A fifteen-minute exercise may help overcome a lifetime of racial stereotyping

i-eca0cf2af9fc3ac4445c7dff7d8aab70-research.gifThe setting was an integrated suburban middle school: nearly evenly divided between black and white students. As is the case in many schools, white students outperformed black students both in grades and test scores. But how much of this difference is attributable to real differences in ability? After all, black kids grow up "knowing" that white kids do better in school. Perhaps this was just an example of kids living down to expectations.

A simple experiment would help find out. A team led by Geoffrey Cohen found a group teachers who taught the same 7th-grade course and were willing to participate in a very brief intervention. In each class, half the students would participate in an exercise designed to affirm their own abilities, while the other half completed a similar exercise that asked them to instead consider the abilities of others.

The self-affirming group of students (evenly divided between African Americans and European Americans) was asked to select from a list of 12 "values," such as having good friends or being good at art, the value that was most important to them personally. They then wrote a paragraph about why the value was important to them.

The other group selected the least important value and explained why it might be important to someone else. Teachers were unaware of which students were in which group. The researchers then tracked the students' performance for the remainder of the school year. Here are the results:


At every performance level, this chart (adjusted for covariates) shows that black students who completed the 15-minute affirmation exercise got better grades than students who did not (control). Interestingly, there was no similar effect for white students, suggesting that the effect of the exercise may have been to remove the handicapping of those students due to racial stereotyping. Even this short intervention asking students to reflect on their personal values appears to cause a significant effect.

How significant? 70 percent of African American students benefited from the intervention. The chances of this effect occurring due solely to chance are less than 1 in 5,000. But why would the effect of such a short exercise be so dramatic? The authors speculate that the benefits are cumulative: when students faced challenges shortly after they participated in the exercise, those who had reflected on their values performed slightly better. This gave them the confidence they needed to do better the next time a challenge was faced. Each successive success prepared students to face future challenges; in the end, this all added up to better performance.

In the second year of the study, the researchers kept more frequent tabs on students, checking grades ten times over the course of the year. They found that students who had affirmed their values did indeed rebound more quickly from setbacks and avoided the downward spiral that students in the control condition often fell into.

Does this study demonstrate that only small interventions are necessary to solve the racial disparity in educational achievement? No. Many black students are in districts that receive less funding than white students, or have parents with less education than white students. For these kids, much more is required than a quick exercise. And these results don't appear to be as effective for the lowest-performing students in this group. But when other factors are equal, it may not take much to eliminate entirely the effects of racial stereotyping for many children.

For more on stereotype threat, see here, here, and here.

Cohen, G.L, Garcia, J., Apfel, N., & Master, A. (2006). Reducing the racial achievement gap: A social-psychological intervention. Science, 313, 1307-1310.

More like this

[This article was originally posted in February, 2007] The setting was an integrated suburban middle school: nearly evenly divided between black and white students. As is the case in many schools, white students outperformed black students both in grades and test scores. But how much of this…
In American high schools, black students typically perform worse than their white peers, which can damage their self-esteem and their future prospects. Studies have found that the fear of living up to this underachieving stereotype can cause so much stress that a child's performance suffers. Their…
As I believe I've said before, if anything good has come from the Larry Summers debacle of a few years ago, it's that it inspired some really interesting research on gender differences in math. If you've been reading this blog for a while, you've probably guessed that one of my favorite topics in…
When Jim and Nora talk about the social groups in their school, they matter-of-factly categorize almost every fellow student into stereotyped pigeonholes. There are the nerds, the rockers, the cools, the goths, and of course, the jocks. The assumption, naturally, is that none of these groups…

I blogged about this too a few months ago. Really amazing stuff.

My question is, what do the Bell Curve advocates think of this? It always seems they require the assumption of an intrinsic and immutable IQ value for anything they say to be taken seriously. Then you have these studies on stereotype threat and it's so obvious that blacks in this country are facing stereotypes with a dramatic affect on performance while whites are not. It's really just amazing research, and should be incorporated into curricula and education schools everywhere, and really throws the "one race is smarter than another" crap that the bigots are trying to propogate eternally out the window.

Thank you for this interesting post! The graph is a bit confusing to me though: Can you tell me how the 'performance clusters' were formed? Did the researchers use a different performance measure for students of differing racial backgrounds? Is their performance difference before and after the intervention measured? I understand that the main point is that the differential effect of affirmation points to a socio-cultural background for performance differences, but I am curious as to what the researchers actually did..

By Jan-Maarten (not verified) on 15 Feb 2007 #permalink


The performance clusters were just even divisions within each racial group according to previous-year's grades.


I'm not sure, but I think the Bell Curve advocates would claim that while stereotyping is a problem, it doesn't explain all the racial differences in IQ. Then they'd start talking about identical twin studies.

Honestly, our side uses some of the same tactics: we claim that IQ doesn't explain all disparities, and that there are other factors, such as stereotyping. I think economic factors are even more important than stereotyping, and they are the root of many of the racial problems we see today. I guess the main difference between me and Charles Murray is that I'd like to remove the obstacles to equality and see how close we can get to true equality. Murray wants to embrace the idea that some people are inherently superior to others.

I once taught undergraduates and had a few classes with a mix of middle-class whites and lower-class blacks. My policy was to always engage every student several times the first week, even if only to ask their opinion. By the end of the second week, the reticence of the blacks -- and the quieter whites -- had vanished, along with any question of differences in performance.

One difference I did see was that blacks volunteered more. I still don't know what to make of that.

I blogged on this at GNXP when the study came out -- it's mostly misleading:


(Darth Quixote's update is just that one paragraph; after that is the continuation of the original post.) Read the whole thing, since it's not overly long, but basically when you look up their descriptive statistics, the effect of eliminating Stereotyp Threat (assuming that's what is happening) is either very weak or 0. The strongest predictor was simply having a particular teacher (maybe that's where the better students are funneled), and previous academic performance (no surprise). Only 50% of those given permission slips returned them & consented, which likely removed many of the less intelligent or less conscientious students. And on and on.

Also, the "target class" was social studies -- a joke if you're going to try to show that you eliminated ST. Try it in a math class and watch the results shrink even more. Also, the authors didn't note that (as Steele has admitted in journals) eliminating ST doesn't shrink the black-white gap: adding ST will exacerbate it, and then un-doing it returns the gap to its typical value of ~1 SD.