Obesity, Time Spent Eating, and Storks Bringing Babies
There has been a lot of blogospheric discussion of the the evidence that has been presented by Catherine Rampell at the New York Times Economix blog showing that countries in which people spend less time per day eating and drinking tend to be places with more obesity. This has been headlined as “fast food equals fat food.”
The author’s take:
Coincidence? Maybe, maybe not. ….
There does seem to be some correlation (although, as we all know, correlation is not causation). And note, of course, where America lies on this chart.
In her prior post on the topic, she had this to say:
Interestingly, of the 18 O.E.C.D. countries have current time-use surveys, the French spend the most time eating and drinking, and yet the French have the fifth lowest obesity rate (sixth lowest among all 30 O.E.C.D. countries, as many countries have data on obesity rates but not time use). Perhaps fast food really is the problem.
Like many people, I questioned whether the scatterplot presented in the post really provided much evidence of causality, so I went back to the original data tables to take a look at it (because I’m so cool).
I recreated the original analysis (minus the inclusion of the OECD average as a data point in the regression, for what I assume are obvious reasons). I get pretty much the same picture, and using a log regression form, get what looks to be the same trend line. The R-Squared on the regression (not noted in the original post, as far as I could see) is 26%. Without the U.S. and Mexico, it goes to about 6%, and becomes statistically insignificant.
But what was really interesting is that there are five other time categorizations provided at the source website. Here’s the same data plot, but using “Time Spent Doing Unpaid Work” instead of “Time Spent Eating and Drinking”:
Huh. This relationship, produced from the same data source, is about twice as strong (R-Squared = 52%) as the one that was reported. It took me literally five minutes of work to discover it. Why do you think that one was reported but not the other? This appears to be a textbook example of the human tendency to accept correlations as “not definitive, but part of the overall picture of evidence for causality” when such data serves to confirm pre-existing beliefs, and to ignore it otherwise.
Just in case the unpaid work doesn’t work for me, what’s are the R-squareds on alcohol consumption and smoking?
— tom · May 11, 05:15 PM · #
Watch out for the multicollinearity! Time(eating) + Time(unpaid labor) + Time(…) = 1 by definition. So if Time(eating) has a negative coefficient for realz, unless a person who spends less time eating has less time in the day than others we’d expect it to tremor out to the rest of the statistics regarding time use. Though the labor could explain how much time is available for eating. What happens if you include both and/or more, minus one time use variable to keep the matrix from imploding?
It’s a blog post, of course, but I think it’s a little unfair to to call it textbook bias – the kneejerk reaction/pre-existing belief is to suspect the opposite, that they’d be positively correlated, since Americans spend all day eating and snacking and adding 4th and 5th meals.
— rortybomb · May 11, 06:02 PM · #
rortybomb:
Yes, I was of course aware of that issue. Time spent eating is almost entirely uncorrelated with time in unpaid labor (R^2 = 1%). It’s logical that the “tremor” effect would be a ripple in the ocean, not a big wave in a still pond, because total time spent eating is very small as compared to any of the other 5 categories.
The reason I put in her quote from her prior post was to indicate that I thought it was clearly Rampell’s prior belief that fast food = fat food.
Best,
Jim
— Jim Manzi · May 11, 06:18 PM · #
I want to see all this graphed against time spend commenting on blogs.
— Tony Comstock · May 11, 06:31 PM · #
This William Easterly post (through Marginal Revolution) covers the same problem with data mining. I suspect I’d be looking for a secret pattern long after the rats settled on their bet.
I’d love to see a study showing how close to 50/50 the green/red ball ratio has to get for the rats to give up on ‘always green’.
http://blogs.nyu.edu/fas/dri/aidwatch/2009/05/maybe_we_should_put_rats_in_ch.html
— tom · May 11, 08:26 PM · #
This post provides stunning information. If fish oil can control obesity, then surely everyone will make a beeline for it. There can be nothing better than losing weight and feeling more fit. With loss of weight, the body feels suppler and many diseases are also averted. Taking fish oil can be very good for our overall health.
— Fish oil benefits · May 14, 10:52 AM · #