Validity and Reliabiliy in Learning and Assessment

1. VALIDITY: an assessment is valid if it assesses the concepts that are supposed to be assessed

1.1. Content Validity Evidence: To be deemed valid, the questions on an assessment must directly correlate with the standards or instructional objectives that were taught.

1.2. Criterion Related Validity Evidence: Includes concurrent and predictive. This type uses evidence from the assessment as well as other sources, such as another well known assessment that evaluates the same concepts or a prediction of a student's future performance, to determine a tests validity. In this method a specifiec numerical value is assigned to rate validity.

1.3. Construct Validity Evidence: This type uses theories or reasonable assumptions to determine if the responses yeilded are what was expected.

2. Reference: Borich, G. & Kubiszyn, T. (2010). Educational testing and measurement: classroom application and practice (9th ed.). Hoboken, NJ: John Wiley & Sons, Inc.

3. RELIABILITY: an assessment is reliable if it yeilds consisten results over multiple attempts

3.1. Test Retest/ Stability: this method involves using the exact same assessment again at a later time. If no instruction has been given on the subject this shows the reliability of the assessment. The assessment is more reliable if the student scores are close to their scores when they first took the test.

3.2. Equivalence: This includes two forms of the same test given to a group. Prior to giving the test an estimate of the students scores is determined. If the students' scores are close to the estimate, the test is deemed more reliable.

3.3. Internal Consistency: this method includes "Kuder Richardson methods", "Split Halves" and "Even Odd Reliability". It can be used by using the one half of a test, then later using the other half. For this method to accurately show reliability, the concepts measured must be equally distributed throughout both parts of the test.