I am sort of stuck with the following problem concerning logistic regression analysis in my research on the determinants of technology adoption.
I have a questionnaire containing 28 items (two ask about company characteristics, one is whether the company has adopted or not adopted the technology (thus the dichotomous outcome variable), the rest are Likert scale items). The Likert scale items refer to a total of 7 factors, which I identified in a principal component analysis. Now, I want to find out how much every single factor contributes to the adoption/non-adoption of the technology. I try to do this by computing the composite score for each factor (taking the mean) and performing a binary logistic regression analysis on all of them.
In SPSS, do I have to indicate that the composite scores of the factors are categorical? Or are they continuous? Is there anything else I have to watch out for?
Thanks for your guys' help!