Actress Geena Davis, perhaps best known for her role in the feminist baseball film A League of Their Own, realized there was a problem while watching movies and TV with her then-two-year-old daughter. She wondered: Where were the female characters? Shouldn’t there be some research out there showing that entertainment for young children is bereft of female characters? And so in 2004, the Geena Davis Institute on Gender in Media was born.
The institute, which analyzes the portrayal of women in children’s media, just got a $1.2 million grant from Google’s Global Impact Awards–part of a larger announcement of $23 million in grants given to nonprofits changing the world with technology– to take its research to the next level by developing software that can do some of the less subjective analysis.
Today, the Geena Davis Institute produces enlightening studies like this one, an analysis led by University of Southern California professor Stacy Smith (a research partner) that analyzes gender roles and occupations of women in family films, prime-time programs, and cable. The study, which examined 11,927 speaking characters in hundreds of films and TV shows, came up with a number of startling statistics: 28.3% of women in the family films analyzed wore “sexy attire,” 14% of employed females in prime-time TV were shown in C-suite positions (compared to 86% of employed males), and just 28% of speaking characters in family films were women.
“There will always be subjectivity that will require humans, but if there’s quite a bit that can be handled by a software tool, we can do [analysis] quicker, cheaper, and deeper,” says Madeline Di Nonno, executive director of the institute. “If we have this great tool, we could we do real-time reporting the way Nielsen does.”
The software is still in the very early stages, but Googlers are helping out, surveying the landscape of open-source audio and video recognition tools to see what they can put together. The institute’s software efforts are so new that it isn’t quite sure what it will be able to measure, but Di Nonno tells us: “One thing we did talk about is screentime. How long is every character on screen? Say there’s a ratio of approximately three to one in terms of male characters to female characters, does that female character talk as much or less than the male characters? That type of minutiae we haven’t been able to capture.”
In the future, the institute could quickly run the numbers on, say, the recently announced Writers Guild Award nominees to see which TV shows are doing the best job of representing females. Or it could analyze a show’s representation of women right after airing. All of a sudden, a previously time-intensive process might be as simple as clicking a few buttons.
We know there is a problem with female representation in media–the institute’s extensive research makes that clear enough–but having a piece of software to run some of the numbers might make it easier to quantify the scale of the issue.
The institute hopes to have a fully baked piece of software ready in under two years.