Discuss the types of psychometric data that can provide evidence of construct validity. Define and give an example of each.

Question

What will be an ideal response?

Answer 1

• Reliability/precision: Evidence of reliability/precision is important; otherwise, the test scores may correspond to theoretical predictions one time but might not do so again. In addition, the theory underlying psychological testing suggests that a test cannot have a stronger correlation with any other variable than it does with itself (which is what reliability is in practice), and therefore we can use estimates of reliability/precision to evaluate the strength of correlations with other variables that are related to the theoretical construct.

• Convergent evidence of validity: If the test is measuring a particular construct, we expect the scores on the test to correlate strongly with scores on other tests that measure the same or similar constructs. This correlation provides us with convergent evidence of validity.

• Discriminant evidence of validity: Just as we would expect some tests to correlate with our new tests, there are other tests we would not expect to correlate with our test. When the test scores do not correlate with unrelated constructs, we can say that the test is demonstrating discriminant evidence of validity.

• Multitrait–multimethod design: Researchers can combine the need to collect evidence of reliability/precision, convergent evidence of validity, and discriminant evidence of validity into one study. The researcher chooses two or more constructs that are unrelated in theory and two or more types of tests—such as objective, projective, and a peer rating—to measure each of the constructs. Data are collected on each participant in the study on each construct using each method.

• Experimental interventions: Experimental interventions in which the test is used as a dependent variable make a substantial contribution to the argument for evidence of construct validity. If the underlying theory predicts that a course of treatment or training will increase or decrease the psychological construct, a significant difference between pretest scores and posttest scores would be evidence of construct validity for the test used to measure the construct.

• Evidence of validity based on content: Because psychological constructs exist primarily in our imaginations and are not observable, it is more difficult to define a test domain well enough to construct a test that contains a representative sample from the construct. In some cases, however, psychological constructs do lend themselves to content validation, and when evidence of validity based on content can be provided, it greatly strengthens the case for construct validity.