Dissertation-to-Book Publication Patterns among a Sample of R1 Institutions

INTRODUCTION A common concern about openly available electronic theses and dissertations is that their “openness” will prevent graduate student authors from publishing their work commercially in the future. A handful of studies have explored aspects of this topic; this study reviewed dissertation-to-book publication patterns at Carnegie Classification R1 academic institutions. METHODS This study analyzed over 23,000 dissertations from twelve U.S. universities to determine how frequently dissertations were subsequently published as books matching the original dissertation in pagination, chapters, and subject matter. WorldCat and several other resources were used to make publication determinations. RESULTS Across the sample set, a very small percentage of dissertations were published as books that matched the original dissertation on pagination, chapters, and subject matter. The average number of years for dissertations in the study to be published as books was determined for broad subject categories and for select academic disciplines. Results were compared across public and private institutions, and books that were self-published or published by questionable organizations were identified. DISCUSSION Dissertation-to-book trends occur primarily in the social sciences, humanities, and arts. With dissertations for which the author is actively working to publish as a book, the commonly offered 6- to 24-month embargo periods appear sufficient, provided that extensions or renewals continue to be available. CONCLUSION This study has implications for librarians providing services to graduate students, faculty advisors, and graduate colleges/schools in regard to dissertation embargo lengths, self-publishing, and what we have termed questionable publishers, as these areas continue to provide opportunities for librarians to educate these stakeholders.


INTRODUCTION
Electronic theses and dissertations (ETDs) have been in existence for more than a generation. The concept was proposed in 1987, and Virginia Tech developed supporting software in the 1990s (Networked Digital Library of Theses and Dissertations [NDLTD], n.d., para 4). Concerns about openly available ETDs-ETDs that are available worldwide through open access instead of behind a paywall-began to surface shortly after ETDs became available, and concerns continue to persist to this day, despite many academic institutions requiring openly available ETDs (Salo, n.d.). One commonly voiced concern is the potentially diminished future publishing opportunities for graduate students who make their thesis or dissertation openly accessible. Similar to others (Ramirez, Dalton, McMillan, Read, & Seamans, 2013), we also had evidence from our own institution that graduate student authors were restricting access to their ETDs and that university faculty were advising students to restrict access due to concerns about publishers' perceptions about prior publication. The University of Oklahoma (OU) began depositing ETDs into its institutional repository in the fall of 2013 and almost immediately received pushback from the university community. This pushback took the form of discussions at faculty senate, in meetings with the graduate college, and in conversations with department liaison librarians. Concerns centered around the fear of diminished publishing opportunities for graduate students if their ETDs were openly available.
Our study sought to determine how frequently dissertations written between 2000 and 2012 were published as monographs, thereby addressing the "diminished opportunities for publishing" concerns using data instead of anecdotal evidence. We also tried to discern whether there had been a shift in the percentage of dissertations that had been pub-lished as monographs when open ETD policies went into effect at Carnegie Classification R1 academic institutions in the United States. We hoped that our findings would enable us to provide better scholarly communication services to our graduate students, graduate advisors, and faculty members, as well as provide a service to the broader academic community, especially for librarians in institutions similar to ours.

LITERATURE REVIEW
Literature on ETDs in general began to be published around 1987 (Fineman, 2003), and the topic of ETDs being openly available online has been published in the professional literature for both higher education and libraries for more than a decade (Clement & Rascoe, 2013;Eaton, Fox, & McMillan, 2000;Seamans, 2003). In particular, Corbett provided a recent and excellent overview of the transition from print theses and dissertations to ETDs (Corbett, 2017).
Over ten years ago, Charles Lowry, the former executive director of the Association of Research Libraries (ARL) and former dean and university librarian at University of Maryland and Carnegie Mellon University asserted the following: In many, perhaps the majority, of cases, the dissertation or thesis is and will be the only version of the work that ever appears. In other cases, the appearance will have no effect on the author's later ability to publish a version(s) of the work. On the other hand, in some cases digital posting may compromise the future prospects for the author to publish. We should not jeopardize access in the former cases to protect the latter. (Lowry, 2006, p. 391). Perhaps Lowry's suggestions weren't shared widely enough, for PhD students continued to hear advice to the contrary, and the topic began to be addressed anecdotally in the professional and trade literature by 2008. In a notable article for the Chronicle of Higher Education, Jennifer Howard argued that openly available dissertations were actually more likely to be published, because the number of readers (measurable by downloads) could signify a potential market for the work (Howard, 2011). Similarly, Fyffe and Welburn (2008, p. 154) surmised that "concerns over the effect of open access on future publication can be exaggerated." In a 2011 survey of academic publishers, researchers found that 72% of the surveyed academic publishers from the social sciences and humanities welcomed submissions of works derived from ETDs, while only 4.5% indicated they would never consider such a submission for publication (Ramirez et al., 2013). Of those surveyed, most publishers expected the ETD to be significantly revised to match their publishing guidelines with regard to length, audience, and voice (Ramirez et al. 2013, p. 374).
In 2012, a similar study was conducted with academic journal publishers in the sciences. In that study, 51% of the respondents indicated that manuscripts derived from openly accessible ETDs are "always welcome for submission," while 19% would consider each submission on its own merit (Ramirez et al., 2014). Comparing the survey results with those of the 2011 social sciences and humanities study, the researchers found that science publishers were more concerned with open ETDs negatively impacting citation rates of derivative works, while the social sciences and humanities editors were primarily concerned with the marketability of the work (Ramirez et al., 2014, p. 818). This difference makes sense given the differences in the market; a journal article based on an open work is not likely to change the revenue stream for a given journal title, while a published book based on an open work might not get purchased in favor of simply accessing the open work through a repository.
The issue resurfaced for a major debate in 2013 when the American Historical Association released a statement declaring that "an increasing number of university presses are reluctant to offer a publishing contract to newly minted PhDs whose dissertations have been freely available via online sources" (American Historical Association, 2013).
The professional and trade articles covered the "dissertations to books" problem anecdotally, and the focus of studies, including both the 2011 and 2012 surveys, was journal editors' and university press directors' attitudes toward online theses and dissertations. These studies were useful in allowing us to frame our study, but the purpose and motivation of the 2011 and 2012 surveys were different from ours. Davis, Eyer, and Butkovich (2016) published an analysis of engineering dissertations that were subsequently converted to patent applications, but their study also wasn't directly applicable to ours. Similarly, an article by David Stern (2014) addressed the topic but was particularly focused on undergraduate work being made openly available instead of ETDs.
In late 2017, a study of dissertations published as books in 2014 and 2015 that became available in open access institutional repositories was conducted by Johnson, Goldberg, and Detmering (2017). Their study focused on books that were clearly identified as revised dissertations in a book vendor database. The resultant data set was then compared to ProQuest Dissertations & Theses Global and to institutional repositories. Within the study, a small percentage of dissertations were found in ProQuest and institutional repositories. The average amount of time between completing the dissertation and publishing a book was found to be eight years, and LC classes represented in the data set were within the humanities and social sciences. A list of publishers producing the most books based on dissertations was also provided. This study differed from ours in that it focused exclusively on books that had been identified as revised dissertations, it used a smaller data set, and it was undertaken within a more limited time frame. Our study included such works but utilized a longer time frame and expanded its scope to identify books based on dissertations that were not necessarily identified as such by a major book vendor.
In order to frame our study, we also needed to review the literature about publishers that sought to publish dissertations. How common was it for publishers to publish dissertations without heavy revisions? Was it more common among certain types of publishers? What kind? We started our research believing that it was more common for "predatory publishers" to publish dissertations without heavy revisions; however, we quickly determined that the term predatory didn't adequately identify the scope of the dissertation-tobook-publishing experience. Berger (2017, p. 207) began to outline our dilemma when she wrote about predatory publishers, "This term is reductionist." Our issues mirrored Berger's: "As to predators and victims, the picture is far more complex than meets the eye" (2017, p. 207).
Based on our experience, we developed a list of the types of publishers and situations we were concerned about, and-indeed-they were more complex. They included: 1. dissertation republication services that scraped openly licensed dissertations from repositories and offered them for sale unbeknownst to the author 2. dissertation republication services that solicited graduate students (or newly minted PhDs), generally via e-mail, with an offer to publish their dissertation 3. graduate students (or newly minted PhDs) who wanted their dissertations published "as is" and contacted a publisher 4. graduate students (or newly minted PhDs) who wanted their dissertations published "as is" and self-published For our literature review, we focused on the ways in which we could identify activity primarily in category 2-services that solicited graduate students with an offer to publish their dissertation. Secondarily, we were hoping to find additional insight into categories 1, 3, and 4. We hoped to find examples and definitions of each of these situations in the literature, but literature about these categories of monograph publishers in relation to our study proved to be problematic. There was a great deal of literature on predatory journal publishers; the term "predatory" did not seem to be used with monograph publishers nearly as frequently in the literature. Again, we turned to the professional and trade literature. Specifically, Stromberg's 2014 article in Slate appeared to be the only work that dealt comprehensively with predatory monograph publishers, noting that "companies such as Lambert Academic Publishing and VDM Publishing have been actively turning dissertations into books, with little or no change from the completed dissertation to the book" and "monographic predatory publishing includes the practice of targeting authors of masters and doctoral theses for potential publication. Theses are published as-is and revenue is generated by library purchases" (Stromberg, 2014).
With limited reference in the library literature to "predatory" monograph publishers, we looked to discussions of scholarly publishing beyond library literature to better define the self-publishing efforts (items 3 and 4 from our list), where we found a focus on "vanity publishing." Vanity publishing has generally been eyed skeptically in academe (Fotheringham, 1998;Henderson, 1984); finding a good publisher for one's work has often been taken as a proxy for the quality of the work (Cargill, Charvat, & Walsh, 1966;Gilman, 2000). Authors who have been otherwise "unable to attract the interest of larger academic presses may seek a to bolster their credentials" by reaching out to an unknown-to-them publisher (Dudley, 2013), which may be a predatory publisher, may also be termed "dubious," "low-credibility," or "deceptive" (Berger, 2017), or may be a "vanity publisher" or "self-publisher." In the end, it was challenging to determine self-publishing vs. questionable publishing, so further research on the open web for specific cases-along with using the named publishers from the Stromberg article-became the most likely way to make such determinations.
In sum: The scholarly literature has addressed ETDs and publishers' attitudes toward them; it has also addressed predatory journal publishers. The professional and trade literature have anecdotally addressed predatory book publishers, using a variety of ill-defined terms to describe their practices. For these reasons, we were unwilling to use the term predatory publisher to describe monograph publishers or their activity. Because of a lack of scholarly literature, a lack of definitions, and many unanswered questions about the publishing practices of book publishers, we chose to apply the term questionable publisher to encompass the types of publishers and situations, outlined above, about which we were concerned. No scholarly literature has addressed the actual publication of derivatives of the thesis or dissertation across disciplines, except the recent Johnson study noted above. Our study is an effort to seek data across disciplines and in one area-dissertation-to-book publication-to supplement the anecdotal literature and thus to provide better scholarly communication services to graduate students, graduate advisors, and faculty members.

METHODS
The 2010 Carnegie Classification, which was in use at the time our research began, listed 108 U.S. institutions as R1-very high research activity (Carnegie Foundation for the Advancement of Teaching, 2011). We categorized these 108 institutions into five geographic regions: Northeast, Southeast, Central, Northwest, and Southwest. To create a manageable sample size, we decided to narrow our survey to approximately one-quarter of the institutions, or 24 institutions total. Twenty-three potential participants were selected from the Carnegie R1 list. We chose these participants in an attempt to best represent each geographic region as well as both public and private universities. This method required professional judgment and was not a scientific sample. We had already collected data for our home institution, University of Oklahoma (OU), for a total of 24 participants. We then identified scholarly communication librarians as likely contacts for the data we sought from these institutions, since we assumed these individuals were most frequently engaged in ETD policies within their libraries. If a scholarly communication librarian was not available, we contacted a collection management librarian. In either case, if that individual was not the correct contact for the information, we asked them to provide us with the name of the appropriate contact. In all cases, we used library websites to find the names of these individuals. Invitations to participate were sent via e-mail. We conducted the analysis itself throughout calendar years 2014 and 2015, so the data reflects publications which were in print during that time period.
For reasons ranging from logistics, to policy, to data quality/completeness, to political climate, 12 of the 24 institutions were unable to participate. The remaining 12 institutions provided data. At that point we made a decision not to pursue another 12 participants; our data set with the 12 participating institutions was so large that we recognized that contacting more institutions would delay our ability to complete this research in a timely manner. We chose to focus on dissertations from 2000 to 2012 because this was the focus of our initial data set at OU. In order to offer faculty and graduate students data rather than anecdotes about the number of dissertations that become books, we undertook an analysis of our 2000 to 2012 dissertation output.  Table 1 lists the final set of participating institutions from which we gathered data, including the coverage dates of the data each institution provided.
We requested each participating institution to provide the following data elements for dissertations submitted to their institution in the years 2000 to 2012: • full student name • college with which the student was associated • department with which the student was associated • dissertation title • student major • year doctoral degree was awarded • year open access ETDs were initiated at institution One institution was unable to provide the dissertation titles, so we searched that responding library's catalog to determine that information.
For those institutions able to provide all the requested data elements, we searched World-Cat to determine if an institution's dissertation author had published any books since completing their degree. Because authors' names are hyperlinked in WorldCat, finding the dissertation there was a way for us to see if they had published anything else. When more than one author shared a name, we looked for works by that author that were writ- After collecting this data set-the dissertation authors, their dissertations, and subsequent books published by those authors-our next step was to determine how closely the dissertation and the resulting book matched each other. In this step, if we found a book providing a close match to the dissertation on title, pagination, chapters and subject matter, we considered the dissertation to be successfully published as a "matched book." Works with very different titles but pagination, chapters, and subject matter closely matching, as subjectively determined by the researchers, were also considered to be dissertations successfully published as "matched books." WorldCat frequently did not provide enough data to make a determination as to whether a published book was based on a dissertation, so we supplemented our research on titles wherever we could find relevant and verifiable information: dissertation abstracts, library catalogs, Amazon, Google Books, publisher websites, and EBSCO's GOBI database. This last source provides records that often indicate whether a book is a revised dissertation, and it allows one to view tables of contents or sometimes entire books, when they are available electronically. The ability to compare the content of a published book with a dissertation was very valuable in making a determination as to whether a published book was based on a dissertation. If it appeared that significant changes occurred from dissertation to publication, we did not consider the book to be a match for the purposes of this study.
After exhausting all these sources of information, if we still could not make a determination if a published book was a match to a dissertation, we simply removed the dissertation from our data set. We believed it would be better to not count it at all versus crediting it as a published match when it was not. We reviewed a total of 23,476 dissertations, and we removed 37 from this study for this reason. Of course, this also means that we may have missed some publications resulting from dissertations. Perhaps the work was published by an overseas publisher and no WorldCat record existed for us to research; or perhaps the author had changed names, and the name authority control had not yet reflected that fact.
A word about items that showed significant changes from dissertation to book is in order here. Many dissertations provide the seed for a monograph publication, but we chose to focus on dissertations that became book publications more or less intact ("matched"). To do otherwise would have required that we make judgment calls we were not prepared to make and which, we believe, would have made our results more questionable. The only way to have included books based on dissertations that had undergone significant revision would be to survey the authors of those dissertations. This would be a worthy project, but there are obstacles that make it very difficult to undertake. One would have to wait several years before surveying the dissertations' authors, and the likelihood of being able to successfully contact them by that time would be substantially diminished.
Our argument in this instance was that dissertations receiving significant revision from degree to publication were not the crux of the issue. In the limited amount of scholarly literature addressing this, publishers did not appear to be overly concerned about open access dissertations that are significantly different from the book they ultimately publish. In the literature we reviewed, that particular concern seemed to arise from scholarly societies. Instead, our question came from the opposite direction: how many dissertations were published, largely intact ("matched"), as books? Was it common for publishers to publish dissertations without heavy revisions? If so, was it more common among certain types of publishers?
We mapped the academic disciplines included in the study into broad disciplinary categories. We mapped American studies, art, Asian studies, classics, English, foreign languages and linguistics, history, literature, music, Near/Middle Eastern studies, philosophy, and religious studies to arts and humanities. Anthropology, architecture, business, communication, economics, education, ethnic studies, family studies, geography, international studies, journalism, library science/information studies, political science, psychology, public health, social work, sociology, urban planning, and women's studies were mapped to social sciences. Applied mathematics, mathematics, and statistics were mapped to sciences. We chose to analyze only two STEM disciplines, mathematics and statistics. Researchers in STEM disciplines tend to rely more heavily on journal literature than on books, except for mathematics and statistics, which commonly use book literature more heavily than most other STEM disciplines (Sinn, 2005).
Thus, we included mathematics and statistics, theorizing that members of disciplines which utilize books more heavily are also more likely to author books. A more recent article indicates that mathematics literature may be moving away from this trend, but the author of that work also notes that by using Web of Science for data collection, book citations may not have been adequately captured (Barsky, 2012). Two additional disciplines, engineering and computer science, have been reported to use journal literature less heavily; their nonjournal use, however, is spread across a variety of works-books, conference papers, technical reports, theses or dissertations, and standards (Musser & Conkling, 1996).
Occasionally it was challenging to determine in which category a discipline might reside.
In such instances, we looked at the actual dissertation titles to determine how to map them as either arts and humanities, social science, or science topics. For example, the architecture dissertations in the data set were clearly on social science and not humanities topics. American studies dissertations most frequently covered topics in the arts and humanities instead of social sciences, and public health dissertations were focused much more strongly on the social or policy aspects of the topic than the medical.
Without a list of criteria or an accepted definition by which to identify questionable, predatory, or low-quality publishers, we could only identify questionable publishers with the tools we had available to us, namely the two publishers that were the focus of the Stromberg article, VDM/Lambert. We did research publishers not readily recognizable to us on the open web, as well as library-related literature and higher education publications such as the Chronicle of Higher Education, to determine whether we should consider them to be "questionable" publishers according to the list we had developed. For example, we were unfamiliar with LFB Scholarly Publishing and thought it might be a questionable publisher. However, our research identified it as a legitimate firm, frequently publishing in the area of criminal justice and criminology. Firms specializing in self-publication were also confirmed by researching them on the open web; those which provided services allowing an author to pay for their work to be published and sold were considered self-publishing concerns.
Participating institutions received a summary analysis of their institutional data. The summary indicated the following data points for the broad disciplinary categories of arts and humanities, social sciences, and sciences (math and statistics): • number of doctoral degrees awarded • total number of books published • number of self-published books or books published by a questionable publisher • longest publication period (years) • shortest publication period (years) • average number of years to publication In this summary analysis we also provided a further breakdown of the number of books published from dissertations, by academic discipline.

RESULTS
Within the study, across all institutions, 3% of dissertations were published as books that "matched" the original dissertation on pagination, chapters, and subject matter. Table 2 lists the summary data we collected across all 12 participating institutions separated by broad discipline. Public and private institutions demonstrated different publication rates for the broad disciplinary categories. In public institutions, the arts and humanities accounted for 58% of the books published that matched dissertations; social sciences accounted for 40%; and science (math and statistics) disciplines accounted for 2%. In private institutions, arts and humanities accounted for 81% of the books published that matched dissertations. Social sciences accounted for 19%, while no science (math and statistics) disciplines published books from the sample of private institutions we examined.
Our analysis demonstrated that self-publishing and publishing with "questionable" publication firms were active dissertation-to-matched-book publishing trends. In the arts and humanities, 7% of books published were from questionable publishers. In the social sciences, the number of books published from questionable publishers rose to 25%, and in science (math and statistics) disciplines it rose to 40%. We must note that the 40% figure for the two science disciplines is based on a very small number of publications and therefore should be approached with caution.
There was a wide range in the time it takes for a dissertation to be published as a matched book. Our study found dissertations written in 2000 that were published as matched books in 2012 and dissertations published as matched books within the same year. Across all participating institutions, we found that the average number of years for an arts and humanities dissertation to become a book was five; in the social sciences it was four years, and in mathematics and statistics it was two years.
Some disciplines were well represented in this study, since they were offered at more than 50% of the participating institutions. Others were not, with some only offered by one or two institutions. The following disciplines were included in the study and are categorized within the broader categories of arts and humanities, social sciences, and sciences (math and statistics). Those offered by more than 50% of the participating institutions are marked with an asterisk. This is important so that we do not draw conclusions from sample sizes that are too small:

Sciences (math and statistics):
applied mathematics, mathematics, statistics Reviewing the dissertation-to-book trends in academic disciplines, the data show that religious studies, history, Asian studies, architecture, and American studies had 10 to 15% of their disciplines' dissertations published as books that "matched" the dissertation. Of these, all are considered arts and humanities disciplines except architecture, which we classified as a social science. Anthropology, classics, English, ethnic studies, foreign languages and linguistics, literature, Near/Middle Eastern studies and political science had 5 to 9% of their disciplines' dissertations published as matched books. Private institutions converted a higher percentage of dissertations into books than public universities in the disciplines of foreign languages and linguistics, history, literature, and religious studies.
As previously indicated, some disciplines were well represented in this study since they were offered at more than 50% of participating institutions. An analysis of these disciplines provided insight into the average number of years it took at these institutions for dissertations within the select disciplines to be published as matched books. This should not be construed as a reason for all dissertations within these disciplines to receive two-or five-year embargo periods; as noted earlier, only 3% of dissertations in the study were published as matched books, and only a handful of disciplines saw dissertation-to-matched-book publication rates in the low double digits. With this caveat in mind, Table 4 summarizes the average publication period by discipline.
Analysis became more challenging when we tried to determine the possible impact of open access on dissertation publication, because open access requirements had not been in place for very long at many of the institutions we studied. In our study, only the University of Arizona had developed an open access policy for dissertations in the middle of the time period included in the study. Looking at the University of Arizona's data, we found that 2.5% of their dissertations were published as matched books, both five years before and five years after their open access policy was implemented. Other institutions simply did not have enough beforeand-after data to report on, and for the same reason, trends are nonexistent in this data set.

DISCUSSION
The data suggest that authors from institutions similar to those included in our study convert their dissertations into monographs without significant revision at a very low rate (~3%). This is an important data point for institutions discussing the potential impact of requirements that dissertations be deposited in institutional repositories and made openly accessible. Our study did not attempt to determine the reason for the low percentage. It may be that authors were not seeking to publish their dissertations as books, or that they were making substantial changes during the publication process that made the book significantly different from the dissertation.
Among both broad and more focused discipline categories, study data showed that dissertation-to-matched-book publishing trends do exist. This may have import for campus discussions of standard embargo periods offered by institutions. Reviewing the crowdsourced Google spreadsheet Institutions requiring electronic thesis/dissertation submissions (Salo, n.d.), as of this writing, most reporting institutions offer embargo periods of between six months and two years, and about one third offer longer embargoes through extensions, renewals, the granting of a perpetual embargo, or the negotiation of an embargo period other than the standard period(s) offered. Our study indicates that with dissertations for which the author is actively working to publish as a book, the 6-to 24-month embargo periods would be sufficient, provided that extensions or renewals continue to be available. If extensions/ renewals will not be made available, longer embargo periods should be negotiable at the time the dissertation is prepared for institutional repository submission.
The study data shows that self-publishing and what we have termed questionable publishers continue to be an area where librarians can educate graduate students and their faculty advisors. As our literature review determined, self-publishing has generally been eyed skeptically in academe. With self-publishing on the rise (Milliot, 2010) this may change in time, but at present not much credence is given to self-published works. Still, some authors opted for this publication route. Of greater concern, however, is the rise of questionable book publishers. Companies such as Lambert Academic Publishing and VDM Publishing have been actively turning dissertations into books, with little or no change from the completed dissertation to the book (Kennedy, 2011;Stromberg, 2014).
For graduate students planning on a career in academia, publication venue is important. Faculty advisors are likely aware of the publishers which are considered reputable and quality publishers in their respective disciplines, but they may not be aware of questionable publishers seeking to publish inexperienced authors' works with little or no concern as to editing, peer review, layout, or presentation. Making faculty members and students aware of questionable publishing practices and the companies that engage in them may help stem the tide of disreputable "predatory" publishers and publications that have not benefitted from quality editing and peer review. Those in the arts and humanities disciplines appear to be most aware of what constitutes a reputable publisher, as the percentage of dissertations in these disciplines published by companies such as Lambert and VDM is the lowest. Librarians have the chance to continue seeking opportunities to make students and faculty advisors aware of questionable publishing practices and the companies that engage in them.
Exploring the potential impact of open access policies on dissertation-to-book publishing would be a fruitful area for future research. The data set used in this study could potentially become more useful to the authors as soon as 2018, when a full five years would have passed since the last dissertations were completed. This would allow us to search for books published in 2016 and 2017, and to have more data for pre-and post-open access book publication. Researchers could also acquire a data set from ProQuest's Dissertations & Theses Global rather than from individual institutions. This product provides the necessary data elements: author, title, publication year, institution, and department. One could further identify institutions upon which to focus, if desired. With this data, researchers could explore dissertation-to-book trends. If, however, researchers wished to compare dissertationto-book trends before and after open access policies go into effect at select institutions, this information is not available from ProQuest. Researchers could use Institutions requiring electronic thesis/dissertation submissions (Salo, n.d.) to identify this information, search for the information on institutional websites, or contact the institutions directly.
Another potential way to explore the dissertation-to-book question would be to survey authors, but as indicated previously, pursuing this option has significant obstacles. It is difficult to keep track of accurate contact information after students graduate; graduate advisors and alumni associations alike find this challenging. Alumni associations, the most likely entities to maintain this information, don't readily share it with others; indeed, they generally have policies designed to protect the privacy of alumni and limit use of alumni contact information to legitimate, university-affiliated entities.
For future researchers, it is recommended that analyses focus on social sciences and/or arts and humanities disciplines only. This study chose to include math and statistics from the sciences, theorizing that since these disciplines use book literature more heavily than most other STEM disciplines, they would demonstrate a dissertation-to-book trend. However, based on this study, it would appear that dissertation-to-book trends are not significant enough to justify the effort.
Finally, there is a gap in the scholarly literature addressing questionable dissertation-to-book publishing practices. Developing accurate definitions and criteria against which all publishers can be evaluated for unscrupulous or questionable activity would be a valuable asset to the scholarly literature.

CONCLUSION
Whether directly submitted to an institutional repository or deposited into ProQuest Digital Dissertations/UMI and then imported into an institutional repository, ETDs are established in the graduate education landscape. Institutions of higher education choose to make dissertations available to the scholarly community for review, to build the scholarly record, and to provide positive exposure for the young scholar and the institution. Many of them have chosen to do this electronically (ETD). At the same time, these institutions may experience internal pressures to provide embargo periods for dissertations and theses. This study provides a first effort to determine the percentage of dissertations that are later published as books matching the original dissertation, as a service to young scholars, their faculty advisors, higher education institutions, and librarians who work with institutional repositories. For those disciplines that emphasize book publication, this study provides initial data as to dissertation-to-book publication rates-data that has been missing from campus discussions. It is hoped that others will build on this study and continue to help the higher education community understand the impacts of open access ETDs.