Character network graph for "Teminator 2: Judgment Day." Source: "Linguistic Differences," Ramakrishna, A. et al, ACL 2017, Canada.An analysis of movie scripts conducted at USC found that stereotypes – about gender, race and age – are created and reinforced in films. While this news may not come as a shock, researchers at the USC Viterbi School of Engineering’s Signal Analysis and Interpretation Lab (SAIL) now have some hard statistics to back it up.

The SAIL study found that, in the nearly 1,000 scripts reviewed, men had over 37,000 dialogues while women had just over 15,000. Men also portrayed nearly 4,900 characters, compared to just over 2,000 portrayed by women. Regardless of race, females also tended to be about five years younger than their male counterparts. And in the majority of films, female roles were not central to the plot.

In addition, the analysis found 7 times as many male writers as female writers, almost 12 times as many male director, and a little over 3 times as many male producers.

To conduct their study, the researchers developed a tool to quantify the sophistication and the tone of language of characters and dialogues in scripts pulled from The Daily Script and IMSDb. The content of characters’ language – and their interactions across gender, race and age – were analyzed. Graph theory was used to determine how central characters are to a movie's plot by analyzing the ties and relationships to the other characters within the film. A character network graph and web of relationships, not unlike a transit hub, could then be examined.

The researchers also looked at the production teams involved in the making of the films. Interestingly, casting directors were an exception to the male-dominated trend – there were twice as many who were female. However, this did not seem to impact the gender of characters in the films. By contrast, female character representation was an average of 50 percent higher when female writers were involved.

“Computational language analysis and interaction modeling tools allow us to understand not just what someone says, but how they say it, how much they say, to whom they speak and in what context, thereby offering new insights into media content and its potential impact on people,” noted Shri Narayanan, senior author of the study.

Other findings from the study included:

• Male dialogue contained more swearing and words related to achievement and death; female characters tended to be more positive, using language connecting with family values (as mapped from a SAIL Lab tool called Emotiword).
• Latino and mixed-race characters had more dialogue related to sexuality.
• Compared to other races, African-American characters had a greater percentage of swear words in their dialogue.
• As characters age on screen, they appear more “sage-like”: intelligent, less excited, with less mention of sexuality and more talk of religion.

The researchers will have forthcoming work based on adaptations of Shakespearean novels – studying how characters traditionally played by males in theatrical productions are played by women in modern movie adaptations, and whether roles now played by females have changed or diminished in importance.

Click here to explore an interactive character network graph of selected films.