College and Research Libraries Interdisciplinary Research in the Sciences: Implications for Library Organization Julie M. Hurd Accounts in both the popular media and scientific literature attest to the increasingly interdisciplinary character of scientific research. The twentieth century has seen the emergence· of problem-centered and mission-oriented research in which discoveries and developments in one discipline are synthe- sized into the research of a very different field, often with dramatic and life-altering results. This paper uses techniques of citation analysis to examine information use by scientists in a university chemistry department and offers a measure of the interdisciplinarity of the research they publish. The chemists whose published research was examined were found to make use of many journals that class outside the discipline of chemistry; over 49 % of the journals cited in a sample of their recent publications are classed in other disciplines. This study will consider implications for university libraries attempting to provide information services to scientists engaged in interdisciplinary research . • niversities are organized ac- cording to the disciplines rep- resented among their faculties and programs, and the aca- demic department is the basic unit in the structure. The research libraries that serve universities frequently mirror this structure in their organization of materi- als and services. Thus, in the sciences librarians maintain chemistry or mathe- matics or physics libraries with focused collections intended to meet most of the needs of the faculty and students in the particular discipline. An alternative or- ganizational structure is the centralized science library that may exist in lieu of or alongside departmental libraries and that serves some larger number of dis- ciplines. A considerable body of litera- ture argues the advantages and limita- tions of each of these types of organization, and a recent article by Leon Shkolnik ana- lyzes both sides of this ongoing debate.1 During the twentieth century new fields such as biophysics, molecular bi- ology, and the environmental sciences have emerged. In these fields scientists trained in diverse disciplines come to- gether to work on problems or projects that demand a broad-based perspective or to apply techniques developed in one field to research in another. These re- search teams frequently share a mission- oriented focus and may hope to solve important health-related problems or to develop new materials or procedures for some particular market. They are often more applications-oriented than their Julie M. Hurd is Science Librarian at the Science Library of the University of Illinois at Chicago, Chicago, Illinois 60680. She expresses appreciation to those UIC faculty colleagues who read drafts of this manuscript and provided helpful comments: William G. Jones, Stephen Wiberley, and John Cullars. Referee Peter Hernon also made valuable suggestions on sampling. 284 College & Research Libraries parent disciplines whose research is more basic or theoretical in nature. Evi- dence of this trend is seen in the estab- lishment of interdisciplinary units on university campuses with titles includ- ing "center," "committee," or "insti- tute," as well as in increasing university alliances with profit-sector organiza- tions either through collaborative activi- ties or through grants from corporations in support of university-based research. The fact that universities now are estab- lishing patent offices and sponsoring the development of research parks to aid in technology transfer also supports this observation. University libraries can expect in the future to serve users ... who may experience difficulties in using infor- mation sources and services organized on a discipline-based model. . On many campuses the increase in the amount of interdisciplinary research car- ried out by faculty and their graduate students has resulted in new and differ- ent needs for library collections and services. This interdisciplinary research has generated information needs that differ in significant ways from those of twenty years ago. Scientists may be using information from more than one field or in nonjournal formats such as patents or standards. University librar- ies can expect in the future to serve in- creasing numbers of users whose needs may not be confined by the boundaries of a single well-established discipline and who may experience difficulties in using information sources and services organized on a discipline-based model. DESCRIPTION OF THE PROBLEM The purpose of this research is to in- vestigate the extent of interdisciplinary research among faculty in a single uni- versity department. The method chosen, described in more detail in a following section, will be that of citation analysis, utilizing the recent publications of the faculty studied. This project is intended July 1992 to serve as a test of the technique that will subsequently be used to study the information needs of a larger group of faculty on campus. The information ob- tained will be useful in planning a new library to support research in the sciences and in designing library ser- vices appropriate for the specific user community. DISCUSSION OF PREVIOUS RESEARCH A. L. Porter and D. E. Chubin observe that "the absence of data on interdiscipli- nary research has been a bane to the study of this phenomenon."2 In fact, a literature search that attempts to identify studies of interdisciplinarity is in itself an illustration of some of the problems of interdisciplinary research. Sociolo- gists of science, information scientists, science librarians, and science policy specialists are some of the authors of research papers treating the topic. Their articles appear in a wide variety of jour- nals and conference proceedings, in- cluding publications in the basic sciences. The materials cited in this paper were found through use of the library and information science indexing services, Science Citation Index, Social Sciences Citation Index, and through cita- tions in other papers. The scatter of this literature, and thus the need to consult more than one secondary service, is a feature common to other interdiscipli- nary investigations. Thoughtful authors who deal with this topic are careful to define the ter- minology they use. A recent book by Julie Thompson Klein provides in-depth discussion on the nature of interdiscipli- nary discourse and devotes several chapters to definitions of interdisciplinar- ity. This volume also includes an exten- sive classified bibliography that Klein considers to be a representative sample of a far larger body of literature. Klein begins her analysis with a discussion of the terminology that has been employed ·• by various authors and notes her prefer- ence for the terms interdisciplinary and integrative for work that seeks to "accom- plish a range of objectives: • to answer complex questions; • to address broad issues; • to explore disciplinary and pro- fessional relations; • to solve problems that are beyond the scope of any one discipline; • to achieve unity of knowledge.''3 Klein's definition seems to apply to many problems under investigation at present in numerous university research facilities. Projects concerned with cures for a disease or focused on space ex- ploration or the environment, for ex- ample, are typically multi person efforts, and a team frequently brings together individuals whose training reflects several disciplines. In a medical labora- tory there may be a physiologist work- ing cooperatively with a biochemist; perhaps a specialist in bioengineering collaborates as well. A. J. Meadows explores diffusion of information across scientific disciplines and distinguishes interdisciplinary and trans-disciplinary activity. He prefers the term interdisciplinary when referring to the "integration of information from two different sources to create something new."4 By contrast he uses trans-discipli- nary to describe the use of information, techniques, or equipment developed in one field by practitioners in another. He cites the use of computers by both his- torians and physicists and observes that no relationship between the subjects is implied by this sharing. Meadows' defi- nition of interdisciplinary research is congruent with Klein's, and he finds in- terdisciplinary information transfer to be of primary interest. That is also the emphasis in this article. Talmon Pachevsky employs the term complex to describe those scientific fields that "have been born at the junction of different branches of knowledge and as a result of the integration of the com- ponent [sic] entirely new sciences have come into existence." He considers bion- ics, engineering psychology, and molecular biology to be examples of highly integrated fields. He uses inter- disciplinary to characterize an interme- diate level of integration below that of complex sciences for fields that represent Interdisciplinary Research 285 "only the sum of the initial intercon- nected scientific branches." 5 He goes on to speak of a still lower level of integra- tion within a particular field between branches and subdivisions of that science. His use of terms differs slightly from that of Meadows and Klein, but he is concerned with very similar issues. His article reports on a questionnaire- based effort to assess the shortcomings of discipline-oriented information sys- tems in the sciences. While his focus was on problems encountered in small developed and developing countries, his findings speak to issues common to all interdisciplinary research. Rustum Roy chose 1960 as the "birth date" of interdisciplinary research on campuses and described the situation prior to that as characterized by a balkani- zation of knowledge, with many "fief- doms, each with its army (departmental faculty), local dialect (journals), and re- ligious establishment (professional societies)." 6 With the New Frontier and the Great Society came an increased pub- lic awareness of societal problems, a re- sulting availability of public funding to address these problems, and the expec- tation that universities would share this mission and shape appropriate research agendas. Roy asserts that societal prob- lems require an "interdisciplinary" or "multi-disciplinary" approach. He con- trasts these two terms by providing oper- ational definitions: "Interdisciplinary activity on a campus is a day-to-day in- teractive mode of research (or study) where, in order to do the best work, each researcher's work demands the use of ideas, concepts, materials, or instru- ments from one or more disciplines." 7 Klein's and Meadows' definitions corre- spond closely. Roy contrasts this defini- tion to multidisciplinary research where a mission-oriented problem is broken down into "separate (typically discipli- nary) components to be carried out by separate investigators with different skills" and where the synthesis of the results is not the responsibility of the primary investigators but rather accom- plished by others at a secondary managerial level. Roy's interdiscipli- 286 College & Research Libraries nary activity is of greater interest here because it is characterized by such fea- tures as interaction and co-authorship among scientists and by local program management. Another approach to the study of in- terdisciplinarity is L. L. Hargens' survey that measured patterns of migration among disciplines and specialties. The population sampled in this study was defined as "all those who had earned doctorates in the sciences, engineering, and humanities during 1938-80 and who were residing in the U.S. in 1981." A 70% response rate provided data from 39,547 respondents. Analysis suggested that the respondents were representative of the entire population. Hargens was able to track migration streams, mapping among disciplinary groups major patterns of movement that seemed consistent with previous research that had utilized data collected from citation studies or from analyses of field similarities. Hargens did not address the issues of information needs that are the focus of the present paper, but he did validate the phenome- non of interdisciplinarity and provide cor- roborating evidence that complements other approaches.8 Greg Marlowe provided a case study of diffusion of scientific knowledge across discipline boundaries when here- counted chemist W. F. Libby's interac- tions with American archaeologists. Informal contacts and collaboration from 1946 to 1948 ultimately led to the application of carbon-14 dating tech- niques to the determination of the age of archaeological artifacts. Marlowe's ac- count, drawn from Libby's correspon- dence and interviews with some of the individuals involved, describes the diffi- culties the scholars encountered with unfamiliar concepts and terminology. Marlowe emphasizes the catalytic role of a key foundation official known to be "risk-taking" in support of cooperative and cross-disciplinary research.9 One of the more extensive studies of cross-disciplinary information use is found in Paul Metz' work that examines library circulation data at a large state university. Metz analyzes data obtained July 1992 from an online circulation system that included information on library patrons' academic status and departmental affil- iation as well as records on the library materials each had checked out during the two-day period selected for study. This provided a detailed library use "snapshot" that was subjected to a statis- tical analysis that enabled Metz to un- derstand better who uses research libraries and what materials are in greatest demand. Metz devotes much of his effort to assessing faculty use of sub- ject literatures, the extent to which such usage crosses discipline boundaries, and the implications for library organization and collection development. This study also provides a thoughtful discussion of the differences between information ob- tained from circulation statistics and that derived from citation analyses.10 Metz re- gards the two approaches as complemen- tary, and that is the view taken here as well. Each method provides a part of the larger picture and, taken together, can inform us more accurately on the elusive and complex concept of use. Other scholars investigating inter- disciplinary research have employed unobtrusive measures that are in the public domain, analyzing citations in the published literature. Citation analyses are based on the assumption that authors' practices of referencing litera- ture in their writings reflect in some fashion the utility of the cited materials. The fact that citation behavior is moti- vated by many factors is acknowledged, and use of data collected by this method should be interpreted in the context of other complementary information. A recent citation study that addresses questions of cross-disciplinary informa- tion use examines indexing of physics literature by major secondary services in other disciplines. K. E. Clark and W. R. Kinyon studied coverage of physics journals by such services as Chemical Ab- stracts (CA), Science Citation Index, En- gineering Index, and Mathematical Reviews. Inclusion of citations to physics journals was considered to measure the importance of physics to the discipline represented by each service. 11 Such a study looks outward from a single field to measure the influence of that discip- line on others. The opposite perspective is provided in Jin M. Choi's study that analyzes the journal literature in anthropology in order to assess its intellectual depen- dence on other fields. Choi analyzed ci- tations in core anthropology journals during two one-year periods separated by a span of twenty years and concluded that disciplinary communication patterns · appeared stable over the time period studied. She also looked at subspecialties within anthropology to identify intradisci- plinary communication patterns and found evidence of isolation of subdiscip- lines from one another. She characterized anthropology as a "receiver'' discipline because her analysis revealed that 70 per- cent of the literature cited was generated in other fields, including history, biomedi- cal sciences, and linguistics.12 Katherine W. McCain also considered the information needs of a single spe- cialty, the history of technology, an4 sampled articles from a core journal in that field to assess patterns of informa- tion use by scholars. She differentiated between primary and secondary sources cited and focused on the interdiscipli- nary nature of secondary source cita- tions because she hoped to trace the flow of information across discipline boun- daries. Her findings also were intended to provide useful data for collection development in the humanities. The his- torians of technology she studied ap- peared to behave like other groups of humanist scholars in their preference for monographic over serial sources; they drew on numerous other disciplines and cited works from such diverse fields as economic history, archaeology, and tech- nology itselfP McCain and James E. Bobick em- ployed citation analysis of faculty pub- lications, doctoral dissertations, and preliminary doctoral qualifying briefs to assess journal use in the Biology Depart- ment at Temple University. Their study described collection maintenance and development decisions in the Biology Li- brary and demonstrated the utility of Interdisciplinary Research 287 citation analysis in a departmental li- brary setting. 14 The present study also employs cita- tion analysis to measure the extent of interdisciplinary research activities in a group of university-based scientists. In- terdisciplinary information use by mem- bers of a science department was investigated, using an indicator of inter- disciplinaryresearch first described by Daryl E. OlUbin, A1art L Porter, and Frederick A. Rossini.15 Chubin's methodology employs a bibliometric measure, citations outside cate- gory, derived from the literature generated by the group studied and/or the litera- ture citing a paper or group of papers as an indicator of cross-disciplinary re- search activity. His studies made use of the massive Institute for Scientific Infor- mation database both to examine specific fields of research and to charac- terize the qualities of those heavily cited papers that have come to be known as citation classics. Chubin believes that this indicator of interdisciplinarity offers potential for application to "micro- level" studies such as those focusing on "the research program of a particular laboratory." That suggestion is explored in this paper in which the measure cita- tions outside category is used to investi- gate the information needs of faculty in a university department through an analysis of citations by these scientists in current publications. This research leads to an improved understanding of the detailed informa- tion needs of a particular user commu- nity and is intended to provide infor- mation useful for planning improved in- formation services for the scientists in the department studied. This paper also dis- cusses problems likely to occur when scientists' interests and information needs cross traditional discipline boun- daries and considers implications for science libraries attempting to provide information services to scientists en- gaged in interdisciplinary research. BACKGROUND The population to be studied is the Chemistry Department at the University of Illinois at Chicago (UIC), part of the 288 College & Research Libraries University of Illinois system, a Research I university with enrollment exceeding 24,000, offering doctorates in 50 fields. The UIC Library is a member of the As- soda tion of Research Libraries and is organized along broad discipline lines into the Main Library, the Library of the Health Sciences, the Science Library, the Architecture and Art Library, and the Mathematics Library. The Science Li- brary, located in one of the science build- ings, serves faculty, staff, and students in chemistry, physics, biology, and geology. It subscribes to approximately 1,600 jour- nals and serials and holds over 150,000 monographs, dissertations, documents, and technical reports in book and micro- form. Its reference collection contains the major indexing and abstracting services that fall within its subject scope and in- cludes most of the important science refer- ence sources such as Gmelin, Beilstein, and the Sad tier spectra collections. Online search services (fee-based) are routinely provided as part of a full public services program. The collection, planned to serve the needs of UIC chemists, has been developed with the guidance of an ac- quisitions policy statement first articu- lated in 1971 by the science librarian with assistance from the faculty. A collection analysis self-study project completed in 1982 for the UIC Library measured the effectiveness of collecting practices and offered suggestions for enhancing the strengths of the collections in years to come. Chemistry was one of three dis- ciplines selected for detailed study. Users saw journals as the most impor- tant component of the collection.16 The self-study employed several techniques including citation analysis to measure the strengths and weaknesses of the journal collection. A comprehensive re- cent review article containing 721 cita- tions served as the basis for the analysis which measured the percentage of the items cited that were available in the UIC Library. Because faculty judged the coverage of the starting article as repre- sentative of their interests, this assess- ment was considered a measure of the relationship between the existing collec- tions and local needs. The study task July 1992 force determined that the library held 80 percent of the journal and serial titles and 65 percent of the books cited in the review article. This study recommended that the library continue to monitor the faculty's need for journal publications and to develop lists for all the scientific disciplines whose research is supported by the campus collections. The present re- search reinforces that recommendation. The study reported in this paper pro- poses to use a technique that measures cross-category citations: • to determine the extent of interdisci- plinary research in a university science department, in this case the chemistry department; • to evaluate the scatter of sources supporting these chemists' research; • to compare the findings in this environ- ment to data reported in the literature; • to suggest implications for library or- ganization and services that follow from these findings. METHODOLOGY The university's staff directory pro- vided a roster of faculty in the chemistry department. It listed 28 individuals with rank of professor, associate, or assistant professor. These individuals were the subjects for this study. Excluded were several others designated as visiting or emeritus faculty because expectations for publication rna y be different for them than for those on the tenure track. The University of illinois Libraries have access to Current Contents databases over the campus computing network, and a search of these databases produced a list of articles authored by the chemistry fa- culty members. A total of 22 faculty had published 119 articles in the journals in- dexed by Current Contents over the two- year period covered by the online files . This represents an average output of 5.41 articles per publishing faculty member over the most recent two years. The range for this group was from 1 to 14 articles. From this population of 119 articles a stratified sample was drawn. For each author up to 3 articles were included, using all an author's publications during the period if there were3 or fewer. For those authors with more than 3 articles, 3 were selected at random from the total output. Only research articles were included in the sample; review articles were ex- cluded· if encountered. Some articles in the sample were categorized as notes or communications by the database but their length and number of references fell within the range for those classed as ar- ticles. The sample drawn for analysis comprised 59 articles. The typical article in the population studied was coauthored by three scien- tists and had 33.89 references in its bibli- ography. The number of references in the sample articles ranged from 0 to 88; two articles had no references so the sample of citations for analysis was drawn from 57 articles. The references in these 57 arti- cles (1,932 total) provided a sample popu- lation to evaluate the extent of chemists' interdisciplinary information use. FINDINGS The 59 articles in the sample were pub- lished in 26 different journals whose titles are listed in table 1. Each journal was identified with a broad subject cate- gory using the assignment in Ulrich's In- ternational Periodicals Directory, 27th edition, and this information is shown in the same table. 17 The distribution of dis- ciplines represented is shown in table 2 and, even at this level, appears to display a high level of interdisciplinary interest. Less than 60 percent of the sample arti- cles authored by chemistry department faculty were published in journals that Ulrich's classifies as chemistry. Each paper in the sample was ob- tained and the references analyzed as fol- lows. First, each cited work was classified by format of publication: journal article, monograph, conference proceedings, ref- erence work (i.e., table, handbook, data compilation, etc.), government document, dissertation or thesis, technical report, computer software, or unpublished document.18 It should be noted here that the designation conference proceedings was ~eserved for compilations of papers presented at symposia or conferences that were not published as a regular issue of a journal. Included in thecate- Interdisciplinary Research 289 gory are those proceedings appearing as occasional supplements to a journal, as an irregularly published monographic series, or as an edited collection not in series. Conference proceedings were counted this way in order to measure use of a class of materials that would not be acquired automatically with a journal subscription; library selectors would need to make individual purchase deci- sions to add these materials to a collec- tion. Furthermore, it was assumed that use of the category conference proceedings as defined here would provide collection managers with information on the im- portance of this type of material. Less than 60 percent of the sample articles authored by chemistry department faculty were published in journals that Ulrich's classifies as chemistry. Another category requiring definition is unpublished document, which includes all those references to items that were "in preparation," "in press," or "unpub- lished," as well as to those identified as a personal communication. No doubt some of these works have seen subsequent publication, although perhaps bearing a title differing from that cited in the refer- ence. Others may not be published and may prove difficult to locate, possibly only available through direct communica- tion with the author who has provided the reference. Table3 summarizes the formats of materials cited in the sample articles. Each journal cited was assigned to an Ulrich's subject category (in the same manner as were the source journals in which the citing article appeared), and a summary of that data is provided in table 4. Although the subjects in this study are affiliated with a chemistry department, their use of the journal literature extends beyond their own discipline. When citing journals outside their primary field, these scientists appeared to make most use of journals in physics and biology, but also occasionally cited materials in a number of other fields. 290 College & Research Libraries July 1992 TABLEt JOURNALS IN SAMPLE POPULATION (N = 59 ARTICLES) Journal Applied Spectroscopy Biochemical & Biophysical Research Communications Biochemical Journal (3 articles) Biochemistry Biochimica et Biophysica Acta (2 articles) Biopolymers (2 articles) Chemical Physics Letters (2 articles) Chemicke Listy Inorganic Cltemistn; (4 articles) Journal of Catalysis Journal of Chemical Physics (12 articles) Discipline· Physics Biology Biology Biology Biology Chemistry Chemistry Chemistry Chemistry Chemistry Journal of Electron Spectroscopt; and Related Phenomena Physics Physics Chemistry Chemistry Chemistry Chemistry Chemistry Chemistry Chemistry Journal of Labelled Compounds & Radioplmrmaceuticals (2 articles) Journal of Molecular Structure Journal of Organic Chemistry (2 articles) Journal of Organometallic Cltemistn; Journal of Physical Chemistn; (2 articles) Journal of the American Chemical Society (5 articles) Journal of tlte Chemical Society-Chemical Communications Nucleic Acids Research Biology Chemistry Photochemistry and Photobiologtj (3 articles) Physical Review B Condensed Matter Polyhedron Reaction Kinetics and Catalysis Letters Synlett (2 articles) Tetrahedron Tetrahedron Letters (3 articles) Tribology Transactions '''Ulrich's subject classification For each of the 57 articles in the sample the data on journal citations were entered into Microsoft File in a spread- sheet format which computed the index citations outside category (COC) where the citations outside category were calcu- lated for each article as follows: COC=(J- CH - UC) I (J - UC) where: J = total journal citations CH =#chemistry journal citations UC =#unclassified journal citations19 The COC, proposed here as an index of interdisciplinarity, ranges from 0 to Physics Chemistry Chemistry Chemistry Chemistry Chemistry Engineering 100% for the articles in the sample, with a mean of 49%. In other words, the articles studied here, published by a group iden- tified through departmental affiliation as chemists, cited only 51% of their journal references from chemistry journals. The citations to journals in the sample were sorted by discipline in order to ob- tain details on the most frequently cited journals. In the field of chemistry, the 10 most frequently cited titles are shown in ranked order in table 5. ·For comparison, the ranking obtained from the Chemical Abstracts Service list of "1 000 Most TABLE2 CHEMISTS' PUBLICATIONS BY DISCIPLINE (N =59) No. of Discipline Journals % Biology 8 13.6 Chemistry 35 59.3 Engineering 1 1.7 Physics 15 25.4 TABLE3 FORMATS OF MATERIALS CITED SUMMARY (N = 1,931 CITATIONS, 57 ARTICLES) No. of Format Citations % Journals 1,685 87.26 Monographs 122 6.32 Conference 36 1.86 proceedings Dissertations 17 0.88 Unpublished 28 1.45 Other . 43 2.23 ,. Other includes government documents, handbooks, tables, technical reports, and software. TABLE4 DISCIPLINES OF CITED JOURNALS SUMMARY (N = 1,685 JOURNAL CITATIONS, 57 ARTICLES) Discipline No. % Chemistry 782 47.36 Physics 481 29.13 Biology 304 18.41 General 35 2.11 science Othert 49 2.97 Unclassifiedt 34 ,. Percentages are based on the number of classified journal citations, 1,651. t Other includes aeronautics, astronomy, ceramics, engineering, metallurgy, mathematics, environment, pharmacy, and medicine. :f: Unclassified journals are not listed in Ulrich's Intemational Periodicals Directon;. Interdisciplinary Research 291 Frequently Cited Journals" is provided. That ranking is based on coverage analy- sis of two volumes of Chemical Abstracts, volumes 109-10 (July 1988-June 1989), a period that corresponds closely to the publication dates of the articles in the sample. Additionally, the lSI Impact Fac- tor (most current value, as appearing in Journal Citation Reports) is shown to pro- vide an indication of the frequency with which the "average" article in that jour- nal is cited in a year. (Institute for Scien- tific Information analyses have deter- mined that the average scientific paper is cited about 1.7 times per year.) The impact factor is the ratio between cita- tions and citable items published and is useful in comparing larger or more frequently issued journals to smaller or less frequently issued ones. The highest impact factor in the volume of Journal Citation Reports consulted was 48.313 for the Annual Review of Biochemistry. Tables 6 and 7 provide comparable data on the five most frequently cited journals in the fields of biology and physics. All other disciplines whose journals were cited were represented by much smaller numbers of references; no analysis of titles seemed called for under such circumstances. DISCUSSION OF RESULTS This study examined the citation prac- tices of a group of chemists associated with a single university department. Classification by discipline of the jour- nals cited in a sample of their recent publications allowed calculation of cita- tions outside category as a measure of the interdisciplinarity of their research. The findings of this study may prove useful in comparing these chemists to other larger groups of scientists and may il- luminate specific interdisciplinary re- lationships that will suggest changes or improvements in library services to this group. This section discusses the implications of the data reported in ta- bles 1-7. Table 3 provides data on the formats of materials cited by chemists arid con- firms that the single most important in- formation source is the scientific journal: 292 College & Research Libraries TABLES MOST FREQUENTLY CITED CHEMISTRY JOURNALS,. (IN RANKED ORDER) No. of Journal Citations CA Rank Journal of the American Chemical Society 133 6 Inorganic Chemistry 66 26 Joumal of Organic Chemistry 51 19 Chemical Physics Letters 49 23 Chemical Physics 45 133 Journal of Physical Chemistry 40 14 Journal of the Chemical Society (all sections) 36 98+ Tetrahedron Letters 34 9 Chemical Communications (Chemical Society) 27 27 Photochemistnt and Photobiology 24 311 July 1992 lSI Impact Factor 4.566 2.691 2.344 2.289 1.884 3.139 2.254+ 2.080 2.418 2.130 ,. A total of 99 titles classifying in chemistry were cited in the sample articles, 782 total citations. The top five journals account for 44% of the total chemistry citations. t Rank for Perkin Transactions, highest ranked of sections. TABLE 6 MOST FREQUENTLY CITED BIOLOGY JOURNALS (IN RANKED ORDER) Journal Biochemistry Journal of Molewlar Biology Journal of Biological Chemistn; Nucleic Acid Research Proceedings of the U.S. National Academy o Science . No. of Citations 66 34 26 21 19 CA Rank 18 131 17 7 lSI Impact Factor 4.006 6.555 6.491 4.298 10.032 ,. A total of 47 titles classifying in biology were cited in the sample articles, 304 total citations. These top five journals account for 54.6% of the citations in biology. TABLE 7 MOST FREQUENTLY CITED PHYSICS JOURNALS* (IN RANKED ORDER) No. of Journal Citations CA Rank Journal of Chemical Physics 236 4 Physical Review (all sections) 46 2+ Molecular Physics 23 279 Surface Science 22 65 Journal of Phttsics (all sections) 15 122+ lSI Impact Factor 3.588 3.820+ 1.964 2.917 2.173+ ,. A total of 52 titles classifying in physics were cited in the sample articles, 481 total citations. These top five journals account for 71% of the citations in physics. . t Rank for Pl1ysicnl Re·l'iew B, highest ranked of sections :f: Rank for foumal of Physics B, highest ranked of sections over 87% of the citations in the sampled articles were to journal articles. (Next most important, as measured by frequency of citation, were books). These figures fall within the range established by earlier work. Herman H. Fussier ana- lyzed citations in the writings of chemists and physicists in one of the first studies of this type and determined a serial citation rate for chemists of 93%.20 Charles H. Brown reported 94% citations to serials in his monograph published several years after the Fussier article.21 Penelope Earle and Brian Vickery col- lected data from over 65,000 citations in a sample of books and journal articles produced over the course of a year by authors in the United Kingdom. For scientific fields the serial citation rate they measured averaged 82%.22 Al- though the serial citation rate is lower in this present sample than those reported for chemists by Fussier and Brown, it does not appear that the scientific jour- nal is about to be replaced by any other publication format. Furthermore, the professional association continues to be the most important publisher of chemi- cal journals: the American Chemical Society with its Journal, Inorganic Chemis- try, Journal of Organic Chemistry, and Journal of Physical Chemistry and the Chemical Society, London with its Journal and Chemical Communications. Other im- portant publishers include the large inter- national firms of Elsevier and Pergamon which have specialized in scientific pub- lication and whose market is primarily li- braries rather than individual scientist subscribers. Table 4 provides summary data on the journal citations in the entire population reporting aggregate counts for the 57 ar- ticles with references. (Two articles had no references thus reducing the citation analysis sample to 57 articles.) It shows a high degree of interdisciplinary use of journals by the chemists in this sample. As a group, these scientists frequently cite not only the journals identified with their own discipline but also other titles identified with biology and physics. In addition, they make occasional refer- ences to journals in a number of other Interdisciplinary Research 293 scientific fields. As in the study by Porter and Chubin there is practically no cita- tion across broad field categories, i.e., to works outside the sciences. The proposed measure of interdiscipli- narity, citations outside category, ranges widely from 0 to 100% but averages 49% for the 57 papers in the study. As a group, these chemists are not discipline-bound in their use of information; and their reading, as measured by the works they cite, is not confined to only those items classified as chemistry. A typical university chemistry depart- ment, such as the one studied here, in- cludes individuals belonging to most of the major chemical specializations. Per- haps cross-disciplinary information use is more prevalent in some specialties of chemistry than others, and that question was examined in a preliminary way with the data gathered for this study. The high level of interdisciplinary information use measured for these chemists appears to argue against the narrow departmental library type of organization. There are 13 papers in this sample with COC values below 10%. This subset of the population includes authors whose information needs seem to be more fo- cused on the materials within their parent discipline than those of their departmen- tal colleagues. Do these individuals belong to any particular branch of chemistry? To explore this question, the ACS Directory of Graduate Research was consulted for information on faculty spe- cialization. This resource provides statis- tical compilations descriptive of the doctoral- and master' s-granting depart- ments of chemistry in the United States; details on enrollment, academic pro- grams, students, and faculty are sup- plied for individual departments. The faculty listed are categorized in the directory according to the major subdi- visions of the field of chemistry. Those faculty in this sample whose papers had COC values below 10% are identified 294 College & Research Libraries with either organic or inorganic chemis- try. At the other end of the COC range are 9 papers with calculated COC values of 85% or greater; these authors are iden- tified with either physical chemistry or biochemistry, specialties that, by their names alone, appear to be more interdisci- plinary in nature. The directory was used to categorize each member of the sample, and average COC values for each of the specialties represented in the UIC Chemistry De- partment were calculated: Biochemistry 85% Inorganic chemistry 29% Organic chemistry 24% Physical chemistry 64% Centralization of science collections and coordinated collection develop- ment offers enhanced potential to supply a campus with the maximum number of unique journal titles. A more detailed analysis of inter- disciplinarity variation by specialty is beyond the scope of this paper and would require a more extensive analysis of a larger set of citations. Nonetheless, these preliminary findings suggest that such an investigation might reveal sig- nificant differences in use of materials by particular specializations. Of course, these data also reflect some of the ambiguities inherent in any effort to organize knowledge along discipli- nary lines. Unsurprisingly, physical chemists are likely to make heavy use of the physics literature, and biochemists rely a good deal on publications in the field of biology. These latter two special- ties are examples of interdisciplinary fields of chemical research that have grown increasingly important during the present century. A perusal of the ACS Directory of Graduate Research demon- strates that these are now well-estab- lished specialties in chemistry and are represented, in varying per,entages, in almost every department listed in the directory. The presence of interdiscipli- nary specializations such as these makes July 1992 it more difficult to define a narrow set of library resources appropriate for chemists and, if studies such as this one are employed, it is clear that chemists' information needs are seen to overlap with those of physicists and biologists. This has implications for library organi- zation, and that issue will be addressed in the following section. Tables 5 to 7list the most heavily cited titles in chemistry, biology, and physics; these journals must be considered among the most important for this par- ticular group of chemists. If one com- pares these ranked lists with other measures of use that describe a larger universe of publishing scientists, both similarities and differences appear. The Chemical Abstracts rankings are derived purely from article counts and therefore rank highest larger, more frequently pub- lished journals. Some of these chemists' most frequently cited journals are among the largest; others, however, rank much lower on. the CA list and may reflect speCialization strengths within this par- ticular department. lSI impact factors at- tempt to correct for sheer volume and size and can be used, in conjunction with other measures, to judge relative utility of titles. Almost without exception the most frequently cited journals are high impact; this is particularly so for the phys- ics and biology titles. In fact, the Journal of Chemical Physics and the Joumal of Biologi- cal Chemistry place near the top of any of the rankings. CONCLUSIONS: IMPLICATIONS FOR LIBRARIES This study of citations to the journal literature by the chemists of a university department has produced findings that should prove useful in the improvement of science library services. This group of scientists, representative of many chemis- try departments in research universities, makes use of a variety of resources but, as for previous generations of scientists whose information use has been docu- mented earlier, they continue to rely most heavily on the primary journal, whether published by a professional as- sociation or by a major commercial pub- lishing house. Their information needs cannot be met with journals that class only in chemistry; they also use materi- als that might just as well be claimed by physicists and biologists. The high level of interdisciplinary infor- mation use measured for these chemists appears to argue against the narrow de- partmental library type of organization. A chemistry library, narrowly defined and stocked, would only partially meet their needs; a .broader, divisional science library seems better suited to support their highly interdisciplinary research. When univer- sities have operated with a departmental library structure, there has often been considerable duplication of materials; a chemistry library for these chemists would very likely feel pressure to dupli- cate some titles held in a physics or a biology library. If acquisitions budgets were open-ended and available titles less numerous, then duplication of sub- scriptions might be a reasonable ap- proach to meeting need. Few institutions can now claim that extensive, or indeed any but minimal, duplication represents the wisest deployment of scarce re- sources. In these times of declining mate- rial budgets, however, centralization of science collections and coordinated col- lection development offers enhanced potential to supply a campus with the maximum number of unique journal titles. Another type of difficulty that may be encountered by scientists engaged in in- terdisciplinary research concerns the use of secondary services to identify materi- als relevant to their research. Many of the oldest and largest indexing and ab- stracting services are discipline-based; they frequently are published by pro- fessional associations and have developed to meet the needs of scientists in the parent discipline. Although a service such as Chemical Abstracts attempts to cover the field of chemistry comprehensively, econ- omic factors eventually limit size and scope for any service. Journals in other discip- lines may be indexed selectively; less im- portant titles may be covered by fewer services or with longer time lags. For some scientists a comprehensive litera- ture search is likely to require the use of Interdisciplinary Research 295 more than one index in order to locate all relevant literature. The emergence in recent years of cross-disciplinary indexes serves those well for whom a "match" occurs: e.g., Pollution Abstracts or Environmental Bibliogra- phy, but there are many more interdisci- plinary fields than there are currently indexing services to assist their research. Given this reality what can science li- braries do to assist the growing numbers of interdisciplinary researchers? Several services seem capable of bridging disci- pline boundaries and merit cons~deration. Online databases offer capability for searching the electronic equivalents of several indexes and abstracts simul- taneously. Although variations in both in- dexing vocabulary and authors' tenninology occur across the files, a carefully designed search strategy will likely offer appreciable time savings over manual use of the same indexes. Furthermore, the availability of master indexes to a vendor's files allows a strategy to be tested for retrieval effective- ness prior to entering the databases. For example, use of Dialog's DIALINDEX can be an important early step in. identifying databases likely to contain relevant cita- tions. Recent enhancements to retrieval software now offer the ability to reduce duplication in multifile searches, and this feature, while not always able to eliminate all duplication, does result in lower print costs. Smaller libraries may also benefit from online services in that these services provide access to large costly files that might not be justifiable as subscriptions. Any library may discover an online database for which it holds no paper equiv- alent and which seems particularly suited to support an interdisciplinary query. End-user searching can be an attrac- tive alternative to mediated searching for scientists working at research fronts, a frequent location of interdisciplinary in- vestigations. When a field of study is growing rapidly, terminology tends to be in flux, and indexing vocabularies may be unresponsive to effective strategy devel- opment. In such situations suitably trained scientists can find it most effi- cient to be directly involved in informa- tion retrieval; they may then make relevance judgments while online and 296 College & Research Libraries modify strategies to reflect their assess- ment of citations found. Science librari- ans can serve as resource persons and consultants to these end-user searchers and may also direct them to suitable thesauri and search guides. Science libraries that operate in a de- centralized environment with several departmentalized collections serving the sciences will very likely hear com- plaints from their users engaged in inter- disciplinary research; these are the patrons whose journals and indexes are scattered over two or more campus loca- tions. These may also be the patrons most inconvenienced by cancellations of duplicate journal $Ubscriptions: what was once in the library in their building can now be consulted only by a trip across campus. Services to consider that address these problems include: • intracampus exchange programs for new journal issues or title pages for browsing use; • photocopy services employing cam- pus mail, couriers, or use of telefac- simile transmission to provide timely document delivery of needed materi- als between sites; • use of campus local area networks for e-mail to remote libraries for transmis- sion of reference questions, online search requests, delivery of books or journals, circulation services, interli- brary loan initiation, etc. Of course, these enhancements bear a price tag, and few library budgets are suffi- ciently expansive to launch such new ser- vices without careful projections of staffing and equipment needs. Even if user fees must be assessed, however, for many pa- trons the convenience factor can encourage use of departmental or grant funds. Finally, the importance of being aware of new research initiatives on campus July 1992 cannot be underestimated. In this regard, library committees and faculty liaisons are sources of valuable early information on new research centers developing on cam- pus. It can be much too late if a library learns from a university press release that an interdisciplinary research center has been established. Such programs have been planned without librarian input on available library resources to support them or without opportunity for library staff to begin long-range planning for acquisition of materials or development of support services. DIRECTIONS FOR FURTHER RESEARCH This study was intended to be the first in a multipart investigation of science departments' information use. The methodology described above will be util- ized to study the publications of those other departments whose faculty and students are considered the primary constituency of a divisional level science library. The findings will support plan- ning for improvements in library ser- vices in the following areas: • developing document delivery ser- vices to minimize inconvenience in use of materials located at other cam- pus libraries; • allocating science library shelving to most heavily used titles and shifting less-used ones to storage sites; • identifying journals suitable for ad- dition to the collections or for can- cellation. This study should provide a better un- derstanding of the relationships among the various science disciplines in the university environment and will in- form those involved in planning new facilities for the scientific research community. Interdisciplinary Research 297 REFERENCES AND NOTES 1. Leon Shkolnik, "The Continuing Debate over Academic Branch Libraries," College & Research Libraries 52:343-51 (July 1991). 2. A. L. Porter and D. E. Chubin, "An Indicator of Cross-disciplinary Research," Scien- tometrics 8(3-4):161-76 (1985). 3. Julie Thompson Klein, Interdisciplinarity: History, Theory and Practice (Detroit: Wayne State Univ. Pr., 1990), p. 11. 4. A. J. Meadows, "Diffusion of Information across the Sciences," ISR: Interdisciplinan; Science Reviews 1:259-67 (1976). 5. Talmon Pachevsky, "Problems of Information Services with Respect to Integration of the Sciences," Journal of the American Society for Information Science 33: 115-23 (May 1982). 6. Rustum Roy, "Interdisciplinary Science on Campus-The Elusive Dream," Chemical & Engineering News 55:28-40 (Aug. 1977). · 7. Ibid., p.32. . 8. L. L. Hargens, "Migration Patterns of U.S. Ph.D.'s among Disciplines and Specialties," Scientometrics 9(3-4):145-64 (1986). 9. Greg Marlowe, "W. F. Libby and the Archaeologists, 1946-1948," American Journal of Science. Radiocarbon 22(3):1005-14 (1980). · 10. Paul Metz, The Landscape of Literatures: Use of Subject Collections in a University Libran;, ACRL Publications in Librarianship, no. 43 (Chicago: American Library Assn., 1983). 11. K. E. Clark and W. R. Kinyon, "The Interdisciplinary Use of Physics Journals," College & Research Libraries News 50:145-50 (Fal11989). 12. Jin M. Choi, "Citation Analysis of Intra- and Interdisciplinary Communication Patterns of Anthropology in the U.S.A.," Behavioral & Social Sciences Librarian 6(3-4):65-84 (1988). 13. Katherine W. McCain, "Cross-disciplinary Citation Patterns in the History of Tech- nology," in ASIS '86, ed. Julie M. Hurd (Medford, N.J.: Learned Information, Inc., 1986), p.194-96. 14. Katherine W. McCain and James E. ,Bobick, "Patterns of Journal Use in a Departmental Library," Journal of the American Society for Information Science 32:256-61 (July 1981). 15. Daryl E. Chubin, Alan L. Porter, and Frederick A. Rossini, '"Citation Classics' Analysis: An Approach to Characterizing 1nterdisciplinary Research," Journal of tlze American Socieh; for Information Science 35:360-68 (July 1984). 16. Collection and Analysis and Management Project. Final Report (Chicago: The Library, The University of Illinois at Chicago Circle, 1982). 17. Ulrich's was used to establish the disciplines of the journals in the sample because it offered the most comprehensive, subject-arranged listing of journals. In their studies of "citations outside category" Porter and Chubin used the subject assignments made by the Institute for Scientific Information; Metz used Library of Congress call number ranges to assign to subject categories. There may be some variations in the subject assignment for an individual title, depending on the source selected, but in most cases there is no disagreement. Where Ulrich's provided Dewey or Library of Congress classification information for a title, subject assignments were usually consistent. 18. The date of publication was noted as well, although those data were not utilized in the analysis reported here. All the citation data were entered into Microsoft File to allow for statistical analysis as described in the section to follow. 19. Unclassified journals were those not found in Ulrich's. 20. Herman H. Fussier, "Characteristi.cs of Research Literature Used by Chemists and Physicists in the United States, Part II," Library Quarterly 19(2):119-43 (1949). · 21. Charles H. Brown, Scientific Serials, ACRL Monograph no.16 (Chicago: Association of College and Research Libraries, 1956), cited in R. B. Devin, "Who's Using What?" Library Acquisitions: Practice & Theon; 13(2):167-70 (1989). 22. Penelope Earle and Brian Vickery, "Subject Relations in Science/Technology Litera- ture," ASLIB Proceedings 21(6):237-43 (June 1969). • Cl) r: ca -A. -ca > 0 .. I:L I:L c ... a:: u c ACRL Approval Plans ... save time and money! ACRL publications provide practical ideas, tools, and methods for: Management Bibliographic Instruction Collection Development Special Collections Statistics & Research The convenient and cost,cutting ACRL Approval Order Plan provides automatic priority shipping of ACRL's new books at a 20% discount to ACRL members ( 10% to nonmembers) . Choose from two categories: Plan P guarantees that you will automatically receive all new ACRL publications* including CLIP Notes. ACRL publishes approximately 5 to 6 new titles each year. · Pe riod icals and subsc ripti on ite ms nor included. Plan PC exclusively for titles in the CLIP Notes (Col, lege Library Information Packets) series. CLIP Notes collect data and sample documents from academic libraries to assist librarians in estab, lishing or refining services and operations. You may change or cancel your approval order at any time . Return privileges guaranteed. Enrollment Form 0 Plan P guarantees shipment of all n ew non-periodical ACRL publications including CLIP Notes . 0 Plan PC guarantee s shipment of all new CLIP Notes. Please give your ALA membership number if you are eligible for the 20% ACRL membership discount. Note : Personal membersh ips may not be used to qualify institutions for the ACRL member dis count . ALA membership number ---------- P.O.# _______ __ AssociAT ION O F Ship to: ----------- COLLEGE Bill to ------------ & RESEARCH LIBRARIES Return this form to: ACRL Approval Order Plans, ACRL/ALA,50 Ea~t Huron Street, Chicago, IL 60611 -2795. To order by phone, call 1 (800) 545 -2433 or fax: 312·944-2641.