Bryan Caplan  

Bastards and Stereotype Accuracy

Which Books Should We Re-Read?... Happy Birthday, David Carden...
I'm a firm believer in stereotype accuracy.  I just finished re-reading my favorite chapters from Lee, Jussim, and McCauley's excellent Stereotype Accuracy: Toward Appreciating Group Differences.  The book was published in 1995; Jussim's recent summary brings us up to date.  When scientists measure the correspondence between stereotypes and actual probabilities, they've found that...
...stereotype accuracy -- the correspondence of stereotype beliefs with criteria -- is one of the largest relationships in all of social psychology.  The correlations of stereotypes with criteria range from .4 to over .9, and average almost .8 for cultural stereotypes (the correlation of beliefs that are widely shared with criteria) and .5 for personal stereotypes (the correlation of one individual's stereotypes with criteria, averaged over lots of individuals).  The average effect in social psychology is about .20.  Stereotypes are more valid than most social psychological hypotheses.
Given my knowledge of this research, I felt a slight pang of guilt when I dismissed negative Westerosi stereotypes about bastards as thinly-veiled misanthropy.  If the people of Westeros consider bastards "wanton" and "treacherous," doesn't the best social science tell us to presume they're right?

Not really.  Here's a key caveat from Clark McCauley's chapter "Are Stereotypes Exaggerated?" in Stereotype Accuracy:
[T]he strongest conclusion from the present review is that purely cognitive mechanisms of stereotype exaggeration are inadequate to explain the pattern of results...

It is still too soon to throw out the exaggeration hypothesis, however.  Exaggeration of stereotype characteristics may be a powerful tendency in perceptions of a group seen as an enemy.  U.S. perceptions of the Japanese in World War II, for example, certainly suggest the possibility of caricature and stereotypic exaggeration in perceptions of an enemy group.  Thus, motivational mechanisms for stereotype exaggeration may be more powerful than the purely cognitive mechanisms featured in current literature.  This possibility is not easily evaluated on the basis of available research, none of which contain any indication that the stereotyping group felt real hostility toward the stereotyped group.
Wise words.  My only quarrel is that lots of "available research" does indeed confirm that popular stereotypes rooted in hostility toward markets and foreigners are not just exaggerated, but false.

COMMENTS (10 to date)

Dr. Caplan,

While you are on the subject, I would love to hear your thoughts on stereotype threat. You are always so eloquent and tactful in your presentations of information.

I heard many strong arguments in favor of the existence of stereotype threat. Posing counter-arguments can be difficult because, in my experience, even the most tactful references to counter-evidence to stereotype threat are met with very charged emotions: For considering the counter, I am perceived as being cruel, insensitive, and making excuses to perpetuate an oppressive "system." Despite such research as...

Anyway, I greatly respect your perspective as well as how you communicate. I would love to hear what you have to say about the matter. Thank you!


Mark M. writes:

What about the stereotype that poison is the preferred weapon of women, cravens and eunuchs?

PrometheeFeu writes:

Wouldn't we expect the caveat to more generally apply to cases where being wrong is cheap? For the average American during WWII, being wrong about the Japanese would be costless. Similarly, assuming that bastards are rare in Westrose, being wrong about bastards would be very cheap for the average resident of Westrose.

On the other hand, being wrong about a very common group will be very costly as you may be cheated by unrecognized cheats, waste wages on lazy laborer, or pass up an opportunity to deal with an honest merchant.

(Also, do not forget to the possibility of a multiple equilibria world where one may be treacherous as a response to being frequently accused of treachery.)

NZ writes:

Don't wait for others--you'll be waiting a long time. The best way to fight stereotypes, if that's something that concerns you, is to break them yourself, making sure lots of other people see you do it. Of course if you're unable (or even better, unwilling) to break a stereotype, then the stereotype may not only be accurate but also have a good reason for existing.

gwern writes:

PrometheeFeu: in your examples, people would also have a lot less information about the group. If you don't live in Japan, pre or during WWII, you're going to lack a lot of information about Japan (even ignoring all the American propaganda and other side-effects of a war).

How do you distinguish between people having incentives to not indulge excessive stereotypes, and people having more accurate information?

Ted Levy writes:

Well the book Stereotype Accuracy: Toward Appreciating Group Differences must be of high scholastic quality. The first editor is named Yueh-Ting Lee...

Clay writes:

I thought the entire Caplan argument for complete open borders immigration was agreeing that many stereotypes are correct, but that they aren't valid reasons to justify physical exclusion. In this post, he suggests that foreigner stereotypes are false, which is completely different.

KnowPD writes:

I think the key difference between our world (as it is or what we would like to think it is) and Westeros is the expectation of positive sum through cooperation. Bastards represent uncertainty - they are unpredictable from an alliance perspective so they go to the Wall to work for the good of all kingdoms. If we assume that unpredictability is worse than predictably evil, it becomes more difficult to determine strategy in a zero sum or negative sum Machiavellian world of Westeros. In our positive sum world, discriminating against bastards is both irrational and immoral. In the Machiavellian world of Westeros, it is immoral but I question whether it is irrational. Why would it be irrational to denigrate the wildcard / unknown? You reduce the uncertainty by reducing their status through denigration. Rational? Yes. Moral? Probably not.

Ă…sehelene writes:

I highly recommend Jussim's new Book "social perception and social reality". (I blogged about it here, I figure an economist wouldn't mind if a psychologist is a bit self-promotional).

Of course stereotypes can be wildly, exaggeratingly wrong - especially if one is motivated for some reason to derogate a group - and it can be very harmful. But, we need to have some everyday guide to other people, which is where the stereotypes we have tend to be reasonable (although not always accurate) bets on how to interact with people.

As I say, I'm happy if you stereotype me as a typical university lecturer, because I really don't want people to always know my inner unique pretty snow-flake.

But, really, read the book. It is expensive (Jussim acknowledges it too). Which is a shame for such a highly readable book, despite the fact that it also throws numbers at you. (I know, would not deter an economist, but I deem it readable also for those who think numbers are like vampires and must be avoided).

isomorphismes writes:

FWIW, I would argue that the better mathematical way to explain why stereotypes are dangerous is not the location of the centre but the lack of variance.

In your linked example, that would be confusing "the median woman is more F and less T than the median man" with one of many common and incorrect misinterpretations like "Each women is emotional" (meaning high F, low T) or "each woman is more emotional than each man" or "Women as a class are characterised only by their low T, high F--they have no other characteristics". "Women are more emotional than men" can be rendered into maths in different ways when we're talking about a distribution.

PS: The Wikipedia article on stereotypes and cliches in TYPE (which is whence the metaphor originates) was enlightening to me and bears on what I just said. It's not that the type-block bears the wrong words; it's that the same type-block is used over and over and over, never changing.

Comments for this entry have been closed
Return to top