This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: Female protagonists in animated films have come a long way over the past several decades. Take the 1959 movie "Sleeping Beauty": Soft-spoken and vulnerable, Princess Aurora is cursed as an infant by the evil fairy Maleficent and ultimately saved not by her own wits, but by a prince's kiss.
Flash forward to 2016's "Moana," whose bold teenage heroine belts out songs as she embarks on a daring voyage across the ocean. She makes her own choices, takes risks, gets knocked down, then pulls herself back up. Yet a stark gender imbalance remains, a new University of California, Berkeley study has found—just not where you might think.
While female human characters receive close to half of the total screen time in animated films, only about 25% of screen time among nonhumans—animals, monsters, robots, magical creatures and other similar beings—goes to female characters. Notably, "Moana" and its sequel give female nonhuman characters no screen time at all. "We knew from prior research that there was a gender disparity in live-action films—male characters generally have three times as much screen time as female characters—but we were surprised that there's an even stronger male skew when looking at the screen time of nonhuman characters in animated films," said David Bamman, a Berkeley associate professor in the School of Information and lead author of the study, "Measuring Gender Representation in Animated Films." For the study, the researchers analyzed 224 animated films that ranked among the top 50 movies at the box office each year from 1937 to 2025, with a particular focus on films released in 1980 and after, when the number of animated films began to take off.
To identify and measure nonhuman characters, they trained computational models to recognize characters on screen, such as Olaf in "Frozen" or Judy Hopps in "Zootopia," and then measured how much screen time each received. The team manually checked the models' work in 100 of those films to ensure they accurately recognized which characters were on screen. Bamman co-authored the study with Allison Cooper, a cinema studies scholar at Bowdoin College, and Berkeley students Reina Kushihashi, Ruby Alvarez Rubio and Madison Mar, who graduated from Berkeley in May.
The research is published on the arXiv preprint server. In the gender analysis, nonhuman characters had to have a speaking part played by a voice actor. Because animated characters do not self-identify, the researchers' gender classifications are necessarily interpretive: They relied on voice-actor gender and fan-community descriptions, including pronoun use, and manually reviewed cases where the evidence was unclear or the sources disagreed.
Using computational models, Bamman said, allowed the researchers to analyze films at a scale that would be impossible to do manually. "We develop computational models with a goal of reproducing what a human sees—when they're watching a movie, for example—to measure some phenomena at a broader scale than what an individual reader or viewer could read or view in a lifetime," he said. Discover the latest in science, tech, and space with over 100,000 subscribers who rely on Phys.org for daily insights. d research that matter—daily or weekly.
Extract — continue reading at the source.
