Hello,
besides the approach where a binary or ordinal outcome is modelled via a threshold model: is it possible to model a regression with a binary outcome using a probit or logit link function like in the Generalized Linear Model (GLM) framework? If so: Could you give me a hint how to implement it?
Thanks!
Printer-friendly version
"is modelled via a threshold model" -- This is the probit model. Link functions, like with the GLM framework, are not available. However, OpenMx can do IRT models. See ifaTools
Alright, thanks for the information!