AVE calculation under CFA
Hi everyone!
I am currently using JSAP for one of my research projects, and I am now using the CFA function to validate my measurement models.
I am very grateful that JASP can directly output the AVE value while I have some questions about the calculation procedure
In the following screenshot, the standardized item loadings are 0.849, 0.799, 0.693, 0.809.
As such, the squared loadings should be: 0.721, 0.638, 0.480, 0.654, and the AVE value should be the average of squared loadings, which is 0.623.
However, it does not equal to the output given by JASP (I am using JASP 0.18.3):
I would much appreciate it if anyone could help me with this issue, thanks a lot in advance! :)
Best Regards,
Bohao
Comments
I'll forward this to our experts.
Cheers,
E.J.
Hi @MaBH,
The AVE is calculated by the package semTools, from its documentation (https://search.r-project.org/CRAN/refmans/semTools/html/AVE.html) we can read that the AVE is calculated as
Now I don't have enough intuition to confirm that that is the same as calculating the average of the squared standardized factor loadings.
From your screenshots, I can't calculate the AVE by hand, as I am missing the covariance matrix of the factors, and the implied covariance of the indicators (alternatively the covariance matrix of the residuals). However, if I try to do the calculations by hand using the formula above in some random datasets, the results match the one of JASP.
Hope this helps.
Cheers,
Simon
Hi Simon,
Thanks a lot for the clarification!
I think the approach I was using is the original formula proposed in Fornell & Larcker (1981) while the semTools package adopts a more generalized formula, which does not impose any assumptions on factor variance.
Thanks a lot for helping me with the concepts, as I have learned something meaningful from this. 😀
Best Regards,
Bohao