3 Takeaways - Unconscious Bias is Real, So Are the Solutions: Harvard Kennedy School Former Academic Dean Iris Bohnet (repost) (#103)

Episode Date: July 26, 2022

How can we reduce or neutralize unconscious bias? It’s a critical question these days – especially with DEI in mind – answered by an expert: Iris Bohnet, the former Academic Dean of the Kennedy School and co-Director of the Women and Public Policy Program. She calls it “unfreezing” our minds, and offers some surprisingly simple solutions.

Welcome to the Three Takeaways podcast, which features short, memorable conversations with the world's best thinkers, business leaders, writers, politicians, scientists, and other newsmakers. Each episode ends with the three key takeaways that person has learned over their lives and their careers. And now your host and board member of schools at Harvard, Princeton, and Columbia, Lynn Thoman. Hi, everyone. It's Lynn Thoman. Welcome to another episode. Today, I'm excited to be here with Iris Bonet. She's a professor and academic dean at the Harvard Kennedy School. She believes that firms are wasting money on diversity training because most programs

just don't work. They don't change attitudes, let alone behavior. And rather than run more workshops or try to eradicate the biases that cause discrimination, she believes that companies and organizations need to redesign their processes to prevent bias. Today, I'm excited to learn how simple, evidence-based changes can reduce and neutralize the bias behaviors in classrooms, police

departments, and boardrooms, and in hiring and promotion. Welcome, Iris, and thanks so much for our conversation today. Thank you so much for having me, Lynn. My pleasure, Iris. Let's start with understanding what unconscious bias is. What is it and why does it exist? a couple of years ago, a few colleagues of mine wrote a case study on her. And many of our listeners, I'm sure, have been in graduate school and participated in these types of cases where we analyze the case and we learn about the protagonists, about what they did, how they performed, how they had success and failures. And that's exactly the kind of case that this was about. But then a few years later, people took that very case and replaced Heidi's name with Howard. And now we use this case study, in fact, across the country to teach our students about the power of implicit or unconscious bias in a matter of minutes.

The students prepare for the case, not knowing that there is a second protagonist called either Heidi or Howard. And they come to class also having filled out a questionnaire where they evaluate how well Heidi and Howard have done. And what we find time and again is that students agree. And when I say students, I do mean male and female students, that students agree that Heidi did a great job and Howard did a great job. In fact, objectively, they were both great. But we just aren't quite as comfortable with Heidi

because she defies our stereotypes of what a typical venture capitalist looks like. And she defies our stereotypes of what a good woman does. She was just a bit too assertive, a bit too successful in her life. And that's unconscious bias. Of course, unconscious bias has nothing to do with gender or race or nationality or religion or body size or height. But we have documented unconscious bias in all of those dimensions. It has to do with the counter-stereotypical individual. So male nurses also experience unconscious bias because people aren't used to seeing men in those types of roles. And female venture capitalists experience unconscious bias

because we're not used to seeing women in those roles. Unconscious bias, does it have anything to do with tribalism, with people associating with their group, either by gender or by race or ethnicity or anything else? It's a very good question, Lynn. I often distinguish between two ways in which we can describe bias. One is the device that I just alluded to, and I want to call those stereotypes, as in, do you belong to the group that I'm looking at? Am I thinking that you are the typical type of character when I think of venture capitalism? That's a stereotype problem. What you are now asking about is the problem that we sometimes call affinity bias or in-group,

out-group preferences. We're more comfortable with people who look like I do. And we, in fact, see both. We see that I judge the kindergarten teacher based on whether or not, to use a counter-sturgical example, whether he fits my expectations of the typical kindergarten teachers. And then secondly, I'm also more comfortable with people who look like me. So if that person was a white Swiss woman who has done synchronized swimming, as in my case, then I'm more comfortable with that type of teacher than if the person looked very different. And both are real. So stereotypes and in-group preferences have been very well documented and often hold us back from seeing talent or promoting talent,

hiring talent, where talent really is. And how early do you see unconscious bias or affinity bias in children? Very, very early. This is not my own research, but it has been documented in newborns almost. They recognize the faces that they are familiar with, the skin tone of the faces they're familiar with. Certainly they recognize their parents really, really early on. And by the way, when I say parents, it doesn't have to be parents. It's just a primary caregiver. Whoever they see most, they are most familiar with. And so it's documented them really early on. And familiarity, again, is not something that we

should be ashamed of. In fact, I am not the one pointing fingers. I am saying that that's human nature, that we have been, so to speak, programmed to have those types of impulses. And so the important question is not to train this out of our minds, but to think about how we can change our environments to make it easier for all of us to get this right. Where do you see unconscious bias or affinity bias in society? Everywhere, Lynn. That could be a very quick answer,

but it has really been documented in many, many different places that people enter a room and they don't know anyone. And so we are much more likely to join a group that kind of looks like we do, whether that's in terms of race,

ethnicity, language, of course, it's not a big one, like we do. Whether that's in terms of race, ethnicity, language, of course, is not a big one, culture, gender, anything that's visible. And that makes us feel like we have an easier entry into that group. But it has been documented as early as with children, in that children kind of look for people who look like they do, including in looks, body size, height. We see that people tend to affiliate with people who are similar in many dimensions. And then of course, we see it in professional life as well. So sadly, my answer is in fact really true. It is everywhere. It is everywhere. Very sadly. Many organizations have tried diversity and inclusion training. How well has that worked?

Not so well, unfortunately. And it is a sad truth for many organizations. I know most all organizations still engage in training. And so I think the bigger question is, how could we improve our trainings? Because I don't think we will get rid of them very soon. And there is a purpose for them too. What good trainings can do is to raise awareness. And I don't want to downplay that. Even though we might think it is more broadly shared, the understanding of unconscious bias is increasingly shared by many people. And people understand that stereotypes are real and that in-group preferences are real.

It's still important for people to understand that it is real and that in-group preferences are real, it's still important for people to understand that it is likely them as well, that it's all of us. This is not a problem that happens elsewhere, but it happens in our very own lives, in our own organization. In a very ideal world, what I sometimes call is happening in those trainings is unfreezing our minds, making us open to different possibilities. The possibility that somebody else might perceive the room that I see in a certain way, very different because they're used to standing in the back of the room. And I might be used to standing in the front of the room. And we've all had that experience. It looks very different. So this perspective really matters. And that's what I

want to call unfreezing. But then we can't just let, you know, go to people back to their daily lives, if they're unfrozen minds, as in being open, maybe a bit more. It's a vulnerable place to be unfrozen. So you will go back to old habits, to old behaviors. This is not an intentional move, but it is often very unintentional, just to go back to our old habits. That's the second part of this awareness raising, that we have to unfreeze the minds, and then we have to give people the tools to, in fact, follow through on their virtuous intentions, and then eventually refreeze, refreeze the new habits, the new ways of behaviors, and that often means a systemic change rather than change just in our brains. Before I ask you about how to get the systemic change, what can we do to reduce and neutralize biased behaviors as we raise our children?

I myself have not done research on children, but I am going to give you some generalizable insights from that type of research. Seeing is believing. It's very important that our children see all kinds of different people in all kinds of different roles. They need to see female astronauts. They need to see black astronauts. They need to see white male teachers. That's really important. And that, of course, affects our media, affects the books that our children read, affects the cartoons our children watch. And that, in fact, is an area where Lynn, I think, compared to when we grew up, we actually have made quite a bit of progress. So seeing is believing, and there's some really good research showing that even short exposures to such of these role models that I just now described,

at an early age and even at a later age, the research has been done in elementary school and has been done in high school of bringing in, for example, female scientists to increase the likelihood that girls think that STEM fields could be for them. Even short exposure can affect career trajectories. I think that's something that parents could easily do. Just make sure that your children have all opportunities available. Again, we don't want to brainwash them. Exposure to the whole world, not in a stereotypical way, would be very useful. And so exposure to the whole world is important for raising our children. Looking at biased behaviors in institutions like corporations or police departments or boardrooms, what can we do to systematically reduce and neutralize biased behaviors?

The first big insight is that we should not sideline our diversity, equity, inclusion efforts in our organizations. And as you say, Lynn, these could be NGOs, these could be for-profits, these could be universities, these could be police departments, these could be for-profits, these could be universities, these could be police departments, these could be agencies. So when I say organizations, I literally mean any system of some sort of collaboration where we work together.

And I don't think we have, in fact, used the same kind of rigor and scrutiny and data and accountability and metrics in our efforts on diversity, equity, and inclusion as we have, for example, even in our marketing departments. That is the type of approach that we should be using. And then in contrast, we really haven't done that for diversity, equity, and inclusion. Really measuring what works and what doesn't work and using data to inform our decision making. It's both a simple and a big first point. It's simple because we do it in many other departments, but we just have to use that same type of business figure in our diversity inclusion efforts.

If you look at the different stages from hiring to mentoring to promotion, what do you see in terms of bias and how can we systematically de-bias at each stage? This is a big question. Let me do the journey of an employee and focus on where bias could creep in and what specific things organizations could be doing. It often starts with the sourcing. Organizations have to identify potential applicants. And so where do you go? Where do you advertise? Is the first important step. We have to cast the net much more widely, much more inclusively, really understand the barriers that people experience and that sometimes along racial lines, but sometimes also along geography. I think casting the net widely and thinking about where you source your talent is very important. And then secondly, I just alluded to these job advertisements that we place, and that's really, really low-hanging fruit in that we now have the

methodologies, the algorithms available to help us de-bias the language that we use in job ads. We now can show which words are particularly gendered. We have a library of gendered words, which we know will be more likely to attract men or women respectively. For example, warm and caring, we're more likely to associate with women, whether that's true or not. And assertive and even leadership, we're more likely to associate with men. So making sure that you don't use very biased language in your job ads, or if you have to use one of these terms, counterbalance them with terms that are more likely associated with the other gender is really, really helpful. And so that's just the entry stage. And then we have to look at

how we in fact evaluate these applicants. And sadly, evaluation is the home of bias in all kinds of different areas. It's very hard for us to be unbiased in those evaluation tasks. And so what do we do? Let me just debunk some myths. Having diversity on your evaluation committee is not going to solve the problem by itself. It's helpful, but it's not true. Let me give you an example about stereotypes, a personal story. My husband and I brought our son when he was a baby to a Harvard Daycare center, and we had to hand the baby to a male caregiver. And the fact that my husband was male didn't protect him from being nervous about a male caregiver. this moment of shock and surprise that that's just not what we expected of course it was wonderful and everything turned out great but these stereotypes are shared independent of your own gender so i think that's just an important message in that that's why the evaluation committee is important it's important to have diversity on the committee but it doesn't protect

us from every sort of bias what it does help with are two things the other bias that we mentioned before and that's this in-group preference. That if we just go with people who look like we do, who we think are in quotation marks, are fit for the organization, then it is helpful to have different representation, different people with different histories, different approaches, different perspectives on that committee who might have different preferences in terms of replicating themselves. The second reason why it helps is that, of course, a candidate also feels encouraged if they can see somebody like themselves in the organization evaluating them. So if everyone looks completely different than they do, that already might be a warning sign that that might not be placed for me. So the message here is diversity under committees is helpful, but don't think it is going to solve the unconscious bias problem because stereotypes are shared.

And the second question is, is it even a good thing to have a committee? And that's actually where I feel quite strongly that we have a lot of evidence that groups often fall prey to groupthink, where we don't benefit from the individual representative's intelligence, but rather fall prey to converging

to maybe the loudest voice in the room or to what the person just before me said. So we have to be very, very careful about groupthink on these evaluation committees. So therefore, my advice typically is to say, do not interview candidates as a group, but interview them as an individual. So that's my second message. Don't do group interviews, do individual interviews. The third important message is that we generally tend to overvalue the interview. Sadly enough, one of the worst

predictors of future success in the organization is the unstructured interview. The unstructured interview is an interview where we just have a free flow conversation, where I might interview somebody. I'm serving as academic dean at the Kennedy School. So one of my jobs is, in fact, faculty hiring, promotion, et cetera. So I might then just use a free flowing conversation where we talk about the person's research, their teaching, but also their hobbies. And then we might both discover that we enjoy hiking in the Swiss mountains. And of course, already I'm going to adore the person. When in fact, Swiss mountain hiking has no predictive power in determining who could be a great professor at Harvard. That's the problem with these unstructured interviews. So do not do unstructured interviews. If you do want to do interviews,

then do them in a structured way. Think about the questions beforehand. Use questions which have been proven to be predictive of future success in your organization and use the same set of questions with all of your candidates. So if these are five questions, use the five questions in the same order and give every question score and then evaluate them comparatively. They often refer to this as horizontal evaluation rather than vertical evaluation. Can you talk about career advancement, how to de-bias that and practical methods people can use?

That's in many ways the elephant in the room. I am not saying that we have solved our diversity problem at the entry level. We have not. So we have to keep working on diversification at the entry level and the hiring stage. But we also know that this pyramid here is real, where we have more diversity at the entry level and then much less at the top. And that's true both for gender and for race in the United States in particular. So the question is, what happens there and how do we lose our talent? My first observation is actually related to something that you might want to call systemic, but really has more to do maybe with culture than with a firm system.

And that's performance support bias. And I'm starting with that because that can happen very, very early on. And I first came across it in a law firm. They have done the important work that has to happen first to diagnose the problem and understand where the issue is and not just throw money at the problem and have realized that they have enormous diversity entry level, different gender, with even more women as first year associates than men, and also more diversity in terms of race and even sexual orientation they could measure at the

entry level than at the partner level. And then we worked on their promotion procedures, their performance appraisals, but we just couldn't quite get it to work. And then we realized that the problem really started the very first year when somebody joins this law firm, where some associates are given more support to perform than others. And it happened unintentionally. These are not bad people making bad choices. This is all of us naturally being drawn to people who look how we do. And what they then did was to formalize this process. And that's, I think, a very important message for every organization, that we have to fix

our formal procedures. And I will go to performance appraisals and promotion process in just a moment. But the first stage is more in the informal, where some of these, I sometimes call them microvalidations happen. So pay a lot of attention to this more informal part of career advancement. Performance appraisals are another place where what I refer to as the model of bias evaluation process is a huge handicap for all of us. Really hard to evaluate performance, to assess performance, in particular in the types of jobs that many of our listeners probably are in, where we in fact refer to these processes as subjective performance appraisals.

We've actually just done some research with a financial services company on their performance appraisals. We've actually just done some research with a financial services company on their performance appraisal systems. They had used a process that many, many organizations, in fact, the majority of organizations that I've ever worked with employ and that I have had concerns about. They asked their employees to self-evaluate and then share their self-evaluations with their managers before managers make up their minds. And that's probably not rocket science to imagine that the employee self-evaluations could in fact impact the manager's assessment.

And if there are differences in either self-confidence or the cultural acceptability of shining the light on yourself, then maybe we will see gender differences. We might see racial differences. We might see differences by geography in how comfortable people are giving themselves high ratings. And so that was my fear. This company also then had a glitch in their system where they couldn't share self-evaluations with

their managers in one year. And of course, that's a bit like, how should I say this, finding Swiss chocolate for a researcher in that you find a research site where you don't have truly A-B testing that we normally like, where we have a control group and a treatment group, but it's close to an A-B test where you have a bit of a shock to the system that was unanticipated, and you can look deeper into what that shock did. And here's what we found. It's actually quite an interesting story showing that a deep dive can really guide the policy choices that a firm makes. We found that all women gave themselves lower evaluations than men, and this was particularly pronounced for women of color. So women of color gave themselves lower evaluations than men. And this was particularly pronounced for women of color.

So women of color gave themselves even lower evaluations than white women. We otherwise didn't find big race dynamics. We also didn't find, although we expected that, we didn't find huge cross-cultural dynamics in this firm. This is a multinational firm headquartered in the U.S. And now the question is, what do managers do with this information in the years where they have self-evaluations available? And what happened in this firm, I think, was happy news for the firm on the one hand, and then surprising news for the firm

in another dimension. The managers closed that gender gap completely. Managers, in fact, equalized men and women on average. However, that did not happen for the race gap that I talked about before, that we in particular found for women. And in fact, men added to what was a gender gap to start with and a race gap for women of color and added a huge race gap. So people of color ended up with much worse evaluations than white employees, and certainly much worse than their self-evaluations would have suggested. The short summary here is we find what I call supply-side-induced gender gaps,

where employees themselves start out with gender gaps, and the demand-side-induced race gap, where something we didn't see on the supply side was added by managers. And so now the question is, what happened in the year where self-evaluations weren't shared? It actually helped women of color, because now we take away that lower self-evaluation that they had started out with. And so we might want to call this an anchor. So we de-anchor managers from their low self-evaluations and managers just assess the employees the way they always would. It did have that impact, but it didn't overall close the race gap because managers independently added this big race gap. And in the end, men of color in particular were doing worse in this organization.

They decided to stop sharing self-evaluations. And in addition, they have introduced calibration meetings where we don't just stop with managers' assessments, but we meet as a group at the end. And we actually look at our data horizontally across all of our departments. And we look at whether we see certain patterns, whether that's by gender, by race, by nationality. And we ask ourselves whether these are in fact based differences in performance or whether

we see the patterns to a degree that we might be concerned about bias, whether that's along race or nationality lines. So these deep dives can really help inform what medicine, I sometimes use the language I know from the natural sciences as the diagnostic of the ailment can really inform the medicine that I then prescribe. Not just throwing money at the problem, but intervening where something is broken. And that's exactly what this organization has been doing. That's my recommendation for organizations. First, assess your performance assessments and how do you rate

them, how they're doing in terms of different demographic groups, and then intervene to fix what you think is broken. Promotions are similar. You first have to assess. We did this for another multinational company. This time they were headquartered in the UK just to see where are they? They were very focused on gender, where are they losing women? Is it because women aren't hired? Is it because they leave either involuntarily or is it because we're not promoting them? We found what we typically find in that we find a gender gap in promotions.

Many organizations who do that type of analysis have noticed, for example, that women have to be enrolled longer before they are promoted. They might also have found that even if there's no gender gap in promotions, women are promoted to less prestigious roles than their male counterparts. And those roles might not translate, for example, into salary increases or in real responsibility changes. So what the company ended up doing was to create a bit of a cheat sheet for each manager to show manager the track record of promotions by gender compared to the available pool. And that actually helped that organization a lot already. Very often, these decisions happen inadvertently, not by conscious decision-making,

but just because of the biases that we had talked about before, Lynn, which again is the affinity bias and the stereotype. And if you don't look the part, I'm just less likely to think of you. Iris, before I ask for your three key takeaways, is there anything else you'd like to discuss that you haven't already touched upon? What should I have asked you that I didn't ask you? The one thing I might want to just reiterate is the piece of culture. That formal processes, and much of my work focuses on formal processes, are hugely important. Absolutely a place to start. And numbers do matter. So whether you are the only one or whether you are part of a group,

it really does matter in how you're perceived. But that now leads me again to this question of culture. Much of this happens at the water cooler or at the golf club or some bar. So we have to think about the informal environment as well. And the one thing that I love for listeners to keep in mind are meetings. So many of us are in meetings a lot, maybe too much, but a lot. It's half informal, half formal. It's not quite like a promotion process, but we still have some

rules in our meetings. And those rules in fact can be shaped. And that's a place where you can have impact on people's experiences of inclusion and exclusion. Sometimes we refer to this as micro behaviors, where some of us might not be given credit for our comments. Some of us might be overlooked. Some of us might feel like they're not in the know because the meeting really has taken place before the meeting and coalitions have already been formed. So that's a place where I recommend people pay a lot of attention. Last question, Iris. What are the three key takeaways you'd like to leave the audience with today? First of all, I would love for you to address diversity, equity, and inclusion the way you address any other business challenge.

Use rigor, use data, use evidence to diagnose what the issue is and then fix what's broken. That also includes accountability. That accountability has to be clearly set on whose responsibility it is to solve this problem and to use metrics to in fact measure whether the problem has been solved. So that's my first message. We also know that people work towards certain goals. Being quite clear about our aspirations, our expectations of performance is as important in diversity, equity and inclusion as it is in our sales departments or in our finance departments. Second, don't try to fix minds, fix systems.

Go after systemic bias in your organizations, make a diagnostic, make a good effort to really understand which systems work, which systems don't work. Often that's inadvertent. It's not a conscious effort to exclude certain people. But there's a lot of research documenting biases that can creep in into how we hire, how we do performance appraisals, and how we promote our employees. So systemic issues are hugely important. And then thirdly, don't stop with the system, but also address your culture. Try to understand as best as you possibly can what's going on in terms of culture of inclusion and exclusion in your organization. And meetings often are a good place to start.

Iris, this has been terrific. Thank you so much. Thank you so much for having me, Lynn. If you enjoyed today's episode and would like to receive the show notes or get new fresh weekly episodes, be sure to sign up for our newsletter at 3takeaways.com or follow us on Instagram, Twitter, and Facebook. Note that 3takeaways.com is with the number three. Three is not spelled out.

See you soon at 3takeaways.com.