skip to main content skip to footer

Validity of Test Interpretation and Use

Author(s):
Messick, Samuel
Publication Year:
1990
Report Number:
RR-90-11
Source:
ETS Research Report
Document Type:
Report
Page Count:
17
Subject/Key Words:
Test Interpretation, Test Use, Test Validity

Abstract

The meaning of the terms "validity" and "scores" are defined and clarified. It is noted that "validity is a matter of degree, not all or none... and that over time the existing validity becomes enhanced (or contravened) by new findings." Convergent (supporting) and discriminant (contradictory) evidence are also defined in terms of test validation and it is concluded that "the key issues of test validity are the meaning, relevance, and utility of scores, the import or value implications of scores as a basis for action, and the functional worth of scores in terms of the social consequences of their use." Three types of test validity, one of which comprises two subtypes, and the limitations of each, are discussed. These are: 1) content validity, including predictive and concurrent validity; 2) criterion-related validity; and 3) construct validity. Salient properties and critical limitations of these traditional types of validity are underscored. Two facets of unified validity are then considered--the source of justification of the testing (either evidence or consequences), and the function or outcomes of the testing (either interpretation or applied use). A four-fold classification based on this is obtained, highlighting both meaning and values and represented in a four-cell table. Each cell is briefly considered in turn and it is concluded that "as different foci of emphasis are added to the basic construct validity appearing in each cell, ...what at first glance was a simple four-fold classification appears more like a progressive matrix." (JGL) (32pp.)

Read More