www.cmm.bristol.ac.uk/forum

Hello,

I am running a multivariate multilevel model with 2 outcomes. I am interested in checking if the correlations between the residual outcomes are significant. As per your example, I have used the following code to calculate the correlation at Level 2:
display [RP2]cov(cons_1\cons_2)/sqrt([RP2]var(cons_1)*[RP2]var(cons_2)

Can you advise on the code or command I can use to check if the correlation is statistically significant?

Many thanks,
Valerie

I asked George about this and his response was that you can use the nlcom command for this, i.e.:

Dear Chris, this works - thanks very much. Best regards, Valerie