Published January 6, 2014

Happy new year, all, and for those of you have something to say about the Eagles’ loss to the Saints, feel free to put remarks in the comments section, and I (plus some guys I know from South Philly) will get to each of you. Please be patient.

Today’s post has nothing to do with football, but rather with something that deeply puzzles me about a new paper in press at Evolution and Human Behavior entitled, "The Effect of Ecological Harshness on Perceptions of the Ideal Female Body Size: An Experimental Life History Approach" by Sarah Hill and colleagues.

For those of you too busy at the gym keeping your new year’s resolution, what puzzles me is why priming works on preferences that are calibrated over the course of a lifetime. You may now get back to your CrossFit WOD or whatever.

Ok, for those of you too lazy to get into CrossFit and so still reading, here’s what Hill did. In a series of three experiments, they primed subjects with “ecological harshness” using fake news items about the present economic downturn or increasing levels of violent crime. The central dependent measure was subjects’ reported preferences regarding the ideal body size—people see an array of bodies that differ in size, and pick the one they think is the ideal.

The idea is, roughly, this. According to life history theory (LHT), optimal body size depends on the ecological conditions. Putting it, again, very roughly, if you grow up in a world in which resources are scarce, then it might not be a bad idea to store energy in your body against the risk that you’re likely to need it. In contrast, if you grow up in a world of rich resources, then storing calories in your body might not be as high a priority. In this way, the features of the environment one grows up in calibrates preferences about how one would like one’s body to be.

The authors were interested to see if priming subjects with respect to the present ecological conditions—are things going well or poorly?—would move around some subjects’ beliefs about ideal body size. As the authors put it:

We predicted that women sensitized to a faster life history strategy would respond to ecological harshness cues by idealizing a heavier body size relative to controls. Because the relationship between energy status and fertility is specific to women, we predicted that these cues would not influence men‘s own body ideals.

Holding aside the details, in three studies, results conformed to this prediction.

Here is why I find this puzzling. One way to think about life history theory is that people are building up representations of what the world is like from the data they gather as they are developing. How much food am I getting? How much social support do I have? How long are the people around me living? Every day, people get a new observation, building up, presumably, an increasingly accurate estimate of whatever parameter is relevant for the life history variable in question.

Updating the world should take place in something like a Bayesian fashion. If you’ve gone through 10 years of near starvation, and then you suddenly are treated to a banquet, you wouldn’t want to update your beliefs about the world you live in based on that one big meal. In terms of the data, that’s a drop in the bucket, one observation set against thousands. After a long period of time, updating should take place very gradually, and only, it seems to me, in the face of a substantial amount of new data.

Consider, for instance, Deb Lieberman’s nice work on incest. She finds that aversion to incest depends on the amount of time that one has lived with an opposite sex sibling. It would be surprising if the system designed to avoid incest changed significantly the day after—or even the hour after—an opposite sex sibling happened not to be around. Wouldn’t it?

In these priming studies, the subject is getting a tiny amount of information about the state of the world, in this case in the form of a newspaper article. Presumably this information is set against all the prior information that the subject had that influenced their view of ideal body size. In this context, the new information seems tiny. Why should such manipulations be expected to work? Is the life history system designed for a world in which abrupt changes occur frequently? Is it a side-effect, having to do with the way that the information is delivered?

I want to be clear that my worry isn’t a more general worry about priming in psychological experiments, which, over the last year, has generated a certain amount of debate. I’m puzzled because my prediction would have been that to move around these sorts of beliefs, whether they are implicit or explicit, would have been pretty difficult, given that the beliefs are like priors with a ton of data behind them.

There’s probably a good theory paper on this to be written. Challenge offered ...