Gender Bias in Tech Media

Jeff Carpenter
4 min readDec 7, 2016

--

Thousands of people and groups are working tirelessly to improve the diversity problem in tech, including mentoring organizations like Girl Develop It, Black Girls Code, Code 2040, leadership organizations like Project Include, and the tech companies themselves.

However, so far nobody has taken a broad look at how tech media is representing our industry to the outside world with respect to gender diversity.

One of journalism’s main goals is to report an unbiased view of reality. In an ideal world, tech journalism would mirror the gender diversity of the tech industry fairly equally — e.g. if the tech industry were 50/50 male/female, tech journalism would write articles about both equally — however, this is not the case today. Gender representation in tech journalism skews even more male than representation in the industry itself.

I want to make it clear that I’m not trying to pick on tech media. Tech journalism is responsible for a lot of good in our industry. For the 4 years I was out in the middle of the woods in Maine in college I read tech media voraciously — it was my only means of keeping up with the industry I loved and wanted to join one day. I want to ensure 100% of people can feel as connected as I did.

Pronoun distribution

I took every article that top tech news websites have published up to October 2016 and analyzed the number of male vs. female gender pronouns (he, him, his vs. she, her, hers). This will give us a rough look at how frequently each gender is talked about. One downside of this approach is that it doesn’t accurately reflect the reality of the gender spectrum today — however, I think it’s a good start. After analyzing 3.9 million pages, here are my findings.

To start off with some background information, here are the latest gender diversity stats for the biggest tech companies:

Figure 1: Men vs. Women at major tech companies

Average percentage of women: 34%.

Now for the pronouns. The top 6 tech media sites ranked according to Alexa are TechCrunch, PCMag, Techradar, Wired, PCWorld, and Ars Technica. Here is the distribution of gender pronouns across all articles per publication:

Figure 2: Gender pronouns in the top 6 tech news sites

Average percentage of women: 17%.

Engadget, Gizmodo, Mashable, and The Next Web are also popular tech media sites — let’s take a look at them:

Figure 3: Gender pronouns in other popular tech news sites

Average percentage of women: 20.1%.

From here you might wonder how this distribution compares to non-tech media like The New York Times or the Wall Street Journal:

Figure 4: Gender pronouns in general news sites

Average percentage of women: 31.5%.

And how about more female-oriented media like Elle or Vogue?

Figure 5: Gender pronouns in female-oriented publications

Average percentage of women: 60%.

The average female pronoun distribution on all tech media sites shown here is 18.2%. Tech media sites skew 15.8% more male than the number of women in the industry, 13.3% more male than The New York Times and The Wall Street Journal, and 41.8% more male than the female-oriented media sites Elle and Vogue.

You can’t improve what you don’t measure

Here’s why I think measuring and improving this bias is important:

  • Tech media has a vast reach. When women STEM majors look through the lens of tech media, what they see will frame their career choices.
  • Gender bias in tech media deprives women of role models.
  • It reinforces bias. It leads us to think that women are more rare than they actually are.

The goal tech media should have is to at least mirror the diversity of the industry, and best-case scenario take the responsibility to push the needle in the positive direction. But we can’t improve what we don’t measure. Going forward in 2017 I will release a quarterly report to look at these numbers across the above publications. Sign up to get updates when they’re released.

Thank you to Jonathan Kalin, Elva Fan, and Jill Carpenter for their ideas and feedback, and thanks to the folks at Common Crawl for providing the dataset.

--

--