Conservative Attrition Bounding and Imputation
Longitudinal panel studies over multi-year horizons face attrition risks that could bias long-term earning estimates if unobserved dropouts systematically differ across treatment and control arms.
Picture this
Imagine testing a health program over four years where some participants move away and stop answering calls—to prove the program works even under the worst-case scenario, researchers assume all missing control participants suddenly got high-paying jobs while missing treated participants got low-paying jobs, verifying if the positive effect still holds true.
What the evidence says
Long-run wage gains remained statistically significant and economically meaningful (>10% of control mean) across all differential attrition scenarios, retaining a positive point estimate even under the most conservative assumption of imputing a full 0.5 SD penalty to missing treated units.
- Who
- **N = 3,052 panel participants** tracked across a 4-year follow-up period in Addis Ababa, Ethiopia.
- How
- Sensitivity analysis imposing differential attrition bounds (Lee bounds and Karlan-Valdivia/Blattman imputation up to $\pm 0.5$ standard deviations of control outcomes) [2, 5].
What to do
Apply conservative differential mean and standard deviation imputation bounds to longitudinal trial data to confirm treatment effect robustness against attrition.
From the source
"However, we are able to estimate economically large and statistically significant effects of the workshop in the large majority of cases. For instance, the size of the effect is above 10 percent of the control group mean in all simulations but one."
Anonymity or Distance? Job Search and Labour Market Exclusion in a Growing African City