[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: -logit- interpretation question

From   Philip Ender <>
Subject   RE: st: -logit- interpretation question
Date   Fri, 27 Nov 2009 15:18:27 -0800

On 11/27/2009  Michael Crain <> wrote:

>I have a data set of 35 observations with the predictor variable (x1)
>using the Likert scale and y = 0 or 1.
>I run -logit- and -logistic- (Stata v 10) and get a result that says:
>note: outcome = x1> 3 predicts data perfectly except for
>     x1 == 3 subsample:
>     x1 dropped and 28 obs not used
>-logit- is dropping 28 of the 35 observations and is using only 7 observations
>for the logit model but I don't understand the theoretical reason for doing so.


Your data has partial separation.  I recommend  the -firthlogit- command
by Joseph Coveney.  It will give you an estimate for your predictor x1.  To
get the program type -findit firthlogit-.

Phil Ender
Statistical Consulting Group
UCLA  Academic Technology Services
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index