Those Self-Fulfilling Prophecies

You may recall that we were recently discussing stereotypes as a result of an earlier post. Particularly, I was talking about the effects those sterotypes can produce as a result of modifying the expectations of others, making it hard for some people to be taken seriously, and resulting in them having to go that extra mile (or several) as a result.

Well, I’d like to point your attention to a recent study about the direct effects of those stereotypes on the stereotyped. Quoting from an article by AP science writer Randolph E. Schmid:

[Steven J.] Heine and doctoral student Ilan Dar-Nimrod wanted to see how people are affected by stereotypes about themselves. They divided more than 220 women into four groups and administered math and reading comprehension tests between 2003 and 2006. Their results are reported in Friday’s issue of the journal Science.

What they actually did was to provide the different groups with different images and reading materials before they did the tests. They seem to have found significant differences in the results that suggest that having a negative stereotype of yourself in mind actually makes things worse. In other words:

It’s a process psychologists call a stereotype threat, Heine explained. “If a member of a group for which there is a negative stereotype is in a position to test the stereotype, they are likely to choke under the pressure.”

So reminding them of the stereotype affects them.

Here’s what they found:

The women in the lower-scoring groups read essays that either contended that there is a genetic difference between men and women in math ability, or discussed the images of women in art — a reading which did not discuss math but was designed to remind them of being female.

Those two groups not only fell short of the other women, but their performance declined between the two math tests, meaning they scored lower after reading the essays than before.


On the other hand, reading essays that contend there is no natural difference between men and women in math skills lets them go ahead and answer the questions without any added pressure.

And that was also the case with those reading essays arguing that any differences aren’t their fault, but exist because of conditions such as teachers giving boys preferential treatment in the early years of learning math.

That, explained Heine, “may allow a woman to say, ‘This stereotype doesn’t apply to me.'”

How much better did they do?

The women who did better in the tests got nearly twice as many right answers as those in the other groups….

Heinie is a psychology professor at the University of British Columbia in Vancouver.

This is certainly very interesting to me (and not surprising), and it is good to see that such a study was done. I’d also like to see results on studies done to look at simialr effects in other groups that are not “supposed to be able to do” mathematics and science, according to popular stereotypes. That might be quite revealing too. I fear that it is a very significan effect across the board, which is why I worry so much about the effects of stereotypes in the mainstream media, and not just by poorly chosen imagery, but also by simple omission.

I don’t have any means just this minute of looking at the journal Science for more information, but I think it would also be interesting to see what the breakdown of socio-economic and educational backgrounds of these women might be. I don’t know if they included such data.

I ask because I would like to know if there is an enhancement or reduction of the effect as a result of thsoe factors. Or is it that the stereotypes are internalised at such a root level that it does not matter? Put differently, does a firm knowledge or belief (from more formal education, for example) that it is “merely a stereotype” make you immune or not? I suspect that it gives you a mild immunity at the very best. Certainly not much to speak of. I know from first hand experience how much useful energy and effort can be wasted worrying about other people’s image of you in the light of a stereotype they might be applying to you. It certainly affects one’s performance. And that’s just the conscious part of the whole process….


Bookmark the permalink.

13 Responses to Those Self-Fulfilling Prophecies

  1. There have been a few of these kinds of studies recently. I posted one on my blog in August — the study found that women do worse in math tests if primed to think about stereotypes (by being asked questions about gender) and better in mathh if primed to tink about things their good at (eg, by being asked what university they go to). When women were primed to think about their strengths, the there was no significant difference between the performances of men and women.

    One of the researchers cites another study in which it was found that black students’ performance on IQ tests was affected by whether they were asked about their race beforehand.

    It would be really interesting to see similar studies done that broke down the date according to other factors, as you suggest.

    Certainly worry about stereotypes has affected me in the past — confidence has a huge effect on how productive and focussed I am when working. So these reslts don’t suprirse me, except in that they emphasise how internalised these stereotypes are.

    Good post.


  2. Warren says:

    I wonder how stereotypes about women affects men taking such tests.

  3. Rob Knop says:

    Warning — anecdotal evidence coming.

    When I teach intro astro to non-majors, I often have students come in and say, “I am no good at math,” or “I can’t do this.” I try to tell them not to approach it that way — because, in general, the students can do what I’m asking. There was one student last summer who would say “I don’t understand any of this,” even though she had reasoned out the entire question. For several of these students, the expectation that they can’t work out something a little bit challenging causes them to freeze up, and give up even if they’re already most of the way there.

    And, yes, I see this much more often with women than with men, but it’s also true in my class that in general more of the women who are struggling seem to be willing to work hard than the men who are struggling. Some other cultural thing.

    (Of course, there was that post on cosmicvariance recently about how studies showed that American children were more confident in their ability to do math, but performed less well overall…. Clearly a complicated system is involved here. My guess is that the children result is probably tied into the whole “self-esteem is all” thing that was going on in elementary schools at least a decade or two ago, and may still be going on. This result you quote shows that self-esteem is something… but it’s not everything.)


  4. anon says:

    The last paragraph of the Science article from these authors:

    Whether there are innate sex differences in math performance remains a contentious question. However, merely considering the role of genes in math performance can have some deleterious consequences. These findings raise discomforting questions regarding the effects that scientific theories can have on those who learn about them and the obligation that scientists have to be mindful of how their work is interpreted. What President Summers perhaps intended to be a provocative call for more empirical research on biological bases of achievement may inadvertently exacerbate the gender gap in science through stereotype threat.

  5. DancingBear says:

    In general, I always worry about the reliability of testing ITSELF, especially in the form of time-limited trials, as a means to rank people in meritocratic contexts (e.g., deciding who is admitted to grad school).

    Clearly the studies discussed in this post show that testing is very sensitive to biases due to gender stereotyping. I would assume (and there must be research to this effect) that it is also very sensitive to one’s perception of performance in previous testing. I know several very talented people, including my partner, whose self-fulfilling self-stereotyping is that they don’t test well. And then they don’t.

    There other dangerous biases that must be controlled when testing… one certainly is that the grading (if it is not anonymous and there is an element of subjectivity to it, as in grading essays) may be biased by students being typecast early in the course as good academic material, or not; another is that it is often possible (and economical in terms of time) to “study to the test” while acquiring little usable knowledge of a subject.

    Testing well does not necessarily correlate with making strong, creative contributions to a field. Slow thinking can be great thinking.

  6. Clifford says:

    DancingBear…. I think that everyone is in agreement with your summary statemnet. But the point of the study is simply to demonstrate that the thinking *was* affected detrimentally. It is not about whether the tests have anything to do with the “real” conditions that a person might meet in the field.



  7. DancingBear says:

    I understand the point of the study. MY point is that the perhaps the policy remedy to some of the negative effects of such stereotyping is to de-emphasize testing.

    And my broader point is to wonder how much the negative effects extend to the overall performance of a person in “real” field conditions. I suppose they must. But that would probably very hard to investigate scientifically, since experimental psychology, almost by definition, must rely on tests.

    You write “The real conditions that a PERSONA might meet on the field.” Or an ANIMA? (Just humoring you, you probably meant PERSON…)

  8. Clifford says:

    MY point is that the perhaps the policy remedy to some of the negative effects of such stereotyping is to de-emphasize testing.

    Wouldn’t a better remedy be to de-emphasize stereotyping?


  9. spyder says:

    I would like to see the expansion of some of these studies to include the use of stereotyping from other cultures; portrayed to evoke either sympathy/empathy for the victims or complicity in the stereotyping. The world is certainly cosmopolitan, but i am sure that the nuances (say from Clifford’s England) between groups stereotyping behaviors has a broad range of diversity. Thus would showing the group used in this study, females from Liverpool dysphemically satirizing those from the northeast, lead North American females to score poorly or better, and so forth? Do we recognize the process of stereotyping for what it is, or do we miss seeing it when it is from other than familiar cultures???

  10. Rob Knop says:

    MY point is that the perhaps the policy remedy to some of the negative effects of such stereotyping is to de-emphasize testing.

    Wouldn’t a better remedy be to de-emphasize stereotyping?

    I’d argue for both…. There’s some chance that de-emphasizing something so sensitive to “irrelevant” factors (i.e. testing) might help the broader problem. But testing is a problem all unto itself, and stereotyping is a problem all unto itself.

    The problem with de-emphasizing testing is… what do you replace it with? When you need some sort of “objective” measure that allows you to compare large numbers of people in a relatively short amount of time, what do you do? In my ideal world, I’d be able to give every student in every class an oral final, so that I can adapt for a given student’s style of working, and so that I can tease out what the student really understands without having to worry that some sort of text anxiety got in the way. (Which of course is a bit stupid, because I remember the oral final I took for a class in college– I was *way* more nervous than before any written class.)

    However… I don’t have that kind of time.

    When it comes to admitting grad students, I’m always dubious about the general GRE and the Physics GRE. I wouldn’t be surprised if those scores *correlated* with how well students tended to do in grad school, but that’s unfair on an individual basis. One given individual might do very well despite weak GRE scores. And, indeed, when I’m reading grad applications, I place the most weight on the letters of recommendation. However, when you have to compare people from very different palaces… what do you do? What sort of general, universal criterion can we have? Alas, testing, whether or not it’s really effective, is very easy to implement, and so we lean on it an awful lot.


  11. Jude says:

    I’ve long been interested in prejudice, partly because my father was virulently prejudiced *against* prejudice (in the 1930s, one of his teachers was a Nazi sympathizer who complained about the impure nature of citizens of the US, in particular those of “mixed breeds” like my dad). I taught a prejudice unit to my son when I homeschooled him briefly last year. His greatest and most annoying prejudice was against flute players (he plays trumpet). I had him take some of the Implicit Association Tests from the Harvard University site (, as well as the prejudice tests at Interestingly enough, he turned out to have a *positive* prejudice towards blacks–interesting because there aren’t a lot of blacks in our community.

    I think that if you’re really smart, prejudice doesn’t affect your performance. Years ago, I was the only pregnant woman in a room filled with men taking their GREs. I felt extremely uncomfortable, but I still aced the tests.

  12. Jude,

    “I think that if you’re really smart, prejudice doesn’t affect your performance. Years ago, I was the only pregnant woman in a room filled with men taking their GREs. I felt extremely uncomfortable, but I still aced the tests.”

    Smart and confident and comfortable aren’t the same things. Most people do internalise prejudices to some degree, but that degree may or may not affect your confidence in something. And these are all statistical concepts — the fact that stereotypes *tend* to affect performance doesn’t mean that everyone’s performance will be affected by stereotyping. I’m not convinced that that intelligence necessarily overrides a tendency to have your confidence affected by stereotypes.


  13. Pingback: Show and Tell - Asymptotia