A blog about problems in the field of psychology and attempts to fix them.

Tuesday, November 27, 2012

Mind-Body Dualism is Bad for your Health!

Hey, you know all that pop-psych BS coming out of positive psychology and similar movements? Things like: "Try forgiveness, your cholesterol level will thank you!"*

Thanks to the good people at Psychological Science, I finally found a result I can support! October's issue brought us: "The mind is willing, but the flesh is week": The effects of mind-body dualism on health behavior by Forstmann, Burgmer, & Mussweiler from the University of Cologne. This clever little study found out that being a mind-body dualist is actually bad for your health. This is great! No more bothering with logical arguments or historic discussions, no more focus on incoherence, no more two part posts. You should all stop being dualist because it will make you gain weight and die young. Problem solved. And this isn't a all a joke, their methods were pretty clever, and the results pretty clear. The researchers start with a simple hypothesis:
Specifically, we hypothesized that the more people perceive their minds and bodies to be distinct entities, the less they engage in behaviors that protect their bodies.
In testing this hypothesis, they present six studies, with a few hundred participants between them. It all went something like this:


Study 1:
66 participants, gathered from a subject pool, read a short text (under the guise of testing potential textbook material) either about mind-body dualism or physicalism. Participants then took a manipulation check and answered a 26 item questionnaire. Participants primed for dualism reported less engagement in health-related behaviors (p = .047).

Study 2:
30 participants, found online through Mechanical Turk, unscrambled sentences, half of which related to dualistic or physicalist (depending on condition). They then indicated how important certain health related attitudes were to them. Participants primed for dualism reported less positive attitudes towards healthy behavior (p = .053).

Study 3:
29 participants, gathered from a subject pool, participants viewed pictures of either unhealthy or healthy items, then indicated the amount of overlap between mind and body. Participants who viewed unhealthy things indicated stronger dualistic belief (p = .042).



Study 4:
42 participants, gathered from a subject pool, were primed as in study 1. An unrelated task earned them a ticket they could use to try winning one of four cookbooks. Two cookbooks were about "unhealthy" foods, two about "healthy" foods - thus this study is supposed to be measuring health-relate behavior, rather than just attitudes. Participants primed for dualism selected the unhealthy cookbooks more often than those primed for physicalism (p = .049).

Study 5: 
53 participants, approached in a university cafeteria, in a were given a modified vignette from study 1 and asked to explain the concept back in their own words and apply the principle to their own lives. They were then given the first part of a faux memory task, and asked to return after eating lunch. After lunch they finished the memory task, and were then asked to describe what they had eaten for lunch, to rate their meal on a scale from 'not healthy' to 'very healthy', and to estimate the calories. A third party later looked at their descriptions and rated the meals on the same variables. An unrelated task earned them a ticket they could use to try winning one of four cookbooks. Participants in the dualism condition claimed to have eaten less healthy meals (p = .022) and coders agreed (p = .046). Results of the calorie count were not reported, and because coders and participant data were reported on different scales (with participant data z-transformed), we can't tell if

Clearly I wish the researchers had used better behavioral measures, but that said, it is a good first shot at the hypothesis. The only thing that really makes me sad about this study is the serial boarder line p values, which suggest something fishy might be going on. Especially for the study where they were recruiting online, it would have been pretty trivial to double the sample size and see what happens. I'm surprised Psych Science let them get away with that, but que sera. Uri, you want a shot at this one?

--------------

* For the record, I actually think positive psychology is a very good thing, as a corrective to our field's historic focus on the more "negative" aspects of personality. I'm only suspicious of it as a stand-alone cure-all, and as a purveyor of plain-language advice.

2 comments:

  1. Doug Candland sent me an email taking me to task for seeming to suggest that gathering more data would improve p-values. Of course, when you gather more data, the effect on the p-value will (on average) depend on whether or not your effect is real, and there is no guarantee that more data will give you a lower p. If your effect is not real, gathering more data probably won't make you happier. If I gave any other impression... then I deserve to be smacked around a little.

    To be clear, though, it is the case that doubling your sample size will (almost always) disambiguate a borderline result. It would be unusual to go from 50 to 100 subjects and still have your p sitting right on the edge of .05.

    ReplyDelete
  2. The just live a simple life without all the worries and other things that will just complicate one's life.
    child psychologist Denver

    ReplyDelete