SE_ATE=sigma_hat_0/sqrt(N_0) + sigma_hat_1/sqrt(N_1)Ĭalculate the 95% CI for the ATE. Hint: Double-check when you’re using standard deviations and when you’re using variances.
Now calculate and display the standard error of the ATE (call this variable SE_ATE). Now calculate and display the ATE (call this variable ATE): ATE = Y_bar_1-Y_bar_0 R has a bunch of commonly-used statistics built in, like sum to count the number of observations, mean to give the average, and sd to give the standard deviation.įor the treatment and control groups separately, find the: -sample sizes -means -standard deviations (first uncomment all of the variable names) # Sample sizes by treatment group Mean(jtpa$age) # 33.14263 mean(jtpa$black) # 0.2596394įirst, we’re going to estimate the ATE and corresponding standard error by hand (it builds character!). What percent of the sample are categorized as Black?.Let’s briefly look at some of the variables. Use the qplot function (in ggplot2) to plot a histogram of earnings. To get some quick facts about earnings, we’ll use the summary command. We’re really intersted in earnings, which is the earnings (in dollars) in the 30 months after randomization. # "earnings" "prevearn" "married" "pbhous" "hsorged" "black" # "sex" "class_tr" "ojt_jsa" "oth_serv" "bdate" "age" First, how big is it? And what variables does it have? dim(jtpa) # 11204 26 names(jtpa) # "recid" "radate" "assignmt" "site" "training" "afdc"
0 kommentar(er)
