Types of Validity and Reliability

1. Reliability

1.1. Test –retest estimates of reliability are obtained by administering the same test twice to the same group of individuals, with a small time interval between testing, and correlating the scores. The longer the time interval, the lower test –retest estimates will be.

1.2. Alternate-form estimates of reliability are obtained by administering two alternate or equivalent forms of a test to the same group and correlating their scores. The time interval between testings is as short as possible.

1.3. Internal consistency estimates of reliability fall into two general categories: split-half or odd–even estimates and item–total correlations, such as the Kuder –Richardson (KR) procedure. These estimates should be used only when the test measures a single or unitary trait

2. Validity

2.1. Content Validity: assessed by systematically comparing a test item within structional objectives to see if they match. Content validity evidence does not yield a numerical estimate of validity.

2.2. Criterion-related Validity: established by correlating test scores with an external standard or criterion to obtain a numerical estimate of validity evidence

2.2.1. Concurrent Validity evidence: Determined by correlating test scores with a criterion measure collected at the same time

2.2.2. Predictive Validity: Determined by correlating test scores with a criterion measure collected after a period of time has passed

2.3. Construct validity evidence is determined by finding whether test results correspond with scores on other variables as predicted by some rationale or theory.