Randy Elliot Bennett

Randy Elliot Bennett is an American educational researcher who specializes in educational assessment.^{[5] [6] [7]} He is the "Norman O. Frederiksen Chair in Assessment Innovation" at Educational Testing Service.^[8] He has received Fellow status in the American Educational Research Association (AERA) in 2017^[5] and was elected to membership in the National Academy of Education in 2022.^[6] Bennett has also previously served as president of the National Council on Measurement in Education (NCME).^[9]

Randy Elliot Bennett
Born	Brooklyn, New York
Occupation	Educational Researcher
Nationality	American
Notable awards	AERA E.F. Lindquist Award[1] AERA Cognition and Assessment SIG Outstanding Contribution to Research in Cognition and Assessment Award[2] NCME Bradley Hanson Award[3] Teachers College, Columbia University Distinguished Alumni Award[4]

Publications

Bennett is author or editor of ten books, as well as over 300 journal articles, chapters, and technical reports. Those publications have concentrated on several themes. The 1998 publication, Reinventing Assessment: Speculations on the Future of Large-Scale Educational Testing,^[10] presented a three-stage framework for how paper-and-pencil tests would gradually transition to digital form, eventually melding with online activities, blurring the distinction between learning and assessment, and leading to improvements in both pursuits. A series of subsequent publications built upon the work of Robert Glaser, Norman O. Frederiksen, Samuel Messick, James Pellegrino, Lorrie Shepard and others to create a unified model for formative and summative assessments under the Cognitively Based Assessment of, for, and as Learning (CBAL) initiative.^[11][12] This work, noted in the citations for both the E.F. Lindquist Award and his AERA Fellow designation,^{[5] [1]} is described in two journal articles, Transforming K-12 Assessment^[13] and Cognitively Based Assessment of, for, and as Learning.^[14] The latter publication articulated assumptions for the CBAL assessment model in a detailed "theory of action," which described the assessment system components, intended outcomes, and the action mechanisms that should lead to those outcomes, predating the generally recommended use of that device in operational testing programs.^{[15] [16]}

The journal article, "Formative Assessment: A Critical Review",^[17] questioned the magnitude of efficacy claims, the meaningfulness of existing definitions, and the general absence of disciplinary considerations in the conceptualization and implementation of formative assessment.^[18] The article encouraged a deeper examination of premises, more careful consideration of effectiveness claims, and a move toward incorporating domain considerations directly into the structure and practice of formative assessment.^{[19][20] [21]}

Two reports--Online Assessment in Mathematics and Writing^[22] and Problem Solving in Technology Rich Environments^[23]--documented studies that helped set the stage for moving the US National Assessment of Educational Progress from paper presentation to computer delivery.^{[24] [25]}

Several more recent articles called attention to the need for testing companies and state education departments to exercise caution in using artificial intelligence (AI) methods for scoring consequential tests. That theme was developed in a book chapter, "Validity and Automated Scoring",^[26] and summarized in The Changing Nature of Educational Assessment.^[27] These publications note that in automated essay scoring, for example, caution is needed because of the inscrutability of some AI scoring methods, their use of correlates that can be easily manipulated for undeserved score gain, and the routine practice of building scoring algorithms to model the judgment of operational human graders, thereby unintentionally incorporating human biases.

Bennett's latest work centers on equity in assessment. The commentary, The Good Side of COVID-19,^[28] makes the case that standardized testing, and educational assessment more generally, must be rethought so that they better align with the multicultural, pluralistic society the US is rapidly becoming. In a follow-up article, "Toward a Theory of Socioculturally Responsive Assessment",^[29] he assembles assessment design principles from multiple literatures and uses them to fashion a definition, theory, and suggested path for implementing measures more attuned to the social, cultural, and other relevant characteristics of diverse individuals and the contexts in which they live. That line of thinking is elaborated upon in "Let's Agree to (Mostly) Agree: A Response to Solano-Flores".^[30]

A logical extension of the ideas explored in Toward a Theory of Socioculturally Responsive Assessment is to personalize so that assessments are adapted to the characteristics of the individual. In "Personalizing Assessment: Dream or Nightmare?"^[31] Bennett articulates why personalized assessment is needed, what precedents exist for it in educational measurement, what it looks like in practice, and why it should be a concern.

In "Rethinking Equity and Assessment Through Opportunity to Learn",^[32] he considers the impact of structural inequity in US society and its implications for educational assessment. That consideration leads to a theoretical model depicting how macro-level systems factors and community, home, and school factors, interactively influence opportunity to learn across generations, thereby creating and perpetuating disparities among racial/ethnic and socioeconomic groups that are reflected in test scores, grades, and life outcomes.