Hello Statalisters,
I am new to Stata and am working on a balanced panel data set of 178 firms over 6 years and want to performa a logistic regression.
To prepare a Hausman-Test, I ran a Random Effects-model and then wanted to run the Fixed Effects model. However, when I run the FE-model, 100 of 178 companies are dropped. Therefore I thought I should switch to Random Effects. However, when I now run the Hausman-Test, to check between Random and Fixed Effects, it suggests me to use a Fixed Effects model. Is this, because stata does not account for the lack of observations with which the FE-model was created? Or does it even make sense to run the Hausman-Test at all, after such a considerable number of observations has already been dropped and I should just immediately focus on the RE-model?
I am looking forward to your replies.
Thank you,
Julian
I am new to Stata and am working on a balanced panel data set of 178 firms over 6 years and want to performa a logistic regression.
To prepare a Hausman-Test, I ran a Random Effects-model and then wanted to run the Fixed Effects model. However, when I run the FE-model, 100 of 178 companies are dropped. Therefore I thought I should switch to Random Effects. However, when I now run the Hausman-Test, to check between Random and Fixed Effects, it suggests me to use a Fixed Effects model. Is this, because stata does not account for the lack of observations with which the FE-model was created? Or does it even make sense to run the Hausman-Test at all, after such a considerable number of observations has already been dropped and I should just immediately focus on the RE-model?
I am looking forward to your replies.
Thank you,
Julian
