## Saturday, December 13, 2014

### Diversity is also an empty lost-and-found bucket

Way back when, an intellectual giant of our time--nay, THE intellectual giant of our time--kindly described a post I'd put together as "an elegant little regression analysis". It would become standard fare for this audacious amateur, and should continue to be to this day.

In that spirit, then, a few correlations with the percentage of respondents by state in a Gallup poll (via TWCS) asking "If you lost a wallet or purse that contained two hundred dollars, and it was found by a neighbor, do you think it would be returned with the money in it, or not?" who answered in the affirmative:

By estimated average IQ from NAEP scores: +.78 (p-value = .000000000031)
By Ice People population percentage: +.72 (p-value = .0000000034)

It would be crass, of course, to show these same results in their inverse, so you'll just have to use your imagination--or at least a little algebra--if you're the kind of miscreant who is interested in them.

It might be countered, however, that what is being insinuated here is misleading. What about poverty and criminal activity? How do these things correlate with the rate of expected wallet return?

By Supplemental Poverty Measure*: -.83 (p-value = .00000000000016)
By murder rate: -.73 (p-value = .000000002)

Is it that intelligent Ice People avoid crime and poverty while Sun People embrace them? That sounds pretty racist. Perhaps instead a dearth of crime and poverty actually turns Sun People into Ice People, ancestry and all, while an abundance of crime and poverty turns Ice People into Sun People? So many unanswered questions, so much to ponder!

Parenthetically, political orientation isn't nearly as informative as proximity to the Canadian border is. The correlation between the percentage of a state's population that voted for Obama in 2012 and the belief that a lost wallet would be returned by a neighbor is an inverse but quite modest and statistically insignificant .17 (p-value = .251).

* A superior measure of genuine impoverishment than the traditional Census definition that only takes income into account.