»  National Review Online

February 1st, 2001

  Stereotypes Aren't So Bad


One of the most heartening features of the times we live in, if you are of a conservative inclination, is the trend of discoveries now being made in the human and biological sciences. Anthropology, psychology, sociology and genetics are all turning up results — good, hard, replicable scientific results — whose broad tendency is to prove that human nature is much more like what conservatives have always said it was like, than it is like what leftists have believed. Every time one of these results escapes from Academia into the awareness of the general public, it is greeted with shrieks of horror and obloquy by the leftist establishment (remember The Bell Curve?) and the researchers who uncovered it are tarred, feathered, and run out of the public square with cries of "racist!" ringing in their ears. Those results are piling up mightily behind the dam of orthodoxy, though, and the guardians of that dam are running out of fingers to stick in the cracks. Meanwhile the academics — normally a timid and retiring lot by nature — are getting bolder and bolder in defense of their hard-won truths.

All of which is by way of introducing an astonishing book I have just been reading. The book's title is Stereotype Accuracy: Toward Appreciating Group Differences. It is a collection of academic papers written by researchers in the field of Social Psychology. The editors are Yueh-Ting Lee, Lee J. Jussim and Clark R. McCauley, and the book was published by the American Psychological Association in 1995. I found it by chance when hunting around for research materials to support an article on racial profiling I was writing for the print version of this magazine. (That article will appear in next week's issure of National Review.)

Before you run out to buy the book, I had better say that this is academic stuff, written in academic jargon, with lots of graphs and tables. These are the kind of folk who write "veridicality" when they mean "truth." Stereotype Accuracy is not bedtime reading, and I personally found some parts of it hard to follow. The main thrust of the book is clear enough, though. These people, all of them full-time researchers at respectable universities, have been carrying out studies of the tendency we all have to stereotype groups of human beings, including our own group. They have been asking very interesting questions, not all of them politically correct. Why do we stereotype? Are stereotypes actually good for anything? Do they blind us to individual qualities? Are they ever true?

Stereotypes, of course, come in both positive and negative varieties, with a single group often being the subject of both simultaneously. Positive: Fat people are jolly, good-looking people are sociable, blacks are athletic, Jews are smart, the English are classy. Negative: Scandinavians are morose, blondes are dumb, blacks are lazy, Jews are pushy, the English are undersexed.

Generally speaking, stereotypes have had a very bad press. They lead, we have been told, to bias, prejudice and discrimination. The editors of Stereotype Accuracy point out that many of the researchers who contributed to the book were first attracted to the field by a desire to fill some obvious gaps. Until recently, most academic work on stereotypes took that hostile point of view, and there was a need to redress the balance. Pure logic suggests that stereotypes fall into four classes (any one of which might, of course, after careful research, turn out to be empty): positive and accurate, positive and inaccurate, negative and accurate, negative and inaccurate. Prior to the work summarized in this book, practically all research had been done on category 4, practically none on the other three. If you are a researcher looking for a topic, you naturally zero in on gaps of this kind.

Well, now the research is being done, and the results coming in are unexpected. Far from being a loathsome aberration that ought to be purged from our behavior, it turns out that stereotypes are essential life tools, are accurate much more often than not, and that we do not use them as much as, from cold practical considerations, we should. Modestly, methodically, with batteries of experimental evidence, these researchers are demolishing most of what you thought you knew about stereotypes.

 •  People ascribe a stereotype to everybody in the subject group.   "All Germans are efficient." "All English people have bad teeth." In fact, these researchers were not able to locate anybody who believes that a stereotype is true of all members of the stereotyped group. Stereotypes are probabilistic tools, and even the most dull-witted human beings seem to know this. People who believe that Mexicans are lazy or that the French don't wash, understand perfectly well that there are lots of industrious Mexicans and fragrant Frenchmen.

 •  Stereotypes exaggerate group characteristics.   No, they don't. Much more often, the opposite is true. For example, the racial stereotypes that white Americans hold of black Americans are generally accurate; and where they are inaccurate, they always under-estimate a negative characteristic. The percentage of black American families headed by a female, for example, was 21 at the time of one survey (1978): the whites whose stereotypes were being investigated offered estimates of from 8 to 12 per cent. It is not true that stereotypes generally exaggerate group differences. As in this example, they are much more likely to downplay them.

 •  Stereotypes blind us to individual characteristics.   Nope. It is not the case that when we pass from a situation where we have nothing to go on but a stereotype (cab driver being hailed by young black male) to one where a person's individuality comes into play (interviewing a black job applicant), our stereotypes blind us to "individuating traits." On the contrary, researchers have found that the individuating traits are seized on for attention, and stereotypes discarded, with rather more enthusiasm than the accuracy of stereotypes would justify. Teachers' judgments about their students, for example, rest almost entirely on student differences in performance, hardly at all on race, class or gender stereotypes. This is as one would wish, but not as one would expect if the denigrators of stereotyping were to be believed.

 •  The real function of stereotypes is to bolster our own self-esteem.   Wrong again. This is not a factor in most stereotyping. The scientific evidence is that the primary function of stereotypes is what researchers very prettily call "the reality function." That is, stereotypes are useful tools for dealing with the world. Confronted with a snake or a faun, our immediate behavior is determined by generalized beliefs — stereotypes — about snakes and fauns. Stereotypes are, in fact, merely one aspect of the mind's ability to make generalizations, without which science and mathematics, not to mention much of everyday life, would be impossible. Researcher Clark R. McCauley:

Standing next to the bus driver, we are more likely to ask about traffic patterns than about the latest foreign film. On the highway, we try to squeeze into the exit lane in front of the man driving a 10-year-old station wagon rather than trying to pull in on the man driving a new Corvette. Looking for the school janitor, we are more likely to approach a young man in overalls than a young woman in overalls. This kind of discrimination on the basis of group differences can go wrong, but most of us probably feel that we are doing ourselves and others a favor when we respond to whatever cues and regularities our social environment affords us.

Ah, the sweet cool breeze of common sense! Wafting to us from Academia, of all places!!

With the courtesy and humility that is proper to all honest scientific inquiry, the editors of Stereotype Accuracy have included a dissenting voice, Charles Stangor of the University of Maryland, who throws a wet blanket over some of the findings. ("Arguments that stereotypes are by and large accurate are premature …") Stangor takes a clear political stand: "As scientists concerned with improving the social condition …" He is the only one in the book to do so, and does not explain why "improving the social condition" is necessarily any business of scientists as scientists — however much it may concern them as citizens. Nor does he offer a definition of "improving the social condition," the precise interpretation of which is, of course, the source of all political differences. Lenin wanted to "improve the social condition" of Russia via a terroristic dictatorship that expropriated all private property; Adolf Hitler undoubtedly believed that he was "improving the social condition" of Germany by exterminating the Jews. Most of us would disagree with both of them, but neither their views nor our disagreements with them belong in the realm of scientific inquiry. It is highly characteristic of political ideologues that they believe "improving the social condition" can have only one possible meaning — theirs.

I found that the overall effect of reading this book was to make me feel more tenderly towards the human race and angrier towards the Left, which, au fond, hates humanity and seeks to wage war against human nature. (Mao Tse-tung denied flatly that any such thing as "human nature" exists.) Here we are, a rather fragile, smelly, two-legged animal with all the soft tissue on the outside, not very fast and not very strong, dropped into the world with few natural defenses and swamped with a continuous tsunami of impressions coming in to all our five senses all day long, that we somehow have to sort out into useful information. To accomplish this stupendous task, we have developed, or been given, marvellous skills. The most marvellous of all, perhaps, is our skill at generalizing, without which, as Clark McCauley points out (see above), life would be impossible.

Yet just as marvellous, in a way, is the power that even the least intelligent of us seems to have, to drop our generalizations when more useful, more particular information, is available — to form individual judgments that violate our stereotypes. We can all do this, and we all do do it, all the time. What a piece of work is a man!



(Yeah, yeah, I know: you don't normally get footnotes in an opinion column. But reading all that academic stuff has put me in a "footnotes" frame of mind. Indulge me.)

(1) Having spoken somewhat slightingly of the style of these authors, I would like to credit one of them — presumably Yueh-Ting Lee, who is a joint author of this particular paper — for reminding me of my second-favorite Confucius quote: Jun jun chen chen fu fu zi zi (君君臣臣、父父子子). Given that jun means "prince," chen "government minister," fu "father" and zi "son," I invite the reader who knows nothing about classical Chinese grammar to try guessing the meaning of this apothegm, without looking at the translation below.**

My first favorite Confucius quote, not entirely inapt here, is: Junzi bu qi (君子不器, at [2-12] here.) — "A man is not a pot."

(2) The business of writing "veridicality" for "truth" reminds me of an anecdote about Bertrand Russell. While the great philosopher was living in America, Harvard University asked him to give an address to their philosophy department. Russell wrote up an address and sent it to them for approval. He gave it the title "Words and Things." Some days later he got a call from the Philosophy Department. "Prof. Russell, we think your address will do just fine. However, there is a problem with the title. 'Words and Things' really won't do for a lecture on academic philosophy. Do you think you could change that title? Make it a little more … professional?" Russell changed the title to something like "Linguistic Correlates of Epistemological Constructs." Harvard was happy and the lecture was a success.

**You can just barely make a direct translation into English, with the first occurrence in each pair being a noun in the vocative, the second a "verb" in the imperative: "Prince — prince! [I.e. behave as a prince should.] Minister — minister! Father — father! Son — son!" See [12-11] here.