I am running a logit model with approximately 2000 observations, one of the
explanatory variables which is a dummy variable =1 or 0 is equal to 1 for only
50 observations or less and the standard error on this variable is huge when I
run the model - the Chi Sq is equal to almost 1.
Other people have run similar models with this data (published research) and
have found this variable to be statistically significant. What am I doing
wrong? I am not a seasoned user of logit models as you can probably gather.
Thank you in advance for any help.
|