Remember that choice advice from Microsoft CEO Satya Nadella when, at a celebration for women in computing, he said that they "should have faith that the system will give you the right raise," rather than be so bold as to ask for one? That if they only practised the "super power" of silent patience, "karma" would get them a fat paycheque – and, presumably, the corner office that comes with it?
Boy, did he get told. In online posts and essays, the criticism was sharp, prompting Nadella to embark on a whirlwind apology tour. It was a boneheaded comment, especially given new statistics showing that all the major tech and social-media companies, including Microsoft, are at least 80-per-cent male when you don't count administrative or support jobs, that many of the women who take software positions in the industry soon split for more welcoming jobs and fair pay, that their tales of workplace sexism are all over Twitter. (What complainers – didn't Apple and Facebook just offer to pay to freeze their female employees' eggs?)
Let's be generous: Maybe Nadella didn't realize what he was saying. Maybe he was suffering from a case of what psychologists call "unconscious bias," the instinctive blind spot created by his own experience. He'd never had to ask for a raise, he later explained. (Considering he earned $84-million in compensation in 2014, that's spectacular karma.) It didn't occur to him that women might need more than faith and talent to get one. "My advice," he conceded in an internal company memo, "underestimated [the] exclusion and bias that can hold people back."
Not just people – but profits, creativity and smart product development. Once the domain of psychology labs, the economic case for eliminating "unconscious bias" from the work force has made it the new buzzword for corporations trying to improve diversity and woo the best talent. At Google, 26,000 employees have taken the company's "unconscious bias" workshop, developed last year – prompting, for example, the should-have-been-obvious realization that the rooms in one building were named after almost entirely male scientists. Another team also realized that a perplexing software flaw in a new video-downloading program was the result of not considering left-handed users in the design. Firms such as Ernst & Young and Deloitte are developing research-based criteria to remove gut feelings from promotion decisions and performance reviews.
Fiona Macfarlane, a managing partner and the chief inclusiveness officer at Ernst & Young, says tackling unconscious bias isn't just important for landing the best people in a competitive global market. It also recognizes that, in a complex, knowledge-driven economy, there are no longer "solutions to problems, there are perspectives on dilemmas." She cites research showing that diverse teams with people from a variety of backgrounds are the most innovative.
"Talent that is more engaged works better – they give you more, they work harder, they are more creative. And they feel engaged when they belong," says Shannon MacDonald, the chief inclusion officer at Deloitte. "By calling it a gender issue, you are absolutely limiting the potential. This is about understanding where people are coming from."
We're all biased one way or another, and often we don't even realize it. You can test your own bias here. In 1995, a pair of psychologists, Anthony Greenwald and Mahzarin Banaji, defined the term "implicit stereotype," to describe how we unconsciously attach characteristics to people from a certain social group, using random bits of acquired information, influenced by our culture, upbringing and previous experience. This creates a bias – even one we might consciously and overtly reject. These "mind bugs," as Banaji and Greenwald put it in their recent book Blindspot: Hidden Biases of Good People, may reveal themselves in subtle microbehaviour, such as avoiding eye contact during a handshake with someone from a different race. Yet of the thousands of people who have taken the Implicit Association Test, an online word game created by Banaji and Greenwald to test the prevalence of stereotypes in areas such as race, appearance and gender, roughly 80 per cent revealed a bias.
The risk of bias is now taught to doctors, who can miss the right diagnosis because of false assumptions – the oversharing, needy patient who is really sick, the high-stressed working mom whose anxiety symptoms are really a heart attack. Given that most human beings have an affinity for people like them, unconscious bias is also one convincing explanation for why white, male executives still sit beside other mostly white, male executives in the corporate boardroom.
"It is very hard for you and me to make absolute judgments about anything. Really what we are trying to do is get our biased minds to get things right," says Iris Bohnet, professor of public policy at the Harvard Kennedy School.
The practice of unconscious bias may be subtle: shuffling a résumé to the bottom of the pile, or including a passing comment on a performance review. Research, for instance, has shown that in traditionally male fields, résumés bearing female names receive fewer callbacks, or are offered lower starting salaries than men – even if the woman's qualifications are better. The same bias has been found for résumés with black-sounding or Muslim names.
A recent analysis of performance reviews published in Fortune magazine found that women were more likely to be critiqued on personality traits, men on their work. Says Margaret Regan, president of the FutureWork Institute, which offers corporate training on unconscious bias: "A woman is described as edgy, a man as rough around the edges." The latter is fixable, the former deemed a character flaw, even though, says Regan, "they are really talking about the same behaviour."
Tall men also get an edge: While roughly 15 per cent of American men are over six feet, almost 60 per cent of CEOs surpass this height. This bias creates a self-perpetuating advantage: Tall boys are seen as potential leaders early, receive more opportunities to learn leadership skills as they grow up, and transition more easily into the role as adults. The same pattern has hindered female prospects in math and science careers.
But how do we tackle instinctive beliefs that people don't even know they have?
"Male executives will genuinely tell me they don't know why their company doesn't have more women leaders," Regan says.
Unconscious bias training, which Regan says has become one of the most in-demand diversity workshops, is designed to help leaders answer that question – by learning the science behind bias, testing their own unconscious judgments and group discussion.
Erick Vandeweghe, a partner at Deloitte in Toronto who participated in bias training, said he was surprised to learn that he unknowingly favoured men over women for analytical tasks. "My first reaction was one of defensiveness – I don't do that," he recalls, admitting that the process had made him more aware of how he selects teams and makes training decisions. "It's a journey to make sure I am more mindful." He says he also became more attuned to bias cropping up at meetings to discuss employee performance. For example, he says, he saw that when partners discussed an employee's level of "presence" with clients they often favoured men, while failing to consider how a woman – or even an introvert – might exhibit different qualities in presentations, but be equally effective.
A smart way to work around ingrained bias, Bohnet's research suggests, is to create relatively simple "nudges" – using strict criteria for evaluations, comparing résumés in a group so that skills nudge out first impressions. In one famous nudge, orchestras began installing screens during auditions so judges could only hear and not see performers, a move credited, in a U.S. study, with increasing the chance a female musician would make it past the first round by 50 per cent.
To prompt managers to be objective when assessing employees, Google has developed a bias checklist. Companies are also using data to weed out biased decisions. Ernst &Young tracks team assignments to ensure all staff get the diverse experience they need for promotions. Deloitte is developing inclusion measurements to evaluate leaders.
As Microsoft's Nadella explained in his memo to staff, his "generic advice was plain wrong." But his mistake was a human one. It's safe to say he's had his consciousness raised. With hope, the salaries and leadership stats for women in tech will be next.
How the science of bias can bring diversity to the workplace
A control group was allowed to choose a snack each day for a month, while another group was asked to choose all 30 snacks for the month. The first group went back to their favourites, over and over again. The second chose a variety of snacks. "Even though you might have a preference for a particular Snickers bar, you can't imagine you would want to have this 30 times in a row," says Dr. Iris Bohnet, a professor of public policy at the Harvard Kennedy School, who studies unconscious bias. "But every morning, you are tempted to go with your favourite."
Bohnet and her colleagues used this principle to eliminate unconscious bias in hiring and annual reviews. By comparing résumés together – instead of presenting them one by one – participants chose more diverse teams, or made decisions between a man and women that were based on performance and not gender stereotypes. Even more helpful, says Bohnet, is when companies create clear criteria for promotions and hiring, requiring managers to follow a checklist, and not rely on factually dubious impressions. "Most of us have a strong bias toward interviews – 'If only I could speak to the person, than I would have a gut feeling for how they do the job.' The correlation between those gut feelings and future performance is non-existent."
A real-life example
In 2003, when Bob Elton took over BC Hydro, he used a similar approach to change how people were promoted. Instead of promoting one position at a time, the executive team looked at the entire pool of employees to find those with strong leadership skills. They set aside the requirement for technical skills, and focused on a specific checklist, including "leadership values," such as team-building, communication. Candidates were approached rather than expected to apply for each job. In one case, Elton recalls, he hired a talented female employee who hadn't considered a manager's role because she only worked three days. By the time he left the post, women held 42 per cent of the top 100 management roles at BC Hydro. Typically, he says, at a utilities company, men would hold as many as 80 per cent of those positions. But Elton never called it a gender-equity program. The natural consequence of removing traditional assumptions about leaders, he says, was that more women were promoted.
Different examples of bias
Human beings have a natural preference for people who look like them, come from similar backgrounds, education. This is a problem in interviews, in particular when, with little else to go on, people might rely on similarities, or the fact that their alma mater is on the résumé, when choosing candidates.
The tendency to judge people based on recent events, rather than long-term records. In the same way, people are more likely to choose the first option when presented with two choices, even if the second option is better.
People tend to seek out information that confirms what they already believe, even unconsciously. For instance, a man might be forgiven for making a math "mistake." A woman might be judged, in the same situation, as "weak in math."
Human beings are more likely to recall a negative experience – with a colleague, say – than a number of positive ones.
Relying too much on intuition. This is a particular risk among doctors, and other experts in the field, prompting them not to ask the right questions.