Method to compare difference in average treatment effect between 2 subgroups and then between two study arms? Difference in differences for RCT data?

Asked Apr 19 '19 at 17:43

Active Apr 19 '19 at 17:43

Viewed 42 times

Essentially I have an RCT with 1 intervention arm and 1 control arm. Within each arm I have 2 sub-groups I want to compare the average treatment effect (drug users and non-drug users) for a binary outcome. I then want to compare the difference between users and non-users in the intervention vs. the difference between these groups in the control.

What is the appropriate method for this comparison of deltas between study arms? While it seems like difference-in-differences, it's not an observational study.

asked Apr 19 '19 at 17:43

Jessica

Hi @Jessica, a question: is your group variable (drug-users vs. not, if I understood well) likely to have been affected by the intervention? Or is it pre-determined/exogenous? – Roland Apr 19 '19 at 19:22
A simple difference-in-differences is just a comparison of means (like you are describing). I think it would do the trick in this case. – Amaziah Apr 20 '19 at 03:47

Method to compare difference in average treatment effect between 2 subgroups and then between two study arms? Difference in differences for RCT data?

0 Answers0