A new analysis of best picture nominees over the last seven years shows that racial and gender representation hasn’t changed all that much
In the six years since the #OscarsSoWhite movement highlighted the dearth of members of marginalized groups in Academy Award-nominated films, the diversity of both on-screen and behind-the-camera roles has barely improved.
Saying it plainly: The #OscarsAreStillSoWhite. They also remain predominantly male, with patterns of inequality both in front of the camera and behind the camera.
A new analysis of all the best picture nominees from 2014 through 2020, shows:
While there were notable increases in the diversity in acting roles in 2016 and 2018, the percentage of major roles played by actors of color in films nominated for best picture in 2019 was smaller than it was in 2014, the year that prompted the #OscarsSoWhite movement.
Any overall increase in representation in a “good” year is strongly affected by just two or three movies. For example, the strong numbers in 2018 were due to the presence of Black Panther, BlacKkKlansman and Roma among the best picture nominees. But every year there are movies nominated for best picture that have no Black or other actors of color – zero – in a notable role.
Black and brown actors only appear in significant numbers in films about the world of Black and brown characters.
Looking at behind-the-camera positions reveals similar results:
Racial and ethnic diversity in major roles such as director or cinematographer continues to be modest and any improvement in the overall numbers is largely due to two or three movies each year. Women are largely confined to the jobs of casting director and co-producer. Directors, cinematographers, writers and executive producers are overwhelmingly male.
I’m a faculty member at Yale University and a computational scientist interested in questions at the intersection between science and society. For several years, I have helped to design a data science-driven study of the Academy Awards related to diversity in Hollywood. This study examined whether patterns of representation have changed since the #OscarsSoWhite movement of 2015 and other messages that emerge from a nuanced analysis of diversity across ethnicity and gender.
Our analysis looked at best picture nominees released between 2014 and 2020 (seven years) for a total of 58 movies. When studying actors, we focused on the main cast, not everyone who appears on screen. The behind-the-scenes roles analyzed included: director, writer, casting director, cinematographer, editor and producers (including people identified as co-producer and executive producer).
Race, ethnicity and gender were determined by viewing photographs of the individuals on IMDb and other online resources. Three people viewed photos of every individual, and the perceived race/ethnicity was determined by majority judgment. (In a few cases, the person responsible for an off-screen job could not be determined and that position was not included in the calculations.)
Why did we use films nominated for the Academy Award for best picture? Why not study independent or smaller productions? For one, the Academy Awards was the basis for the original #OscarsSoWhite critique. Why examine only the best picture nominees and not all categories? Because this category is a proxy for the most recognizable, financially lucrative and/or prestigious films in the world. When examining issues of workplace diversity, these films are good test cases, because they are the arenas where the most visible are operating. The studios that produce best picture nominees dictate the direction of the industry, including who gets hired for leadership positions that dictate the shape of the entire industry.
These methods allow us to analyze broad trends while identifying nuances in the numbers. Notably, even “good news” is tempered by analysis of the overall picture. For example, several of the Oscar favorites in 2021 have predominantly Black casts. Judas and the Black Messiah is nominated for best picture and has two actors, Daniel Kaluuya and LaKeith Stanfield, nominated for best supporting actor. Ma Rainey’s Black Bottom has been nominated in five categories, although not best picture. Only two of the five nominees for best directing this year are white men. All four film actors honored at the recent Screen Actors Guild Awards, often a predictor of Oscars success, are Black or Asian.
These accolades, however, mask the fact that the larger problems persist: The overall representation of actors of color in 2021’s best picture-nominated films is still only 23% and just 26% are women. Only 22% of the major behind-the-camera positions in these films were filled by people of color and 31% were filled by women.
The #OscarsSoWhite movement started with a single tweet in 2015, but has evolved into a method for amplifying long-standing conversations about the lack of diversity within Hollywood. April Reign, an attorney and media strategist, wrote the tweet in response to the 2015 Academy Award nominees for best actor being exclusively white, and has been a leading voice on matters of diversity in entertainment since.
She notes that, “#OscarsSoWhite was not just about race. We want to see people from traditionally underrepresented communities in front of and behind the camera in all films because that’s what the world is.”
Diversity ‘in front of the camera’
Our numbers show that since the start of the #OscarsSoWhite movement, there have been wide swings in racial and ethnic representation in films nominated for best picture. The movement was started in the shadow of the 2015 Academy Awards, where under 18% of main cast members were nonwhite actors (for films released in 2014). Since then, the numbers have grown to just above 40% in 2016 and 2018, to as low as under 13% in 2019.
Importantly, large swings in numbers are not the product of widespread changes in representation but reflect single films that featured Black or Latino casts: Fences (100%), Moonlight (100%) and Hidden Figures (47%) in 2016; Black Panther (86%), BlacKkKlansmen (33%) and Roma (100%) in 2018. That is, even when there was relatively high average representation across films, this representation was driven by single films, rather than generalized representation. Said differently: Even when Academy Award films appear diverse across all nominated films, it is because the diversity is concentrated within single films. Notably, 19 of the 58 films analyzed (just under one-third) have no actors of color in the main cast.
Reign agrees that the year-to-year fluctuations are alarming, and that even in years with good representation, there are reasons for concern.
“We can see incremental progress,” she said. “The fact that we have a Crazy Rich Asians, the fact that we have a Coco, we have a Black Panther – those are all fantastic in and of themselves, but they skew the results for everything else, because those films are so heavily focused on people of color.”
When we consider films with relatively diverse (greater than 20%) main casts, some interesting observations can be made. Films such as 2016’s Hidden Figures and 2018’s BlacKkKlansmen each had prominent Black actors as leads. And while their overall representation numbers (47% and 33%, respectively) weren’t far from the year’s average, neither one reached 50%. That is, even though those films had strong representation in key spaces, their films didn’t employ an especially large number of diverse actors.
Analysis of gender provides an interesting comparison to the data on racial and ethnic diversity. For one, women actors remain underrepresented, comprising just under 30% of main casts. Similar to the race and ethnicity statistics, several films had especially high gender representation, often in films that were about women: 2015’s Brooklyn (73%), 2017’s Lady Bird (60%) and 2019’s Little Women (59%).
While there were modest increases in 2016 and 2017, the overall representation of women was more stable than the race and ethnicity figures across the seven years since #OscarsSoWhite, ranging from a low of just under 26% in 2020 to a high of just above 34% in 2017. Overall, the representation of women across years in the main casts of best picture-nominated films remains low, but is higher overall (29% to 23%), with a more narrow range than the race and ethnicity figures (range: 12% in 2019, 40% in 2016 and 2018). Either way, the reality remains: Most of the fictional worlds of films nominated for best picture are primarily populated by men.
Diversity ‘behind the camera’
The issue of leadership, those who identify and establish priorities, curate projects and cultivate talent, has been central to discussions about why diversity gaps exist. Our study identifies staggering gaps in representation in ethnicity and gender behind the camera, with subtle, yet powerful patterns.
For one, both people of color and women were not only underrepresented (14% and 26%, respectively) in the seven years of best picture nominees, their numbers were both lower than their representation among actors (23% and 29%).
The percentage of women behind the camera was close to their numbers on screen (26% to 29%), suggesting that women are underrepresented relatively equally behind and in front of the camera. But these overall numbers only tell part of the story. An examination of what happens position by position reveals more disturbing patterns.
For example, while women made up a staggering 78% of casting directors, they were only 8% of directors and 5% of cinematographers. People of color — only 14% of those working behind the camera across roles — ranged from 11% of editors to a high of 25% of directors. This is a striking finding, highlighting how women are far behind in technical roles (e.g., cinematographer and director), while racial and ethnic representation remains low in producer roles (11% of producers to 15% of executive producers).
“What happens behind the camera from screenwriting to directing to producing to editing is incredibly important,” Reign said. “When it comes to diversity behind the camera, the most important questions are who is telling the story, and whose story is being told.”
Rashad Robinson, the president of Color of Change, agreed, saying that “when people are shut out from power and the ability to actually have power in terms of being able to tell their own stories, it has a major impact.”
What do we make of these numbers? They highlight how the terrain of diversity, equity and inclusion is complex. Our findings offer several perspectives and potential solutions.
1) For actors, segregation should not be a stand-in for representation. On the one hand, stories such as 2016’s Moonlight, 2018’s Roma and 2019’s Parasite are critical, as they are stories told by, and about, people who are underrepresented in Hollywood. The existence and critical acclaim for these films is a sign that stories about a wide breadth of experience are an unequivocal positive. Their existence, however, highlights another problem in Hollywood: Most best picture nominees portray overwhelmingly white worlds with few Black and brown actors in significant roles.
2) Behind-the-camera, specific initiatives are needed to address specific gaps. The lack of women in roles such as cinematographer and director are telling. Similarly, the relative lack of racial and ethnic diversity among producers is an issue for which films get made. While producers are not creative positions, they are the chief administrators of film projects, and have power over how a film project is carried out.
Robinson said that attempts to address inequalities must be “conversations about changing rules that produce and manufacturer inequality.” He warned that we “don’t want the solution to be driven by authors who have created unequal structures, who can then feel charitable by setting up programs for people to have opportunities.”
#OscarsSoWhite has brought widespread attention to the important issues that it raised, but our findings suggest that Hollywood has a long way to go.
We don’t expect Hollywood’s diversity problems to be any easier to solve than diversity issues in other industries. But entertainment provides a unique opportunity to observe patterns on a large scale in highly visible settings. Our data reveals that the scope and shape of diversity problems are nuanced, and require careful attention moving forward.
This study was pioneered by Nelson K. Bennett, a Washington-based technologist who collaborated with C. Brandon Ogbunu. The two have founded QuantNoir, an initiative that uses quantitative methods to investigate cultural phenomena.
The year described in the data refers to the year that the film was released, not the year that the movie was nominated. For example, the films nominated for Academy Awards in 2021 were released in 2020.
Each film was evaluated for its racial, ethnicity and gender representation by calculating the percentage of main acting roles or “behind-the-camera” jobs filled by people of color or women. Films are compared using percentages of their casts or behind-the-camera roles (i.e., a cast of 100 with 20 actors of color is treated the same as a cast of 10 with two actors of color). Generally, the major actors are identified in the opening credits or the closing title credits before the full list of extensive cast and production crew. Also, they are usually listed in the “billing block” for the film – the list of names on the bottom portion of the official poster (or “one sheet,” as it is called in the movie industry).
Race/ethnicity and gender were determined by viewing photographs of the individuals from various sources. Three individuals viewed photos of every individual, and the perceived race/ethnicity was determined by majority judgment. We recognize that perceptions of race and gender can be imprecise, but this is a limitation of almost all studies of this kind. For issues of diversity and inclusion in the entertainment industry, perception is a key metric. While a few individuals may self-identify differently from how they are categorized in this study, this approach is useful as a method for dissecting patterns across dozens of films and hundreds of individuals.
The data used in the study can be found here. It is available with attribution for non-commercial use under a Creative Commons license. For other questions, please contact C. Brandon Ogbunu (@big_data_kane) or Nelson Bennett (@adibogodi).