Detecting Hidden Bias

You may not see it, but its probably lurking among your managers and perhaps even in you.

Two people—one a nattily dressed young white man, the other a middle-aged black woman who is slightly overweight—apply for a job with your organization. They seem equally qualified, but the hiring manager has an inexplicable and slightly negative reaction to the woman. “I just can’t put my finger on it,” he tells you, “but I don’t think she’ll be a good fit.” You agree, admitting you just have a feeling the male applicant would be a better performer.

Are you, or is your hiring manager, harboring a bias against this female applicant—perhaps one based on age, sex, race or physical appearance? If so, is that bias unduly influencing your collective hiring decision?

According to analysis conducted by a Harvard University-led research team, it is entirely possible that you and your manager are biased—and that you don’t even know it.

Such hidden biases can be disastrous for the employees who suffer as a result of them; they also can damage businesses by leading managers and employees to make flawed business decisions in a number of areas, including hiring, promotion, training opportunities and project assignments. For HR, the task is clear, but daunting: Help uncover and address such bias before problems arise.

Hidden Instinct

Most people are more prejudiced than they think, according to mounting evidence. A web-based test developed by the Harvard-led research team found a significant degree of implicit bias among those tested—despite what researchers say were honest assertions by test takers that they harbored no prejudices.

The team developed the Implicit Association Test (IAT) as part of a project designed to detect bias based on several factors, including race, gender, sexual orientation and national origin. Researchers found that the highest levels of bias—70 percent or more—were directed at blacks, the elderly, the disabled, the overweight and other stigmatized groups. Furthermore, minorities internalized the same biases as majority groups.

Researcher Tony Greenwald, a University of Washington psychology professor, was one of the first to take the test—and was immediately struck by the results. “We were initially surprised to find these biases in ourselves,” says Greenwald. “After finding them in ourselves, we were not so surprised to find them in others.”

Ben Dattner, a psychologist and principal with Dattner Consulting, a New York organizational effectiveness and human resource consulting firm, and a professor at New York University, says the human propensity toward bias may not necessarily have nefarious roots.

“Part of the insights and wisdom of the study is that biases don’t necessarily stem from evil in the hearts of men and women,” Dattner says. “Making a quick categorization of people and situations was important in human evolutionary history, and recognition of this tendency is the first step to dealing with it and overcoming it.”

What It Means for HR

Bias—hidden or overt—can have implications for such decisions as who you select to join your organization and how they are evaluated, promoted and compensated. In turn, these decisions can affect employee turnover and quality of life. They also may lead to lawsuits, because those who score higher on implicit bias have been shown to display greater discrimination.

“As it pertains to the workplace and HR, bias decreases productivity because we spend time spinning in circles producing what I call ‘human capital waste,’ ” says Milton Perkins, SPHR, North Central regional director for the Society for Human Resource Management (SHRM) and staff leader of SHRM’s diversity expert panel. Hidden bias, he says, “will affect turnover, and, at the end of the day, people who are hurting inside will inevitably hurt someone else—they will hurt your business, impact your customers and drain your productivity.”

Quinetta M. Roberson, associate professor of human resource studies at the Cornell University School of Industrial and Labor Relations, says “implicit bias might represent a subtle strategy for establishing intergroup differences and/or facilitating micro-inequities between members of different groups.” The effect? “The level of access to opportunity and to organizational networks may differ for members of different groups.”

As a result, Roberson says, such bias can reduce a company’s day-to-day productivity by affecting employees’ ability to work in teams.

Hidden bias also can leave employers vulnerable to shifting demographics. Labor estimates show U.S. employers will face a shortage of skilled workers by 2010, and organizations that allow hidden biases to infiltrate personnel decisions won’t succeed at properly hiring, training, engaging and motivating certain types of workers, which will put them at a competitive disadvantage in the war for talent.

Help Workers Face Their Biases

Will individuals with an implicit bias always act in biased ways? Not always, argues researcher Greenwald. “We believe that people aware of their implicit biases can, if they wish, choose to suppress their expression by paying attention to their behavior in situations that allow possible discrimination,” he says. “However, most people remain unaware of their implicit biases.”

The key, then, is to start by helping managers and employees recognize these skewed perceptions.

Since implicit bias is—by its nature—subconscious and covert, it may be difficult, if not impossible, to get people to acknowledge that it actually exists. The challenge is that it is hidden from both the perceiver and the target, argues Roberson. As a result, “implicit attitudes are more difficult to assess, monitor and/or influence.”

Greenwald says getting people to take one or more of the IATs is “an excellent device that can be regarded as a first step” in building individual awareness of unconscious prejudices. “We do advocate that it be given in the workplace, just that it be anonymous and that it not be used for selection purposes,” he says.

Paul Steven Miller, a law professor at the University of Washington and former commissioner with the Equal Employment Opportunity Commission, agrees the test may be helpful in challenging people in a nonjudgmental way to think about biases they may harbor.

“If you collect a group of managers and supervisors within your organization and say, ‘OK, we’re going to do antidiscrimination training—is anyone a bigot?’ my sense is nobody will raise their hand because people just aren’t going to define themselves as a bigot or a racist,” Miller says.

“In fact, when someone is accused of discrimination, often people take great offense because it’s an ugly accusation,” Miller continues. “And yet those same people, when you dive in and peel back the layers, may have biases ingrained that are affecting the decisions they make, the assignments they give, and the promotion or hiring they do.

“The test is one way of getting at this issue,” Miller adds. “Whether it’s the only way or the best way, I don’t know, but it’s certainly one way to get people to talk about the real issues.”

The test certainly had a surprising effect on staff members of the Montgomery, Ala.-based Southern Poverty Law Center, a nonprofit organization dedicated to protecting civil rights and promoting tolerance, which was so interested in the IAT research that from 2001 to 2003 it provided funding to develop several tests. As a result, one might expect that center staff members would be more bias-free than other groups—yet the test revealed that they too had hidden biases.

“Bigotry is a persistent social problem in this country, and we can’t escape being socialized in this context,” observes Jennifer Smith-Holladay, the center’s senior adviser for strategic affairs. Smith-Holladay says her own results uncovered a preference for white people—a group to which she belongs—and a preference for heterosexual people, “a group to which I don’t belong.

“I discovered that I not only have some in-group favoritism lurking in my subconscious, but also possess some internalized oppression in terms of my sexuality,” Smith-Holladay adds.

Lesson learned? “In the case of my own subconscious in-group favoritism for white people, for example, my charge is to be color-conscious, not color-blind, and to always explicitly consider how race may affect behaviors and decisions,” Smith-Holladay says.

Spotting the Problem Organizationwide

To systematically spot potential hidden bias across an organization, Roberson encourages HR to use a “diversity dashboard” that looks at a range of diversity metrics. Beyond things such as affirmative action plans, Roberson says HR should look at hiring and promotion rates, career path movement, and compensation among different employee groups to spot inequities. Dattner recommends adding statistical analysis on performance appraisals to look for patterns of potentially biased evaluations.

Roberson also encourages HR to look at language used in various HR contexts—staffing, performance reviews, and the identification of high-potential employees and succession planning or leadership candidates—as a way of auditing HR systems for hidden bias.

The most obvious examples of loaded language involve using words to describe expectations about how people will behave, rather than their actual behavior, says Roberson, or using adjectives in a performance review to describe the employee, rather than simply stating what the employee’s performance was, or whether that level of performance was acceptable. “Joe Smith is lazy” is a much different assessment than “Joe Smith did not complete a single task on time, thereby failing to meet his goals.”

A more subtle case might relate to the verbs a manager chooses to describe the employee’s performance. For example, writing that Joe Smith “exhibited good teamwork skills” suggests that although Joe has demonstrated these skills in the past, he may not necessarily be expected to do so in the future. Writing that Joe “is a great team player” is a broader statement about Joe, not just his demonstrated behavior, and implies that he may be expected to be a team player in the future.

Reducing the Impact of Bias

To address bias head-on, experts recommend two primary weapons: inclusiveness and training.

“Responding to hidden bias is about creating and maintaining inclusive HR systems,” Roberson says. “For example, high involvement work practices such as coaching, mentoring and team-based work arrangements may be effective in increasing access and participation of all employee groups.”

In addition, creating diverse HR teams such as task forces, selection committees and the like may help prevent hidden bias—or at least reduce the effect of any one person’s bias.

Similarly, when it comes to performance appraisals, HR can use systems and methodologies such as multi-rater, 360-degree feedback that make it less likely that any individual or group’s biases will have an undue influence on the evaluation of anyone in the organization.

Roberson and Dattner also recommend diversity training for employees and managers as a way of raising awareness about bias. And Dattner suggests providing additional training for a diverse group of the organization’s best interviewers.

“After a series of job interviews, organizations have, but often miss, the opportunity to evaluate the interviewers,” he says. He advises looking for people who “have better radar to get beyond biases and demographics” and who ultimately “are really good judges of who will be successful,” then giving them training to further improve these skills.

Using multiple interviewers with diverse backgrounds and different perspectives is another way to help ensure that more valid and legally defensible selection decisions are made—and that the impact of any biases held by individuals or groups is minimized.

Dattner also suggests avoiding unstructured interviews in favor of structured ones, which ensure that all candidates are asked the same questions regardless of demographic characteristics or appearance.

Unstructured interviews can be biased and a poor predictor of actual job performance because interviewers “have a tendency to make snap judgments based on superficial criteria, and then spend most of an interview confirming first impressions rather than getting to know the candidate in an open-minded way,” he says.

As a result, such interviews accurately predict on-the-job performance only 20 percent of the time, says Dattner. By contrast, the success rate for structured interviews is 50 percent.

Setting up a blind applicant review system also can help prevent biased selection decisions. A 2004 study of job candidates with white- and black-sounding names by the University of Chicago Graduate School of Business found uniform discrimination across occupations and industries. Federal contractors and employers claiming to be equal opportunity employers discriminated as much as other employers. As a way to counter this bias, employment lawyers recommend masking the names and addresses of applicants before circulating resumes.

Hope for the Future

Greenwald says the extent of an individual’s implicit bias is affected by many factors that can change over time. Such factors include the person’s group memberships, the dominance of a person’s membership group in society, consciously held attitudes and the level of bias in the immediate environment. As a result, says Greenwald, these biases can be modified by experience.

Dattner also argues that biases can erode over the longer term as people get to know each other and move beyond initial impressions.

“One thing that research has demonstrated is that the more interaction you have with someone and the better you get to know them, the less likely you are to rely on stereotypes, instead coming to know an individual as a unique person, rather than as a ‘type,’ ” he says.

In short, Dattner says that as Harvard psychology professor and mindfulness expert Ellen Langer has argued, “you become less biased not by discriminating less, but by being more discriminating and learning about the nuances that differentiate individuals from other members of whatever groups they may be, or you may perceive them to be, members of.”

But could we ever be completely free of bias?

“Not in my lifetime or well beyond,” Greenwald contends. “Maybe it wouldn’t even be desirable,” he says, adding, “I don’t know that we believe that all biases are bad.”

Brian Nosek, assistant professor of psychology at the University of Virginia and a member of the IAT research team, agrees: “We need biases in many contexts—I have a bias against foods that make me sick. This is an adaptive bias. The challenge is to identify those situations in which our implicit biases contradict our explicit values. Those are the ones that individuals, organizations and cultures will likely want to confront and curtail.”

Pamela Babcock is a freelance writer based in the New York City area. She has worked as a reporter for The Washington Post and the News & Observer in Raleigh, N.C., as well as in corporate communications.