img.latex_eq { padding: 0; margin: 0; border: 0; }

Friday, May 11, 2007

Here We Go Again

When I read stories like this, I actually want to call up the author and verbally berate them. Evidently, someone decided to look at a large group of school children and see if there was a correlation between math test scores and birth dates. Not a particular correlation, mind you. There was no hypothesis. They weren't looking for a particular pattern, just any relationship would do. Shockingly, they were able to find one. In this case, it's that summer birthdays lead to poor performance, and they actually attempt to pin the blame on pesticides used during this season.

I don't know how many times studies like this have to be shot down before people get it. If you look at a random group of people, some people will do better and some will do worse. That's what randomness requires. We would not expect those good and bad scores to be evenly distributed. They're going to clump up in some months. But they could easily clump up in different months with different samples. Unless you are looking for something in particular to back up a hypothesis, you aren't doing science.

I don't have to know when this journalist was born to know that they don't no a thing about math. They've proved my hypothesis for me.

2 comments:

BSB said...

I agree with you, but as the mother of an August child I would add that IF there is a correlation at all (and I'm obvioulsy not in a position to do any kind of follow-up study myself), it would seem more likely related to school entrance cut-off dates than pesticides. Children with summer birthdays are younger than their counterparts when learning the basic skils.

Tony said...

I happen to be a September child myself, and certainly know what it is like to be the young one in class. If these findings bear out in follow-up studies, which I don't for one second believe they will, school entrance cut-offs do seem much more plausible than prenatal pesticide damage.