Message/Author 

Clio Berry posted on Friday, July 27, 2012  7:20 am



Hi, How can I look at the effects of gender and age on my model? If I am interested in the effects on both the fit of the measurement model (CFA) and the structural model paths (3 latent variables plus 4 predictors), then is the only option to look at a separate model for males/famales and older/younger? What other options would there be? Any thoughts from anyone much appreciated. Thank you. 


You can compare the standardized coefficients. 

Clio Berry posted on Monday, July 30, 2012  1:51 am



Thanks Linda just to check, do you mean compare the standardised coefficients across groups in multigroup modelling? Best wishes, Clio. 


Standardized coefficients should not be compared across groups because the standardization is done using different standard deviations for each group. 

Clio Berry posted on Wednesday, August 01, 2012  4:07 am



Thanks Linda, does this mean that there is no way to compare the magnitude of the effect of a predictor/covariate across groups? Thank you, Clio. 


The variables are on the same scale so you can compare the raw coefficients across groups. You can test their equality using either a chisquare difference test or MODEL TEST. 

Clio Berry posted on Thursday, August 02, 2012  12:43 am



Thanks Linda 

yobie lam posted on Saturday, October 12, 2013  10:12 pm



Hi, If i wanna to test the effect of a demo variable (3 levels), as an indicator, in the model with good model fit indices, does it correct to use the following command? in fact, i wanna to see whether there is an effect of demo on M & DV if it happened. Categorical = demo; estimator = mlr; model: IV by v1 v2 v3 v4; DV by v5 v6; M by v7 v8; M on IV; DV on M IV; demo on IV; Thank you very much. 


If you want to see the effect of demo on m and dv you would say m dv ON demo; 

Back to top 