Saturday, July 23, 2016

Agnostics more intelligent, trusting than atheists

In a post last month, Vox Day made a couple of assertions: 1) Agnostics are more intelligent than atheists, and 2) Atheists don't trust other people because they're projecting their own lack of integrity onto others.

I've looked at the first issue before and recall the ordering, from most to least intelligent, going agnostic-atheist-uncertain believer-firm believer, but am not able to find the relevant post, so we'll recreate it here by tapping old faithful, the GSS.

For contemporary relevance, all responses are from 2000 onward. To avoid racial confounding only whites are considered and to avoid language fluency issues only those born in the US are considered. Mean IQ, as converted from wordsum scores assuming a white mean of 100 and a standard deviation of 15, by belief in God (n = 4,734):

On GodIQ
Uncertain believer100.8
Firm believer98.4

Parenthetically, "firm believers" make up more than half of the respondent pool, which is why the results appear at first blush to skew above an average of 100.

Without passing comment on the speculative reasons why it might be the case, the second assertion can be evaluated by looking at dichotomous responses to the question, "Generally speaking, would you say that most people can be trusted or that you can't be too careful in life?" Using the same parameters as above, the percentages who say that most people can be trusted, by belief in God (n = 5,294):

On God%Trust
Uncertain believer40.9
Firm believer38.4

If being trusting of others is considered desirable, the ordering runs the same as it does for intelligence. Agnostics are more trusting and more intelligent than atheists, but atheists are more trusting and intelligent than theists (as gauged by the metrics employed here, anyway).

By definition using natural methods to discern, discover, or comprehend aspects of the supernatural are likely to come up short. Some people interpret things as being revelatory, and many more benefit from--and realize that society benefits from--the aspects of unity and teleology derived from trying to make sense of the supernatural. I don't get much out of that myself and find that stoicism applied to the 21st century works better for me, but everyone's mileage will vary. Find what gets you the farthest and ride it.

GSS variables used: YEAR(2000-2014), GOD(1)(2)(3-5)(6), BORN(1), WORDSUM, RACECEN1(1), TRUST(1-2)


Anonymous said...

The trust variable basically correlates strongly with Wordsum and Educ anyway, so you'd expect that order of trust in the absence of any effect from religion on trust.

szopen said...

Tadaaa! (sorry, but after being called liar by VD it's hard not to feel a bit of satisfaction after reading a bit of data like this)

The Z Blog said...

I would be interest in seeing the diversity in the groups. For instance, within the 1 SD above 100, what is the distribution of believers, agnostics, etc. My hunch is the left side of the bell curve is dominated by believers, but the right side would look pretty much like society a a whole.

I'm skeptical of the trust part. I'd expect a higher degree of clannishness among believers, but I don't think that's what they are measuring here.

Anonymous said...

Though, to be fair he said atheists trust other atheists less and that's harder to test. He could probably claim that this result is because the atheists all trust the agnostics and religious people (since he seems like the kind of guy who argues that way, as part of his general shitlord persona posturing).

Although that would be a weird result because it would imply that atheists trust the religious more than the religious trust other religious, which would be very odd. (Perhaps he'd reformulate that the religious distrust other religious because they know that they're repenting sinners, and likely to sin again ;) ? ).

There are reasons you might expect atheists to lie less though. Autism quotient is associated to atheism, and then autistic folk tend to have difficulty telling lies (lack of understanding of other people's mental processes, or the gains of telling lies and a general commitment to truth telling) and tend to be quite literal about how they view other people's statements. Depends on if that's the dominant dynamic for autism.

Most religious people tend to think atheism is about an emotional rejection of God, or of moral constraints, or of the social community of a church. That's the "Depraved Atheist" model. If that's right then maybe atheists might be more dishonest / less pro-social.

But for the most part most atheists I've spoken to just don't find that God a plausible model of reality, cf. Razib Khan for an example of an atheist who basically just never found God plausible, and who is pretty stringent about truth telling and learning the truth over bullshitting. That's the more "Innately Naturally Scientific Atheist" model.

Thales said...

You need to control for another variable: education. Those with more education tend to be less religious.

And there is some correlation between educational attainment and intelligence.

Audacious Epigone said...


Only if the intent is to try and discern a causal relationship between belief in God and intelligence, educational attainment, or trust. That's pretty ambitious for a wide but shallow survey like the GSS. We can do a multi variable regression analysis on these things to see if we can tease out an independent relationship between belief in God and either IQ or trust, but the purpose in this post was to evaluate Vox's assertions in the real world.


I'm not much past the point of being a VD virgin. I'd stupidly pigeon-holed him as a video game guy, a world that I'm ten years removed from, but he's clearly a lot more than that.


Among those who got a perfect score on the wordsum test, a plurality are firm believers, about 40%.

Atheists and agnostics probably tend to be more outbred than the religious, even in outbred societies like those of Northwestern European descent. That's probably what the TRUST variable is essentially measuring--how trustworthy complete strangers are perceived to be, not how trustworthy cousins are or even people in your church or local softball league are.


Right. It's worth trying to distinguish between atheists as people who are simply not functionally theistic and do not feel anything supernatural at any level like Razib Khan and John Derbyshire and those who are anti-theists like Richard Dawkins or Sam Harris.

Anonymous said...

This reminded me of a work by Edward Dutton that I wanted to get around reading, actually:

I remember G√ľnther Anders throwing a tantrum when people called themselves "agnostic" in his "Ketzereien". A leftist, what do you need to know more? (Interestingly, he was the son of Wilhelm Stern -- yep, the inventer of the concept of the intelligence quotient.)

JayMan said...

Interesting. Though I'd imagine the difference between atheists and agnostics here aren't statistically significant.