Are there any alternatives for AVE (Average Variance Extracted) to ensure the convergent validity in CFA and SEM?
Whether, AVE is the most needed technique to ensure the convergent validity in CFA and SEM. Are there any alternatives for AVE (Average Variance Extracted) to ensure the convergent validity in CFA and SEM?
In line with the above facts, Malhotra & Dash (2010) documented that; AVE is more conservative and complex one. Instead of AVE, CR is alone adequate to prove the convergent validity.
Please justify my question with references.
One way of assessing convergent validity of a construct is the factor loadings of items. In CFA, if the standardized loadings of items are higher than .50 and are statistically significant, the construct is said to have convergent validity.
Dear Sivapalan Achchuthan,
First of all, I think it is important to define what convergent and discriminant validity means to you (maybe provide an example?). For instance, going back to the paper by Campbell & Fiske (1959), "Validation is typically convergent, a confirmation by independent measurement procedures (p. 81)." Thus, these terms are closely related to the idea of multitrait-multimethod analyses. Pardy et al. (2005) provide a brief overview; Brown (2015) elaborates on these topics in Chapter 6 of his book. However, we can also think of convergent and discriminant validity over time (e.g., Courvoisier et al., 2008). Different approaches to convergent and discriminant validity might lead to different recommendations.
If it refers to low latent correlations between factors (as an indicator for discriminant constructs) or to high correlations (as an indicator for convergent constructs) only, model comparisons might do (e.g., fixing factor covariances to 1 or freely estimated and interpreting the magnitude of latent correlations, allowing problematic indicators to cross-load, comparing theoretical multi-factor models to single-factor models, etc.). In addition, exploratory structural equation modeling (ESESM; Asparouhov & Muthén, 2009) or exploratory factor analysis within the confirmatory factor analysis framework (ECFA; Brown, 2015; Jöreskog & Sörbom, 1979). And of course, this should be done in a CFA framework, investigating the measurement model(s).
Best,
Frank
References
Asparouhov, T., & Muthén, B. (2009). Exploratory structural equation modeling. Structural Equation Modeling, 16, 397–438. doi:10.1080/10705510903008204
Brown, T. A. (2015). Confirmatory factor analysis for applied research (2nd ed.). New York: Guilford Press.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105. doi:10.1037/h0046016
Courvoisier, D. S., Nussbeck, F. W., Eid, M., Geiser, C., & Cole, D. A. (2008). Analyzing the convergent and discriminant validity of states and traits: Development and applications of multimethod latent state-trait models. Psychological Assessment, 20, 270–280. doi:10.1037/a0012812
Jöreskog, K. G., & Sörbom, D. (1979). Advances in factor analysis and structural equation models. Cambridge, MA: Abt Books.
Pardy, S. A., Fabrigar, L. R., & Visser, P. S. (2005). Multitrait-multimethod analyses. In B. S. Everitt & D. C. Howell (Eds.), Encyclopedia of statistics in behavioral science (pp. 1343–1347). New York, NY: Wiley.
The Convergent Validity can be assessed through AVE and the significance of all items measuring the construct
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis