Convergent vs. Discriminant Validity: Which Is More Important?

I understand your point and I partly agree that convergent validity is important because it ensures that items within a construct are consistent and reliable. Your reference to Fornell and Larcker (1981) is also very relevant in supporting this idea.

However, I would still prioritize discriminant validity. In my research on Kahoot, constructs like motivation and engagement are closely related, so even if each construct has good internal consistency, the model can still be problematic if these constructs are not clearly distinct. Without discriminant validity, different constructs may overlap, leading to confusion in interpretation.

From my perspective, convergent validity ensures that a construct is measured well, but discriminant validity ensures that constructs are not redundant. If constructs are not distinct, the overall model loses meaning.
Therefore, both are important, but I believe discriminant validity is more critical for ensuring clear and meaningful results

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?