The problem with almost all movies

By Ana Swanson

By now, most people have heard of “the Bechdel test.” To pass this famous three-part test, which measures whether female characters in a film are anything more than superficial, a movie has to (1) have at least two female characters (2) who talk to each other (3) about something other than a man.

It seems like a pretty low bar, but at least 40 percent of films fail, according to BechdelTest.com, a site that crowdsources these test results. “Birdman” fails. The "Lord of the Rings” movies all fail. “Harry Potter and the Goblet of Fire” fails. Even “Toy Story” fails. And it’s hard to think of a movie that doesn't pass the Reverse Bechdel test -- where two male characters don’t talk to each other about something other than a woman (according to the IMDB universe, some do exist).

The Bechdel test has its critics. Some films with prominent female roles, like Sandra Bullock in “Gravity,” don’t pass the test, while other films that are male dominated or sexist do. But, as Walter Hickey wrote for FiveThirtyEight.com in 2014, for a long time the crowd-sourced information on the Bechdel test was the best data on gender equity in film that we had.

Two years later, we're amassing more data that gives a clearer look at the real role of women in film. In a new project, Hanah Anderson and Matt Daniels at Polygraph analyzed screenplays for 2,000 popular movies, and broke down the number of words spoken by male and female characters -- “arguably the largest undertaking of script analysis, ever,” they say.

To do the analysis, Anderson and Daniels mapped characters who have at least 100 words of dialogue in a movie from the screenplay to the actor's IMDB page. (They caution that their analysis is based on the screenplay, not the actual film – sometimes directors may cut or change lines from the screenplay, and or even change the name or sex of the characters. So there may be small variations between the screenplay and the final film, though the results should be fairly accurate.)

What they find isn’t that pretty. As the Bechdel test suggested, most popular movie scripts are dominated by male characters.

The graphic below shows their analysis for 2,005 screenplays. The dots on the left, in blue, represent films where more than half of the dialogue is spoken by a man. The dots on the right, in red, represent films where more than half of the words are spoken by women. As you can see, there are few films where even half of the dialogue is spoken by women, and quite a few films where 100 percent of the lines are male. (The original graphic is interactive, so you can click on the dot so see the name of the film.)

Hanah Anderson and Matt Daniels, Polygraph

Overall, a woman ended up having the most dialogue (i.e., being the lead) in 22 percent of the films they looked at. In about a third of the films, a woman was in second-place in terms of having the most dialogue. And only in about 18 percent of films did women occupy at least two of the top three speaking roles. In contrast, men occupied at least two of the top three speaking roles in 82 percent of films.

Men consistently have more lines than women across different genres of film. Perhaps predictably, men dominate action movies:

Hanah Anderson and Matt Daniels, Polygraph

But men also dominate comedies:

Hanah Anderson and Matt Daniels, Polygraph

And dramas:

Hanah Anderson and Matt Daniels, Polygraph

And even horror films. In case you're wondering, the all-female horror film is 2005's "The Descent."

Hanah Anderson and Matt Daniels, Polygraph

Another major critique of Hollywood's gender issues is that women can only obtain big roles when they're young -- which is why movies feature so many female scientists, astronauts and criminal defense lawyers who are somehow all in their 20s. While men can get leading roles into their 40s and 50s, women disappear from the screen.

Anderson and Daniels's data confirms this. The graphic below shows the percentage of total words spoken by male and female actors, broken down by age. For women, 22-to-31-year-old actresses have the most lines, speaking 20 million words; for men, 42-to-65-year-old actors speak the most, with nearly three times as many words as the most outspoken group of women.

Hanah Anderson and Matt Daniels, Polygraph

The trouble with princesses

The project was inspired in part by an academic study that my colleague Jeff Guo wrote about in January, in which linguists analyzed speaking roles in Disney films. That research showed that men spoke more often than woman in most of Disney’s princess films, like “The Little Mermaid” and “Mulan.”

That might not seem like such a big deal, but the researchers pointed out that little boys and girls watch these movies on constant repeat, and learn a lot from them. And one of the things that kids might take away is that most of the people doing anything useful or entertaining in these films are male.“There are no women leading the townspeople to go against the Beast, no women bonding in the tavern together singing drinking songs, women giving each other directions, or women inventing things,” one of the researchers of that earlier study said. “Everybody who’s doing anything else, other than finding a husband in the movie, pretty much, is a male.”Daniels and Anderson also looked at Disney movies, doubling the sample size from the original experiment. They found a similar result: In 22 of the 30 Disney films they looked at (the films listed on the left in the graphic below) men had most of the lines. In about five films (shown in the middle) the dialogue was relatively balanced, and only in the four films on the right did female characters have more than 60 percent of the lines.As the Disney princess movies suggest, one of the most surprising results of this analysis is that men have more lines than women even in movies that are about women or made primarily for a female audience -- like “Pretty Woman” and “10 Things I Hate About You.” Even in the romantic comedies in their data set, men had on average 58 percent of the dialogue, Anderson and Daniels said.