Validity vs Reliability

1. Validity: makes sure that the test measures what it is suppose to measure.

1.1. Content : Does the test measure the instructional objectives.

1.2. Criterion-Related:Test scores are correlated with an external criterion.

1.2.1. Concurrent Criterion: Deals with measures that can be administered at the same time as the existing measure to be validated. Does the new and old version of the same test produce the same things.

1.2.2. Predictive: How well the test predicts some future behavior of the examinees.

1.3. Construct: Does the relationship to other information correspond well with the theory.

1.4. Construct: Has the relationship to other information corresponded well with some theory. Are we measuring what we think we should be.

2. Reliability: makes sure that the test yield the same or similar score rankings consistently.

2.1. Test-Retest/Stability: a retest of the same test given to the same group at a different time.

2.2. Alternate/Equivalence: two equivalent forms given to the same group at different times.

2.3. Internal: results from a single test given

2.3.1. Split-half: divide a single test in half. Usually odds on one page and evens on another page. It correlates the scores.Done in the shortest time possible.

2.3.2. Kuder-Richardson: No time interval. a single test given, split into two equal parts. Scores are correlated.