College and Research Libraries Dissertations: A Study of the Scholar's Approach Joan M. Repp and Cliff Glaviano Four academic libraries in Ohio participated in a study exploring how and why extramural researchers access locally produced dissertations. Abstracts and indexes in hard copy were identified as the resources relied upon most heavily by all academic disciplines, with Disserta- tion Abstracts International and related University Microfilms International products being named most often. Researchers, regardless of discipline, were successful searching by subject more often than using any other form of access. Dissertations were most frequently requested by those writing theses, dissertations, or research papers for publication. These findings may have implications for in-house cataloging of locally produced dissertations or for institutional participation in an indexing and abstracting program. he dissertation is a unique for- , mat for scholarly information that is not available in its origi- nal form through established professional or commercial channels. Though substantive information from the dissertation may subsequently appear in the scholarly literature as a journal article or book chapter, the original dissertation is deposited, virtually without exception, at the degree-granting institution. Al- though an occasional dissertation is pub- lished in its entirety, more often this fol- lows considerable editing of styl~ or content. Since hard copy of the disserta- tion is not widely disseminated, access to the information it contains becomes a con- cern for the scholarly community. As Library of Congress priorities pre- clude cataloging of even depository copies of dissertations submitted for copyright, no LC cataloging for dissertations appears on the bibliographic utilities, and full re- sponsibility for bibliographic control falls to the degree-granting institution. Aca- demic libraries necessarily must either ne- gotiate with their parent institution, or themselves originate, all policies concern- ing preservation, circulation, availability, shelving and degree of bibliographic con- trol for local dissertations. Further, only the library and degree-granting institution determine the extent to which access to the dissertation is shared through an in- dexing service such as that provided by University Microfilms International or through contributing cataloging copy to a bibliographic utility. Implicit in accepting responsibility for bibliographic control of the dissertation is its indexing, so that it might be readily lo- cated when needed. Presently most aca- demic libraries provide full descriptive and subject cataloging for local disserta- tions and input such copy on OCLC, RLIN, or WLN. Since the dissertation for- mat is developed by the university's grad- uate school, title pages are formulaic and descriptive cataloging can easily be done by a paraprofessional cataloger. Subject analysis is an altogether different matter due to the timely, experimental nature of dissertation research. It is often both diffi- cult and time-consuming for a profes- sional cataloger to select LC subject head- ings (LCSH) appropriate to these in-depth, highly specialized studies. This problem of subject analysis is most often Joan M. Repp is Chair, Access Services and Cliff Glaviano is Principal Cataloger, University Libraries, Bowl- ing Green State University, Bowling Green, Ohio 43403. 148 j noted in science and technology though works in all disciplines potentially can deal with concepts for which LCSH terms are nonexistent at the time of cataloging. The problem of adequate classification and subject analysis is compounded by changing staffing patterns in cataloging departments. Since cataloging has been automated, the number of original cata- logers has decreased. 1 Each cataloger is therefore responsible for a wider group of disciplines. This reduces the probability of high subject expertise in any one of them. Coincident with the declining number of original catalogers, the number of disser- tations produced has risen: doctoral de- grees in the United States, for example, rose from 6,600 in 1950 to 32,700 in 1982.2 DISSERTATION USERS Users of dissertations can be divided into two groups: intramural scholars and extramural scholars. Comparatively little is known about either group's behavior relative to seeking and using disserta- tions. Informal observation from the refer- ence desk suggests that intramural scholars writing their own theses and dis- sertations may use those already pro- duced at their institution for reasons be- yond obtaining scholarly information. In addition to checking bibliographies for references and determining the scope of an extant dissertation's coverage of a topic to avoid duplicating another's research, intramural scholars use local dissertations to determine a format acceptable to their institution, to evaluate research ap- proaches that have been approved by par- ticular dissertation committee chairs, and for other reasons not related to scholarly content. The in-house tools for locating a disser- tation written at an institution signifi- cantly exceed those available to the extra- mural scholar, who may be assumed to be interested in the dissertation's content alone. While the extramural scholar can be expected to use standard indexes, ab- stracts, and databases, the intramural scholar, provided by his library with LCSH subject analysis and potentially, additional local access by adviser name, department/major, or other approach, Dissertations 149 probably can ignore standard sources alto- gether for locating local dissertations. Often libraries that do not provide full cataloging for dissertations aid the intra- mural scholar through separate in-house indexes or special shelving arrangements, amenities lost to the extramural scholar. Should the library provide only minimal on-campus access to local dissertations, promoting the use of standard sources, the intramural scholar can still take advan- tage of maintaining close contact with oth- ers in his discipline, many of whom will have good knowledge of the dissertations produced in the discipline and available at the institution. It is reasonable to assume that informa- tion needs of the intramural and extramu- ral scholar can differ in scope and empha- sis and that the intramural scholar has resources in addition to standard sources for accessing local dissertations. Standard tools of potential use to both groups have been enhanced considerably by technol- ogy. Many institutions granting doctoral degrees participate in the Dissertation Ab- stracts International (DAI) program, which generates indexes in both hard copy and online through commercial data- base services. Broad subject access is avail- able in DAI and American Doctoral Dis- sertations (ADD) while LSCH access is available for those cataloged dissertations in the portion of the OCLC database avail- able on BRS. DAI hard copy is indexed also by keyword from title and by author; ADD additionally by author and by insti- tution. The database available on BRS and Dialog is compiled from both ADD and DAI. In addition to searches available from the hard-copy indexes, search terms con- structed from Boolean combinations of keywords from title or the entire abstract can be used on BRS and Dialog to locate relevant dissertations. Meanwhile, the technology for using the commercial data- bases is widely available to the scholar at the academic library: in 1980, all forty- three academic libraries responding to a survey indicated they had access to Dialog for database searching.3 While answers to the basic questions of how a patron gains access to a local disser- 150 College & Research Libraries tation and how a patron uses its informa- tion can be assumed from observations by experienced reference librarians, no pub- lished research exists to affirm or deny these assumptions. Costs associated with participating in the DAI program and costs associated with original cataloging of locally produced materials should gener- ate continuing interest in the question of how best to serve intramural and extramu- ral scholars who seek dissertation infor- mation. Clear reasons to modify catalog- ing procedures to include, for example, access by thesis adviser as suggested by George Harris and Robert Huffman; rea- sons to support the cost of full cataloging of dissertations; and/or reasons to justify the expense of participating in a coopera- tive indexing program need to be deter- mined. Assessment of patron behavior in seeking and using dissertation informa- tion becomes necessary before institu- tional procedures can be modified or pro- gram participation evaluated. LITERATURE SURVEY Little has been published in the litera- ture on the handling, cataloging, classifi- cation, and subject analysis of locally pro- duced theses and dissertations. The literature concerning access is dated enough to preclude mention of the com- mercial databases, and institutional par- ticipation patterns have changed enough since the 1970s to make descriptions of the hard-copy indexes and their coverage quite. misleading. Julie Moore's article, "Bibliographic Control of American Doctoral Disserta- tipns: A History,'' presents a brief histori- cal summary of "national listings of dis- sertations which are a unique alternative to the control of published and un- published dissertations in the United States and Canada. " 4 The same topic, dis- cussed from a slightly different perspec- tive and with considerable perspicacity, may be found in Donald Davinson' s The- ses and Dissertations as Information Sources. 5 The second part of Moore's work, ''Biblio- graphic Control of Doctoral Dissertations: An Analysis,'' focuses on a comparison of the usefulness and limitations of Ameri- can Doctoral Dissertations and Disserta- tion Abstracts. She also found that only March 1987 2.1 percent of the dissertations she stud- ied were published as books, an addi- tional 1.4 percent could be identified as parts of a book, and that 15.2 percent were published as journal articles after heavy rewriting. 6 Calvin Boyer's work, The Doctoral Dis- sertation as an Information Source, itself orig- inally a dissertation, assessed the extent to which dissertations in selected sciences produced between 1963 and 1967 served as information sources by studying cita- tions and dissertation-based publica- tions.7 Kelly Patterson and others, in "Thesis Handling in University Li- braries," reported the results of a survey of ninety universities granting doctoral degrees to determine local practice in ''binding, cataloging, classification, stor- age and checking of format practices" with the intent of recommending process- ing practices. 8 Lois M. Pauch' s "Thesis/ Dissertation Processing and OCLC'' de- tails the impact of OCLC on cataloging procedures for theses and dissertations at the University of Illinois at Urbana- Champaign, speeding the process from three hours to one-half hour. 9 Most recently, George Harris and Rob- ert Huffman completed a study of catalog- ing, classification, and subject analysis of locally produced dissertations. ''Catalog- ing of Theses: A Survey'' summarizes their findings. In addition to surveying ac- tual practices for dissertation cataloging in academic libraries, in effect what currently is done to serve the needs of intramural scholars and reference librarians, they found that without AACR2 rules specific to dissertations and the guidance of LC cataloging ''libraries are forced to impro- vise."10 · Though the literature is informative re- garding current practice in the local biblio- graphic control of dissertation informa- tion and its sharing on the bibliographic utilities, studies exploring the content of the various dissertation indexes, abstracts and databases, and the use of dissertation information within the scholarly commu- nity are so outdated as to be extremely misleading. PURPOSE OF THE STUDY The purpose of this study is to assist in the evaluation of current cataloging prac- tices as applied to dissertations produced in-house at four academic libraries in Ohio. The study is limited to exploring the behavior of the extramural scholar and at- tempts to identify user by type of institu- tional affiliation; purpose for which the dissertation was used; and what tools and approaches were successfully used in identifying and locating the dissertation. Hypotheses 1. Scholars access dissertations as an aid to writing dissertations . 2. Due to the research orientation of the parent institution, more borrowing of dis- sertations is done through Association of Research Libraries than other types of li- braries. 3. As a generalization, the dissertations requested are newer by date than older but this behavior varies measurably be- tween social scientists, hard scientists, and scholars in the humanities. 4. The major access is through Disserta- tion Abstracts International, its related precursors and products . 5. Subject access is least often used to locate a relevant dissertation since subject control of dissertation information is very limited. Assumptions 1. Interlibrary loan usage is an accurate representation of extramural use. 2. The behavior of the extramural scholar differs significantly from that of the intramural scholar. 3. The dissertations of the participating institutions available through OCLC, in- dexes, and databases accurately represent all of each institution's locally produced dissertations. Methodology Four academic libraries in Ohio partici- pated in this study: Ohio State University, University of Cincinnati, University of Toledo, and Bowling Green State Univer- sity. A census was taken of all dissertations borrowed through the interlibrary loan departments of the cooperating institu- tions for a period of one year. A questionnaire was mailed out with Dissertations 151 each dissertation and a record was main- tained in-house listing the borrowing in- stitution, the subject of the dissertation as identified in the dissertation or by the de- partment supervising the dissertation, discipline of the dissertation, and date on the title page. The borrower was re- quested to return the completed question- naire with the dissertation. There was no ·follow-up. The data gathering covered a twelve-month period from July 1, 1983, to June 30, 1984. Characteristics of the-Institutions Selected The four institutions selected repre- sented a mix of academic programs lead- ing to the doctorate and represented medium-sized and large universities. The institutions shared the following charac- teristics. All four 1. fully cataloged their dissertations ac- cording to national standards; 2. entered their cataloging record in a national bibliographic database, in this case, OCLC; 3. participated in the Dissertation Ab- stracts International program; 4. permitted dissertations to circulate through interlibrary loan (ILL). There are some variations among the in- stitutions in the following areas: 1. Each began entering records in OCLC at different times (earliest, 1972; most recent, 1978). 2. Participation in Dissertation Ab- stracts International varied from institu- tion to institution (earliest, 1954; most re- cent, 1973) and by discipline within institutions. 3. One institution limited loans of dis- sertations produced after 1954 to recipro- cating institutions only, since these disser- tations are available through University Microfilms International. 4. Charging practices for ILL varied widely and in<;:luded no charge, postage only, $4.50 flat rate, or reciprocal charg- ing. 5. Two institutions would not lend dis- sertations to high school libraries . There wen. no other restrictions on lending. RESULTS In all, 542 questionnaires were sent out with the dissertations as they were circu- 152 College & Research Libraries lated through ILL. Of these, 269 were re- turned for an overall return rate of 49.6%. The percentage base of several sub- populations was high enough to be reli- able. Results were treated with SPSS sta- tistical package. Table 1 summarizes the number of questionnaires sent and re- tUrned by institution. The borrowing institutions were catego- rized as holding membership in the Asso- ciation of Research Libraries (Bowker An- nual, 1982); non-ARL university libraries; college libraries; other institutions of higher education including junior and community colleges, technical and trade schools; public libraries; governmental corporate libraries; business corporate li- braries; other corporate libraries including museum and hospital libraries; and high school libraries. Institutions of higher education ac- counted for 93.2% of the loans, corporate libraries 5.1%, and public libraries 2.2%. One dissertation was loaned to a high school library. As can be seen in table 2, the most frequent borrowers were ARL member libraries. The number and percentage of disserta- tions borrowed by discipline is reflected in table 3. Of the dissertations borrow~d, the highest number, 235 (43.4%), were bor- March 1987 rowed in academic disciplines from the so- cial sciences. Of the social science loans, 176 (75%) dealt with education, excluding educational psychology. The results were highly institution dependent. No attempt was made to relate program offerings and number of degrees granted in each disci- pline to the observed frequency of bor- rowing of the discipline. Table 4 compares the vintage and disci- pline of the dissertations borrowed. The currency of the dissertations borrowed ranged widely from 1906 through 1983, and some variation among the disciplines relative to the recency of dissertations bor- rowed is apparent. The data support the assumption that recently produced dissertations are more heavily used than those written earlier. Figures produced at the same rate for the remainder of the 1980s would result in an N of 395 borrowings for the decade. The figures of table 4 must be interpreted with some caution, however, since a number of pertinent factors are unknown, including the number of dissertations produced by the cooperating institutions in each disci- pline for each time period. Certainly the trend would be upward for total numbers produced at the four institutions, though program empha_sis would result in skew- TABLE 1 RESPONSES BY INSTITUTION University of Toledo Bowling Green State University University of Cincinnati Ohio State University Total *Total less than 100% due to rounding. Number Sent 49 84 121 288 542 Number Returned 31 61 59 118 269 TABLE2 %Returned 63 73 49 41 N/M BORROWERS BY TYPE OF INSTITUTION Type of Institution Member of ARL Other University College (4 year) Other Institution of Higher Education Corporate Libraries Other Total *Total less than 100% due to rounding. Number Loaned 291 181 20 12 28 9 541 %of Total Returned 11.5 22.7 21.9 43.8 99.9* %of Total 53.7 33.4 3.7 2.2 5.1 1.7 99.8* Dissertations 153 TABLE3 BORROWING BY DISCIPLINE Discipline of Dissertation Social Science Applied Science/Technology Languages/Literature History Business Arts Pure Sciences Philosophy Health Sciences Religion Total *Total less than 100% due to rounding. Number 235 98 65 40 38 26 24 10 4 1 541 Percent 43.4 18.1 12.0 7.4 7.0 4.8 4.4 1.8 .7 .2 99.8* TABLE4 NUMBER OF DISSERTATIONS BORROWED BY DISCIPLINE/DATE Discipline 1906-49 Applied Science/Technology 3 Arts 1 Business 2 History 10 Languages/Literature 3 Health Sciences 0 Philosophy 1 Pure Sctences 5 Religion 0 Social Sciences 9 Total 34 ing of figures for popular courses of study and possibly a decline in number of disser- tations produced in less popular disci- plines over time. Another factor difficult to assess accu- rately from the data is the influence of in- dexes on the findings. Indexes and ab- stracts, most particularly Dissertation Abstracts International (including its pre- cursors and related products) are the tools most heavily relied upon by extramural scholars, yet these tools are neither as comprehensive nor as complete as one would wish. Scholars were asked to identify the pur- pose for which the dissertation would be used, or, in other words, to identify the anticipated end product of their research. Of extramural users, 57% indicated use in relation to the borrower's own disserta- tion or thesis; 30%, use for an academic paper or publishable research project other than a thesis or dissertation; 8.3%, use in solving a research problem for 1950-59 1960-69 1970-79 1980-83 Total 4 5 2 3 3 0 1 3 0 10 31 10 55 26 98 4 9 7 26 2 16 15 37 2 21 4 40 8 39 12 65 0 3 1 4 2 2 4 10 3 9 4 24 1 0 0 1 17 115 84 235 49 269 157 540 which publication was not anticipated; while other uses accounted for the re- maining 4.7%. Respondents indicated that 90% of their anticipated products would be associated with an institution of higher education, the remaining 10% in government or business and industry. Overall, scholars indicated that they were borrowing a dissertation in the same academic discipline as their work in prog- ress 69% of the time. Borrowers from so- cial sciences, history, and languages/liter- ature borrowed dissertations in their academic discipline at a rate of 80% or greater. In addition to indicating the purpose for which the dissertation was being re- quested, the borrowers were asked to name the tools they had used to identify and locate the dissertation, as well as their approach to the particular tools. Table 5 reflects the resource used by type of borrowing institution. Resources borrowers used most often in locating dis- 154 College & Research Libraries March 1987 TABLES GENERAL CATEGORY OF RESOURCE USED BY TYPE OF BORROWING INSTITUTION ARL OtherUniv. Resource N % N % Index/ Abstract 80 57.8 38 44.2 Journal, Book, Newsletter 21 15.1 16 18.6 Automated Database 31 22.3 29 33.7 Word of Mouth 7 5.0 3 3.5 Total 139 100 86 100 sertation information were printed in- dexes and abstracts followed by auto- mated databases. As a group, non-ARL university libraries show the highest rate of database searches at 33.7%. The re- maining non-ARL institutions searched automated databases at rate of 27.8%. Though both categories of institutions ex- ceeded the rate of database searching in ARL libraries (22.3%), the variation from chance distribution was not significant us- ing a chi-square test. The scholar's approach to finding dis- sertation information appears to depend in part on the academic discipline of the borrower. The categories of resources used by those scholars identifying their disciplines as applied science/technology, languages/literature, the arts, or the social sciences are tabulated in table 6. These dis- ciplines accounted for 80% of the total re- sponses and the remainder appeared too scattered to yield meaningful results. Borrowers in these categories relied more heavily on printed indexes and ab- stracts than on any other form of access surveyed. Other printed sources, such as newsletters, books, and journal articles Type of Institution Oilier :igher~. NCorpor~e Other Total N % N % 7 6 5 0 18 38.9 3 18.8 1 25 129 49.0 33.3 8 50.0 2 50.8 53 20.2 27.8 3 18.7 0 0 68 25.9 0 2 12.5 1 25 13 4.9 100 16 100 4 100 263 100 were identified as chief finding tools rang- ing from a rate of 32% for applied science/ technology to a low of 8.6% for social sci- ences of the total number of responses in the respective discipline. Social sciet:ttists indicated using automated databases at a rate second only to their use of indexes and abstracts, while other disciplines mentioned database use at a much lower rate. In fact, scholars from the arts identi- fied word-of-mouth sources more often than automated databases, but not at a rate that the chi-square test showed as sig- nificant. A chi-square test of table 6 data indi- cated variations significant at the .OS level in the source choices of borrowers from applied science/technology and the social sciences, the former group relying on non- index printed sources more heavily than could be expected, and the latter on using automated databases. The significant and near-significant (arts use of word of mouth) results clearly indicate discipline- specific influences on borrowers' use of sources. Though no data were gathered, the influences may include the availability of specialized indexes in certain disci- TABLE6 GENERAL CATEGORY OF RESOURCE USED BY SELECTED DISCIPLINE Applied Discipline Science/ Languages/ Social Technology Arts Literature Sciences Total Resource N % N % N % N % N % Index/ Abstract 22 44.0 14 60.9 21 65.6 53 51.0 110 52.6 Journal, Book, Newsletter 16 32.0 1 4.3 6 18.8 9 8.6 32 15.3 Automated Database 9 18.0 3 13.0 5 15.6 40 38.5 57 27.3 Word of Mouth 3 6.0 5 21.8 0 0.0 2 1.9 10 4.8 Total 50 100 23 100 32 100 104 100 209 100 plines or that scholars in certain disci- plines are more sophisticated in using au- tomated databases either specific or germane to their area of study. Certainly the quality and availability of databases in a discipline influence the scholar's will- ingness to search them, while both the na- ture of the discipline and the nature of dis- sertation information probably influenced all search behaviors. For example, the need for timeliness in applied science/ technology may induce its scholars to use the discipline's journal/newsletter net- work more diligently than scholars use similar resources in other fields. Further speculation suggests that the experimen- tal nature of dissertation research may · lead to dissertations being cited more of- ten in the literature of applied science/ technology, thus making printed sources more valuable to the discipline. Similarly, the performance nature of some music and fine-arts dissertations may result in deficient or nonexistent abstracts and per- haps a higher incidence of nondescriptive titles, leading borrowers from the arts to rely more heavily on word of mouth as an appropriate approach to such informa- tion. A closer examination of three or four broad categories of searching resources (the number of word-of-mouth responses being quite low) is of some interest. Of to- tal responses, the chief source of informa- tion was designated as ''index/ abstract'' 135 times. Of this number, Dissertation Abstracts International and Comprehen- sive Dissertation Index accounted for 99 and 21 responses respectively, or 89% of all indexes and abstracts used. While spe- cialized indexes and abstracts accounted for the remaining 11% reported, the only other titles reported more than once were MLA Annual Bibliography, Music Index, and Agricultural Education, each cited twice. Borrowers reporting use of nonindex print sources to locate dissertations found information in subject bibliographies 30% of the time. Professional newsletters ac- counted for 28%, separately published monographs 27%, and journals 15% of the print sources. Of all responses, only 22% indicated Dissertations 155 finding dissertation information by data- base searching, a lower rate than the 27.3% indicated for the selected disci- plines of table 6. Some 86% did not search the database themselves but had searches performed by third parties. All those who did their own searches were from non- ARL university libraries. Borrowers from the social sciences performed 70.2% of all database searches, their 40 searches ac- counting for 38.5% of all searches in the social sciences. Responses to questions re- lating to use of specific databases and search terms used did not yield useful results. The majority of respondents ei- ther did not know the name of the data- base searched for them by an intermediary or supplied the acronym of the local in- house database. Though several indicated they had accessed Dissertation Abstracts Online, there were no clear trends or im- plications for using it or any other data- base named. TERMS USED TO ACCESS DISSERTATIONS Scholars were asked to indicate the term by which they located the material bor- rowed regardless of the tools they used for access. In examining the responses it be- comes apparent that some confusion ex- isted concerning the differences among subject, academic discipline, and search term used to access the dissertation. Re- sponses received indicated successful searching by subject 53.1 %, by search term 8.9%, by browsing 5.4%, by aca- demic discipline 4.3%, and by other means .8% of the time. Although search term was intended to convey the concept of nonauthor, -title, or -subject searching, such as searching by keyword from title, perhaps combined with subject and/or discipline phrases or dates using Boolean logical operators; it is unclear whether ex- tramural scholars responding made much distinction between subject, search term, or academic discipline in categorizing the terminology by which they located disser- tations. As large numbers of respondents (44.6% of overall returns) list DAI or CDI as their chief source of information, it be- comes necessary to investigate the termi- 156 College & Research Libraries nology used in these indexes. In both, groups are arranged by broad subject terms equivalent or nearly equivalent to academic disciplines (e.g., agriculture; bi- ology; chemistry, organic; ecology; micro- biology, etc.), which are then subarranged by keyword from title of the dissertation that often appears to be the subject of the work. In fact, prefatory material in CDI calls such a keyword from title the ''sub- ject keyword.'' CONCLUSIONS Based on the initial hypotheses, the fol- lowing conclusions can be drawn: 1. Most extramural scholars use disser- tation information to produce other schol- .rrly and theoretical works such as disser- tations, theses, or research papers, rather than applied works. As some disserta- tions are published later in whole or in part, it cannot be assumed that disserta- tions have only theoretical rather than practical applications, however. 2. Though research libraries as identi- fied by ARL membership borrowed the majority of all dissertations loaned during the project, other university libraries also accounted for a substantial number of loans. 3. Overall, more recent dissertations were requested than older works, and, al- though there were variations among the disciplines, the data will not support discipline-specific trends concerning the need for timely dissertation materials. Closer examination of the history of the granting of U.S. doctoral degrees leads to the realization that since 1920 more than half were granted from 1971 to 1983. It may be premature to speculate as to whether scholars actually consider the re- cency of dissertation information as they choose a search tool or whether any disci- pline can establish a preference for recent dissertation material given the extremely skewed distribution of dissertations pro- duced. 4. Access was achieved more often through indexes and abstracts than any other source, in fact, more than all other sources combined. Dissertation Abstracts and its related products were the most heavily used tools within the category de- March 1987 spite limitations such as their lack of com- pleteness in certain institutional holdings, and employment of broad, author- assigned subject categories. 5. Extramural scholars most commonly approach dissertation information by sub- ject. It is somewhat perplexing to realize that most scholars found their information by using subject terms that mimic the names of academic disciplines and that are often less descriptive than keywords from the dissertation's title. DISCUSSION AND SUMMARY Though it is clear that the majority of ex- tramural scholars employ dissertation in- formation in formulating theses, disserta- tions, or academic research projects, there is no obvious indication of how such infor- mation is employed. Specifically, there is considerable difference between the bor- rower's use of a dissertation in ensuring that a project will not duplicate earlier work and use of dissertation data/results/ arguments for comparison or contrast with the borrower's own ideas. Presum- ably, the majority of extramural scholars use dissertation information for compari- son or contrast since these borrowers will have read the abstracts of relevant disser- tations before requesting them on ILL, borrowing dissertations most similar to their own research in relatively rare in- stances in order to determine that a partic- ular aspect of the topic has been fully ex- hausted. Even though access to dissertations is self-selected by the degree-granting insti- tution through the amount of cataloging input in the utilities and the extent of par- ticipation in indexing services, there is an . audience for dissertation information out- side research and university libraries. Al- though access is limited by indexing ser- vices promoted and designed to serve higher education and cataloging data shared chiefly with other academic insti- tutions, a portion of that CilUdience still seeks dissertation information. To follow another line of thought, as- sume that dissertations per se are of inter- est only to other scholars writing their dis- sertations and that the process is an ,. + ~·. exercise in "how to do" research. This po- sition has been supported by Bernard Berelson' s research in which graduate fac- ulty were asked whether the dissertation was an original contribution to knowledge or an exercise in research training. More than 50% responded that it is primarily re- search training; only 15% responded that it is ~rimarily a contribution to knowl- edge. 1 Assuming that authors of dissertations containing significant contributions to hu- man knowledge will disseminate this in- formation in another form, such as a monograph or as a journal article, perhaps the dissertation needs no wider dissemi- nation than it has at present, and needs only minimal bibliographic control. Nev- ertheless, some bibliographic control, in- cluding only access by author, title, and institution, would be necessary for those dissertations produced as exercises in re- search and for those cases in which au- thors may not realize the potential impact of their research. Indications are that if dissertation con- tent remains of value due to unique quali- ties of dissertation research and its rigor- ous methodology (something Davinson and Boyer both commented on in their works), then libraries might expect more interest in the dissertation from outside academia, especially as database search- ing becomes increasingly available to busi- ness. Trends in higher education would also indicate potential increased use of dissertations outside academia as the number of Ph.D.'s working outside higher education increases. Though the survey was unable to gauge changes in use patterns by categories of borrowers and only surveyed use by ILL of the four libraries' collections, there is some feeling of increased use outside academia, espe- cially since so many dissertations are available at reasonable prices through University Microfilms International (UMI). UMI products have a tremendous influ- ence on the availability of dissertation ac- cess and consequently on all those seeking dissertation information. Dissertation Ab- stracts International (DAI), Comprehen- sive Dissertation Index (CDI), and Ameri- Dissertations 157 can Doctoral Dissertations (ADD) appear to be the most comprehensive hard-copy tools available to extramural scholars. The popularity of DAI and CAl over special- ized indexes in the survey certainly invites speculation that extramural scholars pre- fer comprehensive coverage to the more focused approach offered in the special- ized indexes. Unfortunately, the UMI in- dexes, hence the database, are incom- plete. Of potential significance, participation in the UMI indexing programs may lead to decisions against the ILL loaning of disser- tations by participants who may suggest that prospective borrowers obtain such materials from UMI. Considering such possible restriction on ILL borrowing, it may be necessary for future investigations into the use of dissertation information to consider UMI order files as more indica- tive of demand for and use of dissertations than ILL borrowing. If, in the future, rela- tively few academic libraries allow unre- stricted borrowing of dissertations in- cluded in the DAI program, UMI may be the only good source of information on scholarly demand for dissertations. Though the responsibility for providing subject access, as with providing for dis- sertation bibliographic control, falls to li- braries at the degree-granting institutions, these libraries may be well advised to ex- amine policies that prescribe LCSH sub- ject analysis for dissertation cataloging, especially if they participate in the DAI program. Though it may seem an abdica- tion of responsibility to leave subject anal- ysis to commercial indexers or dissertation authors, applying subject analysis at the time of cataloging is of extremely limited value to the extramural scholar. It does not appear certain that including subject access with dissertation cataloging input on the utilities is of sufficient merit to justify its expense relative to its avail- ability. Since only a portion of OCLC is available on BRS, only a sampling of LCSH access is available to dissertations nationwide. The chief benefit of including subject analysis at time of cataloging is still to the intramural scholar, arguably a good investment in providing service to local clientele through manual or online public . 158 College & Research Libraries catalogs, even though these same re- searchers would be expected to behave much as the survey respondents when ex- tending their inquiries to the external scholarly community. As subject access to the OCLC database expands under a configuration expected to be in place by 1988, there may be greater reason for providing subject analysis in in- put cataloging for locally produced mate- rials. On the other hand, as more search- ing hardware and software, more subject databases and offline databases on CD- ROM become available to potential end users, researchers may begin to make less use of the subject approach to dissertation information and make better use of the ca- pabilities of keyword searching whether in the UMI database or a discipline specific database. Because the literature amply supports the superiority of Boolean combinations of keyword search terms over the use of standardized subject headings in the con- text of the online environment, the ques- tion of whether or not to continue subject analysis in cataloging these materials, for potential use by other members of a biblio- graphic utility or potential use in the aca- demic library public catalog, may yet be moot. SUGGESTIONS FOR FURTHER RESEARCH It should be apparent from the discus- sion of this study that there are many top- ics related to, or concerned with, the loca- tion and use of dissertation information March 1987 that are worthy of further study. Among the greatest needs: 1. Further in-depth study of the mani- festations of dissertation information, dis- sertations and their derivations, in the scholarly community. Existing studies are quite dated, and even narrow, discipline- specific treatment on use would be very helpful. 2. A closer exploration of the relation- ship, if any, between particular disciplines and the recency of dissertation informa- tion. This could take dissertation-derived manifestations into account. 3. Examiflation of whether or not the use of dissertation information is increas- ing outside academia. 4. A study of how anticipated use of dissertation information might influence the choice of tools or choice of search term in locating relevant information. 5. Exploration of the relationship be- tween the choice of search tools and a dis- cipline: determining which factors con- tribute to furthering effective research methods concerning dissertation informa- tion; discovering which methods of train- ing and promotion might be furthered by library involvement. 6. Full study of the content and cover- age of various general dissertatirm indexes with the intent of determining their reli- ability in providing full availability of dis- sertation information to respective disci- plines, perhaps with the aim of conveying any discovered gaps to the scholarly com- munity and suggesting more comprehen- sive resources available beyond a given hard-copy index. REFERENCES AND NOTES 1. Peter Spyers-Duran, "The Effects of Automation on Organizational Change, Staffing, and Hu- man Relations in Catalog Departments" in Requiem for the Card Catalog ed. Daniel Gore, Joseph Kimbrough and Peter Spyers-Duran (Westport, Conn.: Greenwood, 1979), p.36. 2. U.S. Bureau of the Census, StatisticalAbstractafthe United States (Washington: U.S. Govt. Pr. Off., 1985), p.157. 3. Connie Lamb, "Searching in Academia" Online 5:80 (Apr. 1981). 4. Julie Moore, "Bibliographic Control of American Doctoral Dissertations: A History," Special Li- braries 63:227-30 (May/June 1972). 5. Donald Davinson, Theses and Dissertations as Information Sources (Hamden, Conn.: Linnet, 1977), p.44-47. Dissertations 159 6. Julie Moore, "Bibliographic Control of American Doctoral Dissertations: An Analysis," Special Li- braries 63:289 Ouly 1972). 7. Calvin James Boyer, The Doctoral Dissertation as an Information Source: A Study of Scientific Information Flow (Metuchen, N.J.: Scarecrow, 1973), p.23. 8. Kelly Patterson, Carol White, and Martha Whittacker, "Thesis Handling in University Libraries," Library Resources & Technical Seroices 21:274-85 (Summer 1977). 9. Lois M. Pausch, ''Thesis/Dissertation Processing and OCLC, '' Research Libraries in OCLC: A Quar- terly 4:1-4 (Oct. 1981). 10. George Harris and Robert Huffman, "Cataloging of Theses: A Survey," Cataloging and Classifica- tion Quarterly 5:no.4:1 (Summer 1985). 11. Bernard Berelson, Graduate Education in the United States (New York: McGraw-Hill, 1960}, p.174.