RMP.7:8 - Common Anti-Patterns and How to Avoid Them

Failure	What changes in practice
More reports are treated as more independent studies.	Recover the underlying contributions and model or remove double counting while retaining supplementary information.
Everything numerical is pooled.	Recover the target and transformation; use a supported narrower synthesis when they do not justify the requested aggregate.
Only significant findings count as agreement.	Synthesize the relevant estimates or other justified contribution, retaining uncertainty and missing evidence.
A theme becomes a population percentage.	Preserve the qualitative interpretation and use actual sampling and measurement support for a prevalence claim.
A single confidence label covers the whole corpus.	Tie appraisal and uncertainty to the finding and the approach whose criteria supply that judgment.
Disagreement prevents any useful answer.	State the supported distinctions or the precise refusal, including what remains unresolved.