Tag
adaptive_experiments
5 findings
Economics (general)Counterfactual Re-Randomization InferenceGenerated exact finite-sample p-values controlling for time-varying allocation history without relying on asymptotic normality, verifying that short-term 6-week employment differences across all treatment arms had p-values well above conventional significance thresholds (p > 0.12).Expert TheoryEconomics (general)Frequentist Variance Bounding in Adaptive TrialsBounded the worst-case standard error of estimated average treatment effects at $\sqrt{2\lambda} = 0.05$ (equivalent to 80% power for detecting effect sizes of 0.124), resulting in an actual empirical standard error bound of 0.016 in the trial.RCTEconomics (general)Myopic Stopping Rules for Adaptive Field ExperimentsCalculates wave-by-wave net return by summing expected participant welfare net of costs (weighted by treatment assignment shares $(1-\gamma)\hat{p}_{t}^{dx} + \gamma/k$) and the marginal decrease in treatment effect estimator variance. The experiment continues if and only if expected net return is greater than zero.RCTEconomics (general)Statistical Surrogacy in Sequential Policy TrialsThe surrogacy condition held statistically for 2-month employment predicting 4-month employment (Wald test p = 0.575 for Syrians, p = 0.527 for Jordanians). Targeting 2-month employment in counterfactual simulations doubled the employment gains compared to a standard RCT (13.4% vs 12.4%), whereas targeting noisy 6-week employment yielded zero adaptive gains.Observational StudyEconomics (general)Tempered Thompson AlgorithmAt gamma = 0.2, the minimum allocation probability for any of the 4 treatment arms is bounded below at 5%. In counterfactual simulations using 2-month employment outcomes, the algorithm doubled the employment gains compared to a standard randomized controlled trial.RCT