I am currently trying to run a binary logistic regression on survey data (n=19.355). One of the independent factor variables "country" has 12 categories overall, meaning 12 different countries were included in the survey. When I run the logistic command, one of the countries is always eliminated from the regression (NOT the base level category). However, this does not happen when I reduce the number of independent variables included in the regression model.
The country dropped is always the one with fewest observations (n=381), the other countries have n=676 and more, up to n=7.000 observations.
Why is stata doing that/do I need to code something differently? Is there a way to keep the category in the regression model?
An example of how i would run the command:
Code:
svy linearized : logistic DREAMING ib(first).Finance ib(first).sex ib(freuent).work ib(first).Country
0 Response to Stata drops category of factor variable in logistic Regression
Post a Comment