Both The Geena Davis Institute on Gender in Media at Mount Saint Mary's University (the “Institute”) and Movio have published extensive research on the topic of representation in media. Movio has traditionally focused specifically on the theatrical exhibition sector and audience data, whereas the Institute primarily captures on-screen data and analysis. Our organizations therefore share an interest in studying the intersection of on-screen data and audience data.

While both the Institute and Movio have explored this subject in the past, doing so independently (limited by our unique data specialities) prohibits the ability to examine the “on-screen to audience relationship” at a meaningful scale.

To produce a macro view of the correlations between content and audience, we’ve joined forces, combining our data sets to enable a comprehensive analysis across hundreds of titles. In our latest white paper, I Want To See Me: Why Diverse On-Screen Representation Drives Cinema Audiences, we explore whether what’s shown on screen affects who shows up in the audience.

We focused our research on three key questions:

• Does the presence of certain groups (Race/Ethnicity, Gender, Age) on screen draw larger numbers of the corresponding audience?
• What negative or positive portrayals of certain groups are different viewers seeing in the most popular films?
• What portrayals of certain groups are child viewers seeing in the most popular films?

Discover the insights and full research below.

Background

Both The Geena Davis Institute on Gender in Media at Mount Saint Mary's University (the “Institute”) and Movio have published extensive research on the topic of representation in media. Movio has traditionally focused specifically on the theatrical exhibition sector and audience data, whereas the Institute primarily captures on-screen data and analysis. Our organizations therefore share an interest in studying the intersection of on-screen data and audience data. In other words:

How does what’s shown on screen affect who shows up in the audience?

While both the Institute and Movio have explored this subject in the past, doing so independently (limited by our unique data specialities) prohibits the ability to examine the “on-screen to audience relationship” at a meaningful scale.

To produce a macro view of the correlations between content and audience, we’ve joined forces, combining our data sets to enable a comprehensive analysis across hundreds of titles.

We wanted to know:

•  Does the presence of certain groups (Race/Ethnicity, Gender, Age) on screen draw larger numbers of the corresponding audience?

•  What negative or positive portrayals of certain groups are different viewers seeing in the most popular films?

•  What portrayals of certain groups are child viewers seeing in the most popular films?

Executive Summary

•  There does seem to be a correlation between the representation of various demographic cohorts* on screen and their share of the total audience to a particular film

•  The relationship seems to be stronger for Race/Ethnicity cohorts than Age or Gender cohorts, where a higher percentage of characters of a given Race/Ethnicity on screen results in a higher percentage of moviegoers of the same Race/Ethnicity in the audience

•  Within Race/Ethnicity it does seem to be most apparent for the Black cohort than other Race/Ethnicity cohorts

• It does not appear that there is any evidence that the relationship is impacted by whether or not the group represented on screen is portrayed in a way that could potentially be considered negative (crime, lazy, visually objectified)

•  Looking at representation among films that moviegoers are bringing their children to see, there is significant under representation of Racial/Ethnic minorities in these storylines (Asian's and Latinx being particularly underrepresented), relative to what we might expect given the prevalence of these groups in the overall US population

•  If we focus on gender we see that both Male and Female leads are roughly evenly split in terms of on-screen representation for films with more child visits. This suggests that young moviegoers are getting a fairly balanced representation of genders, when we consider leads, in their on screen media consumption. It is worth noting there is still big gender gaps when we consider supporting, minor and overall characters.

‍

*For the sake of this analysis, a cohort is a group of moviegoers defined by the specified demographic characteristic (Age, Gender, Race/Ethnicity)

Methodology

We examined both the on-screen data and audience demographic data for the top 100 films released theatrically in the U.S. in 2018 and 2019.

Audience Data — Movio

As a trusted partner of exhibitors worldwide, Movio is the largest aggregator of cinema loyalty data. With direct integrations to exhibitor point-of-sale and loyalty CRM programs, our technology captures the
title-level ticket purchase and attendance behavior data of over 10 million active U.S. moviegoers, while our database holds the demographic profiles (including age, gender, ethnicity, and location) of the millions who attend the cinema each year.

Loyalty programs contain a demographic bias in their composition when compared with the general U.S. moviegoing population. To compensate for this bias, we apply a demographic weight to each member profile so that the composition of the weighted population within our moviegoer database matches the demographic breakdown reported in the MPAA Theatrical Market Statistics Report.

Child attendance is calculated using ticket type data (child tickets), with the proportion of child tickets as a proxy for the proportion of audience as a whole. Movio has previously validated that the percentage of child tickets purchased by loyalty members generally mirrors the percentage purchased overall.

The audience data in this report is session-based, corresponding to the ticket purchaser. For example, if Susan (Female, age 35, Asian) buys three tickets to a given showing, the audience data in this study includes only one person for that showing: Susan herself. The other two tickets purchased are disregarded as the demographic profiles of those attendees are unknown.

On-Screen Data — The Institute

For the on-screen analysis in this report, we employ the Geena Davis Inclusion Quotient (GD-IQ), the only software tool in existence with the ability to measure screen and speaking time through the use of automation (see Appendix). This revolutionary tool was developed by GDIGM at Mount Saint Mary’s University and funded by Google.org. The GD-IQ, which incorporates machine learning technology, was designed by Dr. Shrikanth Narayanan and his team of researchers at the University of Southern California’s Signal Analysis and Interpretation Laboratory (SAIL), along with Dr. Caroline Heldman.

On-screen analysis for this study focused on leading and supporting characters. As such, throughout this report, “characters” refers to those with a leading or supporting role. A leading (or co-leading) character is defined as those protagonists that drive the unfolding storyline. Some movies are “buddy films,” depicting two or more characters on the same journey. When this occurs, the coleads are both selected as story protagonists. For ensemble casts, you would choose the character who is most driving the story. A supporting character is defined as those who appear in more than one scene and are instrumental to the action of the story. Characters deemed essential to the development of the central or ancillary plots are considered supporting characters. All other speaking characters are minor characters.

For each character, we captured any available data regarding the character’s gender, approximate age, and ethnicity or race. Additionally, each character was assessed for exhibiting certain negative traits, such as promiscuity, criminal behavior, revealing clothing, laziness, etc. For a full list of traits categorized as negative, please see the Appendix below.

1. The Relationship Between On-screen Representation And Corresponding Moviegoer Attendance

For each movie, we calculated the percentage of total characters on screen for the each race or ethnicity, gender, and each age group, broken down as follows. We then calculated the percentage of the attending audience by the same factors. For each of the following charts:

•  Ethnicity: Asian, Black, Latinx, White

•  Gender: Female, Male

•  Age: Gen Z, Millennials, Gen X, Boomers

Note, the number of Gender non-conforming characters was too small to analyze and currently the audience dataset only includes binary Gender options for moviegoers. The same is also applicable for other ethnicities outside of the one's above.

All the below charts are interactive.

These charts paint a stark picture of the limited opportunities minority moviegoers have to see themselves reflected on screen. Whereas several movies tally 100% of their characters as White, and the majority having over 50% White characters, for the remaining four Race/Ethnicity groups, the majority of films are clustered at below 25% (if not 0%) representation on screen.

We do see a general trend across all Race/Ethnicities where a higher percentage of characters of a given Race/Ethnicity on screen results in a higher percentage of moviegoers of the same Race/Ethnicity in the audience. This trend seems especially strong for Black moviegoers. Given the high percentage of Latinx audience despite low percentage of Latinx characters, one might assume that an increase in Latinx on-screen representation would similarly strengthen the trend toward increased representation of Latinx attendees in the audience.

Looking at representation across genres we can see that films with Asian characters tend to be over-represented in Action and Comedy genres and under-represented in Drama and Horror genres relative to all films in the dataset. Films with Black character representation are over-represented in Action and under-represented in Drama and Horror relative to all films. Films with Latinx character representation are over-represented in Horror, Animation, and Adventure and under-represented in Action and Comedy relative to all films.

If we look at films with Black representation a bit further, looking at majority and minority representation, we see that films with a primarily Black set of characters are significantly over-represented in Comedy, Drama, Thriller, and (to a lesser extent) Crime and under-represented in Adventure, Animation and Biography compared to all films.

For both the Female and Male gender analyses, the movies are generally clustered toward the center of the chart, indicating that both the characters and audience are generally 50/50, Female/Male. As this is a binary comparison, the charts here are an inverse of one another.

On each, the perceivable incline from bottom left toward upper right suggests that the more Female characters, the more Female the audience, and the more Male characters, the more Male the audience.

What is quickly identifiable, as well, is the effect of genre on the percentage characters Male vs Female, and therefore on the gender split in the attending audience. Action movies, most notably, regularly have well over 50% Male characters and a well over 50% (and frequently over 60%) Male audience.

On-screen representation by age appears to have a small effect on attendance for only certain generational groups. Attendance by Millennials and Gen X is not noticeably affected by increased representation on screen, which may make sense given they are the middle of the age distribution and thus find their way into films aimed at both younger and older audiences. For example they may be the parents in a film aimed at a Gen Z audience and the middle aged children in an audience aimed at the Baby Boomer generation.

The relationship among Gen Z and Baby Boomers is particularly interesting to consider given the importance of these two generations on cinematic exhibition. In the case of Baby Boomers, previous research by Movio found that the Boomer audience was a powerful segment at the box office; often bolstering blockbusters (27% of the total audience for Star Wars: The Force Awakens was aged 50+), driving drama title attendance, and creating lucrative new niches that don’t rely on major budgets and foreign success to be profitable. For the Gen Z audience, it will be crucial for cinemas to attract this cohort to theatrical exhibition as they become a larger share of the potential moviegoing audience in the decades to come.

‍

2. The Impact Of Negative Portrayal Of A Cohort On The Relative Cinema Audience Attendance

To see if a negative portrayal of characters from a given cohort had any effect on attendance from the corresponding group, we again looked at the percent audience and percent characters by race or ethnicity, by gender, and by age. Of those qualifying characters, we then analyzed each to see what percent did or did not exhibit certain negative traits or actions.

The negative traits, actions or portrayals considered for this analysis include the below. For more information reference the appendix.

•  Character is a criminal or exhibits criminal activity

•  Character exhibits promiscuous behavior

•  Character exhibits sex for trade

•  Character is verbally sexually objectified

•  Character is visually sexually objectified

•  Character’s intelligence is considered ‘stupid’

•  Character’s work ethic is considered ‘lazy’

On the following charts, a blue dot indicates a film in which at least one character of the specified demographic cohort exhibits at least one negative trait and a yellow dot indicates a film in which no characters in the specified demographic cohort exhibit a negative trait.

In other words, a film in which any character of the specified demographic ‘checks the box’ for any of the negative traits will be represented by a blue dot – regardless of if that character also exhibits positive characteristics or evolves during the course of the story to overcome or shed any negative traits.

For example, on the charts below, a blue dot might indicate that 25% of all characters in a given movie were Black and at least one of those characters were depicted negatively at some point in the story. A yellow dot might indicate that 25% of the characters in a given movie were Black and none of them were depicted negatively at any point in the story.

For each of these following charts:

•  Ethnicity: Asian, Black, Latinx, White

•  Gender: Female, Male

•  Age: Gen Z, Millennials, Gen X, Boomers (see Appendix for breakdown)

Ultimately it appears from this analysis that across the board in terms of ethnicity, gender, and age, the negative portrayal of characters from a certain group has little bearing on whether or not that group attends a movie.

If it were the case that audiences were negatively affected by negative representation of their demographic cohort on screen we would expect to see one of the following:

1. The trend line for No Negative Representation being consistently higher than Negative Representation by some fixed amount, suggesting that for two titles with the same percent of total characters being portrayed on screen, a film with no negative representation would always see the demographic cohort being represented be a higher share of its total audience than one with negative representation by some fixed amount.

2. The trend line for No Negative Representation having a higher slope than Negative Representation, suggesting that for a given increase in percent of total characters being portrayed on screen, a film with no negative representation would see a larger increase in that cohort as a percentage of it’s audience than one with negative representation.

Looking at the results above we see that neither of these seem to be the case in any consistent manner, rather there seems to be little to no difference between the two categories for each demo cohort considered in this analysis.

However, particularly with regards to race and ethnicity, these charts again reflect how minority audiences are given substantially fewer opportunities to even see characters from their racial or ethnic group on the screen, no less characters from their racial or ethnic group who are also not depicted negatively. On the other hand, the vast majority White characters exhibit some negative trait, which, similarly, has no discernible effect on attendance by White moviegoers.

Note that this binary analysis cannot account for the story narrative and character development and therefore there is some lack of the context around a negative trait. For example, there could be a title where all the heroes are primary Black, and the villains are also Black. Or a story about a criminal who learns lessons to set him straight is probably a positive portrayal of that person, ultimately, but due to the presence of criminal activity, the character checks the “negative” box.

It is also worth noting that, given audiences typically purchase movie tickets before seeing a film for the first time, they may not be aware of any negative depictions within a story until after viewing the film: thus, any potential effect of negative portrayals on audience behavior may not be reflected in audience purchase behavior and any conclusions drawn from this analysis should not be considered absolute. That being said, studios do have ample opportunity to telegraph key story elements in marketing content for a particular film, which would/could theoretically inform moviegoers of character depictions prior to actually viewing the film.

‍

3. The Demographic Cohorts Represented In Films Geared Towards Children

Finally, we wanted to examine the diversity of on-screen representation for films geared towards children, in particular to get a sense of what young moviegoers are being exposed to at the cinema.

As previously mentioned, child attendance is calculated using ticket type data (child tickets), with the proportion of moviegoers purchasing child tickets as a proxy for how many children attended the film. Movio has previously validated that the percentage of child tickets purchased by loyalty members generally mirrors the percentage purchased overall.

For each of these following charts:

•  Ethnicity: Asian, Black, Latinx, White

•  Gender: Female, Male

•  Age: Gen Z, Millennials, Gen X, Boomers (see Appendix for breakdown)

In the below charts, the X axis shows the percentage of a film’s total characters (lead, colead, supporting) that are members of the specified demographic cohort. Films further to the right have more representation of the specified demographic on screen than films on the left.

The Y axis shows the share of total ticket purchasers that purchased at least one child ticket to the film. Films higher up on the Y axis had a higher share of ticket purchasers buying at least one child ticket to the film and are thus the films that parents are actually bringing their children to go see, as opposed to limiting ourselves to what might be considered a “children’s film”. Notice the highest titles are almost all Animated titles (light grey), followed primarily by Action titles (black).

Looking at titles in this perspective we are able to get a sense of what demographic cohorts are being represented in films geared primarily towards child audiences.

If we focus on Gender we see that both Male and Females are roughly evenly split in terms of on-screen representation for films with more child visits (higher up on the y-axis), suggesting that young moviegoers are getting a fairly balanced representation of Genders in their on screen media consumption. In previous research, the Institute found that young audiences are now also seeing a more balanced gender representation in top children’s television programming. Using the Geena Davis Inclusion Quotient (GD-IQ), the report found that in 2018, female characters averaged 55.3% of screen time and 50.3% of speaking time in children’s television programming.

Performing the same analysis by Race/Ethnicity we see less of a balanced representation. As one might expect, White characters are very well represented in films geared towards children, with the majority of films on the top half of the chart having 50% or more White characters represented on screen. For Black moviegoers there isn’t quite the same level of representation, however most of the films on the higher end of the Y axis at least have one or more Black characters represented on screen. Unfortunately we do not see the same level of representation for Asian or Latinx characters in children’s media as the vast majority of films moviegoers are bringing their children to see have no representation of these cohorts. This highlights a clear opportunity for more films, aimed at a younger audience, to diversify the characters represented on screen considering the power media has on influencing values and thoughts, especially children.

Among various Age cohorts we find that films geared towards children generally have at least some representation of Baby Boomers, GenX-ers, and Millennials whereas Gen Z is somewhat rarely represented. However this might make sense as Gen Z is roughly in their teenage years and early twenties which one may not expect to see included in children films.

‍

Conclusion

Movio’s vision is to enable everyone to experience the magic of cinema and therefore to ensure the longevity of the theatrical experience. Meanwhile the Institute works collaboratively within the entertainment industry to engage, educate and influence the creation of gender balanced on screen portrayals, reducing harmful stereotypes and creating an abundance of unique and intersectional female characters in entertainment. With a diversifying population, these two goals are inevitably intertwined.

Our analysis shows that there is some correlation between the representation of demographic cohorts on screen and their share of the total audience to a particular film. In other words, moviegoers being able to identify with the characters in a movie drives their moviegoing behavior. Knowing that this is more prevalent when we look at Race/Ethnicity cohorts, this further strengthens the argument for more substantial diverse representation of characters on screen. This becomes even more apparent when we consider films appealing to children. There is a clear opportunity for more films, aimed at a younger audience, to diversify the characters represented on screen considering the power media has on influencing values and thoughts, especially children.

We know that media like cinema influences not only how we view ourselves but also how others view us. Titles with diverse characters not only attract audiences of a relative cohort but can attract a wide-range of moviegoers, which plays an important educational role, but also provides an opportunity to see box office success. Now more than ever consumers have many entertainment options that speak to them and their lives, so if cinema is to remain relevant and impactful in an increasingly diverse world, the on-screen representation gap must be acknowledged and addressed.
Download Overview PDF here.

About Movio

Movio is the global leader in marketing data analytics and campaign management software for cinema exhibitors, film distributors and studios. A company of Vista Group International Ltd (NZX/ASX:VGL), Movio is the world’s most comprehensive source of moviegoer data, capturing the behavior of over 45 million active cinema loyalty members worldwide. Movie lovers at heart, it’s our mission to connect everyone with their ideal movie.


About the Geena Davis Institute on Gender in Media at Mount Saint Mary's University

Founded in 2004, by Academy Award winning actor Geena Davis, the Institute is the only research-based organization working collaboratively within the entertainment industry to create gender balance, foster inclusion and reduce negative stereotyping in family entertainment media. “If They Can See It, They Can Be It.” For more information visit www.seejane.org

Appendix

Find details and definitions in the pdf below.