Hi,
not sure if it solves your complications, but have you checked the Thurstonian model? I think it can account for covariances (2), and missing items (3) should be straightforward, not sure about (1) though.
1 Like