powerLATE implements the generalized power analysis for the local average treatment effect (LATE), proposed by Bansak (2020). A comprehensive tutorial on using this package can be found here.
Power analysis is in the context of estimating the LATE (also known as the complier average causal effect, or CACE), with calculations based on a test of the null hypothesis that the LATE equals 0 with a two-sided alternative. The method uses standardized effect sizes to place a conservative bound on the power under minimal assumptions. powerLATE allows users to recover power, sample size requirements, or minimum detectable effect sizes. It also allows users to work with absolute effects rather than effect sizes, to specify an additional assumption to narrow the bounds, and to incorporate covariate adjustment.
You can install the released version of powerLATE from CRAN with:
Or the development version from GitHub:
library(powerLATE)
#> powerLATE: Generalized Power Analysis for LATE
#> Version: 0.1.2
#> Reference: Bansak, K. (2020). A Generalized Approach to Power Analysis for Local Average Treatment Effects. Statistical Science, 35(2), 254-271.powerLATE provides two main functions:
powerLATE(), which computes the power of the Wald IV estimator, or determines parameters (e.g. required sample size) to obtain a target power.
powerLATE.cov(), which is similar to powerLATE() but additionally allows the inclusion of covariates.
For a comprehensive tutorial on conducting a LATE power analysis with this package, see here.
For more examples on how to use the package, see here.
For a detailed description of the method see: