Cronbach, Lee J.

(b. 1916, Fresno, California; d. 2001, Palo Alto, California). Ph.D. Educational Psychology, University of Chicago, Illinois; M.A. University of California, Berkeley; BA, Fresno State College.

Cronbach took the Stanford Binet IQ test at age 5 and reportedly scored a 200, and he was then placed in [Page 95]Lewis Terman's landmark study of the intellectually gifted. He finished high school when he was 14 years old and college when he was 18. Thurstone's work on attitude measurement provided early influence on Cronbach's lifelong engagement with educational and psychological measurement. However, Cronbach refused to narrowly specialize. “Weaving strands into a tapestry was what I enjoyed, not spinning the thread.”

His life's work spanned three major domains: (a) measurement—where he invented the most widely used reliability coefficient today, the Cronbach alpha; developed a seminal work in generalizability theory; and developed a significant exposition of the core meanings of construct validity; (b) his interactionist approach to instruction and to educational research, in the pursuit of which he argued convincingly for the complementarity of experimental and correlational approaches to inquiry and in which he contributed significantly to ensuring that complex human phenomena were not studied via simple main effects but rather via interactionist approaches, labeled by Cronbach and his colleague Dick Snow as aptitude-treatment interactions. Cronbach later abandoned his pursuit of stable interactions, but he retained his commitment to honoring complexity. This commitment is evident in his third domain of contribution, (c) his work in program evaluation work.

Cronbach envisioned and championed a major educative role for evaluation and legitimized the formative, program improvement purpose for evaluation. He brought a conceptualization of who the audience for evaluation should be—the policy-shaping community, which included not only decision makers but advocacy groups; educators; the media; and interested citizens. Cronbach envisioned evaluation's primary role as educating the policy-shaping community about the nature and contours of persistent educational and social problems and how best to address them. He understood evaluation as fundamentally political, largely because it is conducted in politicized settings, often with high stakes.

Instead of one massive experiment or quasiexperiment (the “horse race” model of evaluation, said Cronbach), he favored an eclectic, broad-based, open methodological approach to evaluation; a fleet of smaller studies, each pursuing an important case or component of the policy or program under study. Cronbach encouraged evaluators to design evaluations to understand in some depth the nature of each context and the quality of the intervention in that context. Over time, then, with many such studies, the policy-shaping community could learn in some depth about that social problem and how best to address it. In addition, Cronbach encouraged evaluators to involve members of the setting in the evaluation study and to provide feedback throughout the course of the study (for program improvement purposes) rather than just at the end.

Cronbach theorized his approach in his 1982 Designing Evaluations of Educational and Social Programs as the units, treatments, observations, and settings (UTOS) framework. This framework championed the importance of external validity in evaluation—will this intervention repeat its successes in a different context characterized by A, B, C? This was in marked opposition to Campbell's continuing privileging of internal validity in evaluation studies.

...