|
By Lawrence M. Rudner
CAT is a fundamental algorithm underlying the administration of the GMAT exam and many other high-quality assessments. The algorithm estimates the ability of test takers “live” as they respond to test questions—answer a difficult GMAT question correctly, for example, and CAT gives us the ability to make your next question even more challenging. In other words, CAT helps us tailor the GMAT exam to the test taker in real time. CAT is one of the reasons why the GMAT retains its unparalleled effectiveness and validity as an independent indicator of how well students will do in the core business school curriculum.
The Council’s CAT conference, held June 7–8 in Minneapolis, Minnesota, USA, drew 180 people from 16 countries and convened top researchers and practitioners from around the world to talk about their most recent research in CAT applications, procedures, models, and technology.
June’s meeting was the first conference truly devoted to computer-adaptive testing in 25 years. In a sense, our conference continued a series of meetings on CAT that were supported by the U.S. Office of Naval Research a quarter of a century ago. Back then, CAT was more theoretical, and the central questions had to do with applying theory to practice. Now that computer-adaptive testing is widely applied, the key questions are around what the practice shows. (We have sometimes found, for example, that algorithms that look very good on paper don’t work in real-world use.) At the GMAC® conference, we were able to share insights into what works and what doesn’t; research discussed in Minneapolis will help enhance the practice of CAT worldwide—and the GMAT exam.
|