I understand your argument that discriminant validity is more important, especially in avoiding overlap between constructs. However, I would argue that without strong ...
I understand your argument that discriminant validity is more important, especially in avoiding overlap between constructs. However, I would argue that without strong convergent validity, we cannot be confident that the items are even measuring the intended construct. I agree that both are essential, but perhaps convergent validity should be established first before evaluating discriminant validity. What do you think about this sequence?