Hi guys, I am using a cross-section data to run 2sls regression. And there is a bunch of dummies coming from a "country id" variable which I have to control them as covariates both in the first and the second stage of my regressions. I just wonder that may I use the original "country id" variable to be a continues covariate rather than turn it into a bunch of dummies. I assume that since these are just covariates whether or not to be dummies, it should not affect my results. BUT I am wrong. The result of my interest variable under the continues covariates is very different from that of under the dummy covariates. Does anyone know why? Many thanks in advance.

Best

Frank Qu