Validity

Chapter 7 Validity

Valid - literally means "Truthful." Validity deals with the extent to which the test is "telling us the truth."

Compared with Reliability - Analyzing validity and improving it is a much more "difficult" and "subjective" process.

Types of validity:

1. Construct Validity - extent to which the test measures what we think (or hope) it measures.

2. Content Validity - extent to which the test items cover all aspects of the domain or construct of interest.

3. Criterion Validity - extent to which the test correlates with measures of the same or similar domains or constructs (this one gets the most attention)

3a. Concurrent Validity - extent to which the test correlates with a criterion measure at the same point in time (e.g., our new anxiety test is correlated with an established anxiety test).

3b. Predictive Validity - Extent to which the test predicts performance on a future measure / criterion (e.g., GRE scores and Grad. School GPA).

4. Face Validity - extent to which a look at the question tells you what is being measured.

Validation Measure - the test or measure with which we compare our new measure. We might validate a new IQ test by comparing it with the Stanford-Binet.

Validity Coefficient - " rxy " - we are really already familiar with this, Pearson's r between X and Y.

”X" - is always the "predictor"

"Y" - is always the "criterion"

Relationship Between Reliability and Validity - A test can be reliable without being valid. It cannot be valid without being reliable. "Reliability is Necessary But Not Sufficient" for validity.

Maximum Value of the Validity Coefficient - rxy can only go as high as the reliabilities of the two measures (X and Y) will allow.

rxy max = SQRT (rxx) (ryy)

Theory Behind Validity Analysis

Recall from Reliability Theory "X = T + E" Any raw score (X) is made up of True Score or "stable characteristics" (T) plus some Error (E).

In validity theory and...