In article <[EMAIL PROTECTED]>,
David R. McCormack <[EMAIL PROTECTED]> wrote:
>The following is a question posed by a faculty member in the social
>sciences. He has done a logistic regression finding near significance.
>But his Betas are zero. What does this mean? Is it just that the accuracy
>is only to three places? His output table is below.
If I'm reading the output below correctly, the p-value for the test
of the null hypothesis that the beta is zero is 1.000. This is total
NON-significance, which is just what you'd expect if the estimate
for the beta is zero.
In general, an estimate of 0.000 and a p-value of 1.000 would seem
surprising, and worth wondering about. But because of the discrete
nature of the data for logisitic regression, I think it's not too
surprising in this context. I wonder how many informative data
points there are, though. Probably not many.
>Table: Results of a stepwise binary logistic regression with female sexual
>orientation as the dependent variable and monthly prenatal stress, alcohol
>use, and cigarette smoking, and their interactions as predictors.
>
>Order of numbers is:
>Beta Standard Error Wald df Significance Exp(B)
>
>Step 1:
>Overall Stress X Small amounts of alcohol X Cigarettes smoked during the
>5th month
>.000 .000 3.358 1 .067 1.000
>
>Constant
>-3.012 .102 868.727 1 .000 .049
>
>----------
>
>Step 2:
>Overall Stress X Cigarettes smoked during the 9th month
>.000 .000 5.826 1 .016 1.000
>
>Overall Stress X Small amounts of alcohol X Cigarettes smoked during the
>5th month
>.000 .000 3.722 1 .054 1.000
>
>Constant
>3.079 .108 806.203 1 .000 .046
===========================================================================
This list is open to everyone. Occasionally, less thoughtful
people send inappropriate messages. Please DO NOT COMPLAIN TO
THE POSTMASTER about these messages because the postmaster has no
way of controlling them, and excessive complaints will result in
termination of the list.
For information about this list, including information about the
problem of inappropriate messages and information about how to
unsubscribe, please see the web page at
http://jse.stat.ncsu.edu/
===========================================================================
