1
$\begingroup$

What do I need to measure interaction between variables in a particular equation?

For e.g. Me just taking 50 grams of protein everyday will help me health wise. Me just doing exercise for 1 hour everyday will help me health wise. Me just stretching 1 hour everyday will help me health wise. Etc.

But Combing two of the above things will help me more health wise than just doing one. And combining all three will help me even more than just doing two of the above things.

For e.g. Taking 50 gm protein and doing 1 hour exercise everyday is beneficial than just doing one or the other. How does taking 50 grams of protein complement exercising 1 hour every day, by how much?

What I want to find out is, a way to measuring interaction between these things. How can I quantify a relationship between each above?

What type of data-set will I need to quantify how they compliment each other?

Note: I don't know what tags to use for my question. Please suggest.

Any suggestions are much appreciated.

$\endgroup$

1 Answer 1

1
$\begingroup$

You could do this in a mixed model. You would have to gather data on lots of people at multiple times. At each time you would need data on protein intake, exercise, stretching etc. for your independent variables and fitness for your dependent variable. You would need people who do various combinations of the independent variables. It might be possible to do this as an experiment, but it would be very hard (if possible and ethical at all). Probably it would have to be an observational study.

$\endgroup$
4
  • $\begingroup$ Thanks! I do have that info in my data. How should I proceed to finding the interaction between them? $\endgroup$
    – user793468
    Mar 7, 2013 at 0:59
  • $\begingroup$ Mixed models are complex and require considerable expertise. How much statistics have you studied? What software package are you using? $\endgroup$
    – Peter Flom
    Mar 7, 2013 at 1:32
  • $\begingroup$ I have taken few courses in statistics, I've studied advanced numerical analysis using SAS. It is okay if its complex, could you point me to some resources? and How I should approach such problems? Any help is much appreciated. $\endgroup$
    – user793468
    Mar 7, 2013 at 2:00
  • $\begingroup$ If you are a SAS user, and clearly you are fine with math, then Mollenberghs and Verbeke may suit you. Also good is Hedeker and Gibbons, which is less SAS-centric. $\endgroup$
    – Peter Flom
    Mar 7, 2013 at 11:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.