0
$\begingroup$

I have a question about interaction terms in logistic regression. I have a dummy dependent variable and 2 predictors in the model. Predictor A with 2 levels and predictor B with 3 levels. In an output table in SPSS software I can't find all interaction terms, it just shows 2 interactions rather than 5 interactions. Why is that? for more descriptions: I have these data: Protein intake (High intake versus Low intake) Geno-type (AA, AG, GG) The ref are (High intake and AA) The spss outcome showed the odds ratio just for (Low intake protein,GG) and (Low intake protein,AG). I should also can find the odds ratio for (High intake, GG) and (High intake and AG). But I couldn't find the later results. Why is that?

$\endgroup$
  • 1
    $\begingroup$ You seem to have created a second account by accident - see here for hope to merge your accounts. This will let you edit your own posts. $\endgroup$ – Silverfish Apr 7 '16 at 8:04
  • 1
    $\begingroup$ Please register &/or merge your accounts (you can find information on how to do this in the My Account section of our help center), then you will be able to edit & comment on your own question. $\endgroup$ – gung - Reinstate Monica Apr 12 '16 at 3:11
2
$\begingroup$

SPSS is showing the right output. There are only 2 estimable interactions in the situation you describe.

This is similar to the case with one categorical independent variable. If it has p levels you can only have p-1 dummy variables. With two IVs, one which has 3 levels and the other 2, the first has only 2 dummy variables, the second has only one, and so, there are 2x1 interaction terms.

$\endgroup$
  • $\begingroup$ One level is used as reference so there are indeed less odds ratios. It might be that the papers you refer to have also reported the reference categories? $\endgroup$ – Daniel Apr 7 '16 at 7:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.