New paper and video out: The Negative Religiousness-IQ Nexus is a Jensen Effect on Individual-Level Data: A Refutation of Dutton et al.’s ‘The Myth of the Stupid Believer’

A recent study by Dutton et al. (J Relig Health 59:1567–1579. https://doi.org/10.1007/s10943-019-00926-3, 2020) found that the religiousness-IQ nexus is not on g when comparing different groups with various degrees of religiosity and the non-religious. It suggested, accordingly, that the nexus related to the relationship between specialized analytic abilities on the IQ test and autism traits, with the latter predicting atheism. The study was limited by the fact that it was on group-level data, it used only one measure of religiosity that measure may have been confounded by the social element to church membership and it involved relatively few items via which a Jensen effect could be calculated. Here, we test whether the religiousness-IQ nexus is on g with individual-level data using archival data from the Vietnam Experience Study, in which 4462 US veterans were subjected to detailed psychological tests. We used multiple measures of religiosity—which we factor-analysed to a religion-factor—and a large number of items. We found, contrary to the findings of Dutton et al. (2020), that the IQ differences with regard to whether or not subjects believed in God are indeed a Jensen effect. We also uncovered a number of anomalies, which we explore.

Video version in the bottom: scroll down!

I know what you are thinking: isn’t this Dutton vs. Dutton??

Indeed! The prior paper:

Numerous studies have found a negative relationship between religiousness and IQ. It is in the region of − 0.2, according to meta-analyses. The reasons for this relationship are, however, unknown. It has been suggested that higher intelligence leads to greater attraction to science, or that it helps to override evolved cognitive dispositions such as for religiousness. Either way, such explanations assume that the religion–IQ nexus is on general intelligence (g), rather than some subset of specialized cognitive abilities. In other words, they assume it is a Jensen effect. Two large datasets comparing groups with different levels of religiousness show that their IQ differences are not on g and must, therefore, be attributed to specialized abilities. An analysis of the specialized abilities on which the religious and non-religious groups differ reveals no clear pattern. We cautiously suggest that this may be explicable in terms of autism spectrum disorder traits among people with high IQ scores, because such traits are negatively associated with religiousness.

OK, so who is right, Dutton or Dutton? Turns out it is Emil.

The brief summary of the problem with the prior study is that the Jensen’s method is biased towards finding null (nothing) results when the precision of the study is low. The precision is low when the relationship being studied is weak, such as a small IQ gap between different Christian groups. So this is basically why the prior study produced null results. We show this is the case via simulation:

The true correlation (r) on Y axis is always 1.00 here, yet we see with small gaps (the d value in the box headers) and smaller sample sizes (X axis), the results (results of each simulation) average closer to 0 instead of the true value of 1. Downwards bias in this method!

To settle the issue we used a well powered (high precision) design with the Vietnam Experience Study (VES) dataset (which you can download here), and show the usual g-loaded pattern / Jensen effect is seen when precision is high:

Alternatively, one can use the items instead of tests, but results are congruent:

The outlying item in the top is a question about what Genesis (first book of the bible) is about, which, unsurprisingly, religious people get right more often than would be expected by their intelligence.

We also show that these results are congruent with those based on modern approaches, including local structural equation modeling (LSEM) and differential item functioning (DIF):

The relationship is also not due to obvious demographic confounds such as race and age, or even social status measures like income and education. The latter might be somewhat surprising. There cannot be a sex difference bias here because this military dataset only has men. Regression table:

Look at the beta value in the top (G), which is about .21 to .22 throughout despite controls or no controls or subsetting to white people only.

The most interesting finding was that it matters how one measures religiousness specifically. We had 12 questions to use:

As it happens, some of these actually show slightly positive relationships to intelligence! This is despite all these items being highly correlated and loading on a general religiousness factor (which was used above). Odd! It seems the explanation is that items that tap into other constructs as well as opposed to more purely religious beliefs are the ones with odd patterns:

Each line is a set of 3 models with different outcome items. The most positive relationship was seen for the item about going to church every week. It seems the prosocial aspects of this result in some intelligence association despite the religious activity. Maybe some smart prosocial atheists attend a lot of church?

Bottom line is that atheists really are smarter, and this is related mainly to their general intelligence, not the other factors, and this is related mainly to the beliefs themselves, not the various religious behaviors or related beliefs such as intolerance for heathens.

Video walkthrough