How Microaggression Training Could Harm Minority Students

Photo: Glenn Koenig/LA Times via Getty Images

As soon as someone mentions “microaggressions,” things tend to devolve. The concept, defined by its chief promoters as “brief and commonplace daily verbal, behavioral, or environmental indignities, whether intentional or unintentional, that communicate hostile, derogatory, or negative racial slights and insults toward people of color,” has become a campus-culture-war lighting rod: Whether you think microaggressions are real and worth worrying or not is generally a signal of where you stand on a bunch of broader, related questions about race and activism on campus.

But microaggressions aren’t just a culture-war cudgel. As Scott Lilienfeld, a social clinical psychologist at Emory University, writes in an important new article on Perspectives on Psychological Science, they have started showing up in applied settings on campus, both in diversity trainings and in other sorts of educational programming. And when they do show up, they are often accompanied by claims designed to heighten, in students’ minds, just how serious microaggressions are: particularly that minority students who experience more microaggressions are more likely to experience a bevy of negative outcomes, up to and including committing suicide.

Lilienfeld’s long, in-depth paper skirts the culture-war component of microaggressions and homes in on the scientific questions: After all, given that those participating in the microaggression research program, or MRP, have made a number of psychological claims about this phenomenon, surely it’s time to evaluate the quality of those claims and the evidence buttressing them. As Lilienfeld argues, there are, at the moment, extremely serious problems with the MRP, including with basic questions about how microaggressions are defined at the most basic level. He makes a strong case that until these problems are resolved, college and universities should not embed microaggression-related curricula into their diversity and tolerance programs, let alone center those programs on microaggressions — and that to do so could actually be harmful.

Before proceeding, it’s important to note that if you define microaggressions simply as offensive, relatively minor things majority-group members say about minority-group members, of course they exist, and of course members of minority groups frequently find them vexing and annoying and hurtful. None of this should really be in dispute. But as Lilienfeld notes, ever since a group of researchers led by the Columbia Teachers College counseling professor Derald Wing Sue launched the concepts of microaggressions into the modern mainstream discussion with a blockbuster 2007 paper, they and others have gone well beyond this. Their work has led to the increasingly widespread adoption of the microaggression concept in campus inclusion and diversity trainings, often at the expense, given perpetual budget constraints, of other sorts of programming. So saying microaggressions exist is different from saying they should serve as a cornerstone of the conversation about campus diversity and inclusion.

Early on, Lilienfeld runs down what he views as the “five core assumptions” of the microaggressions research program.

1. Microaggressions are operationalized with sufficient clarity and consensus to afford rigorous scientific investigation.

2. Microaggressions are interpreted negatively by most or all minority group members.

3. Microaggressions reflect implicitly prejudicial and implicitly aggressive motives.

4. Microaggressions can be validly assessed using only respondents’ subjective reports.

5. Microaggressions exert an adverse impact on recipients’ mental health.

If you think about this list carefully and then look at how microaggressions are discussed and written about by the architects of the MRP, you’ll likely agree that if any of these conditions fail, it would cause a lot of conceptual confusion. If microaggressions aren’t clearly defined or can’t be identified in a fairly specific way, for example, then it doesn’t make sense to center them in diversity-training settings: One person could contest another person’s claim of having been microaggressed (which as we’ll see would run contrary to one of the key claims about how microaggressions “work”), and it would be quite difficult to study them meaningfully.

Lilienfeld goes on to show — again, convincingly — that there isn’t strong empirical support for any of these five core assumptions. It’s a long, detailed paper that would take a while to sum up in full, so I’ll focus on some highlights.

A huge, confusing range of statements are considered to be “microaggressions,” which makes it hard to pin the subject down. As anyone who has dipped a toe into this world can attest, microaggressions can be just about anything. It’s a microaggression, according to the leading scholars working on the MRP, to say America is a “land of opportunity” or a “melting pot,” or for a clinician to say to a minority client, “I don’t usually do this, but I can waive your fees if you can’t afford to pay for counseling.” It’s also a microaggression to refer to something as “politically correct,” “lame,” “or trash.” At the more extreme end of the spectrum, it’s a microaggression to call someone a “credit to their race,” or to offer similarly backhanded, racialized compliments. (All of these examples come from Lilienfeld’s paper, cited either from the MRP research itself or individual college’s guidelines on potentially problematic speech.)

Lilienfeld argues that “it will be essential to shore up the microaggression concept considerably by better delineating its boundaries” in future research. “It will be especially crucial for scholars to explicate not merely what constitutes a microaggression, but what does not [emphasis his].”

The MRP hasn’t proven that minority group members exhibit any agreement on what constitutes a microaggression. If, as the MRP claims, microaggressions harm minority group members seriously enough to warrant this line of research and interventions stemmed at identifying and reducing such utterances, then the first step toward proving this is to establish that minority group members share some baseline level of agreement on which sorts of statements are and aren’t microaggressions. An easy way to check this is to take a sample of a given minority group, explain to them what microaggressions are (since the term probably hasn’t fully broken out of the academy yet), and then ask them whether they think a series of statements qualify. If just 20 percent of members of a group consider a statement to be a microaggression, then maybe it isn’t one; if, on the other hand, 81 percent agree, then we’re getting somewhere.

Now, in some cases this is fairly obvious — most black people would likely agree that “You’re a credit to your race” is a microaggrassion. But as we just saw, microaggressions encompass many far more ambiguous utterances as well, so it’s key to establish that there is at least some level of consensus among the members of the affected groups.

This is fairly basic legwork, but, as Lilienfeld shows, the MRP hasn’t done it yet. In fact, only “one published study” has evaluated this, writes Lilienfeld, and it didn’t do so rigorously. The study in question examined whether a group of white counselors and African-American clients agreed on statements like “My counselor avoided discussing or addressing cultural issues in our session(s)” and “My counselor may have thought at times that I was overly sensitive about cultural issues.” Each of these behaviors was deemed by a separate focus group to be a microaggression, and the researchers found fairly high agreement between the counselors and the clients on whether each behavior occurred — but the study never concluded that the clients themselves viewed the behavior as microaggressive. “Hence,” writes Lilienfeld, “these findings tell us only that clients agree on whether their therapists performed certain behaviors deemed by MRP proponents to be microaggressions, not on whether they agree that their therapists engaged in microaggressions.”

Along those same lines, the MRP has long claimed that microaggressions do harm by communicating implicit messages, but this, too, has never been proven. To take one characteristic example, microaggression researchers have claimed that a statement like “I am color blind” or “We are all humans” communicates to a minority group member that their racial/ethnic background is being devalued. Again, this is an empirical psychological claim — When minority group members hear X, it triggers Y thoughts. But as Lilienfeld writes, “There is no research evidence that the microaggressions identified by Sue et al. are linked, either probabilistically or inexorably,” to the negative messages in question. It’s more or less just the opinion of microaggression researchers. That is an insufficient basis for an important scientific claim.

The research linking microaggressions to negative mental health outcomes is fatally flawed. Perhaps the most important claims made by microaggression researchers involve the cumulative toll microaggressions can take on their victims. Sue has argued that “The invisibility of racial microaggressions may be more harmful to people of color than hate crimes or the covert and deliberate acts of White supremacists such as the Klan and Skinheads.” Elsewhere, he and other microaggression researchers have claimed microaggressions “have major consequences not only on the mental health of the recipients, but also in creating and maintaining racial inequities in health care, employment, and education”; that the “harmful and detrimental effects of microaggressions … on the health and psychological well-being of individuals is undeniable”; that microaggressions “have powerful detrimental consequences to people of color”; that they “shorten life expectancy”; and that they (this is Lilienfeld paraphrasing) “foster suicidal ideation.”

These are empirical claims — explosive ones — based on a handful of studies showing that people who report having experienced more microaggressions also seem to have worse mental and physical health outcomes. But the intensity of these claims is “surprising,” writes Lilienfeld, “given the inherently correlational nature of the data linking microaggressions to mental and physical health outcomes and the dearth of longitudinal data linking microaggressions to such outcomes.” In other words: Just because people who report experiencing more microaggressions also report worse health and psychological outcomes does not mean there is a causal relationship between the two; correlation does not imply causation, as snooty statistics professors everywhere say.

Lilienfeld goes on to identify a major flaw in the research underpinning the most serious claims about microaggressions: It doesn’t control for personality. Specifically, there’s a well-known psychological characteristic known as “negative emotionality,” or NE — a “propensity toward depression and anxiety, and a tendency to react to stressful situations with unpleasant emotions,” as one press release puts it — that has long been a thorn in the side of researchers trying to make causal claims about various stressors, both race-related and not. Microaggression researchers haven’t accounted for it at all, which makes it very difficult to draw any concrete conclusions from their research linking microaggressions to negative outcomes.

People high in NE, writes Lilienfeld, are more likely “to experience aversive emotions of many kinds, including anxiety, worry, moodiness, guilt, shame, hostility, irritability, and perceived victimization” than people who are low in it, and are also more likely “to be critical and judgmental of both themselves and others, vulnerable to distress and emotional maladjustment, and inclined to focus on the negative aspects of life.” So when a given study shows a link between a stressor and an outcome without controlling for NE, the results can be misleading — it might often be the case, for example, that people higher in NE perceive more slights or stressors than those who are low in it. This is such a persistently annoying ruiner of these sorts of studies that one researcher back in the 1980s “went so far as to remark that one of the foremost challenges in the construction of psychopathology measures is to develop self-report indices that are uncontaminated by NE,” as Lilienfeld puts it.

To take one example, Lilienfeld cites a study which found that “the well-established association between job stress and work-related outcomes, such as job satisfaction, was markedly reduced, often to statistical nonsignificance,” when negative emotionality among workers was controlled for. In other words, a lot of the times when we hear that stress is causing a bad outcome, negative emotionality may be a much more important driver of that outcome than the stress itself.

Okay, so back to the studies linking microaggression exposure to negative outcomes: Not only has this research not controlled for NE, but some of the questions asked to assess microaggression exposure are asked in a particularly sloppy way that seems almost designed to bite off big chunks of NE. For example, the so-called Racial Microaggressions in Counseling Scale asks respondents to note whether a series of microaggressions “never happened,” “happened,” or “happened and I was bothered by it.” It seems safe to say that people with high NE are more likely to say something bothered them, so if it’s also true, as a bunch of research suggests, that high NE is associated with poor outcomes, then studies based on this scale may be seriously confounded. It wouldn’t be the first time NE acted as this sort of confound in research dealing with discrimination. In one study, for example, Lilienfeld notes that subcomponents of NE “accounted for 42% of the association between self-reported discrimination and depressive symptoms.”

Now, to be fair, not every microaggression scale is as potentially contaminated as the Racial Microaggressions in Counseling Scale; Lilienfeld encourages researchers to use the Racial and Ethnic Microaggressions Scale, which he says is better. Plus, Lilienfeld notes that it will likely be the case that once NE is controlled for, there is still some correlation between microaggression exposure and negative outcomes (though that still wouldn’t mean causation). But for now, we just don’t know, and claiming a causal link between microaggression exposure and mental-health symptoms — let alone something like suicide — is scientifically suspect, since there just isn’t yet evidence to support this view, especially in light of all the questions about the woolly way in which microaggressions themselves have been conceptualized.

Microaggression-oriented education could be harming students. As Lilienfeld points out, the history of psychology is littered with interventions that ended up having iatrogenic — harmful rather than salutary — effects. “For example, Scared Straight interventions for adolescents at risk for delinquency and critical incident stress (crisis) debriefing programs for people exposed to potentially traumatizing events have been associated with negative effect sizes in some randomized controlled trials.”

We don’t yet have any evidence microaggression education programming is harmful — we don’t have much evidence about such programming, period — but Lilienfeld expresses worry that it could harm minority students by telling them (1) a bunch of statements they would otherwise have viewed as ambiguous or not worth harping upon are, in fact, microaggressions; and (2) microaggressions have been linked to all sorts of negative outcomes, including suicide. There’s evidence, particularly from the PTSD literature, that if you tell people “You are probably going to get these symptoms,” something like a reverse placebo effect kicks in. And since (1) is hampered by the fact that no one has yet presented a clear definition of what microaggressions are, and (2) lacks scientific evidence, there’s no reason to take this risk.

It’s also noteworthy that one aspect of microaggression training runs directly counter to an important idea from cognitive behavioral therapy: the importance of not falling into spirals of so-called “mind-reading.” Lilienfeld explains:

[M]any of the implicit messages posited by Sue and colleagues appear to reflect quintessential examples of what cognitive-behavioral therapists (Burns & Beck, 1978; Freeman, 1983) term the cognitive distortion of mind-reading, in which individuals assume—without attempts at verification—that others are reacting negatively to them. Cognitive-behavioral therapists typically regard mind reading as a subtype of the broader tendency of individuals to jump to premature conclusions. For example, Sue et al. (2007, p. 276) regarded the question “Where were you born?” directed at Asian Americans as a microaggression because it reflects the assumption that recipients are “different, less than, and could not possibly be, ‘real’ Americans” (p. 76). Yet most cognitive-behavioral therapists would maintain that leaping to this inference without attempting to check it out constitutes mind reading, as the intent of this question is compatible with a host of interpretations. Although it may indeed reflect the aforementioned assumption in certain cases, in many others it may reflect genuine and sincere curiosity regarding an individual’s culture of origin.

The point here isn’t to adjudicate the “true” meaning of an ambiguous question like “Where were you born?” The point is that a cornerstone of CBT protocol is that to help patients who are feeling bad, it’s important to get them to identify and try to avoid maladaptive mind-reading tendencies. Microaggression trainings, as they are currently constituted, may well be doing exactly the opposite. That should worry anyone concerned about mental health.

Until we know more, there should be a moratorium on microaggression trainings. After marshaling a bunch of evidence, Lilienfeld lays his cards on the table:

Based on the literature reviewed here, it seems more than prudent to call for a moratorium on microaggression training, the widespread distribution of microaggression lists on college campuses, and other practical implementations of the MRP (e.g., the insertion of microaggression questions on student course evaluations), at least until the MRP can take heed of many or most of the research recommendations listed here (see Table 1). Not only is the MRP still in a premature state of scientific development, but there is insufficient justification for concluding that the potential benefits of microaggression training programs outweigh their potential risks, including a substantial increase in the number of false-positive identifications of statements as microaggressions. To be clear, this proposed moratorium does not extend to the MRP itself, which should continue without interruption, albeit in substantially modified form.

That last sentence is key: This is a perfectly valid thing to study. But the swift way in which microaggressions migrated from a small group of researchers who determined, effectively by fiat, what is and isn’t a microaggression, to real-world applications on college campuses, is a case study in how not to apply psychological claims and theories. Especially when it comes to subjects as important as race relations and minority students’ mental health, it’s simply irresponsible for colleges and universities to be exposing students to untested, unsupported interventions involving microaggressions. Until we know more, at least.

How Microaggression Training Could Harm Minority Students