> My interest is to know the reference of the size of the sample in logistical
> regr ession, which indicates the number of terms to be
> employees in function of the number of observations.
The more, the better. Logistic regression is usually estimated via maximum
likelihood, and to get anything statistically appropriate, you need an infinite
number of observations. Several thousands would do, but if sample size is less
than a hundred, your Wald or LR test should hardly be trusted.
......Stanislav Kolenikov (7-095)-232-3613p -3739f
[log in to unmask] http://giganda.komkon.org/~tacik/
..................Russian - European Center for Economic Policy
.................................................Moscow, Russia
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|