This study investigates aspects of validity reflected in a large and diverse sample of published measures used in educational and psychological testing contexts. The current edition of Mental Measurements Yearbook served as the data source for this study. The validity aspects investigated included perspective on validity represented, number and kinds of sources of validity evidence provided, overall evaluation of the favorability of the test, and whether these factors varied as a function of the type of test. Findings reveal that validity information is not routinely provided in terms of modern validity theory, some sources of validity evidence (e.g., consequential) are essentially ignored in validity reports, and the favorability of judgments about a test is more strongly related to the number of validity sources provided than to the perspective on validity taken or other factors. The article concludes with implications for extending and refining current validity theory and validation practice.

For tests used in education and psychology-and for high-stakes tests used in kindergarten through Grade 12 education contexts in particular-two references provide authoritative information that can be used for critical evaluation. The first, the Standards for Educational and Psychological Testing (American Educational Research Association, American Psychological Association, & National Council on Measurement in Education [AERA, APA, & NCME], 1999; hereinafter Standards), has provided guidance on appropriate test development and use since its first edition introduced over 50 years ago (APA, 1954). The Standards provide guidelines that apply to a wide range of tests and testing contexts. The second reference, the Mental Measurements Yearbook of the School of Education (MMY) series, has chronicled the technical quality of tests for over 70 years (Buros, 1938). PRIMACY OF VALIDITYAlthough intended to serve different purposes, the Standards and the MMY both acknowledge the central role of validity in testing. Although validity evidence is important in any testing context, it is perhaps especially

No abstract