Why do papers get desk rejected?

The existing literature on peer review focuses on the evaluation process performed by peer reviewers and finds modest evidence for social bias and a reliance on certain epistemic concerns (e.g., accuracy and novelty). However, in many journals, only some submitted manuscripts are ever sent out for peer review. Remiss is research on the initial stage of manuscript review, where editors often decide to “desk reject” submitted papers. What occurs at this stage? Do social biases enter? What epistemic concerns are raised? Are they the same or different from those raised by peer reviewers? This paper follows the discussions of editors as they consider manuscript submissions at an elite life sciences journal. Using a multilevel Cox Proportional Hazard model, this paper finds that 60-70% of submitted manuscripts are desk-rejected annually. Papers from Anglosphere countries and those written by higher-status authors are more likely to pass the initial editorial screening despite editors' attempts to counteract potential biases. Additionally, manuscripts are more likely sent on for review when editorial discussions focus on novelty and accuracy, and succeed in finding reviewing editors to guide the paper through peer review. Our findings reveal that initial editorial screening is a critical stage where social and epistemic factors actively determine which knowledge is even considered as potentially suitable for peer review, let alone scientific literature more generally.