JOURNAL OF EDUCATIONAL MEASUREMENT VOLUME 12, NO. 2 SUMMER 1975 CONSTRUCT VALIDATION: METHODOLOGY AND APPLICATION TO THREE MEASURES OF COGNITIVE STRUCTURE RICHARD J. SHAVELSON 1 University of California, Los Angeles GEORGE C. STANTON Stanford University The purposes of this paper are to briefly review construct-validation methodology and to apply some of this methodology to the problem of validating construct in- terpretations of measures of cognitive structure. A review of construct-validation methodology seemed warranted in view of the paucity of studies concerned with this topic, especially in the recent literature on human learning and memory (e.g., Adams, 1967; Anderson & Bower, 1973; Ausubel, 1963, 1968; Frijda, 1972; Johnson, P., 1967, 1969; Mayer & Greeno, 1972; Neisser, 1967; Norman, 1969, 1970; Tulving & Donaldson, 1972). CONSTRUCT VALIDATION Construct Definition Validating construct interpretations of test scores involves an interplay between construct definition, instrument development, and data collection. In this context, the construct definition sets the boundaries for potential measurement techniques and data interpretations. Although construct validation can proceed with an informal, intuitive definition, a mature construct definition should be formal and explicit. The ideal would be to define the construct with a network of associations or propositions which relate the construct to: (a) observable properties or quantities of the construct (the within- construct portion of the construct definition; cf. Loevinger's (1957) "structural component") and (b) other constructs (the between-constructs portion of the construct definition; cf. Loevinger's (1957) "external component")which are, themselves, related to observables. This network of relationships, called a nomoiogical network (cf. Cron- bach & Meehl, 1955), locates a construct in a conceptual space. The "within" portion of the definition specifies the features of the construct and links them to each other and to observable attributes of the person. The "between" portion locates the construct in a conceptual space which includes many other constructs which are rdated to, or inde- pendent of, the construct. The construct definition often operates like a test plan for the development of instru- mentation. It suggests type of stimuli, methods for observing responses, and methods for scoring and interpreting responses. Interpretations of data collected with instruments developed from the definition may be considered hypotheses, to be challenged repeatedly with counterhypotheses (Cron- bach, 1971; Cronbach & Meehl, 1955). Initial construct validation studies should ~Th¢authors wish to acknowledgethe helpful comments of L¢¢ J. Cronbach and N. L. Gage on parts of this paper. Final responsibilityfor the ideas presented herein is the authors'. Requests for reprints should be sent to Richard J. Shavelson,Department of Education, University of California, Los Angeles,CA 90024. 67