Most businesses place a heavy emphasis on implicit biases in their diversity programs, focusing on what they are, how they cause discriminatory employment decisions, and why most people have them (even when they are sure they are bias free). These diversity programs operate on the assumption that if executives, managers and employees are aware of their implicit biases, they will be on guard against them influencing their decisions in ways that are contrary to their conscious (non-discriminatory) intentions. These programs frequently rely on the Implicit Attitude Test (IAT), hosted online at Harvard University’s Project Implicit, to make people aware of their implicit biases.
The IAT, introduced in 1998, is now the most widely accepted indirect measure of people’s implicit attitudes, that is, their implicit biases. The IAT operates by generating a score indicating the test-taker’s (slight, moderate, or strong) implicit preference for one or the other of two concepts, for example, straight people and gay people, based on the comparative speed with which the test-taker (using a computer keyboard) associates positive, rather than negative, words with those concepts. The theory is that when the test-taker more quickly associates positive values with one of the concepts, (straight people, for example) she or he has an implicit preference for straight people, and an implicit bias against gay people.
The IAT has been taken more than 17 million times, and has been widely praised as a psychological innovation with broad practical implications. For example, in Blink: The Power of Thinking Without Thinking, Malcolm Gladwell characterizes the IAT as not just “an abstract measure of attitudes [but]a powerful predictor of how we act in certain kinds of spontaneous situations.” Nicholas Kristof, a New York Times columnist, argues that through the IAT we can learn people’s “subterranean racial and gender biases that they are unaware of and even disapprove of.”
Despite such praise, the IAT has been controversial since it was first introduced. Serious methodological concerns have been expressed about its reliability and validity. But in the context of a diversity initiative, the most important issue is whether the IAT accurately predicts a person’s real-world behavior; in other words, if a person’s IAT score indicates she or he has a strong implicit preference for white people over black people, is that person, in fact, likely to make racially discriminatory decisions or otherwise behave in a racially biased manner? Mahzarin Banaji and Anthony Greenwald, the developers of the IAT, have answered that question with a resounding “yes,” claiming the test “has been shown, reliably and repeatedly, to predict discriminatory behavior…” Two recent meta-analyses of virtually all past studies of the IAT’s predictive ability, however, have reached a very different conclusion.
A 2013 meta-analysis found that the IAT was a “poor” predictor of biased behavior, judgments and decisions, and that it “provide[d]little insight into who will discriminate against whom…” A 2016 meta-analysis found that the IAT had even less predictive ability than the 2013 study concluded, and found that changing a person’s implicit biases, as measured by the IAT, resulted in little or no change in her or his actual behavior. Perhaps anticipating these results, Greenwald and Banaji have recently acknowledged, contrary to their earlier claims, that it would be “problematic to use [the IAT]to classify persons as likely to engage in discrimination.”
Given this state of affairs, the IAT should be regarded as useless for diversity initiatives. Contrary to what might be supposed, however, this is a highly salutary development, for it allows businesses to stop worrying about their employees’ unconscious psychological preferences and start focusing on their actual practices, policies and culture. By discarding the IAT, businesses will be in a position to concentrate on designing and building workplace structures that assure critical employment decisions are unaffected by implicit biases—regardless of whether they are present in the organization. Four key concepts should guide businesses’ efforts in this regard: demand objectivity, force slow thinking, focus locally, and accommodate differences.
Demand Objectivity
An effective diversity imitative should first identify when, how, and where implicit bias might affect critical career-decisions, and then inject as much objectivity into those decisions as possible. Subjectivity is the enemy of diversity. When career-gatekeepers—those people who give assignments, select team members, submit formal evaluations, and influence compensation and promotions—have broad discretion as to whom to praise, advance and sponsor, they are likely to fall prey to “affinity bias,” the reflexive tendency to unconsciously favor (among similarly situated people) those who are most “like” them. Affinity bias can, however, be largely eliminated by, for example, removing the opportunity for open-ended, free-form evaluative comments; forcing career-sensitive decisions to be justified in terms of quantifiable metrics; collecting periodic diversity scorecards with respect to the demographics of career-enhancing assignments, team and project composition, and changes in authority and responsibility; and assuring promotion criteria are specific and clearly supportive of a variety of leadership styles.
Force Slow Thinking
Nobel Prize winner Daniel Kahneman recently pointed out the dangers of fast thinking and the value of slow thinking for important decisions. When we think fast—automatically, reflexively and emotionally—we are most vulnerable to the influence of our implicit biases. When we think slow, however, that is, carefully, deliberately and with explicit reference to factual information, we are far less likely to be influenced by these biases. An effective diversity initiative, should force career-gatekeepers to think slow when they make career-critical decisions. For example, businesses should require such decisions be made by at least two people who must discuss their respective views before reaching a final decision; establish a compensation and promotion process that requires explicit comparison of candidates; review career critical decisions—with the full knowledge of the gatekeepers—for patterns of implicit (and explicit) bias.
Focus Locally
Demanding objectivity and forcing slow thinking should be led at the local level. Managers best understand local practices and cultural attitudes, have immediate control over career-sensitive decisions, and can easily be held accountable for the success of diversity initiatives. Todd Warmer and Michelle Kia are correct in arguing, in their recent HBR post, that organizations need to “focus on ways to mainstream diversity and inclusion in local team practices.”
Accommodate Differences
All groups potentially subject to workplace bias—racial, ethnic, gender, sexual orientation, etc.—should be comfortable that their unique differences will not negatively affect their career opportunities. Take women as an example: women have babies, men don’t; women are typically expected to perform more childcare and housework than men, and women are far more likely than men to drop out of their careers for an extended period. Businesses need to take steps to assure that these gender differences will not trigger discriminatory career consequences. For example, women who take maternity leave should not lose career ground and mothers should receive the same career opportunities and face the same workplace demands as their male co-workers. Women who take advantage of flex-time policies, have less office face-time, or refuse to work “heroic-hours” should not suffer compensation or advancement penalties if their productivity and work quality remains high, and women who take time out should have “on-ramp” opportunities available when they return.
Undoubtedly, the academic debate over the IAT will continue, but from the perspective of an effective diversity initiative, the present evidence makes clear that businesses should forget about reducing implicit biases, and beef up their efforts to prevent such biases from influencing decisions that can affect career opportunities. The four principles we have identified should help businesses do just that.
About the Authors
Andrea S. Kramer and Alton B. Harris are the co-authors of “Breaking Through Bias: Communication Techniques for Women to Succeed at Work,” and have written more than 90 articles and blog posts on promoting diversity and overcoming negative stereotypes and biases. Ms. Kramer is a partner in McDermott Will & Emery LLP, and Mr. Harris is a partner in Nixon Peabody LLP. Contact them on Twitter @AndieandAl.