Cronbach, Lee J(oseph) 1916-2001

ISSUES IN MEASUREMENT

Cronbach's early work on measurement led to the “Coefficient Alpha” paper (Cronbach, 1951), which provided a widely used formula for estimating the reliability of test scores. Later he developed generalizability theory (Cron-bach, Gleser, Rajaratnam, & Nanda, 1972) that examined systematic variations in test performance, and provided techniques for assessing the relative influences of various aspects or facets of the testing procedure. The theory provided important guidance for test developers on matters such as the number of items or the optimal allocation of raters.

Cronbach and Meehl's 1955 study established the centrality of validity considerations in testing, and “construct validity” as a unifying theme for interpreting test scores. Whereas traditional approaches to test validity were limited to examining test content or simple correlations with other variables, construct validation was to build a firm, rigorous theoretical basis for score interpretation. Cronbach also drew attention to the plurality of points of view, values, and beliefs for test developers, users, and policy-makers as essential ingredients in the process of test validation.

THE INTERACTIONIST APPROACH

Cronbach (1957, 1975) argued that the rift between experimental and correlational psychologies artificially separated the learner from the learning environment. He maintained that aptitudes—individual differences in response to educational treatments—were as important to understanding and improving educational programs as the typical response or average program effect. In earlier work (Cronbach & Gleser, 1957) on personnel-decision theory he and his colleague concluded that optimal placements must acknowledge the interaction of personal characteristics and job demands. His work on aptitude-treatment interactions (ATI) sought to demonstrate the potential benefits of matching the right type of instruction to students' abilities, motivations, and interests (Cronbach & Snow, 1977; Corno, et al., 2002). His efforts to discover general propositions about who would benefit from alternative methods of teaching and how learning environments could be designed to maximize benefits for all students, led him almost two decades later to suggest a radical shift in the goals of social science research toward a more contextualized inquiry (Cron-bach, 1975). He sharpened the sensitivity of educational researchers to the ways different learners cope with the demands and affordances imbedded in different learning environments, and he advocated the use of intensive local studies and field methods that produced rich narratives of teaching and learning.

PROGRAM EVALUATION

Cronbach built on the ideas of Ralph W. Tyler (1902– 1994), about how teachers should fashion their instruction to fit their students' needs, to pioneer the movement of “formative evaluation”—the idea that assessment is not a yardstick against which students should be measured and ranked but a feedback tool to stimulate teachers' efforts to improve instruction. He saw the evaluator as an educator informed by empirical studies rather than as an impartial observer submitting a verdict and drafting a correctional order. For him, evaluation was a pluralist, inclusive, and open-ended inquiry, sensitive to unexpected issues or unforeseeable events, even if such real-time adjustments spoiled the scientific elegance of the study.

Cronbach led the Stanford Evaluation Consortium, a multidisciplinary group whose work was summarized in a volume published in 1980 (Cronbach and associates, 1980). Two years later, Cronbach published his own book on program evaluation (Cronbach, 1982), in which he argued that the academic study of human affairs is not an exercise in abstraction. Research programs, he claimed, are valuable to the extent they serve the purpose of improving some aspect of the social reality.