What Data Analytics Says About Gender Inequality in the Workplace
It’s a cultural marker of sorts when a major Hollywood studio buys the film rights to a self-help book, in this case Sheryl Sandberg’s bestseller Lean In: Women, Work, and the Will to Lead. Sony Pictures has hired a screenwriter to fictionalize Sandberg’s tale of the lessons she learned adapting to the male-dominated world of Silicon Valley, where she’s risen to chief operating officer at Facebook. The central message is an uplifting one: A killer work ethic and unflagging drive can lead to success, even if men still set the rules of engagement in the corporate world.
A less empowering, but arguably more realistic, view comes from Anne-Marie Slaughter, an international lawyer, ex-State Department official, and former dean of Princeton University’s Woodrow Wilson School of Public and International Affairs. No stranger to navigating ultracompetitive work environments, Slaughter argues that women, even when they adhere to Sandberg’s advice, have a difficult time finding a balance between home and the office. The lucky few who scale the peaks of professional stature are wealthy, superhuman, or self-employed. The economic and cultural obstacles standing in the way of equal treatment for women in the workforce are formidable, in her view.
So who’s right? Both arguments are thought-provoking, yet they’re not based on hard data. The current debate over the career challenges facing women is understandably passionate and sometimes ideologically charged. Yet maybe it’s time for data analytics and rigorous social science to play a more prominent role in the conversation.
One data point everyone agrees on is that women and men face huge disparities at the office. In the U.S., female workers are still paid only 77¢ for every dollar their male colleagues make. A mere 4.2 percent of chief executive officers at Fortune 500 companies are women. To my mind, there are several possible explanations for this persistent divide.
The traditional theory about workplace inequality focuses on biology—childbearing, maternity leave, and child care hold women back. Then there are the deeply ingrained cognitive biases that rig the game in favor of men. One of Sandberg’s most interesting observations is that women and men work and collaborate differently, causing variations in career outcomes. These obstacles can be overcome, but doing so will require great effort.
To get a better read on the contrasting styles of men and women in the workplace, I’ve set out over the past year to analyze behavior and career outcomes at three different U.S. companies: a banking call center (more than 10,000 employees), an office products manufacturer (about 10,000), and a pharmaceutical company (about 1,000). A study by my data analytics firm, Sociometric Solutions, measured how people actually collaborate using sensor ID badges. The badges detect conversations and speech patterns using a combination of infrared, Bluetooth, and microphone data. (We ignored the actual content of conversations to protect privacy.) They also contain accelerometers, sensors that monitor physical movement. This data was collected on an opt-in basis and individual information was not shared. All of it was wirelessly transmitted to a base station and then on to our servers for processing.
The tags, worn for at least six weeks by each participant, tracked who interacted with whom. We also looked at e-mail, instant message, and phone call data to get a holistic view of how people were engaging with co-workers. Sociometrics is all about analyzing the patterns of relationships that connect people. In the workplace, interacting with the right people in the right way is vital. That’s why we measure social engagement, or how much people participate in a tight-knit group, and exploration, or how much people interact with different social groups. Admittedly, our three businesses do not constitute an exhaustive sample of U.S. workplaces. And the companies didn’t reveal compensation data about individual employees, though men do dominate their senior ranks. To my knowledge, however, this is the first attempt to analyze behavioral differences between men and women at different organizations using hard behavioral data.
At the call center, the one company where we have quantifiable productivity metrics, women were more productive than men, completing calls on average 24 seconds more quickly. Twenty-four seconds might not seem like much, but that adds up to a 9 percent difference in productivity. No differences in workplace performance or collaborative styles were observed at the company to support the idea that men and women perform or interact differently. Nonetheless, women were disadvantaged when it came to winning promotions and reaching the upper echelons of management.
At the pharmaceutical company my team researched, we were able to model how likely people were to get promoted based on their “exploration” scores. Once again, we found little evidence of contrasting work styles between men and women. When we examined the probability of promotion based on traits such as the ability to interact and lead other groups, there was again no significant difference between men and women. Well, to be precise, women were fractionally (0.2 percent) more likely to be promoted than men based on our model. Yet only 13 percent of top executives at the company are female despite a 50-50 gender split in the overall workforce. Lean In strategies may be great for a Harvard-educated high achiever like Sandberg; they don’t necessarily apply to all women.
What about the “maternity leave” theory? Various studies show that married and single women without children also lag behind male colleagues when it comes to pay and career advancement.
For working mothers, gender bias in hiring and recruitment exacerbates the special challenges they already face. In a 2007 study, Cornell University researchers submitted 1,276 fake résumés for real jobs listed in the classified section of a local newspaper. The résumés were equivalent when it came to educational credentials and work experience, but they varied in personal details about gender and whether or not the candidate had children.
The faux male candidates with kids were the most hirable, according to the study. Next came men and women without kids. The least desirable were women with children. Among job interviewers in the study, women were consciously (or subconsciously) punished for having a family. Subjects told researchers they viewed women as more likely than men to sacrifice work duties for family commitments. At the same time, male candidates with kids were viewed as more responsible and hence more desirable job candidates.
None of this means that Sandberg isn’t right about “leaning in” or that more equitable parental leave policies shouldn’t be encouraged. They’re clearly important to continuing the advancement of women in the workplace. But the cognitive bias is the far greater challenge. Attitudes hard-wired into the minds of men (and women) are very difficult to change. Characteristics admired in alpha male executives—boldness, decisiveness, and intensity—aren’t always valued in female ones. Conflicting attitudes about the primacy of a mother’s role in raising children complicate a woman’s career in ways most men need not worry about.
Business leaders can’t wave a magic wand and suddenly reduce societal bias. Subtle cues such as gender roles in TV shows and films will need to change. Even the language we use makes a difference. Top-Toy, the Swedish licensee of the Toys “R” Us brand, has released gender-neutral toy catalogs including such stereotype-busting examples as boys playing with ironing boards and girls having fun with gigantic Nerf guns. The state of Washington recently started rewriting laws using gender-free vocabulary, replacing “fisherman,” for example, with “fisher.”
Such efforts might seem like amusing cases of political correctness run amok or government overreach. But here’s something not so amusing: Do we really want to tolerate a business culture in which men get bonus points for being fathers and women are penalized for being mothers? Eliminating gender inequality in the workplace will require a sustained and generational effort, not unlike other civil rights movements in U.S. history. It’s all in the data.