Hi Johannes,

I think that I would ask several questions about your stress variables. First, do they measure different aspects of stress, or are they measuring stress in a similar way? If they are measuring different domains of stress, then you may consider using them as separate variables. Of course, check for multicollinearity and assure that they are not too highly correlated as to measure the same construct. If they are highly correlated (e.g., .7), then I would drop one. If they measure the same domain of stress, I would pick one with the strongest theoretical links to your study and empirical support. If you combine two separate scales, you will need a strong theoretical rationale for doing so. I hope this helps.

If you combine them, go to Transform, then Compute Variable, the give the new variable a name on the top left, then add them together on the right and hit Ok. The new variable will be in your data set.

Cory

Multiple regression is the good option. You can enter both predictors simultaneously (enter method) or enter both predictors one by one if you have a literature support (hierarchal method). You can see the details in the book (Discovering Statistics Using SPSS by Andy Field), chapter 7