Science Watch

By pretty much any measure, Cupertino High School in northern California is a successful place. Perched in the heart of Silicon Valley, Cupertino sent 85 percent of its senior class to college in 2009. Hundreds of its students choose among a dozen advanced placement classes each year. By and large, Cupertino's kids are doing well.

But some are doing better than others. On average, students at Cupertino far exceed California's target on the state's Academic Performance Index measure. The target is 800; Cupertino students scored 893. But Latino students, who make up 10 percent of the school's population, averaged a score of 780, just under the statewide goal.

Cupertino High School is far from alone. For decades, educators have struggled to close the "achievement gap," the persistent differences in test scores, grades and graduation rates among students of different races, ethnicities and, in some subjects, genders.

In fact, Cupertino's achievement gap may be smaller than average. Nationwide, the data are striking: 94 percent of white young adults have earned a high school degree by age 24, but only 87 percent of blacks and 78 percent of Latinos have done the same, according to the U.S. Census Bureau. A 2009 U.S. Department of Education review found that black fourth- and eighth-graders scored lower than their white counterparts on math and reading in every state for which data were available. Some of the differences can be explained by socioeconomic factors, but not all.

Educators have been chewing over the problem for decades. Many of the solutions they've proposed have been large-scale teaching or curriculum reforms that are time-consuming, expensive or both. Meanwhile, despite some successes in individual schools and programs, the nationwide achievement gap has narrowed only slightly in the past decade.

But recently, a group of social and cognitive psychologists have come at the problem in a different way. These psychologists' approach is based on the idea that at least some of these academic disparities aren't the result of faulty teaching or broken school systems, but instead spring from toxic stereotypes that cause ethnic-minority and other students to question whether they belong in school and whether they can do well there. While such a major problem might seem to require widespread social change to fix, the psychologists are finding evidence that short, simple interventions can make a surprisingly large difference. Quick classroom exercises that bolster students' resistance to stereotypes and change the way they think about learning can have dramatically out-of-scale effects, these researchers say.

And indeed, they've gotten dramatic results. In one of the best-known studies, low-performing black middle school students who completed several 15-minute classroom writing exercises raised their GPAs by nearly half a point over two years, compared with a control group.

Such astonishing results have struck some observers—particularly nonpsychologists—as nearly magical, and possibly unbelievable. But a growing body of evidence is showing that the interventions can work, not only among black middle school students, but also for women, minority college students and other populations.

"When this was first described to me, I was skeptical," says physics professor Michael Dubson, PhD, of the University of Colorado–Boulder, who worked with psychologists there on a study with women physics students. "But now that I think about it, we all know that it's possible to damage a student in 15 minutes. It's easy to wreck someone's self-esteem. So if that's possible, then maybe it's also possible to improve it."

Stereotype threat

Many of the new interventions are based on the concept of "stereotype threat," first identified by psychologist Claude Steele, PhD, in the mid-1990s. He showed that when people who are about to take a test are reminded of negative stereotypes about their racial, ethnic or other group, the subconscious worry that they might confirm those stereotypes undermines their performance by sapping cognitive resources that they could be using to do better on the exam.

Psychologist Geoffrey Cohen, PhD, now at Stanford University, wondered if there might be a way to inoculate students against the effects of stereotype threat by buffering their sense of self-worth and positive identity. He and his colleagues Julio Garcia, PhD, and Valerie Purdie-Vaughns, PhD, tested their theory at a suburban, low-to-middle-income middle school in Connecticut that was about half black and half white. At the beginning of the school year, about 400 seventh-graders spent 15 minutes doing a classroom writing exercise. Half of the students were asked to pick a personal value, such as athletic ability or relationships with friends and family, and then write about why that value mattered to them. A control group wrote about why a value that didn't matter to them might be important to someone else. The students did the exercise one or two times at the beginning of the school year.

In a study published in Science in 2006 (Vol. 313, No. 5791), the researchers found that the short exercise reduced the achievement gap between the black and white students in the class by up to 40 percent over one school term, and that it was particularly effective for low-achieving black students, halving the percentage of black students who got a D or below in the class.

Three years later, Cohen and his colleagues published a follow-up paper in Science (Vol. 324, No. 5925) in which they tracked the original group of students through the eighth grade. Amazingly, the effect lasted—the low-achieving black students who had completed the values-affirmation exercises raised their GPAs by four-tenths of a point (on a four-point scale) compared with the control group, and were less likely to need to repeat a grade. The intervention didn't have any effect on white or high-achieving black students.

What was going on? Cohen hypothesizes that the exercise started a self-reinforcing loop. It strengthened students psychologically at a crucial time, right at the beginning of the school year. By reminding them of something personal that mattered to them, it reduced their stress level and pushed down the distracting worries, brought on by stereotype threat, that they might not measure up. Because of that, they were able to do better on one crucial first exam or homework assignment. Doing well early on boosted their resilience to stereotypes even more, leading to another successful test—and, perhaps, permanently changing their academic trajectory.

Researchers don't yet know the precise cognitive mechanisms at work in the study, says cognitive psychologist Akira Miyake, PhD, of the University of Colorado–Boulder, who has worked with Cohen on the research. But he hypothesizes that pushing down extraneous worries could increase the amount of working memory that students have available to concentrate on schoolwork.

Since Cohen's first studies, he and other researchers have tested the model in other populations, including among female physics students. Women in science face some of the same stereotypes, and achievement gaps, that blacks and Latinos face in the rest of academia. In 2008, women earned only 20 percent of the bachelor's degrees and 18 percent of the doctorates awarded in physics, according to the American Physical Society, even though they earned nearly 60 percent of bachelor's degrees overall.

Miyake and psychologist Tiffany Ito, PhD, along with Cohen, had about 400 students in an introductory physics class at the University of Colorado–Boulder complete a personal-value writing exercise similar to the one used in the Connecticut middle school study. The college physics students did the exercise twice, once during the first week of the semester and once right before the first exam.

In a study published in Science in 2010 (Vol. 330, No. 6008), the researchers found that women who did the self-affirmation exercise did significantly better in the class: Among the control group, about 60 percent of the women earned C's and less than 30 percent earned B's. In the self-affirmation group, as many women earned B's as earned C's. The exercise didn't affect men's grades in the class.

Dubson, the physics professor who taught the class and who was initially skeptical about the intervention, was blown away by the results.

"Holy mackerel, most of our curriculum is based on the theory that the most important thing in learning is 'time on task,' because learning is messy and hard," he says. "I can't think of anything else that you could do in 30 minutes that would have a measurable effect on exam scores."

Other brief interventions, based on related psychological concepts, have also shown promise for reducing achievement gaps. For example, many students face a tough transition to college, but minority students have an added complication: stereotypes that undermine their sense that they belong on campus. A black student struggling to adapt during his first year of college might subconsciously feel that his social or academic troubles were due to race rather than normal freshman jitters. In one recent study, Stanford psychology professor Gregory Walton, PhD, with Cohen, found that boosting a sense of belonging among black college freshmen could improve the students' grades all the way through their senior year. In the study, published in March in Science (Vol. 331, No. 6,023), Walton and Cohen asked 90 black and white college freshmen to read vignettes, purportedly written by older students, describing how school was difficult at first but how eventually they found their social and academic niche. Then the participants had to write essays about what they had just read. A control group read vignettes unrelated to social belonging.

The goal was to change students' attitudes about their sense of fitting in, and to subtly let them know that their "fish-out-of-water" worries were common to all students, and not a sign that they weren't meant for college. It worked. Over the next three years, the black students in the treatment group earned GPAs nearly one-third of a point higher, on average, than those in the control group—roughly halving the black-white achievement gap.

Scaling up

In science, the proof is in the replication. So right now, Cohen, Miyake and the other psychologists are again working with physics professor Michael Dubson to replicate their "women in college physics" study.

During the original study, the personal-value writing exercise went smoothly. But this time, a student presented Dubson with a problem. After the first writing exercise, the student came up to Dubson and said, "Hey, I know what you're doing—I've read about this research."

Dubson asked her not to tell her classmates about the purpose of the writing exercise, and no other student mentioned knowing anything about the research. But the glitch illustrates a challenge for psychologists: If these interventions became more common, and students—particularly older, savvier students—started to recognize them, would they still work?

Some evidence suggests they wouldn't, at least not as well. In one study, published in 2009 in the Journal of Personality and Social Psychology (Vol. 97, No. 5), Cohen and University of California–Santa Barbara, psychologist David Sherman, PhD, found that a self-affirmation exercise didn't work as well when they told the participants that the point of the exercise was to boost self-worth. That might be because telling them the purpose of the exercise caused them to see it as simply a means to an end, the researchers suggest, which undermined its self-affirming punch.

The question of whether students could "know too much" about these interventions is just one issue among many that psychologists are facing as they aim to scale up the research from a trial in a single classroom to a school-wide or school-district-wide intervention.

Other issues may be even thornier. For example: How do we know that what works in one context will work in another? Most of these interventions have been tested in one classroom or one school. Usually, those classrooms are diverse—they include both men and women, or both black or Latino and white students, in order to let the researchers compare the groups. But many classrooms around the country are ethnically and racially homogeneous. Would an intervention designed and tested in a diverse school also work in a homogeneous one? Researchers don't yet know.

Along the same lines, will something that works with elementary school students translate to high school or college students, and will something that works with college students translate to lower grades? Much more refining and testing needs to be done to find out, the researchers say.

And what about delivering the interventions? The psychologists who design them can't be around to train every teacher, as they are during their experiments. And as Dubson's experience shows, the experiments depend on some subtlety on the teachers' part. They must be able to explain convincingly to students why they're doing the writing exercises, without giving away the real purpose. Could teachers be trained to do that effectively on a large scale?

Figuring out the answers to these questions is important, says Purdie-Vaughns, who has worked on several of the values-affirmation studies with Geoffrey Cohen, because scaling up this research is what really matters in the end. Purdie-Vaughns spent her first years after college working with disadvantaged elementary school students for the I Have a Dream Foundation. Trying to figure out what was holding those children back academically is what drew her to graduate school in psychology. Now that her research and others' is providing some answers, Purdie-Vaughns is eager to figure out how to move the knowledge from the lab to the school district.

"I think the delivery is as important as the basic science behind this research," she says.

She's been traveling the country, talking to teachers and school district officials who are interested in working with her on larger-scale intervention studies.

Stanford graduate student Dave Paunesku, meanwhile, is working on the delivery issue from another angle. He's developing online versions of the interventions that students can access from home or school computers. He's aiming to test the program in at least 50 schools next year.

He's already begun testing in nearby Cupertino High School, where Principal Kami Tomberlain says the work fits perfectly with the school's philosophy.

"I think it points to the self-fulfilling prophecy of stereotypes, and the way that people can swallow what's being said about them without even noticing that they're doing it," she says. "That's one of the things we [teachers] talk about all the time here. How can we communicate that what we're teaching is important, and that they're all capable of learning it?"

The psychologists, meanwhile, know that their research by itself won't solve the achievement gap for Tomberlain and other educators.

"I think it's really important to understand that this is not a silver bullet," says Walton. "If you delivered a great [psychological] intervention but the teaching was terrible, the intervention would have no effect."

But in the right places, under the right circumstances, these interventions might just make a difference in students' lives.