I'm asked to compare the observed covariance matrices of 2 groups. I'm thinking of testing a just identified model in a multi-group design. I think the expected covariance matrices of the 2 groups should then be equal to the observed matrices. In that case, testing the difference between the 2 expected covariance matrices should be the same as testing the observed ones. But I don't feel 100% confident about this approach. And maybe there's a easier way. Any thoughts on this? Regards, Ruben.