I recently saw a paper (biological sciences, genomics) that had such poor figures that made consider if I as a reviewer could have suggested rejection based on that. Issues in said paper:

i. 3D-barpplots;

ii. Multi-layers 3D donut plots ;

iii. red-green color schemes;

iv. inconsistent plots with poor labels.

This goes on for the 8 main figures, add nothing to the paper, and is so distracting that makes reading it and interpreting the results quite difficult. I assume the authors (two) know how to make decent figures because in some panels they used stacked barplots which in my view would have been used as replacement for the 3D donut plots.

So in an hypothetical scenario, is it ethical or acceptable for a reviewer to recommend rejection of a paper because of poor figures? Assumptions:

  • the authors were asked to improve the figures but ignored the request or did not improve enough;
  • the quality of the research is otherwise acceptable.
  • Could you please elaborate why 3D-barplots and Multi-layer 3D donut plots are a no-go for you? They are not common in scientific papers in my area, but they are neither "incorrect" nor "banned". Also regarding red-green color schemes: Due to color-blindness? There are different types of this visual limitation that might also effect other colors. – J-Kun Sep 14 at 12:11
  • @J-Kun there are already some studies showing that "the addition of 3-dimensional (3D) perspective depth cues lowered accuracy" ref and it is generally adviseable not to use them ref. – fridaymeetssunday Sep 14 at 12:35
  • @J-Kun For the layered donut plots, it is pretty easy to see the 10% of the volume in the inner ring is not equivalent to 10% in the outer ring making the comparisons meaningless between rings. Furthermore the perspective it adds changes how the same area in the bottom and top of the ring is perceived (Steve Jobs used to use this trick) – fridaymeetssunday Sep 14 at 12:38
  • @J-Kun re the color scheme, yes there are other types of color-blindness and red-green was just used as an example (it is also afaik the most common). Changing colors comes for free and there is no reason to alienate a % of one's audience. Ultimately it's about making the paper as easy to ready as possible without dumbing down from the research. – fridaymeetssunday Sep 14 at 12:41
  • @fridaymeetssunday Oh my god, who ever okayed wood-styled 3D pie charts? – Azor Ahai Sep 14 at 23:10
up vote 6 down vote accepted

In principle yes. However the situation is, as always, more complex. Usually a paper does not get rejected due to a single thing that is wrong, e.g. if there are problems with the abstract, the references, the figures, the organization, the language… and if it is just one of them, but the overall contribution of the paper is good, it may still get published A few exceptions: If the findings of the paper are wrong, the methodology is flawed, or the premise of paper is false it should be rejected. In the other cases it depends on a combination of the shortcomings and if they are fixable or not.

So I would say that the quality of the figures should play a role in the review to the point that a reviewer should write that they cannot recommend acceptance of the paper paper unless the figures are improved (same applies to the list "abstract, references,…" from above).

Of course a reviewer can point out poor figures, and if they get not improved he or she can surely state that he doesn't recommend the paper for publication in this state.

However, a reviewer does not reject a paper, that is the editor's job, based on multiple reviews and possible other information (e.g. how many papers are needed to fill the journal).

  • "a reviewer does not reject a paper" - I have rephrased the sentence where this was stated. As for the need to fill the journal, this is was in a purely online journal. – fridaymeetssunday Sep 13 at 10:20
  • 4
    "Figure X is illegible" is one of the most common things I write in peer reviews. – Anonymous Physicist Sep 13 at 13:07

I think that it is fairly reasonable to ask reviewers to comment on any technical fault they find. It will be the editor's decision how serious these faults are. In this case it sounds as though the problem is not substantive and that the improvements should have been added to one of the stages of revision.

In some cases convoluted figures could prevent proper review of the paper by misrepresenting the results. In this case they should be fixed at an early stage in the review process, and of course if the standardized figures no longer support the arguments presented in the paper this can lead to rejection.

Most journals have clear standards for figures and the editor and/or editorial staff should make sure that the paper conforms to these standards even if the reviewers don't mention it at all.

Your Answer

 

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.