Table 1 Framework for assumptions on missing data
From: A framework for handling missing accelerometer outcome data in trials
Primary | Key sensitivity | Other sensitivity nalyses: | |||
|---|---|---|---|---|---|
analysis: | analysis: | ||||
Decisions | Plausible | Suspicious | Plausible-no-aux | Replace-days | Dismissive |
Use day-substitution? | ✗ | ✗ | ✗ | ✓ | ✗ |
Impute as censored? | ✓ | ✓ | ✓ | ✓ | ✗ |
Assume MAR? | ✓ | ✗ | ✓ | ✓ | ✗ |
Use auxiliary variables? | ✓ | ✓ | ✗ | ✓ | ✓ |