Statistical Review
Pubrica provides statistical review of manuscripts for journals commissioning a statistical referee and for authors preparing to submit. Our statistical reviewers assess study design, analysis, multiplicity, unit of analysis and reporting. Errors such as a misspecified model, an analysis that ignores clustering or a subgroup finding presented as pre-specified are not addressed by language editing, and they are what a methods referee identifies. Journals receive a referee report with a recommendation, and authors receive a ranked list of findings to address before submission.
- Read by a statistician, against the design and the data, not against a style guide
- Common causes of rejection: unit of analysis, multiplicity, missing data handled by assumption, and effect sizes without precision
- Pre-specification checked where a protocol or registration exists, because a post-hoc analysis presented as planned is the finding that does most damage
- Reports for journals and authors: a referee report for a journal, or a ranked list of findings to address for an author
- Reporting checklist applied: CONSORT 2025, PRISMA 2020 or STROBE, depending on the study type
Design and analysis assessedIncluded
Multiplicity and unit of analysisIncluded
Reporting checklist appliedIncluded
Written report to your formatIncluded
Deciding the paper’s fateThe editor’s
Manuscript → design → analysis → reporting → report
- Read by a statistician
- Design assessed before analysis
- Pre-specification checked where it exists
- Findings ranked by severity
- Reporting checklist applied
Service Scope
Four Components of Every Statistical Review
Each review assesses one manuscript and its design, which determines what the analysis can claim.
From the reviewA sophisticated analysis of a design that cannot answer the question is the most common serious problem, and statistical technique cannot correct it.
From the findingsAnalysis treated 480 measurements from 60 patients as independent.Eight measurements per patient. A mixed model or GEE is required; the reported confidence interval is far too narrow as analysed. This is a rejection-level finding.
From the checkThe subgroup analysis showed a significant benefit (p=0.04).One of fourteen subgroup comparisons, no multiplicity adjustment, not pre-specified in the registration. Reportable as exploratory and hypothesis-generating; not as a finding.
From the deliveryThe same analysis produces two documents. A journal needs a judgement, and an author needs a list of changes to make.
Samples are illustrative and anonymised. Hover any one of them to play it again.
Compliance and Standards
Statistical Reporting Standards and Guidelines
Where a journal specifies a reporting checklist, that governs.
Reporting guidelines (CONSORT 2025, PRISMA 2020, STROBE, TRIPOD+AI)
Which checklist the study type triggers, and whether the manuscript satisfies it. CONSORT and SPIRIT both moved to 2025 versions last year.
Statistical principles with estimands (E9(R1), 2019)
Where the manuscript reports a trial, whether the estimand is stated and whether intercurrent events were handled as pre-specified.
ICMJE Recommendations (authorship and conduct)
Authorship criteria, editorial independence, peer review conduct and the declarations an article carries.
Quality management (cert. H3CR9TCU20251212INDQM10N)
How the work is run, recorded and released, and how confidentiality is maintained.
Responsible Use of AI
Independent Statistical Review by a Human Statistician
A statistical review provides an independent assessment of the analysis by a qualified statistician. Automating that assessment would remove its purpose.
Why we do not publish our own AI accuracy figures
Accuracy figures for AI tools are usually reported by the companies that sell them.
Pubrica therefore does not quote its own accuracy figures, as they would carry the same conflict of interest. Instead, we tell you which tool was used on your project, at which stage, and what a specialist changed afterwards.
Methodology
How Our Statistical Review Process Works
We assess design first, then analysis, then reporting. Reviewing in the reverse order tends to find minor problems and miss the major one.
Design and question
Whether the design can answer the question, whether the sample supports the claim, and what the study was powered for.
Analysis
Model specification, unit of analysis, missing data handling, multiplicity and whether the methods match what the results report.
Pre-specification and reporting
Checked against the protocol, the analysis plan or the registration where they exist, and against the reporting checklist the study type triggers.
Report
Findings ranked by severity — what makes the paper unpublishable, what needs revision and what is presentational — in the format the commissioner needs.
Our Experts
Your Statistical Review Team
Your project is handled by a named team that stays with it from scoping to delivery. The handover records which team member completed each stage, and your project coordinator introduces the team by name before work begins.
Reviews the scientific content for accuracy and confirms that each conclusion is supported by the evidence. Matched to your field rather than allocated from a rota.
Checks that statistical methods, tables and conclusions are consistent with the analysis plan and that the analysis can be reproduced.
Prepares, cleans and documents the data so that every step from raw data to final output can be traced and reproduced.
Independently checks code, tables and outputs against the analysis plan. Discrepancies are logged and resolved before delivery.
Packages
Statistical Review Packages
Cost depends on study type, analysis complexity and the type of report required.
Statistical referee report
Journals commissioning a statistical review
Quoted per projectOn study type and complexity
Delivery To your review deadline
- ✓Design, analysis and reporting assessed
- ✓Findings ranked by severity
- ✓Reporting checklist applied
- ✓Referee report in your format with a recommendation
- ✓Follow-up review of the revision where wanted
Your editor makes the decision, supported by a reliable methods assessment.
Service Scope Request a quotePre-submission statistical review
Authors wanting a second opinion before submitting
Quoted per projectOn study type and complexity
Delivery By scope
- ✓The same assessment, reported for authors
- ✓Ranked list of what to fix before submission
- ✓What to reframe rather than reanalyse
- ✓Reporting checklist gaps identified
- ✓Reanalysis quoted separately where it is warranted
A pre-submission review costs less than revising after a rejection, particularly one received after six months under review.
Service Scope Request a quoteSend the methods and results sections and we will report the three most serious findings free of charge, so you know before submission whether there is a rejection-level problem.
FAQs
Frequently Asked Questions About Statistical Review
If your question is not answered here, ask an expert — no quote required.
Unit of analysis errors — clustered or repeated measurements analysed as independent — followed by unadjusted multiplicity and subgroup findings presented as though they were pre-specified.
These are followed by effect sizes reported without confidence intervals and missing data handled by an unstated assumption.
Yes, and the assessment is identical. What differs is the report: a journal wants a judgement and a recommendation, an author wants a ranked list of what to do about it.
We never review the same manuscript for both. Where that would arise, we decline the second engagement.
Not as part of the review. The review assesses what was done and whether it supports the conclusions.
Where the review concludes that the analysis is wrong, reanalysis is quoted separately. For a journal, this would be a separate engagement with different independence arrangements.
A subject peer reviewer assesses the science; a statistical referee assesses whether the methods support the conclusions. Many journals commission both, and many do not commission the second at all.
This gap is one reason methods errors reach publication, and it is why statistical review is a distinct role.
Related services
Related Academic Editorial Services
Statistical review sits between the analysis and the editorial decision.
1THE ANALYSIS ITSELFif it needs work
BiostatisticsAnalysis from the locked dataset. Statistical analysis planEstimands framed before lock.2FOR JOURNALSwho commissions it
Editorial office servicesPeer review and journal operations. Pre-submission peer reviewSimulated review, for authors.3IF SOMETHING IS WRONGafter publication
Research integrity casesCorrections and retractions, to COPE. Forensic and quality auditIntegrity and documentation review.Request a Free Statistical Review of Your Methods and Results
We will report the three most serious findings free of charge, so you can identify any rejection-level problem before submission rather than six months later.