{"blocks": [{"key": "18c9540d", "text": "Scenario", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "6bc637e3", "text": "Reviewing a completed observational study that used Propensity Score Matching (PSM) to estimate the impact of a UI change on watch time.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "8f4a4032", "text": "Question", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "3031757a", "text": "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.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "fcc12ee7", "text": "Hints", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "9d9db9b5", "text": "Discuss overlap, balance diagnostics, causal estimands, and robust variance.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}], "entityMap": {}}