In a paper published in the Journal of Marketing Research in 1981, a group of statisticians introduced the concept of Average Variance Extracted, a statistic that states how much variance captured by the latent variable in a structural equation model is shared among other variables. The calculation of Average Variance Extracted requires a structural equation model to already exist, since it needs the loadings of the indicators for the latent variable for which it is to be calculated.
List the statistics that will be used for the Average Variance Extracted computation. The statistics needed are the loadings for the indicators on the latent variable of interest, the variance of the latent variable and the variances of the measurement errors for all of the indicators. These statistics should all come directly from your structural equation model.
Compute the sum of squares for the indicators loading on the latent variable. List the loadings. Square these loadings. Sum the resulting numbers. Call this value “SSI.”
Sum the variances of the measurement errors. Call this value “SVe.”
Compute the denominator for the Average Variance Extracted. Multiply “SSI” by the variance of the latent variable. Add “SVe” to the result. Call this value “Denom.”
Compute the numerator for the Average Variance Extracted. Multiply “SSI” by the variance of the latent variable. Call this result “Numer.”
Calculate the Average Variance Extracted. Divide “Numer” by “Denom.” The result will be a number between zero and one. This is the Average Variance Extracted.
References
About the Author
Having obtained a Master of Science in psychology in East Asia, Damon Verial has been applying his knowledge to related topics since 2010. Having written professionally since 2001, he has been featured in financial publications such as SafeHaven and the McMillian Portfolio. He also runs a financial newsletter at Stock Barometer.