imultivariate analysis?

#1
Hi,

I have some data looking at a few variables and whether they are different for patients receiving/not receiving certain treatment. They are a mixture of proportions & means & some of them are probably dependant.

So far I've used t test for the means & z tests for the proportions. Its been suggested I look at multivariate analysis - how would I identify the right test to use given the mixture of variables I have?

Thanks.
 

hlsmith

Omega Contributor
#4
Yeah, likely they were referencing multivariable regression. The type is based on the outcome format (linear, logistic (binary), nomimal, ordinal and survival (time to event)). This can also be called multiple regression. However, multivariate is something different - outcome is time varying or units clustered.