Hello everyone,
I am running into an issue trying to use CEM on a dataset where the treatment variable
can take different values within the same unit. Because countries in my panel data set can
have different values of the treatment variable in different country years, CEM is
matching the same country from before it receives the treatment to the same country after
it receives the treatment. For example, if Panama received the treatment I am looking at
in 1950 (treatment =1), and was coded with treatment = 1 from 1950 - 1970, CEM is matching
Panama to itself. Instead, I'd like to match only countries who always have treatment
= 0 to those that take treatment =1.
Is there a way to prevent CEM from doing this?
Thank you!
P. Zachary
Department of Political Science
George Washington University
Show replies by date