Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: "predicts failure perfectly" in binary logit


From   "David Quinn" <dquinn@gvpt.umd.edu>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: "predicts failure perfectly" in binary logit
Date   Fri, 18 Nov 2005 16:50:34 -0500

Hello...

I have a question.  I am using binary logit to analyze an outcome
variable (formal).

When I enter the following string:

logit formal highfaci highform highman xpower xpowform xpowmanip viol pc
period1 period2 period3 period4 cmgoal

I get the following message about one of my explanatory variables
(xpowfaci):

note: xpowfaci != 0 predicts failure perfectly
         xpowfaci dropped and 11 obs not used

I understand that this message is telling me that there is a perfect
correlation occurring with this variable.  This "xpowfaci" variable is
an interaction term of two of my other explanatory variables --
"highfaci" and "xpower." 

My question: Is it proper to still include the eliminated variable --
"xpowfaci" -- in the equation and report the results with these 11 cases
dropped, or should I re-run the logit excluding "xpowfaci" from the
equation completely?  I notice that the coefficient on one of my other
explanatory variables -- "highfaci" -- changes radically depending on
whether or not I include "xpowfaci" in the equation.

Thank you!

David Quinn
Department of Government and Politics
University of Maryland, College Park   

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index