This question is based on an earlier question of mine given here

i refer to the same paper as above: Card and Krueger (1992)

I have understood the rationale for use of "cohort and state fixed effects" from the earlier question. I have further looked at this answer to another question get more clarity. What I have understood is that when we use "cohort and state" fixed effects by interacting between cohort and state, an individual will have 1 on the interaction if it has 1 on both cohort and state.

(1) So, am I correct in interpreting that "cohort-state" fixed effect compares the effect of treatment between two individuals who are in the same state and same cohort?

(2) If I had only "cohort" f.e, I could have interpreted it as the comparison being between individuals in the same cohort. Similarly, with only "state" f.e, it could be interpreted as comparison between individuals in the same state. Following this rationale, what would be the interpretation of using both "state" and "cohort" f.e separately? How would the interpretation differ from that of "cohort-state" f.e?




Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.