Why you shouldn’t trust 65% of Evolutionary Psychology

Over the past few million years evolution has moulded our bodies; changing us from an animal that looked very similar to an ape into one that looks slightly less like an ape. Evolutionary psychology claims that a similar thing has happened to our minds; and attempts to figure out which of our behaviours have been influenced by evolution and how. EvoPsych an approach with a lot of potential – after all, figuring out why we behave in certain ways could help us understand why our thought processes sometimes go wrong – but it’s also a field that really annoys me because a fairly major issue undermines a lot of research on the topic. About 65% of the research, to be exact.

What makes me especially annoyed is that this flaw is not inherent to EvoPsych. It’s not some fundamental problem that can’t be rectified, but a purely artificial issue that is rendering a large body of work effectively useless. But in order to get you to join me in my frustration (because if I am unhappy, other people should be unhappy) we need to back up a bit.

The first step in any EvoPsych investigation is to try and figure out which of our behaviours have been influenced by evolution, rather than education or culture. After all, our brains are incredibly flexible organs, capable of adapting to many circumstances. How to do this? Natural selection provides the answer. Animals with beneficial traits will fare better in the environment and have more babies, spreading those traits around the population. So if you want to spot something with an evolutionary history, you should look for stuff that appears to be effectively universal in a species. This works for our bodies, as well as behaviour. After all, bipedalism evolved and as a result everyone normally walks about on two legs.

Bipedalism is so prevalent I even made it a header image here. That’s commitment.

So far, so good; but here is where the problem emerges. If you were to design an experiment to try and identify behaviours that were widespread, if not universal, in humans how would you do it? You’d take a sample from a range of populations and test to see if they behaved the same way. The problem is that this costs a lot of money, since you’d have to travel all over the world to get a suitably wide sample. So instead, many psychologists work with local populations. And most psychologists live in the West, so most of them are only looking at Westerners (or as they’re sometimes called people from Western, Educated, Industrialised, Rich Democracies; because that creates an amusing acronym. WEIRD).

In fact, a study in 2008 found that 96% of research published in six of the top American psychological journals only examined WEIRD people. For some context, WEIRD people only make up around 12% of the population of the planet. And this is the big problem: researchers are trying to find “universal” behaviours but are only looking at a small group of us that is not representative of humans as a whole¹.

Now, this study was looking at psychology as a whole. Breaking it down and focusing purely on EvoPsych journals seems to paint a better picture. One researchers response to the 2008 study, for example, examined an EvoPsych journal – Evolution and Human Behavior – and found that they exclusively used WEIRD participants in only 65% of research; a much more respectable number. However, that’s only one journal. I took a look at the latest issue of Evolutionary Psychology (the journal on whose website the aforementioned response was published) and of all the experiments on human participants published, every single one of them used WEIRD people.

It’s clear evolutionary psychologists are aware of the issue using WEIRD people raises (and seem to be on the cutting edge of psychology when it comes to removing it), but it is still seem to be a prevalent problem. This reliance on WEIRD people, whether in 100% of cases, 65% or 96% means that the research is being conducted on an un-representative sample.

If we were to take a similarly small sample with other topics, we might come up with some wacky results like…

• Everyone is left handed (10-30% of the population)
• Everyone is gay (2 – 20% of the population, depending on question asked)
• All Americans have encountered a ghost (22% of the population)
• Everyone in the UK lives in London (11%)

In short, I think it’s pretty clear that studying such a small portion of the population probably isn’t going to tell us much about that population as a whole; but if that isn’t obvious there are numerous cases where people have begun looking for “universal” traits identified in WEIRD populations in other groups and failed to find them. For example, a lot of WEIRD research discovered that a lot of people found a certain facial structure was attractive. So sexual selection influenced the evolution of our faces! At least, that’s what we thought until people from other cultures were asked. Turns out this isn’t actually a universal, evolved preference; just a Western cultural one.

The Muller-Lyer illusion. The two lines are the same length, but don’t look it. People from different cultures are more/less fooled by the illusion

And that’s just the tip of the iceberg. That 4% of non-WEIRD research has revealed numerous more differences between WEIRD people and the rest of the world. Even some stuff you might think is actually innate. The Muller-Lyer illusion (see picture above) affects people from different groups differently. Some hunter-gatherers, for example, aren’t even fooled by it! You’d think that illusions, which are tricking our brains, would be universal. We all have very similar brains after all. WEIRD people are also less driven to conform, more likely to make moral judgements based on consequences and tend to be a bit fairer when divying up prizes between groups of people….the list goes on².

These are all behaviours which most EvoPsych research would identify as universal (and thus perhaps evolved) but is actually influenced by culture, education and experience. And unless you study people from a range of cultures, educational backgrounds and so forth; you’re not really going to be able to figure out which behaviours have evolved. Don’t get me wrong. Examining WEIRD people is cheap and efficient, and is a good way of throwing a lot of ideas at the wall to see what sticks. But without a follow-up demonstrating that whatever does stick is actually universal, it remains a hypothesis.

And that’s why I’m frustrated with evolutionary psychology; and why you should be too. And why, unless you see research that has a broad sample, it can’t tell us too much about our evolution. Which sadly, seems to be the most of the research.

 References & notes

1. Arnett JJ. (2008). The neglected 95%: Why American psychology needs to become less American. American Psychologist , 63, 602–614
2. Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?. Behavioral and Brain Sciences, 33(2-3), 61-83.

EDIT NOTE: This post originally talked about 96% of research using WEIRD people. That was based on studies looking at psychology as a whole. Looking at EvoPsych research in particular seems to suggest they are slightly better, with a study of one journal suggesting it may be as low as 65%. I took a quick look at another journal though, and this seems to be a bit on the low side. Nevertheless, I’ve updated the post with a discussion of all this, and to use the new, lower figure in the title.