Convergent vs. Discriminant Validity: Which Is More Important?

If I had to prioritize one, I would choose discriminant validity as more important.

Reasoning

While convergent validity ensures that items measuring the same construct are strongly related, discriminant validity ensures that different constructs are actually distinct from each other. In structural equation modeling, this distinction is crucial because if constructs are not clearly separated, the entire model becomes conceptually unclear.

For example, in my research (e.g., variables like stress and life satisfaction), these constructs are theoretically different. However, if the measurement model shows poor discriminant validity (e.g., very high correlations between factors), it may suggest that they are essentially measuring the same concept. This would make the model redundant and weaken the interpretation of relationships between variables.