I was running a SEM for my research and when I did the reliability test and look on my Construct Reliability (CR) and Variance Extracted (VE) score, there's an issue. All of my variables has CR score above 0,5. But some of my variables are below 0,5. Is it okay if I say that it is reliable or should I do something else?
I heard that Fornell and Lackner discussed about this issue in their article, but I still didn't catch which part of the journal that said it is okay to have my CR score above 0,5 but VE score under 0,5.
Thank you :)
I am afraid how your CR is less than .70 and your AVE is above .50. Normally, the acceptable value for CR is above .70 and for AVE is above .50.
When your AVE is below 0,50 and CR is above 0,60 you can say it's reliable (Fornell & Larcker, 1981, page 46 and 47). So your CR must be above 0,6 when your AVE is below 0,50.
How is your Cronbach's Alpha?
If AVE is below than 0.5 it means that variation explained by error is more than the observed variable which is unacceptable.
.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling