Publish or practice? An examination of librarians’ contributions to research Craig Finlay Chaoqun Ni2 Andrew Tsou2 Cassidy R. Sugimoto2* 1 Indiana University South Bend craigfinlay@gmail.com 2 School of Library and Information Science Indiana University Bloomington 1320 East 10th St. Bloomington, IN 47405 USA Phone: 812.856.2323 {chni, atsou, sugimoto}@indiana.edu ABSTRACT: This article examines authorship of LIS literature in the context of practitioner and nonpractitioner production of published research. 4,827 peer-reviewed articles from 20 LIS journals published between 1956 and 2011 were examined to determine the percentage of articles written by practitioners. The study identified a decrease in the proportion of articles authored by practitioners between 2006 and 2011. Topic analysis of articles revealed subtle yet distinct differences in research subjects matter between practitioner-authored and non-practitioner authored articles. If present trends continue, the character of LIS literature may shift away from many issues relating to practical librarianship. INTRODUCTION Understanding the population demographics of publishing authors within a discipline is crucial to an understanding of the discipline itself. Documenting authorship of scholarly literature within a discipline allows researchers to assess “sources of strength in research and scholarship and the field’s pace among other disciplines.” 1 In the discipline of library and information science (LIS), there is evidence of a trend of fewer librarian-authored research articles being contributed to the literature. 2 If this trend continues, the domain of LIS, as defined by its collective scholarly output, could look very different over the next two decades. By virtue of their everyday work experience, practitioners (i.e., librarians and archivists) bring a different perspective to the literature than that of academics (i.e., professors and doctoral students). In fact, the editors of the Journal of Information Science and Technology found this dichotomy sufficiently striking that in 2004 they instituted a separate process for review of practitioner research. The new process took 1 into account the additional time required for research that investigates practice-oriented problems. Associate editor Steven John Simon wrote in an accompanying editorial that the project’s “goal is to create a forum that will be shared by ALL members of our community … It is our deepest hope that we will all learn from each other.” 3 It has been suggested that librarians need not conduct research, instead focusing their energies on performing the traditional duties ascribed to librarians. In 1979, as faculty status was increasingly common for academic librarians, Mortimer and Beck opined with Hitchensian glee that: Interminable reader-use surveys may be of considerable local interest, but do we really need to publish quite so many of them? Published questionnaires tabulating the views of other librarians on the topic of faculty status can always, of course, serve to pad our own tenure files, [though] such pieces do little to enhance our general academic standing. 4 Such a position discounts outright the possibility that conducting research and engaging in the creation of new knowledge may in fact be invaluable in a librarian’s professional development. As Swisher wrote, “research is not a process that is the responsibility of others; research is a way of knowing, a way of making better practical decisions that is the responsibility of each of us.” 5 Research questions. Whether or not one is an advocate for practitioner authorship, the importance of understanding the current state of a discipline through its contributors should remain self-evident. Toward this end, to gain a more holistic understanding of the demographic of LIS research and the implications of its evolution, this study seeks to answer the following questions: 1) How stable has the proportion of librarian authors been since 1955? This will inform our understanding of the decrease observed by Wiberly, Hurd, and Weller—specifically addressing whether this is a trend or anomaly. 2) Would a decrease in the proportion of librarian authors alter the topicality of LIS literature? If librarians are found to consistently write about topics different from those most frequently researched non-librarian authors, then we may propose that a loss of practitioner authors would impoverish the literature in that regard. 3) How well-cited are librarians as compared non-librarian authors? How frequently do librarian authors cite non-librarian authors and vice versa? Such findings would inform our knowledge of the dynamics of the LIS publishing community. LITERATURE REVIEW A number of recent studies have examined variations in the scholarly practices of practitioners as opposed to those of academics; practitioners tend to focus on practical, problem-based topics, while academics tend to focus on theory-driven, methodological studies. 6 Stolcis, when 2 comparing practitioner and academic research in law enforcement, argued that academic research tends to be methodological and data-driven, not concerned with best practice: Academics are trained to generate knowledge in their respective disciplines, not necessarily to solve organizational problems. This tendency does not lend itself to conducting research that concerns itself with practical problems. 7 This disparity in research needs and interests may help explain Hildreth and Aytac’s finding that collaboration between academics and practitioners is rare: in a survey of 206 LIS articles, collaborations between librarian and LIS instructional faculty authors accounted for less than 10% of the sample. 8 It has been noted that academic librarians primarily access scholarship pertaining to library services. 9 Haddow and Klobas wrote that such a tendency has led to practitioners being “criticized for their focus on operational or day-to-day information, and ignorance, or at best, a lack of interest in research.” 10 This may be indicative of differing priorities between the two groups, particularly in the context of a 2002 survey of librarians’ usage of LIS literature which found that “the perceived inadequacy of research to address practical workplace problems was a major reason for information professionals not consulting the research.” 11 It is not only the types of research that separate practitioners and academics, but also the venues in which they publish and the nature of citations to and references from articles in such venues. Schlogl and Stock found that practitioner-dominated LIS journals had a much lower citation half-life than their more theoretical information science counterparts, suggesting that “practitioners are mainly interested in receiving quick up-to-date information. Articles with long lists of references are usually not demanded since practitioners may not have the time to perform an extensive literature review in many cases” 12 Indeed, it was recently confirmed that practitioner-authored articles tend to have shorter reference lists than articles authored by academics. 13 If librarians tend to both author and access primarily practice-focused journal articles, then a decrease in librarian-authored literature can be expected to have important implications for the profession. In 1999, Weller, Hurd, and Wiberly examined the number of contributions by academic librarians to peer-reviewed literature between 1993 and 1997. The trio painted a rosy picture of academic librarian publishing, writing that: Academic librarians are contributing a very significant proportion of the LIS literature. The proportion of contributions by academic librarians to the scholarly literature may be increasing. Although the data are limited, it appears that academic librarians who publish do so as frequently as LIS instructional faculty in general. 14 3 Upon revisiting the subject 7 years later, the authors surveyed the relevant literature for the time period 1998-2002, finding across-the-board declines in the number and proportion of peerreviewed articles written by academic librarians, as well as a decrease in the number of academic librarian authors as a whole. 15 Specifically, the number of articles written by academic librarians was found to have declined by nearly 13%. 16 While the study was insufficiently broad to determine whether this was part of a long term trend or a temporary anomaly, the authors did determine that the drop was not due to a decline in the number of academic librarians. They also found that librarian authors were much less likely to collaborate with instructional faculty than with other librarians. One possible reason for this comes from a recent survey of Canadian librarians, which found that only 13% of the librarians considered themselves to be “active scholars.” 17 It was also found that “most universities have not provided their librarians with either formal or informal guidelines concerning an appropriate time commitment to scholarship.” 18 A pair of studies conducted in the 1970s indicated that a gap in productivity may have already been forming at that time. Massman surveyed 224 librarians and 205 instructional faculty members in three Midwestern states, comparing their publishing productivity. It was found that instructional faculty members were more productive, publishing an average of 1.7 articles per person over a 2-year period, compared to the .7 articles per librarian over the same length of time. 19 Similarly, Paula De Simone Watson studied the publication activity of librarians at 10 research universities and concluded that librarians were less likely to publish than their instructional faculty counterparts, publishing an average of less than one article per year. 20 Watson also found that the newest librarians were less likely to publish than more established librarians were, indicating a possible trend. A 2010 survey of LIS researchers indicated a marked disparity in the academic culture of the two groups; whereas 96% of full professors responded that they felt they were expected and encouraged to research and publish, only 19% of academic librarians felt the same way. 21 In addition, 54% of practitioners felt that they were neither expected nor encouraged to publish, compared with 0% of full professors. 22 It was also found that library practitioners indicated a lack of a reward system for publishing in their respective institutions. Amid this alleged limited support for librarian research production, a debate continues over whether the requirement to publish at all is fair for librarians, considering their regimented, yearround, 40-hour work week and various service requirements. 23 A number of individuals have argued that the reality of a working librarian’s schedule is not conducive to conducting research. Such constraints should not be news to any academic librarian; in 1990, Cosgriff, Kenney and McMillan surveyed 97 ARL libraries and found that while academic libraries requiring publication for granting tenure offered more opportunities for time off to conduct research, the 12-month schedule of librarians was still a hindrance to productivity. 24 Floyd and Phillips 4 conducted a survey of academic librarians and found that the most consistent worry on the part of librarians was that “the requirement to publish in order to be a successful academic often competes with the requirement to perform daily work in order to be a successful librarian.” 25 In addition, librarians consistently responded that they did not feel as if they were provided with sufficient blocks of time to conduct research. Brown advocated the use of time logs by librarians in order to allocate small blocks of time each day to do research; she suggested that librarians should be assured that “we do have time: not enough to do everything we’d like to do, but time nonetheless to allocate as we see fit.” 26 The issue of tenure, and whether librarians were experiencing difficulty gaining it, was addressed by Mitchell and Reichel, who surveyed 690 universities on the issue. At the heart of the matter is the ACRL “Joint Statement Regarding Faculty Status for Academic Librarians,” which holds that: Faculty status entails for librarians the same rights and responsibilities as for other members of the faculty. They should have corresponding entitlement to rank, promotion, tenure, compensation, leaves, and research funds. They must go through the same process of evaluation and meet the same standards as other faculty member. 27 Nevertheless, Mitchell and Reichel found that “although most librarians tend not to publish frequently, as a group, they do not have notable problems in achieving tenure.” 28 Specifically, it was found that librarians were able to gain tenure through community and library service in lieu of prolific publishing. This would seem to support Klobas and Clyde’s findings. In fact, Mitchell and Swieszkowski found that academic librarians were more likely to gain tenure than professors were. 29 Smith and DeVinney conducted a survey of 526 academic librarians at 33 universities and found that 47% were granted tenure without any publications. 30 It is the flip side of the coin described by Gillum, who suggested that one of the explanations for the decrease in librarian authorship was the decision of many universities to do away with librarian tenure all together. Gillum argued that “without the lure of promotion and tenure, there is little motivation for librarians to contribute to the body of LIS literature.” 31 Fennewald interviewed publishing academic librarians in regard to motivation; a recurring theme was that “the importance of research is to identify new knowledge that will enhance practice.” 32 It is for this same reason that some have argued for continued librarian scholarship; Stewart wrote that as the responsibilities of academic librarians have expanded to include “active engagement in teaching, content management and development, and even data curation, it has become more important for librarians to engage in scholarship.” 33 We suppose that a loss of practitioner-authored contributions to the literature will necessarily change its overall character. It has also been hypothesized that a decrease in librarian-conducted research could damage librarians’ job performance. Gillum pointed out that “conducting research 5 and scholarly writing deepens one’s knowledge of the subject matter being researched, resulting in enhanced provision of information.” 34 In short, conducting research increases knowledge in a way that simply reading cannot, and increased knowledge cannot help but translate into better practice: Research is not an activity that occurs at the fringes of the field. Rather, it is central to the continued development of library and information science as a profession or discipline represented by graduate programs within academia. 35 METHODS Selection of journals. The sampling frame for this study was the list of journals created by Nisonger and Davis 1. Nisonger and Davis surveyed deans, directors and department chairs of ALA-accredited LIS programs and directors of ARL libraries regarding the prestige and importance of certain LIS journals for promotion and tenure at their institution. 36 The results of the study provided two ranked lists of journals, one from the perspective of the deans of LIS schools and one from the directors of ARL libraries. We identified the 20 highest ranked journals from each perspective, for a list of 29 unique journals. Journals not indexed by Web of Science (WoS) were excluded, as were monograph series, leaving a total of 19 journals (see Appendix A). Selection of articles. Full bibliographic information (title, abstract, authors, cited references, etc.) was downloaded from WoS for all articles published in these 20 journals from 1956 to 2010 (or for the duration that they were in existence and indexed by ISI, if later than 1955). Name changes for all journals were included (e.g., American Documentation, Journal of the American Society for Information Science, and Journal of the American Society for Information Science and Technology were all aggregated into a single journal). Although WoS distinguishes among article types (e.g., review, editorial, research), there is some error in this classification scheme. Therefore, to further limit the study to research articles only, any article with no cited references was eliminated. Systematic sampling was employed in order to select a smaller number of articles to code. All articles from the first issue of each journal for each year were selected. Although it is possible that these could have contained special issues, there is no reason to believe that the first issue of each year is systematically different from other issues over the course of the journal. After this sampling method (and employing the exclusion criteria listed above), 4,827 articles remained for analysis. However, each of the selected journals published with different frequencies (i.e., quarterly, bi-monthly and monthly). If each journals output is evenly distributed among the 1 We used the list of journals compiled by assuming that “not familiar” and blank responses were equivalent to a “zero.” 6 issues published each year, then the first issue of a quarterly accounts for 25% of the journal’s annual output. In contrast, the first issue of a bimonthly accounts for 16.7% of the journal’s annual output, and the first issue of a monthly accounts for 8.3%. To provide reliable estimates for the set of all articles published during the period of the study, we therefore weighted the article counts for those journals that published more often than quarterly. Specifically, the values for bimonthly journals were weighted by a factor of 1.5 and the values for monthly journals were weighted by a factor of 3.0. Coding of author affiliation. Each of the 4,827 articles were examined to determine the affiliation of the authors. Coding followed a simple scheme in which each article was classified in one of three categories: 1) articles written solely by one or more individuals employed primarily in a library/archive setting; 2) articles written by at least one author employed primarily in a library/archive setting and at least one other author not employed primarily in a library/archive setting; 3) articles written solely by one or more individuals not primarily employed in a library/archive setting. Initial coding was done using the affiliation information provided in the WoS records, provided that the data included details on the level of the unit (e.g., “Harvard Univ Lib” or “Sch of Lib & Info Sci”) that made classification possible. Some journals were more likely to provide this level of detail than others were. After this initial classification, physical and digital copies of the articles were obtained and examined for affiliation data in the index or accompanying data for each article. When no information could be found in the journals, bibliographic reference books were consulted (e.g., Who’s Who in Library and Information Services). In addition, online curriculum vitae and bibliographies were used to identify a given author’s employment venue at the time of an article’s publication. Sufficient information for coding was identified for 99% of the sample (n=4,772); that is, only 55 articles could not be coded after these search strategies were employed. Affiliation analysis. Authors’ affiliations were examined to determine the contributions and collaborative behavior of librarians and non-librarians for each of the 20 journals and for each of the 12 five-year periods. This provided a detailed description of both the proportion of librariancontributed literature over time and the sites at which librarians most often chose to publish their research. In addition, simple citation data were assembled, describing mean citation counts for all articles by coded type. Weighting was employed in each of these analyses, as described above. Topic analysis. Latent Dirichlet Allocation (LDA) was originally developed as an unsupervised topic modeling technique concerning the probability distribution of keywords over topics. LDA is particularly helpful in regard to the “classification, novelty detection, summarization, and similarity and relevance judgment” of large-scale data. 37 This modeling technique is based on the assumption that each word can represent part of the semantic meaning of a document, and that no inherent ordering function exists between words and documents in a corpus. 38 Based on a hierarchical Bayesian model, LDA characterizes each topic by a probability distribution of each 7 word under each topic. For this paper, the outcome of the topic model is the probability distribution of each word in a given topic, using words from the titles of the articles. The higher the probability of a word in a topic, the higher extent that the word is representative of the topic. In essence, the topics are comprised of words and the probability value describes how closely associated the topic is with that word. Keywords were drawn from article titles and we have provided the words that are most dominant (i.e., the ten with the highest probability of occurring within the topic). Choosing the number of topics requires interplay between the output of the results and knowledge of the domain. The objective is to find topics that are exclusive and interpretable: too few topics and the organization is too general, too many and a unifying theme among the words is not easily discernible. For this analysis, we ran LDA to yield 4, 5, 6, 7, 8, 9 and 10 topics. The results with six topics were most coherent and yielded the highest face validity. Topic labels are constructed by a subjective analysis of the most dominant words. For more detailed information about LDA, please refer to Blei, Ng, and Jordan (2003). Limitations. Coding was done at a macro level—library vs. non-library and collaborations between library and non-library individuals. The types of libraries and the affiliations of those who were not in a library were not investigated. It is fair to mention that it is not uncommon for librarians to teach courses in a library school setting, in addition to their other duties. While this is true, it may be assumed that, being employed by a library, such an individual would still be subject to the time demands of other librarians. While perusal of the data suggests that most nonpractitioners were academics (instructional faculty and doctoral students in LIS programs), these data were not systematically gathered as the paper wanted to focus on the two binary categories of practitioner and non-practitioner. RESULTS Authorship trends. Non-librarian authored journal articles comprised 67% of the sampled journal articles, while librarian-authored journal articles made up 31% (Table 1). Please note that the number of publications and citations reported here is counted with weights based on journal publishing frequency. The same statistics of journals without weighting by publishing frequency can be found in Appendix B. Fewer than 3% of articles were written collaboratively between librarian and non-librarian authors. Articles written by non-librarians were found to have been cited 9.4 times on average. This is nearly three times as often as articles authored solely by librarians, which averaged 3.5 citations per article. Collaborative articles between librarian and non-librarian authors were cited twice as often than those articles written solely by librarians. The citations to papers of type 1 and type 3 were significantly different at the .01 level. All 20 journals featured articles written solely by librarians, as well as articles authored solely by nonlibrarians. Four journals did not have an instance of collaboratively written articles (Libraries & Culture; Library Collections, Acquisitions, & Technical Services; Journal of the Medical Library Association; and Journal of Information Science). 8 Table 1: Distribution and mean citation of articles by type (weighted by publishing frequency). Description Type 1 Librarian Type 2 Collaborative Type 3 Non-librarian Note: weighted result #Articles 2077.5 156 4427.5 % of total 31.2% 2.3% 66.5% #Total Cites 7334 1132 41485 Mean Cites/article 3.5 7.3 9.4 #journals 20 16 20 A decrease in the number of overall journal articles and proportion of the overall literature authored by librarians (Type 1) was observed between 2002 and 2011 (Figure 1). Between 2002 and 2006, the number of articles authored by librarians decreased by 10%. During this period, the number of journal articles authored by non-librarians increased by 20%. The number of collaboratively authored journal articles nearly doubled, from 26 to 47. Proportionally, the percentage of overall articles authored by librarians was found to have decreased by just over 7% during the same period, from 31% to 24%. The distribution of corresponding increase for type 2 and 3 articles was 1.5% and 5.4%, respectively. Figure 1. Number of articles by author classification, 1956-2011 (weighted by publishing frequency) 9 The data were also examined according to individual journals, revealing a varied spectrum of author demographics within LIS (Figure 3). Of the 20 journals selected, 7 could be considered librarian journals, with more than 50% of the articles written by librarians (Table 2). Library Trends had the most balanced distribution, with 246 Type 1 (librarian) articles, 249 Type 3 (nonlibrarian) articles, and 11 Type 2 (collaborative) articles. The journals with the largest proportion of non-librarian authors were seemingly less diverse than those with the largest proportion of librarian authors (Table 2). Four journals (Information Processing and Management, JASIST, Journal of the American Medical Informatics Association, and the Journal of Information Science) contained less than 10% of Type 1 (librarian) articles. Table 2: Top six journals for highest percentage librarian and highest percentage nonlibrarian authors. Librarian-focused journals CRL LRTS JALib Portal ITLib RUSQ Percentage of librarian authors 71.6 69.7 69.0 67.5 58.6 54.6 Non-librarian-focused journals IPM JASIST JAMIA JIS LISR JDOC Percentage non-librarian 96.2 94.4 92.1 91.4 88.3 81.3 The number of articles by each author type, by journal, is provided in Figure 2. 10 Figure 2: Authorship composition of all journals (weighted by publishing frequency). Topical trends. Epistemic communities are defined by a degree of topical coherence. Therefore, we wanted to examine not only whether journals varied by author type, but whether the topics varied by author type. To this end, we applied LDA to analyze the topics of the Type 1 and Type 3 articles (Type 2 articles were not included in this analysis due to the small number). The results provide six topics that define the literature for each analyzed type. Topics are defined by words that frequently occur in each area. Table 3 provides the six dominant topics generated for the librarian-authored papers, labeled as (1) government information, (2) reference and information services, (3) collections and publishing, (4) medical librarianship, (5) technical services, and (6) academic libraries and librarianship. Table 3: LDA results for librarian-authored journal articles. Government information Word Load Reference and information services Word Load Collections and publishing Word Load Medical librarianship Word Load Technical services Word Load Academic libraries and librarianship Word Load information 0.0232 library 0.0201 library 0.0371 library 0.0209 catalog 0.0228 academic 0.0321 libraries 0.0190 reference 0.0193 information 0.0098 university 0.0136 collection 0.0201 library 0.0242 national 0.0121 information 0.0148 public 0.0085 health 0.0136 library 0.0201 reference 0.0193 public 0.0053 online 0.0143 book 0.0085 academic 0.0131 online 0.0143 0.0148 scientific 0.0048 web 0.0089 services 0.0085 electronic 0.0131 access 0.0080 informati on university government 0.0048 literacy 0.0084 develop 0.0078 information 0.0115 digital 0.0080 faculty 0.0089 0.0094 11 access 0.0048 service 0.0084 collections 0.0075 medical 0.0115 book 0.0074 web 0.0089 policy 0.0048 system 0.0074 publishing 0.0072 state 0.0094 0.0059 literacy 0.0084 science 0.0048 catalog 0.0074 trends 0.0059 impact 0.0089 preservatio n oclc 0.0059 service 0.0084 federal 0.0043 instruction 0.0074 collection 0.0052 case 0.0079 academic 0.0059 librarian 0.0084 Table 4 presents the results of the dominant topics for non-librarian-authored articles. The topics for non-librarian-authored papers may generally be categorized as (1) information retrieval, (2) academic libraries, (3) health informatics, (4) government information policy, (5) public libraries and librarianship, and (6) and information seeking and behavior. Table 4: LDA results for non-librarian-authored journal articles. Information retrieval Word Load information 0.040 3 0.038 5 0.010 3 0.009 6 0.008 6 0.007 9 0.007 7 0.007 2 0.006 7 0.006 5 retrieval systems search citation scientific document text approach indexing Academic libraries Word library academic education online information study collection developmen t survey evaluation Load 0.025 1 0.014 6 0.012 6 0.008 3 0.008 3 0.007 4 0.007 4 0.007 2 0.006 9 0.006 3 Health informatics Word health informatio n data medical clinical system impact electronic case model Load 0.019 4 0.013 7 0.012 1 0.012 1 0.010 8 0.009 9 0.008 3 0.007 6 0.006 1 0.005 7 Govt. information policy Word Load Public libraries Word information 0.0558 library web 0.0094 information management 0.0089 research government 0.0086 services analysis 0.0083 public policy 0.0077 librarian technology 0.0077 publishing public 0.0074 digital 0.0071 preservatio n media internet 0.0071 book Load 0.036 6 0.015 8 0.012 4 0.007 9 0.007 6 0.006 8 0.006 2 0.005 9 0.005 4 0.005 4 Info. seeking and behavior Word Load information 0.0370 research 0.0295 library 0.0154 science 0.0065 social 0.0062 knowledge 0.0059 factors 0.0059 seeking 0.0059 community 0.0056 case 0.0051 Type 1 and type 3 papers share a focus on government information policy and academic librarianship, and both categories contain a topic with an evident focus on medicine. Librarianauthored journal articles feature a greater emphasis on library services, while non-librarianauthored journal articles display a focus on information seeking, use, retrieval, and informatics. DISCUSSION 12 The decrease in the number of librarian-authored journal articles is notable. The long-term data show that while there were slight decreases in the number of type 1 articles from 1962 to 1976, and again from 1987 to 1996, the data showed a gradual and consistent increase in librarian production of articles. The single largest increase in the number of librarians-authored journal articles (1997 to 2006) was followed by a sharp decline in the following 5 years. The number of type 3 articles has shown a steady increase of approximately 25% per 5-year period for the last 20 years, outpacing the longitudinal increase in Type 1 articles. Prior to the recent decrease, the percentage of the literature contributed by librarians had remained rather steady, around 35%, since 1972. The lone exception was a brief decrease between the 5-year periods beginning in 1992 and 1997, and that dip decrease followed by an equal increase in the following time period. Interestingly, this increase corresponds with the time period during which Wiberly, Hurd, and Weller documented a decline in the proportion of journal articles written by academic librarians. 39 The recent decrease in librarian-authored articles may simply be an anomaly, to be reversed over time. However, it may also suggest systematic changes in the scholarly communication practices within the field. A possible explanation for such a migration is that the advent of blogging and social media has connected librarians and opened up new channels for the dissemination of service-oriented literature. Such conduits allow librarians to circumvent the peer-review process and the delay between submission and publication that it demands. In noting the explosive growth of library blogs, Gilman quoted a librarian blogger who wrote that: Libraries are largely dependent on and are competing with technologies that change every nine months. How are we supposed to progress as a profession when it still takes a year and a half for an article to progress from submission to publication? 40 Recall the studies by Mitchell and Reichel and Mitchel and Swieszkowski which found that librarians are gaining tenure without publishing, whether through substituting service or by having the publication requirement waived entirely. If librarians do not feel the same pressure to publish as instructional faculty do, and if librarians are indeed migrating to blogging, it might explain the decrease in authorship documented by Wiberly, Hurd, and Weller. Indeed, the rapid growth of the blogosphere after 2001 has been documented and falls within the timeframe of Wiberly, Hurd, and Weller’s second study. 41 Aharony supports this, with evidence that librarians tend to write essay-like posts, and that their blog tags fit into larger categorization systems, according to subject. 42 In other words, librarian bloggers are not simply keeping personal blogs, but rather are writing about specific research topics. It is also interesting to note that blogs created by librarians independently of their parent institution tend to cluster together in a community of interlinked blogs, while blogs published by academic institutions tend to be more isolated. 43 13 While the extent of a hypothesized migration to blogging is as yet undetermined, there is already a divide among librarians in regard to whether blogging should count toward tenure, as well as to whether blogging should be considered publishing or a public service. 44 This divide seems to align along an age gap, with younger librarians responding positively that blogs should count towards publishing requirements for tenure, and more senior librarians rejecting the idea. Nevertheless, relatively few librarians are willing to publish in non-traditional outlets, given the difficulty of establishing a robust tenure portfolio without publications in peer-reviewed journals and monographs. 45 In any case, if present trends continue, it may be expected that librarians will contribute less than 20% of the scholarly literature in LIS by the end of this decade. It is also interesting to note that there is some truth to “assertions that libraries are no longer the primary research focus at the doctoral level in LIS,” suggesting that library school curricula may also contribute to this shift in focus. 46 Further, faculty status of librarians is currently in a state of flux, with some universities choosing to strip librarians of faculty status all together. 47 If this becomes a trend, expect to see librarian contributions to the literature to further decrease. This study provides further support of a divide in journal readership and authorship along professional lines. Powell, Baker, and Mika found that, among American Library Association members, the most read journals were College & Research Libraries, Public Library, Reference & User Services Quarterly, Library Resources & Technical Services, and Information Technology & Libraries. 48 Excluding Public Libraries, which was excluded from that study, all of these journals were shown to be predominantly authored by librarians. It was also found that, among members of the American Society for Information Science & Technology (ASIST), the most read journals were the Journal of the American Society for Information Science & Technology (ASIST), Information Outlook, College & Research Libraries, Information Technology & Libraries, and Journal of Academic Librarianship. Of these, with the exception of IO (which was not included in this study), only JASIST was shown to be predominantly authored by non-librarians. Of further interest in this regard is the documentation of markedly lower citation counts associated with librarian-authored journal articles. This is likely due, in part, to the shorter references lists present in librarian-authored research. 49 A factor of concern to the study was whether the loss of librarian-authored peer-reviewed research would noticeably alter content of the literature. The results of the LDA analysis yield an uncertain response. At the level of six topics, three were found to be similar, and three were found to be non-similar. The differences may be described as indicative of a library science/information science divide. Whereas type 1 articles displayed a focus on government scientific policy and science librarianship, the type 3 articles displayed a focus on government information policy—similar topic areas, but different perspectives on these areas. Similarly, a focus on medical librarianship on the part of librarians was mirrored in the non-librarian 14 authored research as a focus on health informatics. Finally, a focus on online public access catalogs has a sibling topic of information retrieval. The other three topics were not quite as similar, leading to the conclusion that a loss of librarian authors would in fact alter the content of the literature. These findings are reinforced by other recent bibliometric analyses, demonstrating the decline in library-related topics in the literature. 50 CONCLUSION A decline in librarian-authored research, coupled with significant differences in the topic of librarian-authored research suggests that the character of LIS literature is likely to change in the near future if present decline in librarian-authored research continues. The apparent disengagement of librarians from the traditional channels of scholarly communication will necessarily decrease librarians’ familiarity with scholarly communication, and this in turn may affect how librarians, especially those employed at academic institutions, interact with students and academics who are conducting research. We may assume that this decrease will engender a shift in topicality away from library services. This will no doubt have important pedagogical implications for students and academics alike in the nation’s MLS programs, as students may find themselves as charges in the care of academics whose collective research interests infrequently extend to the very field they are seeking to enter. Of course, whether the current trend is destined to continue, no one can yet tell. Those who believe that librarianship benefits from published research conducted by its practitioners will no doubt hope that it does not. 15 ACKNOWLEDGEMENT. The authors would like to thank Cody Behles for his work on previous versions of this manuscript. APPENDIX Appendix A. Journals selected from Nisonger and Davis study (listed alphabetically) with publication frequency Appendix B. Unweighted results. 1 Steven E. Wiberley, Julie M. Hurd, and Ann C. Weller, “Publication Patterns of U.S. Academic Librarians from 1998 to 2002,” College & Research Libraries 67(2006): 205. 2 Wiberley, Hurd, and Weller, “Publication Patterns.” 3 Steven John Simon, “Rigor Vs. Relevance: Why Can’t We All Just Get Along?” Journal of Information Science and Technology 1, no. 1 (2004): 10. 4 Roger Mortimer and Nelson Beck, “The Librarian’s Role – Two Views,” The Journal of Academic Librarianship, 4 (1979): 448. 5 Robert Swisher, “Focus on Research,” Top of the News, 42 (1986): 176. 6 Gaby Haddow and Jane E. Klobas, “Communication of Research to Practice in Library and Information Science: Closing the Gap,” Library and Information Science Research 26, no. 1 (2004): 29-43; Bill Crowley, Spanning the Theory-Practice Divide in Library & Information Science (Scarecrow Press, 2005); Christian Schlogl and Wolfgang G. Stock, “Practitioners and Academics as Authors and Readers: The Case of LIS Journals,” Journal of Documentation 64, no. 5 (2008): 643-666. 7 Michael J. Bolton and Gregory B. Stolcis, “Ties That Do Not Bind: Musings on the Specious Relevance of Academic Research,” Public Administration Review 63, no. 5 (2003): 627. 8 Charles Hildreth and Selenay Aytac, “Recent Library Practitioner Research: A Methodological Analysis and Critique,” Journal of Education for Library and Information Science 48, no. 3 (2007): 236-258. 9 Dale Montanelli and Collette Mak, “Library Practitioners' Use of Library Literature,” Library Trends 36, no. 4 (1988): 765-784; Kathlyn Turner, “Do Information Professionals Use Research Published in LIS Journals?” (paper presented at the annual meeting of the international Federation of Library Associations (IFLA), Glasgow, Scotland, August 18-24, 2002); Haddow and Klobas, “Communication of Research.” 10 Haddow and Klobas, “Communication of Research,” 30. 11 Kathlyn J. Turner, “Do information professionals use research published in LIS journals?” (paper presented at the 68th IFLA Council and General Conference, Glasgow, August 18–24, 2002: 10. 12 Schlogl & Stock, “Practitioners and Academics,” 661. 13 Jingfeng Xia, Sara Kay Wilhoite, and Rebekah Lynette Myers, “A ‘Librarian-LIS-Faculty’ Divide in Open Access practice,” Journal of Documentation, 67, no. 5 (2010): 791-805. 14 Ann Weller, Julie Hurd, and Stephen Wiberley, “Publication Patterns of U.S. Academic Librarians From 1993 to 1997,” College & Research Libraries 60 (1999): 359. 15 Wiberley, Hurd, and Weller, “Publication Patterns.” 16 Wiberley, Hurd, and Weller, “Publication Patterns,” 205. 17 David Fox, “A Demographic and Career Profile of Canadian Research University Libraries,” Journal of Academic Librarianship 33, no. 5 (2007): 547. 16 18 Ibid Virgil Massman, Faculty Status for Librarians (Metuchen, N. J : The Scarecrow Press, 1972). 20 Paula De Simone Watson, “Publication Activity Among Academic Librarians,” College & Research Libraries 38, no. 5 (1977): 375-384. 21 Jane E. Klobas and Laurel A. Clyde, “Beliefs, Attitudes and Perceptions About Research and Practice in a Professional Field,” Library & Information Science Research 32, no. 4 (2010): 237-45. 22 Klobas and Clyde, “Beliefs,” 248. 23 William K. Black and Joan M. Leysen, “Scholarship and the Academic Librarian,” College and Research Libraries 55 (1994): 229-41; Kathleen Kenny, Linda D. Tietjen, and Rutherford W. Witthus, “Increasing Scholarly Productivity Among Library Faculty: Strategies for a Medium-sized Library at the University of Colorado,” The Journal of Academic Librarianship 16 (1990): 276-9. 24 John Cosgriff, Donald Kenney, and Gail McMillan, “Support for Publishing at Academic Libraries: How Much Exists?” The Journal of Academic Librarianship 16, no. 2 (1990): 94-99. 25 Barbara Floyd and John Phillips, “A Question of Quality: How Authors and Editors Perceive Library Literature,” College and Research Libraries 5, no. 1(1997): 81. 26 Jeanne Brown, “Time and the Academic Librarian,” portal: Libraries and the Academy 1, no. 1 (2001): 6. 27 ACRL, “Joint Statement on Faculty Status of College and University Librarians” (1972). http://www.ala.org/acrl/standards/jointstatementfaculty 28 W. Bede Mitchell and Mary Reichel, “Publish or Perish: A Dilemma For Academic Librarians?” College and Research Libraries 60, no. 3 (1999): 238. 29 W. Bede Mitchell and L Stanislava Swieszkowski, “Publication Requirements and Tenure Approval.Rates: An Issue for Academic Librarians,” College and Research Libraries 46, no. 2 (1985): 249-255. 30 Karen F. Smith and Gemma DeVinney, “Peer Review for Academic Librarians,” Journal of Academic Librarianship 10 (1984): 87–91. 31 Shalu Gillum, “The True Benefit of Faculty Status for Academic Reference Librarians,” The Reference Librarian 51, no. 5 (2007): 323 32 Joseph Fennewald, “Research Productivity Among Librarians: Factors Leading to Publications at Penn State,” College & Research Libraries 69, no. 2 (2008): 107. 33 Christopher Stewart, “Whither Metrics? Tools for Assessing Publication Impact of Academic Library Practitioners,” The Journal of Academic Librarianship 36, no. 5 (2010): 449. 34 Gillum, “True Benefit,” 311. 35 Peter Hernon and Candy Schwartz, “Library and Information Science Research: What Do We Need?” Library & Information Science Research 15 (1993): 116. 36 Thomas Nisonger and Charles Davis, “The Perception of Library and Information Science Journals by LIS Education Deans and ARL Library Directors: A Replication of the Kohl– Davis Study,” College & Research Libraries 66 (2005): 341–377. 37 David Blei, Andrew Ng, and Michael Jordan, “Latent Dirichlet Allocation,” Journal of Machine Learning Research 3 (2003): 993. 38 Ibid. 39 Wiberley, Hurd, and Weller, “Publication Patterns.” 40 Isaac Gilman, “We’re Content Creators, Too: Libraries and Blogging,” Oregon Library Assocaition Quarterly 12, no. 1 (2008): 15. 41 Ravi Kumar, Jasmine Novak, Prabhakar Raghavan, and Andrew Tomkins, “On the Bursty Eveolution of Blogspace,” World Wide Web 8, no. 2 (2005): 159-178. 42 Noa Aharony, “Librarians and Information Scientists in the Blogosphere: An Exploratory Analysis,” Library & Information Science Research 31 (2009): 174-181. 43 S. Craig Finlay, Carolyn Hank, Cassidy R. Sugimoto, and Michael Johnson, “The Structure of the Biblioblogosphere: An Examination of the Linking Practices of Institutional and Personal Library Blogs,” The Journal of Web Librarianship (accepted). 44 Arthur Hendricks, “Bloggership, or is Publishing a Blog Scholarship? A Survey of Academic Librarians,” Library Hi Tech 28, no. 3 (2010): 470-474. 19 17 45 Cassidy R. Sugimoto, Andrew Tsou, Sara Naslund, Alexandra Hauser, Melissa Brandon, Dnaielle Winter, Cody Behles, and S. Craig Finlay, “Beyond Gatekeepers of Knowledge: Scholarly Communication Practices of Academic Librarians and Archivists at ARL Institutions,” College & Research Libraries (accepted). 46 Craig S. Finlay, Cassidy R. Sugimoto, Daifeng Li, and Terrell G. Russell, “LIS Dissertation Titles and Abstracts (1930-2009): Where Have All the Librar* Gone?” The Library Quarterly 82, no. 1 (2012): 29-46; Cassidy R. Sugimoto, Daifeng Li, Terrell G. Russell, S. Craig Finlay, and Ying Ding, “The Shifting Sands of Disciplinary Development: Analyzing North American Library and Information Science (LIS) Dissertations Using Latent Dirichlet Allocation (LDA),” Journal of the American Society for Information Science & Technology 62, no. 1 (2011): 185-204. 48 Sydni Dunn, “As Their Roles Change, Some Librarians Lose Faculty Status,” Chronice of higher Education 59, no. 28 (2013): 6 48 Ronald Powell, Lynda Baker, and Joseph Mika, “Library and Information Science Practitioners and Research,” Library and Information Science Research 24, no. 1 (2002): 49-72. 49 Schlogl & Stock, “Practitioners and Academics”; Xia, Wilhoite, and Myers, “A ‘Librarian-LIS-Faculty’ Divide.” 50 Cassidy R. Sugimoto, Daifeng Li, Terrell G. Russell, S. Craig Finlay, and Ying Ding, “The Shifting Sands of Disciplinary Development: Analyzing North American Library and Information Science (LIS) Dissertations Using Latent Dirichlet Allocation (LDA),” Journal of the American Society for Information Science & Technology 62, no. 1 (2011): 185-204; Stasa Milojevic, Cassidy R. Sugimoto, Erjia Yan, and Ying Ding, “The cognitive structure of library and information science: Analysis of article title words. Journal of the American Society for Information Science & Technology, 62(10), 1933-1953; Craig S. Finlay, Cassidy R. Sugimoto, Daifeng Li, and Terrell G. Russell, “LIS Dissertation Titles and Abstracts (1930-2009): Where Have All the Librar* Gone?” The Library Quarterly 82, no. 1 (2012): 29-46. Journal name ASLIB Proceedings College & Research Libraries 16 1 LIS Deans rank 21 11 Government Information Quarterly Information Processing & Management 14 41 25 7 quarterly bi-monthly Information Technology and Libraries 10 25 quarterly Journal of Academic Librarianship Journal of Documentation 3 20 7 5 bi-monthly bi-monthly Journal of Information Science 26 17 montly. Journal of the American Medical Informatics Association Journal of the American Society for Information Science and Technology 44 19 bi-monthly 7 1 montly Journal of the Medical Library Association Libraries & Culture 15 24 14 13 quarterly quartlery Library & Information Science Research 20 3 quarterly 9 39 quarterly 4 1 quarterly Library Collections, Services Library Quarterly ARL Directors rank Acquisitions, & Technical Publishing frequency bi-monthly bi-monthly 18 Library Resources & Technical Services 6 15 quarterly Library Trends 2 6 quarterly 17 n/a 17 n/a quarterly quarterly 5 10 quarterly Libri Portal: Libraries and the Academy Reference & User Services Quarterly Appendix B Distribution and mean citation of articles by type (without weights) Code Type 1 Description Librarian #Articles 1788 % of total 37.5% #Total Cites 5722 Mean Cites/article 3.2 #journals 20 Type 2 Type 3 Collaborative Non-librarian 126 2858 2.6% 59.9% 636 22975 5.0 8.0 16 20 Figure 1. Number of articles by author classification, 1956-2011(without weights) 19 Figure 2: Authorship composition of all journals (without weights) 20