William Saletan in Slate notes that “[Chinese] households with [motor] vehicles have an obesity rate 80 percent higher than their peers.” (The link in the article doesn’t lead to a good source.) Sure, that correlation probably has two huge confounding factors: income, as rich Chinese both own cars and eat more; and maybe a bit of selection bias.* Yet it’s nice to think that such a big correlation implies at least a little bit of causation.
Incidentally, elsewhere in Slate “Joel Waldfogel”:http://www.slate.com/id/2148759 points to British evidence that taller people are paid more because they’re smarter, while “Daniel Engber”:http://www.slate.com/id/2132990/ links heavy babies to bigger adults. All of which sets me up as an outlier again: against American male norms, 87th percentile by birth weight, 5th percentile by adult height, and, well, an immodestly high figure on intelligence. Of course, earnings — well, never mind. (Of course, a decades-old British data set also wouldn’t account for Jewish or Asian Americans, two short[er] groups with above-average incomes.)
* This seems more likely in the U.S., where not driving is an anomalous choice. Bicyclists as a group may be skinnier than average, but no one claims that the act of playing football makes one hefty.