Criterion-Based Validity Evidence

Criterion-based validity evidence is frequently referred to as criterion-based validity, criterion-related validity, or simply criterion validity. In social science research, understanding the psychometric properties of an instrument is essential. These important psychometric properties include reliability and several types of validity. Thus, a researcher must often assess the evidence for the face validity, construct validity, content validity, and/or criterion-based validity of the instruments used in research. Although all forms of validity evidence indicate how well a measure measures what it is supposed to measure, criterion-based validity evidence is related to how accurately one measure predicts the outcome of another criterion measure. If a measure is a valid indicator of a construct of interest, then that measure could be used to predict the values of other measures related to that construct. Therefore, a measure that has high criterion validity would be one in which knowing the value of the predictor variable would allow the researcher to predict the value of the other criterion measure with high accuracy. Furthermore, the [Page 423]criterion-based validity applies to the validity of the predictor variable, not the criterion variable.

Criterion-based validity evidence is often of primary concern in educational research and assessment because educators are frequently looking for ways to determine whether assessment measures will be able to predict success or failure in later educational endeavors. Although other measures of validity are assessed using the opinions of experts and the similarity of a measure to other useful measures of the same construct, criterion-based validity is generally calculated and reported in quantitative measures from correlation and regression analyses. The general term criterion validity can include measures of predictive validity, concurrent validity, and postdictive validity. It is important to note that there are several threats to the criterion-based validity of conclusions drawn from the use of instruments that researchers must account for when using measures to predict criterion variable values. The remainder of this entry further describes the three types of criterion-based validity, highlights limitations of criterion-based validity evidence, and demonstrates how to calculate a measure of criterion-based validity.