Your role in AAAI’s peer review
AAAI uses a double-blind peer review model in which authors and reviewers are anonymized throughout the review process. To enable blinding, authors provide a separate Title Page (author details) and a blinded Manuscript file (no author details).
As a reviewer, you should avoid attempting to identify authors and should keep all manuscript content confidential.
AAAI’s Reviewer Guidelines page on the legacy OJS site exists as a navigation item but contains minimal public instructions. This updated page therefore expands the guidance using AAAI’s peer-review approach and internationally recognized reviewer-ethics resources.
Reviewer promise (plain language)
When you accept a review, you agree to keep the manuscript private, declare conflicts of interest, provide an objective assessment, and write feedback that helps authors improve the work—even when you recommend rejection.
Before you accept: fit, time, and conflicts
A strong review starts with a strong decision to accept (or decline) the invitation. Declining quickly is better than accepting and missing deadlines. Reviewers also must disclose conflicts of interest and avoid reviews where impartiality is compromised. COPE’s reviewer guidelines emphasize confidentiality and conflicts as core ethical duties.
Accept if you can answer “yes” to all of these
- Expertise I can competently assess the study design, clinical relevance, and methods used (or I can judge the manuscript within my expertise and flag limits).
- Time I can complete the review within the requested timeframe (or I can propose an alternative deadline immediately).
- No COI I have no conflict of interest that could bias my judgment (financial, personal, institutional, close collaboration, or direct competition).
- Confidentiality I can keep the manuscript and all correspondence confidential and will not use manuscript content for personal advantage.
Examples of conflicts of interest (COI)
| Personal or professional relationships | Recent co-authorship, mentor/mentee relationship, close friendship, or personal dispute with an author or team. |
|---|---|
| Institutional conflicts | Same department/institution as an author, or ongoing collaborative project between institutions that could influence judgment. |
| Financial/industry ties | Consulting fees, speaker honoraria, funding, patents, stock ownership, or product-related ties connected to the intervention/diagnostic/biologic being evaluated. |
| Direct competition | You are working on a closely related manuscript, grant, or patent and might benefit from delaying or appropriating ideas from the submission. |
If you are unsure, disclose
If you suspect a potential COI, disclose it to the editor before accepting. Editors can decide whether the conflict is manageable or requires reassignment. Transparency protects you and the journal.
How to review: a practical, step-by-step method
A helpful review is structured, evidence-based, and specific. The goal is not to “rewrite” the paper, but to assess whether the study is reliable, clinically meaningful, and clearly reported, and to guide improvements that make the work stronger for readers and patients.
Step 1: Quick scan (10–15 minutes)
- Topic and scope: Is it within asthma, allergy, immunology, respiratory health, or closely related translational/clinical domains?
- Study type: Trial, cohort, cross-sectional, lab study, systematic review, case report, brief report, clinical image, etc.
- Primary claim: What does the paper claim—and what evidence is used to support it?
- Deal-breakers: obvious ethical problems, missing consent where required, implausible data patterns, or major reporting gaps.
Step 2: Deep read (focus on methods and results)
Most reliability issues are in the methods/results chain. Spend time checking whether the design, participants, exposures/interventions, outcomes, and analysis logically support the conclusions. If you can’t reproduce the logic of the analysis from what is reported, the paper is not yet publication-ready.
Step 3: Write your report (helpful structure)
| 1) Summary (2–4 sentences) | Describe the manuscript in your own words: question, design, and main finding. This proves you understood it. |
|---|---|
| 2) Major issues (numbered) | Focus on validity, missing critical methods, ethical concerns, inappropriate statistics, and over-interpretation. |
| 3) Minor issues (bullets) | Clarity, wording, missing citations, figure labeling, small reporting corrections. |
| 4) Confidential notes to editor | Ethics concerns, suspected overlap, reviewer COI clarifications, or anything not appropriate to send to the author. |
Write like a trusted colleague
Be direct but respectful. Avoid sarcasm, personal comments, or assumptions about competence. Critique the work, not the authors. Editors may remove inappropriate language before sharing the report with authors.
What to evaluate: reviewer criteria for asthma & allergy manuscripts
AAAI publishes a wide range of respiratory and immunology research. Use the relevant checklist below depending on the manuscript type. If parts are outside your expertise, say so explicitly and focus on what you can assess well.
For original clinical research
- Research question: Is it clinically relevant and clearly stated?
- Population: Inclusion/exclusion criteria, recruitment, setting, and representativeness.
- Outcomes: Are outcomes clinically meaningful and measured appropriately (e.g., exacerbations, symptom control, lung function, biomarkers)?
- Bias/confounding: Are limitations acknowledged and mitigated?
- Analysis: Appropriate statistical tests, missing data handling, and transparent reporting.
- Conclusions: Do they match results (avoid over-claiming causality from observational designs)?
For randomized trials or intervention studies
- Randomization method, allocation concealment, blinding, adherence, and handling of protocol deviations.
- Safety monitoring and adverse event reporting (especially relevant for biologics and immunotherapy).
- Primary vs secondary endpoints: avoid “spin” or overemphasis on exploratory outcomes.
For systematic reviews / meta-analyses
- Clear question, explicit eligibility criteria, reproducible search strategy, and screening workflow.
- Risk-of-bias assessment and evidence certainty discussion.
- Appropriate synthesis method (narrative vs quantitative) and heterogeneity handling.
For case reports / clinical images
- Novelty and educational value for clinicians (why this case matters).
- Clear chronology, differential diagnosis, and rationale for management decisions.
- Consent for publication and protection of patient privacy (no identifiers).
Common red flags to note (and how to report them)
If you notice possible plagiarism, image manipulation, improbable data patterns, or missing ethics/consent statements, do not accuse the authors in harsh terms. Instead, provide evidence-based observations in confidential notes to the editor (e.g., “Figure 2 appears duplicated with Figure 3” or “Large text overlap is similar to [citation/link]”).
Confidentiality and ethics expectations for reviewers
Reviewing is privileged access to unpublished work. COPE’s ethical guidelines emphasize that reviewers should respect confidentiality and must not use information obtained during peer review for personal advantage.
The Council of Science Editors similarly notes that material under review should not be shared outside the review process without editor approval.
Do
- Keep it private: do not share the manuscript, figures, or your review with anyone outside the journal process.
- Declare COI: disclose any potential conflicts immediately (before or after accepting if you realize later).
- Be fair: judge the work on evidence and clarity, not on guesses about identity, institution, country, or language background.
- Be constructive: provide specific recommendations that help the authors improve reporting and interpretation.
Don’t
- Don’t “shop” ideas: do not use unpublished content in your own research, grant proposals, patents, or teaching materials without permission.
- Don’t contact the authors: do not reach out directly—even if you think you recognize them—because that can break blinding and fairness.
- Don’t demand citations to yourself: suggesting a citation can be appropriate only if it is truly necessary and relevant, and should never be used for personal metric gain.
- Don’t disclose identity: avoid signing your report or inserting identifying comments. AAAI uses a double-blind model.
Use of AI tools while reviewing
Do not upload or paste any manuscript text into public AI tools or external systems that store or reuse content, because that can violate confidentiality. If you use permitted tools for grammar assistance or note-taking, keep content minimal, de-identified, and within secure environments approved by your institution.
Recommendation categories and what they mean
AAAI’s editorial decision is based on editor judgment, reviewer reports, and journal priorities. You are not “accepting” or “rejecting” the paper yourself; you are advising the editor with evidence and reasoning.
| Accept (rare on first round) | Manuscript is clear, methodologically sound, and needs only trivial edits. |
|---|---|
| Minor revision | Core methods/results are sound; changes are mainly clarity, reporting details, small analyses, or better framing. |
| Major revision | Study may be valuable but requires substantial work: key methods clarifications, additional analyses, re-organization, or stronger limitations. |
| Reject | Work is not reliable, not within scope, lacks ethical compliance, or would require a fundamentally different study to support claims. |
Consistency tip
If your report contains multiple “major issues,” align your recommendation with the effort needed to correct them. A “minor revision” recommendation should not require re-running experiments or re-collecting patient data.
After you submit your review: revisions and follow-up
Re-review requests are common. If the authors submit a revised version, check whether they addressed the major concerns and whether the response letter is complete. A high-quality revision response:
- Responds to each comment point-by-point.
- Indicates where changes were made (page/line/section references).
- Explains respectfully when a request cannot be fulfilled (e.g., data not available), and updates limitations accordingly.
If you cannot re-review due to time constraints, decline quickly so the editor can proceed efficiently.
Frequently asked questions
What if I suspect I know the authors?
Do not attempt to confirm identities. Evaluate the manuscript on its merits. If you believe your suspected identity creates a conflict of interest, disclose it to the editor and request reassignment.
Can I involve a trainee or colleague in the review?
Not without explicit editor permission. Confidential material should not be shared outside the review process unless approved. COPE and CSE emphasize confidentiality and controlled sharing.
May I keep a copy of the manuscript for my records?
No. Treat the manuscript as confidential work-in-progress. Delete local copies and notes that contain manuscript text when the review concludes, unless the editor provides a formal instruction to retain records for an investigation.
What if the manuscript is poorly written in English?
Focus first on scientific validity. Provide a few examples of unclear phrasing and recommend language editing for readability. Avoid dismissing work solely due to language issues if the science is sound.
How does AAAI handle double-blind review?
AAAI states it uses a double-blind model and asks authors to submit a Title Page (author details) and a separate blinded Manuscript file.
