Tag
selection_bias
7 findings
Methods & evidenceAttanasio Selection-Bias Decomposition FrameworkVirtual-Within interaction yields a statistically significant intensive quality increase among motivated Milestone 0 completers ($\beta = 0.138, p < 0.05$ in Large-Country Sample; $\beta = 0.427, p < 0.05$ in Uganda Sample). Under the BAS assumption, this observed conditional effect serves as a strict lower bound for the true unconditional quality impact.RCTMethods & evidenceExogenous Failure AssignmentAmong the subset forced into the difficult task who failed in Round 1 (N = 558 in Sample A, N = 522 in Sample B), treated students re-attempted the difficult task in Round 2 at rates 17.6 percentage points higher in Sample A (p = 0.002) and 13.9 percentage points higher in Sample B (p = 0.064) than control group baseline rates of 36% and 51%.RCTLabour & employmentFixed-Fee Referral InvarianceIncreasing fixed finder's fees from Rs 60 to Rs 110 increased referral participation by 8 percentage points (p < 0.10) but had zero statistically significant impact on referral cognitive test scores (p = 0.865) or referral relationship composition.RCTEconomics (general)Monotonic Selection Bounds AdjustmentUnconditional reading test score differences of 2.06 points increased to an upper bound estimate of 4.36 points (0.33 to 0.66 standard deviations), and math test score differences increased from 0.77 to an upper bound of 2.52 points among vocational voucher winners.RCTEconomics (general)Non-Applicant Selection Bias in Active Labor Market EvaluationsControl application rate was 43% in Experiment 1 and job fair attendance was 5.9% in Experiment 2, establishing that recruitment messaging alters who enters the sample pool before program training impact evaluation occurs.RCTEducationStudent Sorting Bias and Baseline Score AdjustmentPrior to randomization, the within-school standard deviation of teacher mean baseline math scores was 0.382 to 0.520. Evaluating instructors using unadjusted end-of-year student test scores yielded a severely biased predictive coefficient of 0.228 (SE = 0.042), significantly overstating true instructor impact.RCTDevelopmentSocial Restriction-Driven Program Take-Up HeterogeneityOverall program take-up exceeded 70%, but Scheduled Caste women (the least restricted group) were nearly 33% more likely to attend training sessions than Muslim and Upper Caste Hindu women, demonstrating that baseline social mobility dictates intervention take-up.RCT