3.18.2010

If your n=1, it's not a study

Biostatistics was not my strongest suit in medical school. I'm better at some of the concepts than others. For example, I will happily explain sensitivity vs. specificity, and I'm aces on positive predictive value. On the other hand, when I try to understand the conversion from probability to odds, my eyes glaze over and I have a hard time staying vertical.

However, I can recognize when a study has no power. It's generally best not to come to conclusions too broad when your sample size is unimpressive.

On that note, I give you two writers who are generally smarter than this kind of thing. First up, the estimable E.D. Kain at True/Slant, discussing the appalling case of a couple who let their real infant starve to death while "raising" a virtual one on one of those Second-Life-like games:
[O]ne feels almost guilty participating in this virtual world when a story like this surfaces. What other ways could we be spending our time? How may we be neglecting those we love? The average high school student spends five and a half hours a day in front of a screen. This is increasingly true of all age demographics. It’s eerily reminiscent of Ray Bradbury’s Fahrenheit 451 – but in a sense, even more frightening for its lack of anything really sinister. Nobody is out burning books. We’re just creating a world in which they are increasingly irrelevant. And in which family, community, and even our children are becoming increasingly irrelevant.

Then there's Rod Dreher on obesity, enraged by the case of a hugely fat woman who strives to gain yet more weight:
This revolting Donna Simpson person weighs 602 pounds, and is trying to get
to 1,000 pounds. Why? She makes her money with a website in which pervs (like
her kinky boyfriend) pay to watch her shove food in her mouth and jiggle around.

"I love eating and people love watching me eat," she says. "It makes
people happy, and I'm not harming anyone."

[snip]

If... most everybody who is obese can't do anything about it, why is it
that obesity rates have skyrocketed over the course of a single generation? Look
at this map.
Obesity like this isn't something that just happened. We are
eating too much, we're eating the wrong kind of food, and we're not exercising.
Of course some of us will have a more difficult time than others controlling our
weight, owing to our genetic inheritance, or environmental factors (e.g., it may
be harder to exercise, or to access healthier food). But in the main, I simply
don't buy that obesity is something largely beyond the individual's ability to
control.

First of all, I'm sure Elizabeth has plenty to say about the question of obesity and free will. But that's not really where I'm going here.

Every so often, someone will seize upon a particularly horrifying or eye-catching story and use it as a jumping-off point for a broader social commentary. However, the truth is, neither of these stories say anything meaningful about anything other than the particular people in it. Perhaps we spend too much time online, and obesity rates are problematic, whatever the cause. But the deeply disturbed people in these two cases are worthless as proxies for society at large.

Both Dreher and Kain doubtless had the opinions they expressed before they read about these people, and the stories were merely the impetus for posts that were already brewing in some murky sulcus or another. But it weakens what might otherwise be stronger commentaries to predicate them on such outliers.

6 comments:

  1. Totally agreed, and the one woman who wants to be obese of course says nothing about the gazillions who don't want to be, but are.

    Something about the Korea baby story stuck out at me - the infant was premature. Depending on how premature, and how impacted was the infant, this may have nothing at all to do with a family ignoring their kid because they loved a video game, and everything to do with wanting to escape dealing with a severely disabled child. Not that that is an excuse! But the lesson to be drawn may be entirely different than the one Kain is frawing.

    ReplyDelete
  2. Oops, I did it too! No lesson to be drawn at all, probably. Just meant that what Kain thinks was going on psychologically with these people might be very different than what was going on!

    ReplyDelete
  3. Good, so can we all agree that no matter how much we admire Pres. Obama for using terminally ill patients as props in his speeches, it doesn't add anything meaningful to the HCR debate?

    ReplyDelete
  4. first off, ignore the troll, who seems to miss the entire point, it is not case studies it is these specific case studies that are flawed.

    secondly, I think case studies are perfectly acceptable way to gain insight, with statistics peoples eyes glaze over, the pathologies that occur for a significantly overweight child can be illustrative. You simply don't see many fat Koreans (especially north, except the Dear Pumpkin...oops Leader, the psychopath Kim Il Jung, who has slimmed down considerably due to the fact that he is thank God dying). Or many fat Cambodians or Vietnamese or even Japanese (sumo notwithstanding).
    I think the example Dreher used was a bad one.
    The American lifestyle helps make Americans fat. Change the lifestyle and I promise you you can take off the weight and keep it off.
    The only question for people is is it worth it?
    When I lived in Asia I lost 30 pounds and kept if off, I only gained weight back when I moved back to the West, and 15 years later am still lighter than I was I first moved there. Contrast that with a person who was heavy then in the states and how much more they are now.

    charo

    ReplyDelete
  5. GJ, you need to update your material. Your schtick is getting old.

    ReplyDelete
  6. Gee, you are right! Spend a year yammering on about the same old thing that people have already made up their minds about, and it gets old.

    Oh, wait, you aren't talking about Pres. Corpse-man! You are talking about me. As Emily Littela would say, "Got Votes?"

    ReplyDelete