|Home | About | Journals | Submit | Contact Us | Français|
Objectives: We aimed to assess journal authors' current knowledge and perceptions of open access and author-pays publishing.
Design: An electronic survey.
Setting: Authors of research papers submitted to BMJ, Archives of Disease in Childhood, and Journal of Medical Genetics in 2004.
Main outcome measures: Familiarity with and perceptions of open access and author-pays publishing.
Results: 468/1113 (42%) responded. Prior to definitions being provided, 47% (222/468) and 38% (176/468) reported they were familiar with the terms `open access' and `author-pays' publishing, respectively. Some who did not at first recognize the terms, did claim to recognize them when they were defined. Only 10% (49/468) had submitted to an author-pays journal. Compared with non-open access subscription-based journals, 35% agreed that open access author-pays journals have a greater capacity to publish more content making it easier to get published, 27% thought they had lower impact factors, 31% thought they had faster and more timely publicaitons, and 46% agreed that people will think anyone can pay to get published. 55% (256/468) thought they would not continue to submit to their respective journal if it became open access and charged, largely because of the reputaiton of the journals. Half (54%, 255/468) said open access has `no impact' or was `low priority' in their submission decisions. Two-thirds (66%, 308/468) said they would prefer to submit to a non-open access subscription-based journal than an open access author-pays journal. Over half thought they would have to make a contribution or pay the full cost of an author charge (56%, 262/468).
Conclusions: The survey yielded useful information about respondents' knowledge and perceptions of these publishing models. Authors have limited familiarity with the concept of open-access publishing and surrounding issues. Currently, open access policies have little impact on authors' decision of where to submit papers.
The future of scientific publishing based on subscription-based models is uncertain. With growing use of the Internet, free and instant access to scientific literature is increasingly expected. Open access publishing which gives lawful free access to scientific journal content on the Internet with production funded by means other than subscriptions has thus attracted notable debate and publishers are beginning to experiment with new publishing models. The BMJ, for example, has regarded itself as an open access journal since 1998, giving all readers free access to research articles online (and all other content until January 2005).1 The BMJ funds its open access policy through various sources including subscriptions to the print journal and advertising revenue. Other publishers have turned to alternative methods for raising revenue to fund open access to journal content.
One such alternative is the author-pays model that replaces subscription fees with author charges for publication.2 The term `author-pays' reflects the shift in the cost of publishing from the reader to the author. In reality, though it is the authors' funding body that is expected to cover the costs on the authors' behalf. Some author-pays journals waive fees in cases of author economic hardship e.g. BioMed Central. Little is known about authors' perceptions of open access publishing and their perceptions of journals that charge an author fee. Research conducted so far has largely been limited to open access experiments with heavily subsidized author charges, and author surveys with limited sampling or low response rates.3-6
The Wellcome Trust made a conservative estimate of the cost of publishing an open access article to be between US$500-US$2 500, dependant on the journal's level of selectivity.7 However, this figure is arguably an underestimate because it excludes contributions to overheads and profits.8 Publishers' experiments with author charges are of limited value because the charges levied are almost always lower than the true cost of publishing an open access article. Results from these experiments are further limited because the experiments are not taking place in the same market context that would apply if author charges became a widely accepted model. It is thus difficult to generalize the findings of these experiments across journals in the current market context.
The viability of open access publishing models depends on support from a number of stake holder groups, but most obviously authors of academic articles. It is important for publishers considering different business models to gauge the level of support amongst authors for open access publishing, and their perceptions of journals that charge authors. A couple of author surveys have been carried out in an attempt to assess understanding and perceptions of open access publishing and author-pays publishing models.4-6 However, these surveys are of questionable validity because of limited sampling or poor response rates. Some also appear to have used jargon without providing a defining context before asking participants to respond to a subject with which they may not be familiar.
Noting the limitations in previous research and gaps in current knowledge, we conducted a series of open-ended interviews with international authors to assess:
We found a greater awareness of these concepts than previously reported following definitions of terminology. Whilst the interview data documented a range of opinions and perceptions of open access publishing and author charges, the sample was too small to generalize or quantify these findings. In order to establish the extent of knowledge and opinions held about open access publishing and author-pays models amongst the wider academic community, we used our interview data to develop a questionnaire. This paper reports the findings from this survey.
We sampled authors submitting original research papers to three medical journals: BMJ, Archives of Disease in Childhood and Journal of Medical Genetics. We selected two clinical journals from different areas of medicine and a general medical journal to increase generalizability. Data about the authors of research papers submitted to these journals between 1 January 2004 and 31 August 2004 were downloaded from the journals' electronic manuscript tracking systems. Using the random number generation tool in Excel, we took a random sample of 400 papers submitted to each of the three journals during this period. Duplicate authors (n=66) were excluded within and across journals and 21 authors were excluded as we received an automatic response indicating the e-mail address was inaccessible. This resulted in a final sample pool of 1113 authors (374 BMJ, 370 Archives of Disease in Childhood, 369 Journal of Medical Genetics).
An invitation to participate in the online survey was e-mailed to authors with a link to a URL for access to the survey. Non-responders were sent e-mail reminders 1 and 3 weeks after the original mailing.
Questionnaire content was based on data collected in the open-ended interviews about open access author-pays publishing.9 The survey assessed both authors' knowledge and perceptions of author-pays open access publishing by including a series of both factual/knowledge assessing items as well as items assessing the level of agreement with statements. We also asked authors about their direct experience of submitting to author-pays journals.
468/1113 (42%) authors responded (181/374, 48% BMJ; 152/370, 41% Archives of Disease in Childhood; 135/369, 37% Journal of Medical Genetics). The characteristics of the respondents are presented in Table 1. The sample included international authors with a broad age range and level of research experience.
Prior to any definitions being provided, 47% (222/468) reported that they were familiar with the term `open access publishing', 28% (130/468) were not familiar and 25% (116/468) were `not sure'. 38% (176/468) reported that they were familiar with the term `author-pays publishing', 35% (164/468) were not familiar with the term and 19% (91/468) were `not sure'.
After definitions of these terms were provided (see Box 1), only 33% (43/130) of those who said they were not familiar with the term `open access publishing' reported that they knew nothing about the concept (either `completely unfamiliar' or had `come across the idea but knew nothing about it'), Table 2. However, for the term `author-pays publishing' two-thirds of those who had said they were unfamiliar with the term reported that they were still `completely unfamiliar' with the concept, or had `come across the idea but knew nothing about it' (108/164, 66%), following a definition. Of those who had reported that they were not sure or were familiar with the term `open access', 78% (265/338) agreed that the definition we gave was the `same' or `similar' to what they had thought the term meant. For the term `author-pays publishing', of those who had reported that they were not sure or familiar with the term, 89% (239/267) agreed that the definition we gave was the `same' or `similar' to what they thought.
Box 1 Descriptions of open access publishing and author-pays model
Open access describes the publication of journal content online with lawful free access for any reader without the requirement of a subscription or other fee. Currently, a small number of scientific journals publish their content with open access.
Author-pays publishing describes a publishing model where journals charge a fee to authors submitting and/or publishing papers to help cover publication costs, as an alternative source of income to reader subscription charges. This means that journals can publish their content online with open access (free access to any reader). (NB: This is different to journal page illustration or colour charges.) Currently, a small number of journals operate under an author-pays revenue model and charge between £500 to £2000 per paper on acceptance.
Only 24% (112/468) of authors reported that they were aware of a current open access author-pays journal, and 10% (49/468) had submitted to an open access author-pays journal. Of the 49 (69%) who had previously submitted to an open access author-pays journal 30 said they would do so again based on their past experience; 16% (8/49) said they would not do so again; and 22% (11/49) were not sure.
Responses to the statements did not show any clear-cut levels of agreement or disagreement as shown by the distribution of responses across the Likert scales (Table 3). However, compared with non-open access subscription-based journals:
Benefits rated as most important were `more equitable access to science for the developing world' (30% extremely important; 142/468) and `immediate access to research articles when I need them through the Internet' (30%; 142/468) (Table 4). The least important benefit was author ownership of the copyright, with a quarter of the sample stating that it was not at all important.
Across all journals, 55% (256/468) reported they would not submit to `this' journal (Archives of Disease in Childhood 82/152, 54%; BMJ 109/181, 60%; Journal of Medical Genetics 65/135, 48%) if it became open access and started to charge author fees on acceptance in the range of £500 to £2000. Of the respondents, 68% (173/256) reported that their concern over how the author fee could be paid was `very' or `extremely' important to their decision not to submit; 53% (135/256) said it would be pointless to pay to publish in `this' journal when other journals might publish it for free; 29% (138/468) reported they would still submit to `this' journal (Archives of Disease in Childhood 42/152, 28%; BMJ 52/181, 29%; Journal of Medical Genetics 44/135, 33%) if the journal became open access and started to charge author fees on acceptance. The two most important factors influencing this decision (`very' or `extremely' important) were the reputation of the journal (83%; 114/138) and the author's belief in the quality of the journal (81%; 112/138) (Table 5).
The majority of authors (54%; 255/468) reported that open access publishing has `no impact' or `low priority' in their choice of where to submit their paper. Only 13% (60/468) reported that it is `very important' or `the most important consideration' (Table 6).
Thirty-nine per cent (182/468) would only submit to an author-pays journal if it was their first choice of submission or a highly reputable journal (Table 6). A further 18% (86/468) would only submit to an author-pays journal if it was their only chance of getting published and 9% (42/468) would not submit to any author-pays journal even if it was their first choice of journal or highly reputable.
Two-thirds of the authors (66%, 308/468) said they would prefer to submit to a non-open access subscription-based journal than an open access author-pays journal (Table 6).
Twenty-one percent (97/468) thought they would not be required to make any contribution to the author charge—they thought it would be covered by their sponsors or institution (Table 7). Sixteen per cent (74/468) felt that they might have to make a contribution to the cost, and 40% (188/468) thought they would have to cover the full cost (with or without a sponsor for their research).
We found greater awareness of open access publishing and author-pays models than previously reported.4-6 When authors were provided with definitions, many confirmed that they did know about the concept. Knowledge of these publishing models is still not widespread. Journal reputation and perceived quality are more important factors in authors' decisions to submit to a journal than whether it has an open access policy.
This is the first survey with a reasonable response rate about authors' perceptions of open access publishing that used definitions of terminology and provided context. As such it is likely to have resulted in a meaningful response. For example, we were able to demonstrate that providing definitions resulted in more authors reporting familiarity with the concepts. However, we did also observe some contradictory findings. For example 11 respondents having at first said that they were not familiar with the term `open access publishing' later said they were `very knowledgeable about open access publishing'. It is possible that these 11 authors knew a lot about the concepts of open access publishing but were being cautious. These respondents have not been excluded from the analysis as responses to subsequent questions were not reliant on their knowledge of these publishing models.
The main limitation of our study is the response rate (42%), despite strategies to optimize this. However, our response rate was considerably higher than two of the three previous author surveys on this subject (both of these surveys achieved just 4%5,6). An earlier survey achieved a response rate of 58%, but this was limited to 62 social scientists at a single university in Italy.4 We are unable to compare our responders and non-responders as we do not have details of the non-responders.
Our survey was limited to authors submitting to three biomedical journals and so the generalizability of the findings is limited. However, we purposefully included a general medical journal and two specialist journals in different biomedical disciplines to increase the external validity. We also used submitting authors rather than published authors to include authors with varying degrees of experience and authority and took a random sample of authors submitting throughout the calendar year to avoid seasonal effects.
It is also possible that authors biased their survey responses in the hope of persuading the journals in question not to adopt an author-pays model. We did not include control questions for the statements presented in Table 3. By phrasing the questions in opposite terms and sending these to a subsample, we could have tested whether the way the question was phrased had biased the results. However, we did include a similar proportion of positive and negative questions and mixed up the order in an attempt to overcome this potential bias.
This study suggests that open access publication per se is not currently an attractive feature for most authors when selecting a journal to publish in. Whilst academics are encouraged to submit to high impact factor journals, open access is unlikely to be a strong motivating factor.
Authors disliked author charges, and a notable proportion thought they would have to personally contribute to the charge. This may reflect the fact that many authors are not externally funded for their research, and would be unable to access research funds at a time when papers are accepted for publication.10 Publishers should be cautious about introducing open access publishing funded through author charges in the current climate. It is important to note that some funding agencies are proposing support for authors and if money were built into grants for publication charges, funded authors' opinions may change.
Highly reputable journals might suffer less, or not at all, from reduced submissions if many journals switch to this model. `High quality' non-charging journals do exist as an alternative and our findings suggest that there is little that would motivate authors to choose journals that levy a charge.
Further research should be conducted with authors from different biomedical journals and disciplines outside of medicine to determine their opinions and perceptions. Since familiarity with these publishing models remains limited, it is premature to draw conclusions about whether authors will ultimately accept them.
Acknowledgments We thank all the authors who took part.
Competing interests The survey was funded from the BMJ Publishing Group research budget.
Both authors are employed by the BMJ Publishing Group. The BMJ is currently an open access journal and is considering whether to adopt author charges to fund the publication process. The authors do not have a financial interest in what the BMJ decides to do.
Ethics We did not receive ethical approval for this specific survey, but did receive general approval from the BMJ's ethics committee to conduct research with our authors about this topic.