A note on the value of customary measures of item validity

Abstract

This article discusses the value, in construction of information tests of the objective sort, of validity coefficients based upon the ability of an item to differentiate between high and low groups. Data are presented for 50 multiple-choice items and 78 true-false items that had been used 5 or more times in the course in general psychology at Ohio State University. Despite wide variation in coefficients of validity for the various items, these coefficients were found to have low reliability; that is, the validity coefficient for a given item varied widely from time to time. The author concludes that in an educational system where the students in a course are divided into small classes taught by different instructors and where the content of the course is being constantly modified, it is questionable whether coefficients of item validity have enough value to justify the labor and expense involved in their calculation. (PsycInfo Database Record (c) 2025 APA, all rights reserved)