For reviewers

1. Do you have the expertise

   Not for the whole manuscript necessarily — say which part you can assess. A reviewer who covers the design but not the clinical context is useful if the editor knows that is what they are getting.

2. Do you have a conflict

   Decline where an author is a current or recent collaborator, a colleague, someone you supervised or were supervised by, or where a financial or personal interest could reasonably be perceived to affect your judgment. If you are unsure, say so and let the editor decide.

3. Can you do it in the time

   If not, decline promptly and suggest someone else. A late decline costs the authors more than a fast one.

• Do not share the manuscript, discuss it outside the review, or use its content or ideas before publication
• Do not involve a colleague or a student without the editor’s agreement; where the editor agrees, name them so their contribution is on record
• After submitting the review, reviewers should securely delete downloaded manuscript files and must not retain or use unpublished material.
• Reviewers must obtain journal permission before using AI to assist with a review. Any permitted use must be disclosed, and confidential manuscript content must not be uploaded where confidentiality cannot be assured.
• For double-blind manuscripts, reviewers must not attempt to identify the authors.

Fit

Does the work make a substantive contribution to behavior therapy or mental health? Assess the contribution in relation to the question and study design.

Question and contribution

Is the question clear, and does the study add something to what is known?

Design

Is the design appropriate to the question? Where it is not, say what would have been, and whether the existing data can still answer something.

Methods

Is there enough detail to repeat the study? This is where most manuscripts in this field fall short.

Analysis

Is the analysis appropriate and pre-specified? Are effect sizes reported with precision rather than p values alone? Is missing data handled and described?

Results

Are all pre-specified outcomes reported, including unfavorable ones? Do the numbers in the text, tables and abstract agree?

Interpretation

Do the conclusions stop where the design stops? Flag causal language attached to a non-causal design specifically, quoting the sentence.

Ethics

Are approval, consent and, where applicable, registration reported? Where participants were at risk, are safeguards described?

Limitations

Are they specific to this study rather than generic?

Reproducibility

Are data, materials and instruments available, or is their unavailability explained?

Intervention studies

Intervention described in enough detail to be repeated; comparator described as fully as the intervention; dose intended versus dose received; provider training and supervision; fidelity monitoring and what it showed; blinding where it was possible; follow-up schedule and retention; harms, deterioration and differential dropout.

Observational studies

Sampling frame and response; validated measures and their properties in this sample; confounders named and justified; missing data; effect sizes with precision; generalizability claimed only to the population studied; and causal language checked sentence by sentence.

Psychometric studies

Construct definition; item development; sample size and composition against the analysis; reliability with confidence intervals; structural, convergent and discriminant validity; measurement invariance where groups are compared; and the intended research or clinical interpretation of the score.

Digital and computational studies

Dataset provenance and quality; how training, validation and test sets were separated, and whether leakage is possible; leakage-resistant internal or held-out evaluation appropriate to a prediction claim; external validation for transportability claims; subgroup performance and uncertainty where data permit, with limitations stated; privacy and consent arrangements; engagement and attrition as findings; and whether the conclusions exceed what the evaluation supports.

Reviews and meta-analyses

Search strategy and its reproducibility; inclusion criteria applied consistently; risk-of-bias assessment; heterogeneity examined rather than averaged away; and whether the synthesis supports the strength of the conclusion.

Case reports

Is there transferable reasoning — a formulation, an adaptation, an instructive course? A case report is clinical observation and must not be presented as evidence of effectiveness.

• Open with a short summary of what you understand the study to have done, so the editor can see whether it was read as intended
• Number your points, and separate what is essential from what would improve the manuscript
• Quote the sentence or point to the table when you raise something
• Say what would fix a problem, where a fix exists
• Write about the work, not the authors
• Do not require citations to your own work
• Note language that needs attention, but do not treat it as a scientific fault

Confidential comments to the editor are for concerns you cannot put to the authors — a suspicion of misconduct, or a conflict you noticed while reading. They are not the place for the substance of your assessment.

Your recommendation advises the editor, who decides. Editorial decisions are based on scope, scientific quality, methodological rigor, ethical compliance, reporting quality, and relevance to the journal.

Single-blind by default; double-blind review is available on request.

Accept

The work is sound and reported adequately. Rare at first review.

Minor revision

The conclusions hold; specific reporting or analysis points need addressing.

Major revision

The study may support its conclusions but the manuscript does not yet show it, or an additional analysis is needed. Say what would have to change.

Reject

The design cannot answer the question, the reporting is too incomplete to evaluate, or the work is outside the journal’s subject matter. Say which, and say it in a way the authors can use elsewhere.

If you suspect plagiarism, duplicate publication, image manipulation, fabricated or falsified data, an undisclosed interest, or an ethical problem, raise it in the confidential comments with what you observed and any evidence you have.

Do not contact the authors, and do not investigate yourself. The editorial office handles it, corresponding with the authors and, where appropriate, their institution. A concern raised in good faith that turns out to be unfounded carries no consequence for the reviewer.

Reviewers who would like to review for the journal, or to be considered for the board, are invited to write to the editorial office at [email protected] with their areas of expertise and an ORCID iD.