I want to test for correlations in the brain with a set of variables (such as personality measures). I have previously used regAna to do this, but I recently learned I could also do it using 3dttest.
I am not sure though about 2 things:
- are these methods both achieving the same thing?
- can I do more than one variables at the same time using ttest? and if I do, are the results equivalent as doing it separately for each variable
for example if I want to see the variation with age and then variation with sex, as opposed to variation with age after controlling for sex?
Thanks a lot for your advice!
Actually I went ahead and tested both (Ttest using 2 covariates together as compared to only 1) and they look different - I am sorry, I should have tried that first! I guess this is we are seeing the effect of covariate A after controlling for covariate B, or am I missing something?
I guess this is we are seeing the effect of covariate A after controlling for covariate B, or am I missing something?
Yes, that’s correct. Suppose that you have a model with two explanatory variables:
y = a + b[sub]1[/sub]* x[sub]1[/sub] + b[sub]2[/sub] *x[sub]2[/sub] + e
Unless that the two variables x[sub]1[/sub] and x[sub]2[/sub] are fully uncorrelated (orthogonal) with each other, it’s expected that the effect estimates b[sub]1[/sub] and b[sub]2[/sub] are different from their counterparts in a model with one explanatory variable.