This Little Piggie Did Calculus, This Little Piggie Stayed Home

Scientists in Great Britain have just completed a study that shows a correlation between finger length and math ability. When I first came across this article, I was all set to be annoyed. I thought for sure that it was going to be another adventure in data mining, but I was wrong. The finger length is not the proximal cause, but rather a result of prenatal hormone surges. Testosterone, which is known to affect growth the right half of the brain where spatial reasoning is located, also causes the ring finger to be slightly longer than the index finger. A digit ratio of less than one correlates with left brain development, or language skill.

Now from the little bit I've read about this, I can find no obvious fault in the methodology. There was a predetermined hypothesis, evidence was gathered, and a conclusion was made. I do find major fault in the reporting. Here is how the article begins:

Parents may be able to predict how well kids will do in math and reading by measuring their fingers, British scientists claim.

Now I understand that journalists have to front load the hook. Their goal is to get people to read the story. Fine. But shouldn't they feel some obligation to present the material in a responsible manner as well? Most researchers don't get the opportunity to speak directly with the public and are forced to use the media as a go-between. For most people, the only science the read about is in their local newspaper. So placing a statement like this right up under the byline can be very dangerous. Sure the author used passive tense and employed words like may and claim to slightly erode confidence in the conclusion. But they should know that when it comes to predicting test scores, parents will latch on to anything remotely scientific, especially if it comes from the British, because we all know how smart they are. I mean, just listen to them talk.

Here is how those same British researchers present their conclusion in the final paragraphs;

Finger length is by no means a measure of intelligence or ability, Brosnan stressed.

"You may well look at a finger and say, all other things being equal, they may well have a leaning towards mathematics," Brosnan says.

"Or if they have a child who has a digit ratio that suggests very good verbal skills and a struggle with mathematics, this might suggest that, to explain mathematics to that child, you use a lot of verbal strategies and you cut down on the graphical, visual information."

By their own admission, the results of the study are less meaningful than this article would imply. The scientists appreciate the power that their work might have on desperate parents and rightfully suggest that not too much be read into it. These results are interesting at best. There is by no means a single definitive predictor of skill or intelligence.

I'm also very sensitive to boy vs. girl related math myths lately. After reading Rebecca's post earlier in the month, I have been thinking a lot about all things sexist in mathematics. Now to be fair, this was not a gender study, per se. It was a study of androgen surges, and because both males and females have both testosterone and estrogen, a continuum is possible which matches the evidence. Generally speaking, if the measured difference between groups is less than that within groups, then the phenomenon is fairly impotent. So if there is more variance between girls scores than there is between the average of girls and boys, than we can't really link the difference to gender. But that's not what this study is doing, so I can still only argue with the reporting.

Numerous studies have been done lately that conclusively demonstrate the harmful affects societal stereotypes have on test scores. Girls are known to do worse on a math test if you tell them before hand that boys do better. It's a horrible reality, and as a future male teacher (currently male, not yet a teacher,) I desperately want to avoid contributing to this problem. I wish that the authors of these articles felt the same way.