Reviewer selection: author or editor knows best?
http://www.100md.com
《胸》
Thorax Editorial Office, BMJ Journals, BMA House, Tavistock Square, London WC1H 9JR, UK
Correspondence to:
Professor J A Wedzicha
Thorax Editorial Office, BMJ Journals, BMA House, Tavistock Square, London WC1H 9JR, UK; J.A.Wedzicha@medsch.ucl.ac.uk
Differential behaviour of author and editor suggested reviewers
Keywords: peer review
Thorax is now the second highest ranked respiratory journal and submissions are increasing year on year.1 To attract high quality manuscripts, authors must have confidence in our editorial process and we are committed to continuing and improving the transparency, efficiency, and fairness of peer review.
When the current Editors took over the journal in 2003, Thorax changed to online manuscript submission2 and, for the first time, authors were formally invited to suggest up to four suitable peer reviewers for their work. Indeed, we encourage authors to suggest reviewers, as set out in our advice to contributors.3 Peer review is a vital step in the editorial process, guiding the selection of appropriate papers for publication and providing constructive comments for authors to improve their work. So, what are the benefits of suggesting reviewers (or not) on the editorial process? There is, in fact, little evidence on the behaviour of reviewers suggested by authors in comparison to editors, and we have recently investigated this issue.
We examined the submission of 229 original papers to Thorax, representing a random sample of half of those submitted during the 6 month period from October 2003 to the end of March 2004. Data were collected on the presence, number, and utilisation of author suggested reviewers, together with information regarding the outcome of the submission, subject area, and country of origin. These 229 papers generated 430 reviews, and the behaviour of reviewers suggested by authors and editors was compared. Statistical reviews were excluded.
57% of authors suggested reviewers, most commonly providing four names. There were differences in the suggestion of reviewers across country of origin and subject area. Authors from English speaking countries were more likely to suggest reviewers than authors from elsewhere (58% v 34%, p<0.001). Regarding subject area, while the two most common topics were COPD and asthma (together accounting for 38% of submissions), asthma researchers were more likely to suggest reviewers than not, and the reverse was true for COPD.
Manuscripts externally reviewed were more likely to have been written by authors who had suggested reviewers (54% v 29%, p<0.001). Perhaps the presence of suggested reviewers encourages external review? This seems unlikely for two reasons. Firstly, on average, from four suggested reviewers only one was actually used, and each paper was reviewed twice. Secondly, if papers with suggested reviewers were being sent out more often because this was easier to achieve, then one would expect a higher proportion of reject decisions by reviewers on these manuscripts. This was not the case. Perhaps, then, the presence of author suggested reviewers is a surrogate marker of manuscript quality?
Of the 430 reviews, 21% were from author suggested reviewers, with the remainder from reviewers selected by the editors. 75% of review solicitations were accepted and returned complete, and this did not vary with the source of reviewer. Interestingly, it seems that author suggested reviewers generally return a more favourable response, recommending publication in 64% of cases compared with 54% of cases for reviewers selected by editors (p = 0.138). Reflecting this, reviewers suggested by the editors were statistically less likely to return a review discordant with the final editorial decision than author suggested reviewers (25% v 41%, p = 0.010). In all cases of discordance between the author suggested reviewer and the final decision, the reviewer had recommended publication. These data are shown in fig 1.
Figure 1 Differential behaviour of reviewers suggested by authors or editors. *p = 0.01.
In summary, there are intriguing differences by country of origin and subject area in the proportion of authors who suggest reviewers, and the suggestion of reviewers by authors appears to be a surrogate marker of manuscript quality. The likelihood of a review solicitation being accepted and completed did not differ between reviewers suggested by authors in comparison with editors, but reviewers suggested by authors tended to be more favourable to publication. We trust these results—the first to be published on this topic—will interest and inform authors, reviewers, and editors alike. We wish to actively encourage the suggestion of reviewers by authors.
REFERENCES
Wedzicha JA, Johnston SL, Mitchell DM. Thorax annual report: 1 October 2003 to 30 September 2004. Thorax 2004;59:1012–5.
Anon. Thorax online submission. Thorax 2002;57:1004.
Anon. The peer review process on Thorax. Available online at http://thorax.bmjjournals.com/misc/ifora/peerreviewrevs.shtml (accessed 4 May 2005).(J R Hurst, E C Howard and)
Correspondence to:
Professor J A Wedzicha
Thorax Editorial Office, BMJ Journals, BMA House, Tavistock Square, London WC1H 9JR, UK; J.A.Wedzicha@medsch.ucl.ac.uk
Differential behaviour of author and editor suggested reviewers
Keywords: peer review
Thorax is now the second highest ranked respiratory journal and submissions are increasing year on year.1 To attract high quality manuscripts, authors must have confidence in our editorial process and we are committed to continuing and improving the transparency, efficiency, and fairness of peer review.
When the current Editors took over the journal in 2003, Thorax changed to online manuscript submission2 and, for the first time, authors were formally invited to suggest up to four suitable peer reviewers for their work. Indeed, we encourage authors to suggest reviewers, as set out in our advice to contributors.3 Peer review is a vital step in the editorial process, guiding the selection of appropriate papers for publication and providing constructive comments for authors to improve their work. So, what are the benefits of suggesting reviewers (or not) on the editorial process? There is, in fact, little evidence on the behaviour of reviewers suggested by authors in comparison to editors, and we have recently investigated this issue.
We examined the submission of 229 original papers to Thorax, representing a random sample of half of those submitted during the 6 month period from October 2003 to the end of March 2004. Data were collected on the presence, number, and utilisation of author suggested reviewers, together with information regarding the outcome of the submission, subject area, and country of origin. These 229 papers generated 430 reviews, and the behaviour of reviewers suggested by authors and editors was compared. Statistical reviews were excluded.
57% of authors suggested reviewers, most commonly providing four names. There were differences in the suggestion of reviewers across country of origin and subject area. Authors from English speaking countries were more likely to suggest reviewers than authors from elsewhere (58% v 34%, p<0.001). Regarding subject area, while the two most common topics were COPD and asthma (together accounting for 38% of submissions), asthma researchers were more likely to suggest reviewers than not, and the reverse was true for COPD.
Manuscripts externally reviewed were more likely to have been written by authors who had suggested reviewers (54% v 29%, p<0.001). Perhaps the presence of suggested reviewers encourages external review? This seems unlikely for two reasons. Firstly, on average, from four suggested reviewers only one was actually used, and each paper was reviewed twice. Secondly, if papers with suggested reviewers were being sent out more often because this was easier to achieve, then one would expect a higher proportion of reject decisions by reviewers on these manuscripts. This was not the case. Perhaps, then, the presence of author suggested reviewers is a surrogate marker of manuscript quality?
Of the 430 reviews, 21% were from author suggested reviewers, with the remainder from reviewers selected by the editors. 75% of review solicitations were accepted and returned complete, and this did not vary with the source of reviewer. Interestingly, it seems that author suggested reviewers generally return a more favourable response, recommending publication in 64% of cases compared with 54% of cases for reviewers selected by editors (p = 0.138). Reflecting this, reviewers suggested by the editors were statistically less likely to return a review discordant with the final editorial decision than author suggested reviewers (25% v 41%, p = 0.010). In all cases of discordance between the author suggested reviewer and the final decision, the reviewer had recommended publication. These data are shown in fig 1.
Figure 1 Differential behaviour of reviewers suggested by authors or editors. *p = 0.01.
In summary, there are intriguing differences by country of origin and subject area in the proportion of authors who suggest reviewers, and the suggestion of reviewers by authors appears to be a surrogate marker of manuscript quality. The likelihood of a review solicitation being accepted and completed did not differ between reviewers suggested by authors in comparison with editors, but reviewers suggested by authors tended to be more favourable to publication. We trust these results—the first to be published on this topic—will interest and inform authors, reviewers, and editors alike. We wish to actively encourage the suggestion of reviewers by authors.
REFERENCES
Wedzicha JA, Johnston SL, Mitchell DM. Thorax annual report: 1 October 2003 to 30 September 2004. Thorax 2004;59:1012–5.
Anon. Thorax online submission. Thorax 2002;57:1004.
Anon. The peer review process on Thorax. Available online at http://thorax.bmjjournals.com/misc/ifora/peerreviewrevs.shtml (accessed 4 May 2005).(J R Hurst, E C Howard and)