- The Validity of a test tells us how well the test measures what it claims to measure for the population for which it was developed. Validity information will not apply to uses of the test with different populations.
- Validity is not a property of a test, but an interpretation of the scores resulting from the test.
- In order for a test to be viewed as a valid test, it must contain content-, criterion-, and Construct-related evidence.
- A test´s validity information needs to be reviewed and revised on a regular basis.
- Content-related evidence for validity tells us whether the test´s sampling of items fairly represents the content Domain. This type of evidence is necessary in making inferences from examinees´ scores to larger domains of interest.
- Criterion-related evidence for validity tells us whether the test can predict performance on some other test, measure, or real-world domain. This is usually determined by measuring the strength of association between the test and the other measure, test, or domain of interest.
- Predictive-related Criterion validity is a type of criterion evidence that predicts a student´s performance on another measure at a future time.
- Concurrent-related criterion validity is a type of criterion evidence that predicts a student´s performance on another measure at the current time.
- Construct evidence for validity tells us how well the test measures the characteristic, or psychological construct, it was intended to measure.
- The Unified view of validity suggests that each source of validity evidence should be considered as contributing to a test´s overall validity. Evidence for each type of validity needs to be confirmed before a test can claim to be valid.
- A Table of specifications helps test developers map out the objectives of the test, the corresponding items, and the relative weights of these items. The table of specification serves as a blueprint for the test. Content validity evidence can also be drawn from the table of specifications.
- Consequential evaluation of validity of a test examines both the intended and unintended impact of the test, its interpretation, and subsequent consequences.
|
