Error performing Logistic Regression

Hi wonder if anyone can help me?

I am currently doing my dissertation and studying fear of crime comparing victims and gender.

I performed an ANOVA for my first part as I had another DV which was physiological from a galvanic skin response.

My other DV was my fear of crime scale which is ordinal (1=no fear, 2=a little fear, 3= do not know know, 4=moderate fear and 5= fear)

I was told to do a non parametric test with my gender and victim status which was simply yes or no answer.

So i go to perform a logistic regression (read it was the appropriate test to conduct) and this message appears on the output:

The dependent variable has more than two non-missing values. For logistic regression, the dependent value must assume exactly two values on the cases being processed.

I am completely lost to what this means or how to correct this issue :shakehead

Can someone please explain perhaps what needs doing or what it is telling me?

Thanks :)


Fortran must die
I am not sure which DV you are using as you mention multiple variables. The fear one has five levels and I am guessing that is the DV you are talking about running above.

If your DV has more than two levels you can not run binary logistic regression. That is what the warning message is telling you I think - that you are tryng to run binary logistic regression with more than two levels.

You need to run ordinal logistic regression (if you think the data is ordinal which I would looking at your levels) or multinomial logistic regression if you think the levels are unordered.
Thanks for the reply :)

Not to long ago my supervisor who helps me out has emailed and told me to do two ANOVAs and discuss why in the discussion section of my dissertation so he's decided to keep it simple it seems.

But thanks anyway for the reply i'll keep that in mind if i ever have to do one again


Fortran must die
If you do it again you might Look at the "Logistic Regression Using SAS" by Paul Allison (2nd ed). He has a chapter for ordinal logistic regression and one for multinomial logistic regression.

Good luck.