Making up statistics for good causes

I was unsurprised to learn that Gary Gates’ careful estimate of the proportion of Americans who are LGBT prompted a huge controversy. The answer is that 3.8% of people self-identify as a member of one of these groups, which is an average across eleven population-based surveys. One common reaction from the LGBT community was anger at such a low estimate; Larry Kramer called Gates a “horse’s ass”. The probably reason is that much LGBT activism has seized on “one in ten” as the true frequency (I first heard it as the share of homosexuals), as a way to emphasize how the importance of the group and to promote LGBT status as part of the mainstream. Those are both noble goals, and worth pursuing, and these pesky facts assembled by Gates threaten the means chosen to promote them.

This is far from the only situation where activists or intellectuals take a noble goal as justifying the means of statistics that are misrepresented or just fake outright. For example, the semi-anonymous Chris of Aspiring Economist deconstructs the claim that one in four women are raped here. It’s probably not true, and the top comment is telling – Angie more or less admits the statistics are a means to an end, rather than facts. And it happens on all sides – in the Gates article above, he talks about conservative groups just ignoring the bisexuals in the LGBT figures to downplay the percentage of people in the overall group.

I’m increasingly obsessed with the way people are misinformed or misled about statistics, by well-intentioned activists and policymakers, due to my own research – lots of people have totally wrong impressions about how HIV is transmitted and the course of the disease once you get it. I’m not sure how intentional the creation of those misconceptions is.

Gates’ whole piece defending the original estimate is carefully considered and worth looking at. Fascinatingly, the share of people who admit any same-sex attraction in their lifetimes is just over one in ten – with a slightly different initial spin, this study might have been heralded by the same activists who denounced it.

7 thoughts on “Making up statistics for good causes”

  1. I remember your Uncle Scott telling me once (yes, he actually spoke to me) that you could take the same stats and make them argue for either polar end of an argument, just by manipulating and interpreting the numbers and stating your case in very strong language. He was right.

  2. I take your point, but perhaps a different comparison example should be used than the rape study – Chris’s article seems to miss (crucially) that the definition of rape is misunderstood (i.e. in my view, any unwanted sex from anyone, versus the image of a stranger going around violently attacking unsuspecting girls), that rape victim status is stigmatized and that women either don’t or wouldn’t self-identify as such for complicated reasons, and that rape is actually not talked about in frank or open terms except in private (maybe). Highlighting the frequency of rape was a way for the anti-rape movement to try to get the topic more out in the open and more talked about… and it is, as a policy matter, but not so much at the level of individuals, at least in my experience.

    1. Those are all good points, and I don’t want to downplay them, but what I’m trying to get across is exactly what your last sentence says: “Highlighting the frequency of rape was a way for the anti-rape movement to try to get the topic more out in the open and more talked about”. Chris is just looking at the same study that activists cited, to figure out where their figures come from. 73% of the quarter of women who describe themselves as being raped categorized as rape but don’t themselves consider the event to have been raped. There’s stigma going on here for sure, but they’ve already discussed the sexual act and the circumstances.

      I’m not at all opposed to what anti-rape activists have chosen to do. If it works, then it’s probably a good thing – and I think it has been a highly effective strategy. I can see why they want to defend the numbers they are using. But I’m also interested in the prevalence of rape for other practical reasons – if we can’t measure it accurately, how can we see whether it’s gone down (and therefore our efforts are working)?

      A more honest way of doing the same thing would be to point out that rape is under-reported and stigmatized so that we really don’t know how common it is. But that would be less effective, which is probably why it’s not done. What about pointing out that nearly everyone has close friends or family members who have been raped? That seems equally powerful, and I’m sure it’s true.

      1. Pointing out that rape is under-reported and stigmatized has in fact been done quite a lot. Perhaps the message is not reaching the general public as widely as anticipated, and/or perhaps my impression is skewed because I have become more immersed in that and related issues in recent years.

        Anyway, I’m glad you’ve got this blog going! Following along with every post 🙂

  3. Foods like meat, cheese, milk, salt, grains and fish are predominately acidic.
    But rather than scold them into eating their fruits and vegetables,
    I find that being creative works much nicer. cup frozen strawberries, and one
    scoop vanilla protein powder in a blender and blend until smooth and frosty.

Leave a Reply

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