Peer review is the cornerstone of scientific research publication. The review process frequently improves a manuscript,¹ and there is a call for reviewing to be acknowledged and rewarded as a scholarly activity.^2,3 Authors, however, also receive reviewer feedback that they consider ‘ill-informed or biased’.⁴

Datta⁵ provided a list of points to ‘help researchers and scientists to focus their energy to improve the quality of peer reviews’. Although most are important issues for authors and editors, an author is unaware whether a specific reviewer fulfilled the points related to reviewer timeliness, continuity and confidentiality. But authors can judge whether a reviewer is courteous, constructive or has mistaken the review role as that of being the copy editor. The Committee on Publication Ethics (COPE) has an overarching guideline focusing on the ethical aspects of reviewing.⁶ Journals or journal publishers usually provide their guidelines, frequently based on the COPE guidelines.

During the study period, 67 submissions were reviewed. Reviewer comments could be located for 65 submissions (97%). For 63 submissions, complete review documentation was available for data collection; for two, only author responses to reviewer comments were available.

Table 1a and Table 1b summarises submission, review and outcome characteristics. Submissions were made to 36 journals, of which 23 were South African-based. Nearly 60% of reviews contained some positive comments, and 74% of submissions received some positive comments. For rejected submissions, this percentage was 56% (10 of 18 submissions). Reviewers frequently (75% of reviews, 88% of submissions) required some expansion of information (without commenting on how this would impact word counts). Incomprehensible comments comprised mainly garbled typing in comment boxes or poor punctuation or phrasing, making understanding difficult or impossible. Comments indicating that the reviewer had misunderstood the information provided or had not read the manuscript closely consisted mainly of the reviewer indicating that details were missing, whereas the details were stated in the manuscript.

Two reviews contained possibly rude, derogatory and/or destructive comments:

Comments that implied changes to the ethics committee-approved protocol comprised mainly requests for additional types of data to be collected or data to be collected in a different way. Copy-editing occurred frequently. Five reviews (4%) comprised of only language editing.

Reviews were done in a variety and combination of formats, the most common being an unstructured narrative report (59%). Forty-seven percent (n = 54) comprised only an unstructured narrative report, 19% (n = 23) only a report structured by the reviewer and 10% (n = 11) only comments and/or track changes in the manuscript. Twenty-four reviews (20%) contained answers to items and/or questions specified by the journal.

Significant associations were found between review format and types of reviewer comments:

For 10 (15%) submissions, the editor or associate editor added their own comments to the reviewers’ feedback or was the sole reviewer of the manuscript. For 13 (20%) submissions, the editor or associate editor gave input at some point during the review process.

Reviewer comments were received in a wide range of formats. As highlighted by researchers in the surgical field, there is little uniformity regarding the review requirements stipulated by journals.⁷ To ensure that reviews are thorough and cover all aspects of the manuscript, some structured format seems preferable. However, reviewers do not necessarily give feedback in the format required by a journal. As editors of Educational Studies in Mathematics, Mesa et al.⁸ conceded that few reviewers answered the questions stipulated by their journal, and therefore, they had decided that for their journal, no specific review questions or criteria would be stipulated in future.

It was encouraging that most reviews (even those leading to rejection) contained positive feedback. This confirmed why such a structured investigation was necessary; otherwise, authors may remember only extreme or odd cases. As reviewers, it is important that a review should assess not only the weaknesses but also the strengths of a manuscript.⁷

Journals’ review instructions regarding copy-editing differ widely with some stating explicitly that no comment is needed regarding spelling, grammar and layout⁹ while others state that all spelling, grammar and typographical errors must be pointed out at the review stage as no further copy-editing is done later.¹⁰ COPE guidelines⁷ state that a reviewer should ‘not attempt to rewrite it to your own preferred style if it is basically sound and clear’.

In a review of 1491 sets of reviewer comments of manuscripts in the fields of Ecology and Evolution and Behavioural Medicine, Gerwing et al.¹¹ found that 12% contained at least one unprofessional comment towards the author or their work. The two potentially rude comments identified in the current study would be considered unprofessional according to the definitions of Gerwing et al.¹¹ Gerwing et al.¹² report on the social media debate after their 2020 publication and state that the debate indicated that there is a negative cultural zeitgeist in peer review. Reviewers need to ‘put more effort into considering the perspective of authors when wording their comments’.¹² In the current study, there was little evidence of such negativity. Open review, a developing trend in scientific publication,¹³ may also influence the tone and type of reviewer comments. Only one of the journals included in this study made use of transparent open review.

As pointed out by Roediger¹⁴ and Eva,¹⁵ reviewers, as readers of a submission, may misinterpret or misunderstand information provided. As reviewers frequently scrutinise submissions more closely than the usual reader, authors need to acknowledge that readers of the published manuscript may experience the same difficulties. Authors should regard such comments as flagging sections that need to be written more clearly.

Gerwing et al.¹¹ reported that 41% of reviews contained incomplete, inaccurate or unsubstantiated comments. Generally, if authors can motivate in their response why they consider comments inappropriate or incorrect, the motivation will be accepted by the reviewers and/or editor.

Feedback implying changes to the ethics committee-approved protocol places the researcher in an ethical dilemma. The revision time allowed by journals is often 2–4 weeks, whereas approval by the ethics committee of a protocol amendment and thereafter completing the required new work (data collection, analysis) may require extensive time.

The American Speech Language Hearing Association describes the role of the editor to include monitoring the peer-review process ‘to ensure fairness, timeliness, thoroughness, and civility’.¹⁶ Tennant and Ross-Hellauer¹⁷ highlighted that the editorial role is, unfortunately, seemingly focused on decision-making rather than on the process leading to that decision. In a survey of editors’ opinions on peer review,¹⁸ most editors responded that a reviewer’s comments may be edited without his/her permission when a reviewer has used inappropriate or offensive language (58%).

Arthur¹⁹ has called for editors to become more active in pre-screening manuscripts, particularly in terms of language, scope, format and scientific quality. This would lessen the load on reviewers. The review pool is often limited or overburdened² or, as has been shown through mathematical modelling, only ‘a small portion of the scientific community is carrying a disproportionate load of the peer review’.²⁰

Despite its small size and particular profile of manuscripts reviewed, this study gives some insight into what authors can expect from the review process in terms of type and tone of comments. These findings can guide authors during manuscript preparation and authors, reviewers and editors during the review process.