Methods reading room: what analytical reports actually establish · RESEARCH-26765

Case 3 — Correlation in a small observational dataset

Report sample size, potential confounders and measurement uncertainty. Association does not identify direction or causation, and exploratory results need independent evaluation.

Case 3 — Correlation in a small observational dataset

Two variables move together in a small sample. What would a careful discussion acknowledge?

Exercise RESEARCH-26765/3: address this question within its stated limits.

Case 3 — Correlation in a small observational dataset

Clarification for “Correlation in a small observational dataset”: the point beginning “Report sample size, potential confounders…” 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 3 — Correlation in a small observational dataset

Method note for “Correlation in a small observational dataset”: use “Report sample size, potential confounders…” 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 3 — Correlation in a small observational dataset

Limit of the answer in “Correlation in a small observational dataset”: “Report sample size, potential confounders…” 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 3 — Correlation in a small observational dataset

Correction note for “Correlation in a small observational dataset”: retain the wording beginning “Report sample size, potential confounders…” 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 3 — Correlation in a small observational dataset

Summary draft for “Correlation in a small observational dataset”: the answer addresses “Report sample size, potential confounders…”; the related point is “Association does not identify direction or causation, and…”. 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 4 — Units and scale: prevent a silent comparison error

Two figures use different units and one axis is logarithmic. How should a comparison be documented?

Exercise RESEARCH-26765/4: address this question within its stated limits.

Case 4 — Units and scale: prevent a silent comparison error

Record the original units and axis transformation, show any conversion explicitly and retain the original values. Do not compare apparent visual distances across incompatible scales.

Case 4 — Units and scale: prevent a silent comparison error

Method note for “Units and scale: prevent a silent comparison error”: use “Record the original units and axis transformation, show any conversion…” 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 4 — Units and scale: prevent a silent comparison error

Clarification for “Units and scale: prevent a silent comparison error”: the point beginning “Record the original units and axis transformation, show any conversion…” 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 4 — Units and scale: prevent a silent comparison error

Limit of the answer in “Units and scale: prevent a silent comparison error”: “Record the original units and axis transformation, show any conversion…” 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 4 — Units and scale: prevent a silent comparison error

Correction note for “Units and scale: prevent a silent comparison error”: retain the wording beginning “Record the original units and axis transformation, show any conversion…” 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 4 — Units and scale: prevent a silent comparison error

Summary draft for “Units and scale: prevent a silent comparison error”: the answer addresses “Record the original units and axis transformation, show any conversion…”; the related point is “Do not compare apparent visual distances…”. 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 5 — Reading an abstract without skipping the methods

A strong abstract conclusion is being discussed without the study methods. What should the reading note cover?

Exercise RESEARCH-26765/5: address this question within its stated limits.