Myers’ Critique of EP: Strong Language But Weak TeaPublished 18 January, 2011
A reader asked me to reply to PZ Myers’ post about Bering’s post, so I thought I would do that. (I apologize for the length of this one. Replying to posts is like grading papers. The very good ones take little time and require few comments, but…)
Ok, on to it… Myers’ first worry is about the samples and materials in the four studies Bering cites. Myers writes: “All of the studies involve small numbers, typically of college students at American universities (and even more narrowly, of psychology students), and all involve responses to highly subjective stimuli.” Sure, all studies would benefit from larger sample size, and of course generalizability from a sample is always an issue. While I’m happy to say that inferences must be cautious from any given sample, how is this a critique of evolutionary psychology? That is, does the use of such samples make any field (social psychology, perception, etc.) “awesomely trivial drivel?” Or did this only become a reprehensible scientific practice when evolutionary psychologists started doing it? I mean, people were using college sophomores to do social science before evolutionary psychology existed. (Zimbardo’s prison experiment: 24 American undergraduates…) Why don’t we see apoplectic blog posts every time someone reports on a study about stereotyping with college students? Why doesn’t research in other areas make bloggers moan that they “cannot bear the entire field of social psychology or personality psychology?” as Myers writes about evolutionary psychology. It’s very interesting that this dog doesn’t bark. If you want to test a hypothesis that has no functional model behind it, any method you choose is just fine. If your hypothesis derives from an idea about function, suddenly this method becomes intolerable.
In any case, of course these are convenience samples, and yes, that’s a reasonable worry and yes, one can have a sensible conversation about generalizing from any given study. That doesn’t make the field “tainted” or not rigorous. In fact, there’s a sense in which they’ve got it backwards. If you look at a journal such as Evolution and Human Behavior, many people who contribute are gathering data from field sites that other social scientists studying the same phenomena don’t. This is why you see such a greater fraction of anthropologists contributing to evolutionary psychology journals than to social psychology journals. (Anyone want to guess what fraction of papers in EHB compared to JPSP included an anthropologist in 2010?) If the issue is subject population, it seems to me that critics such as Myers should be applauding evolutionary psychologists in this respect.
As to the issue of using “highly subjective stimuli,” I’m not sure what a “subjective stimulus” is, but it seems to me that if one is going to study rape, you’re going to need stimuli of some sort, and I suppose they’ll be “subjective.”
Second, he writes about the studies Bering reviews that “[n]one test anything to do with inheritance, none try (or even can) look at the genetic basis of the behaviors they are studying. Yet somehow evolutionary psychologists conclude that ‘women may have been selected during human evolution to behave in ways that reduce the likelihood of conception as a consequence of rape.’” In a previous post, I talked about Coyne’s discussion of observations of bee behavior – which had nothing to do with genetics or inheritance – and his inference that the bees were designed to kill hornets by roasting them. Why didn’t Myers get steamed about this inference, condemning behavioral ecology as lacking in rigor? There’s no genes or inheritance there, right? Similarly, how did Harvey conclude that the heart is for pumping blood without looking at the genetic basis of the heart? Darwin didn’t know anything about genetics, yet was able to make plenty of inferences about function. How did he manage? Why should the logic of inferring function from form end with human social behavior? I know I say this a lot on this blog, but the logic of adaptationism is that one can make inferences about function from form. Bee behavior allows the inference of design to kill hornets. Sure, one can have a debate about this logic, but if so, then the worry applies just as much to Coyne’s claim about hornet-killing as to any similar claim about humans. Surely the argument isn’t that the inference from form/behavior to function is licensed only if one doesn’t style oneself an evolutionary psychologist. Right?
Next, Myers writes, “Another way to look at it is that they are hypothesizing that women are more likely to behave in ways that invite physical attack and brutal abuse when they aren’t ovulating. That is a remarkable assertion.” Superficially, that is a strange assertion. (I’m just going to address this claim. The material he quotes is actually about reducing the probability of conception conditional on being raped, not the probability of taking risks conditional on ovulating.) Let’s look at this in terms of risk sensitive foraging. Organisms will take fewer risks (going foraging, which risks predation) when they’re not in need of calories. Does that mean they are more likely to behave in ways that invite attack – more precisely, entail risk of attack – when they’re hungry? Yes, of course. The point is that there are always tradeoffs. Organisms have to make tradeoffs, which means taking risks in some cases but not others, depending on the offsetting probabilistic benefits. (Now, whether they are correct or not is a separate issue. I don’t really know what the tradeoff might be in this case, but I suppose there could be one. In any case, one wants some explanation for the pattern of data (N=232), and I’m sure Petralia and Gallup would welcome ideas for distinguishing their hypothesis from alternatives.) And yes, sure, maybe it’s a spurious result. That’s always a possibility. But the point here is that it’s not illogical to use the notion of tradeoffs to frame hypotheses.
Myers also has a comment about “culturally conditioned behaviors,” but I confess I don’t know what he means, so I just want to be explicit that I didn’t miss this, I just don’t know what he means. I think he draws the usual learned/innate and cultural/evolved distinction that most of us have abandoned, but it’s hard for me to tell.
Here he seems to shift from criticism of the primary literature to criticism of the post. He writes, “Then many of the studies that are described with such enthusiastic certainty as having definitive results turn out to be subjective, pointless messes.” I don’t quite see what he means. He quotes Bering here, and in the quoted material all Bering does is relate the results, not a hypothesis about function. I invite readers to look at the quoted material; does it do more than summarize the findings? I agree what one can infer from them is open to interpretation, but the data are the data.
At the top of his post, Myers writes that the field is “poor science propped up by a conviction that plausibility is sufficient support for certainty.” His objections are based on the use of undergrads (and “small” samples, in these cases in the hundreds), the confusion about the inference from behavior to design (in human social behavior, anyway), and the confusion about tradeoffs. But I think a more important point is this accusation about certainty. Again, I went back to the papers Bering discusses. Remember that Myers’ concern here seems to be (and maybe I’m wrong) that the field is “poor science propped up by a conviction that plausibility is sufficient support for certainty.” To take the first study Bering cites, about handgrip strength, the authors make a consistency claim, along with the modal “may” in there: “These results are consistent with other evidence that women may have been selected during human evolution to behave in ways that reduce the likelihood of conception as a consequence of rape” (my italics) and they even add, explicitly this line: “Our findings, however, must be interpreted with caution.” What more do you want in terms of being tentative in one’s conclusions? Similarly, Garver-Apgar and colleagues write, “women may possess specially designed perceptual counter-strategies that guard against male sexual coercion.” That modal “may” seems to be pretty distant from the certainty claim. Where is the “conviction” or the “certainty” that Myers hangs on evolutionary psychologists?
Now, it could be that he’s really saying that Bering is “too certain” in his blog post. Bering writes: “I’m riveted, and convinced, by much of the logic in this anti-rape area. And researchers are just getting started. Above is a set of astonishing truths.” He’s saying that he’s convinced, which I think Bering is entitled to be. Maybe he shouldn’t be; that’s fine. But in any case he’s clear that he thinks that the research is just getting started, rather than being all done. (You can tell by when he writes, “And researchers are just getting started.”) So maybe it’s the “astonishing truths.” Are those two words why he’s so upset? If so, then that’s fine. Bering is writing for Slate, not the primary literature, and I think a little enthusiasm is quite understandable. If Myers would have preferred a different flourish to punctuate Bering’s post, he’s certainly entitled to his opinion. I hardly think this is cause for his remarkably strong language about the field as a whole. I mean, the fact that these studies persuaded Bering doesn’t speak to the status of the discipline as a science.
And, since this post is so long anyway, another digression. Do you think the enthusiasts who commented on Coyne’s and Myers’ posts looked at the four papers in question, or even one of them, and then decided that, yes, these particular papers are representative of the discipline of evolutionary psychology as a whole? Did they then did they study the papers and determine that they were, in fact, poorly executed, thus licensing the strong claims Myers makes about the field? For that matter, if the paper by Navarette et al had used exactly the same method, but had no functional grounding at all, would the paper be better? And, speaking of that paper, what alternative framework predicts cycle effects?
Ok. I guess I’ll just leave it there, and apologize again for the length of this post.
(But just one ps, just as a bit of an aside. Myers touts Elisabeth Lloyd’s book on the female orgasm as “a wonderful example of solid, rigorous, scientific thinking.” I find that interesting in so many ways, but I’ll restrict myself to an observation or two. First, Lloyd’s thesis is the same one that Don Symons articulated a quarter century earlier in his book The Evolution of Human Sexuality. I find this particularly amusing in the context of Myers’ claim that “Too often investigators start with the assumption that a feature absolutely must have been selected for” since Symons exactly did not do this; he, and Lloyd, make a byproduct claim. I would argue that this is, in fact, typical of evolutionary psychology, which treats adaptation as the “onerous concept” Williams advised it should be. In any case, for a very different take on the book, I strongly recommend David Barash’s review here in Evolutionary Psychology, which includes classic lines such as this one: “Elisabeth A. Lloyd has taken a really terrific topic and written a really terrible book.”)