With the analysis dataset I created, info_rct_analysis.dta, do standard balance checks to show the randomization worked.
Regress a student (or school) level outcome, on randomization block dummies, and the treatment indicators. Create a paper-ready table with the following characteristics:
Rows: Different variables you are checking balance on
Columns: (1) Control group mean, (2) Treatment P difference, (3) Treatment S Difference, (4) Treatment B difference, (5) p-value of a test that (2)-(4) are jointly zero
The bottom two rows should report the sample size and another p-value that tests all coefficient estimates within a given column (only 2-4) are jointly zero. Look into seemingly unrelated regression for examples on this. We can also talk about this in some detail when you get to it.
With the analysis dataset I created, info_rct_analysis.dta, do standard balance checks to show the randomization worked.
Regress a student (or school) level outcome, on randomization block dummies, and the treatment indicators. Create a paper-ready table with the following characteristics:
Rows: Different variables you are checking balance on Columns: (1) Control group mean, (2) Treatment P difference, (3) Treatment S Difference, (4) Treatment B difference, (5) p-value of a test that (2)-(4) are jointly zero
The bottom two rows should report the sample size and another p-value that tests all coefficient estimates within a given column (only 2-4) are jointly zero. Look into seemingly unrelated regression for examples on this. We can also talk about this in some detail when you get to it.
┆Issue is synchronized with this Asana task by Unito