A Closer Look at Media Diversity: Should TV Reflect Reality?

April 18, 2016

There has been much talk about the importance of media diversity, especially over the past few years and in the wake of the controversy over a Hollywood dominated by old White men.

Two recent reports have helped describe and quantify the issue, one from GLAAD (formerly the Gay and Lesbian Alliance Against Defamation), and the other from the USC Annenberg School for Communication and Journalism. The reports (Where We Are on TV and Inclusion or Invisibility? Comprehensive Annenberg Report on Diversity in Entertainment, respectively) are the results of painstaking tabulation and analysis. Like any measure they are not perfect, but they do provide insight into media representation trends. As a longtime researcher on the media (and author of Media Mythmakers: How Journalists, Activists, and Advertisers Mislead Us), I was interested to review the findings, specifically focusing on the status of sexual and racial minorities depicted onscreen.

According to the GLAAD report, “Counts are based on scripted series which air or are expected to air in primetime between June 1, 2015 and May 31, 2016 for which casting has been announced. This season marks the 20th year GLAAD has tracked the presence of LGBT characters on television by calculating their numbers in scripted primetime programs across both broadcast and cable networks. In 2005, GLAAD expanded this count into the Where We Are on TV report, which has allowed us to track trends and amass statistics for all series regular characters on broadcast television with regard to sexual orientation, gender identity, and race/ethnicity.”

Part of GLAAD’s aim is making sure that minority characters are represented not only in a positive light (i.e., not as offensive stereotypes), but also accurately and proportionally in comparison to the demographics of America and its television audiences. In several places GLAAD specifically compares ratios of fictional characters with real-life populations in its report, for example noting that “This year, 43% of series regulars on primetime scripted broadcast series are women, an increase of three percentage points from last year. According to the 2010 census, the country’s population is 51% female” (p. 13).

The Where We Are on TV report begins with a one-page series of highlights, and the first one states “Of the 881 regular characters expected to appear on broadcast primetime programming in the coming year, 35 (4%) were identified as gay, lesbian, or bisexual” (p. 5). GLAAD found another 35 recurring characters-people who make multiple appearances in a series but are not part of the main cast-who fit that description (pp. 4-7).

Interestingly–and counterintuitively for many people–the GLAAD report found that gay, lesbian, and bisexual characters are if anything overrepresented on broadcast and cable television compared to their numbers in the general population.

Sexuality is of course fluid and difficult to quantify (the responses vary by who is asking the question, whether sexual behavior is self-reported and self-identified, and so on; for a discussion of the challenges, see pages 2 and 3 of the Gates report, discussed below), but most of the research finds that well under 5% of the U.S. is gay, lesbian, or bisexual.

For example Joel Best, professor and chair of sociology and criminal justice at the University of Delaware, finds that “the incidence of homosexuality among adults is between 1 and 3 percent” (for a detailed analysis see pp. 89-93 in Best’s 2001 bookDamned Lies and Statistics: Untangling Numbers from the Media, Politicians, and Activists, University of California Press). A widely-cited 2011 study by UCLA demographer Gary Gates of the Williams Institute on Sexual Orientation Law and Public Policy concluded that “An estimated 3.5% of adults in the United States identify as lesbian, gay, or bisexual and an estimated 0.3% of adults are transgender.”

As for gays and lesbians specifically (excluding bisexuals) the Gate report found that estimates range from 0.7% to 2.5%, with an average across nine studies of 1.34% (p. 3). Thus depending on which statistic you use for comparison, according to GLAAD’s research there are several times as many gays and lesbians on television as there are in real life.

Interestingly, a 2015 Gallup poll found that Americans greatly overestimate the number of gays and lesbians: “The American public estimates on average that 23% of Americans are gay or lesbian,” a number which has remained about the same since 2002. Thus most people believe that about one in four Americans are homosexual when in fact about one in seventy are.

The Annenberg study used a different methodology and data set: “The Comprehensive Annenberg Report on Diversity (CARD) assesses inclusion on screen and behind the camera in fictional films, TV shows, and digital series distributed by 10 major media companies (21st Century Fox, CBS, Comcast NBC Universal, Sony, The Walt Disney Company, Time Warner, Viacom, Amazon, Hulu, and Netflix). Movies theatrically released in 2014 by the major studios or their art house divisions were included in the sample, provided they met a certain threshold of domestic box office performance. Prime-time first run scripted series as well as digital offerings airing from September 1st 2014 to August 31st 2015 were sampled on broadcast, popular basic cable, premium channels or streaming services associated with the companies listed above. In total, the sample included 414 stories or 109 motion pictures and 305 broadcast, cable, and digital series. The major unit of analysis was the speaking or named character. Each speaking character was assessed for role, demographics, domesticity, and hypersexualization” (p. 1).

The CARD report concluded that “Of the 11,194 characters that could be evaluated for apparent sexuality, a total of 224 were coded as Lesbian (n=49), Gay (n=158), or Bisexual (n=17)” (p. 11). For homosexual (gay and lesbian) characters specifically, the Annenberg study found that 1.8% of TV characters were represented. Since the number of gays and lesbians in the United States is estimated at around 1.34%, this number falls in the range of these estimates as well as the 1% to 3% calculated by Prof. Best.

Thus both the GLAAD and the Annenberg reports found that gays and lesbians are, if anything, overrepresented on television as compared to American demographics. When bisexuals are included the studies differed, though in opposite directions: Of broadcast primetime programming characters, GLAAD found that 4% were identified as gay, lesbian, or bisexual, which is 0.5% higher than reflected in the U.S. population.

The Annenberg report, using a different dataset and methods, found half as many LGB characters as the GLAAD study did: “Only 2% of all speaking characters across the 414 movies, television shows, and digital series evaluated were coded LGB. This point statistic is below the 3.5% of the U.S. population that identifies as Lesbian, Gay, or Bisexual, as reported by the Williams Institute at UCLA” (p. 11). Thus there are 1.5% fewer LGB characters on television than there are in real life by the CARD calc

As for transgender characters in the CARD study, “A separate measure assessed whether characters were transgender. Only seven speaking or named characters identified as transgender sample wide, which calculates to <1%” (p. 11). This statistic seems to accurately reflect the 0.3% transgender U.S. population found by the Williams Institute on Sexual Orientation Law and Public Policy. The GLAAD report found that 2% of the characters in primetime programming on cable networks were transgender (p. 9), which would mean that transgender characters on television outnumber their real-life counterparts by a factor of about six to one. The study found no transgender characters on broadcast television, and 7% of LGBT characters on streaming content providers were transgender.

The fact that both organizations found that gay and lesbian characters are accurately represented (if not overrepresented) in scripted entertainment media would seem to be a remarkable finding worthy of note in the reports (if not national news), touting how successful GLAAD and others have been in helping get accurate representations on broadcast television. Obviously there is a long way to go and much more needs to be accomplished, but I found it strange that the result wasn’t mentioned in any news reports that discussed the studies, including the New York Times and in an otherwise fairly in-depth piece on Slate. I reached out to several spokespersons at GLAAD and USC Annenberg School for Communication and Journalism (including authors of the two reports), but all declined comment.

Dueling Demographics

The GLAAD and Annenberg reports raise provocative questions about the comparison of real-life to fictional minority characters. It’s important to recognize that having a disproportionate number of sexual, racial, or other minorities on television is neither inherently good nor bad: The question of whether there are “too few” or “too many” is a subjective value judgment depending on your goals and agenda.

There is no magic or “correct” number for how many gays, lesbians, or other minorities “should” be main characters on television shows; it all depends on what you are comparing that number to. One goal might be to have a 1:1 ratio of fictional minority characters to real-world minority characters, as the GLAAD report suggests for women’s roles. Another goal may be to simply have more minority characters, in both positive and negative roles; yet another goal may be to have more such characters solely in positive roles.

The studies raise a curious dilemma: If accurate demographic representation of fictional characters is the goal (as is stated explicitly in both reports), then, for example, homosexual and transgender characters will continue to be nearly invisible on television for the simple reason that they make up such a statistically small segment of the population.

Racial Representation

The GLAAD report also found that African-Americans are about accurately represented as TV characters compared to U.S. demographics for the 2015-2016 season: it reported “the highest percentage of Black regular broadcast characters it has ever counted: 16%” (p. 19). Blacks constitute about 14% of the United States population according to the CDC. Latinos, however, remain underrepresented by about half, according to GLAAD: “The percentage of Latino/a characters has actually dropped by one point to 7% (61) from last year’s all-time high. This is a drastic difference from the actual U.S. population, which was estimated to be 16% Latino/a in the 2010 census” (p. 21). Thus the GLAAD report found that, demographically, gays and Blacks are about accurately represented (or overrepresented) in TV, while Latinos and others are significantly underrepresented. There are of course other ways to measure fictional TV representation, but that is the primary measure that GLAAD chose to use.

The Annenberg report found that “Of those speaking or named characters with enough cues to ascertain race/ethnicity (n=10,444), 71.7% were White [and] 12.2% Black…. The number of shows featuring ‘racial/ethnic balance’ was evaluated. If a show featured any underrepresented characters within 10% of the U.S. Census statistic, it was considered balanced. Only 22 stories depicted racial/ ethnic balance on the broadcast networks (19%), 18 on cable (13%), 1 on streaming (2%), and 8 in film (7%). Clearly, most stories fail to reflect or match the demographic composition of the U.S.” (p. 7). (It’s not clear whether or not the Annenberg report considers the racial balance of Black characters to be balanced or not; since Blacks comprise 14% of the population, the 12.2% the study found would seem to be within the plus or minus 10%, though that threshold may apply only to within shows, not between them.)

Quantification Complication

Quantifying TV shows is complicated by the fact that some of the best and most interesting characters are neither all good nor all bad but display a blend of good and bad traits as their character moves through a compelling story arc; Breaking Bad’s Walter White, for example, at various times is sympathetic, heroic, and repellent. Most actors–of whatever gender, sexuality, or race–relish playing those complex characters. Is a courageous and hard-working Mexican day laborer in a TV series a positive role model because of his character, or a negative one because of his menial job? What about a successful but morally corrupt and predatory Mexican lawyer? With a few black-and-white exceptions, quantifying whether a given television character is “good” or “bad, or a “positive” or “negative” role model-or representative of a given minority-is an unenviable task. These issues are further complicated if we try to quantify multi-racial characters or those with not-fully-explained (or shifting or confusing-to-categorize) sexualities. (The CARD report ends with five pages of footnotes describing in detail how the sampling was conducted, and how these and other vagaries were quantified.)

Another issue is that in any study of artistic representations of life such as TV or movies, LGBT will naturally appear unrepresented if characters whose sexuality is not made explicit are included because they will be coded heterosexual by default. A given character’s sexuality may not be referenced for the simple reason that it’s irrelevant to the script, and not every TV or film character has a story arc that would necessarily reveal his or her sexual orientation. For example is the Soup Nazi’s sexuality germane to the Seinfeld character? That character may in fact be homosexual in his backstory, but unless a script specifically refers to his sexual or relationship status–a potentially jarring tangent–then he will be counted as straight for no reason other that it wasn’t mentioned.

In real life, of course, we interact with dozens of people each week who don’t explicitly reveal their sexuality to us, from coworkers to store clerks to customers. In today’s sexually diverse world it would be foolish–perhaps even homophobic–to assume that everyone is heterosexual unless told otherwise, but that seems to be the assumption behind the methodology used in these studies.

GLAAD has acknowledged the inherent difficulties and shortcomings of using TV shows to measure LGBT representation. Shortly after its report was issued last year, the organization announced that it would no longer be compiling its annual Network Responsibility Index, ranking TV networks on their LGBT inclusiveness. Slate‘s June Thomas noted, “As GLAAD CEO and president Sarah Kate Ellis explained in a slightly less than c
rystal-clear column in Variety, the organization decided that counting the TV queers wasn’t a satisfactory way to measure the medium’s gay-and-trans-friendliness.”

Fiction vs. Real Life

Quantification difficulties aside, there is of course no reason why fictional entertainment should necessarily accurately reflect real life–in dialogue, plot, or percentage of characters who come from different demographic categories. Television is escapist entertainment, and the vast majority of characters in scripted shows lead far more interesting, dramatic, and glamorous lives than the audiences who watch them.

While fictional cops on television shows regularly engage in gunfire and shootouts, in reality over 90% of police officers in the United States never fire their weapons at another person during the course of their career. TV doctors seem to leap from one dramatic, life-saving situation to another, while most real doctors spend their careers diagnosing the flu and filling out paperwork. (There are of course many completely fictional characters on television including superheroes, zombies, and vampires with no corresponding real-life counterparts to compare the numbers to–but if they did exist, it’s a safe bet that real zombies, superheroes, and vampires would also probably lead much less exciting lives than their on-screen role models.)

Countless real-world things, people, and situations are under- or over-represented in scripted television drama as compared to real life. TV shows about police officers and doctors, for example, greatly outnumber those about accountants and teachers for the simple reason that some careers are inherently more dramatic than others.

Young people are more likely to be main characters in dramas for similar reasons; a character in her twenties or thirties is (rightly or wrongly) assumed to be more socially and physically active and have more personal, career, and other opportunities for drama and character growth than the same character in her sixties or seventies. There are many exceptions, of course–Matlock and Murder She Wrote being prime examples–but screenwriters are, on average, more likely to create roles for younger actors because the dramatic possibilities seem more open. An eight-year-old lead character may have adventures in school, with her family, on the playground, at an amusement park, and so on, while an eighty-year-old lead character is less likely to have such adventures. This narrative bias may or may not be a form of age discrimination, though in any event there is no reason why any number of minorities couldn’t be successfully cast in those roles regardless of age.

Complicating the issue is the fact that television entertainment, perhaps moreso than most other media, is consumer-driven. Shows that are interesting, well-acted, and well-scripted (and, of course, given the opportunity to build an audience–increasingly a luxury in today’s entertainment world) will attract viewers and therefore advertising dollars. Since gay characters, like gay marriage, have been overwhelmingly accepted by the majority of Americans, there’s no reason to assume that the same will not happen with racial minorities as well as sexual ones–and TV depictions may certainly play a role. As scientists know, correlation does not imply causation and thus it’s not clear whether widespread social acceptance of gays and lesbians led to their increased visibility in scripted dramas, or whether more LGBT characters led to more social acceptance, or some combination of the two.

Audiences who demand historical accuracy implicitly limit the race, age, and gender of the actors who can portray certain historical figures. A mainstream Hollywood film about Alexander the Great, for example, won’t be portrayed by a woman, nor an African-American–nor even a middle-aged White male, since the conqueror died at 32–without complaints. (Art or experimental films, of course, have more latitude, as in the 2007 Todd Haynes film I’m Not There, in which Bob Dylan was played by six actors including Cate Blanchett.)

For this reason screenwriters and filmmakers often highlight lesser-known minority historical heroes for their films, as when Chiwetel Ejiofor portrayed Solomon Northrup in 12 Years a Slave, which was not only a commercial success but won three Academy Awards including Best Picture. There is a rich vein of untapped stories of such heroism among all minorities, and surely many more will follow with equally capable execution and success. Hopefully the trends in increased diversity that both reports highlight will continue; having more minority characters telling their own stories and offering diverse viewpoints can only make TV entertainment richer and more interesting-regardless of whether or not their numbers reflect real-world demographics.