PSY 370 Lecture Notes - Lecture 38: Concurrent Validity, Construct Validity, Factor Analysis
Document Summary
The degree to which a test measures a theoretical construct. That is, the construct it was intended to measure (think of the nomological et) Hopefully, all of our evidence should point us toward the same conclusion. But here, we"re focused on relationships with other variables. Other variables (group membership, situational factors) not already covered under criterion relationships. A factor analysis is not evidence of construct validity. We used to consider these as part of construct validity. Now they have their own category: internal structure. Do all the items really measure the same thing. ***internal structure analyses tell you about the internal structure of the items within themselves, but not withing the nomological net and bigger picture of theory. Convergent = things coming together and lining up. Correlate with other measures of the same thing. Has nothing to do with item discrimination (or bias) Relationships between our test and tests that should be unrelated.
