The measurement of the conceptual model includes 7 factors.
Instead of using the AVE values and correlation (Fornell-Larcker, 1981), the reviewer requested me to use the chi-square difference test to confirm discriminant validity as suggested by Anderson & Gerbing (1988). I used AMOS to perform the Chi-square tests.
However, there is an inconsistency between the two methods. The result with Chi-square tests is not as good as that with AVE values.
Why did this inconsistency happen?
Where did I do wrong when performing the Chi-square test?
Could you suggest to me ways to justify non-significant differences in chi-square tests?
Thank you very much!