Issues in Science and Technology Librarianship
Earth Science & Engineering Outreach Librarian
Ithaca, New York
Anne E. Rauh
Engineering & Computer Science Librarian
Syracuse, New York
Discussions of the potential of open access publishing frequently must contend with the skepticism of research authors regarding the need to pay author fees (also known as publication fees). With that in mind, the authors undertook a survey of faculty, postdocs, and graduate students in physical science, mathematics, and engineering fields at two research universities (Cornell University and Syracuse University) asking for their experience with and opinion of paying author fees for publication of research in open access journals. The results of this survey indicated that most respondents had not decided against publishing in an open access journal due to the author fee requirement. Those who had paid them only requested or received coverage for those fees in grant line-items or from institutional sources in a few cases. Responses seemed to combine cautious optimism about open access journals with intense skepticism about their quality and intense opposition to the idea of having to pay any additional costs from their own pockets.
Nariani and Fernandez (2012) wrote "there have been few studies of author perceptions on publishing in OA journals." In recent discussions with faculty regarding open access, we have heard comments in opposition to being asked to pay to publish in the form of author fees. The dearth of research literature in this area combined with anecdotal "local" sentiments inspired us to expand the body of research by conducting our own study, focusing on perceptions of author fees in open access publishing.
Cornell University has encouraged and materially supported open access publishing among its own faculty and even faculty at other institutions by several means. Cornell University has, since 1991, operated the arXiv preprint web site, which now hosts nearly a million articles, primarily on astrophysics, physics, mathematics, and some other quantitative fields (ArXiv 2014). While not quite an open access venue, arXiv serves in much the same capacity, giving access to research for free to readers and supporting itself financially by soliciting voluntary financial contributions from institutions, with suggested support levels pro-rated based on usage (2012). At least anecdotally, among many physicists, arXiv has in many ways become the only "publication" that matters.
Other, more recent steps to support open access by units of Cornell University include the creation in 2009 of the Cornell Open access Publishing fund (COAP), with a combination of funds from the university provost and library. This fund pays some or all of the costs involved with publication of student and faculty research in open access journals (Peng 2009). The provost has periodically renewed and the funding since that time.
Cornell's University Assembly passed a motion to study open access publishing as a campus-wide policy in April of 2014, prior to which similar motions in support of open access had been passed by the Faculty Senate in 2004, 2005, and 2006. The Assembly motion recommended study of further expanding Cornell's own available open access repositories as an alternative to (though perhaps not replacement of) traditional publishing venues (Carozza 2014).
The context of open access publication at Syracuse University is somewhat different. In general, Syracuse has not had the history of supporting open access to the same extent as Cornell. Syracuse University does not have a fund to assist with open access publishing fees nor has its University Senate passed motions to support open access. Much of the advocacy for open access on campus comes from the libraries though, due to recent staffing changes, open access advocacy is no longer a priority.
The genesis and significance of OA journals and OA publishing has been chronicled in many publications. The specific implications for some specialized libraries as well as for the publishing industry were ably summarized by Albert (2006) who proposed a motivating ideal behind OA journals: making the results of research freely available online and therefore liberating institutions from having to subscribe to expensive journals that monopolize access to the research. Albert's paper is an early record of the shift in attention to this issue, when the National Institutes of Health joined librarians and scholars in their interest in the issue.
There is a longitudinal record of the attitudes of research authors toward OA journals extending back to the 1990s. Xia (2010) compiled and compared surveys extending back to that time and found increasing awareness of OA journals combined with a steady association of them with low prestige and a (false) impression that content in such journals was published without peer review. Xia's longitudinal examination did not delve into the issue of author fees. Butler (2013) found ample evidence in circulation to support both legitimate fears and false impressions about the nature of individual OA journals and of the OA "movement" overall.
Nariani and Fernandez (2012) found generalized resistance to the idea of research authors paying OA fees in their survey of researchers at York University. They found that funding for OA publication was a main concern.
In a study commissioned by the Publishers Association in 2003, Rowlands, Nicholas, and Huntington (2004) found that authors were aware of unsustainable price increases in academic journals and that unrestricted access to scholarly information was seen as a positive shift. However, they concluded that authors would prefer "open access at both ends" meaning that authors would neither pay to publish nor pay to access.
Schroter, Tite, and Smith (2005) found similar attitudes in their study of authors' perceptions. While some authors agreed that publishing charges would be acceptable if covered by grant-funding agencies or universities, many believed that "authors themselves should not be required to pay." Coonin and Younce (2010) received similar responses to their survey; one quarter of authors they surveyed would consider publishing in a journal that required a publication fee if their institution or grant-funding agency would cover such an expense, while slightly more than half of the respondents (56.1%) would not publish in such a journal.
A study by Warlick and Vaughan (2007) was the one example in which some respondents were not concerned with author fees. Biomedical faculty at UNC-Chapel Hill were unfazed by this trend and seemed to expect it. However, faculty at Duke surveyed in the same research had results similar to the previous studies. The authors concluded that "one difference between the two institutions is that UNC-Chapel Hill covers the cost of publishing in BioMed Central journals whereas Duke does not".
Solomon and Bjork (2011) received replies to a survey from 429 authors who had published in 69 OA journals and thereby paid what they refer to as "article-processing charges" (APCs) or "publishing fees." The disciplines of the research authors surveyed were classified into seven categories: health sciences, biology, and life sciences; education, social sciences, law, and political science; history, archaeology, art, architecture, languages, and literatures; technology, engineering, mathematic, statistics, and computer science; business and economics; chemistry, earth sciences, physics, and astronomy; agriculture and forestry. Their responses regarding amounts paid for APCs and the sources of funding for such costs varied broadly, particularly between agriculture and forestry (lowest cost APCs) and health sciences, biology, and life sciences (highest). The results also showed a difference in who used grant funds to cover APCs. Thirty percent of respondents from chemistry, earth sciences, physics and astronomy said that they used grant funds to cover APCs whereas only 10% of business and economics respondents did likewise.
As to the decision to publish in an OA journal, they found that speed of response, disciplinary fit, and journal quality were all more important considerations for researchers than was the fact that a journal was OA (2011).
Lastly, Fowler (2011) looked at publishing trends in mathematics and found that authors' unwillingness to pay author fees was a larger factor than their inability to do so when deciding against publishing in an OA journal.
Our home institutions (Cornell University and Syracuse University) both had licenses to the Qualtrics Survey software product ('Qualtrics'). This product enables the creation and dissemination of online surveys and in-process examination of results. In addition, Qualtrics had a number of features commending its use in this study. First, when designing a survey, it enables "skip logic" that can either ask follow-up questions or else skip them, contingent upon a respondent's answers. For example, when asked "Within the past three calendar years, have you ever paid author fees for publishing an article in an open access journal?" if the participant answers "Yes," he or she is then presented with two follow-up questions asking the names of such journals and then the largest author fee paid. If the respondent answers "No," those two questions are skipped and the survey proceeds.
Qualtrics also enables a relatively high degree of control over survey distribution and participation. Surveys created in Qualtrics can be hosted as web sites, with the URL publicized only as widely as desired and with authentication using institutional IDs and passwords enabled in order to ensure that only certain groups can give responses. The surveys can also be taken down once the survey period is over.
The surveys used for this study (formatted for print) are found in Appendix 1. Exporting the survey causes it to lose some specific Qualtrics formatting, but the representation is adequate. Individuals with a Qualtrics license who wish to see the surveys as they originally appeared may contact the authors to receive an exported version in .qsf-format.
We were familiar with the best methods of communication with the departments we wished to survey at each institution and we chose varying means accordingly. At the same time, we tried to keep the survey distribution and presentation methods as similar as possible at both institutions. At Cornell University, we identified the mailing lists of their academic departments (see below) and distributed the survey link to them via their department administrators (not department chairs). The survey was available to respondents between Monday, September 30, 2013 and Monday, November 4, 2013, with a renewed call for participation (viz. "There is still a chance to participate in this survey") distributed in mid-October. We selected this time period on the basis of it being late enough after the beginning of the semester and far enough from final exams and the holidays at the end of the semester that participants would have the maximum likelihood of being able to participate.
We approached these departments at Cornell University for participation:
At Syracuse University, the authors contacted faculty members directly with an invitation to participate in the survey via e-mail. Prior to distributing the survey, department chairs were notified about the upcoming invitation and made aware that this survey was part of the authors' research agenda. A college-wide mailing list was used to contact engineering and computer science faculty as the author is the liaison to that college and had access to send e-mail to that list. For the other STEM faculty surveyed, department web pages were used to collect e-mail addresses and the author e-mailed those faculty members directly. The survey was live from October 1, 2013 to November 1, 2013. We invited faculty to participate on October 1, and reminded them to participate on October 14 and October 28.
The departments at Syracuse University we invited to participate were:
We chose these departments at both institutions based upon several rationales. We wanted to broadly survey the entire physical sciences, mathematics, and engineering communities at both institutions. We wanted to select departments that were similar across both institutions so that we could make some one-to-one comparisons. Finally, we wanted to avoid the confusion that surveying the life sciences would entail. Cornell University has a diverse array of biological science departments, including some affiliated with its veterinary school, agriculture school, and medical campuses. To avoid any confusion entailed by surveying them, we elected to exclude them (with the exception of the departments of biomedical engineering at both universities).
The rates of response for almost all of the individual engineering fields at both Cornell and Syracuse were too low to allow for meaningful discussion or statistical power (the exception was the Department of Applied Physics at Cornell, which is affiliated with the College of Engineering). We decided to simply roll all such findings together for both institutions under the single heading of "engineering." This obscures important interdisciplinary differences, but was done to enable some use of the data.
We did not calculate response rates because it was not practical to take a "census" of the departments surveyed since there are far too many graduate students, postdoctoral researchers, adjunct faculty, limited-term researchers, lecturers, and cross-appointed faculty at both institutions to get a total count.
From 123 total responses across both universities, the largest number of survey respondents came from the faculties of physics at Syracuse (15%), applied physics at Cornell (11%), mathematics at both Cornell (11%) and Syracuse (11%). No other single department at either institution returned more than nine responses. The combined "synthetic" category of engineering received a total of 36 replies (29%) across both institutions.
Tenured faculty was the most common professional status of the respondents at both institutions (46% from Cornell, 39% from Syracuse). The majority (79% from Cornell, 82% from Syracuse) of respondents said that they had not paid an author's fee to publish an article in an OA journal within the past three years.
Of those respondents who had paid such a fee to publish, there was little overlap in the journals identified. Of 15 journal titles identified by respondents at the two institutions, two were PLOS ONE. No other individual title was identified more than a single time. At both institutions, the most common response among those who said they had paid an author fee in the past three years was that they fee was more than $1,000.
At both institutions, the majority of responses (51% at Cornell, 66% at Syracuse) indicated that respondents did not forgo publishing in an OA journal due to an author-fee requirement. Unfortunately, we noted some lack of clarity in this question after the fact. It could be that respondents had considered publishing in an OA journal but elected against it due to the author fee or that they had simply never considered publishing in an OA journal and so author fees were a non-issue.
At both institutions, the vast majority of responses (87% at Cornell, 85% at Syracuse) indicated that respondents had not requested line-item funding to cover author fees. Of the those who indicated that they had done so, the largest single group, though not majority, (just 14% at Cornell, 11% at Syracuse) said they had done so as part of a grant application (for example, as part of an application to the NSF).
Since affirmative responses to the line-item funding question were so low, the two succeeding survey questions -- asking whether respondents had received such funding and what was the largest amount awarded -- had correspondingly low response rates: just two from Cornell, who received "More than $1,000" and three from Syracuse, who received "Between $500 and $1000".
The final structured-response question was "If a student, postdoc or other individual associated with you or your lab wished to publish in an open access journal, would you be willing to pay their author fees?" The most common response at both institutions (39% at Cornell, 49% from Syracuse) answered "Not sure/Don't know". A further 24% from Cornell and 37% from Syracuse simply answered simply "No."
In the context of the other findings, this might indicate that many of the survey respondents are not yet broadly publishing in OA journals themselves, much less encouraging individuals in their labs or research groups to do so. The tabulated data for all survey questions can be seen in Appendix 2.
The complete set of responses to the two free-text questions at the end of our survey can be found in Appendix 3. The responses can be broadly categorized. The first free-text question was "Overall, do you have any opinions or feelings about the model of OA publishing that supports journals by the model of charging contributors author fees?" The responses were strongly negative (70% at Cornell; 62% at Syracuse) or of mixed or uncertain opinion.
To add some overall qualitative analysis to the quantitative results, in both the case of Cornell and Syracuse, there were two general strands to the 'negative' responses. First were those respondents (43% of all responses from Cornell, 27% from Syracuse) who identified some aspect of cost as the deciding factor: they were unwilling or unable to pay author fees in order to publish in OA journals, and/or felt that it was unjust to expect research authors to bear any additional costs at a time of declining research funding, and/or some other variation upon these themes. The second strand of criticism for open access publishing can be generally categorized as doubts about the quality of OA journals; 38% of responses from Cornell and 57% of those from Syracuse gave responses of this sort. Survey respondents expressed concern that mechanisms of refereeing and editing were not sufficient, that the Journal Impact Factors of such journals were too low, and that "the quality simply isn't there."
What positive comments there were in the first free-response question (just 19% at Cornell, 16% at Syracuse) can be characterized as tentative. The general sentiment of these comments was that the model is an emerging and may be flawed now; hopefully it will improve in the future.
For the second free-text response question, "Would you recommend publication in OA journals to a colleague and/or student? If so or if not, for what reasons?" the responses can be categorized as somewhat negative at Cornell (40%), overwhelmingly negative at Syracuse (78%).
To generalize from all of the responses above with an eye toward comparison of the two institutions, we note more similarities than differences. At both institutions, mathematics and physics faculty together formed the majority of respondents (possibly reflecting the much more active scene of alternatives to conventional journal publishing in these disciplines, either through open access or such venues as arXiv). Similar numbers reported that they had not decided against publishing in a journal due to the presence of author fees. Similar numbers (and similar overwhelming majorities) indicated they had not requested line-item funding to cover OA author fees. Finally, similar numbers answered either that they would advise a student or junior colleague that they should not publish their findings in an OA journal or else, at best, that they were unsure whether they would do so or not.
One of the only notable differences we found in the responses from the two universities was that, in the first free-response question, significantly more Cornell faculty than Syracuse faculty identified unwillingness to pay, shortage of funds, or some similar financial concern as their concern about the model of open access publishing. This is counter-intuitive given the higher research profile of Cornell University (and, presumably, the greater financial resources available to faculty there) as well as the availability of the Cornell Open Access Publishing (COAP) Fund to pay at least some of the costs of OA publishing. From quick searches in Scopus and Web of Science, it is evident that Cornell faculty publish more frequently than those at Syracuse, and so use of open access journals which require author fees for each submission might be a greater aggregate expense for Cornell faculty.
Concerns about the quality of open access journals seemed matched by a general unwillingness by respondents to accept any added costs. There seemed to be little understanding on the part of these respondents that they were already paying for library resources through overhead on their grants, or that engaging in OA publishing was a worthwhile form of "activism" toward a more equitable, accessible research literature landscape.
Given our findings, if OA publishing is to gain greater currency among faculty, we believe that librarians will need to do a better job of communicating the value of open access and of the existence of high-quality OA journals. Contrary to the perceptions of some faculty, there are good quality, well edited, peer-reviewed OA journals and not just fly-by-night, "predatory" ones. We believe it is also clear that if faculty are to bear additional costs, they will want additional, specific financial support from funders and/or their home institutions. Merely rhetorical appeals to faculty to publish in OA journals will likely not be sufficient without such incentives, at least at the outset.
It may be that future moves by non-OA journal publishers may further undercut the appeal of OA publishing. As was mentioned by some respondents in the final questions on our survey, some non-OA journals already charge author fees and research faculty may wonder why they should forgo the "known quantity" of a traditional publisher and journal title for the unproven model of OA if they are going to have to pay an author fee either way. Certainly there are some fields in which OA journals and practices have penetrated deeply and become part of the landscape. But there are other fields in which the concept is still unfamiliar and high-quality OA journals may not yet exist or be widely known. This disparity may hinder them from ever gaining traction equally across the STEM disciplines.
Finally, as regards future directions for research, our survey strongly points to highly disparate "cultures" of OA publishing within STEM fields. It is evident that fields such as mathematics and physics are far more accepting of OA publishing and, presumably, author fees than are fields such as chemistry and engineering. In mathematics, several societies have begun OA journals parallel to existing conventional ones. In physics, the influence of arXiv is broadly felt. We believe that likely explanations for chemistry and engineering faculty reluctance to publish in OA journals include a paucity of good-quality OA titles in these fields and the desire of faculty to publish in journals that are well-known not only in academia but also in industry. We believe investigation of these disparate OA or anti-OA cultures within disciplines would be fruitful ground for future study.
Jeremy Cusker would like to thank Dianne Dietrich and Gaby Castro-Gessner of Cornell University Library for their extensive and invaluable assistance with Qualtrics and the design of the survey; Steve Rockey and Jill Powell of Cornell's Engineering Library for their assistance and encouragement; and his family for their love and support.
Anne E. Rauh would like to thank Scott Warren of Syracuse University Libraries for his assistance with the dissemination of this survey and Janet Pease for her encouragement of this project and librarian-driven research at Syracuse University.
Albert, K. M. 2006. Open access: Implications for scholarly publishing and medical libraries. Journal of the Medical Library Association : JMLA 94(3):253-262.
Arxiv. 2012. Initial 5-Year Support Pledges. [Internet]. Available from: https://confluence.cornell.edu/pages/viewpage.action?pageId=174295285.
Butler, D. 2013. Investigating journals: The dark side of publishing. Nature 495(1442): 433-435.
Carozza, J. 2014. Cornell scientists encourage open access publishing. Cornell Daily Sun. May 7, 2014.
Coonin, B. and Younce, L.M. 2010. Publishing in open access education journals: The authors' perspectives. Behavioral & Social Sciences Librarian 29(2):118-32.
Fowler, K.K. 2011. Mathematicians' views on current publishing issues: A survey of researchers. Issues in Science and Technology Librarianship 67. Available from: http://www.istl.org/11-fall/refereed4.html
Nariani, R. and Fernandez, L. 2012. Open access publishing: What authors want. College and Research Libraries 73(2):182-95.
Peng, Y. 2009. University creates $50K fund to aid open-access publishing. Cornell Daily Sun. October 1, 2009.
Rowlands, I., Nicholas, D., and Huntington, P. 2004. Scholarly communication in the digital environment: What do authors want? Learned Publishing 17(4):261-73.
Solomon, D. and Bjork, B. 2011. Publication fees in open access publishing: Sources of funding and factors influencing choice of journal. Journal of the American Society for Information Science and Technology 63(1): 98-107.
Schroter, S., Tite, L., and Smith, R. 2005. Perceptions of open access publishing: Interviews with journal authors. BMJ 330(7494):756.
Warlick, S.E. and Vaughan, K.T. 2007. Factors influencing publication choice: Why faculty choose open access. Biomedical Digital Libraries 4:1.
Xia, J. 2010. A longitudinal study of scholars' attitudes and behaviors toward open-access journal publishing. Journal of the American Society for Information Science and Technology 61(3): 615-624.
This work is licensed under a Creative Commons Attribution 4.0 International License.