Abstract

Objective – To determine whether library and information science (LIS) doctoral research at North American institutions has, over the last eighty or so years, displayed a clear trend toward addressing topics other than those associated with librarianship and traditional library functions; and whether one can discern, in this regard, any significant differences among those institutions.

Design – Conceptual content analysis of dissertation titles and abstracts.

Setting – North American universities with American Library Association accredited LIS programs in the period 1930 through 2009.

Subjects – The titles and, to the extent available, the abstracts of 3,230 LIS doctoral dissertations completed at these institutions during this period.

Methods – Having opted for a directed, single-category type content analysis, the researchers began by pre-establishing a group of terms which they assumed could “represent the core curriculum of the master’s in library science”: terms which they surmised would therefore be able to function, where they appeared in “the records of doctoral output”, as good indicators that that output itself can rightly be judged to have had “an explicit focus on libraries/librarianship” (pp. 36, 44). The terms selected were: “librar*”, “catalog*”, “circulat*”, “collection develop*”, “collection manag*”, “school media”, and “reference” (where “*” indicates truncation, and that any term beginning with the respective letter string was acceptable).

The researchers then simply tallied for each of the 3,230 dissertations under investigation how many times one or more of the pre-chosen terms occurred in its title and in its abstract, not recording which term or terms that occurred. (They do not make entirely clear to what extent data collection was computerized.) They subsequently analyzed the data longitudinally and by institution, with only one, nominal and dichotomous, variable for the title as well as for the abstract: whether or not any of the pre-chosen terms occurred at least once. Multiple occurrences, whether of the same term or of varying terms, played no role.

Their analysis for the entire period of 1930 through 2009 was based on title data only, and did not take doctorate-granting institution into account. The separate analysis (N=2,305) for the period 1980 through 2009 excluded the thirty cases in which one or more of the terms occurred in the title but none of them occurred in the abstract.

Main Results – One occurrence of any of the specified terms in the title was, for the overall period of 1930-2009, enough for any given dissertation to be qualified as having an explicit focus on libraries/librarianship. The percentage of such dissertations remained fairly stable from the 1930s through the 1980s, at between 56% and 62%, with the exception of an unexplained dip for the 1950s to 44.1%. Then, for the 1990s, the researchers discovered a fall-off from 57.9% to 36.0%, and in the following decade a further decrease, down to a level of 21.5%.

During the separately-analyzed period 1980-2009, the percentage of dissertations with at least one of the specified terms in the title as well as in the abstract diminished steadily from well over half (58.4%) for 1980-1984 to less than 1 in 5 (19.8%) for 2005-2009. A chi-square test revealed that the relationship between year of dissertation and term occurrence is statistically significant. By far the greatest decrease, of 15 percentage points, was that between the first half and the second half of the 1990s. Interestingly, during the whole thirty-year period, the percentage where a term appeared not in the title but only in the abstract remained fairly constant, at around 20%, give or take about 2.5 percentage points. Yet when one looks at how many of the dissertations displayed none of the terms in the title and none in the abstract, one sees a continuous increase starting at 20.7% for 1980-1984 all the way up to 61.0% for 2005-2009, with the sharpest climb, of more than 17 percentage points, occurring around the mid-1990s. The distinction between the year 1980 and the year 2009 is even greater: from just over 1 in 7 (14.7%) to more than 3 out of 5 (62.2%).

The analysis by institution revealed a statistically significant relationship for the period 1980-2009 between institution at which the dissertation was written and the occurrence of any of the terms at least once in both title and abstract. Certain institutions (most notably SUNY-Albany, Syracuse, Missouri, Hawaii, Montréal, and Long Island) showed a much higher than average overall level of no occurrence, and some (Michigan in particular, but also, for example, Florida State and the University of North Carolina) displayed a remarkably consistent decline in occurrence.

Conclusion – The researchers conclude that their study, insofar as North America is concerned, “has provided empirical evidence for . . . the lessening focus in LIS dissertations on topics commonly associated with librarianship” and that it “supports the assertion that this focus varies significantly between schools—with some schools demonstrating a more explicit connection to library-related topics than other schools” (p. 43). They are unsure how best to interpret these findings or how they could be applied, but they do offer certain suggestions for possible interpretations and pose a few questions regarding what those interpretations might imply (p. 44). One could, they suggest, argue that the terms employed in the study “are themselves antiquated, and dissertations are charting new territory, pushing the boundaries of both research and practice.” Another possibility is that “while the dissertations may not be immediately applied work, the work could be utilized for application.” On the other hand, it may simply be the case that the selected terms indeed remain trustworthy indicators, and that doctoral candidates “are no longer studying topics that are relevant to the practical field” of librarianship. One could perhaps even justifiably assert that LIS is in effect no longer a single unified discipline but, rather, has split into a library field and an information field, whereby the latter has been steadily gaining the upper hand in LIS programs, albeit less so at some institutions than at others.

In pondering the above alternatives for interpretation of this study’s results, the library practitioner will probably also be inclined to reflect, the authors suggest, on the prospects for adequate academic research support of actual library practice, while keeping in mind, furthermore, that the formal education of future practitioners will largely remain in the hands of those trained as LIS doctoral students. To what extent will these educators feel an affinity with, and possess an understanding of, the world of practical librarianship?

Commentary

This article is the most recent contribution to a literature, extending back roughly half a century but of very modest extent, dealing with the topical orientation of North American LIS dissertations. The study’s importance and uniqueness lie in the attempt to reveal significant trend data over an exceptionally long period, indeed starting with the year in which the first LIS doctoral degree was granted, to less than three years ago. The data offered are nevertheless less ample and expressive, and their analysis less refined, than in the cases of those previous, diachronically more restricted studies.

There is to my knowledge no critical appraisal tool available that has been expressly designed for the evaluation of content analysis studies, in LIS or elsewhere. However, generic checklists, and even many of those created for other fields or for other study types, can certainly be of use. One major stipulation in many such checklists is that a research study publication should begin by unambiguously indicating what the reason for the study was: why was it carried out, and why is that important? (e.g., Connaway & Powell, 2010, pp. 310, 314-317). Ideally, this research purpose, along with a clear “analytic story” and an evident “underlying logic”, should inform the entire research presentation, through to the discussion and the conclusions (Treloar, Champness, Simpson, & Higginbotham, 2000), and it should furthermore be sufficiently placed in the context of related previous scholarship (Connaway & Powell, 2010, pp. 314-317). The researchers should be sure to “distinguish adequately between the problem and the purpose of the study”, that is, between what was studied, and why (p. 316). The article under review unfortunately fails to meet these basic requirements. In their introduction, and throughout the greater part of the text, the authors repeatedly suggest that the aim of the research was to investigate the evolution and boundaries of LIS’s disciplinary identity, and to clarify its present status. However, the specific conclusions that they ultimately draw imply, very strongly, that their principal goal was in fact to establish that there is a growing disconnect between doctoral research and the actual concerns of library practice. In their final paragraph, nonetheless, they play down any question of a gap between research and practice and again emphasize the issue of how “to more precisely map the evolution of our discipline” (p. 44). Thus, they leave the reader in complete uncertainty as to the actual purpose and import of their study.

The reliability of the study design is open to little question. Its validity, however, is quite another matter. How appropriate and trustworthy should we consider a measurement approach that depends on, respectively, the minimal and isolated occurrence, or the absence, of an arbitrarily limited number of words or two-word combinations in document surrogates, without regard to syntax or context or to frequency of occurrence, for peremptorily establishing whether the documents themselves do or do not qualify as dealing significantly with a topic of interest to a given field of professional activity? (Janis, 1949; Krippendorff, 2004, pp. 22-25, 213) And that while at least some of those terms in themselves can have (widely) varying meanings and usages in the language, by no means limited to the domain of librarianship? Face validity (Krippendorff, 2004, pp. 313-314) is, therefore, an obvious concern here. What lies behind that is a fundamental problem of empirical validity: in this case, more specifically, of semantic validity (pp. 319, 323-324), one that renders the findings considerably less illuminating and useful than they otherwise could have been. Remarkable is that the researchers apparently neglected to subject their method to any kind of validity test. Krippendorff (1980) already long ago suggested ways in which content analysts can carry out semantic validation of their research designs even without expending the extra time and costs required for pilots or for pre- or post-testing. And had the researchers indeed chosen not a conceptual, but for example a relational or a contrastive or a contextual variety of directed content analysis, their results would surely have proven considerably more enlightening.

The inclusion of additional terms could have broadened their results to encompass further important aspects of contemporary library practice, such as instruction, liaison, management of electronic resources, scholarly communication issues, special collections, etc. The original researchers do admit (p. 43) that they were using a “blunt instrument”, but apparently were content with achieving, rather easily and quickly in this way, at least a modicum of empirical evidence for what they saw as the broadly shared anecdotal impression that North American LIS education at its highest level has in the last decade and a half been more and more abandoning its interest in traditional library functions and, possibly, in “librarianship” as a professional or occupational identity.

Whether this abandonment should be seen as a bad thing, or indeed as a good thing, for LIS practitioners will depend entirely on how one feels about, and perhaps on how dependent one perceives oneself to be on, that occupational identity and that traditional library world. In any case, what would now be particularly useful, and what this study unfortunately did not even attempt to provide, is some good trend data not on what LIS doctoral research is focusing less and less on, but on what things it is in fact focusing more and more on. Data that preferably go this time beyond just North America. Only then shall we be in a position to decide whether there exists a significant gap between the kinds of research that contemporary LIS practice could benefit from, and the kinds of research that doctoral students, and therefore our future LIS educators, are actually carrying out.

References

Connaway, L. S., & Powell, R. R. (2010). Basic research methods for librarians (5^th ed.). Santa Barbara, CA: Libraries Unlimited.

Janis, I. L. (1949). The problem of validating content analysis. In H. D. Lasswell, N. Leites, and Associates, Language of politics: Studies in quantitative semantics (pp. 55-82). New York: Stewart.

Krippendorff, K. (1980). Validity in content analysis. In E. Mochmann (Ed.), Computerstrategien für die Kommunikationsanalyse (pp. 69-112). Frankfurt/Main; New York: Campus.

Krippendorff, K. (2004). Content analysis: An introduction to its methodology (2^nd ed.). Thousand Oaks, CA: Sage.

Treloar, C., Champness, S., Simpson, P. L., & Higginbotham, N. (2000). Critical appraisal checklist for qualitative research studies. Indian Journal of Pediatrics, 67(5), 347-351. doi: 10.1007/BF02820685