Evolutionary Psychology, Causality, and Conjecture
Like most pop-Evolutionary Psychology articles, it leaves me grinding my teeth. Normally I hesitate to harp on other people’s fields, but Evolutionary Psych, especially as it appears in popular media (which may be no fault of the actually researchers, heaven knows I have my issues with science reporting), is one that bugs me. It seems based largely on finding faint statistical effects, and then “explaining” them as they might relate to our evolutionary past, using, as Mike puts it “Just-So Stories”.
That is what it feels like, and it’s essentially a massive exercise in confounding. While yes, your explanation might be true, it is also possible there is a third factor associated with both, that is generating a correlation that isn’t there. The canonical example I was taught was this:
- Smokers are more likely to die of gunshot wounds
Now, I can invent some stories about why this would be true, and if I could somehow relate it back to cavemen and sex, preferably lurid sex, I have the makings of an evolutionary psychology article. Smoking is sexy as we all know, which leads to jilted lovers, torrid affairs, and thus death by firearm.
The actual answer, at least in this example, is that smokers are no more likely to die of being shot. But, both smokers and gunshot victims are more likely to be poor. If you don’t look at the victim’s socioeconomic status, what you get is an association between two unrelated phenomena. And if you do…the association goes away, and we are left with one of the main conclusions of social epidemiology, that being poor is bad for you.
Evolutionary Psych seems to do this alot – substituting a rigorous exploration of actual causal effects with a conjecture built on a gut feeling, or something where, if I came up with it over dinner, people would go “Yeah, that’s probably true…”
For example, this article suggests several behaviors in ovulating that might be explained by rape avoidance – but could just as easily be explained by being selective about mates, which presuming strangers apt toward violent sexual assault aren’t considered “good mates”, would induce the association between ovulation and rape avoidance despite the actual cause being mate selectivity. Whether this association is true, or confounded, is of course something that could be – and needs to be – answered with careful study design and rigorous statistical analysis. Not a suggestive narrative – that isn’t actually science.
Others cover other problems with the article better than I can (in brief, that the focus of the consequences of rape are all pregnancy based, which seems at odds with the trauma of the experience, the idea that evolutionary adaptation is that strong – and that transferable to modern life (are the same things risky behavior now and generations ago?), etc.), but given the name of the blog is Confounded by Confounding, I thought that bit deserved special mention.
Filed under: General | 1 Comment