Message/Author 

Anonymous posted on Sunday, May 29, 2011  7:54 am



Greetings! For my analyses, I am interested to evaluate whether different levels or benchmarks of a psychological construct measured by two widely used assessment measures (x1 and x2) differentially relate to another construct that is a continuous variable y, at two assessment points (post treatment and follow up). The benchmarks are set at mean value, 1 Standard Deviation above the mean, and 1 Standard Deviation below the mean, separate for child and parent. How do I analyze this in MPlus? Thank you! 


It seems the model you want is: MODEL: f BY x1 x2; y WITH f; I would first fit the model separately at post treatment and follow up. If the model fits at both time points, I would then put them together. The benchmark issue of plus and minus one standard deviation is not clear to me. 

Anonymous posted on Sunday, May 29, 2011  11:28 am



Thank you very much for your prompt response. The benchmark issue of plus and minus one standard deviation is to examine at what point (i.e., at mean, 1SD+, 1SD the mean value) a specific psychological disorder measured by two assessment tools becomes problematic enough to cause functional impairment and hence require psychotherapeutic services. 

Anonymous posted on Sunday, May 29, 2011  2:36 pm



How do I analyze in MPlus the research question stated in my previous post? Thanks! 


You said "differentially relate to another construct that is a continuous variable y, at two assessment points (post treatment and follow up)."  If the word "differentially" refers to a comparison across the two time points of the relationship between the construct and y (the WITH statement  although it may be an ON statement depending on how you think of this), then the model given above can be applied to the two time points and the relationship equality tested. 

Back to top 