On March 9, Wilkes presented his findings during the "Quality Education: Evolution and Revolution" conference in Orlando, Florida, sponsored by the Center for Applications of Psycological Type (CAPT). "Our initial studies, completed in 1994 and 1995, have led us to undertake a major research program to ascertain the nature and size of these cognitive-style biases in the SATs," says Wilkes. "We are putting the field on notice that a major change in how we look at the SATs is on the horizon."
Wilkes' presentation was based on two studies completed under his direction by WPI undergraduates. The first, by former WPI student Benjamin Kibler, David Kingsland 93, who will receive his master's degree in electrical engineering in May 1996, and Charles McTague '95, who formed the theoretical framework for the work; the second, by Danielle Batey '97 of Fairfield, Maine, Paula Brezniak '97 of Cherry Valley, Mass., and Ashwin Purohit '95 of Albany, NY, focused on the actual testing. WPI, CAPT and the Worcester public schools collaborated on the projects, which were sponsored by Keith McCormick '91, founder of Comprehensive College Preparation Services, a Worcester-based company that offers a range of college preparation classes and tutoring and has a natural interest in the PSATs and SATs.
The authors gathered data from 530 high school students in two waves of data collection at four Worcester-area urban and one suburban high school. The first group of 280 students took the last round of the old SAT exam in 1993; the second group of 280 students, took the new version of the SAT in 1995, the year it was introduced in response to complaint of gender, race, and cultural bias in the exam.
"The average score had gradually declined over the years as the pool of students taking the exam became larger and more socially and culturally diverse," says Wilkes. "We concluded that despite revisions that focused on efforts to increase equity in these areas, the fundamental nature of the SAT task environment a prestructured and timed test still creates a bias in favor of some cognitive types in favor of others. The bottom line for us, in our admittedly limited studies, was that SAT scores differences related to cognitive diversity were enhanced in the new version not lessened as the exam's creators and those who pushed for the change expected.
"Our analysis of the changes suggests that time pressures on students may have increased by replacing a percentage of the analogies with additional reading comprehension sections. Our samples show that differences in SAT score gaps are as large or larger than previous gaps in scores attributed to race, social class and coached vs. uncoached students, which began the current SAT debate."
Wilkes is working with teams of students to gather data on 4,000 Worcester-area high school students using the PSATs to determine the size and nature of cognitive biases. "Should these pilot-study findings replicate on the larger database, they will kick off a national debate because, for some cognitive types, the SATs do tell you something, but for others there is no correlation between these exams and college performance."