Validity and Reliablity

Get Started. It's Free
Validity and Reliablity

2. Validity Evidence

2.1. Content Validity Evidence

2.1.1. This is done by ensuring that the test questions corresponds to what is being measured by the test (Kubiszyn & Borich, 2010, p. 330).

2.2. Criterion-Related Validity Evidence

2.2.1. Concurrent criterion-related validity evidence

2.2.1.1. This can be figured by giving both a new and established test and figuring the correlation between the two test scores. This provides us with a validity coefficient. Then if the new test provides similar results as the established test, then the new test would be considered a good and valid test (Kubiszyn & Borich, 2010, p. 330).

2.2.2. Predictive validity evidence

2.2.2.1. This explains how a test can be used to determine the success of the persons who are taking the test will do in the future in certain situations. This can be done by having people take a test, then after a period of time in the situation that is being predicted measure how well that person is doing with a test and comparing. The set of data from both tests and this correlation provides us with the predictive validity coefficient (Kubiszyn & Borich, 2010, p. 331).

2.3. Construct Validity Evidence

2.3.1. This type of validity is measured if there is a relationship that corresponds to some theory (Kubiszyn & Borich, 2010, p. 331) . Such as if a test is administered about long division, then you would expect an improvement in scores after intensive coaching in long division.

3. Reliability

3.1. Test-Retest

3.1.1. This can be determined by having students take a test twice in a measured period of time. The scores are then compared to one another and this provides a correlation score, the higher the score the more reliable the test.

3.2. Alternate Forms

3.2.1. This reliability test provides two similar tests to the students to determine the reliability of the test. The higher the correlation between the two scores the more reliable the test would be considered to be.

3.3. Internal Consistency

3.3.1. Split-half methods

3.3.1.1. This can be determined by having the students take a test with only the odd questions and then with only the even questions and comparing the two scores to one another.

3.3.2. Kuder-Richardson Methods

3.3.2.1. This method determines whether one form of a test is similar to the other form of the test using the KR21 or the KR 20 mathematical equations.