Because most philosophies that frown on reproduction don't survive.

Thursday, April 02, 2009

Adventures in Regression

Megan McArdle provides a perfect example of why a turning a fool loose with a regressiion or graphing tool can get you all sorts of strange places.
I've seen a lot shakier plots used to justify some sweeping conclusions, and if those were justified, well, then I'm forced to conclude that Mexican lemons have improved highway safety a great deal. The vitamin C, maybe? The fragrance? Bioflavanoids?

This is particularly tricky when you bring time into it, because things trend--as we get richer, we buy safer cars, get better emergency rooms, etc. We also import more lemons to make our chi-chi cocktails and lemon meringue pies. Overlay the two, and you've got a hell of a causal relationship.

But I expect that four years from now, we'll still be having the same conversations with proponents of "cancer clusters" and Democrats convinced that they can scientifically prove that Democrats are better for GDP by doing ham-fisted regressions of Democratic presidencies with a few tightly correlated economic variables. What's the mechanism? What makes electric power lines cause cancer, but not the earth's vastly more powerful magnetic field? What policies did Harry Truman and Bill Clinton have in common (but not with Richard Nixon) that caused this marvelous confluence? Well, maybe we don't know the mechanism exactly, but never you mind: just look at that bee-yoo-ti-ful correlation!

On less political issues, I find myself dealing with these kinds of "analysis" at work all the time. As in all things, common sense is in order. If there's no reasonable explanation for any sort of causal relationship between two factors, then consider very seriously the possibility that there is none.

I wish I had the years behind me to get away with the Professor's recurring line from The Lion, the Witch, and the Wardrobe: What do they teach them in schools these days?


John Henry said...

Beat me to it. This is an instant classic. Of course, when these graphs are used in politics, the proponent believes they have a reasonable causal explanation (e.g. the stupidity of the other party's policies). And so the graph doesn't really do any work; it's usefulness is dependent upon the acceptance of the same premises which are under dispute.

But this is a useful example of the perils of regression qua regression.

Darwin said...

As Paul Krugman has clearly has clearly shown, lemon imports increase less slowly during Republican administrations.

Jokah Macpherson said...

People used to be good at analyzing regressions for a reasonable relationship but more and more these days they are only average.

Jen said...

Thank you! Since my days as a Sociology major this has been one of my pet peeves. When I got to grad school (public policy school, no less) and heard "grown ups" making crazy correlations I nearly blew my top! We were all required to take Advanced Quantitative Analysis (mainly regression) but I can't count how many times we were back to whether or not there was any reason to assume causality.

Have you read Freakonomics? Makes me think of that book & the "link" between abortion & crime . . .

Donald R. McClarey said...

As Churchill memorably observed, statistics are like lampposts: better for support than illumination.