Microaggression and Macrononsense

Sophisticated, affluent people in the United States (SAPs) have been trained through years of education to respect whatever is presented to them as “science,” even if it’s not very good science, even if it’s not science at all. Their years of education have not trained them how to tell the difference. Sophisticated and affluent Americans, as a group, are pretty gullible.

So when their leaders in journalism, academia, and business announce a new truth of human nature, SAPs around the country are likely to embrace it. The idea of microaggressions is one of these. It was first popularized a decade ago, and now the pervasiveness of microaggressions in American life is taken as settled fact.

We could have seen it coming. Already, by the time microaggressions became widely known, social scientists had invented the Implicit Association Test (IAT). The test, administered online and to college students throughout the country, pretended to establish that anti-black and anti-Latino prejudice among white Americans was ever-present yet, paradoxically, nearly invisible, often unrecognized by perpetrator and victim alike. Even people who had never uttered a disparaging remark about someone of another color were shown by the IAT to be roiling cauldrons of racial animus. You know who you are.

The IAT thus laid the predicate for microaggressions. They were the outward, unwitting expressions of implicit racism; not only were they evidence of it, they were offered as proof of it. (Circularity is a common tool in cutting-edge social science.) Microaggressions are usually verbal but they don’t have to be. In their pathbreaking paper “Racial Microaggressions in Everyday Life” (2007), the psychologist Derald Wing Sue and his team of researchers from Columbia University helpfully listed many common microaggressions. Saying “America is a melting pot” is really a demand that someone “assimilate to the dominant culture.” Having an office that “has pictures of American presidents” on the wall announces that “only white people can succeed.” Also, an “overabundance of liquor stores in communities of color” carries the microaggressive message that “people of color are deviant.”

Sue’s paper came at a crucial time, just as it was becoming obvious that public, systematic, and institutionally enforced racism was declining toward zero. The paper perfectly matched the prejudices, if you’ll forgive the expression, shared by nearly every credentialed social psychologist and cognitive researcher. “Racial Microaggressions in Everyday Life” has been cited in more than 1,700 published papers over the last 10 years, better than three times a week on average. It also inspired more sequels than X-Men. You can fill in the blank: “Racial Microaggressions and .  .  .” “the Power to Define Reality” .  .  . “Daily Well Being Among Asian Americans” .  .  . “Difficult Dialogues on Race” .  .  . “Psychological Functioning” and so on. The pursuit and explication of microaggressions is a cottage industry in the larger industry of social-science research, and guaranteed grant-bait.

The SAPs of higher education tumbled first to the idea. As Lilienfeld points out, microaggressions now include using the phrase “politically correct” at the University of Wisconsin and describing America as “the land of opportunity” anywhere in the University of California system. From higher ed the idea quickly migrated to the world of business, where more SAPs were waiting. ­Lawsuit-shy executives and easily excited personnel directors instituted elaborate and expensive programs to convince employees, first, that their workplace was lousy with microaggressions, and second, that they were about to be deloused, ready or not.

As always in corporate America, political correctness (I’m not in Wisconsin) comes packaged as good business sense. One commonly cited figure—that microaggressions cost U.S. businesses $450 billion to $550 billion in lost productivity—is predictably dubious. It came from a Gallup finding that “actively disengaged employees” diminish U.S. productivity by that amount. A former Harvard dean, citing “Gallop,” then told readers of the left-wing website Huffington Post that microaggressions contributed to the $450 to $550 billion loss in productivity. And now, from a legal website: “The ‘soft bigotry’ of microaggressions has real costs—$450 to $550 billion in U.S. workforce productivity according to Gallup.” At least they’ve fixed the spelling.

And recently, in a solemn act of validation, lexicographers included the word “microaggressions” in the Merriam-Webster Dictionary’s website.

You’d think that an idea like this, passing from social science into the larger culture with astonishing speed, would have received a lot of scientific scrutiny from social scientists: how microaggressions work, socially and psychologically; how frequently and under what conditions they occur; how they can be avoided or ameliorated .  .  .

But you would be wrong. This is where Lilienfield and his paper come in. What he calls the “microaggression research program” has been a daisy-chain of confirmation bias rather than a rigorous pursuit of scientific truth. Indeed, he says, the very concept of microaggressions is too hazy and ill-defined to be studied systematically, even by social scientists.

His “Microaggressions: Strong Claims, Inadequate Evidence,” published in the journal Perspectives on Psychological Science, is a tour de force. Lilienfeld undertakes a critical review of the literature that social scientists have produced on microaggressions. He begins, as he should, by conceding the fact of ongoing racism in the United States. He knows too that words can be unintentionally wounding. He points to an incident in which an engineering professor expressed surprise before his classroom when an African-American student got a perfect score on a test, apparently on the assumption that no black student could be expected to do such a thing. All of us have been witness to similar incidents of callousness and gaucherie. I once heard of a well-loved public servant who referred to a rival politician as “the first mainstream African American who is articulate and bright and clean.” Racial prejudice can indeed be insidious.

But is it ubiquitous? Does it soil even the most innocent encounter between persons of different races—or rather, between a white person and a person of another race? (In the scheme of microaggressions, a white person is ipso facto the perp.) The answer is, and Lilienfeld’s point is, we don’t know. Anecdotally it’s clear that words taken as a microaggression by one Latino or African American may go unnoticed by another. Despite such common experiences, which would seem to undermine many definitions of the word, no scientist has “challenged the central assumptions that microaggressions, as currently conceptualized, .  .  . comprise a psychologically meaningful construct.” The most important and worrisome claim of micro­aggression researchers is that microaggressions cause consistent and measurable harm to the mental health of their recipients—and yet no one has done a systematic review of the empirical evidence that supposedly establishes such a causal link.

Lilienfeld says that the incoherence of the concept begins with the word coined to express it. “It is doubtful whether an action that lies largely or exclusively in the eye of a beholder can legitimately be deemed ‘aggressive,’ ” he writes. To be aggressive, an act must be intentional. Yet microaggression scholars insist that many micro­aggressions are involuntary, an unconscious ripple from the depthless sea of racism in which we have been soaking all our lives. And there’s the oxymoron problem. A truly aggressive act is unlikely to be best described as “micro.” Aggressions tend to be big.

And so do the holes in microaggression research. It passes over the most basic questions of how the concept should be applied. Researchers have yet to use Sue’s list of microaggressions to test precisely what proportion of minority students are offended by them and whether they see in the microaggressions the same offensive message that Sue’s researchers did. “The association between microaggressions and specific implicit messages remains conjectural,” Lilienfeld writes.

He finds in the literature all the methodological flaws that we have come to expect from politically motivated social science: small sample sizes, self-selected nonrandom samples, self-reporting of results, the embedded bias of researchers, the lack of an agreed-upon terminology and system of measurement, and an inadequate use of control groups. Most obviously, the all-important matter of personality traits has not been factored into the research. If a microaggression is “in the eye of the beholder,” then we should know something about the beholder. What traits are shared by people who are most likely to sense and be offended by microaggressions?

The research, says Lilienfeld, “has all but ignored the potentially crucial role of negative emotionality [NE] in shaping perceivers’ judgments of microaggressions.” NE is one of those complicated social science-y terms that describe something utterly commonplace. It refers to the gloomy disposition of someone who is given to worry, irritability, and feelings of victimization. Often researchers investigating psychological reactions to others will try to control for NE. This way they can be sure they are measuring an independent reaction and not just the person’s normal disposition. The single microaggression researcher who did consider NE found that it could not entirely account for the mental harm attributed to microaggressions. Otherwise, however, “in light of the virtually wholesale neglect of NE in [microaggression research] it seems especially premature to advance strong causal assertions regarding the ties between microaggressions and mental health outcomes.”

If NE and not microaggressions account for some of the mental harm experienced by people who claim to perceive them, should we be surprised? None of what Lilienfeld suggests is novel, after all; none of it would have come as a surprise to anyone lucky enough to have lived long before any social scientist felt the need to invent the term “negative emotionality.” Just about everyone understands that some people are ornery, and orneriness, by my own unscientific calculation, seems evenly distributed among human beings whoever they are. After many thousands of years of human behavior, men and women in authority should know that priming such people when they are young and impressionable to detect insults and slights even where they don’t exist will result in a great deal of general unhappiness.

Given our estrangement from such home truths, and our hypnotic devotion to whatever we are told is “science,” Lilienfeld’s approach may be the only one that will return the conversation to the realm of life where it is actually lived. He emphatically wants research into microaggressions to continue, assuming the term can be meaningfully defined. At the same time, he’s using modest and plausible social science to undo the damage that bogus and extravagant social science has done to life on campus and elsewhere. He is translating basic human understanding—common sense—into the only language many of our SAPs understand and accept.

And he is not completely alone among social scientists. Just within the field of race relations, a handful of researchers are casting doubt on once-unquestioned findings. New research, for example, threatens the conceptual underpinnings of the IAT, one of the most widely used tools in social science (see sidebar below). “Stereotype threat,” a hazy phenomenon uncritically accepted by students of racial disparity, has failed multiple attempts at rigorous replication (see sidebar on page 27). Even diversity training programs, which now include initiation into the ideology of microaggression, are increasingly understood to be worthless. “The positive effects of diversity training rarely last beyond a day or two,” wrote two researchers in the Harvard Business Review last summer, “and a number of studies suggest it can activate bias or spark a backlash.” (“Nonetheless,” they add puckishly, “nearly half of midsize companies use it, as do nearly all the Fortune 500.”)

Cutting-edge thinkers who boast of their devotion to science may soon have to concede that many fashionable assumptions about how human beings behave are as unscientific as racism itself; so is the vast belief system based on them. And if this process of self-correction continues, with brave researchers willing to question the assumptions at the foundation of their discipline, we who are skeptical of the whole enterprise may have to concede that some social sciences do, after all, qualify as real science.

I hope nobody takes that the wrong way.

Andrew Ferguson is a senior editor at The Weekly Standard.

IAT or BS?

Few inventions of the social sciences have been as influential and unavoidable as the Implicit Association Test. Introduced at a press conference by a team of social scientists in 1998, it has since has become an essential tool for marketing researchers, corporate trainers, and anyone else who hopes to discern, and hence manipulate or improve, the views and behavior of other people. The powers of the IAT are said to be large and various, but it is most famous for divining hidden racial attitudes, or what the trade calls “implicit bias.” The bias is implicit because the poor subject who takes the test isn’t aware of his own racism. He may even pride himself on his egalitarian racial views—indeed, such people are often the worst offenders, implicitly.

You can take the test online, and millions of people have, at a site called “Project Implicit” run by Harvard University. The subject presses one key or another as faces and words flash on his computer screen. Sometimes bad words appear with pictures of black people, sometimes with pictures of white people. Measured to the millisecond, the test-taker’s reaction times—how quickly he associates bad words with black faces and good words with white faces—are deemed to convey his true racial attitudes. The mechanical proc­ess by which implicit racism adjusts its travel time from the brain to the fingertips has never been traced, or at least publicly revealed. Still, the Harvard researchers insist that implicit racism is found in nearly everyone and, unchecked, will often lead to brutish behavior. The implicit racism uncovered by the IAT is routinely brought out to explain any disparity between white people and black people, from per capita income to criminality. When a white police officer kills a black suspect, journalists and lawyers will lunge for the IAT. Rare is the cop who has escaped schooling in the science of his own implicit racism.

But for several years the science behind the test has been under pitiless assault. In 2015 two social scientists, Gregory Mitchell of the University of Virginia and Philip Tetlock of the University of Pennsylvania, undertook a clear-eyed review of the vast, mostly credulous literature on the IAT. I might as well quote their comprehensive conclusion directly rather than paraphrase. “On issue after issue, there is little evidence of positive impacts from IAT research: theories and understandings of prejudice have not converged as a result of the IAT research; bold claims about the superior predictive validity of the IAT over explicit measures have been falsified; IAT scores have been found to add practically no explanatory power in studies of discriminatory behavior; and IAT research has not led to new practical solutions to discrimination.”

The best science reporters are on the case. In an exhaustive and heroic account of the IAT controversy, Jesse Singal of New York magazine concluded that the critics had the better evidence and the better argument. He quoted one of the inventors of the test, now a much-decorated professor at Harvard. “I don’t read commentaries from people who are non-experts,” she wrote in an email. Then she suggested her critics might be suffering from racism themselves, or something still darker. Their study of the race IAT, she wrote, “seems not to be about the evidence. It will need to be dealt with by them in the presence of their psychotherapists or church leaders.”

The ‘Stereotype Threat’ Reconsidered

“Stereotype threat” is another discovery of social science designed to account for a racial disparity, in this case the academic achievement gap between black students and white students. The gap shows up in high school graduation rates, proficiency levels, and especially in average scores on standardized tests.

Stereotype threat emerged in psych lab experiments under the direction of Claude Steele, a Stanford social psychologist (and twin brother of the great essayist Shelby Steele). Claude Steele theorized that African-American students, confronted with an achievement test, were held back by the internalized stereotype that whites were more intelligent than blacks.

In his basic experiment, two groups, each with a mixture of “high-achieving” white and black students, were given the same SAT-like test. The first group was told the test was a measure of intelligence, thus conjuring up a stereotype threat. The second group was told that the same test was a simple problem-solving exercise; no stereotype threat was introduced.

According to Steele and his colleagues, the test scores of the group under stereotype threat showed the typical black-white disparity in scores you’d expect to find in such tests. But in the second group, they said, the disparity effectively disappeared. Their paper, published in 1995, has been cited more than 5,000 times and spawned yet another cottage industry. Social scientists eagerly embraced the idea that intellectual performance, and other kinds of performance, could be improved or discouraged by merely tweaking a few environmental factors. Stereotype threat has come in especially handy in explaining the gap in math scores between girls and boys.

It wasn’t until 2004 that another team of researchers discovered that Steele’s findings had been misinterpreted, owing to a statistical confusion. The African-American students in the non-threat group had indeed scored better on the test than those in the threat group. But the most exciting finding, the one that had made Steele’s paper famous, was false: The gap between white and black scores was proportionally the same in both groups. The achievement gap had not been closed by removing the stereotype threat. Nevertheless, many psychology text books and other researchers continue to report the original misinterpretation.

Replications of stereotype threat experiments have rarely been done in real-world settings. Variations on the original experiment are usually conducted in university psych labs, with small, nonrepresentative samples, under the watchful eye of sympathetic researchers. In 2014, two researchers finally did the first meta-analysis of studies of stereotype threat and girls’ math scores. When they adjusted for “publication bias”—the tendency of researchers and journals to publish papers that reach a popular conclusion—they found that the effects of stereotype threat on math scores were statistically insignificant.

In social psychology generally, stereotype threat remains too good to check. One renegade psychologist, Lee Jussim of Rutgers, thinks he knows why. “Stereotype threat,” he wrote last year, “is a great rhetorical tool in the quest for egalitarianism. .  .  . It is, therefore, professionally risky to challenge ideas that serve egalitarian rhetoric.” ¨