For Reviewers
1. Purpose and Review Model
Data & Science thanks scholars who contribute independent, careful and constructive peer review. The journal operates double-blind peer review: authors and reviewers are not informed of each other’s identities during assessment, insofar as this can be maintained. Manuscripts selected for external peer review receive at least two independent expert reports before acceptance.
The Editorial Office coordinates invitations and correspondence. The Editor-in-Chief or an appropriately delegated academic editor is responsible for substantive editorial decisions. Reviewer recommendations are advisory, not binding.
2. Who May Serve as a Reviewer
Reviewers must have relevant expertise, a credible record of research or professional scholarship, and the ability to assess the particular manuscript impartially. A doctoral qualification or equivalent demonstrable expertise is normally expected. Reviewers should have recent relevant scholarly engagement and a verifiable institutional or professional affiliation.
As a safeguard for independence, reviewers should not be from the authors’ institution or have coauthored publications with them during the preceding three years. Reviewers must also disclose other circumstances that could reasonably affect independence. The handling academic editor may apply stricter safeguards where necessary.
An invitation to review does not authorize a reviewer to delegate the assessment to another person. A proposed co-reviewer must be approved by the journal before being given access to the manuscript.
3. Invitation and Timeliness
Please accept or decline an invitation promptly after examining the abstract and subject area, and before accessing sensitive materials where possible. Decline if the topic is outside your expertise or if you cannot provide a substantive report within the requested period.
If you require an extension, notify the Editorial Office as early as possible. If declining, you may suggest appropriately qualified alternative reviewers without assuming that they will be appointed. Review deadlines are indicated in the invitation and may be adjusted to support rigorous assessment.
4. Conflicts of Interest and Impartiality
Disclose financial, personal, academic, institutional, professional, competitive, or intellectual interests that might influence—or appear to influence—your judgment. Examples include a recent collaboration, joint grant, direct supervisory relationship, close friendship or dispute, direct financial stake, or competition for the same result.
If you identify a likely author despite anonymization, do not attempt to investigate the identity further. Inform the handling editor if this creates a conflict. If in doubt whether a relationship is relevant, disclose it and allow the editor to decide whether the review may proceed.
Assess the manuscript on its merits regardless of the author’s identity, nationality, affiliation, gender, beliefs, or ability to pay publication charges. Do not request irrelevant citations to your work, the journal’s work, or that of associates for citation-boosting purposes.
5. Confidentiality and Use of Manuscript Material
Manuscripts, figures, tables, datasets, supplementary files, identities, editorial communications, and unpublished findings obtained through peer review are confidential. Do not share, circulate, reproduce, post online, use for personal research or competitive advantage, or discuss them with unauthorised parties.
Do not contact authors directly about a manuscript under review. Use the journal’s designated review channel. Preserve anonymity in the report text and in document metadata. Where an approved co-reviewer contributes, identify that person privately to the editor and ensure equivalent confidentiality.
Review reports are confidential unless a separate, voluntary open-peer-review arrangement is agreed. A report may be published alongside an accepted article only with express consent from the relevant author(s) and the individual reviewer. A reviewer’s name is published only if that reviewer separately agrees.
6. Use of Artificial Intelligence by Reviewers
Peer review is a human scholarly responsibility. Reviewers must not use generative AI to generate substantive evaluations, conclusions or recommendations. Under no circumstances may a reviewer upload, paste or otherwise transmit any part of a confidential manuscript, figure, table, dataset, editorial correspondence or identifiable review content to an external generative-AI service.
Basic local spelling and grammar checking may be used. Any other limited AI-assisted language polishing of reviewer-authored, non-confidential wording requires prior editorial permission and disclosure to the handling editor and must not transmit confidential information. The reviewer remains accountable for every statement in the report.
If misuse of AI or breach of confidentiality is suspected, the editor may disregard the report and take appropriate follow-up action. The journal’s advisory AIGC percentage for manuscripts is not a reviewer acceptance or rejection rule.
7. Reviewing the Scholarship
Evaluate the work in light of its article type, research question, and contribution to the journal’s scope. A useful review considers: (a) relevance and originality; (b) significance and relationship to prior research; (c) validity of design and methods; (d) appropriateness of data, analyses, assumptions and interpretation; (e) transparency, reproducibility and ethics; (f) clarity and organization; and (g) whether conclusions are supported by evidence.
For empirical or computational work, assess dataset suitability, sampling, analytic design, robustness, replicability, and limitations. For qualitative research, consider appropriateness of design, evidence, reflexivity and transparency. For reviews, evaluate the search and selection strategy where applicable, coverage, synthesis and intellectual contribution. For data/resource papers, consider documentation and potential for meaningful reuse.
Consider ethics approval, informed consent, data availability, conflicts of interest and authorship or AI disclosures if presented. Reviewers need not repeat the journal’s automated similarity-screening procedure, but should report apparent plagiarism, duplicate publication, manipulated images, fabricated data, unreliable citations, or undisclosed use of AI when supported by identifiable concerns.
The journal requires manuscript similarity scores to be strictly below 20%. AIGC detection scores below 25% are only an advisory reference; the AIGC score alone must never be used as a reason to recommend acceptance or rejection.
8. Preparing a Constructive Review Report
Write the report in English, using professional and evidence-based language. A helpful report normally includes:
- A short, accurate summary of the manuscript’s main question, approach and contribution.
- An overall assessment of novelty, significance, strengths and principal weaknesses.
- Major comments focused on substantive issues, linked to sections, figures, tables, or line numbers where possible.
- Minor comments that improve clarity, reporting, presentation, and precision.
- A reasoned recommendation selected from the options available in the system.
Provide specific and feasible suggestions. Distinguish essential corrections from optional improvements. Refrain from hostile, disparaging, discriminatory, or personally critical remarks. Do not insist on unnecessary experiments, methods, datasets, or citations that do not materially improve the work.
Any confidential concerns that should not be communicated directly to the authors—for example, evidence of an integrity issue—should be communicated privately to the handling editor through the appropriate system field.
9. Recommendations and Revisions
Reviewers may recommend: Accept; Minor Revision; Major Revision; or Reject. A recommendation should be supported by the substantive assessment in the report. The handling academic editor, not the reviewer or editorial office, makes the final decision.
If revisions are requested, reviewers may be invited to assess the response and revised manuscript. Pay particular attention to whether essential concerns have been resolved; do not continually introduce unrelated new requirements without justification. The journal normally allows no more than three major-revision rounds, with discretion in exceptional cases.
Authors are informed of the specific revision deadline in the editorial decision letter. Reviewers should not impose independent fixed deadlines or insist on withdrawal solely because a substantial revision requires additional time.
10. Research Misconduct and Ethical Concerns
If you encounter potential plagiarism, fabrication, falsification, inappropriate image or citation manipulation, duplicate submission, unethical research, or undisclosed conflicts, report the concern privately and promptly to the editor with the evidence available. Do not conduct an independent investigation, contact institutions or authors, or publicize an allegation.
The journal considers concerns in accordance with its Editorial Policies, provides an opportunity for an appropriate response, and may seek further expert or institutional input. Reviewers are expected to cooperate confidentially if follow-up clarification is requested.
11. Recognition and Contact
The journal values reviewer contributions and may acknowledge service or provide certificates where administratively supported. Such recognition does not influence manuscript decisions, nor does it entitle reviewers to editorial appointment or indexing recognition. Any disclosure of reviewer identity or service must be authorized by the reviewer.
For questions about an invitation, a conflict of interest, confidentiality, report submission or possible misconduct, use the editorial system or contact authorsupport@cscholar.com, citing the manuscript identifier. See https://ds.cscholar.com/editorialpolicies for the journal’s full policies and https://ds.cscholar.com/about/privacy for personal-data handling.