Methods reading room: what analytical reports actually establish · RESEARCH-09337

Case 20 — Exploratory subgroup analysis: keep the label visible

State whether it was prespecified and how many comparisons were considered. Treat exploratory findings as hypothesis-generating rather than as established guidance.

Case 20 — Exploratory subgroup analysis: keep the label visible

Clarification for “Exploratory subgroup analysis: keep the label visible”: the point beginning “State whether it was prespecified and how…” belongs to the question as written. It should not be read as a report that a new check, purchase or experiment has been completed. Review focus: scope. Keep the stated question separate from any broader claim.

Case 20 — Exploratory subgroup analysis: keep the label visible

Method note for “Exploratory subgroup analysis: keep the label visible”: use “State whether it was prespecified and how…” as a row heading in a proposed working record. Put the statement, its supporting source or measurement, and any unresolved detail in separate fields. This proposes a format; it supplies no additional evidence. Keep the stated question separate from any broader claim.

Case 20 — Exploratory subgroup analysis: keep the label visible

Limit of the answer in “Exploratory subgroup analysis: keep the label visible”: “State whether it was prespecified and how…” is an excerpt from the discussion, not an independent verification. A reader needs the relevant original material before treating an interpretation as established. Review focus: scope. Keep the stated question separate from any broader claim.

Case 20 — Exploratory subgroup analysis: keep the label visible

Correction note for “Exploratory subgroup analysis: keep the label visible”: retain the wording beginning “State whether it was prespecified and how…” in the record. If later material changes it, add the date, the changed wording and the source for that change; do not invent a completed correction here. Keep the stated question separate from any broader claim.

Case 20 — Exploratory subgroup analysis: keep the label visible

Summary draft for “Exploratory subgroup analysis: keep the label visible”: the answer addresses “State whether it was prespecified and how…”; the related point is “Treat exploratory findings as hypothesis-generating rather…”. Keep both connected to the opening question, and mark any conclusion needing further records as unresolved. Review focus: scope. Keep the stated question separate from any broader claim.

Case 21 — Research figure captions: tell the reader what was measured

A chart caption names a mechanism but omits the observed endpoint. How can the caption be made precise?

Exercise RESEARCH-09337/21: address this question within its stated limits.

Case 21 — Research figure captions: tell the reader what was measured

Describe the measurement, population or model, units and uncertainty shown. Interpretive discussion belongs alongside the evidence, not in place of it.

Case 21 — Research figure captions: tell the reader what was measured

Clarification for “Research figure captions: tell the reader what was measured”: the point beginning “Describe the measurement, population or model,…” belongs to the question as written. It should not be read as a report that a new check, purchase or experiment has been completed. Review focus: scope. Keep the stated question separate from any broader claim.

Case 21 — Research figure captions: tell the reader what was measured

Limit of the answer in “Research figure captions: tell the reader what was measured”: “Describe the measurement, population or model,…” is an excerpt from the discussion, not an independent verification. A reader needs the relevant original material before treating an interpretation as established. Review focus: scope. Keep the stated question separate from any broader claim.

Case 21 — Research figure captions: tell the reader what was measured

Method note for “Research figure captions: tell the reader what was measured”: use “Describe the measurement, population or model,…” as a row heading in a proposed working record. Put the statement, its supporting source or measurement, and any unresolved detail in separate fields. This proposes a format; it supplies no additional evidence. Keep the stated question separate from any broader claim.

Case 21 — Research figure captions: tell the reader what was measured

Correction note for “Research figure captions: tell the reader what was measured”: retain the wording beginning “Describe the measurement, population or model,…” in the record. If later material changes it, add the date, the changed wording and the source for that change; do not invent a completed correction here. Keep the stated question separate from any broader claim.

Case 21 — Research figure captions: tell the reader what was measured

Summary draft for “Research figure captions: tell the reader what was measured”: the answer addresses “Describe the measurement, population or model,…”; the related point is “Interpretive discussion belongs alongside the evidence, not…”. Keep both connected to the opening question, and mark any conclusion needing further records as unresolved. Review focus: scope. Keep the stated question separate from any broader claim.

Case 22 — Data provenance: keep source, transformation and output together

A table has been copied into a new format without its source or transformation notes. What should accompany it?

Exercise RESEARCH-09337/22: address this question within its stated limits.

Case 22 — Data provenance: keep source, transformation and output together

Keep the source citation, retrieval date and a record of conversions or exclusions. Make synthetic examples explicit and never combine them silently with observed data.