Discriminant validity is evidenced when a measure shows which pattern with dissimilar constructs?

Discriminant validity is evidenced when a measure shows which pattern with dissimilar constructs?

• A. High correlations with dissimilar constructs
• B. No relationship with dissimilar constructs
• C. Moderate correlations with dissimilar constructs
• D. Low correlations with dissimilar constructs

Answer: (D)

Discriminant validity is shown when a measure is distinct from other constructs it should not be related to. For dissimilar constructs, you want the measure to share little to no variance with them, so the correlations are low. This low shared variance indicates the measure is capturing a different attribute rather than something it shouldn’t be measuring. If the correlations were high, that would suggest the constructs aren’t as distinct as hoped, undermining discriminant validity. Moderate correlations with dissimilar constructs would imply some overlap, which isn’t ideal for demonstrating clear separation between constructs. So the best pattern is low correlations with dissimilar constructs.

Discriminant validity is shown when a measure is distinct from other constructs it should not be related to. For dissimilar constructs, you want the measure to share little to no variance with them, so the correlations are low. This low shared variance indicates the measure is capturing a different attribute rather than something it shouldn’t be measuring. If the correlations were high, that would suggest the constructs aren’t as distinct as hoped, undermining discriminant validity. Moderate correlations with dissimilar constructs would imply some overlap, which isn’t ideal for demonstrating clear separation between constructs. So the best pattern is low correlations with dissimilar constructs.