【Knowledge】What to do if the AVE is large than the shared variance

请问一下如果AVE大于shared variance怎么办呢？

Answer by Wen-Lung Shiau(萧文龙) 

有在英国的一位教授老师问了一个分析上的好问题，请问一下如果AVE大于shared variance怎么办呢？

Ans: A solution with simple and complex way to solve this problem. 

Under a good scale validity, a simple way includes a. delete items' factor loading <0.7  b. delete large residuals C. increase sample size.

A good scale validity is the indicators of the focal construct (1) are accurate representations of the underlying construct, (2) adequately capture the multidimensional nature of the construct, (3) are distinguishable from the indicators of other constructs (discriminant validity), and (4) are related to the measures of other constructs specified in the construct’s nomological validity (MacKenzie et al. 2011).

A complex way is to follow the appendix A&B from MacKenzie et al. (2011) and procedures from Shiau et al.(2022).

Note 1. Please take CB-SEM (model fits) and PLS-SEM into consideration. 2 Please also take reflective and formative into a critical issues.

Reference：

MacKenzie, S.B., Podsakoff, P.M. and Podsakoff, N.P. (2011), “Construct measurement and validation procedures in MIS and behavioral research: integrating new and existing techniques”, MIS Quarterly, Vol. 5 No. 2, pp. 293-334.

Wen-Lung Shiau and Li-Chun Huang (2022) Scale development for analyzing the fit of real and virtual world integration: An example of Pokémon Go, Information Technology & People, (SSCI, 2020 IF= 3.879 ABS 3 * INFORMATION SCIENCE & LIBRARY SCIENCE Q1, 23/164. Online https://doi.org/10.1108/ITP-11-2020-0793