Clementine Jacoby has a most unusual background. She graduated from Stanford in 2015 and subsequently reported, “I left Stanford thinking that I would be a professional circus performer.” Indeed, she had worked as a circus performer in Mexico and Brazil, specializing in aerial hoop. She spent her first year after college teaching acrobatics, not to Cirque du Soleil wannabes but rather in a Brazilian gang-diversion program, to encourage those individuals to put aside lives of crime.1
The experience left her with a profound understanding of the problems of crime and lawlessness, and it also taught her that rehabilitation was in fact possible. She then went to work for Google as a product manager for four years, working on Google Maps and Android, which gave her organizational experience and helped her meet and learn from a lot of tech talent.
But when 2018 rolled around, she found herself nagged by a feeling that she should be doing more to improve the world. She developed the idea for Recidiviz, a nonprofit designed to identify individuals in prison who might be eligible for early release without endangering the broader community. More generally, this was part of a bigger movement to bring data analysis to the U.S. criminal justice system. She applied to Tyler’s Emergent Ventures program, asking for enough money so that she could quit her job and start the nonprofit. Tyler liked her pitch and within a few days sent a substantial sum of money her way, no further questions asked. Clementine did indeed quit her job and proceed with her plan.
Recidiviz really took off with the pandemic, when states wanted to release many prisoners in order to limit their COVID cases in penitentiaries. Many states consulted with Clementine and Recidiviz as to who could be released safely, and the result was that tens of thousands of prisoners were released and many lives saved. North Dakota, for instance, released 25 percent of its prison population in one month in an orderly manner. Recidiviz is now a highly successful nonprofit, and it has attracted many millions of dollars of additional funding.2
When Tyler thinks back to his video interview with Clementine, he was impressed by a few things. First, she had the vision of building Recidiviz with highly talented individuals only, rather than relying on a series of bureaucratized roles filled from the typical nonprofit pool of employees. Instead she drew upon her friends in the tech world. Second, she seemed determined and willing to do something “weird” (it seems less weird now that it has succeeded), even though it meant a big cut in pay and unclear future job prospects, or perhaps the lack of any future job trajectory whatsoever. She really believed in the project, and she was taking the plunge with no obvious parachute, usually a good sign.
At the same time, it’s safe to say that Tyler and Clementine didn’t entirely click during that first meeting. The call went just fine, but she wasn’t playing the strategy of “the charmer.” It was the facts and the facts alone, and that matched her vision for Recidiviz as a data-driven enterprise to advise policymakers. If she was going to pull that off, surely her pitch should be rational and data-driven as well, and it was.
Fortunately, Tyler had appropriate expectations for the conversation. Though Clementine didn’t seem warm in a superficial sense, Tyler realized that she was doing her best to navigate the fairly limited range of behaviors allowed women in professional settings. We’ll discuss this issue further in what follows, but often when a woman is interviewed by a man, there is less emotional space for her to inhabit in ways that would be considered both appropriate and impressive. With many interviewers, she would not be allowed to be too assertive, or for that matter “too smiley.” Tyler felt she was, given these possible constraints, pitching herself and the plan just right. The substance of her case was thorough and impressive. And so the grant was made.
That is just one example, but we would like to teach you how to overcome a broader set of biases when it comes to gender and minorities. This is a chapter about how to overcome or at least limit your own biases. Out in the streets, you can be a social justice warrior all you want, but when it comes to talent selection and also the workplace, you may well be biased nonetheless. It is not possible to unpack each and every instance of particular bias (“What exactly is the right way to interview people from urban areas in Mozambique?”), so we will focus on some of the most general lessons. You can take that as a kind of superstructure for filling in particular pieces of learning depending on the issues you face in your own circumstances.
The first part of this chapter will consider biases related to women, who are a majority group in most societies. Later on, we will take those general approaches to bias and discuss how to be more perceptive and less biased when it comes to race, noting that most of this treatment will be set in an American context, rather than covering, say, bias against ethnic Chinese in Southeast Asia. Still, our goal is to uncover generalities rather than to detail each and every bias that might exist, so we hope the analysis will help you fight your own prejudices, or the prejudices of those you work with, no matter what your cultural setting.
And to be clear, we are writing this chapter as two white guys, if we may be perfectly blunt about that. We recognize that there is much about these topics we will never have an experiential understanding of, and that bias will remain in our treatment no matter what we do. We also are writing this chapter with the tone that we would use if we were speaking to people occupying positions of relative power in any situation and trying to educate them about their biases. We don’t intend that as a prejudicial decision, but we think that approach will give this chapter the greatest impact and influence. If you, as a reader, need to make some adjustments, please try to do so, and in the meantime, please don’t think we are intending to exclude you from this discussion. The whole point of this chapter is to help create the conditions for broadening the discussion.
Unfortunately, the arguments about the differences between men and women are usually frustrating and unproductive. Too often the core debate is about whether the differences between men and women are genetic and intrinsic or, alternatively, the result of socialization and thus perhaps malleable.
While we recognize those debates as important, we are deliberately leaving them aside as overdone and also as a distraction from our main mission. Rather than relitigating the James Damore memo (Damore was a Google employee who wrote about the intrinsic differences between men and women), we instead focus on a more practical question: How can we do a better job understanding women’s initial presentations of themselves and then harnessing their talents in the workplace? And if you are a boss or talent selector, how can you do a better job hiring women while improving your workplace and easing whatever injustices might exist in broader society? The points in this chapter are intended for the interview process, the promotion process, how you communicate with your employees, and how you should understand the workplace roles and environment you are creating.
It is better for everyone—yourself included—if you side with an emancipatory perspective that improvement is possible and you can be an agent of positive change. This holds even if you harbor very strong conservative views about the intrinsic differences between the sexes.
Here is why even the most conservative gender theorist among us should adopt and, furthermore, embrace an emancipatory perspective. Even if men and women exhibit systematic differences from birth, there still will be unfair discrimination against individual women. There is still considerable variation within each gender, and intrinsic gender differences will make it easier to miss the standout performers who do not belong to the favored group. Take tennis, where we do know that men on average have stronger and faster serves, almost certainly for intrinsic biological reasons. In that setting, it would be easy to neglect the talented women players, or to forget that women’s matches can be more interesting (longer, more complex rallies) and sometimes more popular with the crowd as well, all as evidenced by the world of women’s tennis. Yet in this setting, statistical discrimination can feel justified, and that can make it more stable. Nonetheless, you, as an entrepreneur and talent spotter, will have opportunities to spot the possibilities in women’s tennis, and also to spot the women who can beat most of the men, or perhaps the women who can innovate in tennis in ways the men do not. When it comes to gender, those opportunities may be all the greater precisely because many of the other talent spotters are wrapped up in the generalizations, even if those are correct generalizations on average. Indeed, it did take a long time for women’s tennis to rise to its current popularity and status.
Or say it turned out that, in fact, women were on average worse at some other workplace activity. Still, likely many women would be better at that activity than most of the men, and you would gain from spotting and enabling those women. So even if you think “women are worse at [fill in the blank],” it is still a mistake to dismiss female talent. These may be instances when the returns on talent spotting are going to be high, because other people are too attached to their statistical discrimination.
Thus, there is a strong case for believing that much more can be done to elevate the prospects of women in society and in the workplace. That case does not require any particular view on whether the observed differences between men and women are biologically or socially determined.
Before we proceed, a few points.
First, we are going to focus on women, but we also will offer some remarks on dealing with individuals from other cultures—“cultures” being broadly defined in a way that may include cultures from within your own country.
Second, we are going to make the discussion deliberately dispassionate and for the most part drained of anecdote and moralizing. We are well aware there is a truly massive literature on women in the workplace, including numerous personal accounts of bias, discrimination, harassment, outrage, and more, with some of those accounts being highly personal. We consider these presentations important but feel we do not have much to add to them, and so we will not attempt to systematically present all relevant injustices. If it sounds like we don’t “care enough,” that is because we are trying to mobilize your analytic, talent-spotting side toward practical, beneficial ends, rather than playing to your emotions.
Our drier approach will focus on a few key results from data science and what they can teach us about finding and mobilizing talented women. That means taking the kinds of research studies already presented and asking what they might imply for women in the workplace. That also means most of the arguments presented here are ones you don’t necessarily hear in the more popular discussions, and that is to ensure we are adding to the debate rather than just repeating what you already can find elsewhere.
Let’s now look at a few of the central empirical results relevant to women and talent.
First, note that women measure as having somewhat different personality profiles than do men. Women score higher than men on the traits of agreeableness, neuroticism, extraversion, and openness, with the largest differences coming on agreeableness and neuroticism. As always, we are using those terms in the formal sense specified in personality theory, and you should not assume either that agreeableness is intrinsically good or that neuroticism is intrinsically bad.3
In many cases, men and women have broadly similar scores on the Big Five qualities but for different reasons, and thus, there are still underlying personality differences. Male openness and female openness are basically the same, but within that category men score much higher on self-assertion and women score much higher on being outgoing and friendly. Women are also more sensitive and socially flexible, and men are more likely to form larger competitive groups with relatively stable hierarchies. Men also have a higher variance of agreeableness, while women have a higher variance of extraversion. Just from personality differences as measured by scores, whether a person is male or female can be forecast with about 85 percent accuracy. That predictability is so high because the algorithms are looking at the entire constellation of personality traits, not just trait-by-trait differences between women and men.4
One striking feature of the research literature is that personality for women predicts earnings with more power than personality for men. That is a result from multiple papers, not just one; it seems to be robust, and it holds for Canadian data as well. For instance, in what is arguably the most systematic study of this question, by Ellen K. Nyhus and Empar Pons, personality matters much more for women in the workplace. (In technical language, in the unadjusted regressions, personality has an adjusted R-squared of 0.7 percent for men and 5.0 percent for women—a big difference.) For women, emotional stability measures as a greater factor determining wages, as does agreeableness, which in many studies affects wages negatively for women. That is, the agreeable women seem to earn less, for whatever reason. In the Canadian data, a one-standard-deviation increase in agreeableness for women is associated with a 7.4–8.7 percent income penalty, but there is no corresponding income penalty for men.5
Melissa Osborne Groves wrote her dissertation on gender and earnings and found some striking results. Moving beyond Five Factor personality theory, she considers other factors that have predictive power for the earnings of women in the United States and the United Kingdom. For instance, she considers the factor of “externality,” “the belief that outcomes are the result of fate or luck,” and shows that a one-standard-deviation increase in a woman’s externality score is associated with a more than 5 percent decrease in wages. This may mean that a sense of agency is good for productivity, but be careful that such an active temperament doesn’t go too far in the wrong direction. A one-standard-deviation increase in the trait of “aggression” is associated with an 8 percent decrease in wages, while a one-standard-deviation increase in the trait of “withdrawal” correlates with a decrease in wages of more than 3 percent. Again, these and other coefficients in the paper show a stronger connection between personality and wages than what is typically found for male earners.6
It is striking to view the results on the personality trait of aggression as it relates to male and female earnings. For men, the trait of aggression correlates with higher earnings in high-status professions but correlates with lower earnings in lower-status professions. If you are a bowling alley attendant, you are not supposed to act like a temperamental founder, but a CEO can get away with this. Aggressive women, in contrast, have earnings penalties correlated with both high- and low-status professions.7
What should we make of the apparent fact that personality matters more for how much women earn? One obvious insight is that the talent search for women is more difficult or requires a subtler set of skills. For instance, effective, talented women seem to boast less and show less overt aggression, for fear of incurring workplace and broader reputational penalties. That arrangement is unfair, but still, it is a potential opportunity for you to find those individuals by other means—for instance, by asking a candidate and her references more explicitly about the candidate’s skills and degree of dedication.
It’s also likely true that some employers do not like all of the personality traits of women and/or fear that their workers or customers do not like those traits either. The employers then offer lower wages (or no job at all) to women who seem to possess the undesirable personality traits and higher wages and better posts to women who they think will be more popular. We call that “the nice girl hypothesis,” presuming that employers might wish to hire “nice girls.”
As an employer, you can take advantage of other people’s prescriptive stereotypes. If a woman, or for that matter a man, has a personality trait that the marketplace finds not entirely desirable, a potential arbitrage opportunity, as well as a chance to undermine stereotypes, arises from hiring women with those traits. To be sure, some of the less-desired traits in question may actually hamper job performance. For instance, a male salesperson may be more persuasive in some contexts through his greater exercise of authority, even if unjustly so, due to the prejudices of his audiences. Still, in other cases where the market is less taken with the personality traits of women, they still can do just as effective a job. If you can view the matter more objectively than the rest of the market, you can make some relatively good hires.
There is yet another reason to give many supposedly “non–nice girls” an extra consideration. Even if the more aggressive women will alienate your customers or their co-workers, you don’t have to let them alienate you. The market as a whole is penalizing the more aggressive women by measuring their impact on customers plus co-workers plus the boss or hiring authority (you). At the very least you can become more detached and eliminate that source of bias. It then remains the case that you should pay more attention to the talented women whose personalities do not always fit perfectly what the market currently favors.
There is pretty clear evidence, coming from a variety of quarters, that bosses dislike some personality traits in women for reasons unrelated to job performance, and some of that research gives clues to the possible mechanisms at work. We don’t regard any of these pieces of research as bulletproof, but in combination they do seem to suggest some common patterns of differential treatment for women.
In one study, by economist Martin Abel, 2,700 people are hired to do transcribing work. Associated with the task is a fictitious manager who offers what is in essence fictitious feedback. In this setting, if the (fictitious) boss criticizes the worker, the worker’s job satisfaction falls, and furthermore, the worker starts assigning less importance to the job task. Criticism from the boss isn’t enjoyed, which is hardly big news. But here is the striking result: these effects are twice as large when the negative messages are perceived as coming from female bosses compared to male bosses. Since this is online communication, it cannot be that these (fictitious) bosses are behaving any differently. Rather, it seems that many people have a harder time being criticized by an entity they perceive to be a woman.8
The literature on female voices, and how female voices are perceived, provides further evidence that some kinds of reactions to women are often more negative than reactions to men. In general, deeper voices are perceived as more authoritative, and that makes it harder for women to exercise authority with the voices they usually have, at least not without appearing bossy or strident. Women have to work especially hard in this regard; for instance, Margaret Thatcher hired a speech coach to, among other things, dramatically lower the pitch of her voice. More generally, it is striking that the pitch of women’s voices has dropped significantly over the course of the postwar era. Women’s voices used to register a full octave higher than men’s voices, but now they register only two-thirds of an octave higher on average. That suggests women are trying to fit into a more “managerial” set of social roles, but it is not always easy for them, in part due to voice bias.9
Finally, consider the evidence, cited earlier, that both neuroticism and agreeableness affect the wages of women negatively. This is consistent with the feminist critique that women are supposed to be tough but not too tough, firm but not obnoxious, like men but not too much like men, and that they are asked to walk an almost impossible middle line in the workplace (and perhaps when running for political office too). It is less consistent with the view that difficult women are simply troublesome to deal with, and troublesome employees are costly for the company. If that were the dominant effect, one might expect agreeable women to command a premium in the workplace, yet that is not the case. So again, be open to the view that the “difficult” women may be undervalued in the marketplace. That may mean women are forced into a somewhat narrower set of acceptable workplace roles, and you can do better by cultivating a broader conception of how they might fit in. Think a bit less in terms of “women taking care of problems in the workplace” and more in terms of “women as innovators.”
Once you get past anecdotal observation, we find in the research literature several (interrelated) major gender differences that consistently survive replication challenges, including in papers authored or co-authored by women. These results also hold up on multiple methods, including real-world data from the field and also from lab experiments.10 These differences are the following:
Let’s start with the evidence that, perhaps in response to the labor market penalties suffered by aggressive women, women as a whole do not self-promote as aggressively as men do. One carefully done study recruited nine hundred workers on Amazon Mechanical Turk, both men and women. The men and women were given tasks with possible bonus payments. They also were asked to assess their own performance on the task. When the metric was “I performed well on the test,” with a scale of 1 to 100, women reported an average of 46. The men on average reported an average of 61 on the same scale, even though the actual performance of the men was no better than that of the women. That is a remarkably large difference in self-assessment. Furthermore, note this gap persisted even when both the men and the women had perfect information about their actual performance on the task.11
Another recent study looks at written proposals submitted to the Gates Foundation and finds a different kind of evidence for women being more hesitant. It turns out that women and men have different communication styles: women are more likely to use narrow, highly specific words, and men are more likely to use broader, bigger-picture words. (Are the men “more conceptual” or “blowhards”? Depends!) And reviewers, it turns out, favor broad words, which are more commonly associated with more sweeping claims, and disfavor the use of too many narrow words. It is important to note that the research associated with the more narrowly worded proposals, once created, did just as well in the intellectual marketplace as the research associated with the broader, bolder claims; furthermore, ex ante applicant quality was also measured as equal. This seems a clear case of intellectual bias, one where the average style of the women had a harder time competing. So don’t be too worried if the women you are interviewing aren’t making such sweeping claims.
The net result is that “even in an anonymous review process, there is a robust negative relationship between female applicants and the scores assigned by reviewers.” This discrepancy persists even after controlling for subject matter and other variables. Notably, however, it disappears when controlling for different rhetorical styles. This is consistent with the view that men often do not “read” very well the different rhetorical, intellectual, and, yes, perhaps personality styles sometimes used by women.12
Perhaps most significantly, much of the gender gap in wages seems to be mediated by the personality factor of self-confidence. In a large variety of workplace settings, women, on average, are less self-confident than men and exhibit less self-confidence in public settings. Yet labor markets often reward confidence, sometimes even excess confidence. Some of what appears to be discrimination against women is actually discrimination against lower-confidence individuals, with the burden falling on women disproportionately. Note that the wage gap between men and women is highest at the higher levels of achievement, consistent with this hypothesis about confidence really mattering for wages, as presumably high confidence is demanded more at the higher job positions.13
Finally, there is some evidence that this confidence gap comes from very early in our lives, from high school or perhaps earlier yet. In one paper, looking at school grades 7–12, young women exposed to “high-achieving boys” perform worse academically and show lower confidence and aspirations. In contrast, they are helped by exposure to high-achieving girls. The young men, in contrast, are not affected by their exposure to either high-achieving boys or high-achieving girls. That too suggests that female discouragement is possible and occurs too frequently.14
That said, confidence gaps to some extent are self-fulfilling prophecies. As younger women see that there are relatively few female role models in a particular area, that may lower their confidence all the more. A self-perpetuating cycle is set in motion, and that cycle can be hard for a society to break.
All of these results point to the notion of a confidence gap as one of the main differences between men and women in the workplace, especially for higher-level jobs.
What does that mean concretely for an employer or talent seeker? First, for some jobs, such as asset trading, a lower level of confidence may be an advantage rather than a drawback. The less bullish presumably will be less willing to trade so frequently and less willing to take outrageously risky positions with their portfolios. For many jobs, including in politics, diplomacy, and prudential supervision, epistemic humility is more important than risk-taking. There is evidence from economics, for instance, that the gender confidence gap comes mainly from male economists making proclamations about areas they don’t know much about.15
Second, if you are thinking of hiring a woman who is genuinely self-confident, you may well be underestimating her abilities because you may not so readily perceive the strength and virtues of her confidence, given the gender stereotype of lower female confidence. So look extra hard for such candidates, because women who are more competitive and less risk-averse than their gender average might be neglected by the broader market. By appreciating the true and higher value of those female hires, you can gain from the world’s statistical discrimination and in the process rectify an injustice. Furthermore, note that the confidence gap accounts for more of the gender wage gap at relatively high levels of achievement. So this point is especially important for top jobs and much less significant for, say, simple service jobs or lower-level managerial tasks.
Third, very often jobs frame risk-taking and competitiveness in manners shaped by the male leaders of a company, or perhaps those framings are a holdover from earlier practice and leadership. More concretely, that means that many tasks within an organization may be framed more competitively than they need to be, or there may be a “rhetoric of risk” surrounding activities and tasks that are not really all that risky. (Ever try reading one of those breathless magazine articles about how everything in business is so dynamically changing everything all at once? It’s not actually true.)
In short, one way to mobilize female talent within your company is to remove cultural barriers to the advancement of women. One study, for instance, found that it was possible to induce more women to compete simply with nudges and changing the choice of architecture to make competing in a particular situation “opt out” rather than forcing people to “opt in.” In this constructed experiment, women were induced to apply for more potential promotions and experienced no adverse consequences regarding either their performance or their reported well-being. That is hardly a definitive result about real-world institutions, but it does communicate the basic point that if you encounter a generalization that seems true to you, you’ve got a chance to do better and outperform the market, rather than accept a brief for fatalism.16
So far we’ve focused on women in typical workplaces, but a separate literature considers the role of women as inventors. The data on patenting also show that women could, if given better opportunities, contribute more to innovation than is currently the case. But this present shortfall also may have to do with the aforementioned confidence gap.
First, women patent less than do men. For instance, in the data from 1998 only 10.3 percent of U.S.–origin patents have one or more female inventors. European data from 2009 show that women are involved in 8.2 percent of patents, and in Austria and Germany the numbers are as low as 3.2 percent and 4.7 percent, respectively. You don’t have to think that patents are such a wonderful measure of inventiveness to realize that these numbers nonetheless are reflecting very real differences in what men and women are up to.
The common response to this reality is to suggest that there is a serious “pipeline problem”—that not enough women from a young age are encouraged to be engineers or to hold other comparable positions in the innovation pipeline. There is some truth to this charge, but the numbers support it less than we would have expected.
If you look more closely at the gender gap in patenting, it has multiple causes, but here is the surprising fact: “Only 7% of the gender gap is accounted for by the lower share of women with any science or engineering degree, while 78% of the gap is explained by lower female patenting among holders of a science or engineering degree.” (The remaining 15 percent of the gap comes from a lower rate of patenting among women who do not have a science or engineering degree.) The single biggest specific source of the patent gap is that women are underrepresented in patent-intensive fields of engineering such as electrical and mechanical engineering, and also underrepresented in development and design, which are the most patent-intensive job tasks. So practically speaking, to whatever extent the confidence gap might be closed, women might become more interested in entering the more patent-heavy fields.17
It seems highly unlikely to us that the distribution of talent across gender represents some kind of inviolate and immutable natural order. There is a feasible future where more women work in riskier, patent-intensive jobs, just as we have had ongoing shifts of women into many other professions. Furthermore, a better allocation of talent really could matter. According to the study just cited, if the gender imbalance could be fixed entirely, the rates of patenting and innovation would go up, and GDP per capita would rise by 2.7 percent. Attaining even a fraction of those gains would be a big deal in a $20 trillion economy. And they could be a big deal for a talent seeker.
Researchers also have studied women in the venture capital process. Here too the results point our attention toward some biases relevant to the confidence gap. One study, by Sabrina T. Howell and Ramana Nanda, found that exposure to venture capital judges following a VC competition increased the chance that male participants would go on to found a company. The chance that women in similar circumstances would go on to found a company was much smaller. A follow-up survey found that men attained this greater success in part because they were much more willing to proactively reach out to the VC judges compared to the women. That may indicate the men have greater confidence in their ideas—and also greater confidence in getting a fair hearing from the judges, fear of harassment being another factor in this equation.
More generally, confidence gaps will lead to network frictions— meaning that it is harder to build out a large, very effective, and very diverse set of contacts. The resulting weaker ties can harm entire groups and in turn create the conditions that make confidence gaps seem at least partially justified.18
It is important for employers—especially male employers—to understand where the gender confidence gap might come from, at least in part. Both anecdotal and statistical evidence indicate that women who negotiate for promotions are more likely to be seen as intimidating, bossy, or aggressive. It is harder for ambitious women to be seen as likable. Sometimes the problem is directly sexual in nature. A group of guys can hang out together at work, sometimes even engaging in rhetorical horseplay and some amount of sexual joking and innuendo. In more extreme (but not uncommon) cases, they may go out together and visit strip clubs, or maybe get drunk together. A woman just doesn’t fit into those groups in the same way. Furthermore, a woman who fully engages in off-work socializing runs the risk of being propositioned, and in extreme cases she can be assaulted. There is also the risk of coming under suspicion from the spouses of co-workers. It is a more difficult social and networking environment for women to negotiate. Mentoring can be psychologically more complex across the sexes, and these days, in an environment of #MeToo, many men are reluctant to mentor younger women in a close or intense manner. In such a world, many women are not entirely certain how they fit into the workplace picture.19
One recent study of 1,139 venture capital pitches from 2010 to 2019 applied machine learning techniques to categorize pitches in terms of their styles, which in turn was correlated against how venture capitalists received those pitches. We’ve already discussed some of the broader results in Chapter 5, but the gender results are interesting as well. When women are pitching as part of single-gender teams, the women are judged more stringently on pitch quality than are the men. That is consistent with the view that women have to walk a thinner tightrope when presenting themselves to the outside world. But the really striking result was this: when women pitched on mixed-gender teams, the quality of the women’s pitch didn’t really matter at all. It seems the potential investors paid attention only to what the men said.20
Finally, given all of these constraints, talent spotters should pay greater heed to women coming from nontraditional backgrounds and also to women who are late bloomers. Precisely because women are in some ways different from men, and many talent-spotting mechanisms are more geared toward males, it is easier for super-talented women to go unselected. It is also possible that many women, in the earlier stages of their careers, had negative experiences with sexual harassment, or had children, and then made their “comebacks” much later in their careers. For those and other reasons, talented women might take longer to find their true avocations.
Consider N. K. Jemisin, a black female science fiction author who has sold millions of copies and won Hugo and Nebula Awards. When she started out, she believed that a career in fantasy writing was closed to her due to her identity. Instead she pursued a graduate degree in psychology and ended up working as a career counselor in a college in Springfield, Massachusetts. Still, she kept on writing, often anonymously online. When she was thirty, she hit a wall: she was in debt, she disliked Boston, where she then lived, and she disliked her boyfriend. Only after she committed herself to writing for a living did things begin to turn around.21
For a more unusual example, consider Sister Wendy (Wendy Beckett), who in the 1990s wrote bestselling books on art history and hosted a BBC television documentary program. Almost single-handedly, she interested an entire generation in the Western art classics and was described by The New York Times as the “most famous art critic in the history of television.” Her career background was as a nun, and she appeared on TV in full habit and with her signature buck teeth. She was born in South Africa in 1930, spent much of her earlier career living under a religious code of silence, and later devoted herself to solitude and prayer. She suffered from periodic bouts of ill health and spent years translating medieval Latin scripts. That hardly seems like an auspicious beginning for a television celebrity. Yet one day she was talking about artworks in a museum, a film crew asked to videotape her, the clip came to the attention of a BBC producer, and the rest, as they say, is history, with the core of her artistic career starting after the age of sixty. Needless to say, no one else on TV or in the field of art history had the aesthetic or historical perspectives of Sister Wendy.22
So far we have focused on personality across genders, because we know personality effects are real and there is less of a case that intelligence differs between men and women. Still, there are some interesting results about intelligence and gender, and they are of direct relevance to the hiring process. Perhaps most importantly, it seems easier for many bosses and talent scouts to pick out the smarter men than it is for them to pick out the smarter women.
In one study, people who looked at photographs of men and women were, on average, better able to spot the men who measured as smarter in tests than the women who did so. Some people “look smart,” and even if that judgment is highly subjective, some of the time it is correct. The countenances of those men offer socially accessible cues about their intelligence, but the countenances of women do not, at least not on average. In other words, it is harder for people, including male bosses, to pick out the smart women just from their looks. That result is open to several interpretations, but one obvious possibility is that smart women fit into stereotypes less, and also that many people are more accustomed to learning how to spot intelligence in men rather than women, perhaps because smart women do not always have the same high social status as smart men.23
Another interesting result from the paper is that women are in general better at assessing the intelligence of both men and women. We are not sure why that might be, but it is one reason (among others) to ensure that enough women have feedback into your hiring process.
There are also people who look smart but aren’t—watch out for those candidates. One study, for instance, suggested that people give higher intelligence ratings to those who are smiling and to those who are wearing glasses, even when those traits show no particular correlation with actual intelligence. Those results are based on strangers rating 1,122 Facebook images, where the individuals who supplied the images had been given IQ tests. It is possible you have those biases too, so consider how you might limit them. Maybe give the frowning candidate with contact lenses a closer look, but at the very least don’t be too confident in your looks-based assessments of how smart other people really are. These particular results about glasses and smiling have not been established through repeated replication, but we agree with the general idea that our intuitions about and judgments of intelligence can be led astray, so don’t get too cocky.24
Anecdotally, we have found that men have a harder time judging the intelligence of women because women often present themselves as more agreeable in an interview setting than men do. The agreeableness may be pleasant to interact with, but it obscures critical judgment and smooths over the transmission of “data” about the intelligence of the interviewee. Many men, in particular, will incorrectly downgrade the intelligence of an especially agreeable woman. They may find her likable or “pretty smart,” but the men won’t be sufficiently open to the idea that perhaps she is very smart indeed. This is one bias that men (and many women) should try to avoid.
In essence, male judgment often goes astray when women are (a) quite agreeable or (b) not very agreeable. That’s a lot of cases! One of the virtues of Five Factor personality theory, for all its limitations, is simply that it gives you some categories to help you think through and overcome some of your possible biases.
If evaluations of women’s intelligence are “smoothed out” in this manner, as it seems they are, women may end up being favored for many midlevel jobs, especially those requiring high conscientiousness, because the woman seems like a safer choice. At the same time, it will be harder for a woman to demonstrate that she deserves to be considered for much higher positions. There might not be any bias on average, but still, many evaluators will find it harder to perceive and identify the highest reaches of the female talent distribution in a given endeavor. That result can hold even if male (and many female) evaluators are not prejudiced on average against women in the workplace.
On net, many of the biases against women in the workplace can be thought of in terms of this idea of smoothing. When it comes to personality, it seems that observers smooth out too little when forming their impressions of women, and instead embrace the exaggerated impressions. The “difficult women” are considered more difficult than they really are, and the “good girls” are favored and viewed as meek and cooperative more than is actually the case. When it comes to intelligence, likely we see the opposite, excess smoothing: really smart women are undervalued and the not-so-smart women are overvalued, with too many impressions bunched near the mean. So as a single, simple recommendation, you might try smoothing out your personality impressions of women more and your intelligence impressions less.
It is interesting that Y Combinator, a leading venture capital firm, always has at least one woman as part of a three-partner interview panel. Historically, that position was established by Jessica Livingston, one of the original four founders of YC. Jessica’s acumen is legendary within YC circles, as she has a powerfully accurate gut instinct for talent, and in particular for weeding out bad apples. Jessica has since stepped back somewhat, but the organization has realized that much of that special je ne sais quoi wasn’t unique to her. Women partners seemed better than men at detecting deceit or disingenuous founders. The addition of a woman also changes the conversational dynamic of the screening quorum’s post-interview discussion in subtle and profound ways. We are not sure why but find it interesting that one of the most successful and durable talent screeners in the world requires that women be part of the screening process.
Finally, in terms of general principles, what should you do when interviewing or otherwise evaluating individuals from another country, culture, or religious or linguistic background? How do you pull out the noise from the signal? Let’s consider those questions, but we will focus on race as the relevant application, most of all in an American setting, because that’s what we know best.
Being black in the United States is an increasingly diverse experience, in part because of significant immigration from Africa and the Caribbean and Latin America. The Washington, D.C., area, where Tyler lives, is now by far the world’s second-largest “Ethiopian city,” to provide a simple example. If Tyler encounters a black person in the course of daily life, the chance that person is an East African is fairly high. The Bay Area, where Daniel lives, also has a relatively high representation of East Africans. And within African American communities more narrowly defined, histories and experiences differ greatly. For instance, Clarksburg, Mississippi, is a very different environment from Los Angeles, and both are different in turn from Boston. In 2020 many Americans were surprised to learn that Minneapolis has such a serious problem with racism, although if you know the history of the city, you’ll be aware that racial inequities have been a major theme there for a long time. On top of all those diversities, black men and black women in America may face very different kinds of racial barriers.
Our first piece of advice—and we mean this for individuals of all races—is not to pretend that you understand race as an issue very well. Don’t approach the problem, and the issue of bias, with some pet theory about how the world works with respect to race, because the diversity of racial issues, problems, and biases likely will defeat your schema. Mostly, as an outsider, you want to shed many of your preconceptions, whether explicit or implicit ones, and open yourself up to the talent possibilities in minority communities, particularly communities you may have no connection to personally.
The interview setting is one straightforward way to see the relevance of racial issues. To provide an example, both Daniel and Tyler have noticed that interviewees from many foreign countries and cultures are much more polite—and also more distanced and formal—than are white Americans and most other white Anglo-American interviewees (Canadians, British, Kiwis, and so on). American blacks are often similarly polite and formal. These groups of interviewees are sometimes not sure what cultural rules they are operating under, or what kind of impression they are making, and so they respond with risk-averse strategies of politeness and formality. That eases communication in some regards, but it also makes it harder to understand them and to judge their talent strengths and weaknesses.
Quite simply, individuals from different cultures are harder to read. Furthermore, a basic question arises: If these individuals are behaving more politely, is that because their culture places greater value on politeness, and so maybe on the actual job they will actually be more polite? Or is it merely a temporary exigency, designed to deal with the unfamiliar situation of being interviewed by a person from a different culture? Or maybe it is a permanent exigency, to deal with what will continue to be an unfamiliar situation on the job? Very often you do not know the answer.
To the extent there is a cultural gap between whites and blacks (or other groups) in an interview setting, it is a common strategic response—on both sides—to take fewer chances. To be less natural. To tell fewer jokes. To reveal less about one’s personal life. And so on. It is thus harder to move into the highly productive conversational mode that we discussed in Chapter 2. The end result is that you—even if you have no prejudices in the narrow sense of that term—are less likely to see the true talent strengths of the people you are talking to.
You will note that this mirrors some of the problems faced by women, as we discussed earlier. Women often feel—correctly in many cases—that there are fewer personality roles they are allowed to comfortably fill in terms of asserting “bossiness,” dominance, or other qualities. Nor are they necessarily allowed to show a comparable degree of emotional weakness. And so they often (rationally) respond by putting themselves out there less, role-playing more, being blander and superficially more pleasant or perhaps super-formal, or even trying to hide behind particular styles of makeup and dress. Those information-obscuring behaviors are a reaction to workplace bias, even if no single person in a particular workplace has strongly sexist opinions about women. The response to broader social pressures is to choose and invest in particular modes of self-presentation to the world, and those modes cannot be changed at the drop of a hat in all circumstances, even when such changes would be advantageous for all parties involved.
Returning to race, when President Obama ran for office and served, he had the sense (probably correctly) that his options for displaying angry behavior were far more limited than white candidates and presidents have experienced. He always had to sound reasonable and to act calmly, in a way that never constrained his predecessor, George W. Bush, or the assortment of other politicians who use ranting and raving and outrage as rhetorical tools. For a black leader in America, those strategies are much harder to pull off without alienating or even frightening a significant share of the electorate. And so Obama remained famously cool. It is no accident that America’s first black president is a personality type to whom such a cool mode of presentation came fairly naturally anyway.
The idea that “all white people are racist” may be upsetting to some white people who make an earnest effort to oppose racism, but it is important to see the truth in it. It is not that all white people intend racist outcomes. But it is the case that in a society with some racism and with some very definite cultural differences, one group of people—the wealthier and more powerful majority—will systematically be unable to see a lot of the talents of the less wealthy minority, in this case blacks. That is an obstacle that many, many black talent candidates face, and which is very difficult for the majority group to understand and to emotionally internalize as being real.
In sum, even in the absence of outright prejudice, blacks and other minority groups can face very real obstacles in making their talents known.
So what then to do? We don’t have cure-all answers, but we would like to present a few steps you can take to better improve your perceptions of the talents of people outside of your immediate, closely connected groups. These are not silver bullets, but at the very least, marginal improvements are possible.
The first thing you can do is to understand this problem. Plant in your mind the thought that a significant subclass of potential workers goes around with many of their talents invisible or at least significantly harder to spot. You can and should believe that, no matter what your exact view on the degree of explicit prejudicial intent in the modern world. The number of people who believe and internalize that truth simply isn’t high enough, and you should make a point of being on the proper side of this ledger.
Again, this point applies to people of all races. No matter what your background, a significant portion of the hiring pool probably comes from people of races and backgrounds different from yours. This will be all the more true as “work from a distance” spreads and American companies continue to hire the best talent from all over the world, even if those individuals do not immigrate to the United States.
The next step is to adjust your behavior accordingly. Look harder for talent, and learn to look better, including across racial (and other) divides. We’re not saying this is always easy, but it is remarkable how many people still have not even arrived at the stage of having any consciousness of the underlying problem at all.
Toward these ends, one concrete step you can take is to put yourself in environments where other people do not perceive your talents very readily, not only to get a sense of what that is like, but also to make the idea emotionally more vivid to you. For example, if you go to Finland, don’t assume that everyone there hates you, is pissed off at you, or doesn’t want to speak with you. The prevailing norms there are for people to be more distant and taciturn. When Tyler visited Finland, he felt he was crude and loud much of the time, and he went out of his way to moderate his behavior so as not to stick out so much. Maybe the Finns could not spot his talents or his articulateness as a result. That is the kind of experience you are trying to have—to feel what it is like when the perception of your talents is blunted. In turn, you will have a better sense of the possibly hidden abilities of others, whether the barriers be those of race, culture, religion, gender, or whatever.
Trying to learn a foreign language—and getting far enough to actually communicate—will serve this same end, though it can be very costly. For a long time, you just won’t sound that smart or that clever. Tyler found it instructive, in his twenties, to live in Germany. At the time, he spoke good but not perfect German. From his demeanor and dress, he clearly was not an American serviceman, and furthermore, most servicemen don’t learn much German. As a result, many Germans assumed that he was Turkish or someone from some other area that often sent migrants to Germany, like the Balkans. Once he heard an angry “Get out of here, you Turk!” (in German) in response to one of his queries. That is, again, the kind of feeling you wish to experience and grow to understand, though we do not recommend having to stay in such a world all the time (to be clear, most Germans were very nice to Tyler during his year living there).
Try to see what it feels like to be the one assessed by a very different culture and not always assessed so generously. See how helpless or clueless you can at times feel. Without summoning up whatever external markers of status and wealth you might possess, try to ask someone in a (very) foreign culture to do you a significant favor. Try to measure whether the responses you get to this sort of test are different from what you might do at home when asked by a comparable stranger. Then emotionally internalize those lessons and recall them next time you interview people from very different racial or cultural backgrounds. Also keep in mind (depending on your circumstances) that you might always have the option of stepping out of such situations and returning to a relatively privileged life, but maybe the person you are interviewing does not have a comparable liberty.
Part of Daniel’s success at talent selection may spring from the dual backgrounds in his personal history. He was born and grew up in Israel, but his parents are American Jews, and Daniel has a closer connection to American culture than most Israelis typically would have. Still, he is an outsider. Living in America, he has the perspective of a Jewish person who grew up in Israel and who also had Arab and Christian friends. Living in Israel, however, he was at least partially American in outlook, due to his parents and cultural ties and his flawless English and accompanying American accent (you can notice that Daniel does not have the American accent or vernacular of any particular region—often a giveaway that someone grew up abroad). An Arab swimming instructor taught him much more than the backstroke: the real lesson from Mr. Amos was that the inspiration to excellence could be found with someone who had a very different outlook on life. There was also no single religious community where Daniel felt at home, and so his whole life he has been used to looking at things from the outside. And his native bilingualism reinforced the notion that there is always more than one way to express or frame a particular idea—a source of mental flexibility and a natural entry point into the notion of multiple perspectives.
Maybe you were not born into such an environment, but at the very least, you should consider traveling to countries that would count as unusual relative to your native culture, as that can help you unlock and better understand the wide range of cultural variations. So instead of taking your next vacation on the North Carolina shore, try India or Tanzania (conditions permitting). Don’t think that you now “get India,” however. You had only a short period of time there, you probably don’t speak the major native languages, and India itself is amazingly diverse, with multiple major religions and language groups, just as Tanzania is. What you can see, however, is just how far cultural variation can extend. How much your earlier presuppositions were based on contingent facts rather than human universals. How easily and how frequently you can be surprised by your interactions with people from other cultures.
Visiting Africa in particular may be useful for understanding race in the United States. It is not that any African countries are “like” the United States with respect to race, but rather you can learn from the contrast. In most parts of Africa (parts of southern Africa excepted), people do not grow up feeling race as an issue in the same way that an American black person might, in part because everyone around them is black. Often African migrants report that they first “learn race” only when they come to the United States. To spend time in an environment where most people are black but race is not the same kind of issue as it is back home in America can be remarkably instructive. And if you are white or Asian or Latino, you will be the one who sticks out and who is frequently conscious of that—another instructive way to learn something about what it is like to feel different all of the time.
More generally, making such trips will mean you will be caught less unaware when you are assessing and interviewing people from other cultures. Working off these principles, do send your kids to study abroad, or to live for a while in another country, if at all possible and affordable. If nothing else, you are contributing to their long-run success as a manager and talent selector.
It usually doesn’t suffice on its own, but reading books is another way to broaden your horizons. When it comes to reading on race issues, we do have some very specific pieces of advice (again, keep in mind this is written from the perspective of two white guys).
First, read autobiographies, as their first-person narrative gives you a direct pipeline into the thoughts, feelings, and talents of some people very different from you. American history is remarkable for having so many superb African American first-person narratives. You can start with Frederick Douglass, Booker T. Washington, Zora Neale Hurston, Malcolm X, and James Baldwin. It is better to start with works from the more distant past, as they are less likely to collide with your current political views, and thus, you can absorb the content with fewer distractions. Reading President Obama’s autobiography, in contrast, might be useful and instructive, but your opinion of it possibly will be too caught up in whatever you thought of Obama as president. So deliberately seek out a bit of distance, and history enables you to do that.
If you would like another concrete suggestion that goes beyond reading, visit a black church at least once, arguably more times if you find the experience to be meaningful (again, assuming public health conditions allow). It is one accessible way to see a very open side of black America, and you will be welcomed heartily.
If you wish to get more intellectually daring, here is another step you can take. Read or listen to some of the more radical takes on race, whatever might be outside your comfort zone. Recognize that you don’t have to agree with them. But try to spell out in your own mind why somebody might possibly believe and promote those claims. If necessary, write out, if only privately, what you think are the best arguments for those points of view. That is one good way of trying to put yourself in the minds of others. See if what you come up with is at all convincing—not in the sense that you have to agree with it, but whether you think you really have come up with the strongest possible version of those arguments. Would the people on the other side of the debate recognize this as a smart and good-faith attempt to represent their perspective? Keep on working at this until you sincerely believe you have done your best in building a plausible case for views you disagree with.
We are recommending this as a means to deal with racial issues, but the method is general. If you have trouble getting along with people of different political or religious views, or have trouble spotting their talents, try to talk or write out their views in what you consider to be the most persuasive form possible. If only for a very short period of time, you will feel yourself, at least intellectually, put into their shoes.
We don’t pretend to have covered any more than a sliver of the relevant issues when it comes to race and culture. At the very least, know what you don’t know. The next step is to realize there is a problem and that there also is something you can do about it by improving your talent selection abilities. We hope we have helped you to make a few steps in that direction.