top ten library peri- odicals in circulation. Its success is primarily help to develop the A.C .R.L. into a strong and mature professional organization . 1 Certainly most, if not all, of these purposes have been adequately served b.y C&RL. Katz attested to its high quality by stating, ''In many ways the best of the American Library Association publications, this is profession- ally edited and contains articles and features not only of interest to college and university libraries, but to anyone dealing with the problem of bibliography, cataloging, acqui- sitions, and the whole range of professional librarianship. "2 due to the fact that throughout the years it has not strayed from its originally stated pur- poses, which were to: (1) serve as the official medium of communication between the association and its subsections and their members; (2) make available selected articles presented at conventions at which college and research li- brarians gather , and publish other profession- ally significant articles; (3) serve as a clearing house for educational, re- search , and library news of interest to college, university, and reference librarians; (4) seek to bridge the gap between these librarians and the faculties, college administrators, and research workers whom they serve; (5) integrate efforts of college, university, and ref- erence librarians with those of kindred groups such as educational and research agencies and learned societies; (6) review and abstract such books , pamphlets , and current periodical literature as would be of interest to the personnel of the A.C.R.L.; Gloria S. Cline is assistant director for public services, University of Southwestern Louisiana , Lafayette . 208 I During its first forty years, seven men served as editors of C&RL: A. F. Kuhlman, 1939-41 Carl M. White, 1941-48 Maurice F. Tauber, 1948-62 Richard B. Harwell, 1962-63 David Kaser, 1963-69 Richard M. Dougherty, 1969-7 4 Richard D. Johnson, 1974-80 Changes, of course, occurred in its pl}blica- tion practices. For example, during the first six years of publication, C&RL dated its vol- umes with combined years, e.g., volume one was dated 1939-40. Then issued as a quar- terly, each volume contained the December, March, June, and September issues. In 1945, however, no Decem her issue was published, and from volume seven, 1946, to date, the volumes have adhered to a calendar year. C&RL continued as a quarterly publication until 1956, at which time it changed to bi- monthly. In March 1966 it gave birth to ACRL News, later renamed College & Re- search Libraries News. The News was to publish". . . News from the Field, Personnel profiles and notes, classified advertising, and other matters of a timely nature ... ,"3 thereby providing rapid news dissemination to the academic library profession and free- ing C&RL to publish scholarly papers. Because C&RL has gained a national rep- utation as a leading library periodical, it is a likely target for retrospective analysis. One wonders what it has accomplished in its long history. What topics did it cover? Whom did it publish? How did it change? The purpose of this study is to answer these and other questions by: (1) describing the literature both published and cited in C&RL, 1939 through 1979; (2) identifying interesting and significant changes or trends in publication patterns; (3) comparing the findings of this study with the results of similar studies of various subject disciplines, especially land- mark studies of scientific literature; and (4) reaching conclusions concerning the scholar- liness of C&RL from the ascertained trends and comparisons. DATA AND METHODOLOGY Data of two types were collected from vol- umes one through forty of C&RL: those con- cerned with the source documents and those concerned with the cited documents. The study was limited in a number of ways, in- cluding: 1. No issues of C&RL News were ana- lyzed; 2. Editorials, news items, programs for meetings or conferences, and the like were excluded; 3. Review articles were excluded; 4. Only bona fide articles, recognizable because of known authorship, were included in the study whether or not they listed refer- ences; 5. References added by the editors were excluded; 6. All references listed as "Ibid." or "Op. cit." were included; First Forty Years I 209 7. Where multiple references were listed in a single footnote, all were included; and 8. References given in the text of the arti- cles, but not listed as footnotes, were in- cluded. Altogether, the completed database con- sisted of 1,775 source documents (i.e., the ar- ticles published in C&RL, volumes one through forty) and 11,658 cited documents (i.e., items cited in the articles published in C&RL). Data items for both sources and cited documents included much that was identical: principal author, sex of principal author, coauthor(s), title, and date. For source documents, additional information that was gathered included author's institu- tional affiliation, number of pages, number of references, subject classification, and number of author and journal self-citations. For cited documents, the additional infor- mation included a single letter from the LC classification scheme to indicate the subject, country of publication, language, publisher for monographs or journal title for articles, and form of the document. Because the database is large and extends over a long period of time, this study is di- vided into two parts: analysis of the source documents and analysis of the cited docu- ments. The data are tested in many ways uti- lizing frequency distributions and cross- tabulations, as well as means, percentages, etc., where meaningful. When possible, comparisons of the results are made with lit- erature from other subject areas. PART ONE: SouRCE DocuMENTS There were 1, 775 articles published in C&RL from December 1939 through No- vember 1979. Characteristics of these articles are identified in two areas: of the articles themselves and of the authors who published them. To ascertain changes and trends in the literature and to smooth out anomalies from year to year, data are usually presented in eight five-year spans. Number and Length of Articles Three questions concerning trends in the publishing habits of C&RL will probably be of some interest to its readers. First, How many articles did C&RL publish during each five-year period of this study? Second, What 210 I College & Research Libraries· May 1982 was the average number of articles per issue? And third, What was the average number of pages per article? Table 1 summarizes the findings for each of these questions. The number of articles published in any journal is, of course, controlled by the editors and board of the journal itself. The years 1945-49 were the most productive in terms of articles published (297). This could be ac- counted for in some measure by the fact that C&RL published its third issue in two parts in 1944-45 (a combined year), 194 7, and 1949. Each extra issue was dedicated to a sin- gle theme: communication and cooperation, essays in honor of Charles Harvey Brown, and rare books in the university library. Even if the articles appearing in these special issues (37) were subtracted from the overall total for 1945-49, the results would remain essentially the same, with that time span pro- ducing both the greatest number of articles and, the highest average number of articles per issue. (C&RL was, at this time, a quar- terly publication.) Twenty years later, in 1965-69, the next greatest number of articles appeared in C&RL, 263 (the journal was then bimonthly), which surprisingly was fol- lowed in 1970-74 by a record low of 163. Whether the increase to 194 in 1975-79 indi- cates a restabilization of production remains to be seen. An obvious trend toward fewer but longer articles per issue can be identified in the ta- ble. The trend is most apparent in the aver- age number of pages per article which in- creased steadily from 4. 79 in 1945-49 to 8.08 in 1975-79. In 1963, Garfield and Sher pub- lished the results of a study of scientific litera- TABLE 1 PUBLICATION CHANGES IN C&RL No . of Avg. No . of Time Articles Articles/ Period Published Issue 1939-44 207 10.35 1945-49 297* 14.85 1950-54 230t 11.50 1955-59 206t 7.22 1960-64 215 7.17 1965-69 263 8.77 1970-74 163 5.43 1975-79 194 6.47 •Three issues appeared in two parts. lOne issue appeared in two parts . Avg. No. of Pages/ Article 5.76 4.79 4.93 5.19 5.37 6.30 7.66 8.08 lNumber of issues/volume increased from four to six in 1956. ture in which they found that 17.3 articles per issue was the norm, while the average number of pages per article was 5.4. 4 Al- though C&RL did not publish as many arti- cles per issue as scientific journals, its average number of pages per article, in the 1960-64 time span, was identical to that given for sci- entific journals, 5.4. Unreferenced Articles and Average Number of References Per Article Of greater importance as measures of the scholarliness of a journal are two characteris- tics for which standards have been estab- lished for scientific literature: (1) the inci- dence of unreferenced articles and (2) the average number of references per article. C&RL experienced increased adherence to these standards in both areas for 1939 through 1979. Table 2 presents supporting data that contrast the steady decrease in un- referenced articles with the steady increase in the average number of references per article. The percentage of articles in C&RL hav- ing no references whatever was excessively high (more than 40 percent) during the early years of the study and remained well above the average for scientific literature (10 per- cent)5 throughout the first thirty years. It is encouraging to note, however, that the per- centage of unreferenced articles steadily de- creased after 1945-49, and in the last decade of the study either approached or fell below 10 percent, the standard for scientific litera- ture. As the percentage of unreferenced arti- cles decreased, the average number of refer- ences per article increased correspondingly. In 1970 Price found the norm for the average number of references per source article (de- fined as AR) for scientific literature to be in the range of 10 to 22. 6 Ten years later, in TABLE 2 REFERENCING CHARACTERISTICS OF C&RL o/o of ~;f~r~~~~ Time Unreferenced Period Articles Article 1939-44 45 2.89 1945-49 47 3.23 1950-54 41 3.61 1955-59 39 4.09 1960-64 33 5.88 1965-69 25 9.16 1970-74 13 10.56 1975-79 9 15.46 1980, AR for all literature indexed in the Sci- ence Citation Index was found to be 15.9. 7 However, when AR was calculated for arti- cles only, excluding meetings, notes, editori- als, etc., it was found to be 24.28 For the en- tire forty-year span of this study, AR was computed to be 6.57 (i.e., 11,658 cited docu- ments divided by 1,775 source documents), a figure well below even the low norm of 10 for scientific literature. It was also well below Barnard's finding of 16 for library litera- ture.9 Barnard's study, however, analyzed the citations in seven library periodicals, not one; covered two years, not forty; and omit- ted all articles having no references, unlike the current study which included them. The reader should note that AR increased steadily through the years, and in the last fifteen years of the study, either approached or fell into the norm of 10 to 22 for scientific litera- ture. Journal Self-Citations In 1979 Garfield reported that self- citations were contained in about 20 percent of a journal's references. 10 In the same paper, he explained the difference between two kinds of journal self-citation rates: There are two self-citation rates, the self-citing and self-cited rates. The self-citing rate relates a jour- nal's self-citations to its total references. The self- cited rate relates a journal's self-citations to the number of times it is cited by all journals including itself. For example, journal X made reference to 10000 items, including 2000 items it itself had pub- lished. Its self-citing rate is 2/10 or 20 % . On the other hand, journal X was cited 15000 times in the references of all journals, including its own. Its self-cited rate is 2/15 or 13.5 % . In this study it was possible to examine two phenomena of journal self-citations in C&RL: (1) the self-citing rate (the data for this study did not provide a means for ana- lyzing the self-cited rate), and (2) the per- centage of source documents containing journal self-citations. Table 3 presents the changes that occurred in both over the years. As an example of how the percentages in table 3 were calculated, in 1975-79, C&RL made reference to 2,999 items, including 335 items it itself had published; thus its self- citing rate was 11.17 percent. In the same time span ,. 194 source documents appeared in C&RL, 101 (52.06 percent) of which in- First Forty Years I 211 TABLE 3 JouRNAL SELF-CITATIONS, 1939-79 o/o of Source Documents Self-Citing Containing /ournal Time Period Rate Sel -Citations 1939-44 7.64 12.56 1945-49 8.95 14.82 1950-54 12.41 23.04 1955-59 10.74 22.82 1960-64 6.26 18.14 1965-69 6.54 26.24 1970-74 5.85 34.97 1975-79 11.17 52.06 eluded one or more references to items it had published. Note that the self-citing rate fluc- tuated a great deal and never approached the 20 percent that Garfield reported. Although the percentage of documents containing journal self-citations had increased consist- ently and rapidly from 1939 to 1979, the to- tal number of journal self-citations was mini- mal with respect to the total number of citations (1,001 out of 11,658). Activities Discussed in C&RL In order to determine the activities dis- cussed in C&RL during its first forty years, each article was scanned by the writer who then used a somewhat modified version of a classification scheme developed by Saracevic and Perk to categorize the source documents according to the subjects they covered. 11 The scheme is outlined in appendix A. Table 4 illustrates the distribution of source documents according to the activity discussed in each time span and overall for 1939 through 1979. Organization and ad- ministration was the major topic most often discussed in C&RL throughout the forty years of this study, and comprised, overall, 33.6 percent of all activities. It was followed by general topics (18.7 percent), resources (14.3 percent), public services (13. 7 per- cent), and technical services (12.6 percent). Combined, the activities of automation and information retrieval ( 4 .1. percent), library instruction (2 .1 percent), and photorepro- duction (. 9 percent) were discussed less than 10 percent of the time. . Few trends in the activities discussed could be identified because the changes in relative frequency from one time period to another were inconsistent. Until 1975-79, there ap- 212 I College & Research Libraries· May 1982 ;! C') 00 C'l t- t- 12 ~ ~ o) g:i "' c;o O"HO C'l c;o cx:i-.:t<......i......ici _M __ _ o:> Q")C')C'I-C'I OOM ;! cilr.icx:i~lr.i ......ic-l.....; ~ C'IC'I __ _ ~ .,. t-t-l!)-0") 000 0> cilr.i~o-o ooo § C'IC'),....;-- peared to be a slight trend toward less em- phasis upon general topics. However, with the publication in 1976 of a large number of historical studies, this trend seemed to termi- nate. A very slight increase in emphasis upon public services and a slight decrease of inter- est in automation and library instruction were also noted in the latter years of the study. Six of the eight major activities were di- vided into subtopics. Table 5 presents the data on subtopics for 1939-79. Several areas represented by the subtopics are noteworthy. First, general administra- tion received the most emphasis because it in- cluded topics of continuing interest to librar- ians, i.e., finance, personnel, salaries, etc. Second, special types of materials were the resources most often discussed. These in- cluded government publications, rare books, indexes, and abstracts. Third, not surpris- ingly, library cooperation was the area of public services that received the most atten- tion. And last, as one might expect, catalog- TABLE 5 SUBCLASSIFICATION OF ACTIVITIES DISCUSSED IN C&RL, 1939-79* Subclassification Organization and administration General administration Professional education Architecture and equipment Resources Book Serial publications Special types of materials Subject literatures Audiovisual materials Public services Circulation Reference Library cooperation Use and user studies Reader services Technical services Acquisitions and selection Cataloging and classification General activities Automation and information retrieval Automation of library processes Informatiom retrieval and documentation Photoproduction and microfilming Copyright law Microfilming techniques and equipment No. 447 56 92 30 25 99 59 41 50 46 96 45 7 95 103 25 46 26 7 9 % 31.8 4.0 6.6 2.1 1.8 7.1 4.2 2.9 3.6 3.3 6.8 3.2 0.5 6.8 7.3 1.8 3.3 1.9 0.5 0.6 •The activities classified as " General" and " Library Instruction" had no subactivities . ing and classification received the most em- phasis among the subtopics of technical services. Source Author Productivity A total of 1,240 principal authors contrib- uted 1,775 articles to C&RL, 1939 through 1979. This averages to 1.43 articles per au- thor over the forty-year span of this study. Figure 1 illustrates the wide range in author productivity. In 1977 Watson, reporting on the publication output of librarians at ten large university libraries, found the median productivity to be two publications in five years, one of which was a book review. 12 In his landmark study of 1926, Lotka de- 992 tl) 1-l 150 0 .c: +l ~ ~ Q) () 1-l ~ 0 (/) lH 100 0 1-l Q) ~ ~ z so 0 0 . 2 4 6 First Forty Years I 213 scribed the productivity of scientific au- thors.13 He found: (1) that the proportion of all persons making a single contribution to chemistry and physics journals was about 60 percent, and (2) that the number of persons making n contributions was about 1/n2 of those making only one contribution. In the current study, 80 percent of the principal au- thors made a single contribution to C&RL. It was obvious then that librarians were not as productive as scientific authors, a conclusion that was in agreement with Schorr's findings for library literature. 14 A discussion of Lotka's law and a detailed statistical analysis of the data from C&RL appear in appendix B. 8 10 12 14 16 18 Number of Articles Fig.l Number of Articles Contributed by Source Authors L---.-L--------------------------------------------------------------------------~ 214 I College & Research Libraries· May 1982 Leading Authors A very weak core of productive authors was identified. Only six authors contributed ten or more articles to C&RL during the forty years of this study. They were: Number of Author Articles Downs, Robert B. 24 Metcalf, Keyes D. 17 Muller, Robert Hans 14 Ellsworth, Ralph E. 12 Shaw, Ralph R. 11 Tauber, Maurice F. 10 These six authors, representing 0.48 percent of all source authors, contributed about 5 percent of the articles appearing in C&RL. Only Robert B. Downs contributed articles during each of the eight five-year spans, while Keyes D. Metcalf published in seven of the time periods (in the first six and in the eighth) and Maurice F. Tauber published ar- ticles in each of the first six five-year periods. These authors obviously had unusually long and productive careers. Sex of Authors Sex was recorded for the principal author only, and data were tabulated for all in- stances when sex was known, rather than for each unique author. Sex was known in 1, 768 of 1, 775 cases. For seven cases (0.39 percent) sex was unknown because initials were used for given names and no photograph or bio- graphical information accompanied the source article. The incidence of unknown au- thorship was considered to be negligible, and thus it was felt that omission of the data in the discussion would not distort or bias the results in any way. From 1939 through 1979, principal _ authors were overwhelmingly males (78.85 percent), with females consti- tuting only 21.15 percent of all contributors to C&RL. Surprisingly, this balance re- mained almost constant throughout the years, as illustrated in table 6. Institutional Affiliation The name of the institution with which the principal author was affiliated was recorded in every instance where the information was available. Over the forty-year span of this study, only 66 of 1, 775 cases were unknown. The top ten institutions are listed in table 7. Although it seemed obvious that the ma- jority of contributors to C&RL would be as- sociated with academic libraries, it was nec- essary to categorize institutions according to type to see if this assumption proved to be true. Nine categories were used: academic li- braries, special libraries, library associa- tions, government libraries, public libraries, library schools, other (nonlibrary institu- tions), foreign institutions, and unknown. The frequency of distribution for institu- tional affiliation (by type) of source authors for 1939 through 1979 is given in table 8 in descending order. As anticipated, almost 60 percent of the contributors did come from ac- ademic libraries. Kim and Kim found that between 57 and 61 percent of the contribu- tors to C&RL from 1957 to 1976 were aca- demic librarians. 15 As seen in table 8, there- mainder of the distribution seemed reasonable also, except for the category of "other." However, it was understandable when one realizes that the majority of con- tributors who fell into this category were in- vited conference speakers whose texts were later published in C&RL. Additional con- TABLE 6 SEX OF SouRcE AuTHORS Time Period 1939-44 1945-49 1950-54 1955-59 1960-64 1965-69 1970-74 1975-79 TABLE 7 Male Sex (in Percent) 78 77 78 87 85 77 80 79 INSTITUTIONAL AFFILIATION OF SouRCE AuTHORS No. of Institution Articles University of Illinois ·73 Columbia University 58 Library of Congress 53 Harvard University 47 University of Chicago 37 University of California, Berkeley 34 Stanford University 30 University of Wisconsin 23 American Library Association 19 Purdue University 19 No. of Authors 44 40 41 26 27 28 20 13 14 15 TABLES TYPES OF INSTITUTIONS OF SouRcE AuTHORS, 1939-79 Type of Institution No. Academic libraries 1,042 Other (nonlibraries) 200 Library schools 152 Government libraries 111 Unknown 66 Foreign 63 Speciallibraries 60 Public libraries 56 Library associations 25 Total 1,775 % 58.70 11.27 8.56 6.25 3.72 3.55 3.38 3.16 1.41 100.00 tributors within this category were often ei- ther retired or unemployed librarians. Li- brary school faculty provided a good many of the articles published in C&RL, as did government library employees. Contributors from foreign countries, as well as from spe- cial and public libraries, were few in num- ber, as might be expected. The smallest per- centage of contributors were those affiliated with .library associations. Collaborative Authorship Collaborative authorship was recorded for each of the 1, 775 source documents pub- lished from 1939 through 1979. From the fig- ures presented in table 9, one can see that'the vast majority of the articles had no coau- thors. However, a trend toward increased collaborative authorship was easily identi- fied by tracing the decrease in the percentage of articles having no coauthors through each of the eight five-year periods as seen in table 10. For twenty-five years there was no break in the extent of collaborative authorship. In 1965-69, however, the rate of articles having no coauthors dropped over 8 percent, and was followed by two additional consecutive TABLE9 NuMBER oF CoAUTHORs CoNTRIBUTING TO ARTICLES IN C&RL, 1939-79 No. of No. of Coauthors Articles % None 1,586 89.35 1 165 9.29 2 17 .96 3 4 .23 4 or more 3 .17 Total 1,775 100.00 First Forty Years I 215 decreases of over 6 percent. The trend to- ward increased collaborative authorship seemed to parallel a similar increase in the sciences. In 1963, for example Garfield and Sher reported an average of 2.1 authors per source document, 16 a figure that had in- creased to 2.56 by 1980. 17 For C&RL, the av- erage number of source authors per article had increased from 1.04 in 1939-44 to 1.36 in 1975-79, still far below the average for sci- entific literature. Author Self-Citations The practice of author self-citation has re- ceived little attention in the literature. Gar- field and Sher reported, "In this index [Sci- ence Citation Index], 8% of all citations are first-author self-citations."18 Table 11 pre- sents the data for two facets of author self- citations in C&RL. The findings for author self-citations par- alleled those discussed earlier for journal self- citations. The author self-citing rate fluctu- ated a good deal and never reached the 8 percent Garfield mentioned for scientific lit- erature, indicating that the total number of author self-citations (414 out of 11,658 total citations for 1939-79) were minimal. It was TABLE10 ExTENT OF CoLLABORATIVE AuTHORSHIP Time Period 1939-44 1945-49 1950.,...54 1955-59 1960-64 1965-69 1970-74 1975-79 TABLEll Articles Having No Coauthors (in %) 95.65 95.62 93.48 92.72 93.95 85.93 79 .14 72.68 AuTHOR SELF-CITATIONS, 1939-79 Time Period 1939-44 1945-49 1950-54 1955-59 1960- 64 1965-69 1970-74 1975-79 Self-Citing Rate 5.09 4.16 5.45 2.65 2.30 2.55 4.93 3.30 %of Source Documents Containing Author Self-Citations 9.7 7.1 10.0 8.7 10.7 12.9 22.7 29.9 216 I College & Research Libraries • May 1982 interesting to note that although author self- citing rates were small, there was a marked increase in the percentage of contributors to C&RL who cited themselves, from 9. 7 per- cent in 1939-44 to 29.9 percent in 1975-79. PART Two: CITED DocuMENTS Variables investigated in this part of the study include: (1) growth rate of cited docu- ments; (2) authorship of cited documents, in- cluding the identification of leading authors, sex of authors, and the extent of collaborative authorship; (3) bibliographic form, with em- phasis upon the most-cited periodicals and monographs; ( 4) language/ geographic/sub- ject distributions; and (5) time span of cited documents. Documents totaling 11,658 were cited in C&RL during the forty-year span of this study. Some of these documents had authors but no titles (for example, correspondence); conversely, some were anonymous, having titles but no authors. Other variables such as publisher and date of publication were also sometimes missing from the data; this situa- tion usually occurred when they were omit- ted from the author's reference and could not be readily ascertained. Therefore, the num- ber of cases analyzed for each of these varia- bles will vary. Growth Rate of Cited Documents It has been widely accepted for years that world literature has grown at an exponential rate. 19 While Danton20 and others have dis- cussed the proliferation of library literature, they have not shown whether it too has in- creased exponentially. Table 12 provides the frequencies for documents cited in C&RL during 1939 through 1979. Note in the table that the number of cita- tions per five-year period was averaged in or- der to reduce the "noise" that would have re- sulted in wide yearly fluctuations in the citation patterns. By plotting these eight five- year averages on linear graph paper, a curve was obtained that illustrates in part an expo- nential rise in citations. Observe in figure 2 that four of the eight points on the curve lie on what resembles an exponential curve·; four points form peaks above the exponential portion. The initial aberration occurring in the curve during the 1945-49 period resulted · TABLE12 FREQUENCY OF DOCUMENTS CITED IN C&RL Time Yearly Period No . Average 1939-44 589 118 1945-49 961 192 1950-54 862 172 1955-59 829 166 1960-64 1262 252 1965-69 2430 486 1970-74 1726 345 1975-79 2999 600 primarily from C&RL having published one issue in two parts during three of the five years, thereby increasing the total number of citations by 131 (or 26 when averaged). If these were subtracted from the 1945-49 to- tal, the curve would have approached a true exponential in its early part (see the dashed curve). An explanation for the second peak in 1965-69 is more difficult to formulate. If the figures for 1965-69 and 1970-74 were re- versed, there would be little need for an ex- planation. A plausible reason for the unprec- edented growth in citations in 1965-69 (which really began in 1960-64) is that the 1960s were years of great expansion in li- braries. Funds were available from the fed- eral government to support research and experimentation in such ventures as automa- tion, as well as collection development and buildings. 21 As a consequence, the literature undoubtedly expanded to report the results of research in the field. In the early 1970s, however, fu~ding slowed22 and the number of citations decreased markedly as if to re- flect a slowdown in library research activity. In the mid-1970s, the citations seemed tore- sume their earlier rate of growth. Whether or not the overall rate for 1939 to 1979 was truly exponential, it did indicate that the archive of available literature in library science had expanded enormously during those forty years. Authorship of Cited Documents A total of 4,072 individual authors was cited from 1939 to 1979. Of the 11,658 cited documents, 1,158 (9.93 percent) were anon- ymous, a rate considerably lower than Brace found in his study of the citations in .library science dissertations (25 percent). 23 Over 60 "' -~ .., Q) Po. .., 700 600 500 ~ 400 >< I ~ "' c:: 0 "j "' ~ 300 u .... 0 .., Q) .0 6 :.E 200 Q) 00 "' .., Q) ~ 100 First Forty Years I 217 0 ~--._--~------~--_.--~----~--._--- 1940 45 so 55 60 65 70 75 80 Time (Y e ars) Fig. 2 Average Number of Citations per Five-Year Period in C&RL, 1939- 79 percent (2,531) of the cited authors were cited only once, and almost 98 percent (3,985) were cited fewer than fifteen times each during 1939 through 1979. Figure 3 il- lustrates the frequency of authors cited one to fourteen times. Table 13 provides the actual data for figure 3. Leading Authors The writer arbitrarily decided to include among the core of leading authors only those who had been cited, on the average, once a year for the entire forty·-year span of this study. Thus , to be included, an author had to be cited forty or more times. Only 17 (0.42 percent) of the 4,072 authors cited in C&RL qualified as leading authors. Table 14 lists the leaders in rank order. Three of these 17 authors were among the 6 leading contribu- tors to C&RL: Robert B. Downs, first with twenty-four contributions; Keyes D. Met- calf, second with seventeen contributions; and Ralph E. Ellsworth, fourth with twelve contributions. Sex of Cited Authors Sex was recorded for principal author only; data were not gathered for coauthors. Four categories of sex were established: (1) male, (2) female, (3) corporate, and (4) un- known (for personal authors whose sex could not be determined because initials were used for given names in the citations). Sex was tabulated in all instances when it was known 218 I College & Research Libraries • May 1982 500 400 300 200 100 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 10 12 14 Number of Times Cited Fig.3 Frequency of Author Citations (Less than Fifteen Times per Author), 1939-79 (10,500 cases) rather than for each unique author. Overall, males were cited more than six times as often as females, i.e., 72.75 per- cent to 11.29 percent. Corporate authors constituted 9.39 percent of the total, while 6.57 percent were personal authors whose sex was unknown. Table 15 analyzes the data on sex by five-year periods. Few strong trends or changes in the sex of cited authors were read- ily apparent. There was an increase of almost 7 percent in females cited during the years from 1955-59 (6.31 percent) to 1975-79 (13.03 percent). During the same twenty- five-year span, there was a decrease in the ci- tations of corporate authors from 15.03 per- cent in 1955-59 to 6.50 percent in 1975-79. The use of initials-only for personal authors hovered around the 4 to 6 percent mark in six of the eight time periods, with a high of over 11 percent reached in 1975-79. Extent of Collaborative Authorship The majority of authors (88.06 percent) cited in C&RL over the forty-year span of this study did not collaborate with others in writing their papers. It was often difficult to determine the exact number of coauthors, because the citing authors frequently used et al. to indicate two or more coauthors rather than their listing each by name. However, the writer recorded individual names for up to three coauthors and then used et al. to in- dicate four or more coauthors. Because of TABLE 13 NuMBER oF CITED AuTHORS CROSS-TABU LA TED BY THE NUMBER OF TIMES EACH wAS CITED No . of Times Cited 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 24 25 26 27 28 29 30 34 36 39 41 43 44 46 49 51 52 54 55 62 66 73 82 111 No . of Authors 2,531 637 315 157 96 60 40 31 27 24 25 16 17 9 10 7 9 4 6 4 5 3 3 2 4 3 1 3 1 3 1 1 2 1 1 1 1 1 2 2 1 1 1 1 1 1 these variations in citing patterns, the data may be somewhat inaccurate. However, it was thought that these inaccuracies would not greatly affect the results. Table 16 pro- vides an overall picture of the changes in col- laborative authorship during 1939 to 1979. There was a definite trend toward greater collaboration among cited authors, increas- ing from a low of 4.3 percent in 1955-59 to 15.7 percent in 1975-79. This trend paral- First Forty Years I 219 Name TABLE14 LEADING AUTHORS CITED IN C&RL, 1939-79 Downs, Robert B. American Library Association Wilson, Louis R. U.S. Office of Education Metcalf, Keyes D. Branscomb, Harvie Lyle, Guy R. - Rider, Fremont Ellsworth, Ralph E. Library of Congress Randall, William M. McAnally, Arthur Knapp, Patricia B. Fussier, Herman H. U.S. Congress Danton, J. Periam Williams, Edwin E. No. of Times Cited 111 82 73 66 62 55 54 54 52 52 51 49 46 44 43 41 41 leled a similar increase noted in table 10 for collaboration among source authors. Bibliographic Form of Cited Documents At the beginning of the data-gathering phase of this study, eighteen categories of bibliographic form were established for the cited documents. It turned out that only one form, patents, was not used at all. Periodi- cals (44.65 percent) and monographs (36.41 percent) together provided the bulk of all cited documents. Table 17 presents an over- view of the forms of documents cited from 1939 through 1979. Table 18 presents the same data in five-year spans and shows that there were only small fluctuations in the cit- ing patterns. A slight trend toward increased use of periodical literature and decreased use of monographs was evident. Table 19 provides a comparison of the bib- liographic forms cited in various subject fields. The difference in citation patterns be- tween sciences such as chemistry and physics and disciplines such as speech and library sci- ence was striking. The sciences cited serial literature more than 85 percent of the time, while the social sciences divided their cita- tions almost equally between serials and monographs. Bibliographic Form: Periodicals Leading Titles. Periodicals comprised the bibliographic form cited most frequently in C&RL. From 1939 through 1979, 642 differ- 220 I College & Research Libraries • May 1982 TABLE15 SEX oF CITED AuTHORS (IN PERCENT) Time Period Male Female Corporate Unknown 1939-79* 72.75 11.29 9.39 6.57 1939--44 77.44 13.78 8.75 5.03 1945--49 68.73 13.95 13.39 3.93 1950-54 72.60 10.65 12.60 4.15 1955-59 76.38 6.31 15.03 2.28 1960-64 74.54 8.55 12.87 4.04 1965-69 75.32 10.33 8.26 6.09 1970-74 75.28 11.68 7.44 5.60 1975-79 67.06 13.03 6.50 11.41 •overall percentage totals for 1939-79. ent periodicals were cited. Table 20 lists the core of ten (1.6 percent) that provided almost 55 percent of the total number of periodical citations. C&RL led with almost 20 percent of the periodical citations. It was followed by Library Journal, Library Quarterly, Ameri- can Libraries, and other leaders in the field. Only one of the top ten periodicals was not a library periodical in the strictest sense, American Archivist, although its relevance to libraries was readily apparent. Periodical Title Dispersion. Stevens de- fined title dispersion as ". . . the degree to which the useful literature of a given subject area is scattered through a number of differ- ent books and journals. If there is much scat- tering, the title dispersion is high; if a large portion of the literature is contained in a few journals, the title dispersion is low."24 Table 21 contrasts the title dispersion of the litera- ture of several subject disciplines. In the two studies of library science litera- ture listed in table 21, only two titles were re- quired to produce 25 percent of the refer- ences. At that point, it was not possible to differentiate the literature of library science from the literature of the sciences, chemistry and physics. Nor was the distinction readily apparent at the next level, where seven jour- nals included 50 percent of the cited litera- ture. The higher title dispersion for library science became obvious at the level where 75 percent of the literature was contained in a much greater number of journals than was true for chemistry or physics. While the title dispersion of library science literature was not as low as it was for the sciences, it was not as high as that of United States history. It seemed instead more comparable to the title dispersion of biochemistry. Bibliographic Form: Monographs Monographs comprised over 36 percent of the total documents cited. This figure was comparable to the 36.9 percent that Brace found in his study of library and information science dissertations. Eight-hundred eight (808) publishers supplied the monographs cited. These publishers were widely scattered and, altogether, the ten leaders listed in table 22 provided only about 34 percent of mono- graphic citations. Most of the leaders were predictable: the American Library Associa- tion, the official voice for the library profes- sion; the University of Chicago Press and Co- lumbia University Press, leaders by virtue of the fact that they nurture two of the best li- brary schools in the country; and the Associa- tion of College and Research Libraries, an important division of the American Library Association. If the same criteria were applied to the cited monographs as were applied to the cited authors in order to determine a core of leading titles (i.e., each must be cited forty or more times) only two titles would be in- cluded. The first, Teaching with Books: A Study of College Libraries, by Harvie Brans- comb, received fifty-two citations. It was cited most heavily (twenty-one times) from 1940 through 1945. However, its importance to librarianship was demonstrated by the fact that it continued to be cited through 1975-79. The second document cited more than forty times was Administration of the College Library by Guy R. Lyle, three edi- tions of which were cited forty-three times in C&RL from 1947 through 1979. Although they were not cited forty or more times, two additional titles were cited heavily enough to Q., x "' CJ:: 0 ::r: ~ f-o ;J < w j;oJ ...:l ;> ~ != E-< ~ 0 IXl < ...< ...< 0 u 1: ~ ~ OOtr.lC'IM r--:o~o~ oo- ~ ~~~~~ o MC'I-OC'I r- 00- "' oo---oo "' I ooo:i~oo II) "' 00 ~ o:>OOMC'It- 0~~00 "CI _gg ~ ~ s E=: '0 ci z o:> "' t-tn-.:t<--.:t' II) I 1Cc<:iooo II) II) o:> i1 ~<;Ot-OC'I 0 c:-i1Coo~ II) o:> 1 ~~~~~ II) ot--oo ..,. o:> ~OOo:>C'Itr.l 0~00~ o:> First Forty Years I 221 TABLE17 BIBLIOGRAPHIC FoRM OF DocuMENTS CITED IN C&RL, 1939-79 Form Periodicals Monographs U.S. govt. publications Proceedings Theses and dissertations Correstondence Annua continuations Unpublished materials Annual reports Reports State govt. publications Newspapers Laws and statutes Audiovisual materials Archival materials Legal cases Miscellaneous Unknown Total No . Times Cited 5,205 4,245 464 299 226 200 194 153 137 117 64 60 32 13 12 4 227 6 11,658 % 44.65 36.41 3.98 2.57 1.94 1.72 '1.66 1.31 1.18 1.00 0.55 0.52 0.28 0.11 0.10 0.03 1.95 0.05 100.01 deserve recognition: The Scholar and the Fu- ture of the Research Library: A Problem and the Solution by Fremont Rider, and The Uni- versity Library: Its Organization, Adminis- tration and Functions by Louis Round Wilson. The Scholar was cited thirty-six times from 1944 through 1979, and two edi- tions of The University Library were cited thirty-two times between 1946 and 1979 . Language/Geographic/Subject Distributions Stevens defined language distribution as "the extent to which the research worker will find useful articles in languages other than his own. "25 In this study, documents. in six- teen different languages were cited from 1939 through 1979. English language docu- ments were cited overwhelmingly (97 .4 per- cent). This strong preference for English lan- guage materials contrasted sharply with that of the sciences as illustrated in table 23. It should be noted, however , that some of these studies of scientific literature were old and had not been updated .· The 1981 Science Ci- tation Index Guide did not indicate the per- centage of cited documents that was in for- eign languages. However, there was some reason to believe that English had become more important to the sciences than ap- peared to be the case in the data presented here. Garfield found , for example, that al- 222 I College & Research Libraries· May 1982 TABLE 18 BIBLIOGRAPHIC FoRM OF DocuMENTS CITED IN C&RL BY FIVE YEAR PERIOD, 1939-79 Form 3 Time Period • (in Percent) 4 5 6 8 Periodicals 40.0 45.1 45.8 46.4 39.4 41.1 46.2 48.8 Monographs 42.9 32.9 31.3 38.7 35.7 39.4 36.4 35.0 Proceedings 1.2 3.3 4.5 1.6 1.8 3.3 1.3 2.8 Reports 0.0 0.2 0.6 0.2 1.0 1.1 2.1 1.1 Annual reports 1.0 1.0 1.3 1.4 2.1 1.2 0.3 1.2 U.S. govt. publications 3.4 6.1 2.9 3.4 4.6 3.7 2.8 4.6 State govt. pubs. 2.2 0.3 0.5 0.2 0.2 0.3 1.1 0.4 Annual continuations 4.4 1.8 1.3 1.6 1.3 1.3 1.7 1.6 Theses and dissertations 1.7 2.1 1.7 0.7 1.7 1.6 2.3 2.5 Correspondence 0.7 3.1 4.2 2.4 4.3 0.8 1.4 0.4 Archival materials 0.0 0.1 0.0 0.2 0.0 0.1 0.3 0.1 Laws and statutes 0.5 2.0 0.1 0.0 0.1 0.1 0.2 0.1 Audiovisual materials 0.0 0.0 0.0 0.4 0.0 0.0 0.4 0.1 Legal cases 0.0 0.0 0.0 0.2 0.1 0.0 0.0 0.0 Newsbapers 0.0 0.2 0.5 0.7 0.8 0.6 0.9 0.3 Unpu lished materials 1.0 0.9 4.3 0.7 2.5 1.1 1.5 0.4 Miscellaneous 1.0 0.8 0.9 0.9 4.2 4.5 0.9 0.6 Unknown 0.0 0.0 0.1 0.1 0.1 0.0 0.0 0.0 •1 = 1939-44 ; 2 = 1945-49; 3 = 1950-54; 4 = 1955-59; 5 = 1960-64 ; 6 = 1965-69; 7 = 1970-74; 8 = 1975-79 . TABLE19 PERCENT oF SERIALS AND MoNOGRAPHS CITED IN VARIOUS SUBJECT FIELDS* Author and Subject Fussier- Chemistry, 1948 Fussier- Physics, 1948 Garfield- Science, 1980 Broadus-Speech, 1953 McAnally- U.S. History, 1951 Popovich- Business/Mgt . , 1978 Stewart- Politics, 1970 Barnard- Library Science, 1957 Brace- Library Science, 1975 Cline- Library Science, 1980 *Table adapted from Barnard, p . 16. Hncludes journal items only. Serials 92.7 91.8 86.5t 45.7 9.2 58.6 23.0 51.8 33.0 50.1t Bibliographic Form (in %) Monographs 5.2 7.7 47.9 45 .6 31.9 66.0 37.7 38.9 36.4 Others 2.1 0.5 6.4 45.2 9.5 11.0 10.5 28.1 13.5 !Includes periodicals, proceedings, annual continuations, and annual reports. though 56 of the 100 most cited papers in the physical sciences had been published outside the United .States, all100 were published in English. 26 Geographic Distribution. Four categories of geographic distribution were established: (1) United States; (2) English-speaking for- eign countries, including the British Isles, Canada, Australia, New Zealand, and other British colonies; (3) .other foreign countries; and (4) unknown. Table 24 presents the geo- graphic distribution of documents cited throughout the years 1939 through 1979. Note that over 90 percent were published in the United States. Subject Distribution. Subject distribution was defined as the extent to which writers in a particular discipline draw on sources out- side that discipline. Low subject dispersion indicated that the cited literature belonged to a well-defined field that drew little from outside sources. High subject dispersion ex- isted when the literature was largely com- posed of materials from outside the subject area to which the literature rightly belonged. In order to determine the subject dispersion of cited documents, they were assigned a sin- gle letter of the Library of Congress classifi- cation scheme. Table 25 presents the findings for 1939 through 1979. Almost 65 percent of all cited documents fell into the Z class, the special subject of library science. First Forty Years I 223 TABLE20 TEN PERioDICALS CITED MosT FREQUENTLY IN C&RL, 1939-79 No . of Periodical Times Cited College & Research Libraries 1,001 Library Journal 550 Library Quarterly 379 American Lib: :·;ries (incl. ALA Bulletin) 205 Library Trends 183 Library Resources and Technical Services 117 ASIS Journal (incl. American Documentation) 114 Special Libraries 110 American Archivist 108 Wilson Library Bulletin 81 Others (632 periodicals) 2,357 TABLE21 Percent 19.23 10.57 7.28 3.94 3.52 2.25 2.19 2.11 2.08 1.56 45.28 NUMBER OF PERIODICALS CONTAINING THE RESEARCH LITERATURE OF DIFFERENT SUBJECT FIELDS Cum. Percent 19.23 29.80 37.08 41.02 44.54 46.79 48.98 51.09 53.17 54.73 100 .01 No. of No . of Periodicals Containing Various Percentages of References Author and Study References 25 % 50 % 75 % 100 % Gross and Gross- Chemistry, 1927 3,633 2 7 24 247 Fussier-Chemistry, 1948 1,085 1 5 19 131 Fussier- Physics, 1948 1,279 1 3 17 134 Henkle- Biochemistry, 1938 17,198 3 12 56 851 McAnally- U.S. History, 1951 452 14 54 149 259 Barnard- Library Science, 1957 863 2 7 36 183 Cline- Library Science, 1980 5,205 2 8 52 642 TABLE22 LEADING PuBLISHERS OF MoNOGRAPHS, 1939-79 No . of Cumulative Publisher Citations Percent Percent American Library Association 428 10.08 10.08 University of Chicago Press 182 4.29 14 .37 Columbia University Press 163 3.84 18.21 McGraw-Hill 112 2.64 20.85 Wiley 104 2.45 23.30 Wilson 102 2.40 25.70 Scarecrow 100 2 .36 28.06 American Council on Education 81 1.91 29.97 Harper 80 1.89 31.86 Association of College & Research Libraries 75 1.77 33 .63 TABLE23 LANGUAGE DISTRIBUTION OF SUBJECT LITERATURES* Total Author and Subject English German French Other Non-English Fussier-Chemistry, 1948 64.5 25.0 3.0 7.5 35.5 Wood-Chemistry, 1967 50.3 6.4 7.3 36.0 49.7 Fussier-Physics, 1948 66.6 22.1 2.9 8.4 33.4 Wood-Physics, 1967 73.0 4.0 4.0 19.0 27.0 Wood-Biology, 1967 75.0 3.0 3.0 19.0 25.0 Stewart- Politics, 1970 82.1 17 .9 McAnally- U.S. History, 1951 96.4 3 .6 Barnard- Library Science, 1957 94 .5 1.2 1.9 2.4 5 .5 Cline- Library Science , 1980 97.4 1.0 0.5 1.1 2.6 •Table was adapted from Stevens, p.l7. 224 I College & Research Libraries • May 1982 TABLE24 GEOGRAPHIC DISTRIBUTION OF CITED DocuMENTS, 1939-79 Country Number United States 10,573 English-speaking foreign 664 Other foreign 415 Unknown 6 Total 11,658 Percent 90.69 5.70 3.56 0.05 99.90 Table 26 analyzes the subject dispersion by five-year period. The highest dispersion oc- curred in the fifteen years between 1960 and 1975 when the percentages of documents classified as Z were the lowest at 53, 54, and 62 percent. In all other time periods, the per- centage of documents in the Z's hovered at or above 70 percent. Few trends were discernible in the subject distribution of cited documents. In 1960-64, when subject dispersion was at its greatest, education documents (L's) were cited more frequently than any classification other than Z. From that time period to 1975-79, how- ever, the use of education materials de- creased 8.4 percent and the use of social sci- ence items (H's) increased 7.0 percent. Only the H's and L's provided a substantial per- centage of documents outside the Z's, the spe- cial subject of library science. Subject dispersion of the cited literature is compared with that of other disciplines in Table 27. This study adopted Barnard's defi- nition of subjects closely related to library science and considers them to be: history (D, E, F), sociology (H), education (L), litera- ture (P), and technology (T) . The two studies of library science literature showed a surpris- ingly low subject dispersion, Barnard's being lower than that which Fussier found for chemistry, while the subject pispersion of this study was comparable to that of physics. Time Span of Cited Documents Studies of the time span between the publi- cation of a document and its use (i.e., cita- tion) have produced some interesting obser- vations. For example, upon finding that chemists used chemical literature for a longer time than physicists used chemical literature, Fussier said, "It is reasonable to .suppose that this is a natural phenomenon in which spe- cialists in a field use the literature of the field over a longer time span than do non- TABLE25 SUBJECT DISPERSION OF CITED DocuMENTs, 1939-79 LC Classification Number A 197 B 162 c 142 D 49 E 52 F 16 G 16 H 962 J 176 K 68 L 1,113 M 104 N 34 p 106 Q 358 R 33 s 4 T 102 u 11 v 1 z 7,534 None 418 Total 11,658 Percent 1.69 1.39 1.22 0.42 0.45 0.14 0.14 8.25 1.51 0.58 9.55 0.89 0.29 0.91 3.07 0.28 0.03 0.88 0.09 0.01 64.63 3.59 100.01 specialists. . It may well be that the liter- ature of 'outside' fields is used only so long as it is new and fresh. "27 In 1970, Price examined data from the Sci- ence Citation Index and found that for each year from 1964 through 1968 more than 50 percent of the citations were to the last five years. 28 The data for 1965 provided a high markof55.3percent. Forthatsameyear, the percentage of citations dated in the previous five years for selected social science journals was lower: American Sociological Review , 35.2 percent; Psychological Bulletin, 37.8 percent; and Journal of Abnormal and Social Psychology, 39.2 percent. American Docu- mentation (now Journal of the American So- ciety for Information Science) was the only social science journal that had a higher per- centage (at 59.8 percent) than that of the Sci- ence Citation Index journals (55.3 percent). Table 28 presents the time span of various subject literatures from studies completed over a thirty-year span. The highest percent- ages of citations dated in the previ<;)us five years were: 69.4 percent for physics; 56.4 percent and 52.0 percent for library science; and 51.3 percent for chemistry. There is rea- son to believe that Fussier's data may no longer hold true. In 1980, Garfield reported that: . . . the references [in biochemistry articles] are to a higher proportion of older material than was the case previously .... In 1969, 54.5 percent of [the references in Acta Biochim. Biophys.] were more than five years old. This steadily increased to 71.7 % in 1977. . . . A similar increase has also taken place in mathematics and botany. 29 It is difficult to explain the high percent- age of citations to the last five years in docu- ments cited in C&RL. It cannot be, as Fussier suggested, that new and fresh mate- rial was cited from outside fields. Recall that subject dispersion was low for the documents cited in C&RL throughout this entire study, indicating that librarians had not drawn fre- quently from outside sources. A reasonable explanation was that only current materials First Forty Years I 225 were of importance to librarians because they discussed current topics- e.g., what the current inflation was doing to budgets, what technological changes had occurred, etc. SuMMARY AND CoNCLUSIONS College & Research Libraries has been one of the leading publications in the field of li- brary science since its inception in 1939. Pub- lished by the ACRL, it has served many pur- poses, chief among which was to be the avenue of communication between the asso- ciation and its members. When C&RL News originated in 1966, it assumed that impor- tant responsibility, publishing news items, personnel profiles, and other notes, in the process freeing C&RL to publish more schol- TABLE26 RELATIVE FREQUENCY oF DocuMENTS CITED IN C&RL CROSS-TABU LA TED BY LC CLASSIFICATION Classi- Time Period (in Percent) fication 39-44 45-49 50-54 55-59 60-64 A 0.9 1.1 1.7 2.8 3.9 B 1.0 1.3 0.8 0.6 0.9 c 0.3 0.0 1.3 0.4 0.1 D 0.2 0.0 0.9 0.4 0.4 E 0.9 0.4 0.6 0.2 0.5 F 0.3 0.3 0.5 0.1 0.0 G 0.0 0.5 0.0 0.5 0.1 H 4.1 1.1 3.8 2.9 4.7 J 0.3 1.3 0.5 0.7 6.5 K 0.2 2.0 0.1 0.2 0.8 L 14.3 11.3 10.8 10.3 14.8 M 0.2 0.0 0.1 0.1 0.0 N 0.3 0.0 0.2 0.1 0.7 p 1.2 1.1 2.4 1.7 1.5 Q 1.2 2.5 2.1 1.0 3.5 R 0.0 0 .2 0.0 0.2 1.0 s 0.0 0.1 0.0 0.0 0.1 T 0.3 0.6 0.6 1.3 0.8 u 1.0 0.0 0.1 0.0 0.0 v 0.0 0.0 0.0 0.0 0.0 z 71.6 71.7 68.4 71.0 53.1 None 1.7 4.4 5.0 5.4 6.7 TABLE 27 SuBJECT DISPERSION OF SPECIAL LITERATURES* Author and Subject Fussier- Chemistry, 1948 Fussier- Physics, 1948 Sengupta- Biochemistry, 1973 McAnally- U.S. History, 1951 Voigt- Metallurgical Engineering, 194 7 Voigt- Mechanical Engineering, 194 7 Barnard- Library Science, 1957 Cline- Library Science, 1980 •Table adapted from Stevens, p .l5. Special Subject( %) 71 63 55 31 61 27 78 65 65-69 1.6 3.6 2.7 0.2 0.2 0.2 0.2 7.3 1.1 0.6 9.6 4.0 0.2 1.2 6.2 0.2 0.0 0.9 0.1 0.0 54 .3 5.7 Closely Related Subject( %) 19 25 7 16 60 14 20 70-74 2.0 1.3 0.5 0.6 1.2 0.1 0.1 16.5 1.6 0.3 7.6 0.1 0.7 0.1 2.5 0.3 0.0 1.4 0.1 0.1 61.6 1.4 75-79 0.7 0.4 1.7 0.6 0.2 0.0 0.0 11.7 0.6 0.5 6.4 0.1 0.1 0.2 2.1 0.2 0.0 0.7 0.0 0.0 73.0 1.0 Other( %) 10 12 62 23 13 8 15 226 I College & Research Libraries • May 1982 TABLE 28 TIME SPAN OF vARIOUS SUBJECf LITERATURES* Author and Subject 0-5 Fussier-Physics, 1948 69.4 Fussier- Chemistry, 1948 51.3 Garfield- Science, 1980 45.6 Broadus-Speech, 1953 27.4 McAnally- U.S. History, 1951 10.4 Popovich - Business & Mgt., 1978 41.1 Barnard- Library Science, 1957 52.0 Cline- Library Science, 1980 56.4 •Table adapted from Barnard , p.30 . arly papers. The purpose of this study was threefold: (1) to describe the literature both published and cited in C&RL from 1939 through 1979; (2) to identify significant changes and trends in its publication pat- terns; and (3) to compare the findings of this study with the results of similar studies of various subject disciplines, especially land- mark studies of scientific literature. The un- derlying goal in pursuing the final purpose was to determine how C&RL, as a leading li- brary journal, compared with scholarly jour- nals in other fields. In this study, the characteristics of both source and cited documents were examined in detail , revealing many changes that had occurred from 1939 to 1979. The vast major- ity of changes boded well for the journal, in- dicating higher standards and increased scholarliness. This summary will briefly enu- merate the changes and draw some conclu- sions concerning the present status of C&RL. From its beginning in 1939 until the mid- 1950s, C&RL published many short, newsy articles. After 1954, however, it began to publish fewer but longer articles. The prac- tice of referencing earlier works was sadly neglected by the contributors to C&RL until the 1970s, when the percentage of unref- erenced articles dropped to 13 percent in 1970-74 and 9 percent in 1975-79, figures in line with the average of 10 percent for scien- tific literature. As a natural consequence of increased referencing, the average number of references per article also increased, from 2.89 in 1939-44 to 15.46 in 1975-79. This latter figure compared favorabl y with the norm of 10 to 22 references per article for sci- entific literature. There were many possible reasons for librarians to have neglected the practice of referencing for so long. For exam- 0- 10 Age of Cited Documents (in % ) 0- 15 0- 20 88.2 93 .9 71.3 78.7 70.1 82.5 88.8 43.4 62.3 21.6 28.3 33.8 70.3 84.7 91.3 67.3 74 .5 81.9 73 .7 81.8 86.3 0--50 79.9 62.1 91.6 94.3 0-100 92 .0 89 .7 96.9 98.0 ple, in the earlier years of the study, the body of library literature was not very large and access to it probably was limited for many li- brarians. Price, however, pointed to the most plausible explanation for this phenome- non: Trivially and quite typically, such unreferenced papers occur when an experienced scientist or li- brarian makes an ex cathedra pronouncement out of his innate knowledge of what should be or what is. 30 Both journal and author self-citing rates in C&RL consistently fell below the average percentages for scientific literature (20 per- cent and 8 percent, respectively). This was simply because the total number of journal and author self-citations were minimal with respect to the total number of citations. However, despite the consistently low self- citing rates , the practice of self-citation in- creased steadily and substantially from 1939 to 1979 for both the source journal and source authors . These self-citations indicated the existence of increased numbers of related materials in C&RL as well as the existence of further documents produced by its contribu- tors. Throughout the forty years of this study, one library activity was discussed more fre- quently than any other , organization and ad- ministration (about 34 percent). One reason for the heavy emphasis on this topic was sup- plied by Kim and Kim, who looked at the au- thorship of articles published in C&RL in terms of library position . 31 They found that library administrators contributed a very large percentage of the articles (65. 7 percent in 1957-66 and 47.2 percent in 1967- 76). It was only natural then that administrators should write about administration. Unfortu- nately, few significant trends in the topics discussed were discernible, because the rela- tive frequencies for each generally varied by less than 5 percent from one time span to an- other. The slight decline of interest in auto- mation and information retrieval, a topic of importance to most librarians, could be at- tributed to the publication of new, special- ized periodicals such as the I ournal of Li- brary Automation. Source authors were characterized by sex, institutional affiliation, and extent of collab- orative authorship. Further, an effort was made to identify a core of productive au- thors, and Lotka's law was applied to the em- pirical data to ascertain whether the contrib- utors to C&RL were as productive as scientific authors. An overwhelming major- ity (about 80 percent) of principal authors were males, and that rate remained rela- tively constant over the years. The Olsgaards pointed out that females had failed to publish up to the normal level, which should have been 84 percent for the general library popu- lation or 61.5 percent for academic librari- ans. 32 As expected, the majority of authors (about 60 percent) were librarians affiliated with academic institutions. Collaborative authorship increased from less than 5 percent in 1939-44 to over 20 percent in 1975-79. Price jndicated that collaborative authorship was useful as a means of analyzing invisible colleges and in-groups, but added that col- laboration arises more from economic rather than intellectual dependence. 33 While it was not the purpose of this paper to investigate the existence of invisible colleges or economic dependence among contributors to C&RL, it seemed likely that both were reasons for the observed increase in collaborative author- ship. A very weak core of productive authors was identified, consisting of only six authors who contributed ten or more articles through the forty years of this study: Robert D. Downs, Keyes D. Metcalf, Robert H. Mul- ler, Ralph E. Ellsworth, RalphR. Shaw, and Maurice F. Tauber. When Lotka's law was applied to the source author data, the results proved that, overall, the contributors to C&RL were not as productive as scientific authors. Whereas Lotka reported that 60 percent of scientific authors published only one article in a given period of time, this study found a much higher rate, 80 percent for contributors to C&RL. First Forty Years I 227 The documents cited in C&RL grew at a rate that appeared in part to be exponential from 1939 to 1979. However, a true expo- nential curve was disrupted by an unusually high growth rate during the decade of the 1960s. This undoubtedly resulted from in- creased funding of research during those years. Many of the characteristics of the cited au- thors paralleled those of the source authors. For example, an overwhelming majority of both source (80 percent) and cited (73 per- cent) authors were males. A trend toward in- creased collaboration among source authors was also observed for cited authors, although at a somewhat reduced rate. Just as the source authors were widely scattered (about 80 percent contributed only one article), so were the cited authors, over 60 percent of whom were cited only once throughout the forty years of this study. Thus it was inevita- ble that out of more than 4,000 individual authors, only 17 were cited often enough to be identified as a core. Three of these 17 au- thors were among the 6 leading contributors to C&RL: Robert B. Downs, first among both source and cited authors; Keyes D. Met- calf, second among the source authors, fifth among the cited authors; and Ralph E. Ellsworth, fourth among the source authors and ninth among the cited authors. One of the significant characteristics of any subject literature is the form in which most of its material is published. Periodical literature is unquestionably the most impor- tant bibliographic form for science. Both Price34 and Garfield and Sher35 reported that 80 percent or more of all references in scien- tific papers were to periodical articles. For C&RL, only about 45 percent of the cited documents were periodicals. However, a slight trend toward increased use of periodi- cal literature was observed. The periodicals cited in C&RLwere widely scattered, a char- acteristic typical of the social sciences but in contrast to that of the sciences, where much of the literature on a topic was contained in a few journals. Language and geographic distributions were low for the documents cited in C&RL, with over 97 percent of them appearing in the English language and about 91 percent published in the United States. Several prob- able reasons for these low distributions pre- 228 I College & Research Libraries· May 1982 sented themselves. First, accessibility of the documents was probably limited to the large libraries and library school libraries. Second, indexing of foreign language material in li- brary science was also restricted, with the only widely available index, Library Litera- ture, including only a limited number of non-English language materials. And last, it was probable that very little library research of importance was completed outside the English-speaking countries. Subject distribution is another characteris- tic widely examined for various disciplines. Scientific literature generally has low subject distribution, i.e., the majority of documents fall into the special subject with little drawn from outside sources. The documents cited in C&RL were classified with a single letter of the Library of Congress classification scheme, with Z designated to be the special subject of library science. Over the forty- year span of this study, about 65 percent of all cited documents fell into the Z's. This low distribution , characteristic of scientific liter- ature, seemed to indicate a well-defined field. It should be pointed out, however, that some researchers feel that this great concen- tration of cited documents in the special sub- ject of library science is a weakness rather than a strength. Saracevic and Perk36 felt that the nature of librarianship was too re- strictive, too self-contained, and that inter- action with other disciplines was needed to broaden the subject. The final characteristic examined for cited documents was their time span. For this study, over 56 percent were cited within five years of their publication. This was a larger proportion than was reported in 1980 for lit- erature indexed in the Science Citation In- dex, 45.6 percent. 37 Librarians thus seemed to require current, up-to-date information as much or more so than did scientists. In 1939, when it commenced publication, C&RL filled a definite need for academic and research librarians. It immediately be- came a leader in circulation among library periodicals. However, in scholarliness, it did not initially adhere to the norms observed for other disciplines, especially the sciences. From 1939 through 1979, positive changes occurred in the documents both published and cited in C&RL, pointing to both an awareness of the need for higher standards and a greater adherence to those standards. If the trends ascertained in this study con- tinue, the future of C&RL as a truly schol- arly library journal seems assured. REFERENCES 1. A. F. Kuhlman , "Introducing 'College and Research Libraries,' " College & Research Li- braries 1:9 (Dec . 1939). 2. William Katz, Magazines for Libraries (2d ed.; New York: Bowker, 1972), p.478. 3. [Publication Information] , ACRL News 1:3 (March 1966). 4. Eugene Garfield and I. H. Sher, "New Factors in the Evaluation of Scientific Literature through Citation Indexing," American Docu- mentation 14:199 Quly 1963). 5. Derek J. deSolla Price, "Networks of Scientific Papers, " Science 149:510 Quly 1965). 6. Derek J. deSolla Price, "Citation Measures of Hard Science, Soft Science, Technology, and Nonscience," in Carnot E. Nelson and Donald K. Pollock, eds., Communication among Sci- entists and Engineers (Lexington, Mass.: Heath, 1970), p.7 . 7. "Science Citation Index 1979-1980; Compar- ative Statistical Summary, " in Science Cita- tion Index Guide and Lists of Source Publica- tions 1980 (Philadelphia: Institute for Sci- entific Information, 1981), p.27. 8 . Ibid. 9. Walter Monroe Barnard, "Characteristics of the Literature Used by American Authors of Journal Articles in Library Science," (Master's thesis, Univ. of North Carolina, 1957), p .8. 10. SCI Journal Citation Reports; a Bibliometric Analysis of Science Jo urnals in the lSI Data Base , 1979, V.14 (Philadelphia: Institute for Scientific Information, 1980), p.13A. 11. Tefko Sarcevic and Lawrence L. Perk, "Ascer- taining Activities in a Subject Area through Bibliometric Analysis," Journal of the Ameri- can Society for Information Science 24:130 (March-April1973). 12. Paula de Simone Watson, "Publication Activ- ity among Academic Librarians," College & Research Libraries 38:375-84 (Sept . 1977). 13 . A. J. Lotka, "The Frequency Distribution of Scientific Productivity," Journal of the Wash- ington Academy of Science 16:317-23 (Dec. 1926). 14 . Alan Edward Schorr , "Lotka's Law and Li- brary Science," RQ 14:32-33 (Fall1974). 15. Soon D. Kim and Mary T. Kim, "Academic Library Research: A Twenty Year Perspec- tive," in Robert D. Stueart and Richard D. I Johnson, eds., New Horizons for Academic Li- braries; Papers Presented at the First National Conference of the Association of College and Research Libraries, Boston, Mass., Nov. 8-11, 1978 (New York: K. G. Saur, 1979), p .379. 16. Garfield and Sher, "New Factors," p.199. 17. "Science Citation Index 1979-1980," p.27. 18. Garfield and Sher, "New Factors," p.199. 19. Derek J. deSolla Price, Little Science, Big Sci- ence (New York: Columbia Univ. Pr., 1963). 20. J. Periam Danton, "The Library Press," Li- brary Trends 25:153-76 Uuly 1976). 21. David Kaser, "The Effect of the Revolution of 1969-1970 on University Library Administra- tion," in Herbert Poole, ed., Academic Li- braries by the Year 2000; Essays Honoring jer- rold Orne (New York: Bowker, 1977), p.64-75. 22. Ibid., p.65. 23. William Brace, "A Citation Analysis of Doc- toral Dissertations in Library and Information Science, 1961-1970," (Ph.D. dissertation, Case Western Reserve Univ., 1975), p.29. 24. Rolland E. Stevens, Characteristics of Subject Literatures, ACRL monograph no.6 (Chi- cago: American Library Assn., 1953), p .12. 25. Ibid., p.16. 26. Eugene Garfield, "1978 Articles Most Cited in 1978 and 1979. 1. Physical Sciences," Current First Forty Years I 229 Contents: Physical, Chemical & Earth Sci- ences20:5 (Nov.17, 1980). 27. Herman H. Fussier, "Characteristics of the Research Literature Used by Chemists and Physicists in the United States. Part II," Li- brary Quarterly 19:124 (April1949). 28. Price, "Citation Measures of Hard Science," p.12. 29. Eugene Garfield, "Trends in Biochemical Lit- erature," Current Contents: Physical, Chemi- cal & Earth Sciences 20: 15 (March 17, 1980). 30. Price, "Citation Measures of Hard Science," p.8. 31. Kim and Kim, "Academic Library Research," p.379. 32. John N. Olsgaard and Jane Kinch Olsgaard, "Authorship in Five Library Periodicals," Col- lege & Research Libraries 41:50 U an. 1980). 33. Price, "Citation Measures of Hard Science," p.7. 34. Price, "Networks of Scientific Papers," p.510. 35. Garfield and Sher, "New Factors," p.199. 36. Saracevic and Perk, "Ascertaining Activities," p.133. 37. "Chronological Distribution of Citations to Authored Items (Non-Patents)," in Science Ci- tation Index 1980 Guide and Lists of Source Publications, p.28. BIBLIOGRAPHY oF DocuMENTS REFERENCED IN THE TABLES IN PART Two Barnard, Walter Monroe. "Characteristics of the Literature Used by American Authors of Journal Articles in Library Science." Master's thesis, Univ. of North Carolina, 1957. Brace, William. "A Citation Analysis of Doctoral Dissertations in Library and Information Sci- ence, 1961-1970." Ph.D. dissertation, Case Western Reserve Univ., 1975. Broadus, Robert N. The Research Literature of the Field of Speech. (ACRL monograph no. 7) Chi- cago: American Library Assn., 1953. Cline, G. S. "A Bibliometric Study of Two Selected Journals in Library Science, 1940-1974." D.L.S. dissertation, Univ. of Southern Califor- nia, 1978. Fussier, Herman H. "Characteristics of the Re- search Literature Used by Ch.emists and Physi- cists in the United States: Pt. 1," Library Quar- terly 19:19-35 Uan. 1949). ----·"Characteristics of the Research Liter- ature: Pt. II," Library 'Quarterly 19:119-43 (April1949). Garfield, Eugene. "Science Citation Index 1979-1980; Comparative Statistical Summary," in Science Citation Index Guide and Lists of Source Publications 1980, p.27. Philadelphia: Institute for Scientific Information, 1981. Gross, P. L. K., and Gross, E. M. "College Li- braries and Chemical Education," Science 66:385-89 (Oct. 1927). Henkle, H. N. "The Periodical Literature of Bio- chemistry," Bulletin of the Medical Library As- sociation 27: 139-4 7 ( 1938). McAnally, A. M. "Characteristics of Materials Used in Research in United States History." Ph.D. dissertation, Univ. of Chicago, 1951. Popovich, Charles J. "Characteristics of a Collec- tion for Research in Business/Management," College & Research Libraries 39:110-17 (March 1978). Sengupta, N. "Recent Growth of the Literature of Biochemistry and Changes in Ranking of Period- icals," Journal of Documentation 29:192-211 Uune 1973). Stevens, Rolland E. Char'acteristics of Subject Lit- eratures. (ACRL monograph no.6) Chicago: American Library Assn., 1953. Stewart, June L. "The Literature of Politics: A Ci- tation Analysis," International Library Review 2:329-53 (1970). Voigt, Melvin J. "Scientific Periodicals as a Basic Requirement for Engineering and Agricultural Research," College & Research Libraries 8:354-59,375 Uuly 1947). 230 I College & Research Libraries· May 1982 APPENDIX A SuBJEcr CLASSIFICATION OF AcriVITIES DiscussED IN SouRcE DocuMENTS IN C&RL l. General: Includes background studies, historical studies, biography, philosophy, censorship, ethics, intellectual freedom, etc. 2. Organization and Administration General Administration: Includes finance, personnel, unions, salaries, etc. Professional Education: Includes institutes, in-service education , etc. Architecture and Equipment: Includes buildings, furniture, equipment, supplies, etc. 3. Resources Book: Includes printing, illustration, book trade, etc. Serial Publications: Includes periodicals, journals, newspapers, etc. Special Types of Materials: Includes government publications, rare books, indexes and abstracts, etc. Subfect Literatures: Includes, e.g., music, folklore, black literature, etc. Audio- Visual Materials: Includes recordings, films, television, picture collections, etc . 4. Public Services Circulation: Includes access to shelves, fines, inventories, etc. Reference and Research Services: Includes reference interview, telephone information service, biblio- graphic searching, etc. Library Cooperation: Includes interlibrary loan, union catalogs, networking, etc. Use and User Studies Reader Services: Includes special services for such groups as the handicapped, cultural programs , etc. 5. Technical Services Acquisitions and Selection: Includes ordering, cooperative purchasing, gifts , etc. Cataloging and Classification: Includes cataloging, classification, indexing, etc. General Activities 6. Automation and Information Retrieval Automation of Library Processes: Includes application of the computer, punched cards , etc. Information Retrieval and Documentation: Includes coordinate indexing, selective dissemination of information , etc. 7. Library Instruction: Includes methods of instruction, slide/tape productions, etc. 8. Photoreproduction and Microfilming Copyright Law and Fair Use Doctrine Techniques, Methods , and Equipment APPENDIXB APPLICATION oF LoTK.A' s LAw Lotka's law has been widely recognized as a measure of the productivity of scientific authors. In 1926 Alfred Lotka analyzed the number of publications of chemists listed in Chemical Abstracts, 1907-1916 (actually he considered only 6,891 names beginning with the letters A and B) and those of physicists listed in Auerbach's Geschachtstafeln der Physik. He found: (1) that the number of persons making n contribu- tions was about lln 2 of those making only one contribution, and (2) that the proportion of all persons mak- ing a single contribution was about 60 percent . These findings can be generally stated as: y = c/xn = cx-n where y percent of authors x number of articles c constant -n slope of the log-log plot of the above equation When y, the percent of authors, is plotted on log-log paper versus x, the number of articles, an essentially straight line results. The slope of the line, -n, can be calculated using a least-squares approach. Lotka found that for n = 2.0, the constant, c, equaled .6079 or 60.79 percent; thus, = 60 · 79 % = 60. 19x-2.0% Y x2 .o First Forty Years I 231 Lotka also found that a least-squares analysis of Auerbach's data yielded n = 2.0, which resulted in the percent of authors given by the above equation. However, for the Chemical Abstracts data, the least- squares analysis yielded a fractional exponent, n = 1.888, which was found (from mathematical tables) to correspond to c = 56.69 percent. Thus, in this case, Lotka's law becomes: 56.69 Y = -- % = 56.69x- 1·888 % xt.sss Coile discussed several instances in which Lotka's law had been misinterpreted by the author's having assumed a value of n = 2.0 regardless of the slope of the log-log plot. 1 He stressed that whatever the data-humanistic or scientific-the manner in which they are collected must be consistent with Lotka's data for statistical tests of goodness-of-fit. Coile stated that the data must include senior authors only, thus eliminating all coauthors. He also found that the chi-square test was not an appropriate goodness-of-fit test for this type of data. Instead, he recommended the use of the Kolmogorov-Smirnov test. The first four columns of table 1 represent the empirical data concerning the number of articles per source author. They can be read as follows: 992 source authors (80.02 percent) contributed one article to C&RL during 1939-79; 142 (11.44 percent) contributed two articles; through 1 source author (0.08 per- cent) having contributed twenty-four articles. Column 4 represents the observed cumulative distribution function, S~x), for the percentages of column 3. A curve fitting linear repression analysis2 of the data of columns 1 and 3 yielded n = 2.44 and c = 51.29 percent. See figure 1 for a plot of these data. Using these values in the above equation for Lotka's law, an estimated percent of authors, y (column 5), was calculated for each corresponding value of x (column 1). From these values of i), the calculated cumulative distribution, F 0 (x), was obtained (column 6). Coile showed that the Kolmogorov-Smirnov test for goodness-of-fit was appropriate for this type of data. 3 The maximum deviation (D max) between the theoretical and observed cumulative distribution functions is given by: Dmax = I SN(x) - Fo(x) I max which is obtained from column 7. This value is compared with the value found at the .Ollevel of signifi- cance (a.o, = 1.63/VN) with N = 1,240. Since Dmax = .3147, which exceeds a. 01 = .0466, one must con- clude that the empirical data do not adhere to Lotka's law. TABLE 1 FREQUENCY OF OBSERVED NuMBER OF SouRcE AuTHORS IN C&RL, 1939-79* No. of Total Articles/ No. of o/o of o/o of Author Source Authors Authors X Authors y SN(x) y F.(x) D 1 992 80.02 .8002 51.29 .5129 .2873 2 142 11.44 .9146 9.42 .6071 .3075 3 46 3.71 .9517 3.50 .6421 .3096 4 26 2.10 .9727 1.73 .6594 .3133 5 9 0.73 .9800 1.00 .6694 .3106 6 13 1.05 .9905 0.64 .6758 .3147 7 2 0.16 .9921 0.44 .6802 .3119 8 4 0.32 .9953 0.32 .6834 .3119 10 1 0.08 .9961 0.18 .6852 .3109 11 1 0.08 .9969 0.15 .6867 .3102 12 1 0.08 .9977 0.12 .6879 .3098 14 1 0.08 .9985 0.08 .6887 .3098 17 1 0.08 .9993 0.05 .6892 .3101 24 1 0.08 1.0001 0.02 .6894 .3107 •y was calculated using Lotka's law with n = 2.44, c = 59.29, and N = 1240. 232 I College & Research Libraries· May 1982 ~ Ul 1-l 0 ;:: +.1 ;:l ~ .-t Ill +.1 0 E-< Il-l 0 Q) tn Ill +.1 1:: Q) u 1-l Q) ll< 100 10 1 .1 1 9 59.29 x2.44 10 Number of Articles/Author (x) Fig.1 Number of Articles/ Author as a Percentage of Total Authors in C&RL, 1939-79 REFERENCES 20 1. Russell C. Coile, "Lotka's Frequency Distribution of Scientific Productivity," Journal of the American Society for Information Science 28:366-70 (Nov. 1977). 2. "Power Curve Fit," in Hewlett-Packard HP-25 Applications Programs (Cupertino, Calif.; Hewlett- Packard,1975),p.98-100. 3. Coile, "Lotka's Frequency Distribution," p.369-70.