Problem 2: Power Calculations For this exercise, use the data set called "Power Calcs Exercise data.dta. Set alpha=0.05 and power=.90. Assume in all cases that there are two study arms of equal size (treatment and control) I 1. Across all households, estimate the mean and standard deviation of per person out of pocket health expenditures. 2. If the health insurance intervention were randomly assigned at the household level to 50% of households, what total sample size would you need to be able to detect a 10% reduction in expenditures? 5%? 1%? Note, to find the necessary sample size you would use "power twomeans x1 x2, sd() power()" wwwwwwwwwwwwwwm