Validity is necessarily a major concern of any testing program. It is in the interest of the user that a test measure what it is supposed to measure and that it bear a reasonable relationship to the criteria it is intended to predict. It is the responsibility of a test sponsor to insure that these qualities prevail in the testing program. This principle of responsibility applies with a special force to a national program such as the Graduate Record Examinations (GRE), which affects large numbers of people. The stakes are high for both the individual and society. The main purpose of this paper is to facilitate discussion of important issues concerning validity and to work toward a framework that the GRE Board Research Committee will find useful in assigning priorities and initiating projects. Toward that end, subsequent sections of the paper provide background, define the scope of the problem, and outline six major objectives that might guide the Board's efforts in this important area of research: to encourage and facilitate institutional validity studies; to deal effectively with methodological issues concerning validity that require, the GRE program's initiative; to develop improved criteria of success in graduate study; how to improve population validity and enhance understanding of it; to improve institutional use of summary program data; and to systematically insure the validity of revised or new measures resulting from program renewal.