Skip to main content

Film dialogue often reinforces stereotypes about race and gender, a machine learning study confirms

Film dialogue often reinforces stereotypes about race and gender, a machine learning study confirms

/

There’s a way to help fix the problem: put more women in writer’s rooms

Share this story

If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.

hidden-figures-movie

In news that should be no surprise to anyone who follows the film industry, a new study has found that men speak more than twice as much as women in film, and often, the dialogue that is given to women helps reinforce gender and racial stereotypes.

The new study comes from the Signal Analysis and Interpretation Lab at the University of Southern California’s Viterbi School of Engineering. Researchers examined more than 53,000 dialogue segments from 7,000 characters in 1,000 scripts, and used machine learning to study how characters spoke and interacted along racial, gender, and age lines. They also examined the makeup of the crews that produced the films, including the writers, directors, and casting agents.

The study had a couple of takeaways. First, male characters spoke far more than their female counterparts, with 37,000 dialogues vs. just 15,000, while women portrayed 2,000 characters, as opposed to 4,900 played by men. The team also found that of the thousand scripts they examined, male writers were seven times more likely to be involved in a project, male directors were 12 times more likely to direct a film, and men were three times more likely to produce a film than female producers. While the gender of casting directors is the exception to the pattern — women outnumbered their male counterparts in casting-director positions two to one — the report found that this had no impact on the gender of the characters they were casting.

Hollywood is currently contending with criticism of the discrepancies between men and women in the film industry, and recent studies have shown that this bias extends to the characters onscreen. Researchers found that between 2014 and 2016, women spoke 47 percent of the time in the top 100 grossing films. But this study didn’t look at the content of those lines. This new study does, and finds that women and minorities tended to use language that reinforced stereotypes.

According to The New York Times, the researchers found that dialogue from female characters “tended to be more positive, emotional and related to family values,” while the dialogue from their male counterparts was “closely linked to achievement,” and to death. At the same time, the study found that black characters swore more, Latino characters spoke more about sexuality, and older characters spoke more about religion.

Researchers also wanted to use machine learning to figure out how important characters are to the film, based on the relationships they have with their counterparts. Removing female characters from most films, they found, often didn’t alter the film’s plot significantly, except in horror films, which the researchers noted was because women tend to be “portrayed as victims.”

“Writers consciously or subconsciously agree to established norms about gender that are built into their word choices,” says the study’s author, Anil Ramakrishna, who also suggests there could be a relatively easy way to fix the inequality: The authors found that when women were part of a writer’s room, “female representation on screen was on average 50 percent higher.”