26 August 2025
Let’s say you are a scientist wanting to share your latest findings with the broader community. The typical practice is to submit a written description of your research project to a journal. The editor of the journal will either reject your manuscript, or send it to two or three referees (aka reviewers) to evaluate it based on their expertise. The reviewers typically ask for many changes to the paper to address their concerns. You then fix as many of the issues as possible and after several rounds of back and forth they may eventually be satisfied. Then the editor usually accepts the paper for publication.
The goal of peer review is to ensure that works that are published are scientifically valid by having experts closely examine the results and the methods used to find them. It is not a perfect system, as evidenced by the existence of retracted papers. These are papers that are found to be incorrect (sometimes even fraudulent) after already going through the pre-publication peer review system. Peer review doesn’t guarantee a study is correct, but it’s probably better than the alternative of everyone just posting whatever they want with no quality control whatsoever.
Being a referee is not an easy task. You need to spend a lot of time and concentration combing through dense scientific jargon that may not be exactly in your area of expertise. Furthermore, in many cases the referees are not paid for this service directly. Rather, it is considered “service to the profession” that is a component of their jobs as researchers themselves. Sometimes they are also compensated by increased status in that the editor will appreciate their insights and tell others, which improves the referee’s reputation in the research community.
Despite these modest rewards, it is often difficult for editors to find reviewers to take on the task of peer review. The editor may ask numerous people to serve as referees before finding the needed 2-3 to agree to the task. Even then, the busy schedule of the referees will often cause them to need more time to complete the reviews. All of this can add up to long delays between when an article is submitted to a journal and when the reviews are returned to the authors.
A further complicating factor is the rapidly increasing number of papers being submitted to journals over time. Since the crude metric of number of publications is sometimes used to decide whether researchers advance in their careers, many are incentivized to publish as many as possible. If every paper needs a few reviewers, and there are more and more papers to consider, the demand for reviewers will also increase. Authors are also incentivized to first submit their work to the highest status journals and then move down the rankings if the paper is rejected by the top journal. This causes editors to have to “desk reject” huge numbers of papers without even sending them to peer review.
I have often wondered if referees should be paid directly for their service. While many researchers may be primarily motivated by curiosity rather than money, at the margin monetary compensation could speed up the review times. Furthermore, if authors were required to pay a small fee just to submit their paper, it might encourage them to only submit the paper to the most suitable journal rather than the highest status journal. A possible proposal would be to have the authors pay \$20 per article submitted. If about 20% of the submitted papers progress to peer review, that would mean two referees could be paid \$50 each for the review. The payment amount could be reduced if the review was not turned in on time. The fee could be on a sliding scale to avoid blocking researchers with financial hardships.
There are plenty of other issues with the peer review system that I am not addressing here. For example, some for-profit publishing companies charge exorbitant fees (thousands of dollars) to the authors after the publication has been accepted. In other cases people are not allowed to read the publication unless they purchase a costly subscription. These fees are usually paid with grant money, which primarily comes from the taxpayer. There have been numerous efforts by the scientific community to set up alternative journals and other publishing platforms that are less costly and universally accessible, with mixed success.
Another, more aggressive proposal, would be to dispense with journals and pre-publication peer-review altogether and rely entirely on post-publication review. For example, everyone could just post their work on a preprint server like Arxiv and then others could post commentaries criticizing or advocating others’ work. The big challenge with that would be to prevent things from degrading into the “comments section” as commonly observed on social media sites.