For Reviewers

Referees are the core of the journal. This page explains what JEPR asks of them.

What to assess

Assess a clear contribution to representation, measurement or decision. Relevant formal results, replications, reanalyses, methodological resources and findings that identify failures or limits are eligible. Positive, negative and limited findings receive the same standards.

1. Object and claim. What is represented or measured, and what is proved, tested or proposed?
2. Evidence. Are formal results relevant and proofs complete? Are formal illustrations, experiments, observations, reanalyses and simulations clearly distinguished? A formal illustration is not empirical validation.
3. Literature and comparison. Are appropriate alternatives considered, including probabilistic, calibration, selective-prediction or established measurement methods where relevant? Are information access and comparisons fair? Distinguish representational advantages from predictive or decision advantages.
4. Results and limits. Assess uncertainty, unit of analysis, exclusions, reused data, sensitivity and protocol deviations. Equivalence claims require a justified margin and a suitable analysis. For abstention, distinguish empirical risk estimates from proved guarantees and check that the implementation satisfies any guarantee's assumptions.
5. Verification. Where feasible, run the supporting code and check key tables or figures, or assess proofs and computational artefacts. Record which results were independently checked, which could not be checked and why. Consider justified data restrictions and alternative assessment routes.

The report

A useful report contains a short summary of the contribution, the changes the referee considers necessary, suggestions that are optional, and a recommendation: accept, minor revision, major revision or reject. Please keep required changes and suggestions clearly apart. State the verification status of the main results and the limits of the evidence; the direction of an empirical result is not a reason for rejection.

Timing

Referees are asked to reply to the invitation promptly and to submit the report within four weeks. If more time is needed, please tell the editor.

Anonymity, confidentiality and generative AI

Review is single-anonymous: referees know the identity of the authors, and authors do not know the identity of the referees. A referee who prefers to sign a report may do so. Manuscripts, data and code under review are confidential: they must not be shared, and they must not be uploaded to generative AI services.

Conflicts of interest

Referees must declare any relationship with the authors or the work that could influence their assessment, and decline the invitation where one exists. See the Publication Ethics and Malpractice Statement.

Recognition

On request, the editor provides a letter confirming the review.

Becoming a referee

Researchers who wish to review for JEPR can register on the site, tick the option to be contacted with review requests, and list their areas of expertise.