In my objective there are three variables. Where one variable effects the other two. Through SEM I found both relationships are statistically significant but I want to know among two variables in which the first variable effects the most.
you can use squared multiple correlation to evaluate which items measure variable more than others. squared multiple correlation is like R square in SPSS. please check the attached file to understand better. you can check factor loading for this purpose as well
Please chech attached file to follow a question like your question and given responses
Dear Uttam
Since beta is a standardized value, you may simply look into the magnitude of two different betas, For the effect size, you may refer to the squared multiple correlation.
Thank you Prof. GHOLAMREZA for your kind attention and comments.
Thank you Dear Wing Yip Chui for your comments.
I have understood your comments and its really valuable. Dear Wing Yip Chui if two beta values are significant at different p values then should I only look into magnitude?
In some statistical programs like Mplus, you can constrain the two paths you are comparing to be the same and use chi square different test, if the model degraded significantly, meaning the two paths should really be estimated separately, then whichever path has the higher value is the stronger one.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling