Dear Stata users,

I am writing this post hoping that someone could help me.

Here the story, I am aiming to identify the causal effect of an intervention. I have a panel covering two periods in time. During the first period, one group is already treated and during the second the whole population is. So the prediction would be that the outcome of the non treated group in the first period catch up the one of the treated group in the second period.

Normally, the fact of being treated in the first period is totally random, but documentation on the specific intervention is quite rare… Furthermore doing a t-test shows that on many points the two groups are not similar. An RCT is then maybe not appropriated. Thus, I was thinking about launching a Difference-in-Difference method, but I am aware that two points in time are not sufficient. But maybe proceeding at robust placebo test is enough.

If any of you have an idea about which method I should adopt I will be very grateful.

Thank you!

Jeanne Ronsard