Khizar
Check:
1. Multicollinearity (though not often a problem in logistic regression)
2. Linearity of the logit
3. Case to variables ratio should ideally be 10 - 15; you are a little
short. Try removing more variables.
Max
-----Original Message-----
From: A UK-based worldwide e-mail broadcast system mailing list
[mailto:[log in to unmask]] On Behalf Of khizar azam
Sent: 09 December 2010 06:52
To: [log in to unmask]
Subject: binary logistic regression perfect model
Dear all greetings,
i hope you all are fine, i have conducted a survey(120 questions) and am
trying to get some meaningful results.I am stuck up with my data analysis
and now i have used pasw18 to analyze my data also i have reduced my 120
variables to 30 variables and my sample size is 259 but i have a problem now
as in model summary -2log likelihood=0, my Negelkerke R value is 1 and 100%
correct predictions is what i am getting, but when i move down to the
variables in the equation table and under the column of sig i get a mix of
0.99 and 1.0 values indicating no significant variable.
your experteze is saughted and would be duly acknowledged. thanks.
regards,
Khizar
:
.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|