DOC PREVIEW
UW-Madison SOC 357 - Surprises From Self-Experimentation - Sleep, Mood, and Weight

This preview shows page 1-2-3-4 out of 12 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

CHANCE 7Many take it as a given that randomized experiments withmany subjects are the only way to learn. Is it true?Surprises From Self-Experimentation:Sleep, Mood, and WeightSeth RobertsI read Exploratory Data Analysis by JohnTukey while I was a graduate student inexperimental psychology. I enjoyed read-ing it and absorbed at least its most basiclesson, the value of plotting data. At thetime, I was doing experiments with ratsto learn how they measure time. One ofthe book’s main points is that “restrict-ing one’s self to the planned analysis —failing to accompany it with exploration— loses sight of the most interestingresults too frequently to be comfortable”(p. 3). My data supported this view. Per-haps 1% of my exploratory graphsshowed something surprising and inter-esting, and much of my research aftergraduate school, including some of theexperiments described here, derivedfrom ideas generated this way. It wasalso in graduate school that I began todo self-experiments. This article is aboutmy slow realization that self-experi-mentation and exploratory data analysishave something in common. Both areways of generating ideas. Self-experi-mentation seems to be better for gener-ating ideas than more conventional waysof collecting data, just as exploratory dataanalysis seems to be better for generat-ing ideas than more conventional waysof analyzing data.My interest in self-experimentationbegan when I read an article aboutteaching mathematics by Paul Halmos,a professor at Indiana University. Hal-mos emphasized that “the best way tolearn is to do.” I was trying to learn howto do experiments; I took this advice tomean I should do as many as possible.I could do more experiments, I realized,if I not only did rat experiments but alsodid experiments with myself as the sub-ject. So I started doing small self-exper-iments. Most of them were trivial andled nowhere (e.g., experiments aboutjuggling).At the time I had acne. My derma-tologist had prescribed both pills (tetra-cycline, a wide-spectrum antibiotic) andWWW.PHOTODISC.COM8 VOL. 14, NO. 2, 2001a cream (active ingredient benzoyl per-oxide). Simply for the sake of doingexperiments, any experiments, I did sim-ple tests to measure the effectiveness ofthese treatments. I believed the pillswere powerful and the cream had littleeffect. To my great surprise, the testsshowed the opposite: The cream waspowerful and the pills had little effect.It was very useful information. Manyyears later, an article in the British Jour-nal of Dermatology reported that antibi-otic-resistant acne is common.Sleep and BreakfastA few years after graduate school I beganto have trouble sleeping. I would wakeup early in the morning, tired but unableto fall back asleep for several hours —a type of insomnia called early awaken-ing. There was no good treatment for it,and it did not go away. My experiencewith acne made me think self-experi-mentation might help.Sleep was harder than acne. My acnestudies had uncovered useful factswithin weeks; my first 10 years of sleepresearch, however, merely showed thatall my ideas about the cause of earlyawakening were wrong, or at least notright enough to make much difference.Among the failed treatments weredietary changes, exercise, and changesin the timing of bedroom lights that wenton in the morning. What should you dowhen all your theories are wrong? I hadno idea. In 1990, I got a personal computerat home, making analysis of my sleepdata much easier. In early 1993, whileexploring the data, I looked at a graphsimilar to the upper panel of Fig. 1,which shows sleep duration over time.The 1993 graph had less data and wasnoisier than the upper panel of Fig. 1 butnevertheless revealed the same thing —that my sleep duration had decreased byabout 40 minutes/day in the middle of1992. I had not noticed the change. Inever used an alarm clock, so the changeimplied that my need for sleep haddecreased.The change in sleep duration hadhappened at the same time I had lost 5kg (lower panel of Fig. 1) by changingmy diet. Before the change, I had beeneating a conventional low-fat healthydiet. The dietary change was a reductionin processing (e.g., cooking, blending,adding spices). For instance, I ate rawfruit instead of fruit juice, brown riceinstead of bread, and stopped eatingalmost all prepared foods, including del-icatessen food, baked goods, and frozenFigure 1. Upper panel: sleep duration (including naps) over time. Each point is a 10% trimmed mean. Error bars show standarderrors determined by jackknifing. Lower panel: weight over time. The measurements started when the dietary change began; theystopped because the scale broke.CHANCE 9entrees. I did not try to change (or keepconstant) how much I ate; I always ateas much as I wanted to. The weight losswas not a surprise; based on rat experi-ments, I had been telling students foryears that processing food usually makesit more fattening.The next time I lectured on weightcontrol to my introductory psychologyclass I showed a graph similar to Fig. 1,with its suggestion that if you lose weightyou may need less sleep. A few weekslater, a student named Michael Leecame to my office to tell me that heknew another way to lose weight andsleep less: Eat a diet high in water con-tent. In practice, this meant eating lotsof fruit and salad. It had worked for him,he said. So I tried it. After a few weeks,however, it was clear the new diet hadlittle effect. I told Michael the results.He asked how much fruit I was eatingeach day. Four pieces, I said. “I eat sixpieces,” he said. So I started eating six pieces of fruiteach day. This required changing mybreakfast — instead of oatmeal, I hadtwo pieces of fruit, such as a banana andan apple. After about 10 days of the newbreakfast, I noticed that I was waking uptoo early much more often. While eat-ing oatmeal, I had been waking up tooearly about a third of the time. (I definedan instance of waking up too early as amorning when I fell back asleep withinsix hours after getting up.) Now I waswaking up too early every morning. Iswitched back to oatmeal and earlyawakening returned to its earlier level. Istarted eating fruit breakfasts again andearly awakening again became muchmore common — leaving no doubt itwas cause and effect. This was exciting;after 10 years of failure, I had finallyfound something that made a difference,albeit in the wrong direction. Tests ofother breakfasts suggested that anybreakfast with a substantial number


View Full Document

UW-Madison SOC 357 - Surprises From Self-Experimentation - Sleep, Mood, and Weight

Documents in this Course
Syllabus

Syllabus

12 pages

Sampling

Sampling

35 pages

Class 7

Class 7

6 pages

Review

Review

3 pages

Load more
Download Surprises From Self-Experimentation - Sleep, Mood, and Weight
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Surprises From Self-Experimentation - Sleep, Mood, and Weight and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Surprises From Self-Experimentation - Sleep, Mood, and Weight 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?