Holding reviewers to a code of conduct would be a mistake in my opinion, because it implies that the peer-review process should facilitate an author’s research (see L. J. Beaumont Nature 572, 439; 2019). Reviewers volunteer their time to judge the validity of a paper as a favour to the scientific community, not to the authors.
A code of conduct typically works best in situations that rely on volunteering and mentoring, where outcomes are not clear cut. For a research paper, this could preclude outright rejection by the reviewer, whose mandate would instead be to offer only constructive criticism to the authors. The role of a reviewer is to advise journal editors on a paper’s suitability for publication, not to advise authors on how to make their work more acceptable to the journal. We already have mechanisms for providing some measure of constructive criticism — for example, when reviewers require major revisions.
Asking referees to keep their criticism positive could exacerbate the overall shortage of researchers willing to review manuscripts, particularly if they feel uncomfortable about reining in negative comments. The onus should instead be on the authors — to make their results clear and compelling in the first place.
Nature 574, 176 (2019)