Are tests a fair way to predict individual performance?
Daniel Robinson, professor at the University of Texas at Arlington, delves into this question.
Dan Robinson is currently the Interim Chair of the Department of Higher Education, Adult Learning, and Organizational Studies and also the K-16 Mind, Brain and Education Endowed Chair at the University of Texas at Arlington. Previously, he served as Associate Dean of Research from 2020-2024 and Chair of the Department of Curriculum and Instruction at UTA from 2017-2020, and as Director of the School of Education at Colorado State University 2012-2013. He received his Ph.D. in Educational Psychology in 1993 from the University of Nebraska where he majored in both learning/cognition and statistics/research. He has taught at Mississippi State University (1993-1997), the University of South Dakota (1997-1998), the University of Louisville (1998-1999), the University of Texas (1999-2012), and Colorado State University (2012-2015).
Dr. Robinson is past Editor of Educational Psychology Review and Associate Editor of the Journal of Educational Psychology. He has also served as editor of Educational Psychology Review and as an editorial board member of nine refereed international journals: American Educational Research Journal, Contemporary Educational Psychology, Educational Technology, Research, & Development, Journal of Behavioral Education, Journal of Educational Psychology, Journal of Experimental Education, Reading Research and Instruction, Research in the Schools, and The Open Education Journal.
He has published over 100 articles, books, and book chapters, presented over 100 papers at research conferences, and taught over 100 college courses. His research interests include educational technology innovations that may facilitate learning, team-based approaches to learning, and examining trends in articles published in various educational journals and societies. Dr. Robinson was a Fulbright Specialist Scholar at Victoria University, Wellington, New Zealand in 2011.
Testing and the Paradoxes of Fairness
How can admissions officers, employers, and scholarship committees maximize the accuracy of prediction of individual performance while minimizing adverse impact due to group differences? Testing offers a straightforward solution to the first half of this problem. Tests are the best way to predict how someone will perform in school, in the military, in medicine, or while controlling airline traffic and flying a plane. Tests are also useful beyond personnel selection, such as for selection of a college major or courses. However, the other side of this problem is more complex. Using tests is always accompanied by group differences that could result in continued systemic discrimination by limiting opportunities for those who are marginalized. We discuss an approach to using tests that incorporates evidence, transparency, and societal values to maximize efficiency and fairness.
Throughout their long history, the chief complaint against tests is that they divide people:
• In admission they divide those who get in from those who don’t,
• In hiring they divide those who get the job from those who don’t,
• In licensing they divide those who get to practice their trade from those who cannot,
• In the granting of scholarships, they divide those who win the prize from those who do not.
There are winners and losers, and so it is natural that those who lost will often blame the test. And, in a very real sense, they are right. Often, if there had been no test, the selections would have been done less accurately, there thus would have been a greater chance that those who were not chosen would have been selected. But there are costs associated with an approach whose natural endpoint is random selection; the costs of less efficient utilization of both human and physical resources.
Read More:
[Cambridge University Press] – Testing and the Paradoxes of Fairness
