[John Adler]

Thanks for your feedback Carlo. I'm reading up on this now so I can be more informed. What do you think of the following statement:

In most linear probability models, R 2 has no meaningful interpretation since the regression line can never fit the data perfectly if the dependent variable is binary and the regressors are continuous. This can be seen in the application below. Which is from: https://www.econometrics-with-r.org/...ity-model.html

Kindest regards,

John

[Carlo Lazzaro]

John:
yes, it may happen.
Hower, in your case most of the predictors were categorical.
You may want to see what happens to your results if you switch to -xtlogit,fe-, remembering that -xtlogit- produces conditional -fe-.