Greetings! The link to the Todd Grande tutorial below demonstrating the calculation of AVE and composite reliability to factor loadings appears to be popular viewing.
My question is would you recommend calculating these after EFA? (on the basis that you are not following up with a CFA). Im struggling to find an academic reference to this. Hair et al (2014) only talk about AVE in the context of CFA. many thanks
https://www.youtube.com/watch?v=8tyjdfpiAJQ
I also know that only used in CFA, AVE is only optimal when the loads are greater than .70, this would serve to determine the good fit of the proposed structure taking into account the coverage of factors in addition to its square root can also determine discrimination in relation to other correlated variables.
Hello Nick,
I agree (with your query and Cristian Ramos-Vera 's response) that AVE/CR are probably most useful for CFA, not EFA.
AVE is a descriptive statistic, which some might find useful. However, for me, in EFA, the more important questions are: (a) does the resultant chosen structure make sense; (b) which variables appear not to affiliate with any factors to a meaningful level (and, what to do with them: abandon, modify, replace?); and (c) how stable is the structure across samples (which a CFA can help answer).
None of these questions is informed by AVE.
Composite reliability is another descriptive statistic, and derives from variable-factor loadings, as does AVE. If you're choosing McDonald's omega as an indicator of reliability over older measures (such as Cronbach's alpha), then this, too, might be useful.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability