[marsa]

a desk rejection is when the editor in chief (or managing editor, or whoever is the one first receiving the submitted paper) decides to reject the submission without sending it out for peer review

basically a judgment call by the person in charge of a journal that the paper is not interesting or impactful enough to warrant going through with the rest of the review/publishing process

[swyx]

ouch. but also, surely that has to happen to the majority of papers, meaning the snap judgment call of effectively one person greatly colors the quality of the whole process. as a conference organizer this is something i worry about. is there a better process proposed out there - that respects the constraint that high value people have limited time to review things?

[brutusborn]

I think preprints provide a good way to solve this problem, but it requires some cultural shifts.

Imagine a world where all papers start their life as preprints. Researchers read these preprints (like arxiv) and comment on them (like PubPeer). Journal editors then search for the best papers (or hear via word of mouth within their field) and journals compete for the rights to publish the most interesting papers. When an agreement is reached, the journal organised reviewers and asks for any changes to the paper.

I may be missing something critical but I dream of a day when the actual people producing academic output have power over journals who produce nothing but fees.

Historically journal controlled peer review didn’t exist and science still progressed perfectly well.

[marsa]

sadly no, it is an unsolved problem of scholarly publishing imo. on the one hand you have the reputable journals following the traditional publishing process that take pride in their high rejection rates -- these require a large percentage of desk rejections to avoid flooding their reviewers with sub-par papers. thus they'll inevitably have some quality papers fall through the cracks + some flashy sub-par papers making the cut.

on the other hand you have the pay-to-publish journals that have a financial incentive to push as many papers through peer review -- these thrive on sub-par papers that are technically just barely 'good enough', but the upside is that the real good ones will also make it through. however, they inevitably face reviewer fatigue, and the most valuable ones will quit reviewing if they often send them low-quality papers. so basically once in a while they'll publish top notch research without being aware of it.

i'm not aware of any middle-ground solutions out there and it certainly feels like a tough problem to solve.