Support Provided ByLearn More
Body + BrainBody & Brain

Psychologists Discover Easy Ways to Make Standardized Testing Less Biased

ByConor GearinNOVA NextNOVA Next

Receive emails about upcoming NOVA programs and related content, as well as featured reporting about current events through a science lens.

When Ahmed Mohamed, 14, was arrested in Texas for bringing the clock he made to school to show his teacher, he encountered just one of the many ways that students can suffer from stereotypes, especially in science and technology. Ethnic minorities and women often experience this pressure because they do not match expectations for what scientists looks like. But a new psychology experiment shows how stereotyped students can use that pressure to their advantage.

Support Provided ByLearn More
brazil-standardized-test
High school students in Brazil take a national standardized test.

Deeply ingrained stereotypes like “girls can’t do math” tend to make people underperform. It’s an effect known as stereotype threat, which arises when people fear that they’ll confirm a negative stereotype about their group. Negative stereotypes about testing can lead to lower scores on standardized tests like the SAT and GRE, harming students’ chances of getting into top schools. Yet that stereotype threat can motivate students to score higher than their socially privileged peers, according to a study coming out this November in the Journal of Experimental Social Psychology by psychology professors Allison Seitchik of Merrimack University and Stephen Harkins of Northeastern University.

“When they’re threatened, they can do better—as long as they know what to do,” Seitchik said.

The researchers handed women college students a 120-problem math test. For some students, the instructions read that the test “has been shown to produce gender differences.” Reading this was intended to induce stereotype threat. And indeed, when students were threatened, they solved 12 problems less on average than non-threatened students.

But when students were threatened and also told to use a more efficient problem solving method, they not only did as well, they correctly answered 13 more problems than non-threatened students also told to use the better method—reversing stereotype threat’s usual effect.

“I thought it was pretty cool that [we]were finding the opposite of what people had found with stereotype threat,” Seitchik said. “If you just tweak their approach…you eliminate the stereotype threat effect.”

Seitchik used the same experimental design as other studies of stereotype threat, but took things a step further to look at specific problem solving strategies.

While psychologists have documented the stereotype threat’s effects, they still disagree on the mental process that causes it. Some argue that worries about confirming a negative stereotype take up short-term memory space which could be used for problem solving. This is called the working memory explanation.

Other psychologists claim that, first and foremost, stereotyped students want to beat expectations, and that it’s this extra effort which traps them. This is known as the mere effort explanation. Giving the test their all, stereotyped students tend to choose problem solving strategies they trust most instead of the most efficient ones. Students then try to solve the whole problem on math tests instead of using logic to eliminate wrong answers—ultimately slowing them down.

“In this study, the authors beautifully identified a situation” where the mere effort model explains stereotype threat, said University of Waterloo psychology professor Christine Logel, who co-wrote a recent review of stereotype threat studies. But in other situations in which stereotypes are in play, the way stereotype threat works could depend on many other factors, she said.

Seitchik’s experiment showed a way to make tests better reflect the students’ true abilities. Instructing students on the best test-taking methods—for example, using logic to eliminate answers—lets them put their extra effort to its best use.

“If we can do this with females that are younger…then we can keep them in STEM,” Seitchik said.

But these findings are not just relevant for gender stereotypes. It works the same for students of any gender facing negative ethnic stereotypes about academic performance. For any stereotyped group, if researchers find the problem-solving method students use when trying their hardest, a simple instruction to use a different one can raise scores, Seitchik said.

Another way to make tests more accurate assessments of ability is to stop asking questions about gender and ethnicity before standardized exams, Seitchik said. Though this would not eliminate stereotype threat, it could lessen it.

“Don’t ask students’ race and gender in the front page of the test,” Logel said. “It’s like flashing a neon sign that says, ‘Race is a factor on this test. Gender is a factor.’ ”