Friday, June 26, 2015

My talk in Quito - Sample size and causal effects

In this talk I introduce the concept of design-based estimation for the average treatment effect in public policy evaluation. In the final section of the talk I propose to use the samplesize4surveys R package in order to compute the sample size required to 1) properly quantify the effect of the program and 2) properly test if the program had effect over the population of interest. 

1 comment: