QUESTIONS FOR Joan C. Williams

This combination of having to be more competent, to show your competence more and having far more complicated politics is a very powerful (and often lethal) combination. And that’s even before we talk about the differential effects based on motherhood versus fatherhood. In my research I have found that the biases triggered by motherhood are even greater than the big three I just described.

Are individual leaders able to interrupt these biases?

Not alone, but leaders are an important part of the picture. The way to interrupt bias is through everyday workplace interactions, and that requires at least two different sets of stakeholders. One is HR, which should be designing formal systems like hiring and performance evaluations to interrupt bias. An informal study of the tech sector found that personality was mentioned in 66 per cent of women’s evaluations, but only in one per cent of men’s. That is a good example of what we call a ‘process metric’. If you want to pinpoint where exactly bias is playing out in your organization, you need to track these evidence-based process metrics — and HR has a central role to play.

Then there are the individual leaders who control the informal systems that play a powerful role in shaping the workplace experience, including access to career-enhancing assignments and access to mentors and sponsors. Bias breeds in informal systems, which is one of the reasons why it’s concerning that so many tech companies have abandoned performance evaluations. Instead, they just provide feedback on an ongoing basis. That is concerning, especially since one of the strong patterns we have found is that people of colour are not given developmental feedback in the same way as white people are.

You believe rethinking the way we recruit and hire is a critical first step in breaking down bias. What else can leaders do to drive diversity in the workplace?

The first message is, if you’re serious about diversity and inclusion, you need to be keeping process metrics. If your company has a problem with its sales function, what do you do? You look to the metrics, and you keep trying things until you nail the desired metrics. In recruiting you should be analyzing the demography of your pool, who survives resume reviews, who gets interviews, who survives interviews, who gets hired, and the startup packages provided. There might be biases playing out at every single one of those stages, in none of them or in only one — but you won’t know until you examine the process metrics.

With informal systems like work assignments, you have to ‘pin down the jellyfish’ because you likely don’t have a formal system to study. At biasinterrupters.org, we have a simple survey that leaders can use to find out who is doing the office ‘housework’: who is sending the follow-up emails, who is finding the time to meet, who is planning the celebrations? That’s all good team building stuff, but typically it’s not the path to promotion. This is another situation where you have to measure who is doing what.

The second step — and in some ways this is even more important — is to find out who is doing the ‘glamour work’ — those assignments that lead to promotions. Pin down what they are and then have individual managers keep track of who gets them. The research suggests that even a simple level of accountability on these things is a gentle bias interrupter that could have a big effect.

Tapping into people’s existing networks has always been a go-to way to fill open roles within organizations. Why should we be limiting that kind of referral hiring?

We shouldn’t, necessarily, but in most cases the single strongest determinant of who is in your social network is similarity. If your goal is to reproduce the demography of your current workforce in a certain level or a certain job, by all means, use referral hiring. If your goal is to take steps towards diversity, referral hiring is probably going to have to either be balanced out by different kinds of hiring or eliminated altogether. As indicated, you need to keep track of who ends up getting hired through referral networks and who ends up getting hired in other ways. If you have a lopsided demography, then you have a problem and need to be proactive about it. Businesses treat every other problem by seeking evidence and then using it to solve it. Why aren’t we doing that with workplace bias?

Why is it so important to ‘manage for bias’ on a daily basis?

Because bias is happening on a daily basis. To de-bias a workplace, you have to teach people to recognize it and arm them with strategies that interrupt it. One-off bias training can’t change a culture because doing anything once cannot change a culture.

If you’re serious about diversity and inclusion, you need to be keeping process metrics.

My colleagues and I have developed a next-generation approach to bias training called the Individual Bias Interrupters Workshop that presents the research, the bias, and an example of how it commonly plays out. Prove It Again, for instance, can play out as ‘the stolen idea’. We put people in groups and get them to brainstorm about how they could comfortably interrupt an instance of a stolen idea.

This accomplishes two goals. First, it leads to a reset in company culture. Once The Stolen Idea has been named as a bias and everyone has talked about it, that makes it a lot harder to do. Second, you end up with a toolkit for how to interrupt bias in a low-key way that doesn’t require you to spend too much political capital. This is really different from some companies’ mindset of, ‘if you see a bias, call it out!’ If that’s your jam, God bless you, but most people will not do that.

You have found that bias is ‘baked into’ performance reviews. What can be done about that?

There are ways to structure performance evaluations so that they can help interrupt bias, rather than help perpetuate it. These are best practices that help an organization be a more effective meritocracy. For example, you shouldn’t accept a performance evaluation with only a global rating. That is a petri dish for bias.

What makes for a good performance evaluation is what industrial psychologists have been telling us for decades: Figure out what the job is, measure the person against the job and provide evidence and accountability. This is important because in-group favouritism often means that objective rules are applied rigorously to out-groups, but quite leniently to in-groups.

That can also happen in a hiring context. You’ll have a national search and the rules of the search are applied very rigorously until a golden boy comes in and all of a sudden, he’s the ‘search of one.’ He doesn’t meet any of the criteria, so he’s never been judged against anybody who does, and yet he gets hired. This happens all the time.

As you have indicated, most organizations aren’t doing anti-bias training properly. How do we build a business case for doing it right?

I actually stopped working on a business case ten years ago. No matter how much evidence you have, it’s never enough to persuade the disbelievers. There will always be some reason why the business case doesn’t apply. I don’t think the issue is the business case. I think most people are convinced of the need to achieve diversity; but they don’t know how to get there. It’s more of a change management issue.

Can your prescription for interrupting bias at the team level be scaled up and applied enterprise-wide?

Absolutely. Individual managers can get started by figuring out who’s doing the ‘office housework’ and who’s getting the glamour work. They can say, ‘Okay, in filling out performance evaluations I’m going to make sure I measure people against objective criteria, rate them based on those criteria and provide an explanation, if only to myself, when I waive criteria for person A but apply it to Person B.’ They can ask HR to gather the process metrics to find out whether there’s a certain demography to their hiring and, if so, where the problem enters in. They can use individual bias interrupters: When the Stolen Idea occurs, they can say, ‘Hey Tim, so glad you picked up on Katie’s idea. You’ve added something important. I wonder if this is the next step.’ That’s a good example of a low-key bias interrupter. Such efforts, applied consistently, make enterprise-wide initiatives possible.

Which organizations do you think are setting the bar high around anti-bias right now?

The organizations that are setting the bar high are those that are keeping metrics and insisting that people meet them. At the risk of repeating myself, if you’re serious about this you need to do what a business does for any other problem: Use evidence, establish the current and desired metrics, and keep trying until you meet them.

Joan C. Williams is a Distinguished Professor of Law, UC Hastings Foundation Chair and Director of the Center for Worklife Law at the University of California, Hastings. Described as having “something approaching rock-star status” in her field by The New York Times Magazine, she has played a central role in reshaping the conversation about work, gender, and class over the past quarter-century.

In most cases, the single strongest determinant of who is in your social network is similarity.