I have collected data for three variables (2 IVs and 1 DV). All are measured through multiple items measured on 5 point likert scale. I want to run CFA and SEM through AMOS. AMOS uses Maximum likelihood method for estimation, which assumes the data to be normally distributed and endogenous variable is to be continuous. However, in my case, the endogenous variable is ordinal (measured on 5 point likert scale), which cannot be normally distributed. What alternative method of estimation in AMOS available?