Dear all,

I want to perform a DiD regression but as my treatment was not randomly assigned, I am using propensity score matching beforehand and am having some trouble linking the two. After using psmatch2, it has generated several new variables which are: _pscore, _treated, _support, _weight, _y, _id, _n1, _nn, _pdif

I have read that I should use the weights when performing the DiD regression but one problem I am having with that is many of my observations have not been assigned a weight at all and this would reduce my sample size by almost 1/4! However, I suppose I will have to do this if it truly does mean that only that much of my sample was matched.

My other question is about the _treated variable. This does not have nearly as many values missing as the _weights variable and they all have a pscore and are "on support". I am hoping this variable is the new group of treated and untreated that psmatch2 has created for me after matching and if I just drop the missing observations, I can use this to perform the DiD regression. Please let me know if I am interpreting this variable correctly!

Lastly, I have performed PSM on the pre-treatment observations but I am not sure how to give those values to the same households post-treatment (I hope this question makes sense, English is not my first language). So my data currently looks like this:

HH | year | _pscore | _treated | _support | _weight
1 | 0 | .1263 | 0 | 1 | .
2 | 0 | .1076 | 0 | 1 | .
3 | 0 | .1103 | 0 | 1 | .
4 | 0 | .2083 | 0 | 1 | .
5 | 0 | .1151 | 0 | 1 | 1
6 | 0 | .1508 | 0 | 1 | 1
1 | 1 | . | . | . | .
2 | 1 | . | . | . | .
3 | 1 | . | . | . | .
4 | 1 | . | . | . | .
5 | 1 | . | . | . | .
6 | 1 | . | . | . | .

Thank you very much in advance. Please let me know if I need to clarify anything further

Edit: I would also like to ask one more question! There is another user-written command -diff- which does the kernel PSM diff-in-diff for you. Could I perhaps use this command instead?