

MODEL INDIRECT with binary outcome 

Message/Author 


I just took an mplus class last week and am trying to run MODEL INDIRECT with binary x, z, and y using Y MOD M Z (low, high, increment) XZ X; First, I ran it without defining my DV as categorical and it ran but I did not get odds ratios. Then I tried defining the DV as categorical and I received an error "The MOD statement is only allowed with estimators ML, MLF, MLR, MLM and MLMV." How can I run this so it recognizes my binary DV? Also, since my z is binary, I just typed in (1, 1, 1) in parentheses for my Z but is this correct? It looks like mplus recognizes the first 2 numbers as standard deviations? Thanks! 


Answers to your 2 questions: Q1. Say Estimator = MLR; Q2. The low/high of the (low, high, increment) statement should give the range of interest for Z. So if Z is scored 0/1 you should say (0,1,1) to get only those 2 values. 

Back to top 

