TIA Tuesday: More “fun” with statistics.

Have you ever been in a conversation with someone and seen them change their minds about something right in the middle of explaining it to you? If you have, then you might experience a bit of déjà vu when you read the following from Vox Day’s discussion of Sam Harris’ Red State/Blue State argument. See if you can tell how Vox’s attitude changes between the first excerpt, from page 115ff of TIA, and the second from just four pages later:

There are several layers of problems with this apparent proof of Christian immorality. The first is that political identity is a very poor substitute for religiosity… [I]t is absurd to credit all of the supposedly law-abiding behavior of “blue” voters to the 16 percent of them who lack religious faith…. If this isn’t sufficient evidence of the foolishness of trying to equate Democratic votes with atheism, the ARIS 2001 survey reported a higher percentage of Democrats among Jews, Baptists, Catholics, Methodists, Pentecostals, Episcopalians, Buddhists, and Muslims than among the not religious, of whom only 30 percent reported a preference for the Democratic Party…. So while the data may be striking, the argument based upon it can only be described as strikingly stupid.

and

What is much more important is the way in which using the more accurate county data demonstrates that Harris’s conclusions are precisely backward. Thirteen of the twenty-five safest cities are situated in RED counties and twenty-one of the twenty-five most dangerous cities are located in BLUE counties. This provides precisely the information that Harris claimed to have sought in vain, it is definitive proof that the social health of Red America is significantly superior to that of Blue America by Harris’s own chosen measure.

Did you catch that? On page 115 and following, Harris’ technique for correlating social health with Christian conservativism is “strikingly stupid.” Yet just four pages later, it’s a “definitive proof” (and not just “a sign,” as Harris called it). What made Vox change his mind so completely? Simple: he found a way to make the voting record say something that he wanted it to say.

Let’s get one thing clear right from the start: as seductive as it might seem to those who think it supports their side, the Red State/Blue State argument is utterly bogus. As Vox correctly observes (before realizing it could be used to support his own claims), it’s wrong on so many levels it’s hard to know where to begin. Vox already pointed out the non-equivalence between Republicans and Christians, and between Democrats and atheists. Add to that the influence of 9/11, and the fact that many of those who voted “red” were not even Republicans, plus the dirty tricks of the Swift Boat Veterans for “Truth,” and it’s doubtful that the 2004 results are even a reliable indicator of the actual liberal/conservative ratio, let alone anyone’s religious affiliation.

Likewise, Vox has pointed out the non-equivalence between the population of those who were voters in the 2004 election, and those who committed the crimes that were tallied in the 2005 statistics. He could also have mentioned (though he didn’t) that the Red State/Blue State argument completely fails to take into account other, more significant variables like population density, unemployment, poverty levels, education, social involvement of Christians vs. non-Christians, and so on. This is not to say that it would be impossible to investigate the statistically-relevant variables looking for a positive or negative correlation between “conservative Christianity” (whatever that is) and social metrics like crime rates, divorces, teen pregnancies, and so on. But if anyone ever does do such a study, it won’t be a matter of merely counting the colored polygons on the CNN map of election results.

Vox seems to have been under no illusion that Red State/Blue State was anything but a thoroughly bogus metric, at least when it was being used to deny a correlation between Christianity and social health. He even seems to have held onto that idea as he moves into the “devil’s advocate” portion of his rebuttal.

But just for kicks, let’s pretend that it is not a measure so ridiculously inaccurate as to be completely useless. Let’s imagine that Harris’s metric really is relevant, that an American voter’s 2004 presidential vote truly is indicative of his religious faith, or the lack thereof, and that statewide criminal statistics are a reasonable measure of an individual’s predilection for immoral behavior. This exercise in imagination is necessary, in fact, because only by accepting his measure at face value and examining it in detail can one fully grasp the true depth of Harris’s exceptional incompetence.

The irony here is absolutely delicious: according to Vox, the true depth of Harris’ incompetence is demonstrated by the fact that his methods lead to the conclusion that conservative Christianity promotes good social health! Talk about your shooting yourself in the foot! He starts by setting out to prove that Harris’ method leads to brain-dead conclusions, and ends by showing that the brain-dead conclusion is, in fact, an oft-repeated Christian claim.

Of course, Vox might be tempted to reply that Harris did indeed create a valid measure, and was incompetent only in applying it to the wrong set of figures. But if that were the case, why did Vox begin by (correctly) pointing out that the method is “a measure so ridiculously inaccurate as to be completely useless”? By trying to have his cake and eat it too, Vox ends up ridiculing the myth that Christian influence is a benefit to society as a whole.

I rather doubt that Vox intended to build such an incongruous case against conservative Christianity, and I think that what may have happened is that Vox began his argument with every intention of constructing a classic reductio ad absurdum, and gradually found himself seduced by the results he started getting when he used a different set of numbers. Instead of the intended “absurd” conclusion, he was getting an argument in favor of Christian conservativism, at which point he suddenly stopped mentioning the flaws in the method, and started crowing about the “definitive proof” of the results instead.

And to put the icing on the cake, Vox goes on to accuse Sam Harris of intellectual dishonesty, on the grounds that he should have known about the figures Vox cited in building his own “proof.” Not that he necessarily did know, but that he should have known. Vox has only just finished building a “definitive proof” of the benefits of Christian influence, using a method that he knows is bogus and that he explicitly said was completely useless, and now he’s going to accuse Harris of being intellectually dishonest. Call it what you will, that’s high-grade chutzpah.

Then again, I could be mistaken. Vox has promised to respond to this series as soon as I’m done with it. Let’s see if he does indeed publicly acknowledge that his version of the Red State/Blue State argument is at least as bogus and dishonest as he accuses Harris’ version of being. Anyone want to lay the odds?