by Matt Barr

The law's power to combat implicit bias

Unless you're knee-deep in contemporary legal scholarship, which certainly I'm not, the paper available for download here has to give you the creeps.

A fair restatement of its premise is: While discrimination law generally prevents or remedies employment and other decisions based on explicit, i.e. intentional, bias, it is inadequate for preventing or remedying similar decisions made under the influence of invidious implicit bias -- bias the decisionmaker may not know he or she even has. A "general strategy of debiasing through law," to quote the paper's abstract, might help put an end to this.

An example, provided in the paper, may be helpful.

An employer is deciding whether to promote Jones or Smith to a supervisory position at its firm. Jones is white; Smith is African-American. The employer thinks that both employees are excellent, but it chooses Jones on the basis of a "gut feeling" that Jones would be better for the job. The employer is not able to explain the basis for this gut feeling; it simply thinks that "Jones is a better fit." The employer did not consciously think about racial issues in making this decision; but, in fact, Smith would have been chosen if both candidates had been white.

Your immediate reaction (I hope) is, "according to whom?" The paper makes it a point to note that this is not as troubling an example if it's part of a demonstrable pattern of such decisions -- that is, the white employee always getting the promotion form the same employer in similar circumstances. In such a case, antidiscrimination law obviously provides a remedy. So "according to whom?" is not answered by an objectively observable pattern of results.

The answer is "lawyers and judges," of course. The paper does not go so far as to suggest Vulcan mind melds as a discovery tool, but seems, if I may editorialize, frustrated such a thing isn't available.

One of the goals here is to "encourage employers to adopt general decisionmaking structures or processes that reduce the intensity and frequency of implicit bias, implicitly biased behavior, or both." One of the steps toward achieving that goal would be "making decisionmakers accountable for their decisions." That phrase seems to understate what we're talking about, in the case of someone whose bias is unknown even to him. Does it mean making a decisionmaker fear punishment, increased compliance costs, or something else and taking that disincentive into account in the decision? Presumably, decisionmakers are already accountable for their decisions; this seems to be suggesting muscling them into deciding the way a lawyer or judge favors.

We could also hang pictures of Tiger Woods in the break room.

People show significantly less bias on the IAT in the presence of Woods's picture -- and also when tested again twenty-four hours after exposure to the picture. Thus, in the real world, if portraiture in the workplace or elsewhere consistently reflects positive exemplars, it is likely -- though certainly not guaranteed -- that those present will show less implicit bias, with likely mechanisms once more being the availability and affect heuristics.

Footnotes omitted. It is "beyond the scope of the present discussion" whether this would violate the First Amendment, which it surely would, at least as I read it.

The cherry picker's objection to all this is the "thought control" argument, which I dipped a toe into when I snarked about Vulcan mind melds. The paper notes that no one complains when the law outlaws certain biased behavior, nor gets involved when actors are mistaken in fact:

The normative problems are least severe when government is counteracting either factual mistakes or discriminatory behavior such as hostile work environments; and if efforts to combat such forms of biased behavior also reduce implicit bias, no one should complain in light of existing law.

But really, the alarming part of all this isn't bound up in some hyperbole about "thought control." (Don't get me started on this, though: "some people engage in biased behavior inadvertently or despite their own ideals. Such people want, in a sense, to be debiased... ")

The approach here springs from the following proposition: If there is something wrong with people, we need to figure out how to use the law to fix it. Starting from this premise, there is literally nothing the force of law can't be deployed to do, if any structural problems it might have can be overcome. Forget "thought control." This is a theory of pervasive government reach that is insidious enough not to need exaggeration.

If we're to use the force of law to (a) discover biases in people they themselves don't even know exist, and (b) correct them, what exactly is there that law can't -- shouldn't -- do?

Can it deicde what you read? Maybe there's no harm comparable to implicit bias in reading. But maybe there is, depending on what you're reading. Can it decide where you work? You may have a job at a chemical plant that pollutes the air and water and contributes to global warming, and you may not even realize it. Can it decide who you associate with? There may be harm in planning a family with someone predisposed to a genetic disease. Or cancer. Or autism. Or obesity.

Can it decide you can't abort a pregnancy? No, that's a Constitutional right! Of course it is. The paper though points out that "the law tolerates some government prohibitions on discriminatory behavior, even when they relate directly to speech," such as prohibiting certain posters or jokes in the workplace. It's a Constitutional right may be persuasive but is certainly, in this area, not determinative. Private decisions are different from ones made in the workplace. Maybe. So then we'll have to settle for saying you can't get a job working for a polluter.

The point isn't that law professors are running around trying to figure out how to control your thoughts, it's that they think they're omnipotent, or would be if they could just work out what each button does. There are places the law shouldn't go. Is anyone asking whether the law is an appropriate weapon in cases like this? Or is everybody too busy asking how can we to ask should we?

Browse books from Amazon.com: