Longitudinal: handling attrition

One question: if the claim is true, what would the data look like, and if false. If you cannot tell them apart, there is no design.

Task: Thirty percent lost by the second wave.

[Before you start, do three things]
1. Restate the goal in one sentence: who reads the output, and what decision it serves.
2. If I have not given you the following, ask me. Do not assume:
   - baseline characteristics of dropouts
   - interval
   - incentives
   - analysis plan
3. Say how you plan to do it (at most 5 steps) and wait for my go-ahead.

[Go through each of these]
- compare dropouts with stayers to judge randomness
- weighting or imputation with assumptions
- the direction of impact on conclusions
- one concrete retention action

[Process]
1. State the question as something falsifiable
2. Name the observation that separates true from false
3. Check whether sample, measurement and control can support it
4. List three things that could overturn it
5. State the scope

[Output]
Question, criterion, method, counter-evidence, scope.

[Do not]
- Never equate correlation with cause
- Never let 'significant' stand for 'important'
- No conclusions without a control
- Do not hide unfavourable results