Most Popular

    Sorry. No data so far.

Winner of Amgen Patients | Choices | Empowerment Competition Emerging Star of HealthCare Engagement Award
Mayo Clinic Award - LeftA winner of the Mayo Clinic iSpot Competition for Ideas that will Transform HealthcareMayo Clinic Award - R

The Blog of Tim Ferriss

The Value of Self-Experimentation

Written by Tim Ferriss, originally appeared Dec 18 at the link above.

Today (Saturday) is the last day that The 4-Hour Body has to beat Guinness on the New York Times’ bestseller list. Although my book has done well on Amazon, Guinness OWNS retail, which is at least 60% of our total reported sales.

For three years, I went through hell and back for this book (100s of experiments like the above), so I’ll kindly ask for your help one last time.

If you’ve been on the fence, I implore you to please take a look at the book and consider ordering. If you’ve been enjoying the book but haven’t told any friends, I implore you to please share your enthusiasm or favorite tips with them in the next 12 hours, if you can. If you really liked it, please consider getting copies as a gift this Saturday.

That said, I really hope you enjoy the following excerpt from the appendices, which explores a common question: isn’t self-experimentation valueless if it’s just one person’s experience?

As we shall see, self-experimentation need not be extreme (I do the extremes so you don’t have to), and you can make significant discoveries with a sample size of one.

I’ll let a professional explain how: Dr. Seth Roberts…

The Value of Self-Experimentation

“All life is an experiment. The more experiments you make the better.” —Ralph Waldo Emerson

“It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If it doesn’t agree with experiment, it’s wrong.” —Richard Feynman

This is an excerpt from The 4-Hour Body, written by Dr. Seth Roberts, professor emeritus of psychology at the University of California–Berkeley and professor of psychology at Tsinghua University. His work has appeared in the New York Times Magazine and The Scientist, and he is on the editorial board of the journal Nutrition.


I started self-experimentation when I was a grad student. I was studying experimental psychology; self-experimentation was a way to learn how to do experiments.

One of my first self-experiments was about acne. My dermatologist had prescribed tetracycline, an antibiotic. Just for practice, I did an experiment to measure its effect. I varied the dosage of tetracycline—the number of pills per day—and counted the number of pimples on my face each morning. First I compared six pills per day (a high dose) and four pills per day (the prescribed dose). Somewhat to my surprise, they produced the same number of pimples. I tried other dosages. Eventually I tried zero pills per day. To my shock, zero pills per day produced the same number of pimples as four or six pills per day. The conclusion was unavoidable: the drug had no effect. (Many years later, research articles about antibiotic-resistant acne began to appear.) Tetracycline is a prescription drug; it’s not completely safe. I’d been taking it for months.

My dermatologist had also prescribed benzoyl peroxide, which comes in a cream. When my self-experimentation started, I believed that tetracycline was powerful and benzoyl peroxide weak, so I rarely used the cream. One day I ran low on tetracycline. Better use the cream, I thought. For the first time, I used the cream regularly. Again I was shocked: it worked well. Two days after I started using it, the number of pimples clearly went down. When I stopped the cream, two days later the number of pimples rose. When I restarted the cream, the number of pimples went down again.

My data left no doubt that (a) tetracycline didn’t work and (b) benzoyl peroxide did work—the opposite of my original beliefs. My dermatologist thought both worked. He’d seen hundreds of acne patients and had probably read hundreds of articles about acne. Yet in a few months I’d learned something important he didn’t know.

This wasn’t the usual line about self-experimentation. Read any book about it, such as Lawrence Altman’s Who Goes First? The Story of Self-Experimentation in Medicine, and you will come away thinking that self-experimentation is done by selfless doctors to test new and dangerous treatments. My experience was different. I wasn’t a doctor. I wasn’t trying to help someone else. I didn’t test a dangerous new treatment. Unlike the better-known sort of self-experimentation, which usually confirms what the experimenter believes, my self-experiments had shown I was wrong.

From my acne research I learned that self-experimentation can be used by non-experts to (a) see if the experts are right and (b) learn something they don’t know. I hadn’t realized such things were possible. The next problem I tried to solve this way was early awakening. For years, starting in my twenties, I woke up early in the morning, such as 4 a.m., still tired but unable to go back to sleep. Only a few dreary hours later would I be able to fall back asleep. This happened about half of all mornings. It showed no sign of going away. I didn’t want to take a pill for the rest of my life—not that there are any good pills for this—so I didn’t bother seeing a doctor. The only hope for a good solution, as far as I could tell, was self-experimentation.

So I did two things:

  1. I recorded a few details about my sleep. The main one was whether I fell back asleep after getting up. How often this happened was my measure of the severity of the problem. In the beginning, I couldn’t fall back asleep about half of all mornings.
  2. I tested possible solutions.

The first thing I tried was aerobic exercise. It didn’t help. Early awakening was just as common after a day with exercise as after a day without exercise. I tried eating cheese in the evening. It didn’t help. I tried several more possible remedies.

None helped. After several years, I ran out of things to try. All my ideas about what might help had proved wrong.

Yet I managed to make progress. For unrelated reasons, I changed my breakfast from oatmeal to fruit. A few days later, I started waking up too early every morning instead of half the time. The problem was now much worse. This had never happened before. I recorded the breakfast change on the same piece of paper I used to keep track of my sleep, so the correlation was easy to see. To make sure the correlation reflected causality, I went back and forth between fruit and oatmeal. The results showed it was cause and effect. Fruit for breakfast caused more early awakening than oatmeal for breakfast. After ten years when nothing I’d done had made a difference, this was a big step forward. I eventually figured out that any breakfast made early awakening more likely. A long experiment confirmed this. The best breakfast was no breakfast.

I was less surprised than you might think. I knew that in a wide range of animals, including rats, a laboratory result called anticipatory activity is well established. If you feed a rat every day at the same time, it will become active about three hours earlier. If you feed it at noon, it will become active about 9 a.m. I had been eating breakfast at about 7 a.m. and waking up about 4 a.m. I had essentially found that humans were like other animals in this regard.

Not eating breakfast reduced early awakening but didn’t eliminate it. In the following years, self-experimentation taught me more about what caused it. By accident, I found that standing helped. If I stood more than eight hours in a day, I slept better that night. That wasn’t practical—after trying to stand that much for several years, I gave up—but the realization helped me make another accidental discovery 10 years later: standing on one leg to exhaustion helps. If I do this four times (left leg twice, right leg twice) during a day, even in the morning, I sleep much better that night. More recently, I’ve found that animal fat makes me sleep better.

Both effects are dose-dependent. I can get great sleep if I stand enough and great sleep if I eat enough animal fat.

How much animal fat is “enough”? I’ve just started trying to figure this out using pig fat, which I consume in a cut called pork belly (the part of the pig used for bacon). I found that 150 grams of pork belly had a little effect; 250 grams of pork belly had a much clearer effect. The effect seems to get larger with more pork belly (e.g., 350 grams). Because pork belly may be more than 90% fat by calories (there is great variation from one piece to the next), it’s a lot of calories of fat to get the maximum possible effect. I need to burn a lot of calories per day to make that many calories easy to eat, but it’s in some respects more convenient than standing on one foot.

Acne and sleep were my first self-experimental topics. Later I studied mood, weight control, and the effects of omega-3 on brain function. I learned that self-experimentation has three uses:

  1. To test ideas. I tested the idea that tetracycline helps acne. I tested ideas about how to sleep better. And I’ve tested ideas derived from surprises. A few years ago, while trying to put on my shoes standing up, I realized my balance was much better than usual. I’d been putting on my shoes standing up for more than a year; that morning it was much easier than usual. The previous evening I’d swallowed six flaxseed-oil capsules. I did self-experiments to test the idea that flaxseed oil improves balance. (It did.)
  2. To generate new ideas. By its nature, self-experimentation involves making sharp changes in your life: you don’t do X for several weeks, then you do X for several weeks. This, plus the fact that we monitor ourselves in a hundred ways, makes it easy for self-experimentation to reveal unexpected side effects. This has happened to me five times. Moreover, daily measurements—of acne, sleep, or anything else—supply a baseline that makes it even easier to see unexpected changes.
  3. To develop ideas. That is, to determine the best way to use a discovery and to learn about the underlying mechanism. After I found that flaxseed oil improved balance, I used self-experimentation to figure out the best dose (three to four tablespoons per day).

One complaint about self-experimentation is that you’re not “blind.” Maybe the treatment works because you expect it to work. A placebo effect. I have never seen a case where this appeared to have happened. When treatment 10 helps after treatments 1 through 9 have failed to help (my usual experience), it’s unlikely to be a placebo effect. Accidental discoveries cannot be placebo effects.

My experience has shown that improve-your-life self-experimentation is remarkably powerful. I wasn’t an expert in anything I studied—I’m not a sleep expert, for example—but I repeatedly found useful cause-and-effect relationships (breakfast causes early awakening, flaxseed oil improves balance, etc.) that the experts had missed. This isn’t supposed to happen, of course, but it made a lot of sense. My self-experimentation had three big advantages over conventional research done by experts:

  1. More power. Self-experiments are far better at determining causality (does X cause Y?) than conventional experiments. Obviously they’re much faster and cheaper. If I have an idea about how to sleep better, I can test it on myself in a few weeks for free. Conventional sleep experiments take a year or more (getting funding takes time) and cost thousands of dollars. A less obvious advantage of self-experimentation is that more wisdom is acquired. We learn from our mistakes. Fast self-experimentation means you make more mistakes. One lesson I learned stands out: Always do the minimum—the simplest, easiest experiment that will make progress. Few professional scientists seem to know this. Finally, as I mentioned earlier, self-experimentation is much more sensitive to unexpected side effects.
  2. Stone Age–like treatments are easy to test. I repeatedly found that simple environmental changes, such as avoiding breakfast and standing more, had big and surprising benefits. In each case, the change I’d made resembled a return to Stone Age life, when no one ate breakfast and everyone stood a lot. There are plenty of reasons to think that many common health problems, such as diabetes, high blood pressure, and cancer, are caused by differences between modern life and Stone Age life. Modern life and Stone Age life differ in many ways, of course; the fraction of differences that influence our health is probably low. If so, to find aspects of Stone Age life that matter, you have to do many tests. Self-experiments, fast and cheap, can do this; conventional experiments, slow and expensive, cannot. In addition, conventional research is slanted toward treatments that can make money for someone. Because conventional research is expensive, funding is needed. Drug companies will fund research about drugs, so lots of conventional research involves drugs. Elements of Stone Age life (such as no breakfast) are cheap and widely available. No company will fund research about their effectiveness.
  3. Better motivation. I studied my sleep for 10 years before making clear progress. That sort of persistence never happens in conventional health research. The reason is a difference in motivation. Part of the difference is how much the researcher cares about finding solutions. When you study your own problem (e.g., acne), you care more about finding a solution than others are likely to care. Acne researchers rarely have acne. And part of the motivation difference is the importance of goals other than solving the problem. When I studied my sleep, my only goal was to sleep better. Professional scientists have other goals, which are enormously constraining.

One set of prison bars involves employment and research funding. To keep their jobs (e.g., get tenure, get promoted, get jobs for their students, and get grants), professional scientists must publish several research papers per year. Research that can’t provide this is undoable. Another set of prison bars involves status. Professional scientists derive most of their status from their job. When they have a choice, they try to enhance or protect their status. Some sorts of research have more status than others. Large grants have more status than small grants, so professional scientists prefer expensive research to cheap research. High-tech has more status than low-tech, so they prefer high-tech. As Thorstein Veblen emphasized in The Theory of the Leisure Class (1899), useless research has higher status than useful research. Doing useless work, Veblen said, shows that you are higher-status than those who must do useful work. So researchers prefer useless research, thus the term “ivory tower.” Fear of loss of job, grant, or status also makes it hard for professional scientists to propose radical new ideas. Self-experimenters, trying to solve their own problem on their own time, are not trapped like this.

Acne illustrates the problem. The dermatological party line is that diet doesn’t cause acne. According to a website of the American Academy of Dermatology, “extensive scientific studies” show it’s a “myth” that “acne is caused by diet.” According to “guidelines for care” for dermatologists published in 2007, “dietary restriction (either specific foods or food classes) has not been demonstrated to be of benefit in the treatment of acne.” In fact, there is overwhelming evidence linking diet and acne. Starting in the 1970s, a Connecticut doctor named William Danby collected evidence connecting dairy consumption and acne; it is telling that Danby wasn’t a professional scientist. When his patients gave up dairy, it often helped. In 2002, six scientists (none a dermatologist) published a paper with the Weston Price–like conclusion that two isolated groups of people (Kitava Islanders and Ache hunter-gatherers) had no acne at all. They had examined more than 1,000 subjects over the age of 10 and found no acne. When people in these groups left their communities and ate differently, they did get acne. These observations suggest that a lot of acne—maybe all of it—can be cured and prevented by diet.

Why is the official line so wrong? Because the painstaking research needed to show the many ways diet causes acne is the sort of research that professional researchers can’t do and don’t want to do. They can’t do it because the research would be hard to fund (no one makes money when patients avoid dairy) and because the trial and error required would take too long per publication. They don’t want to do it because it would be low-tech, low-cost, and very useful—and therefore low-status. While research doctors in other specialties study high-tech expensive treatments, they would be doing low-cost studies of what happens when you avoid certain foods. Humiliating. Colleagues in other specialties might make fun of them. To justify their avoidance of embarrassment, the whole profession tells the rest of us, based on “extensive scientific studies,” that black is white. Self-experimentation allows acne sufferers to ignore the strange claims of dermatologists, not to mention their dangerous drugs (such as Accutane). Persons with acne can simply change their diets until they figure out what foods cause the problem.

Gregor Mendel was a monk. He was under no pressure to publish; he could say whatever he wanted about horticulture without fear for his job. Charles Darwin was wealthy. He had no job to lose. He could write On the Origin of Species very slowly. Alfred Wegener, who proposed continental drift, was a meteorologist. Geology was a hobby of his. Because they had total freedom and plenty of time, and professional biologists and geologists did not (just as now), Mendel, Darwin, and Wegener were able to use the accumulated knowledge of their time better than the professionals. The accumulated knowledge of our time is more accessible than ever before. Self-experimenters, with total freedom, plenty of time, and easy access to empirical tests, are in a great position to take advantage of it.

The above is an excerpt from the new book The 4-Hour Body


Tools and Tricks

Seth Roberts, “Self-Experimentation as a Source of New Ideas: Ten Examples Involving Sleep, Mood, Health, and Weight,” Behavioral and Brain Science 27 (2004): 227–88 ( This 61-page document about self-experimentation provides an overview of some of Seth’s findings, including actionable sleep examples.

The Quantified Self ( Curated by Wired cofounding editor Kevin Kelly and Gary Wolf, a managing editor of Wired, this is the perfect home for all self-experimenters. The resources section alone is worth a trip to this site, which provides the most comprehensive list of data-tracking tools and services on the web (

Alexandra Carmichael, “How to Run a Successful Self-Experiment” ( Most people have never systematically done a self-experiment. And yet, it’s one of the easiest methods for discovering what variables are affecting your well-being. This article shows you the five principles that will help you get started in running successful self-experiments. Bonus: an 11-minute video from Seth Roberts, discussing experiment design.

CureTogether ( CureTogether, which won the Mayo Clinic iSpot Competition for Ideas That Will Transform Healthcare (2009), helps people anonymously track and compare health data to better understand their bodies and make more informed treatment decisions. Think you’re alone with a condition? Chances are you’ll find dozens of others with the same problem on CureTogether.

Daytum ( Conceived by Ryan Case and Nicholas Felton, Daytum is an elegant and intuitive service for examining and visualizing your everyday habits and routines.

Data Logger ( Data Logger for iPhone enables you to store and graph any data of your choosing along with a time-stamp and location. It can be used for anything, whether food-related, animal sightings, or temperature sensor readings around your neighborhood. If you can think of it, it can be recorded and tracked.


  1. [How Seth Roberts’ self-experimentation began]. Roberts, Seth.  Surprises from self-experimentation: Sleep, mood, and weight. Chance.  2001; 4(2):7-18. UC Berkeley: Available from:
  2. Roberts, Seth. Self-experimentation as a source of new ideas: Ten examples about sleep, mood, health, and weight. Behavioral and Brain Sciences.  2004; 27(2), 227-288. UC Berkeley: Available from

Original article at