Most good theories start with a practical problem, according to social psychologist Claude Steele, PhD.

The theory that has become Steele's greatest contribution to psychology began with a problem he encountered at the University of Michigan in the late 1980s. As a new minority faculty member, he found himself placed on the student recruitment and retention committee. On their first day, committee members were handed a stack of material that included a disturbing chart. It showed the average grades of Michigan students, graphed as a function of their SAT or ACT scores. Not surprisingly, on average, students with higher test scores earned higher grades.

But a different trend jumped out at Steele. At every level of SAT or ACT score, black students got lower grades than white students—even though they entered college with the same skills, at least according to the tests.

"That was a surprise," said Steele, who in September became dean of the Stanford University School of Education.

Eventually, the data led him to a series of experiments that crystallized his theory of "stereotype threat." He showed that when people in a negatively stereotyped group—such as blacks in academia or women in math and science—are about to take a test, the subconscious worry that they might confirm those negative stereotypes undermines their performance by draining their cognitive resources away from the exam. So, even black students who came to Michigan well-prepared for college got lower grades than similarly prepared white students.

Since then, experiments by Steele and others have demonstrated the effects of stereotype threat on minority groups and have tested ways teachers and other educators can lessen stereotype threat's effects.

The interventions, he said, are based on the premise that what makes stereotype threat a strong or weak force is all about the context in which a test or other task is performed.

"Being under stereotype threat is like having a snake in the house," said Steele. "You don't know what would happen, or when it would happen, or where it would happen. But you can't quite relax because something could happen. And so you need some assurance ... to help you downgrade the probability that something bad could happen to you based on your identity."

That assurance can come in different forms. In one study, for example, Steele gave a group of men and women a difficult math test. Normally, women do worse than men with similar math backgrounds on standardized math tests, he'd found. But when Steele and his colleagues simply told the women that on this particular test women do better than men, the reinforcement boosted the women's confidence—and the women performed as well as the men.

Other studies, by Stanford University professor Geoffrey Cohen, PhD, among others, have shown that interventions to reduce stereotype threat can have even bigger, longer-lasting results. In his most well-known research, Cohen found that an intervention that boosted black junior high school students' sense of self-worth—by asking them to reflect on a personal value that's important to them—improved their GPAs by almost half a point over two years.

"What's impressive to me is the notion that in real-world academic settings, these [stereotype threat] processes are not just a small part of the problem, but are turning out to be a major part of the problem," Steele said.

And now, through his work and that of others, understanding stereotype threat may turn out to be a part of the solution.

For more on interventions to beat stereotype threat, read the article "How to Close the Achievement Gap" in the September 2011 Monitor.