Many journals, including the Journal of Comparative Physiology A, allow or even request authors, as part of the submission process, to suggest potential reviewers (often referred to as ‘preferred reviewers’) for assessing the merits of their manuscript. Whereas editors are not bound to choosing reviewers from this list, many routinely make use of these recommendations. However, author-suggested reviewers are only one of several sources from which editors select reviewers. Others are experts they know personally; members of the journal’s advisory board; authors of articles cited in the submitted manuscript; authors identified through searches in scientific literature databases, such as PubMed or Web of Science; names of past reviewers stored in editorial management systems; and individuals suggested as alternate candidates by reviewers who declined the editor’s invitation for review.

While authors might frequently have a more intimate knowledge of experts in their fields than editors, there is no question that they have little interest in suggesting reviewers whom they suspect will provide negative reviews of their papers. The assumption of such a bias in the authors’ recommendation of potential reviewers is in line with studies reporting that author-suggested reviewers rate manuscripts frequently more positively than editor-selected reviewers (Schroter et al. 2006; Bornmann and Daniel 2010; Helton and Balistreri 2011; Moore et al. 2011; Kowalczuk et al. 2015; Fox et al. 2017), and thus increase the chances that a paper will be published. Similar concerns have been raised in terms of the assessment by applicant-nominated referees of research grant proposals, characterizing their ratings as “biased, inflated, unreliable, and invalid” (p. 33; Marsh et al. 2007) and leading the Australian Research Council to abandon the use of reviewers suggested by grant applicants.

It is unclear how much the results of the above studies on potential author-suggested-reviewer bias can be generalized to other journals and funding agencies, and, in particular, whether the possibility of suggesting reviewers has benefited some authors of the Journal of Comparative Physiology A. Addressing the latter question is especially important when considering whether the option that authors can suggest reviewers should be maintained in the future.

To examine whether the recommendations of author-suggested reviewers differed from those not suggested by the authors, I carried out a retrospective analysis of 162 manuscripts submitted via Editorial Manager to the Journal of Comparative Physiology A between 2015 and 2021. In each of these submissions, a decision by the Editor-in-Chief was made after peer-review by two reviewers. Submissions in which only one reviewer, or more than two reviewers, had been consulted were not included in this analysis, as were submissions in which a ‘reject’ decision had been made without peer-review. (The latter ‘reject-without-peer-review’ procedure is common practice in the case of manuscripts that clearly do not meet the minimum requirements for publication, based on initial review by the Editor-in-Chief.) I was not involved as Editor-in-Chief in any of the final decisions. After anonymizing all data used for analysis, the nominal recommendations by the reviewers were transcribed into ordinal scores using the following scale: 0 = reject; 1 = major revision; 2 = minor revision; 3 = accept.

In 135 (= 83%) of the 162 submissions analyzed, the author(s) suggested potential reviewer(s) (range 1–8 reviewers; mean 3.3 reviewers; median 3 reviewers). There was no significant difference in the number of suggested reviewers (including cases in which no reviewer was suggested) between submissions that were finally rejected and submissions that were finally accepted for publication (Mann–Whitney U Test, U = 2780; p = 0.088, 2-tailed; N1 = 83 rejected manuscripts; N2 = 79 accepted manuscripts).

To evaluate possible differences in recommendations made by reviewers suggested by the authors and by reviewers not suggested by the authors, submissions were analyzed separately for the two different final decisions made by the Editor-in-Chief, ‘reject’ or ‘accept’. A total of 83 submissions were analyzed for which a final ‘reject’ decision was made. Out of the 166 reviewers involved in the peer-review of these manuscripts, 44 had been suggested by the author(s), whereas 122 had not been suggested by the author(s). The median recommendation scores were identical in each of the two reviewer groups (0 = ‘reject’). The score distributions in these two groups did not differ significantly, as shown by the Mann–Whitney U Test (U = 2418; p = 0.247, 2-tailed; Fig. 1a).

Comparison of relative frequencies of ratings by reviewers not suggested by authors (blue) and by reviewers suggested by authors (orange). a Manuscripts (N = 83) were finally rejected. b Manuscripts (N = 79) were finally accepted for publication. Ratings of the reviewers’ recommendations: 0 = reject; 1 = major revision; 2 = minor revision; 3 = accept

In the submission group that resulted in a final ‘accept’ decision, a total of 79 manuscripts were analyzed. Out of the 158 reviewers involved in their peer-review, 69 had been suggested by the author(s), whereas 89 had not been suggested by the authors. The median recommendations made during the initial round of review (ignoring possible subsequent rounds of review) were identical in each of the two reviewer groups (2 = ‘minor revision’). A Mann–Whitney U Test showed that the distribution of the recommendation scores in these two groups did not differ significantly (U = 2799; p = 0.288, 2-tailed; Fig. 1b).

The comparison of the recommendation scores of suggested reviewers vs. non-suggested reviewers provides indication of similar overall trends of these two groups in their assessment of submitted manuscripts but does not compare how a suggested reviewer and a non-suggested reviewer score the same manuscript. Such a comparison was carried out by selecting from the data used above only those submissions in which the same manuscript was evaluated by 1 suggested reviewer and 1 non-suggested reviewer. Using this subset, a Wilcoxon Signed Ranks Test did not detect a statistically significant difference in the median recommendation scores between suggested and non-suggested reviewers in this subcategory, both for submissions that were finally rejected (Z = − 0.568; N = 30; p = 0.565, 2-tailed; Fig. 2a) and for submissions that were finally accepted (Z = − 0.645; N = 33; p = 0.508, 2-tailed; Fig. 2b) for publication by the Editor-in-Chief.

Comparison of relative frequencies of ratings of the same manuscript within pairs of reviewers not suggested by authors (blue) and reviewers suggested by authors (orange). a Manuscripts (N = 30) were finally rejected. b Manuscripts (N = 33) were finally accepted for publication. Ratings of the reviewers’ recommendations: 0 = reject; 1 = major revision; 2 = minor revision; 3 = accept