Which stats test to use

[hello_kitty]

Hello

I am really rubbish at stats, so now in my 3rd year project I am totally lost.

I was told I need to do multiple regression in SPSS, as my questionnaires were measured with likert scales.

So, I had 30 questions about how a person feels about sports. And 3 questions about how many times they did sports in childhood, adolescence and now (also in likert scales).

Thus, those 30 questions should predict how many times per week person did sports at different points of her life.

the hypothesis is: factors causing drop out of sports at university. so the different times of life were measured to find out whether people really drop out of sports after college. and then those 30 questions are measuring factors why they drop out of sports.

But if I will do multiple regression, I can have only one dependent variable, how should I deal with that?

Help needed really seriously, so any comments will be appreciated. Searched loads, of books but have no idea how they can answer me that.

Cheers

[duskstar]

You can have whether they drop out of sport as the outcome/dependent variable (yes/no) and then whichever of theother variables you want as independent variables.

If you do it this way the outcome will have meaning (probability of dropping out of sport) and you will be able to interpretate each dependent variables coefficient.

Oh, and I'd use a logisitic model to do this personally, if you know about regression methods.