Explaining Statistical Disparities in Who Gets Hired for the Plum Hollywood Movie Roles

That's not a valid work email account. Please enter your work email (e.g. you@yourcompany.com)
Please enter your work email
(e.g. you@yourcompany.com)

HOLLYWOOD QUOTA MACHINE? I am not a statistician, so maybe that’s why I find it intriguing that observed statistical frequencies and percentages are what they are, rather than some other number, e.g., why, according to a recent report, exactly 71% of all protagonists in all 2013 Hollywood movies were male—nearly five times the percentage for females (15%), rather than, say, 57% or 90%; and 75% of them were likelier to have stated occupational goals (as opposed to personal life goals and a figure of 25% for female characters).

On the surface, it may look as though some invisible hand is setting Hollywood quotas—an idea that, nonetheless, sounds as implausible as suggesting that the frequency of getting a “4″ on a roll of one die is 16.66% (1-in-6) only because some cosmic quota has been set by the Force.

This quirky concept of a quota was whimsically pondered in an analysis I once read about the number of people bitten by dogs per year in some big city—New York, as I recall, during the 1970s, apparently never varying from 32 incidents by more than +/- 2, and, on top of that, perfectly matching the theoretical statistically predicted and calculated average generated by the raw data fitted to what is called a “Poisson distribution”. As I recall the author, M.J. Moroney (in his well-received popularized textbook Fact from Figures), saying, it was as though the 35th dog that might otherwise have bitten someone knew that the quota had been met (or, I might add, that the 35th person was smart enough to cross over to the other side of the street).

So, reading the Hollywood data, I couldn’t help imagining some producer or director somehow intuiting that the invisibly set quota of big-lead parts for males or bit parts for females (on analogy with dog bites) had been filled at 71% for males or whatever for females, respectively.

Mysterious 5-to-1 Odds in Bars, on Last Trains and at 5,000 Feet

It was the same feeling I always had years ago while working in Japan and waiting for the last train or going into a karaoke bar: Mysteriously the ratio of males to females hovered around 5-to-1 in the wee hours, irrespective of where in Tokyo or Kobe I was. It was as if the Force had decided that in dubious or dangerous late-night contexts, Japanese females should be only 1/5th as numerous as males. To this day, that still puzzles me.

One empirically testable hypothesis: whether Japanese females are likely to appear in that ratio when the context is somehow disreputable, risky, etc., perhaps suggesting that a female is five times as precious or vulnerable as a male. On a whim, while writing this, I decided to test my own hypothesis and searched for the percentage of women registered as members of the American Parachuting Association.

If my hypothesis is correct, it should be around 16.66%, i.e., 1 in 6, given a 5-to-1 ratio. Here’s what I immediately found at Sisters in Skydiving : “About half of first-jump students are female, yet women make up only about 15 percent of USPA members.” See what I mean? That’s a pretty close match. (Consider doing some testing of your own: Hang around until closing at a disco or its equivalent, a subway or at a 7-11 at 4 a.m. Note the male-female ratios.)

Cosmic Quotas?

So, working with a cosmic or otherwise weird quota concept, why would the stats for Hollywood gender and race casting be what they are if they were determined neither randomly, not by any a priori probability (such as the number of faces of a die or aces in a deck), not by some hidden mathematical algorithm (such as the Golden Section or Fibonacci sequences), nor by any obvious correlated data—such as demographic proportionality (ruled out here given that females represent more than a much smaller fraction of the population than males).

Let’s start with one of the biggest disparities: According to that report, It’s a Man’s (Celluloid) World: On-Screen Representations of Female Characters in the Top 100 Films of 2013 “, 15% of the movie protagonists (and therefore the actors hired for those roles) in 2013 were female. Notice that is exactly the same percentage for card-carrying USPA female parachutists. Hmmm…1 in 6 once again, i.e., a 1- to-5 ratio. Note that this was not a “cherry-picked” pairing: I investigated the skydivers first and only because of the common element of some kind of perceptible risk I conjecture is part of the 5-to-1 ratio.

Even assuming that the Hollywood casting game is one of “perfect information”, i.e., that each casting director or producer knows exactly how many males and females all the others have picked or are planning to, that would do nothing to explain why it’s a 5-to-1 male-to-female ratio for the movie protagonists (as well as for Japanese women out at night and American women jumping out of a plane on a regular basis).

A Sociobiological Rationale for Hollywood Casting Discrimination?

Taking the kind of sociobiological perspective that I favor, I’m tempted to look for some analogue and combination of “parental investment theory”, sex-ratio theory and kin-selection theory, which accurately predicts statistical sex ratios and the quantity of resources various species will invest in offspring and more distant relatives (rather than those unrelated to them).

Might it be that a female movie protagonist requires five times as much investment as a male, e.g., what with the time and cash required for wardrobe, makeup, and their shorter “shelf life” before being shelved because of age bias and sequels and long-term contracts in mind? If that were so, the averaging out at 5-to-1 might be as unsurprising as having 1/6th of random rolls of a die turn up a “4”.

That’s probably hard to believe. But, if A-list actresses or female ACTRA members made, on average, 1/6th of what the males make as income, the casting ratios might, from this investment perspective, be more approximately equal.

In fact, the top 10 female A-list stars, e.g., Angelina Jolie and Meryl Streep, earned collectively $181 million vs. $465 million for the likes of Robert Downey, Hugh Jackman, et al. That’s about 39%. So the 5-to-1 ratio mystery lingers. (Note: in the porn industry, the stats are approximately reversed, with males earning much less than females, despite the health and other risks to males and females alike.)

These percentage disparities are in sharp contrast to the fundamental sex ratio of approximately 1-to-1 for human males and females, which from the perspective of both parental investment and reproductive success must be what it is, despite the fact that for well-off parents investment in a son who can afford many wives, mates, etc., will be rewarded with more grandchildren, i.e., copies of family genes, than investment in a daughter.

The reason the 1-to-1 general population ratio is stable is that any deviation from it would, over time, create a surplus of one sex over the other, e.g., males, through female infanticide, thereby increasing competition among them for the relatively scarcer and therefore now more valuable female offspring, perhaps resulting in selecting against males.

(An analogous argument applies for a surplus of females and in both cases involves some natural mechanisms that require no conscious intervention, e.g., the greater vulnerability of male fetuses when the intrauterine or external environment is unfavorable and unlikely to make multiple, successful matings easy for males, while every female is virtually assured of having at least one offspring).

As an illustration of automatic natural “quotas”, consider the case of one species of fish, the adult bluehead wrasse, which actually switches gender under the influence of favorable or unfavorable changes in the environment and sex ratios. Guppies are also known to have automatically altered sex ratios in response to environmental changes, with more females being produced in unfavorable environments—exactly as sociobiological theory explains and predicts.

Hollywood-Role Race Ratios

Regarding Hollywood race ratios, take Asian stars, as an example: the report states that, “Moviegoers were as likely to see an other-worldly female as they were to see an Asian female character.”

That’s because the percentage of female characters in Hollywood movies was tied by non-terrestrial character castings, at 3%. Here, however, the numbers are less mysterious, since, the industry seems to think that at 5% of the population, audiences of Asian descent aren’t large enough to cater to —as much of Hollywood production does, namely, to the demographics of mass audiences.

(Figures for Asian males were not provided, but if available should be examined to see whether they bring the total close to the Asian population percentage.)

Hence, although there is no mystery in this case of race-based casting, there is a tacit, if not cosmic quota—a quota defined by movie-industry estimates of population percentages and projected race-based box-office revenues.

But then, manipulation of sex ratios in China that has resulted in a huge surplus of males and distorted sex ratio because of preferences for sons should make this kind of bias comprehensible to its Hollywood A-list Asian victims.

Mutant Reptiles in the Theaters?

But that kind of studio thinking forces a tantalizing question about the movies featuring “other-wordly” characters, such as E.T. or the non-human menageries in the Men in Black, Mutant Ninja Turtles, Short Circuit (featuring a lovable robot) and Ghostbuster series: Do Hollywood studio research stats prove that extra-terrestrials, ghosts, robots and mutant reptiles are seated next to us in darkened theaters, generating huge chunks of their box office revenues?

Perhaps only the Hollywood movie moguls and their reptilian overlords know for sure.

By Michael Moffa