Cconstructing a counterfactual

Hi all,

I have a question about constructing a counterfactual sample.

Using the migplac1 variable, I would like to construct a counterfactual sample (pre- state migration). For example, to make a counterfactual on California, people in California who immigrated one year ago are excluded, and people in other states who were in California one year ago are included. That is, after excluding samples of the California inflows and including those of the outflows into the California observations, the observations are set as the counterfactual.

Is it a valid counterfactual? Hope to hear from you.


I can’t speak to whether or not this is a valid counterfactual, only the direct meaning of these variables. It sounds like you are looking for an estimate of the previous-year’s state population in order to compare with the population at the time of the survey, please correct me if I’m wrong. If you are using decennial samples, you can accomplish this through a combination of STATEFIP and MIGPLAC1. For instance, following your California example, people who have not immigrated into California (i.e. they lived in California at the time of the survey and in the previous year) will have STATEFIP==06 & MIGPLAC1==006. People who have immigrated to California (i.e. they lived in California at the time of the survey but not in the previous year) can be identified using STATEFIP==06 & MIGPLAC1!=006. If you are using yearly ACS samples for years after 2000, you should look at the previous year’s ACS sample to estimate previous year’s population.

1 Like

Yes, that’s exactly what I did. Thank you so much for your kind explanation. Great help!

1 Like