Observational Causal Study: Reminder Program With Staggered Market × Channel Launch

Context

You are evaluating the causal impact of medication-subscription reminders on two outcomes:

• User experience (CSAT on a 1–5 scale)
• 4-week retention

The reminder product launched at different times across markets and channels (push/email/SMS). You cannot randomize who receives reminders. Users can opt out of certain channels (e.g., push). Adoption may spill over within households. Design an observational causal analysis that leverages staggered rollouts.

Tasks

Answer precisely:

1. Identification strategy and model
   • State your primary identification strategy (e.g., staggered-adoption DID with event study).
   • Write the regression(s) you would estimate, including fixed effects, time trends/seasonality, and how you will cluster standard errors.
2. Treatment definition and risk sets
   • Define treatment and risk sets to avoid immortal-time bias.
   • Explain how you handle not-yet-treated users.
3. Diagnostics and modern DID
   • Specify pre-trend diagnostics.
   • Explain how you would detect and mitigate treatment-effect heterogeneity bias in TWFE.
   • Choose a modern DID estimator (e.g., Sun–Abraham or Callaway–Sant’Anna) and justify.
4. Matching/weighting backup plan
   • Propose a backup plan using matching or weighting (e.g., PSM or overlap weighting).
   • List covariates needed, caliper/ratio, and balance metrics/thresholds.
5. Negative controls and falsification
   • Propose two negative controls (one outcome, one exposure) and one falsification test.
6. Threats and remedies
   • Address household spillovers/interference, missing CSAT, and channel selection (users can opt out of push).
7. Power/MDE outline
   • Provide a minimal power/MDE calculation outline with assumptions on baseline variance, intra-household correlation, and expected adoption rate.