Lee Jussim, Social Perception and Social Reality: Why Accuracy Dominates Bias and Self-Fulfilling Prophecy, Oxford University Press, 2014, 404 pp., $84.00.
This is a very important book. It is an extensive and painstaking refutation of a set of mistaken assumptions that have dominated social science research since the 1950s, and continue to bias the thinking of most professionals in the field. It is a book for specialists–exhaustive and meticulously documented–but it systematically dismantles the illusions that helped give rise to today’s reflexive suspicion of whites. The author, Lee Jussim, is chairman of the psychology department at Rutgers University, and has spent his entire professional life as a social psychologist.
The mistaken assumptions this book demolishes are that biases, stereotypes, and self-fulfilling prophecies are so powerful that social perception plays a central role in creating social reality. The theory is that ordinary people have entrenched biases and negative stereotypes they refuse to give up. They inflict these negative stereotypes on various out-groups, who then conform to those stereotypes through a process of repeated self-fulfilling prophecies that create caste-like distinctions between groups. It is mistaken perception that creates individual and group differences.
No social scientist may ever have put the case quite that baldly, but Prof. Jussim quotes academic authorities who come pretty close: “Social perception is a process dominated far more by what the judge [observer] brings to it than by what he takes in during it.” (1955) “Our beliefs and expectations have a powerful effect upon how we notice and interpret events.” (1987) We should be aware of “the impressive extent to which people see what they want to see and act as others want them to act.” (2002)
Prof. Jussim writes that “social psychologists saw self-fulfilling prophecies and expectancy-confirming biases everywhere” and that “the extraordinary emphasis on the power of expectations to create social reality . . . has become part of the distilled wisdom of social psychology.” “[T]o this day,” he adds, “one of the shortest routes to success in social and cognitive psychology is to be the discoverer of a new bias.”
Bias and stereotype
The construction of group differences begins with false, negative views people are supposed to have about certain groups. In every instance, the unspoken assumption is that these deluded “people” are white. They are the dominant majority, and use stereotypes to hold everyone else down.
The word “stereotype” was invented in 1922 by the columnist Walter Lippmann. He wrote that people are unable to understand the world in its full complexity, so they simplify in self-serving ways. It was Gordon Allport, however, who brought the concept to full flower in his 1954 book, The Nature of Prejudice. Prof. Jussim writes that Allport “characterized stereotypes as unjustifiably resistant to change and steeped in prejudice, and concluded they were a major contributor to social injustice.” Prof. Jussim adds that the book “set much of the research agenda on stereotypes and prejudice for the next 50 years and remains widely cited today.”
The data appeared to be obliging. In a 1968 study, two different groups were told to read the same essay. One group thought the author was a man; the other thought it was by a woman. The group that thought it was written by a man found it more convincing. In a 1976 study, when students were shown a movie clip of a black and a white bumping into someone in exactly the same way, 75 percent of the students thought the black was acting aggressively but only 17 percent though the white was being aggressive. In 1983, two groups of Princeton students watched identical clips of a child answering questions on a test. The ones who were told the child was from the inner city were convinced she got more wrong answers than the ones who were told she was from a middle-class background. Prof. Jussim says these studies are still widely cited.
But were the data obliging or was it the researchers? The sex bias study could never be replicated. Nor could the bumping study. Prof. Jussim notes that the bumping paper did not contain the usual language about carefully teaching the black and white bumpers to behave in exactly the same way, and that it was the only paper that author ever published. He calls for a “moratorium on citing it.” Prof. Jusssim notes that the Princeton research could not be replicated either. “Psychology is filled with examples of individual researchers having a knack for demonstrating some phenomenon that proves difficult for other researchers to replicate,” he notes delicately.
What these studies–phony or not–were supposed to prove was that stereotypes are so powerful that people cling to them no matter what. In fact, as Prof. Jussim notes, research findings are clear: people use stereotype when they have no other information to go on, but as they get more “individuating” information they treat people as individuals. Prof. Jussim notes that this is still a controversial conclusion among social scientists, who are convinced that most people are blinded by stereotypes and ignore information about individuals.
At an even more basic level, social scientists widely believe that stereotypes are, virtually by definition, wrong. They believe stereotypes rationalize discrimination, so, as Prof. Jussim notes, “crediting any accuracy to stereotypes is tantamount to endorsing bigotry.” He points out that there used to be plenty of research into the accuracy not only of stereotypes but of what people thought about themselves and other people. This research stopped dead from about 1955 to 1985. The assumption that all stereotypes are wrong was so widespread that it was considered futile–even immoral–to study their accuracy.
Many people still bridle at the idea that stereotypes could be accurate. Prof Jussim writes that when he mentioned to his Jewish in-laws that “Jews really are, on average, richer than other people,” they reacted as if he had said, “Jews are all a bunch of cheap, corrosive, money-grubbing vermin who should be exterminated.” Prof. Jussim says it is common to assume that “if stereotypes are associated with social wrongs, they must be factually wrong.”
But people don’t just cook up nasty, baseless stereotypes. If I wanted to spread the word that Japanese-Americans are shiftless, crime-prone layabouts who live in crack houses and rob people, it wouldn’t get very far. Stereotypes are part of mankind’s ability to generalize. Logicians call it inductive reasoning. As Prof. Jussin notes, “scientific research evidence pervasively demonstrates extraordinary levels of accuracy in social stereotypes.”
Of course, they are not 100 percent accurate. Prof. Jussim reports that people are good at putting groups in rank order (who is more or less likely to be criminals, have illegitimate children, graduate from high school, etc.) but they are not nearly as good at quantifying the differences. This is to be expected. Most people don’t pore over census data. But they have a pretty good idea of the differences between races, sexes, nationalities, people of various professions, etc.
Particularly interesting are the ways in which people are inaccurate. Whites and blacks are both about as likely to know the rank order in which the races differ in outcomes, but whites tend to underestimate the extent of the differences while black overestimate them. Prof. Jussim notes that this is not consistent with the view that the majority uses harshly inaccurate stereotypes as a tool of oppression. Blacks, on the other hand, have two reasons to exaggerate differences. First, they are told repeatedly that white society oppresses them and they believe it. Second, since any alleged evidence of oppression is leverage for special treatment, it is in their interests to exaggerate differences.
Prof. Jussim finds that liberals are especially likely to disbelieve in group differences:
[T]hose most likely to inaccurately underestimate real differences were liberals in denial about group differences. . . [I]ntelligence did not matter for this group. Brainy liberals were just as likely as dumb liberals to inaccurately minimize real differences.
It’s the opposite for non-liberals. The smarter they are, the more likely they are to have an accurate understanding of group differences. This makes sense, since smart people are usually more knowledgeable about the world. Prof. Jussim wonders whether smart liberals have actually managed to block reality or whether they know the truth but refuse to admit it.
Prof. Jussim cites one study in which subjects were given a test to see where they ranked on a scale of Right-Wing Authoritarianism, which is supposed to indicate susceptibility to fascism. If people who scored high on that scale turned out to have exaggerated stereotypes the study would no doubt have trumpeted that fact. They didn’t, so the results ended up in a footnote.
Interestingly, Prof. Jussim found that political stereotypes are among the least accurate. Republicans and Democrats assume crazy things about each other, with Democrats slightly more crazy than Republicans.
Given that stereotypes are generally accurate, should we act on them? In the absence of other information, of course we should. Prof. Jussim concocts a theoretical scenario in which we walk onto a train platform where we find ballerinas at one end and a biker gang at the other. If we happen to know that bikers are, say, ten times more likely than ballerinas to be convicted of crimes, we are justified in edging towards the ballerinas. Of course, Prof. Jussim could have given us a much more realistic example, with crime differentials straight from the Justice Department, but let us give him credit for at least hinting at a rational justification for racial profiling.
Prof. Jussim points out the inherent absurdity of the traditional view that stereotypes are all inaccurate and that it is futile to assess their accuracy. If there are any genuine group differences at all, some beliefs about them will be accurate and some not. If inaccurate beliefs really are a social problem why don’t we find out which ones are inaccurate and correct them? That would be intolerable because it would imply that some beliefs are accurate. Of course, as Prof. Jussim points out the very people who claim to hate stereotypes actually love them–in the right context. He notes:
As long as we are demonstrating how open-minded, tolerant, sensitive, and caring we are, it is permissible, even good for us to ‘understand group differences.’ So, in contrast to a social-problems context, where believing in group differences constitutes lowdown dirty stereotyping, in a (multi-)cultural context, recognizing and being ‘sensitive’ to group differences shows how benevolent and egalitarian we are.
To promote “diversity” is to acknowledge group differences; otherwise, what is the advantage of having Mexicans on campus? “Cultural awareness training” is a celebration of stereotypes. Prof. Jussim is no doubt right to conclude that “the current common belief in stereotype inaccuracy appears closer to religion than to science.” Although he does not say this, opposition to the idea of stereotypes is rooted in the refusal to admit even the possibility of any group difference except for The One and Only True Group Difference, namely, that white heterosexual men are evil. Blacks are really as good at math as Asians, and women would make cracker-jack Green Berets, but white men keep them all cruelly oppressed.
This is the final, key step to maintaining white male hegemony. Once white men have convinced themselves that blacks are no good at math and that women are unfit for combat, they beam these malicious stereotypes at blacks and women, which forces them to conform to those stereotypes. As Prof. Jussim points out, the very idea of self-fulfilling prophecies (SFP) is remarkable–people (white men, anyway) can create reality merely by believing something. Prof. Jussim notes that this idea has extended into all sorts of improbable areas: If enough people are told that introverted Sally is actually extroverted, they won’t change their beliefs when they meet her; instead, the power of SFP will turn Sally into an extrovert!
SFP is supposed to be at its most powerful in schools. Prof. Jussim notes that it is common for liberals to believe that “[social]-class-based teacher expectations help create a ‘cast-like’ system that benefits middle-class children and undermines children from lower social class backgrounds.”
The famous 1968 study called Pygmalion in the Classroom is still credited in many circles with having established once and for all the power of SFP. Experimenters told real teachers at a real school that a certain number of the children entering grades one through six had been given a special test that showed they were “late bloomers” and could be expected to make dramatic gains. In fact, there had been no testing; a few children were randomly labeled “late bloomers.”
Some “late bloomers” made astonishing gains. In the first year of the two-year study, the first-grade late bloomers improved an average of 15 points on a standard IQ test, and the second-grade bloomers improved by 10 points. These are the results that have filtered down to posterity and are quoted even to this day. But there is a lot more to the study. The non-bloomers in those grades improved by about 10 and 6 points respectively, too. Furthermore, the gains were found only in the first and second grades, with nothing statistically significant in the other four grades. During the second year, teachers were again fed imaginary test results and there was a bloomer gain in only one grade of the five studied. Thus, out of eleven school years, only three showed any Pygmalion effect, and that effect was weaker in the second year.
Prof. Jussim is rightly suspicious of average IQ gains of 10 or 15 points for any group. He notes that teachers have tried nearly everything to increase IQ scores–in vain. Can a simple false expectation possibly do what heroic enrichment can’t? But there’s worse. Once critics pried into the actual Pygmalion data, they found that the bloomers who pushed up the first- and second-grade averages in the first year had colossal IQ gains as follows: 17 to 110, 18 to 120, 133 to 202, 111 to 208, 113 to 211. “Does anyone really think that the first two kids went from vegetables to reasonably smart, or the last three went from reasonably smart to extraordinary genius, as a result of teacher expectations?” asks an appropriately skeptical Prof. Jussim. “[T]he entire self-fulfilling prophecy effect hinged on the occurrence of bizarre outliers,” he notes, which is a polite way of saying the whole thing was probably a fraud.
This study gave rise to a great deal of controversy, and a number of attempts to replicate it–which failed. The idea of SFP is not, however, complete bunkum. There may be a bit of it, here and there, especially in younger children. Studies have found that there can be SFP in athletic performance; perhaps some young athletes work harder if coaches think they are stars. But for schools, Prof. Jussim concludes that: “Although the scientific evidence may be equivocal regarding whether teacher expectation effects on IQ are nonexistent or reliably very small, it is completely unequivocal that such effects, if they occur at all, are not very large by any standard.”
Even when the libs concede that effects of SFP may not be very strong, Prof. Jussim notes that “the belief that small expectancy effects accumulate over time is very widespread, at least within social psychology.” The idea is that whites and blacks enter the world with equal potential, but positive SFPs raise up whites while negative SFPs bring down blacks. Prof. Jussim says there is no scientific evidence for this. People’s abilities are usually well established, and people around them find that out:
[E]ven among targets from stereotyped groups, disconfirming behavior is far more likely to be noticed and to influence perceptions and judgments than it is to be ignored and dismissed. Such a process, too, will typically increase the accuracy of expectations for individuals. If accuracy increases over time, it will limit and reduce the potential for self-fulfilling prophecy.
If a teacher or anyone else has an accurate assessment of someone, there can be no SFP because SFP is, by definition, based on an inaccurate assessment. Furthermore, in the real world, even if teachers have the wrong idea about someone, it will almost never be as wildly wrong as the “late bloomer” baloney teachers were fed in the Pygmalion study.
Prof. Jussim notes that teachers often do expect more from middle-class students than from slum-dwellers because it is accurate to do so, but they quickly figure it out if they have a dim middle-class kid or a genius from the slums. Even children usually have a pretty good sense of what they can and can’t do, and don’t mold themselves to some mistaken view. At the same time, there is a great deal going on in students’ lives that teachers can’t control. “When all the factors operating against expectancy effects in the classroom are thoughtfully considered,” Prof. Jussim explained, “it is wonder that they occur at all, not that they are typically small.”
All told, unlike the traditional, grim view of (white) people forcing out-groups into oppressive little boxes through the power of stereotype and SFP, people make reasonable generalizations from experience but set them aside when they meet an exception. As Prof. Jussim puts it, when it comes to social perception, “The glass is 90% full. People are not perfect, but they are pretty damn good.” He even adds: “It behooves us [scientists] to undo the erroneously dark image of human social thought that we have perpetrated all these decades and replace it with one that is more appropriate to the evidence.”
So where did that “erroneously dark image” come from? Prof. Jussim’s explanation is pretty damn good, and is worth quoting at length. Here, he uses the word “target” to mean someone being evaluated and “perceiver” to mean someone who is observing and evaluating:
To characterize a belief that some kid is not too bright, is a klutz on the basketball court, or is socially inept as ‘accurate’ has a feel of ‘blaming the victim.’ Blaming the victim is a bad thing to do–it means we have callously joined the oppressors and perpetrators of injustice.
If the belief is ‘accurate,’ then we cannot point to perceivers’ errors, biases, misconceptions, egocentrism, or ethnocentrism as explanations for target difficulties. The unintelligent, unattractive or socially awkward target, in these cases, really is flawed in some way. This is especially true if the negative belief is applied to large demographic groups (i.e stereotypes). Acknowledging this is difficult and distasteful. People who publicly declare that two groups differ in some societally valued attribute (intelligence, motivation, propensity for alcoholism or crime, morality, etc.) run the risk of being accused of being an ‘ist’ (racist, sexist, classist, etc.) or, at a minimum, of holding beliefs that do little more than justify existing status and hierarchy arrangements.
In contrast, an emphasis on expectancy effects or other errors and biases (including but not restricted to prejudice) implies a benevolent and egalitarian concern with injustice. Such an emphasis suggests that so-called ‘real’ differences between groups do not result from any actual attributes of members of those groups (their cultures, their religions, their histories, their social conditions, their geography, their practices, their politics, their genetic predispositions)–they result solely or primarily from the oppressive effects of others’ self-fulfilling prophecies, prejudices, and expectations. Furthermore, this perspective suggests that many differences alleged to be real are not real at all–they simply reflect the ists’ own expectancy-confirming biases.
In addition, an emphasis on expectancy effects provides a clear villain–the holder of the false expectation. It also points to a relatively straightforward way to ameliorate some social inequities–change expectations, stereotypes, etc. In contrast, not only does accuracy seemingly justify inequality (‘they have lower status because they are less skilled, competent, intelligent’ and so on), but also its relevance to solving social problems is not as readily apparent.
If my belief that you are incompetent is inaccurate, all that you need to do is change my belief to ameliorate the problem. But if my belief is accurate, then changing the situation requires much more work–to make us equal, we have to upgrade your actual competence.
This is all very fine, but what Prof. Jussim does not say is that the “clear villain” who turns up in these studies is always and inevitably white people. The conceptions he is dismantling are not just wrong: They are an important part of the conviction that white men are a scourge and that the world would be much better off without us. That is what makes this book so important.
Social Perception and Social Reality reminds me of Arthur Jensen’s 1980 classic: Bias in Mental Testing, which also completely demolished fashionable illusions. Jensen showed beyond any reasonable doubt that mental tests, including IQ tests and the SAT, are not biased against blacks, and accurately assess people of all races. This book is just as exhaustive and just as persuasive.
The trouble is that hardly anyone read Bias in Mental Testing, and hardly anyone has read this book. Oxford University Press priced it at a scandalous $84.00, and although it has been out since April 2014, it has only five reviews on Amazon.com (all give it five stars). Prof. Jussim has a regular column at Psychology Today but piecemeal blogging isn’t enough to promote his arguments.
This book is a rare achievement; it is even more important than the author himself realizes.