Observational studies · manuscript crosswalk
How to complete a STROBE checklist: map manuscript locations and reporting gaps
Entering a page number does not establish complete reporting. Fix the manuscript version, link each requirement to a passage and location, and distinguish not located, planned only and genuinely not applicable. Here is a reviewable crosswalk workflow.
This guide concerns reporting of cohort, case–control and cross-sectional studies. Select the design-appropriate checklist from STROBE and check current journal instructions and applicable extensions. STROBE supports transparent reporting of what was planned, done and found; it is not a study-design prescription, quality score or permission to submit.
Manuscript v1, M2–M4 and the table below are original synthetic teaching material. No AI run, patient-data analysis or customer-manuscript review was performed. Four items are illustrated, not all 22 items and their subitems.
Prepare three sets of materials
Applicable checklist and scope
Keep its name, study design, download date and version. Address asterisks, group-specific requirements and subitems individually; similar guideline names do not establish applicability.
Stable manuscript locations
Supply an authorised manuscript v1 and supplements. Keep section names, paragraph IDs and table/figure numbers. For PDF locations, record the file version and physical page. If only an abstract is available, limit the review to it.
Actual conduct and author-owned facts
To assess whether a method was carried out, also supply the plan and verified outputs. Responsible people confirm ethics, permission, funding, authorship and conflicts. Do not give AI patient-level rows or infer that an unreported procedure was not performed.
Synthetic passages: read the evidence first
Manuscript v1, M2: “Outcome Y is readmission within 12 months; exposure X is the baseline scale score.”
M3: “Exposure X came from the enrolment questionnaire.” M4: “We plan to use multiple imputation for missing data.”
No funding statement, analysis outputs, questionnaire version or funding records were supplied. Do not invent them. M2 and similar labels are paragraph IDs for this example, not a STROBE-mandated section order.
Output: item–location–status–gap crosswalk
This is an editorial location draft, not an official model form or a pass certificate. Item 7 concerns variables, 8 sources and measurement, 12(c) missing-data handling, and 22 funding sources and funder roles.
| Item | Manuscript location | Visible evidence | Reporting status | Next step and responsibility |
|---|---|---|---|---|
| 7 | v1 / M2 | Brief outcome and exposure definitions | Partly reported | Authors verify outcome ascertainment, scale definition and other analysis variables; do not guess absent definitions. |
| 8 | v1 / M3 | Only “enrolment questionnaire” | Partly reported | Authors verify the source, assessment method and between-group comparability where applicable. |
| 12(c) | v1 / M4 | Multiple imputation is a plan | Planned only; conduct unverified | Researchers/statisticians confirm actual handling and evidence; do not rewrite this as completed imputation. |
| 22 | v1 / not located | No funding passage in supplied material | Not located; author confirmation needed | Authors confirm funding and funder roles, including the original study where applicable; do not write no funding or not applicable. |
Partly reported, not located, planned only, pending and not applicable are suggested editorial labels, not official STROBE scores. One item may need several locations. A keyword hit does not establish coverage of every requirement.
Map evidence, not keyword matches
- Expand each item and subitem before locating evidence. Explain design- and group-specific applicability instead of entering one general page number.
- Keep a short exact quote and a complete locator for every judgement. An item can point to several passages, tables or supplements. STROBE does not prescribe a uniform location or order.
- Separate what the manuscript reports from what the study actually did. Absence of text is not proof of nonperformance, and a plan is not proof of execution. Without outputs, retain unverified conduct.
- Researchers must justify not applicable from the design. Missing information, an unsuccessful search or work not yet performed is not automatically not applicable.
- After author-confirmed changes, create v2 and relocate affected passages, figures, tables and pages. Keep before/after records and unresolved items; do not reuse stale page numbers.
A complete AI request: locate before revising
Use only my authorised manuscript [file/version], supplements, applicable STROBE checklist [design/version/access date], and verified plan/outputs. For each item and subitem provide: requirement, applicability rationale, section/paragraph ID/table/figure or PDF physical page, short exact quote, reported content, reporting gap, execution status, proposed revision, responsible person and pending questions. Bind every judgement to visible evidence; state when only an abstract was read. Not located is not not performed; planned is not completed. Do not turn missing information into not applicable or invent definitions, statistics, results, ethics, funding or authorship facts. Allow multiple locations per item. Deliver the crosswalk and gaps before evidence-based wording suggestions. Do not provide a quality score, certification or submission promise. Materials: [manuscript, supplements, checklist, permission scope, necessary aggregates/plan/actual outputs].
Human checks before handoff
- Can each location be reproduced in the same version, with an exact quote?
- Are subrequirements, design and groups covered rather than merely keywords?
- Are not located, partial reporting, plans and actual conduct separate?
- Are not-applicable decisions and author-governance facts explicitly supported?
- After revision, were pages, abstract/text/tables/figures and affected items checked again?
Frequently asked questions
Do 22 completed rows prove study quality?
No. Reporting completeness is not design, data or statistical validity, peer review or a guarantee of acceptance.
May AI write “none” when funding or ethics text is absent?
No. Ask the authors to confirm. Check item 22 separately from journal and ethics requirements; not supplied does not mean nonexistent.
Must evidence appear in a prescribed section?
STROBE specifies no uniform location or order. Follow the journal form, retain versioned reproducible locators and check them again after edits.
Verified sources
The official sources below were checked on 2026-10-18. Access conditions and update dates are rechecked before publication.
- STROBE: design-specific checklists
Checked design-specific and fillable versions; this page is not an official translation or certification.
- STROBE Explanation and Elaboration: scope and items 7, 8, 12(c), 22
Checked flexible locations/order, reporting versus conduct, and the limits of the four examples.
From materials to reviewable outputs
Bring a manuscript to RigorMed for a location-and-gap task
Describe this research-manuscript task in the browser workspace, supply authorised text and the applicable checklist, and adapt the request above. Check actual task support, required materials and credit notices before starting. Authors still review every output; automatic official-checklist completion or submission permission is not promised.