Hi all,

I'm using SPSS to analyse my research data. Usally i'm building logistic regressions in which I find confounders by noticing a change bigger than 10% in my standardized beta coefficients (i.e β, OR), when the suspected confounder is added to the model.

BUT, now that I need to build a LINEAR model I find myself lost. Should I use the exact same approach with multivariate linear models?
(it is after all an additive model, in wech the standardized beta coefficients do not stand for OR).

Any help will be much appreciated,
O