Monday, July 25, 2011

Misusing the scientific method


Last year — March 21, to be precise — I posted a long diatribe on this blog on a study published in Social Psychology Quarterly on the influence of IQ on religious, political and sexual behaviors. Then, two weeks ago, on The Impractical Catholic, I wrote a post trying to remind various people that, for all the benefits it's brought us, the scientific method is still very much a human art, and therefore prone to misuse and abuse.

I don't know if Denyse O'Leary of Mercator.net was reading either my mind or my blog, or I was reading her mind, or if it was just fantastic timing. But lo and behold, she brought forth a couple of splendid examples of how the scientific method can be misused to promote an agenda.

On two posts, which you can access here and here, O'Leary looks at a couple of studies which claimed to show a negative correlation between religion and IQ, and fisks them for methodological problems. Although the link to one of the studies is broken at this time, I commend them to your attention; suffice it to say that neither of the researchers in question seems to have heard of a false-cause fallacy.


Stepping outside the actual debate between Christianity and atheism, it's fair to say that just as people can become or remain Christians and even Catholics for many reasons, not all of them good or bad, people have many motives for abandoning or never taking up religion, with just as much (or as little) reason. The only true idiocy of either side is to presume that the other side consists solely of irrational idiots … no matter how tempting it may be at times.

With that in mind, it's befuddling and sometimes exasperating to see social scientists, who are normally predisposed to explain behavior in terms of environment and social pressures, opt for genetic explanations out of the blue when it suits their purposes. Suddenly, race, sex, education, income, class, political environment and other factors no longer bear on the question; outcomes are hard-wired rather than variably programmable.

Consider the college environment. Without painting with too broad a brush, it's safe to say that the arts and science departments at many traditional four-year colleges have developed and even indirectly encouraged an atmosphere of antipathy or even hostility to religiosity; this can exert (statistically) significant pressure on undergrads to conform to the irreligious/antireligious subculture. If the student lives on campus away from her family, or isn't well-instructed in her church's beliefs, or if she's religiously indifferent coming into college, these may all negatively impact her ability to withstand the desire to conform to the prevailing subculture.

Now consider that college is becoming ever more expensive, taking it further out of reach of poor people even as the divide between the rich and the poor grows ever larger. Poor people have less access to educational support systems, such as programs for learning disabilities or dyslexia, and often live in social environments that discourage academic achievement. As a result, they generally lack not only the financial resources to attend traditional four-year schools but the educational background to succeed if they do attend.

Furthermore, they tend to score lower in IQ tests, which presume an even distribution of education resources and thus aren't reliable indicators of native or raw intelligence. This means that they're less exposed to the anti-religious culture of traditional colleges and experience less pressure to conform. Overall, the effect of income on education combined with a subcultural bias in the sciences towards atheism/agnosticism could, if adequately included in a study, explain most of the IQ difference without appealing to the bigoted, self-congratulatory assumption that atheists are by definition intellectually superior.

Another environmental influence is the dominant religion of a culture. One of the studies attempted to compare the average IQ of different nations with the percentage of people in those nations who didn't believe in God. As an example of what I'm talking about, the two dominant religions of Japan are Shintō and Buddhism, neither of which require a belief in God (although Shintō recognizes spirits, or kami). Therefore, to conclude that the smarter people in Japan are irreligious because the median IQ is 105 and 65% of them don't believe in God is to argue from false premisses.[1]

This is not as much of a problem in the West. However, the other study grouped people according to their religious denominations into two categories, "Liberal" and "Dogmatic". As O'Leary points out, not only were the religions divided into these categories by rather sloppy and unsophisticated reasoning, the author made no attempt to find out how regular the peoples' religious practices were or whether they adhered to the core doctrines of their churches; for all he knew, the "dumb" ones could have been the more liberal or more indifferent people.

The bottom line is that we have two studies which are infuriatingly similar to attempts by white-supremacist "scientists'" to prove the intellectual inferiority of African Americans. Had, in fact, Helmuth Nyborg and Richard Lynn attempted to present their analyses in the context of race or gender, or tried to prove that Jews were inferior to Christians or Moslems, they would have been tarred, feathered and rode out of academia on a rail.

I will grant for argument's sake that this kind of failure of the peer-review system doesn't occur so much within the natural scientists, and that the social sciences are more prone to confirmation bias and egregious number-juggling when political and social issues are at stake. Nevertheless, they should still serve as reminders that science is a human method, not a self-operating mechanical device for generating knowledge.

In war, politics, debate and science, the gravest mistake one can make is to underestimate your opponent. It's an even bigger mistake to try to write that under-estimation into your textbooks.


[1] I'm embarrassed … would you believe I've been misspelling premiss (basis for an argument) for over twenty-three years?