Tables I and II in Haushofer and Shapiro (2016) erroneouslycalculated FWER-adjusted p-values using only 1,000 bootstrap iterations. The corrected tables report bootstrapped p-values with 10,000 iterations. Table III in Haushofer and Shapiro (2016) reporting spillover results for summary indices erroneously weighted the index of psychological well-being with household and village weights. The corrected version removes village and household weights. Table III also erroneously omitted observations with missing outcome variables when calculating differential attrition rates for Lee treatment effect bounds. The corrected version includes the entire original sample. Table III reports slightly different standard errors for the Lee bounds from the original publication as a result of bootstrapping. The corrected versions of Tables I, II, and III are included below as Tables I, II, and III, respectively. (printed in green online) indicates changes due to increasing the corrected p-value bootstrap iterations to 10,000. (printed in red online) indicates changes due to the removal of weights. (printed in orange online) indicates changes due to the inclusion of the entire sample to calculate Lee bounds. (printed in blue online) indicates changes due to pseudo-random results from bootstrap estimation.
|Original language||English (US)|
|Number of pages||4|
|Journal||Quarterly Journal of Economics|
|State||Published - Nov 1 2017|
All Science Journal Classification (ASJC) codes
- Economics and Econometrics