# Why would I get 0 estimates for coefficients in logistic regression?

Why would I get 0 estimates of coefficients in logistic regression? And the coefficients are exactly 0. These are for variables of different levels of a categorical variable.

-
Context please? Coefficients that are exactly zero? Or, just close to zero? Classical logistic regression or some regularized version? – cardinal Nov 13 '11 at 23:14
Friendly reminder: The more effort you put into stating your question clearly and completely, the more easily others can provide clear and complete answers. :) – cardinal Nov 13 '11 at 23:15
If it's a categorical predictor then this probably just means the mean within that level (the one that gives a $\hat{\beta}$ of 0) of the categorical predictor is the same of that in the reference level. – Macro Nov 13 '11 at 23:35
@Macro: (+1) This is true if there is only a single categorical predictor. – cardinal Nov 13 '11 at 23:56
Might be cool if you pasted your data somehow. – Adam Nov 14 '11 at 2:12