Google, like many tech companies, is a man’s world.
Men make up 83 percent of Google’s engineering employees and 79 percent of its managers. In a report to the Equal Employment Opportunity Commission last year, Google said that of its 36 executives and top-ranking managers, just three are women.
Google’s leaders say they are unhappy about the firm’s poor gender diversity, and about the severe underrepresentation of blacks and Hispanics among its work force.
And so they are undertaking a long-term effort to improve these numbers, the centerpiece of which is a series of workshops aimed at making Google’s culture more accepting of diversity.
There’s just one problem: The company has no solid evidence that the workshops, or many of its other efforts to improve diversity, are actually working.
Google is attacking the problem with its considerable resources and creativity. But it does not have a timeline for when the company’s work force might become representative of the population, or whether it will ever get there.
“I think it’s terrific that they’re doing this,” said Freada Kapor Klein, an entrepreneur who has long studied workplace diversity, and who is the co-chairwoman of the Kapor Center for Social Impact. “But it’s going to be important that Google not just give a lecture about the science, but that there be active strategies on how to mitigate bias. A one-shot intervention against a lifetime of biased messages is unlikely to be successful.”
Google says its plan isn’t one-shot. It points out that it has been trying to improve its diversity for years by sponsoring programs to increase the number of women and minorities who go into tech, and meticulously studying the way it hires people in an effort to reduce bias.
In May after pressure from civil rights leaders, the company published a report documenting the sex and race of its employees “to be candid about the issues,” Laszlo Bock, Google’s executive in charge of human resources, wrote at the time.
Google’s diversity training workshops, which began last year and which more than half of Google’s nearly 49,000 employees have attended, are based on an emerging field of research in social psychology known as unconscious bias. These are the hidden, reflexive preferences that shape most people’s worldviews, and that can profoundly affect how welcoming and open a workplace is to different people and ideas.
Google’s interest in hidden biases was sparked in 2012, when Mr. Bock read an article in The New York Times about a study that showed systematic discrimination against female applicants for scientific jobs in academia. The effect was so pervasive that researchers theorized the discrimination must be governed by unconscious cultural biases rather than overt sexism.
Mr. Bock wondered how such unconscious biases were playing out at Google. “This is a pretty genteel environment, and you don’t usually see outright manifestations of bias,” he said. “Occasionally you’ll have some idiot do something stupid and hurtful, and I like to fire those people.”
But Mr. Bock suspected that the more pernicious bias was most likely pervasive and hidden, a deep-set part of the culture rather than the work of a few loudmouth sexists.
Improving diversity wasn’t just a feel-good goal for Google. Citing research that shows diverse teams can be more creative than homogeneous ones, Mr. Bock argued that a diverse work force could be good for Google’s business. Could Google investigate how biases were affecting people’s work — and, more important, could it change its own culture?
The lecture begins with a dismal fact: Everyone is a little bit racist or sexist. If you think you’re immune, take the Implicit Association Test, which empirically measures people’s biases. Dr. Welle goes on to explain that some of the most damaging bias is unconscious; people do the worst stuff without meaning to, or even recognizing that they’re being influenced by their preferences.
The effect of bias is powerful, and it isn’t softened by Silicon Valley’s supposedly meritocratic culture. In the lecture, Dr. Welle shows a computer simulation of how a systematic 1 percent bias against women in performance evaluation scores can trickle up through the ranks, leading to a severe underrepresentation of women in management.
Finally, Dr. Welle points to research showing that we aren’t slaves to our hidden biases. The more we make ourselves aware of the role our unconscious plays in our decision-making, and the more we try to force others to confront their biases, the greater the chance we have of overcoming our hidden preferences.
Google offered several anecdotes that seem to indicate a less biased culture as a result of the training. Not long ago the company opened a new building, and someone spotted the fact that all the conference rooms were named after male scientists; in the past, that might have gone unmentioned, but this time the names were changed.
During one recent promotion meeting in which a group of male managers were deciding the fate of a female engineer, a senior manager who had been through the bias training cautioned his colleagues to remember that they were all men — and thus might not be able to fully appreciate the different roles women perform in engineering groups. “Just raising the awareness was enough for people to think about it,” Mr. Bock said. The woman was promoted.
Another time, in an all-company presentation, an interviewer asked a male and female manager who had recently begun sharing an office, “Which one of you does the dishes?” The strange, sexist undertone of the question was immediately seized upon by a senior executive in the crowd, who yelled, “Unconscious bias!”
Mr. Bock saw all of these actions as evidence that the training was working. “Suddenly you go from being completely oblivious to going, ‘Oh my god, it’s everywhere,’ ” he said.
But whether that will lead to a long-term change at Google and, in turn, the rest of the tech industry, remains an open question.