Full disclosure: I think Evolutionary Psychology is a pseudo-science. This isn’t because the field endorses a flawed methodology (relative to the norm in other sciences), nor because they come to conclusions I’m uncomfortable with. No, the entire field is based on flawed or even false assumptions; it doesn’t matter how good your construction techniques are, if your foundation is a banana cream pie your building won’t be sturdy.
But maybe I’m wrong. Maybe EvoPsych researchers are correct when they say every other branch of social science is founded on falsehoods. So let’s give one of their papers a fair shake.
Ellis, Lee, et al. “The Future of Secularism: a Biologically Informed Theory Supplemented with Cross-Cultural Evidence.” Evolutionary Psychological Science: 1-19.
ABSTRACT: For over a century, social scientists have predicted declines in religious beliefs and their replacement with more scientific/naturalistic outlooks, a prediction known as the secularization hypothesis. However, skepticism surrounding this hypothesis has been expressed by some researchers in recent decades. After reviewing the pertinent evidence and arguments, we examined some aspects of the secularization hypothesis from what is termed a biologically informed perspective.
Biology has something to say about religion?
Based on large samples of college students in Malaysia and the USA, religiosity, religious affiliation, and parental fertility were measured using self-reports. Three religiosity indicators were factor analyzed, resulting in an index for religiosity. Results reveal that average parental fertility varied considerably according to religious groups, with Muslims being the most religious and the most fertile and Jews and Buddhists being the least. Within most religious groupings, religiosity was positively associated with parental fertility.
Muslims?! Oh dog, did Breitbart launch their own journal?
While cross-sectional in nature, when our results are combined with evidence that both religiosity and fertility are substantially heritable traits, findings are consistent with view that earlier trends toward secularization (due to science education surrounding advancements in science) are currently being counterbalanced by genetic and reproductive forces. We also propose that the inverse association between intelligence and religiosity, and the inverse correlation between intelligence and fertility lead to predictions of a decline in secularism in the foreseeable future.
No, this isn’t fair! This MUST have been a crank paper published in some shady journal. I mean, c’mon, its two authors are “semi-retired.” Time for a detour.
No Cranks Allowed
Aha! “Evolutionary Psychology Science” is a whopping two years old, and has a similar name to the journal “Evolutionary Psychology.” This smells like a predatory journal with no real interest in peer review or solid science. Who’s the editor?
…. Oh. Todd Shackelford runs an EvoPsych lab at a major university, has been an author on hundreds of papers and is heavily cited, shared a credit with David Buss a few dozen times, is editing a thick reference on EvoPsych, and even is the editor-in-chief of “Evolutionary Psychology.” Either this paper was published in a legit EvoPsych journal with sufficient peer-review, or a big scandal is brewing.
Words Have Meanings, People
Shit. I actually have to fisk this paper, don’t I? All right, let’s get on with it…
As the term will be used here, secularism refers to the idea that as mankind’s scientific knowledge continues to grow, thereby explaining more and more of the natural universe and the evolution of life, supernatural (religious) explanations will gradually fade into history (Berger 1967; Bruce 2002; Wallace 1966). Other uses of the term have to do with simply keeping any given religion from dominating governmental policies.
Wow, these authors can’t even define “secularism!” It has nothing to do with the death of religion, nor the removal of religious dominance, it is simply acting independently of religion. Policy should be decoupled from any particular sect’s beliefs, in favour of a broad consensus. A religious group is free to monopolize a secular government, so long as they represent their constituents instead of their holy texts. Their second definition is more on-the-mark, but the authors aren’t fans of it.
Without questioning the legitimacy of this latter use of the term, it is not what was meant by those who first began using the term in social science. In particular, several nineteenth century French and German sociologists such as Saint-Simon, Comte, Marx, Durkheim, and Weber all argued that all phenomena, including human affairs, were natural and could be ultimately understood without assuming any supernatural guidance or intervention (Gorski 2003, p. 111).
That’s naturalism, an entirely different -ism! They only intertwine when a consensus view of science is invoked. “Naturalism” tells us human beings are changing the climate on Earth; “secularism” ignores some Christian sect’s beliefs and sets environmental policy to benefit both current and future generations. More generally, secularism is violently agnostic to religion, while naturalism is accidentally atheist.
While correct about the shift to naturalism, Lee Ellis and Anthony Hoskin seem to think that everyone on that list wanted to wipe religion out. Saint-Simon is a glaring counterexample.
His major work, Nouveau Christianisme (1825), announced that the world had arrived at the crisis, predicted by the Old Testament, which was to end in the establishment of a truly universal religion, the adoption by all nations of a pacific social organization, and the speedy betterment of the condition of the poor. Saint-Simon attempted to clear away the dogma which had developed in Catholicism and Protestantism, and to reduce Christianity to its simple and essential elements.
From the mid-1700’s to the mid-1800’s, Europe was revolting against corrupt and powerful elites, including Christian churches. There was a strong desire to knock them down and tear out useless dogma, and the rapid advancements in science and industry suggested a non-scriptural source of truth: science and reason. Rather than toss out Christianity, most intellectuals wanted to retool it into Deism. This isn’t a move away from religion unless you have a narrow view of what “Christianity” “religion” means.
And it’s a bit disturbing that the prior views of long-dead intellectuals are considered relevant today. Astronomy developed from astrology, and famous astronomers like Kepler earned a living from dishing out fortunes. That says nothing about how scientific astronomy is today. It’s also worth remembering the scientific theories that have been around a century or more: General Relativity. Evolution. Science is all about self-scrutiny, so if “secularism” had hung around for over a century in sociology then either A) that theory is on solid ground or B) sociology isn’t a science. EvoPsych researchers fall more into column B.
A Method To The Madness
Wait, I’m only four sentences in? A proper fisk is going to take centuries. Lemmie shave off some time and jump right to the Methodology section.
Each of the seven religiosity questions were measured by asking participants to respond using an 11-point scale ranging from 0, meaning “not at all”, to 10, meaning “the most extreme degree possible.” That’s quite an expansive scale; could you tell a 7 from an 8? Still, I did my homework, and while you don’t gain much by adding more choices the only real harm seems to be to response rates.
The authors almost get props for going outside of North America, as a guard against a very WEIRD cohort, Ellis and Hoskin canvassed people in Malaysia, in addition to their US dataset.
No actual props have been earned, though. To quote their study,
Because our sample predominantly consisted of college students, no claim can be made regarding the representativeness of either sample.
Yep, they only sampled college students. One study found that about 67% of psychology papers published in the USA only sampled undergraduate students, and 80% of those published outside the USA did the same. It doesn’t help that colleges outside the US and Europe are inspired by universities in the US and Europe, keep in regular contact with them, and are also biased towards the young and wealthy. It’s about as WEIRD as you can get outside of official WEIRD territory. Even the authors fess up to the problem!
In fact, Ellis and Hoskin rattle off five flaws in their own study: the aforementioned exclusive use of college students, a lack of childless cohorts “that could substantially alter some of our conclusions,” comparing geographic trends instead of changes over time, a simplistic view of religiosity when other studies have shown it to be more complex, and no measure of a variable that’s critical to their theory. Humility is great, but this list is buried on pages 9 and 10, well away from the abstract and after the main Result section. It’s not mentioned in the abstract, Methodology, nor the Discussion. Few people are ever going to see this list, so the show of humility rings hollow.
Let’s forget about all these flaws, though, because there are sooooo many more to come.
78.9% (1315 + 666)
The Malaysian cohort is dominated both by Muslims and high fertility rates, while the US sample is dominated by Christians and low fertility rates. How much of that is due to cultural differences that are independent of religion? It’s tempting to point to the relative fertility or religiosity of Christians and Muslims within each country, but either way we’re relying on a small sample. This wouldn’t be bad if the authors kept each country separate, but instead they mush them all together for the final analysis. The data on American Muslims is swamped by that of Malaysian Muslims, Malaysian Christians are swamped by American Christians, and culture becomes hopelessly cross-correlated with fertility and religiosity.
The authors make a weak attempt at controlling for this:
Some of our analysis included controlling for parental social status. To measure parental social status, each student respondent was asked to report the years of education for both of their parents and to rate their family’s income on a scale from 1 (meaning “very low”) to 10 (meaning “very high”).
A subjective metric of wealth won’t cut it. A Malaysian who feels rich probably earns substantially less than a US citizen, and that could interact with health care costs or financial stability to bias the stats. Give me a bracketed income instead, adjusted for purchasing parity.
Is contraception more expensive in Malaysia, or abortion more restricted? Then the fertility rate will be higher, independent of religion. Do Malaysian families pool their income to support one kid’s education, gambling they’ll return the favor? Then their income estimates will skew downwards, falsely inflating the fertility-religion correlation. Ellis and Hoskin never consider these factors, which could torpedo any of their findings.
Oh right, their findings. With all these flaws in play, they must look spectacular.
Nope! I’ve converted the listed correlation coefficients to Cohen’s d, and in this form they’re small. Provide me with the fertility of a Catholic person, and I’ll guess their religiosity with no better odds than 55%; in the general case, my odds improve to 60%. As I pointed out earlier, small correlations are actually quite common due to a mix of messy biology and poor statistical methods, and paired with large sample sizes the net result is to make everything “significant.” Most sociology studies have sample sizes between fifty and a hundred; this one had 4,569.
How significant were the findings? Ellis and Hoskin never give numbers, they do not list any p-values or other measures of significance. They also never give error bars, confidence intervals, nor credible intervals. The omission is strange, since they talk about which categories reach statistical significance, and some of their numbers have asterisks floating nearby. The latter is a common shorthand for significance; one star typically means p < 0.05, two is p < 0.01, and three p < 0.001. These omissions make it difficult for other researchers to add to their work, and may be hiding weak results.
You might be wondering why the overall correlation was a lot larger than the ones for Christians or Muslims. The answer is more obvious when you chart all their categories.
The overall correlation is being driven by outliers! Only one group had a greater correlation than was found overall, Jewish people, and the groups closest to the overall figure were the “nones” and “no religion given.” Yet if you tally up those three, and add the nearby Hindu group in for good measure, you account for less than 13% of the total sample. None of the first three are good representatives of religious people, thanks to the high number of secular Jews (roughly 46% of Jewish Israelis call themselves secular, for instance) and religious non-believers, so it’s disingenuous to generalize from them. If anything, this data shows a correlation between non-belief and fertility!
Padding Out The Runtime
But I get the impression Ellis and Hoskin weren’t in it for the data. Read a few scientific papers, and the basic format gets burned into your brain: Abstract (a summary), Introduction (what’s the problem?), Related Works (who else has looked at it?), Methodology (what are you gonna do and how?), Results (what happened), Discussion (how the results fits with your theories), Conclusion and Future Work (summary and/or what’s next), Citations. The Discussion/Conclusion parts are usually quite thin, a page or so, since by then you’ve had ample explanation of the hypotheses in play and what they mean. If you have a good memory, you remembered that I said the Results section ended on page nine.
What I didn’t mention is that this is a nineteen page paper. And that as of the Results section, they followed the standard format.
What’s in the next ten pages? I estimate pages 15 to 19 contain 160 citations, roughly four full pages worth. That is pretty obscene, I can’t remember the last time I read an ordinary paper with more than 40. My peak is around the 190 mark, but that’s because a) my document was a long and wide-ranging summary of other work, b) I was working outside my area of expertise (computer science) and thus had a higher evidential bar to clear, c) I was expecting a hostile audience who would be skeptical of my claims, and d) I’m a bit of a perfectionist. This many citations in a non-review is a red flag; the authors knew their hypothesis was weak, and tried to prop it up by stuffing in citations.
Roughly one of those ten pages is devoted to Muslims and their growth rate.
While there may be multiple causes of high Muslim fertility, we believe that one important factor is the unusually high degree of religious fundamentalism among Muslims, higher in fact than any other religion (Ellis in submission; Westoff and Frejka 2007). […] Because Muslims are more fundamentalist on average than members of any other religion, Muslims will continue to reproduce at relatively high rate.
Jainism is a small religion, yet their most fervent members do not drink alcohol because they would be responsible for killing yeast and bacteria. Torajans have an annual ritual, Ma’nene, where they exhume their dead so they can clean out their coffins, provide a fresh change of clothes (if possible), and take new family photos; in some cases they keep the dead in their homes and treat them as if they were alive. Sun dances take place over four grueling days, where First Nations dancers forgo food or pierce their skin with hooks in order to receive visions. Various Hindu sects do that too, plus crawl along scalding roads, toss babies, and burn widows. The Aztecs make all other religions look like wimps. Even Shinto can get pretty extreme. And yet Muslims are “more fundamentalist on average” than “any other religion?” I’m not buying it, if only because the data suggests they’re pretty heterogeneous.
One page is devoted to “Trends in Average Intelligence.” It’s as bad as you think.
The decline in IQ at the genetic level in Western countries is evidently due to both greater fertility by low-IQ couples than high-IQ couples and to migration from low-IQ countries to high-IQ Western countries (Nyborg 2012).
Setting aside the racism and crass demographic falsehoods (Hans Rosling is an excellent antidote to the latter, as he shows fertility is rapidly declining to 2-ish kids per couple across the globe), these authors claim to be able to tease apart the genetic influences on intelligence, something that no other researcher has conclusively done:
Studies of twins have repeatedly confirmed a genetic basis for intelligence, personality and other aspects of behaviour. But efforts to link IQ to specific variations in DNA have led to a slew of irreproducible results. Critics have alleged that some of these studies’ methods were marred by wishful thinking and shoddy statistics. A sobering editorial in the January 2012 issue of Behavior Genetics declared that “it now seems likely that many of the published findings of the last decade are wrong or misleading and have not contributed to real advances in knowledge”. […]
In a 2013 study comparing the genomes of more than 126,000 people, the group identified three gene variants associated with with how many years of schooling a person had gone through or whether they had attended university. But the effect of these variants was small — each variant correlated with roughly one additional month of schooling in people who had it compared with people who did not.
A follow-up study found each influenced a person’s IQ score by 0.3 points (for scale, two-thirds of us range between 85 and 115 points). Those three genes are the upper limit on how much influence any single gene could have over intelligence; the smart money is that hundreds of genes contribute, each with a near-negligible tug and most affecting more than just intelligence. I love bringing up a famous Russian breeding experiment, where foxes were bred to behave nicer towards humans but as a side effect their fur, ears, and other physiology changed as well. This is the typical case for a gene’s influence on an animal, a messy constellation of subtle effects rather than a precise switch for a specific attribute.
I also wonder how anyone can talk about the average IQ score falling, when an IQ score is relative to the average score of people taking the test. It was never intended to be an absolute measure of intelligence. Still, if we buy into that idea then the best evidence still points to an upward trend in IQ scores, not vice-versa.
But back to the paper. One page is devoted to the history of secularism.
By the mid-nineteenth century, secularist thought had become fairly common among European intellectuals, especially at prominent universities. Presumably, secularist thought was then (and still is) largely limited to individuals with (a) genes for unusually high intelligence (i.e., in the upper 15% of the bell curve) and (b) few religiosity-promoting genes. Persons with any other genetic configurations would have rarely felt comfortable reasoning as secularists do.
Emphasis mine. Ah yes, another example of genetic determinism in an EvoPsych paper, something EvoPsych researchers claim never happens.
And finally, there’s one page repeating their main assertions.
We believe that the breadth and depth of this knowledge will continue to grow throughout the twenty-first century while the proportion of the world’s population who think in secularist terms declines. In other words, we join those who believe that the evidence for a nontheistic understanding of the universe is compelling (Dawkins 2009; Espinosa 2015; Gribbin and White 2016; Power 2012; Stenger 2012). However, due to genetic influences on religiosity, intelligence, and fertility, we predict that declining proportions of the human population will actually accept this type of secularist understanding over the long term. To defend this line of reasoning requires thinking in evolutionary terms, i.e., in terms of differential rates of reproduction.
Examining the Key Stone
I think you get the gist: this isn’t really a scientific study, it’s a theory paper masquerading as one. Over and over, the authors assert that there are genetic links to religiosity and intelligence, and that fertility will lead to a net increase in those genes within the human gene-pool, which is well beyond what their dataset revealed. So if I really want to give this paper a fair shake, I shouldn’t have bothered treating it as an honest study; I should instead have gone for a core premise of their theories.
Numerous studies (based primarily on twin designs) have indicated that religiosity is genetically influenced (Bouchard et al. 1999; D’Onofrio et al. 1999; Eaves et al. 2008; Kendler et al. 1997; Koenig et al. 2005). The extent of this influence is substantial, with most heritability estimates being in the .4–.5 range (reviewed by Bouchard and McGue 2003, pp. 29–31; Koenig and Bouchard 2006). However, a more recent study estimated the heritability to only be .26 (Lewis and Bates 2013).
That Nature article already hinted at some problems with twin studies and genetics. A paper by Evan Charney goes into more detail. To sketch out their methodology, twin studies search for a group of twins that were raised apart and pair them with a group that was never separated. If the group of twins raised apart behaves significantly different than those raised together, the theory goes, those behavior differences must be down to genetics.
A key part of this is finding twins that were raised apart. Yet as Charney points out, many twins “raised apart” in these studies actually spent years sharing the same household, in rare cases as many as eleven. Even once separated, nearly all twins in these studies kept in touch with one another and thus influenced each other’s behavior. Those that didn’t keep in touch would have been less likely to sign onto a twin study too, and twins that don’t resemble one another would face social stigma and be less likely to want to cooperate with their twin for a study. Speaking of social stigma, we also face different social pressures depending on how we look; a black man would have more shared experiences with another black man than a white woman, and this too would shape our beliefs and behavior. Twins, who resemble each other far more than would be expected by chance, would encounter this effect as well.
These studies also assume there’s no bias in which families adopt the wayward twin, but Charney points out that thirty years of research has shown adoptive families tend to resemble birth families. This is an especially big problem with religiosity, as religious institutions are a major player in adoption and sometimes zealously pursue “unwanted” children to give to good Christian couples. A secular person aware of this would be less likely to use Christian adoption agencies; a Christian would be attracted to them, and to adoption in general. This would bias “raised apart” twins to be more religious than “raised together” ones, inflating any religious similarities.
Charney wasn’t focused on this study, but another one that used fraternal and monozygotic/identical twins to examine the heritability of political leanings. That study didn’t just ask about politics, though, it surveyed these twins on twenty-eight separate questions. Care to guess how many of them show significant heritability?
Twenty-eight. In twenty four cases, heritability was stronger than the environment these twins shared.
|heritability||shared environment||unshared environment|
That should make you uneasy. Even the most die-hard genetic determinist would be happy to concede some things are primarily environmental. If that’s the case, we should be able to find areas where attitudes are not heritable. But if everything from attitudes toward X-rated movies to attitudes towards unions and busing show a significant heritable component, it’s time to question the validity of your heritability test. Nobody bats 1.000, not even Quantum Mechanics.
But there’s an even bigger problem lurking in the citations of Ellis and Hoskin’s paper.
Retrospective reports of religiousness showed little correlation difference between MZ (r=.69) and DZ (r=.59) twins. Reports of current religiousness, however, did show larger MZ (r=.62) than DZ (r=.42) similarity. Biometric analysis of the two religiousness ratings revealed that genetic factors were significantly weaker (12% vs. 44%) and shared environmental factors were significantly stronger (56% vs. 18%) in adolescence compared to adulthood.
The average of the estimates across all items leaves a striking impression that is remarkably consistent across boys and girls. On average, the estimated contribution of genetic influences is quite small, of the order of 10 percent of the total variance. By contrast, the shared environment of the family accounts for more than 50 percent of the variance, on average, leaving some 40 percent of the variance due to the unique environmental experiences of individuals within the family (e2) and measurement error.
Evolution lies on a bedrock assumption that genes are constant within the individual; a gene for colour blindness has the same effect from birth to death, unless you have another gene which counteracts its effect. This assumption of constancy is false, as otherwise cancer wouldn’t exist, but it’s still a good approximation of the truth. It is possible to have a complex phenotype that varies over time, but that dilutes the forces acting on the associated genes, and that phenotype has to be constructed from the mostly-static genes present in your germ line. Picture building a tower out of Legos; now picture building a tower out of Legos that automatically opens its front doors at 7 AM and 5 PM, without any external intervention. The latter is ridiculously tough and therefore unlikely to happen by chance, unless you have some magic Lego blocks which start moving at those times.
Human beings do have the equivalent of moving Lego, mind you: puberty, pregnancy, and menopause. Note though that two of them only apply to half of all humans, so they’re useless at explaining a trend in all of us. That leaves puberty, but it’s a poor fit for this. The time immediately after puberty is when human beings are putting the most effort into looking for high-quality mates, and before the advent of birth control it was also when we started popping out babies. Any behavioral gene should be in high gear at that point. In contrast, by our mid-thirties we’d have a large litter behind us; fertility would have dropped off, and in monogamous cultures we’d be cemented together with the partner we picked up much earlier. A behavioral gene would have minimal impact on the next generation, and thus would have a negligible evolutionary tug.
A theory about a adaptive behavior gene that has little effect around puberty but a great effect in adulthood should have been consigned to the scrap heap. It’s far more likely the observed effect is a cross-correlation with environmental or social factors.
There’s also a perverse genetic determinism at play here. Human beings are notorious for having a long adolescence that permits an incredible amount of learning and experimentation, all of which is primarily driven by cultural responses to the environment we live in. To argue for a significant genetic effect to behavior in later life, you have to argue that humanity forgets everything they’ve learned, outright ditches decades of experiences and all the effort we put into making big plastic brains, so that we can be tugged around by our genes.
It sounds ridiculous when put that way, but when put this way in a serious scientific paper…
The shifting trajectory from social to genetic influences on religious behavior from adolescence to adulthood implies a mechanism in which genetic differences gradually mediate the selection of adult behaviors from those presented by the adolescent social environment. The results are consistent with an evolutionary model for extended parental care in which there is ultimately a switch from a parental-based model of behavior to an individual based model of behavior.
… it looks legit on first glance. EvoPsych researchers are stuck trying to find ways to advocate for genetic determinism while appearing not to, and after years of experience they’ve gotten quite good at it.
These findings are consistent with the hypothesis that developmental change occurs around the time an adolescent leaves the family home. When this happens, parents begin to lose influence over their child since they cannot monitor his or her behavior to the same degree as they previously had. While growing up, a child is likely to be given little choice over attending church, celebrating religious holidays, or discussing religious teachings; but when he or she leaves the home, these activities are more difficult, if not impossible, for the parents to influence directly. Thus, the adolescent begins to decide for him- or herself whether religion will continue to be an important aspect of his/her life.
The authors must never have heard of endogamy and exogamy; typically, societies push people to pair up with non-relatives but social peers. This means that even if someone leaves the family home, they didn’t leave their cultural influences and thus didn’t have much choice in what to worship or when. Even in cities, immigrants tend to form their own enclaves which preserve much of their original culture. The universal freedom to worship they propose has only been easy to accomplish in certain European cities in the last hundred-ish years, not enough time to counterbalance the influence of the rest of the globe and tens of millenia. Even if we buy into their genetic determinism, their explanation for varying gene expression falls flat.
The Never-ending Story
Ugh. I could probably write another dozen posts on the flawed assumptions in this paper, but there’s no point. If we have good reason to doubt a strong genetic driver for religion, the foundation of their “secularization” theory is a sturdy as wet cardboard in a tsunami. What’s scary is that this paper appears to be a representative sample; being authored by two “semi-retired” academics and published in a two-year-old journal are red flags, true, but there’s a big name attached to the journal.
This is what the EvoPsych field deems to be sufficiently rigorous to be added to the scientific record, as far as I can tell. I remain comfortable calling it a pseudo-science.
 Matell, Michael S., and Jacob Jacoby. “Is There an Optimal Number of Alternatives for Likert Scale Items? Study I: Reliability and Validity.” Educational and Psychological Measurement 31, no. 3 (1971): 657–674.
 Matell, Michael S., and Jacob Jacoby. “Is There an Optimal Number of Alternatives for Likert-Scale Items? Effects of Testing Time and Scale Properties.” Journal of Applied Psychology 56, no. 6 (1972): 506.
 Henrich, Joseph, Steven J. Heine, and Ara Norenzayan. “Beyond WEIRD: Towards a Broad-Based Behavioral Science.” Behavioral and Brain Sciences 33, no. 2–3 (June 2010): 111–35. doi:10.1017/S0140525X10000725.
 Arnett, Jeffrey J. “The Neglected 95%: Why American Psychology Needs to Become Less American.” American Psychologist 63, no. 7 (2008): 602.
 Westoff, Charles F. “Recent trends in abortion and contraception in 12 countries.” (2005).
 Callaway, E. “‘Smart Genes’ Prove Elusive.” Nature News (2014).
 Trut, Lyudmila, Irina Oskina, and Anastasiya Kharlamova. “Animal Evolution during Domestication: The Domesticated Fox as a Model.” BioEssays : News and Reviews in Molecular, Cellular and Developmental Biology 31, no. 3 (March 2009): 349–60. doi:10.1002/bies.200800070.
 Trahan L, Stuebing KK, Hiscock MK, Fletcher JM. The Flynn Effect: A Meta-analysis. Psychological bulletin. 2014;140(5):1332-1360. doi:10.1037/a0037173.
 Charney, Evan. “Genes and ideologies.” Perspectives on Politics 6.02 (2008): 299-319.
 Alford, John R., Carolyn L. Funk, and John R. Hibbing. “Are Political Orientations Genetically Transmitted?” American Political Science Review 99, no. 02 (2005): 153–167.
 Koenig, Laura B., et al. “Genetic and environmental influences on religiousness: Findings for retrospective and current religiousness ratings.” Journal of personality 73.2 (2005): 471-488.
 Eaves, Lindon J., et al. “Social and genetic influences on adolescent religious attitudes and practices.” Social Forces 86.4 (2008): 1621-1646.