

Comparison of Path Coefficients in Ne... 

Message/Author 


Hi, I have two models. They differ such that Model A contains an additional predictor, whereas Model B doesn't (i.e., I set one factor loading in Model B @0). Now I have two hypotheses: First, I want to test if Model A explains more variance than Model B. Therefore, I calculate a chi square difference tests and expect it to become significant (hope that is correct). Second, I want to test whether the path coefficients of Model B (except for the path @0) differ from the corresponding path coefficients in Model A. Is there any test I can use? My idea would be to simply compare the confidence intervals of the path coefficients as given by the CINTERVAL command. Is that appropriate? And if so, (how) would it be possible to also obtain e.g. 20% confidence intervals (because testing a H0)? Thanks for your help! 


You want to ask this question on SEMNET. Be sure to explain how a factor loading not fixed at zero for a factor indicator implies that this indicator is an additional predictor. 

Back to top 

