Friday, July 3, 2009

Correlation, Cause, And So Forth

Correlation between pirate populations and global warming. Image credit: Church Of The Flying Spaghetti Monster/Wikimedia

Television station KABC published this informative report a couple of weeks ago:

LOS ANGELES (KABC) -- Babies who are breastfed are more likely to do well in high school and go on to college than infants raised on a bottle.
The breastfed kids showed a slight increase in their high school GPA and in the probability of college attendance.

Breastfed kids show increase in GPA

Other than who performed the study, there's no other information of note in the report.

So, do you suppose this is because mother's milk is brain food, or because breast-fed children overachieve to overcome the social stigma?

I suppose it's possible that there's another explanation ...

(h/t Barbara Drescher)


Emily S. said...

There is a part of me that can't help but wonder just who thought of this oh-so-bright idea. Then there's another part of me that's thinking that I don't want to know.

I am not a science whiz by any means, but there are so many factors that could play into this that's it hardly definitive that breast milk makes you smarter or have better grades.

My biggest nitpick is "define slightly". "Slightly" higher GPA is extremely vague. Slightly higher by .1? Or a whole point? There's a big difference. I can't even tell if it's just the article that's being vague and uninformative, or if it's the actual study itself.

Overall though, my reaction is basically: okay, so what? While some children definitely do better in a school or academic setting than others (and therefore not necessarily smarter), if people want their kids to have higher GPAs, they need to study. And actually work for it. It's that simple. Whether you breastfeed your child or not I doubt will make a difference in the long run. If they actually show more concrete evidence down the road, then maybe I'll reconsider.

Cujo359 said...

Hi Emily S.,

Actually, in this particular field of science, I'm no whiz, either. Some things are blindingly obvious here, though.

That Wikipedia link ("another explanation") provides a good overview of what the hazards are of thinking this way. It's always possible that there actually is a connection between the two variables, but it's not necessarily one of cause and effect. They might both be affected by another thing, or the relationship might just be a coincidence. Coincidences aren't unheard of.

What I think the study really found is that there's a slight relationship, which is how correlation is expressed. What difference in GPA that might point to is anyone's guess.

If you want to actually prove that there's a link between two things that display a correlation, then you need to at least show a physical reason why the two should be. Some theory, backed by evidence, needs to explain the relationship. Either that, or an experiment needs to be performed under controlled conditions, with all the variables that might affect an outcome controlled or accounted for.

The KABC article did mention any such thing, which is the other reason why it's useless.

It's even conceivable that there actually is a link between breast feeding and performance in school. There's just no reason to think so beyond there being some statistical correlation.

Cujo359 said...

Sorry, the KABC article did not mention any such thing.

Dana Hunter said...

Well, that led to a long and unexpected blog post, which shall be posted after the holiday. Best illustration ever. And a valuable lesson for all who tend to take graphs, charts and statistics at face value.