0
$\begingroup$

I have one outcome/dependent variable that can be ordinal or nominal and 3 independent variables. I have learned so far how to perform ordinal and multinomial logistic regression in SPSS between a single independent variable and the outcome variable.

However, I am more interested in examining the effect of the combinations of independent variables on the outcome variable. I could not find an answer online and I am new to regressions.

For example, the first independent variable X1 levels are friends and public, and the second independent variable X2 levels are location and time. So, I want to examine the effect on the outcome var. (what is the coefficient or estimate) in the following four cases that as a mix of X1 and X2 level:

X1= friends and X2= location
X1= friends and X2= time
X1= public and X2= location
X1= public and X2= time

I would really appreciate it if you have any thought of to perform such nested analysis whether ordinal or multinomial regression.

The only way I thought of is to have the different combinations of X1 and X2 merged in one column and have another column for the outcome variable and run the analysis in SPSS but I am not sure if this the right way. If there is no way to do that is SPSS, I would be happy to know to do it in R.

Many thanks in advance.

$\endgroup$
1
$\begingroup$

You could try an interaction term between the levels of the independent variables whose combination you're interested in.

I assume that the levels in X1 and X2 are stored as categorical variables. Then, in your regression, you could add a term X1* X2, which represents the product of the levels in X1 times those of X2.

$\endgroup$
  • $\begingroup$ Thanks. That is what I thought of but my question is would this way be correct. I mean would it give the right results-coefficients? $\endgroup$ – Fatma S May 31 '16 at 12:06

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.