Journal of Big Data Research
Guidelines for reviewers
What the journal asks of a review: whether the work belongs here, whether the evidence supports the claims, and what the authors would need to do. These are two different questions and the report should keep them apart.
- Review model
- Single-blind by default; double-blind review is available on request.
- Your report
- Advice to the editor, who makes the decision
- Confidentiality
- The manuscript is a confidential document
- Full policy
- Editorial Policies
Accepting or declining an invitation
Accept where you can judge the substance of the work and can return a report in the time agreed with the editor. Where you can judge part of a manuscript but not all of it, accept and say which part — a review that states its own limits is more useful than one that does not, and the editor can pair you with a complementary reader.
Decline where a conflict exists: recent collaboration with the authors, a shared institution, a supervisory relationship, a competing manuscript in preparation, or a financial interest in the outcome. Declining is not a judgment on the work, and suggesting an alternative reviewer is helpful.
An invitation is confidential in itself. Do not share the manuscript or its existence with colleagues, and do not delegate the review to a student or a colleague without the editor's agreement; where a co-reviewer is involved with the editor's agreement, name them in the report.
Decline if
- You have collaborated with an author recently
- You share an institution or a supervisory relationship with one
- You are preparing competing work
- You have a financial or personal interest in the outcome
- You cannot judge the substance of the manuscript
- You cannot return the report within the agreed period
The two questions
Fit and evidence are judged separately
Question one
Does the work belong here?
Does it make a substantive contribution to big-data research — in theory, methods, systems, data resources, interfaces, governance, or an applied field? Eligibility does not depend on novelty, on dataset size, on a mechanism, or on demonstrated clinical benefit. Established methods and replication are in scope.
Question two
Is the claim supported?
Read the evidence against the claim the manuscript actually makes, using the standards the scope page sets out for that contribution family. A comparative claim needs a comparison; a theoretical result needs its assumptions; an applied finding needs validation suited to the setting.
Keeping them apart
Why it matters
Weak evidence is a reason to ask for more evidence, not a reason to call the work out of scope. Saying which of the two questions drives your recommendation lets the editor write a decision letter the author can act on.
What to look at
The data
Is the material described well enough to judge the work — its provenance, coverage, scale and preparation? Is there a data-availability statement, and where access is restricted, is the restriction explained? Would another group know what the method ran on?
The method
Is the method described in reproducible form, with parameters and the environment the results were produced in? Are the assumptions stated? For a theoretical result, does the reasoning support what is claimed?
The evaluation
Are training and test material kept apart? Are the baselines the ones a reader would reach for, and were they given a fair setting? Is uncertainty reported? For a causal claim, are the identification assumptions justified?
The statistics
Do the tests suit the design and the data? Are effect sizes and intervals reported rather than significance alone? Are multiple comparisons handled? Does the sample support the inference drawn from it?
The claims
Does the discussion stay within what the evidence supports? Watch for a computational result presented as an effect on patients or on an operation — that is a different claim and needs different evidence.
The reporting
Are limitations stated? Are ethics, consent and registration reported where they apply? Are figures and tables legible and self-explanatory? Are the references accurate and the prior work fairly represented?
Writing the report
Open with a short summary of what you understand the manuscript to claim and contribute. It tells the editor whether the paper communicated its own point, and it tells the authors how they were read.
Then separate major points — things that must change for the conclusions to stand — from minor ones. Number them, cite the line or section, and say what would resolve each. Where you ask for more work, say what result would satisfy you.
Write to the author, about the manuscript. Confidential remarks for the editor go in the separate field. Requests to cite your own work are appropriate only where the reference is necessary to the argument, and should be justified as such.
Recommendation
- Accept
- Sound, and the claims are supported.
- Minor revision
- Conclusions stand; specific points need clarifying.
- Major revision
- Worth pursuing, but the evidence or reporting is not yet sufficient.
- Decline
- The claims are not supportable by evidence the authors could reasonably add, or the work makes no contribution to big-data research.
The recommendation is advice. The editor weighs it with the other reports and takes the decision.
Confidentiality, files and AI
Confidentiality
The manuscript and your review are confidential. Do not share either, do not use unpublished material, and do not cite the manuscript until it is published.
Files
After submitting the review, reviewers should securely delete downloaded manuscript files and must not retain or use unpublished material.
AI tools
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.
Raising a concern
Where you suspect plagiarism, duplicate submission, image or data manipulation, or an authorship problem, tell the editor in the confidential field with the specific evidence. Do not contact the authors, and do not raise it publicly.
Revisions, and reviewing for the journal
Where you reviewed the first version, you may be asked to read the revision. Read it against your own points: were they addressed, and does the response hold? A revision is not an invitation to open new lines unless something in the changes raises them.
Researchers who would like to review for the journal are invited to write to the editorial office with a short statement of expertise and a link to their publication record. Reviewing confers no advantage in the review of the reviewer's own submissions.
Related pages
- Scope and evidence
- Aims and Scope
- Policies
- Editorial Policies
- Editors
- Guidelines for editors
- Authors
- Instructions for Authors