Does hiding author names make science fairer?

(Press-News.org) Catonsville, MD, Oct. 10, 2025 — A new study has tested whether hiding authors’ identities in the peer review process makes academic evaluations more fair, reliable and valid. The answer: it’s complicated.

The research was published in the INFORMS journal Management Science in an article entitled, “Blinded versus Unblinded Review: A Field Study on the Equity of Peer-Review Processes.” The study was conducted by Timothy Pleskac of Indiana University; Ellie Kyung of Babson College; Gretchen Chapman of Carnegie Mellon University; and Oleg Urminsky of the University of Chicago.

The researchers compared single-blind review (where reviewers know the authors’ names but authors don’t know reviewers) with double-blind review (where neither side knows the other’s identity). They found that in single-blind review, the process favored senior coauthors and first authors who were Ph.D. students or research scientists, while disfavoring Asian (versus White) first authors.

Male first authors received slightly higher ratings in both systems, with the gap somewhat larger under double-blind review. Yet, submissions with a greater proportion of male coauthors received lower ratings in the double-blind condition.

Both review systems only moderately agreed on which papers were strongest. In fact, less than half of the top-rated papers under one system were also top-rated under the other. This means conference programs could look very different depending on which review style is used.

“Peer review is the backbone of science, but it’s also often criticized for bias and inconsistency,” said Pleskac. “Our research shows that while blinding can reduce certain inequities, it’s not a cure-all. Both systems leave room for bias and for random ‘noise’ in decisions.”

The study also examined whether either system better predicted real-world outcomes such as talk quality, audience engagement, or eventual publication. They found limited differences: neither system strongly predicted which papers went on to succeed.

“Our biggest surprise was the noise,” added Pleskac. “By far the strongest effect we saw was variability between reviewers, regardless of whether they were blind. That tells us there’s something fundamental about how peer review works that needs rethinking. One possibility is to embrace that noise — for example, by taking the top group of submissions and using a lottery to decide among them. Random selection could help bring in ideas and perspectives we might otherwise miss. Science, after all, is about exploration.”

For their research, the study authors examined 530 real conference submissions for a major academic society. They recruited over 100 reviewers, randomly assigning half to single-blind and half to double-blind conditions. To ensure a fair head-to-head comparison, each paper was reviewed by at least three reviewers under each of the two systems analyzed. The reviewers not only tracked review scores, but also real-world outcomes in terms of whether a submission was accepted and later published as full-length journal paper.

The study compared the single-blind process to the double-blind review process on three criteria: fairness, reliability and validity.

The findings highlight both the promise and the limits of anonymity in peer review, offering evidence-based directions for designing fairer and more reliable evaluation systems.

About INFORMS and Management Science

INFORMS is the world’s largest association for professionals and students in operations research, AI, analytics, data science and related disciplines, serving as a global authority in advancing cutting-edge practices and fostering an interdisciplinary community of innovation. Management Science, a leading journal published by INFORMS, publishes quantitative research on management practices across organizations. INFORMS empowers its community to improve organizational performance and drive data-driven decision-making through its journals, conferences and resources. Learn more at www.informs.org or @informs.

# # #

END