- Blog post
Can people train their way out of unconscious bias?
I want to start by saying that I cannot give a definitive, yes-or-no answer to the question I’ve posed in the headline to this post. Sorry about that. And based on my reading, I don’t think anybody can.
But I think it’s a question worth mulling over for anybody who is in charge of employee training about diversity, discrimination or related topics.
Neuroscience has pretty well established over the past few decades that unconscious bias exists, and that it’s stuck deep in our brains. To what extent unconscious bias is “hard-wired” — evolutionarily innate to our species — or learned is an open question whose answer matters more to the behavioral scientist than to the average organization’s HR manager or director of training.
What SHOULD matter to such folks is to recognize that they, all their line managers and all their employees probably harbor biases of which they’re completely unaware, or imperfectly aware. This is a critically important concept: It means that no matter how much people have listened to moral teachers, preachers, or plain old corporate trainers expound the evils of discrimination, and no matter how much they’ve taken these precepts into their conscious minds, these same people are likely to exhibit unconsciously-generated bias against some group(s) of employees at some point in their careers, and maybe more regularly than that.
Thing is, though, that people don’t want to admit, to others or even to themselves, that they harbor such biases. So maybe a useful start to training would be to get people to understand that down deep, they may not be as unbiased as they’d like to think.
That’s where something called the Implicit Association Test comes in. The test (actually a series of related tests) is based on research from scientists at four universities — Harvard, Yale, and the universities of Virginia and Washington — and lives here online.
Categories of bias
The site offers to test your biases across a variety of categories — gender, race, color, ethnicity, disability, weight, sexual orientation, etc. — some of which involve characteristics legally protected against discrimination and some of which don’t.
I decided to take an implicit association test on race, specifically on attitudes toward African-Americans as opposed to whites. (You can see which group I fall into from my photo on this blog.) Although I’m aware that it’s pretty hard to be completely unbiased on racial issues if you live in America, I thought that someone who, like me, is married to a black woman and has non-white children might exhibit only a very small degree of unconscious bias.
Well, color me surprised. Once I’d taken the test, the site informed me that I evidenced the second-highest possible degree of automatic — i.e., unconscious — preference for white faces over African-American ones. I had a “moderate” preference for the former, putting me among 27% of the people who take the on-line test. (At least I wasn’t in the “strong” preference for whites, as a further 27% of the test-takers were!)
The site further informed me that 16% of those who took the test showed a “slight” preference for white people and 17% showed no preference. Just 12% of test-takers, all told, exhibited any preference at all — slight, moderate or strong — for African-Americans.
Good vs. bad pairings
What the test did was to pair a series of good terms, like “wonderful” and “glorious,” with photos of people of one race, and bad terms like “evil” and “unpleasant” with the other. It asked me to categorize a series of these pairings as quickly as I could by punching one or the other of two keys on the computer keyboard. Then the pairings were switched, with the good and bad adjectives switched to the other race, and I did my categorization thing again.
The test uncovers unconscious bias by measuring how quickly participants categorize the good and bad pairings. A test-taker — like me, apparently — who is slower to respond when black faces are paired with good adjectives, and quicker to respond when the same faces are paired with bad adjectives, is displaying some degree of bias.
Let me hasten to add that I’m not 100% sold on the test, and some critics in the wide world aren’t either. The Implicit Association Test has been attacked for purported shortcomings including a susceptibility to manipulation by cheaters, and a skew against older test-takers who take longer to answer.
Nonetheless, there’s serious science behind the test, and people in charge of training may find it useful as a tool to sensitize employees and managers to issues of individual and organizational bias.
Giving fairness a chance
So, to come back to the original question, can people train their way out of unconscious bias? Maybe.
Even if such bias proves so deeply rooted as to be impossible to eradicate, when you make people aware of it, and ask them to examine their decisions about co-workers or subordinates in light of it, you’re giving them a chance to consciously moderate their bias and act more fairly toward all.
And that’s worth the trouble, don’t you think?