Validity Flashcards
(15 cards)
What is validity?
How accurately a tool measures what it intends to measure.
What are the types of validity?
Face validity, Content validity, Criterion validity.
What are the types of Criterion validity?
Concurrent validity, Convergent validity, Predictive validity, Discriminant validity
What is Face validity?
How well a tool/item measures its intended concept on a surface level.
What is content validity?
How well the items measure their intended construct– how relevant and clear the items of the tool are.
What is the difference between content validity and reliability?
Content validity is about the clarity/relevancy of underlying concepts for each item. Reliability is about getting consistent results for one participant among different raters.
What is criterion validity?
A type of validity that measures how a current scale relates/is associated with other scales.
What is concurrent validity?
How well a new test correlates with an established scale of the same measure. (New scale vs. a gold standard scale).
What is predictive validity?
How much does a test score of the scale predict future performance/behaviour? (ex. IQ score and academic performance)
What is convergent validity?
How well does the new test correlate with another measure that is theoretically related to the same construct? (ex. A new test measures self-esteem should correlate strongly with a test measuring self-confidence or self-concept).
What is discriminant validity?
Whether the current scale is not associated with a scale that is conceptually different. (ex. A scale measuring poor sleep should have strong discriminant validity with a scale measuring personality).
What is construct validity?
How well a tool measures the theoretical construct it intends to assess.
What are the ways of testing construct validity?
Testing hypothesis, structural or factorial validity, or cross-cultural studies.
How can you test hypothesis?
Compare the construct between 2 groups via independent samples t-test– ex. administering your tool to 2 groups, one w/ the construct, one w/o the construct (with this, your results would need to show significance p > 0.5). Use convergent validity.
