September 17, 2012

Correlation without mechanism

Sometimes you just know that apparent associations just have to be spurious: the price of tea in China really doesn’t affect NZ violent crime rates, and you’re no more likely to win at lotto if you buy your ticket from a shop that’s sold winning tickets in the past.  There’s no way that one could affect the other.

The Herald has a story on functional foods that’s very similar to one I commented on in May.  Again, before even looking at the claims, how could it be true that the shape of a food was a guide to its function? It would be necessary for God to have set it up that way (and, apparently, without bothering to tell anyone).

Match foods to parts of the body for optimum health benefits:

Also, let’s look at some of the examples

1. Healthy Bones: Bony-looking foods such as rhubarb, rich in vitamin K, and celery, rich in silicon, are both good for bones and healthy joints.

Vitamin K was once suggested as involved in bone health, based on observational studies, but a large randomised trial didn’t find any beneficial effect.  Also, although raw rhubarb has moderately high Vitamin K levels (about half as much as cooked broccoli), cooked rhubarb has much less.

3. Sight for Sore Eyes: Slice a carrot and the round circle will show a likeness to an eye, complete with pupil and iris. They contain beta-carotene and antioxidants, both helpful for eyesight issues.

Parsnips have the same shape (as do many roots), but lack the carotenoids.  Orange/golden kumara and pumpkin, which don’t have the eye shape, do have the carotenoids.  And I’m sure you could think of other body parts with more similarity to the carrot than eyes…

5. Round Fruit: Lemons and grapefruit with limonoids and vitamin C are believed to be helpful in preventing breast cancer.

How about quinces, mangoes, coconuts, and melons, which all have historical and cultural support for looking like breasts (or, if we follow Biblical authority, perhaps venison is the relevant food)

avatar

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »

Comments

  • avatar
    James Guthrie

    The whole ‘carrots for good eyesight’ thing was apparently kicked off as a WWII propaganda/misinformation scheme by the British, who were trying to throw the Germans off the scent with of their newly developed radar technology.

    British Intelligence didn’t want the Germans to find out about the superior new technology helping protect the nation, so they created a rumor to afford a somewhat plausible-sounding explanation for the sudden increase in bombers being shot down. News stories began appearing in the British press about extraordinary personnel manning the defenses, including Flight Lieutenant John Cunningham, an RAF pilot dubbed “Cats Eyes” on the basis of his exceptional night vision that allowed him to spot his prey in the dark. Cunningham’s abilities were chalked up to his love of carrots.

    More at…
    http://www.snopes.com/food/ingredient/carrots.asp

    12 years ago

    • avatar
      Thomas Lumley

      Good point. That’s what made it famous (but it’s not the only basis, as the Snopes article says).

      It’s certainly true that eating carrots is not going to help unless your diet was previously deficient, which is unlikely in Herald readers (although quite common worldwide).

      12 years ago

  • avatar
    James Stanley

    This sounds like sympathetic magic as described out in J G Frazer’s “The Golden Bough” some 100+ years ago — it looks-like-this, so it will work-on-this. Plus ça change… new jargon can work wonders for pseudo-science.

    12 years ago

    • avatar
      Thomas Lumley

      At least sympathetic magic has a rationale: if you believe in magic, then you already believe that human perception and will can make or reveal connections between physically unrelated things.

      Magic is a sufficient explanation for sympathy. Nutrient content, not so much.

      12 years ago

      • avatar
        James Stanley

        Fair enough.

        I was also left wondering from the carrot example in the Herald article how much vegetable topiary is acceptable while keeping the link alive between food shape and function…

        12 years ago

  • avatar
    Scott Macleod

    Strange to see two references to religious mythology on a statistics website. I would expect interpretations to err on the scientific side.

    12 years ago

    • avatar
      Thomas Lumley

      Not at all. Statisticians often have to work in fields where they don’t personally have expert knowledge, so being able to work out the observational consequences of other people’s assumptions is a key skill. If someone wants to design an experiment to test the benefits of homeopathy, or magic, or vitamin C, we’re happy to help. They are likely to be disappointed in the result, but that’s not our problem.

      In this specific case, a sympathetic-magic explanation for the food resemblances would still be wrong, but it would be much more logically coherent than a materialist explanation.

      12 years ago

    • avatar
      James Stanley

      Also, statisticians enjoy spreading around random bits of knowledge.

      (which is a roundabout way of saying that my post wasn’t about suggesting that magic might be a better explanation of these findings, but rather that I picked up similarities between the logic as presented in the Herald article and an anthropological perspective on how “magical” phenomena might work.)

      12 years ago