Every time I write something about diversity in Silicon Valley, and the gross disservice that today’s tech giants are doing to women and people of color by consistently overlooking them for jobs and promotions, I get comments to this effect: What about the pipeline problem? Less than one-fifth of computer science graduates are women. It’s not Google’s fault it can’t find qualified women to hire.
To which I say: Pleeeease.
So, okay, yes, colleges and universities need to do more to make computer science programs inclusive, but that doesn’t let off the hook the most powerful tech companies on the planet for saying one thing and doing another.
Either you want a diverse workforce, or you don’t. You can be someone who pays lip service to a problem; or you can be someone who tries to understand how you’re contributing to that problem, then take corrective action.
That goes for journalists, too. Which is why, two years ago, I set out to better understand gender representation in my own work. And it’s why, with the help of Nathan Matias, a Ph.D. student at the MIT Media Lab and a fellow at Harvard's Berkman Center, I did it again.
But before we get into what I found—and it wasn’t good—let me explain why I wanted to do this in the first place.
Male dominance in global media is well documented, and has been for many decades. Both in newsrooms and in news articles, men are leaders—they make more money, get more bylines, spend more time on-camera, and are quoted far more often than women—by a ratio of about 3:1. I notice male biases in journalism all the time. Which means I know that readers of The Atlantic do, too.
“When women do show up in the news, it is often as ‘eye candy,’ thus reinforcing women’s value as sources of visual pleasure rather than residing in the content of their views,” wrote a group of researchers from the University of Bristol and Cardiff University published earlier this month. They analyzed more than 2.3 million articles published by 950 news outlets over a six-month period, and published their findings in PLOS One. Men and boys were represented more often than women and girls in both images and text, they found.
“Moreover, the proportion of females was consistently higher in images than in text, for virtually all topics and news outlets,” the researchers wrote,“women were more likely to be represented visually than they were [to be] mentioned as a news actor or source.”
Organizations like Forbes and the BBC were particularly egregious, referring to men in their stories 81 percent of the time. Part of the problem, clearly, is that the people who are considered newsworthy—presidents, governors, military leaders, CEOs, and so on—are often men. “The media focuses nearly exclusively on the highest strata of occupational and social hierarchies, in which women’s representation has remained poor,” a group of researchers wrote in another study published last year by the American Sociological Review.
But here’s another way to look at it: Women represent about half the global population, and yet they’re dramatically underrepresented in stories meant to help people understand much of the complexity in the world.
“Inequality defines our media,” said Julie Burton, the president of the Women’s Media Center, in a statement last year. “Media tells us our roles in society—it tells us who we are and what we can be.”
In 40 years of studying this issue, researchers have found the same result over and over again, according to the authors of last year’s American Sociological Review study:
The findings of these studies are consistent: They all report substantial underrepresentation of female names, and they typically find that female names constitute approximately one fourth of all mentions.
That was what Matias and I found when we assessed a year’s worth of my work in 2013. Of the 2,075 people I mentioned in 136 articles over the course of a year, about 25 percent were women. (We counted not just the number of women I named, but the number of times I mentioned them.) That put me on par with mainstream news media, generally. Internationally, about 24 percent of news subjects were female in 2013, according to the Global Media Monitoring Project.
This time, I fared slightly worse.
Last year, Matias wrote a bit of software that scraped all of my articles published by The Atlantic in 2015 and identified proper names as male or female. We then manually combed through the list and cleaned up the data: omitting some names that didn’t need to be counted (like Loch Ness, for instance) and assigning gender in cases where the computer didn’t know what to pick.
We included 192 of the articles I wrote for The Atlantic last year. All in all, I mentioned 736 different people and only 165 of them were women—meaning women accounted for just over 22 percent of the unique individuals I named or quoted in my work last year. (If you look at total mentions instead of unique mentions, it’s even worse. Out of the 2,301 names that appeared in my work last year, 1,839 of them were men. That leaves just 428 mentions of women: under 19 percent of total mentions. This suggests, and we were later able to confirm, that even when I do mention women, I give men more space in my stories.)
A couple of caveats: One difference between this year’s analysis and the one we did in 2013 is that we included fictional characters in the overall count. There weren’t too many of these, but enough that it merits mention. (For instance, Luke Skywalker, Batman, Voldemort, Mario, Phoebe Buffay, Chandler Bing, Goldie Wilson, and Marty McFly all got counted.)
Also, because the vast majority of the people I quoted or named only showed up in one story, it was hard to identify clear patterns at first. To solve this problem, Matias did a simple logarithmic transformation so we could better understand how women were represented among names that appeared in multiple articles. You can see how doing so transforms the visualization of our findings below:
Turns out I didn’t mention any of the same women in more than two articles last year, whereas some of the same men were mentioned in as many as six stories.
These numbers are distressing, particularly because my beats cover areas where women are already outnumbered by men—robotics, artificial intelligence, archaeology, astronomy, etc. Which means that, by failing to quote or mention very many women, I’m one of the forces actively contributing to a world in which women’s skills and accomplishments are undermined or ignored, and women are excluded. (Assessing my work for racial diversity would be more complex, but, I suspect, similarly revealing and disheartening.)
Some people would argue that I’m simply reflecting reality in my work. That’s an overly generous interpretation. Another popular reaction is that my job as a journalist isn’t to actively seek out diverse sources, but to find the most qualified people to help me tell the best possible story. I only agree with that in part: Yes, my job is to serve readers by finding the best sources for my stories, but why assume that the best source isn’t a woman? By substantially underrepresenting an entire gender, I’m missing out on all kinds of viewpoints, ideas, and experiences that might otherwise sharpen and enhance my reporting.
I’m not excluding women on purpose, but I can’t say it’s an accident, either. Reporters choose whom to interview. We carefully parcel out our time as we work toward deadlines. I spent several weeks working on this story about self-driving cars, for instance, and it occurred to me as I was reporting that I hadn’t interviewed any women. In the end, deadline pressure and decisions about what to leave on the cutting-room floor trumped diversity.
It seems that happens a lot in my work. I asked Matias to run a follow-up analysis to see how many times I filed stories with no women mentioned. He looked at a slightly different dataset, expanded to include the 198 articles I wrote in all of 2015 and into part of January of this year. The result: Zero women mentioned in 119 of those stories, amounting to about 60 percent of my work. On top of that, I mentioned men more times than women in nearly three-quarters of my articles. Blergh.
Okay, so, what do I do? There are three steps, Matias suggests, that would make a difference. First, I could actively look for more stories about newsworthy women. “The key is that there are two major factors shaping who you mention: the people that your stories are about and the people who you rely on to make sense of those stories,” he said.
Second, I could try to be more inclusive in what he calls the “one-off” stories, the pieces where I interview or mention a person once, but don’t necessarily expect that person to be a fixture in my ongoing reporting. Third, and here’s the area where I think putting in the effort could most improve my work in the long run, is trying harder to cultivate more women sources on my dedicated beats.
When I think now, off the top of my head, of some of the experts I routinely turn to—for comments on net neutrality, or artificial intelligence, or natural-language processing, or self-driving cars, or digital preservation—the first person on my to-call list is almost always a man. I need to change that. (Matias is also working on building a browser plug-in that could help tally gender diversity in real time as a person’s writing—a tool that could be helpful for individual journalists like me, and also could be useful for editors and publishers who want to look more broadly at gender diversity in a publication’s larger body of work.)
I’m also going to put more of an effort into asking for help. For example, I’m working on a big series of stories about robots and artificial intelligence right now. I’ve done more than a dozen in-depth, fascinating interviews for my first story—and every single source so far has been a man. (Let me be explicit: Please help me. If you’re a roboticist and a woman, or a woman who writes science fiction about robots, or a robot who identifies as a woman—only half-kidding about that one—I would love to talk to you. My email address is email@example.com.)
I’m not sure what a reasonable goal is for the representation of women in my work. But I do know that the next time I do this kind of analysis, I need to see significant improvement. I also know that making that happen is ultimately up to me.