Exercise Sizzle Works Sans Steak

I was stunned to learn from Psychological Science that exercise has an apparent placebo effect: 

84 female room attendants working in seven different hotels were measured on physiological health variables affected by exercise. Those in the informed condition were told that the work they do (cleaning hotel rooms) is good exercise and satisfies the Surgeon General’s recommendations for an active lifestyle. Examples of how their work was exercise were provided. Subjects in the control group were not given this information. Although actual behavior did not change, 4 weeks after the intervention, the informed group perceived themselves to be getting significantly more exercise than before. As a result, compared with the control group, they showed a decrease in weight, blood pressure, body fat, waist-to-hip ratio, and body mass index. These results support the hypothesis that exercise affects health in part or in whole via the placebo effect.

This could also plausibly be a status effect; status has huge health benefits, and perhaps good exercise jobs are seen as having higher status.  This result is stunning because exercise had previously been my clearest example of something people could do to improve their health.  Yes, the strong correlation between health and exercise is partly because healthy people feel more like exercizing.  But it seemed so plausible that exercise also improves health.  Now I am not so sure.

This is also stunning because we already saw that at least 2/3, and perhaps all, of the benefit of anti-depressant drugs is a placebo effect.

Added: The paper reports no apparent change in related health behavior:

The room attendants did not report any increase in exercise outside of work, nor did they experience any increase in workload over the course of the study. In addition, the subjects reported their habits had not changed over the past 30 days with respect to how much they ate (including servings of sugary foods and vegetables) and how much they drank (caffeine, alcohol, and water).

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • billswift

    I have never been convinced that exercise is necessary for health, although I am sure it is helpful. I exercise for fitness, which is beyond “mere” healthiness. But this study is too singular and too small to be convincing either way, it is interesting and should be followed up, but there are too many possible confounding factors and the effects are too small for a single study. One possible confounding factor they don’t seem to have checked is that the experimental group, once the “exercise” benefits of their work were pointed out, acted so as to emphasize those benefits in how they did their work (I regularly work that way, moving so as to enhance the benefits of my regular work, in addition to exercising).

  • http://amnap.blogspot.com/ Matthew C

    It really makes you think, doesn’t it. Wonder what the placebo multiplier is for any number of different medications? I bet over 50% for most of them. . .

  • http://profile.typekey.com/sentience/ Eliezer Yudkowsky

    You know how often such studies fail to replicate. Effect size? Statistical significance / likelihood ratio?

  • http://profile.typekey.com/robinhanson/ Robin Hanson

    Eliezer, I did give the link so you could see for yourself. Of nine effects considered, all were significant with the predicted sign at the .05 level, and three were significant at the .001 level.

  • rcriii

    Could being told that their work was exercise have motivated the experimental group to take other steps towards health – e.g. reducing calories, smoking or drinking? Note that this still might leave room for a (possibly significant) placebo effect.

  • http://amnap.blogspot.com/ Matthew C

    Robin,

    Any versions of the paper available that don’t require me to pay $29 to read?

  • http://profile.typekey.com/bayesian/ Peter McCluskey

    Your comment about antidepressants could easily be misinterpreted, and the evidence is more complicated than your reference suggests.
    See http://www.bayesianinvestor.com/ovb28.html for comments that Typepad’s annoying spam filter won’t let me post here.

  • http://profile.typekey.com/robinhanson/ Robin Hanson

    Peter, the studies you point to don’t seem to conflict with the study I cited.

  • billswift

    Self-reporting exercise and diet makes the study of even less worth, self-reports are notoriously unreliable.

  • http://areasonableman.com Gil

    It’s even stranger because you’d expect people who thought they were getting more exercise to feel like they could afford to eat worse than they would otherwise (regardless of what they self-reported to have eaten). That, absent the placebo effect, would make me expect their health to become worse than the control group, rather than better.

  • http://profile.typekey.com/xp39c/ Hei Lun Chan

    “This could also plausibly be a status effect; status has huge health benefits, and perhaps good exercise jobs are seen as having higher status.”

    I don’t think that’s right, since “good exercise jobs” is just another phrase for manual labor.

  • http://www.nancybuttons.com Nancy Lebovitz

    There could be a status effect because exercise is considered very virtuous/high status. Being told that their work counts as exercise could make maids think of themselves and their work as higher status than they did previously.

  • Walt Guyll

    Perhaps there are health benefits to being deceived by Harvard researchers.

  • mk

    This is very interesting, but I wonder. Maybe it suggests that people who are mentally focused on getting exercise get more out of it.

    Is this study really an example of “sizzle without steak”? Is it false that cleaning is pretty good exercise? The study may instead indicate that “steak without sizzle” is not as effective as “steak with sizzle.”

    “Sizzle” here would mean goal-directed attitude, a purposive intent to get positive benefits from exercise. I could imagine there being differences in heart rate, etc. People who have no intent to get good exercise are more lackadaisical, perhaps.

    I dunno. It’s very interesting but I don’t see that it conclusively indicates a placebo effect.