Evaluate Propensity Score Matching Alternatives and Diagnostics
Company: Netflix
Role: Data Scientist
Category: Statistics & Math
Difficulty: hard
Interview Round: Other
##### Scenario
Reviewing a completed observational study that used Propensity Score Matching (PSM) to estimate the impact of a UI change on watch time.
##### Question
Why is standardized mean difference (SMD) ≤ 0.1 often used as a balance threshold in PSM? If logistic regression is not appropriate for the propensity model, what alternatives would you consider and why? How would you diagnose residual confounding after matching? Describe one method to estimate treatment effect variance under PSM.
##### Hints
Discuss overlap, balance diagnostics, causal estimands, and robust variance.
Quick Answer: This question evaluates causal inference skills focused on propensity score methods, balance diagnostics, overlap/positivity assessment, estimand selection (ATE vs ATT), and variance estimation in observational studies.