Appendix 3 contains additional tables with simulation results for various datasets. Each table includes statistics based on 10,000 simulations of each method, with details on methods and variables provided in Table A2. The tables include results for different sample sizes (30, 100, 300 observations) and datasets such as Sri Lanka data, IFLS school data, IFLS expenditure data, LEAPS math test score data, and LEAPS height z-score data. The simulation results include measures of power for detecting treatment effects, with specific simulated treatment effects described in each table. For example, the Sri Lanka data includes a simulated treatment effect of a 1,000 Sri Lankan Rupee increase in profits, while the ENE data includes a simulated treatment effect of a 920 Peso increase in income. The LEAPS math test score data includes a simulated treatment effect of an increase of one quarter of a standard deviation in the test score. The tables also include p-values for the difference in outcome variables at follow-up versus baseline, with figures illustrating these p-values for different datasets and sample sizes. The results show whether there is any difference in follow-up balance after controlling for baseline variables. The tables and figures provide a comprehensive analysis of the simulation results for various datasets and sample sizes.Appendix 3 contains additional tables with simulation results for various datasets. Each table includes statistics based on 10,000 simulations of each method, with details on methods and variables provided in Table A2. The tables include results for different sample sizes (30, 100, 300 observations) and datasets such as Sri Lanka data, IFLS school data, IFLS expenditure data, LEAPS math test score data, and LEAPS height z-score data. The simulation results include measures of power for detecting treatment effects, with specific simulated treatment effects described in each table. For example, the Sri Lanka data includes a simulated treatment effect of a 1,000 Sri Lankan Rupee increase in profits, while the ENE data includes a simulated treatment effect of a 920 Peso increase in income. The LEAPS math test score data includes a simulated treatment effect of an increase of one quarter of a standard deviation in the test score. The tables also include p-values for the difference in outcome variables at follow-up versus baseline, with figures illustrating these p-values for different datasets and sample sizes. The results show whether there is any difference in follow-up balance after controlling for baseline variables. The tables and figures provide a comprehensive analysis of the simulation results for various datasets and sample sizes.