Measurement, 6: 124, 2008 Copyright © Taylor & Francis Group, LLC ISSN 1536-6367 print / 1536-6359 online DOI: 10.1080/15366360802131635

WEBSITE COMMENTARY

How Cognitive Science Challenges the Educational Measurement Tradition Robert J. Mislevy University of Maryland The procedures through which measurement specialists investigate validity, establish reliability, and ensure fairness are enmeshed with the language and worldview of trait and behavioral psychology. The resulting narrative space articulates poorly with an emerging integration of individual, situative, and social perspectives on cognition—a “sociocognitive” perspective, in the terminology of Atkinson, et al. (2007). Cognitive science challenges the educational measurement tradition to bridge this widening chasm. Doing so entails a broader conception of the nature of proficiency and ways it is evidenced. It means rethinking what we are actually doing when we use measurement models. It requires an articulation between, on the one hand, coarser-grained, between-persons measurement models for studying patterns in behaviors that are at the right level for many practical educational problems, and on the other hand, finer-grained, within-persons models for studying the genesis of those behaviors. In this comment, I sketch key elements of a sociocognitive perspective, note some of its implications for assessment, describe a compatible view of measurement modeling, and call attention to issues that require immediate attention. We see that Borsboom (this issue), Markus (this issue), and Michell (this issue) wrestle with problems that touch on this fundamental challenge, but staying within the measurement tradition limits their progress. Correspondence should be addressed to Robert J. Mislevy, Department of Measurement, Statistics, and Evaluation, University of Maryland, 1230C Benjamin Building, College Park , MD 20742-1115. E-mail: [email protected] This commentary appears in full on our website: http://bear.soe.berkeley.edu/measurement/pubs/ toc612.html