Giving computers the jitters helps explain human behavior

Penn State researchers have developed a computer program that mimics the results when a human appraises a task as threatening and feels worried before starting.

Dr. Frank Ritter, associate professor in the School of Information Sciences and Technology (IST), says, “In this study, we used our program to mimic the results achieved by people who can be seen as having a type of math anxiety. However, the program could also be used to study the effects of feeling threatened or worried before driving a car, using a computer or other stressful task — and to help develop remedial strategies.”

The results were described in a paper, “Using Cognitive Modeling To Study Behavior Moderators: Pre-Task Appraisal and Anxiety,” presented this month at the 48th Annual Meeting of the Human Factors and Ergonomics Society in New Orleans. The authors are Ritter, Andrew Reifers, an IST doctoral student; Laura C. Klein, assistant professor of biobehavioral health, College of Health and Human Development; Karen Quigley, New Jersey Medical School, University of Medicine and Dentistry of New Jersey; and Mike Schoelles, Rensselaer Polytechnic Institute.

Ritter explains that, in general, when people appraise a task as threatening, they have greater difficulty processing information. On the other hand, people who appraise a task as challenging perceive themselves as having more than enough resources to perform the task, don’t respond with stress and perform well on the task.

For their study, the Penn State researchers developed a computer program that simulated how people who felt threatened would respond when asked to perform a serial subtraction problem (i.e., start with a four-digit number such as 2,522 and repeatedly subtract from it a specific one or two-digit number such as 7 or 13). The handicapped program matched fairly well the published data on humans threatened by the task: performance accuracy remained the same but performance speed decreased about 25 percent.

“The correspondence suggests that the modifications we made to the program to simulate feeling threatened and worried provides a plausible explanation for the decreased performance in people on this task,” Ritter explains.

He adds that the results also suggest two ways to help individuals to switch their pre-task appraisal from threatening to challenging. First, they could increase their knowledge through instruction and practice. Second, they could continuously check their results to bolster their confidence while they work. The study was supported by a grant from the Office of Naval Research.