Seth's Blog

At Tsinghua University, which is like a Chinese MIT, I am teaching a small class (25 students) called Frontiers of Psychology. It is required of freshmen psychology majors. There are a few students from other majors. So many of my students do brilliant work that it is hard to keep track. For example, two classes ago I started having presentations (short talks related to the reading). In the very first one, a student talked about her dysmenorrhea and self-experimentation to stop it. Later, during a discussion of how to give a talk, another (female) student said, “I could not have given such a talk.” “That’s a compliment, right?” I said. “I don’t know,” she said. Which is only to say what a radical and stunning talk it was.

For this week’s class I assigned several readings, from which students chose one. The shortest and most popular paper, by Joshua Knobe, a Yale professor of philosophy, was about judgments of intentionality. Knobe showed subjects various scenarios and asked them whether the side effects of a action described in the scenario should be considered intentional or not. Changing one word had a big effect. Knobe concluded that we tend to see bad side effects as intentional, good side effects as unintentional. I assigned it because the effect of changing one word was large and I liked the source of data (“Subjects were 78 people spending time in a Manhattan public park”).

Here is one student’s comment:

When I was in primary school, we had a very kind English teacher who was quite close to me. After she left school, she sent some photos to me and I found it a great honor to deliver them to my classmates. Later on, a math teacher got married and she gave another pupil some sweets to deliver the class. I felt unpleasant since not every student could get a sweet. I thought it unjust.

However, in both cases, photos and sweets, there weren’t enough for the whole class. The only difference was who passed them out. When I did, the main issue I cared about was “I’m the one to deliver them”; in the other case, “Why can’t everyone get one?”

She titled her comment “We are the Heroes, They are the Villains”. Her point was that Knobe’s results could be explained by the idea that we slant our judgments of others and ourselves to make them look worse and us look better — an explanation that Knobe didn’t consider.

Knobe isn’t the only one who didn’t think of it. Other students proposed other plausible explanations. But I think the “we are heroes” explanation is quite plausible because three other students made the same point in other ways. One of them repeated a story from a test preparation book:

A teacher had a student do ten math problems on the board. Then she asked another student to describe what he saw. “Two of the answers are wrong,” he said. “What about the eight correct answers?” said the teacher.

Not a true story but surely based on actual events. Another student told of the time her teacher had made her push her fellow students to exercise for an half-hour per day. The students complained to her about their loss of time. Later, however, her class had finished first in a physical competition — much better than usual. Her classmates did not give her any credit for this.

To emphasize how unobvious this idea is, here is what two professors make of Knobe’s results:

This asymmetry in responses between the ‘harm’ and ‘help’ scenarios, now known as the Knobe effect, provides a direct challenge to the idea of a one-way flow of judgments from the factual or non-moral domain to the moral sphere. ‘These data show that the process is actually much more complex,’ argues Knobe.

My students disagree. Their proposed explanations, such as the “we are heroes” idea, were not “much more complex”.

I believe they have noticed a broad truth about human nature that has escaped many psychologists, not just Knobe. In this excerpt from his new book, my former colleague Danny Kahneman describes what he calls “the illusion of validity”: personality judgments were considered more predictive than they actually were by the people who made them. Could this be another example of “we are heroes”? The “we are heroes” idea also explains the Lake Wobegone Effect: Most people consider themselves above average. The technical name for this is illusory superiority. The Wikipedia article about illusory superiority does not mention the Knobe Effect and vice-versa. In this important aspect of human nature, professors (including me) have had trouble seeing that the trees make a forest.

About fifteen years ago I had lunch with Richard Muller, a Berkeley professor of physics, at the Berkeley Faculty Club. He told me his theory that the “miracles” that the Bible says Jesus performed, such as changing water into wine, were magic tricks. He was writing a novel about it, he said. He also said he had submitted to Science a new theory of climate change based on Milankovitch cycles (cycles of changes in the Earth’s distance and tilt relative to the sun). The editor liked it; the problem was getting it past the reviewers. This press release shows the editor succeeded. So Muller was nice enough or curious enough to have lunch with a stranger (me) who could not possibly help him and was/is creative about big questions. He is now retired. He’s had great career success, including a MacArthur Fellowship (in 1982). He’s won a teaching award. A talented and decent person. (Steve McIntyre, whose comment I read after I wrote this, also says good things about Muller: “one of the few people in this field I regard as a friend.”)

Two years ago he started the Berkeley Earth Surface Temperature Project, devoted to improving the climate record. Fine. In March I liked a talk he gave about climate change. Fine.  Now he has done something astonishing. In a recent Wall Street Journal article titled “The Case Against Global-Warming Skepticism” he took “skepticism about global warming” to be skepticism that the Earth has warmed recently. In it, he describes several problems with surface temperature measurements. Then he says:

Without good answers to all these complaints, global-warming skepticism seems sensible. But now let me explain why you should not be a skeptic, at least not any longer.

The vast majority of skeptics, including me, believe the Earth has warmed substantially since the Little Ice Age. That’s not the issue. Here’s the issue: We are skeptical that we understand why it has warmed and in particular skeptical that humans have caused recent warming.  A big difference. Muller has ignored  the obvious: what skeptics actually think.

Muller’s view of “global warming skepticism” is so strange let me state what might be obvious. For me, and many others, there are three issues: 1. Can we trust climate models? I say no: They have never been shown to be good predictors of what they are being used to predict. The physics of clouds isn’t simple or well-understood. 2. Is it unusually hot now? I say no: The Medieval Warm Period was roughly as hot or hotter. 3. Has recent warming been unusually fast? (Which is what Michael Mann’s discredited Hockey Stick seemed to show.) I say no. Over the past 200 years, the temperature has increased as fast or faster at least twice. Muller’s new data doesn’t address any of these concerns. Whether surface temperatures are higher now than in 1950 (which is what Muller’s new data shows more conclusively than before) is not a big issue.

Why did Muller misrepresent so badly what skeptics say? I don’t know. Maybe he wanted to make his results seem more important than they are. Maybe he has never met a skeptic. I truly don’t know.  Lots of famous scientists (e.g., James Watson) have said what I consider wacky things about unverifiable stuff. But there is nothing vague or unverifiable about this. It is as if Muller had said Shanghai is the capital of China.

James Fallows, whose work I like, has taken Muller seriously. Paul Krugman has taken Muller seriously. Marc Morano, who runs Climate Depot, has responded at length and created a special Muller page.  In March, Morano points out, he (Morano) complained about exactly the same thing from Muller: “Who denies that warming has taken place?” Yes. Morano links to many scientists who are displeased by what Muller has done. One says, “It is not true that the Berkeley group has found relevant evidence for the core questions in the AGW debate.” Yes. “Doubts about the validity of the surface temperature record constitute something like 1% of the issues that climate skeptics as a community have ever raised.” Yes.

Muller’s error interests me because I can’t explain it. Perhaps  it illustrates how unwittingly we shape reality, as shown in a famous split-brain anecdote:

The split-brain patient had to point with his two hands at pictures of two objects corresponding to two images that he had seen on the divided screen (one with each of his two separated hemispheres). The patient’s left hand [pointed] at the card with a picture of a snow shovel, because the right hemisphere, which controls this hand, [had] seen the projected image of a winter scene. [The left hemisphere had seen a picture of a chicken. When asked why he chose a shovel, the patient said (via the left hemisphere, which controls speech):] you use a shovel to clean out the chicken house.

Split-brain patients do not have more mental tricks than the rest of us. Surely we all do this. My question is: When?

Thanks to Tim Beneke.

I have had tea daily for the last ten years, ever since I discovered the Shangri-La Diet. A few weeks ago, I started doing side-by-side comparisons of similar teas or the same tea prepared two ways (e.g., different brewing times). Would the Willat Effect make me a tea connoisseur?

Since then I have done at least one side-by-side comparison every day. It’s almost as easy as making an ordinary cup of tea and a lot more fun. These comparisons have taught me more about tea preparation than the previous ten years. I’ve learned:

1. The black tea I have (an Earl-Grey variant) tastes better when brewed for 3.5 minutes than 4.0 minutes.

2. The black tea tastes better when I use 1.5 grams of tea than when I use 2.0 grams of tea. (After starting these comparisons, I bought a scale for weighing tea.)

3. One of the green teas I have tastes better when “rinsed” for 30 seconds before brewing 1 minute than when simply brewed for 1 minute. In China, this preference (rinse green tea before brewing) is common. I was reminded of it by this comment and Paul Jaminet’s post about tea. Black tea is different, as I noted earlier.

4. I have a caffeine-free tea blend called Choco Late made of cacao husks, vanilla, and rooibos. The package says brew 5 minutes. Which is nonsense. It tastes better (fuller, more rounded) when brewed 30 minutes than when brewed 15 minutes. (I’ve noticed the same thing with caffeine-free chai blends. Enormous brewing times, like 60 minutes, produce much better results than short times.)

5. My most interesting discovery is when I brew Choco Late for 30 minutes it tastes so good I no longer want to sweeten it. It is pleasant enough already and sweetness would distract from the complexity, fullness, and slight bitterness. (At first I wrote “lovely complexity, fullness …”) I was shocked when I noticed this. It has never happened before.

This tea-selling website mentions the Willat Effect under the heading “Do you want to be a tea connoisseur?” I hope this means the idea will spread among the fancy-food community. They have a lot to gain from better understanding of how to make people connoisseurs. Many times I have asked people in that community what makes someone a connoisseur? The usual answer is education. In my case, Willat-Effect comparisons (side-by-side comparisons of similar teas) were far more powerful than reading about tea, drinking a variety of teas, going on tea tours, going to ordinary tea tastings (where you taste a wide range of teas), and talking about tea with experts. I have been to five or six Fancy Food Shows and have visited thousands of booths. Exactly one booth offered side-by-side comparisons of similar products. It was their product made with and without a special ingredient.

Willat-Effect comparisons are mini-science. They aren’t quantitative but they include three other things central to science: 1. Close comparisons. This is the essence of experimentation. 2. You don’t know the answer. 3. You care about the answer.

At the Shangri-La Diet forums, Anima writes:

I have been diagnosed with ADHD and Bipolar II disorder.  I am also a Non-24, a chronic circadian rhythm disorder where one’s body thinks a day is longer than 24 hours. . . .I’ve been using amber safety glasses (around $3 in the hunting section of the sporting goods store) for dark therapy.  I put them on 3 hours before I want to go to sleep.  They block blue light, allowing dark therapy without the dark.  I also wear an eye mask while I sleep.  The glasses make me look like a big weirdo, but they really work.  It’s easier to get to sleep, and they prevent hypomania (the milder form of mania that people with Bipolar II experience) better than any medication I have tried.  It makes sense that almost anyone could benefit from them, because our ancestors were not exposed to blue light after dark. (more…)

• Interview with me on Jimmy Moore’s Livin’ La Vida Locarb
• This article about natto helped its author win a prize for best newspaper food column
• great QS talk about self-measurement by John Sumser. “It all started when I quit smoking. Bad idea. Since I quit smoking in 2004, every quarter for 7 years it has rained shit on me.”
• In a QS talk, I compare the Quantified Self movement and the paleo movement.
• Chinese high-school students in America: Not what was promised. Lack of “rigor” has benefits, as I have blogged: “Dismayed by the school’s [poor] college placement record, Chen considered transferring. Instead, he began to enjoy himself. Because his courses were undemanding, he had time for friends and outside interests. He took four Advanced Placement tests on his own.“I’ve developed my personality a lot,” Chen said. “Everything turned out for the best.””
• If you read The China Study by T. Colin Campbell, a pro-vegetarian book, you may remember the big role played by some casein experiments with rats. Rats that ate a low-casein (= low animal-protein) diet were supposedly in better health than rats that ate a high-casein (= high animal-protein) diet. In this article Chris Masterjohn shows how misleading that was. “One thing is certain: low-protein diets depressed normal growth, increased the susceptibility to many toxins, killed toxin-exposed animals earlier, induced fatty liver, and increased the development of pre-cancerous lesions when fed during the initiation period of chemical carcinogenesis.”

Thanks to Janet Chang.