Cluster assumptions worksheet
StatMedAnalysis | Version 2026-10-31

Purpose: Record the design and evidence behind a cluster approximation; this is not a power calculation.

Instructions: Save a working copy, fill each prompt, and retain source locations. Use NOT REPORTED, NOT APPLICABLE or UNRESOLVED rather than an ambiguous blank. Do not enter patient identifiers. This worksheet is an original planning aid, not a validated instrument.

1. Study question, outcome and assessment time
Response: ______________________________

2. Allocation unit and observation unit
Response: ______________________________

3. Number of clusters by arm
Response: ______________________________

4. Planned participants per cluster; equality or variation
Response: ______________________________

5. Source and population for ICC
Response: ______________________________

6. ICC estimate, uncertainty and plausible sensitivity range
Response: ______________________________

7. Total participants before loss
Response: ______________________________

8. Simple equal-cluster design effect, if applicable
Response: ______________________________

9. Expected cluster loss versus individual loss (separate)
Response: ______________________________

10. Unequal size, repeated measures, stratification or other complications
Response: ______________________________

11. Appropriate power method and small-cluster considerations to review
Response: ______________________________

12. Assumptions owner, independent checker, version and resolution
Response: ______________________________

Next action / owner / due date: ______________________________