Hey guys,

I have a few hypotheses in a randomized control trial that I would like to test:

1) Effect of a treatment on several outcome variables (some of them are continuous, some of them are dichotomous)
2) Moderating effect of commitment in the treatment on effect 1) - Commitment variable obviously only measured for participants of treatment group

Do you know whether I can run a multiple regression in the following way?

y = alpha + beta_1 * treatment (1 if yes, 0 otherwise) + beta_2*treatment * commitment (0 if participant is in control group, otherwise attendance percentage) + error

Or can I only run this kind of regression if I add "+ beta_3 * commitment"?

If none of these is possible, how can I investigate my hypotheses?

I highly appreciate your help!!
Floriisco