Balance Before Effect
Assess balance before interpreting a propensity-score adjusted causal estimate.
Estimate whether seller coaching increases 30-day GMV when coaching was assigned to sellers with stronger baseline activity. Propensity score design: model treatment probability, match or weight, check balance, then estimate effect. The common shortcut is celebrating a significant treatment effect without checking whether coached and uncoached sellers became comparable. Step 1 Model treatment probability from baseline GMV, seller age, category, listings, region, and prior growth. Use only pre-treatment variables that affect assignment and outcome. Step 2 Create weights or matches using the propensity score. The score is a design tool, not the final causal answer. Step 3 Check balance: standardized…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in