You want to be powered to measure a 0.4 standard deviation change in normalized math scores. How many students would you need in total in order to measure an effect of this size assuming that you can administer tests in 36 schools? Calculate your answer assuming power of 80%, an alpha of 0.05, even numbers of students in the treatment and control groups, and the same number of students in each school on average. Assume that you do not take baseline covariates into account and that take-up is 100%. If using Stata, use the "power, cluster" command. If using R, use the "twomeans" and "clustered" commands available as part of the exercise materials. HINT: Normalized math scores are given by the variable "pre_mathnorm" in the Balsakhi dataset and the school variable is "schoolid". Use the ICCest() command in R or loneway command in Stata to estimate the intra-cluster correlation