Exercise and the Placebo Effect

August 25, 2008 9:48 AM

Cleaning hotel rooms is a strenuous business. However, when Alia Crum and Ellen Langer talked to 84 maids, most were under the impression that they did not get enough exercise. Furthermore, when they were measured for tests such as BMI and blood pressure, their results were typical of couch potatoes. The researchers let half the group in on the knowledge that they were getting more than enough of a daily workout and kept the rest in the dark. After a month results showed the former group were healthier on every single one of the objective health measurements tested - despite claiming to have been doing no more exercise or to have changed their diet. The study raises the possibility that mindset alone can influence our metabolism. Christopher Shea in the New York Times and Ben Goldacre in The Guardian have articles discussing the original paper.
posted by rongorongo (48 comments total)

this is amazing! sadly, since i'm a cynic, you could tell me anything and i wouldn't believe you.

so your damned placebo effect is wasted on me...
posted by punkbitch at 9:55 AM on August 25, 2008

Great! Now I just need to convince myself that I'm a maid!
posted by Astro Zombie at 9:58 AM on August 25, 2008 [9 favorites has favorites]

The paper link doesn't work for me and this wasn't addressed in the article: "...claiming to have been doing no more exercise or to have changed their diet"?

It seems like knowing how much exercise or food your subjects are getting is pretty central to a health and exercise study. For instance, people who think they are in poor health are more likely to say "what's another burrito?" Convince them them they are at or close to good health and they are more likely to try to "preserve" the "gains" they've made (i.e. stop digging, the first rule of getting out of a hole).
posted by DU at 10:01 AM on August 25, 2008 [3 favorites has favorites]

OK, so if I can count my 20 minutes of cleaning a day as exercise, then I'll only need an extra 10 minutes! Someone find me a study that says rough-housing and playing tag with your kids is exercise, even if you share a popsicle with them later.
posted by mitzyjalapeno at 10:06 AM on August 25, 2008

i just read at a very reputable scholarly journal that participating in community web forums makes you perform better in the sack...
posted by punkbitch at 10:08 AM on August 25, 2008 [4 favorites has favorites]

Diet and exercise studies are tough. The key is to not necessarily tell people what specific aims you are investigating, so that they aren't tempted to change anything to improve their results. This one, can't say, article isn't coming up in the link for me either.

I know my mom worked one summer as a maid in the Grand Hotel on Mackinac Island. She always carried the sheets with her left arm. Ended up with a freakishly strong left arm for a few years, which amused her to no end. She used to challenge guys to arm wrestling matches, getting them to use their left arm because she was "only a weak girl", thus pitching her strong arm against their weaker arm. She had a lot of fun with that.
posted by caution live frogs at 10:10 AM on August 25, 2008 [4 favorites has favorites]

Placebo effects are becoming a bit of a speciality of Goldacre - he even has a show on them. Which is god because it;s an area that needs more attention from people who aren't snake-oil salesmen.
posted by Artw at 10:16 AM on August 25, 2008

Cleaning hotel rooms is "tough" in the same sense that working on an assembly line (like my dad did) or as a carpet layer (brother in law) or industrial carpenter (brother) is. It's backbreaking and makes you tired but is not "exercise." Heart rate never goes up for more than a few minutes at a time; postures are painful with no range of motion; and of course you're working in hazardous environments to boot. When my brother in law ended up in hospital he weighed 300 lbs (and that at 5'6") and the doc read him the riot act: among other things, laying carpet is NOT the sort of "exercise" that would make him healthier.

Thing is, it's hell to try to fit in any sort of structured workout (even walking, for many people) when you're bone tired from doing blue- or pink-collar physically draining shitwork all day.

All that said, interesting links, thanks.
posted by ethnomethodologist at 10:21 AM on August 25, 2008 [12 favorites has favorites]

Mindset influences pain. That's proven. So why not this.
posted by Zambrano at 10:22 AM on August 25, 2008

The placebo effect is so damned cool. I'm always fascinated by cognitive neuro-bio stuff.
posted by paisley henosis at 10:22 AM on August 25, 2008

Mind over flab. Neat.
posted by Hairy Lobster at 10:29 AM on August 25, 2008 [1 favorite has favorites]

The article is Mind-Set Matters: Exercise and the Placebo Effect. Or via IngentaConnect.

A draft of the article.

It may be possible to obtain a copy of the article “Mind-Set Matters Exercise and the Placebo Effect,” by contacting Catherine West at [e-mail]. (via)
posted by zonem at 10:41 AM on August 25, 2008 [1 favorite has favorites]

Mindset influences pain. That's proven. So why not this.

Are you serious? Pain is perceived. Weight and blood pressure have very objective measurements. For mental attitudes to shape other mental attitudes is fascinating, but mental attitudes influence complicated physical characteristics of the body much more so.
posted by phrontist at 10:43 AM on August 25, 2008

And in addition to the experimental weaknesses pointed out by DU, if I were told that my job was good exercise, I might start doing it more gracefully and vigorously, because hey - it's making me healthy!
posted by phrontist at 10:44 AM on August 25, 2008 [1 favorite has favorites]

I'm at my desk, typing gracefully and vigorously RIGHT NOW!
posted by Artw at 10:54 AM on August 25, 2008 [1 favorite has favorites]

Like much of MeFi, we are gathered round, discussing - I hate to spoil the fun - a placebo paper. You have been told the effects of reading it by the post author, but in fact, it's simply a generic error page. This has been a control FPP.
posted by davemee at 10:55 AM on August 25, 2008 [9 favorites has favorites]

if I were told that my job was good exercise, I might start doing it more gracefully and vigorously

So if scientists told everyone that their jobs made them healthy and that singing instead of talking cured cancer, we would all suddenly be living in a 1950s technicolor musical?
posted by burnmp3s at 10:59 AM on August 25, 2008 [2 favorites has favorites]

Metafilter is more critical of research than any graduate seminar (which are notorious for their nitpickiness and zeal to find fault) in which I have participated. It's truly remarkable how many experts there are here that are quick to spot incredibly obvious and fatal flaws that somehow slipped through the cracks of peer review!
posted by proj at 11:00 AM on August 25, 2008 [9 favorites has favorites]

Are you serious? Pain is perceived.

Millions of people with "compressed nerves", bulging discs, herniated discs walk the earth pain-free. Most "runner's knee" pain is caused by worry that something is "wrong" with the knee.

It's been proven that if you are not told by the "expert" in the white coat and the MRI machine that "there is something wrong with you" that you will not have any pain.

Most non-trauma induced pain is psychosomatic. It's mental stress and worry that causes the pain.
posted by Zambrano at 11:11 AM on August 25, 2008

And in addition to the experimental weaknesses pointed out by DU, if I were told that my job was good exercise, I might start doing it more gracefully and vigorously, because hey - it's making me healthy!

You Might, indeed. But you might not realize it, either. Your mind may be thinking 'hey, I'm exercising now, better get that heart rate up and start making some energy out of this fat I have stored away" - in essence, this _is_ the placebo effect. And it seems to be very real.
posted by Space Coyote at 11:14 AM on August 25, 2008

So if scientists told everyone that their jobs made them healthy and that singing instead of talking cured cancer, we would all suddenly be living in a 1950s technicolor musical?

Even if it didn't cure cancer, it might dramatically improve quality of life.
posted by weston at 11:23 AM on August 25, 2008 [4 favorites has favorites]

Metafilter is more critical of research they don't like the outcome of...

FTFY. If they do like the outcome, it's a chorus of Films At 11.

(I excuse my own nitpick because they specifically said they aren't measuring the variables they are most interested in knowing.)
posted by DU at 11:33 AM on August 25, 2008

It's been proven that if you are not told by the "expert" in the white coat and the MRI machine that "there is something wrong with you" that you will not have any pain.

QUOTING SCIENCE
posted by unSane at 11:42 AM on August 25, 2008 [2 favorites has favorites]

Most non-trauma induced pain is psychosomatic.

Cite? From a creditable source (i.e. somebody who isn't to pain research what Rupert Sheldrake is to psychology)?
posted by Lentrohamsanin at 11:44 AM on August 25, 2008

It's been proven that if you are not told by the "expert" in the white coat and the MRI machine that "there is something wrong with you" that you will not have any pain.

This from the fool that thinks that people with ragweed and pollen allergies are delusional.
posted by GeekAnimator at 11:52 AM on August 25, 2008 [6 favorites has favorites]

Most non-trauma induced pain is psychosomatic. It's mental stress and worry that causes the pain.

You forgot wandering uteri.
posted by sondrialiac at 11:58 AM on August 25, 2008 [1 favorite has favorites]

So what do we take for mental stress and worry? Could that $10 hour of yoga truly be more effective than a jar of ibuprofen?
posted by debbie_ann at 12:06 PM on August 25, 2008

Millions of people with "compressed nerves", bulging discs, herniated discs walk the earth pain-free. Most "runner's knee" pain is caused by worry that something is "wrong" with the knee.

It's been proven that if you are not told by the "expert" in the white coat and the MRI machine that "there is something wrong with you" that you will not have any pain.

So what you're saying is that people don't go to the doctor because they're in pain, but rather, they were fine until they went to the doctor and now they're in pain?

Are you here shilling for homeopathy or for Rikei?
posted by Pope Guilty at 12:06 PM on August 25, 2008

Oh how I hate to introduce personal anecdote into this interrogation of scientific method, but here goes anyway. My daughter recently had an ear infection that hurt so bad she was sobbing through dinner (and this is a kid who loves to eat). As soon as I took her to the doctor and we found out it was an ear infection (she had been doing some high-diving and we were concerned there could be some damage to her ears), the pain immediately lessened to the point where she didn't even want to take the painkiller the doctor prescribed. Although I am not sure pain can be entirely managed in the brain, I do think stress and worry definitely make pain worse.
posted by Fennel B. at 12:11 PM on August 25, 2008

The usual mistake is being made frequently here, namely, distinguishing "psychological" and "physical" effects as two totally distinct and separate things.

Think carefully about it: nothing that happens to your "body" is unmediated by your mind.
posted by fourcheesemac at 12:25 PM on August 25, 2008 [2 favorites has favorites]

This intriqued me, because it so flew in the face of what I understand from the literature to date. Unfortunately, it's a case of invalid analysis of an invalid study design. Quoting from the paper:

Four hotels were assigned to the informed condition, and three were assigned to the control condition.
...
Paired-sample t tests indicated that on each of these variables, the informed group differed significantly between Time 1 and Time 2 ...

It was a cluster-randomized design (with an inadequate number of clusters to boot), but was analyzed as though it were individually randomized, thus way, way overstating the significance levels. You notice the differences are small for both groups, and there's no way they are statistically significant given the cluster randomization. This is a common error in poorly educated scientists and cuts across multiple disciplines. They need to read this book.
posted by Mental Wimp at 12:54 PM on August 25, 2008 [8 favorites has favorites]

I just did a quick read-through of the paper. Seems solid enough in its reporting of the results, but the researchers' conclusion (and thus subsequent media coverage) is a bit misleading.

Lets take the study design and results at face value: the informed group really did see a significant decrease in biologically relevant and objectively measurable health indicators. There are (at least) three possible ways that this might have occurred:

(1) The informed group did not change their diet & exercise behavior at all, in or out of work. This suggests that their mental state had a direct, unmediated impact on their physiology.

(2) The informed group did change their diet and exercise behavior at work, at home, or both, but did not notice it and did not report doing so. This suggests that their mental state had an impact on their behavior, though they did not perceive it.

(3) The informed group did change their diet and exercise behavior at work, at home, or both, but lied. This suggests that their mental state had an impact on their behavior.

The data support all three possibilities equally. There is nothing in this paper to suggest which of these three possibilities is most likely, because the researchers never actually directly measured how much exercise the women got or what they ate. Everything was self-reported.

Now here's the critical point: as the study itself demonstrates, the women themselves were incredibly unreliable observers of their own behavior. For example, even after being told by the researchers that their work easily qualified as exercise, only half of the informed group said that work was exercise. And even after being told that they exercise every day at work, the percentage of workers who said they exercised regularly rose from about 42% to 68%. - e.g., about 1 in 3 still didn't think they were exercising.

It is thus entirely reasonable to conclude that the study participants were just not very reliable in reporting their own behaviors. So DU's intuition - that many did in fact work a bit more vigorously, watch what they eat a bit more, etc. - is just as likely as some kind of direct mind-body link.

Nevertheless, the researchers try hard to push conculsion #1:

Given this knowledge, one interpretation of our results regarding the relationship between increased perceived exercise and improved health would be that they were mediated by a change in behavior. The data collected in this study, however, do not support this conclusion. As mentioned, 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). Thus, neither increased exercise nor decreased caloric intake was reported by the subjects.

Just to drive the point home: the study itself demonstrated that people are unreliable reporters of how much they exercise, yet depends on peoples' reporting of the amount that they exercise to reach its conclusions. This doesn't mean that option #1 is invalid - just that it is no more proven by this study than options #2 and #3.
posted by googly at 1:21 PM on August 25, 2008 [2 favorites has favorites]

&lt taps fourcheesemac's knee with a rubber hammer &gt
posted by anthill at 1:21 PM on August 25, 2008 [1 favorite has favorites]

Wait, they told everyone in a given hotel the same thing and never stopped to think that this might affect the results? It does so greatly, because any random changes in the hotels are now probable causes for changes in the individuals.
posted by ikalliom at 1:25 PM on August 25, 2008

Most "runner's knee" pain is caused by worry that something is "wrong" with the knee.

Nonsense.
posted by Pax at 1:48 PM on August 25, 2008

because any random changes in the hotels are now probable causes for changes in the individuals.

That's exactly why my comment about group-randomized design and analysis was made. The correlation induced by random or systematic changes in the hotels means that individuals' responses from the same hotel aren't as informative as if they were in separate hotels or were randomized individually to separate treatment. They overstate the precision of the means by pretending that each person's experience is uncorrelated with the others. Rookie mistake. I shame the journal editors for not catching this. They must not have had statistical reviewers or the statisticians were asleep when the looked at this.
posted by Mental Wimp at 1:55 PM on August 25, 2008

Mental Wimp: Yes, I was restating it because I don't consider myself a statistician and I wasn't sure if I understood your comment correctly. Thank you for the clarification.
posted by ikalliom at 2:07 PM on August 25, 2008

From a practical point of view, it's not easy to give different information to maids in the same hotel. Just imagine the coffee break conversation.
posted by ikalliom at 2:14 PM on August 25, 2008

From a practical point of view, it's not easy to give different information to maids in the same hotel. Just imagine the coffee break conversation.

Are you now restating the linked material to make sure you understand it?
posted by Bokononist at 2:43 PM on August 25, 2008

Great comments in this thread.

My favorite experiment linking physical and mental exercise given in a classic study by Yue and Cole (1992). I was thinking about it earlier today, so this thread is timely. Yue and Cole compared increase in muscle strength among participants who actually underwent physical training to increase in muscle strength among participants who merely imagined exercising. Training produced a 30 percent increase in maximal force; imagining did not produce as much, but still produced a 22 percent increase!

The experiment is subject to some of the excellent worries that people in this thread have mentioned, but there is theoretical reason to be unsurprised at the results. It's easy to prima facie think that exertion causes muscles to grow without any mediation from the brain: muscle-building is like a callus forming on the soles of your feet, or like pressure hardening coal into a diamond. Of course, that's not the way it works: exertion causes your brain to send commands that bulk you up. It appears that the same neural mechanisms that are responsible for motor action are also responsible for motor imagery. (There are the "mirror neurons" that the science sections of newspapers are so excited about all the time.) Some theorists (simulationists, believers in enactive cognition) posit that when we imagine doing something, we "simulate" the action by using the exact same cognitive mechanisms that we would use in actually performing the movement, but we run the process "offline" so that it doesn't actually result in behavior. However, it might not be possible to quarantine all the behavioral effects. Maybe kicking the appropriate mirror neurons into gear, whether in online or offline mode, is all that's needed to get the brain to send the PUMP YOU UP signal.

Following this line of thought can lead to some pretty entertaining hypotheses. It's fun to imagine that someone could get totally ripped without ever working out, simply by meditating really hard. Of course, such a person would need to be amazingly adept at motor imagery. They couldn't let distractions get in the way. It's actually pretty difficult to continually imagine doing crunches or whatever without the brain getting bored and turning to something else. Constant motor imagery requires meditation aids, and it's pretty clear what the best meditation aid would be. The best way to keep thinking of doing crunches is... to do crunches! So here's my surely wrong but kinda interesting way of thinking about exercise. Increasing muscle mass is a fully mental activity; actually working out is a peripheral or incidental way of encouraging that mental activity. Muscleheads are Zen masters who pump iron for the same reason that yoga instructors hum mantras and listen to CDs of rainforest sounds.

If motor imagery leads to increase in muscle mass, I wonder if distracting yourself by watching TV or listening to an iPod at the gym makes the workout less efficient? If so, you could explain the placebo effect of the maid study. The ones who were told that they were exercising possibly paid more attention to their duties and treated them as exercise, whereas the others were distracted.
posted by painquale at 4:47 PM on August 25, 2008 [6 favorites has favorites]

My friend was a maid for a year or two at a nice hotel in New England maybe 5 years ago. He said he would routinely lock the door and take a breather and sometimes a nap when he didn't have too much to do.

One day, another, longer-serving maid walked in on him and found him asleep. He was nervous, obviously, be she assured him that it was cool and that everybody did it.

I would like to put forth the theory that these maids in the study, upon hearing that they should be getting more exercise than they seemed to be, decided to spend the month working extra hard and moving around so as to show that they were, indeed, working hard.
posted by crazy finger at 5:32 PM on August 25, 2008 [1 favorite has favorites]

Increasing muscle mass is a fully mental activity; actually working out is a peripheral or incidental way of encouraging that mental activity. Muscleheads are Zen masters who pump iron for the same reason that yoga instructors hum mantras and listen to CDs of rainforest sounds.


I'll buy that for a dollar.
posted by From Bklyn at 8:12 PM on August 25, 2008

Are you now restating the linked material to make sure you understand it?

All my trials can be considered perfectly random.
posted by ikalliom at 10:33 PM on August 25, 2008

Mental Wimp - thanks for your insight on the stats. I dimly recall this from the days when I was a psychology student. I guess it is possible we are all looking at the draft of the paper and that changes to the design were made in the final, peer reviewed version.

In terms of interpreting the results Ellen Langer certainly has an agenda of pushing mindfulness as an influence on health in her research which points towards googly's #1 option.
posted by rongorongo at 9:54 AM on August 26, 2008

From a practical point of view, it's not easy to give different information to maids in the same hotel. Just imagine the coffee break conversation.

Yes, I agree. One approach would be to only select one maid per hotel, but then you'd have to go to nearly 200 motels. Another would be to select enough hotels and then analyze using hierarchical mixed models or generalized estimating equations that model the correlation within hotels.
posted by Mental Wimp at 9:59 AM on August 26, 2008

I guess it is possible we are all looking at the draft of the paper and that changes to the design were made in the final, peer reviewed version.

No, I was looking at the published paper. They really did screw up. It doesn't mean their conclusion is wrong, it just means that their study doesn't support it or refute it. It is irrelevant until they do the proper power calculations, and then, if the power is adequate (which I doubt), the appropriate analysis to adjust for the correlated outcomes.
posted by Mental Wimp at 10:07 AM on August 26, 2008

Another anecdote: even though I walk around a lot going about my daily business (no car), one of my doctors told me while we were discussing an exercise plan that my body would respond differently to going for a walk with the specific goal of exercising than going for a walk on the way to the grocery store. He didn't get too technical about it, but the gist was that the anticipation of exercise would make my body gear up before I actually started, making the workout more effective. That's if I understood correctly.
posted by bettafish at 6:06 PM on August 26, 2008

The short answer is that no one understands this stuff very well. Anyone who says so is selling you something, even if it is just selling their name on a research paper.

There are far too many variables to track.

You can take two individuals and feed them EXACTLY the same food, and have them do the EXACT same exercises, and you could have WILDLY different outcomes.

I've seen 5'2" women eat 2000+ calories per day with no exercise of any kind and still struggle to keep from losing weight, and I've seen 6'4" men eat 1400 calories per day, do 3 spin classes a week, and still go up a pant size.

I need to drop a good 50lbs. I do not eat like a cow. Based upon observation of other people I eat about the same, if not less, than most others.

I just finished an 8 month exercise routine, where I did modify my eating some (on the good side), but I dramatically modified my physical activity. Cardio 3 days a week, weight training 2 days per week.

After 8 months, practically no change. The cardio did help my stamina, but there was no change to my physique.

My doctor says that I can radically change my eating and probably loose some if not most of the weight, but that if I ever revert to "normal" eating, then I will immediately ricochet back to my current weight.

The "yo-yo" effect is very well observed and basically the entire diet industry is built upon it as a foundation!

A good friend of mine was eating normal meals and dropping weight like crazy. They tested for everything, nothing showed up. His doctor had him eat 2 peanut butter and honey sandwiches and wash it down with a bodybuilder shake every night right before bed, and he still kept losing weight.

It was anxiety. And the anxiety over him losing weight made his weight loss accelerate, despite taking in almost 1000 extra calories per day.

Counseling, some anti-anxiety pills, a few years later he eats like a bird and has a nice little middle-aged paunch in the front.

It's all about genetics. You can, with dedication, deviate slightly from your target, but seldom very much, and seldom permanently.
posted by Ynoxas at 12:32 PM on August 27, 2008