To run correlation within or between my groups? Help thank you!

[Catherine101]

Hi there-

I have two independent groups a patient and control group who both completed an executive function test and a social cognition test.

I want to look at whether executive function correlated with social cognition.

Should I be running my correlation on the whole sample? or two correlations one on the control group and one on the patient group separately?

Thanks in advance for your help!

Iona

 

[Karabiner]

If you assume different correlation coefficients between groups, then you
could consider an analysis of covariance with social cognition as dependent
variable, and group, exectuve function, and their interaction as predictors.

With kind regards

Karabiner

 

[Catherine101]

Thanks Karabiner thats really helpful.

Best wishes,

Iona