Lagged analysis with zero-inflated co... PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Scott Baldwin posted on Friday, June 12, 2015 - 8:06 am
We've got two variables measured 3 times each. X is a normally distributed and Y is a zero-inflated count variable (# of drinks). We are interested in examining the cross-lagged relationships between X and Y. I'd like to lag the relationships by one time point only and constrain the relationship to be equal across time.

!! Lagged relationships;

Y2 ON X1 (a);
Y3 ON X2 (a);

Y2#1 ON X1 (b);
Y3#1 ON X2 (b);

X2 ON Y1 (c);
X3 ON Y2 (c);

This doesn't appear permissible given the fatal error I received: "RECIPROCAL INTERACTION PROBLEM". It seemed like a long shot given the zero-inflation but I wanted to see if it was possible. Thus, I am considering splitting the analysis into two. One for the X predicted by Y and one for Y predicted by X. If you have any other additional suggestions, I'd welcome them.

However, my question has to do with what does Mplus do with a zero-inflated count variable that appears on the right hand size of an ON statement. For example, if I do something like this:

COUNT ARE Y1-Y3 (nbi);

Y1-Y3 ON gender;
Y1#1-Y3#1 ON gender;

X2 ON Y1 (c);
X3 ON Y2 (c);

Does Mplus just insert the raw Y1-Y6 values as predictors in the regressions? (no distributional assumptions)

Thanks!
 Bengt O. Muthen posted on Friday, June 12, 2015 - 5:51 pm
Don't see what's going on here - please send to Support; data too, if possible.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: