Demographics and stories: how numbers lie

I’ve been reading The Halo Effect: … and the Eight Other Business Delusions That Deceive Managers by Phil Rosenzweig. It’s all about how businesses create stories to explain success and failure instead of looking at hard data and analyzing what really happened. For example, when Cisco was riding high, its employees were happy, its CEO had a brilliant sense of making acquisitions, etc. When Cisco’s fortunes flopped, suddenly those same employees had been miserable and the CEO had been making acquisitions foolishly.

And did you ever wonder how the newscasters on Bloomberg know precisely why some index is up today, before they’ve had time to study it, or poll people buying stocks, or something? The answer, of course, is they have no idea: they’re just picking a plausible explanation and hoping you’ll buy it.

This is what film has been doing for the past thirty years. In a recent article, If audiences don’t want women as leads, why did Aliens succeed?, I pointed out some of the serious questions that somebody ought to be asking before proclaiming that mainstream movies with female leads don’t do so well, except when they do, which doesn’t count – and doesn’t raise any questions. That’s a story, not an analysis of actual data with sound conclusions.

You can certainly look at numbers and see that most of the best-performing movies feature male leads, and lots of movies featuring female leads perform poorly. Then again, lots of movies featuring male leads also flop. And, as William Goldman explained, lots of movies featuring women exceed industry expectations, but instead of investigating them to see why they did well and how their success can be replicated, they get dismissed as non-recurring phenomenon. Wow. With jargon like that, it almost sounds scientific.

It’s not. When a movie featuring a white man flops, the industry at least concocts a story to explain it, and sometimes there may even be genuine analysis involved: perhaps it was marketed badly, or it was released at the wrong time or the film itself was bad. Rarely do you ever hear, “I guess no one likes [insert male lead’s name] anymore.” Never, ever do you hear, “The audience just doesn’t want to see movies about white guys.”

But when it’s a movie about a woman (or a man of color), that’s all the story we get: it had a woman, or a man of color, so it flopped. This, by the way, is what belies the industry’s claims that it’s the audience’s bigotry, not the industry’s, forcing women (and others) out of lead positions in films. If the industry was so bigotry-free, they’d be asking the same questions about female-led flops as they ask about male-led ones. They’d be investigating to see if there’s a way to turn it around to their advantage.

So before you keep accepting this alleged bit of “numbers don’t lie” logic, consider this. Here’s a list of questions the industry would be able to answer (or at least show us their work so far) if they had ever scientifically examined the claim that merely casting a woman in a non-“chick flick” film causes it to flop:

  • Who precisely isn’t watching these movies?
  • Why don’t they want to see them?
  • What, if anything, would persuade them to see them?
  • Are they influencing others not to see them?
  • Could there be other factors influencing the film’s success, such as bad marketing or the film sucking or, you know, any of those things they examine when a male star’s latest film flops?
  • ETA: what did the movie earn per theater, or per seat (ticket sold)? These are granular metrics the industry considers when trying to figure out why a film did unexpectedly well or poorly. Sometimes these numbers indicate stuff like whether the movie would have profited more with a wider or narrower release (how many theaters played it).

You can probably come up with others; stick them in the comments, and I’ll edit the post. These are precisely the kinds of questions the industry bothers to ask when films starring white males flop. And whenever I asked these questions of film pros claiming films featuring women just can’t perform, they had clearly never considered them before. Must be sexism, they’d admit with appropriate sadness in their tone. I’d point out that I grew up in a very sexist place and time, yet all the hardcore misogynist guys saw Aliens. “Must’ve been for the special effects,” they’d suggest. I’d point out that there was no shortage of SFX bonanzas in the 80s for misogynists to watch instead. They’d suddenly have someplace else to be.

An extremely important part of data analysis and science in general is: asking the right questions. Without that foundation, the best data in the world won’t inform you of anything meaningful. Sometimes the questions a researcher poses (or lack thereof) reveals a bias on his or her part. Biases make numbers “lie” through the humans that interpret them. The desire to make the numbers tell a particular story is the worst kind of bias an analyst can have.


  1. cycles says

    You can probably come up with others; stick them in the comments

    What contributed to the success of movies with a woman as lead? I mean, really look at them as something other than anomalies. Do a deep analysis to figure out why people liked them. If producers crack the code, then they could consistently produce successful movies with female leads, and it becomes a no-brainer instead of a huge risk.

    It’s easy to rest on one’s laurels and keep producing male-led movies that make money. There’s a formula. It works. The barrier seems to be the lack of interest in researching alternate formulas to make lots of money. And while that lack of interest may not necessarily hurt Hollywood (they’re gonna make money either way), it hurts the types of people who are not being represented on screen.

  2. says

    Other question to consider: Did the film starring a white woman or person of color, in fact, flop?

    I’m sure we remember how after The Princess and the Frog did “poorly”, Disney decided to redo Tangled and promised never again to make the mistake of thinking women can star on the big screen. But according to Wikipedia, it was “the highest-grossing start to date for an animated movie in December.” In other words, it did better than any other film in a comparable situation (and that’s even ignoring the fact that it opened one week before Avatar), but was still a “failure”.

    And heck, let’s examine the gross revenue for each recent Disney animated feature film leading up to that:
    Home on the Range (2004): $103,951,461
    Treasure Planet (2002): $109,578,115
    The Emperor’s New Groove (2000): $169,327,687
    Meet the Robinsons (2007): $169,333,034
    Atlantis: The Lost Empire (2001): $230,053,725
    Brother Bear (2003): $250,397,798
    The Princess and the Frog (2009): $267,045,057
    Lilo & Stitch (2002): $273,144,151
    Bolt (2008): $309,979,654
    Chicken Little (2005): $314,432,837
    Dinosaur (2000): $349,822,765

    So TPatF ranks the 8th highest-grossing Disney animated feature film in 10 years, but is considered a “failure” because it starred a woman and had the word “princess” in the title.

    • says

      Just FYI, Hollywood only cares about the NET revenue (grosses minus budget). You can get those two numbers easily at Wikipedia for any movie you want, and crunch them yourself. For example, Treasure Planet was a flop – its budget was $140m.

      But most of the movies you picked were indeed quite profitable – as much as comparable white-male led movies – so your argument is valid. I just wanted you to know how they look at it.

      You’re also right that calling TPatF a failure is just… it netted $162m! That’s very good by any standard.

      • says

        Um, yeah. I know.

        I originally added the net revenue for each movie, but (A) I couldn’t find the actual budget figures for two of the them; (B) that would’ve made the comment longer than necessary; and (C) The Princess and the Frog slipped to seventh-highest when considering net revenue.

  3. says

    In our sexist culture, women are for sex first, support second and everything else a distant third. Media roles reflect this. If they make a female superhero, for example, they will emphasize her Sex above all else. If they can’t do this, they will make her Support to a more important, male superhero. Sometimes she will be both. And, like it was said in the Aliens article, being reduced to Sex and Support means they don’t get the respect as full human beings that more well-rounded characters get. Add a healthy does of Madonna/Whore and a nice dollop of the Othering women get subjected to, and you get dehumanized roles that are unsatisfying.
    Of course there are examples that go against the grain, but like you said, people will find any justification to dismiss them.

    Men play Characters. Women play Types of Women.

  4. says

    How about “If it’s based on an existing property, how does it differ from the original, and what did that do to the preexisting audience?”

  5. M.C. says

    Everyone knows that women don’t watch movies and men only watch action movies about other men. Which is why Paramount wanted to turn Twilight into a vampire action flick, according to Mark Morgan: “They would have gone for someone bigger. Boys. More action. I mean, one of their drafts literally had a Korean FBI agent who was hunting and tracking vampires across the coast.”

    I mean, obviously it’s unbelievable that millions of teenage girls would go to the cinema over and over again to watch a simple love story…

  6. SunlessNick says

    If it’s clear that no one wants to see women as leads, yet people claim to love Ripley, Sarah Connor, and Agent Starling, what’s motivating them to lie?

Leave a Reply

Your email address will not be published. Required fields are marked *