Steve, the introverted farmer, bugs me so much that I think it's invalid as an example of the phenomenon. The study was conducted by presenting people a word problem describing Steve. The problem is that we've all got 12 years of schooling, training us that word problems are not objective descriptions of the world - they're carefully worded clues that are supposed to point you towards an answer. So if a word problem expects you to pull in your outside knowledge of statistics about how male farmers outnumber male librarians 20:1, then the reader has been groomed to expect some tip of the hat that the problem relies on outside information. In essence, it's a trick question, because it's not playing by the implied rules of the format, and so it illustrates jack squat.
There are twenty times as many farmers as librarian, so introvert farmers must outnumber librarians.
This can't possibly be true. I know many librarians, but I've never met a single farmer.
1: Maybe, but it illustrates well the broader principle of least effort. Reaching for the pattern-matching answer is easy, thinking about population sizes isn't. Also, the social cues don't necessarily push the way you say. If one presents an apparently easy question to an adult, the adult can reasonably be expected to think it's a trick question, and think before answering. Yet, errors. Kahneman mentions this line of thinking in the section on the ball-and-bat problem.
BTW, I'm surprised there are that many librarians in the US.
The thing that blew me away was the glucose. That shit needs to be operationalized. How much glucose are we talking? My job is basically maintaining motivation through a series of smallish moderately difficult mental tasks, and refraining from killing people who deserve to be killed. I need a non-fattening brain food regimen.
My family has more farmers than librarians.
: Maybe, but it illustrates well the broader principle of least effort. Reaching for the pattern-matching answer is easy, thinking about population sizes isn't.
Not if it's gaming out the situation and trying to figure out what the researcher is expecting. Then we have no idea which system the reader is using.
I had a few thoughts reading these chapters:
Ch 1: First, reading the S1 vs. S2 descriptions, I kept thinking "Donald Trump is all S1 and no S2". The more descriptions came up, the more correct it seemed.
Ch 2: I wondered about the mechanics of the pupillometry experiments - they were done in labs, with indoor lighting? Would the same thing happen outdoors, with brighter light, and at what point would it actually cause too much light to enter the eye? Or is the effect here just smaller?
Ch 3: The parole study thing was interesting to see, though I thought I had seen it seriously questioned more recently - that there were effects from timing and how long each case took and so on that could explain things just as well. Can't find the link now, of course.
I was curious about the mention of "flow", though it's kind of a side note here. In particular, I was thinking about how it relates to video-game playing, which I think of as more S2-like, but maybe that's wrong. The steady stream of internally-generated rewards (dopamine, rather than glucose?) might help keep focus going. Alternatively, there's something else about compulsive/addictive behavior that needs to be worked into here.
7: Doublethinking every survey response ever collected is a quick road to madness. Specifically here, you think eagerness to please overwhelms the desire to be right, which I think in American adults it won't be. You could check for an effect with double-blinded and non-double-blinded researchers asking the questions.
I was thinking about how it relates to video-game playing, which I think of as more S2-like, but maybe that's wrong.
I think video games illustrate the process by which S2 thinking becomes S1. At the beginning of the game you have to spend a bunch of time figuring out how to do stuff, but with practice you get better and more instinctual at it, such that S1 can take over.
8 last: I was wondering about flow v glucose too. The best flow I've ever had was writing essays in a library where I couldn't do anything but write, and so went for hours without eating anything. Yet I was able to work very effectively, at a demanding level.
Clarifying my 9.1: Heebie is right, subjects gaming experiments is a major problem and needs to be controlled for, but in this case I doubt that it matters.
Heebie is biased by the resemblance to her own students.
I was a bit sceptical of the "Steve" example too. Just because there are more farmers than librarians doesn't mean that an introverted, orderly-minded man is more likely to be one.
Oh, also, Chapter 1 goes out of its way to say that S1 vs S2 "isn't real", and I wondered what that was supposed to mean. That they don't represent physically distinct areas of the brain? If they work as an abstraction for thinking about how brains behave, that seems like all that's necessary for real-ness, but perhaps the author has a different philosophical standard.
14: Yes, there is an implicit assumption that traits are evenly distributed through the population. Which is not unreasonable; to defeat the numbers, social selection among librarians would have to raise introversion to 20 times the background rate. Possible, but the reasonable guess is that it doesn't.
I mean, to take a really extreme example, "Steve is from Wigan, is six foot six, weighs 280 pounds and has a broken nose. Is he more likely to be a rugby league player or a manicurist?" Now, of course, there are lots more manicurists around than there are rugby league players. But, still.
Everyone is getting the Steve example wrong. Do you know anyone who decided to move out to the country and take up farming? Did they strike you as extroverts? Does farming strike you as an occupation that doesn't require careful planning to succeed at?
The puzzle only works if you remember the US is not a meritocracy.
18. What you point out is that the example is wrong, except maybe on the planet Zog, where crops mature by everybody partying among them for months at a time. s/farmers/PR flacks/, is it still wrong?
18: Yes, but I think that "everybody" includes the authors, since they're implicitly agreeing with the association, and using demographics, not actual psychological profile data, to disprove it.
That is, the flaw is supposed to be that people don't consider broader factors beyond their stereotypes, not that people are uninformed about the actual character of farmers.
The transition from controlled experiments with pupillometry to real-world examples is in my mind the main question. The careful descriptions and controlled experiments are super-interesting.
The example of the parole decisions is plausibly explained, but other explanations than attention fatigue leading to S1 dominance are definitely possible there. In particular, both S1 and S2 are not known to be constant over time within a single individual.
I'm comfortable with the idea that I understand Kahneman and Tversky better than they understood themselves.
I have their whole book, from back when I was attempting to apply myself.
Do you know anyone who decided to move out to the country and take up farming?
Yes. Also lots of people who have lived out in the country all their lives.
Did they strike you as extroverts?
Yes, much more so than the average.
Does farming strike you as an occupation that doesn't require careful planning to succeed at?
No; but farming also doesn't strike me as a job that requires an intense commitment to tidiness and order, as witness the state of the average farmyard.
That is, to apply myself by writing a dissertation. I wasn't reading the book for lifehacks.
23: I assumed it was pre-emptive (or post-emptive) defense against strawman attacks: homunculi theories which he mentions, more specific scientific charges like 15 mentions. For heuristic purposes I don't see that it matters.
27.last: I don't think you can figure a lack of tidiness and order from the state of farmyards any more than you can figure from the state of an office worker's desk. You want to see how well the computers that run the tractor are setup.
Anyway, more along the lines of what K&T actually meant, there's no reason to think that farmers are unlikely to be introverts and careful. So the Steve example is very different from 17, where manicurists really are unlikely to be built like a linebacker. The human brain is not saying "No! Farmers are gregarious and sloppy!" We're just bad at accounting for the effect of the size of the relevant population. And we are. They have to teach nurses and doctors and medical researchers not to make the mistake of ignoring the rareness of the condition when we evaluate a positive result in a medical test.
Farmers ad nauseum: my actual thought process on reading the question was roughly: 1. Librarian; 2. Really?; 3. Is it possible for a farmer to match this description?; 4. Yes; 5. Demographics; 6. Farmer. If one had a well-developed habit of assuming mediocrity one could presumably jump from 1 straight to 5.
You can't jump into the same farmer twice.
But if you could, it might well be nauseating.
I am from the 19th century, so I assume everyone is a farmer until they prove otherwise.
8 - (Note: I didn't read these chapters, but know a little about such studies)... The pupil studies would have been conducted indoors under very controlled lighting conditions. Although it's true that things like cognitive effort or emotional arousal will impact pupil size (increase it), effects of lighting will have even larger impacts on pupil size. So if you were outside in bright light, pupils would be very small. They should still dilate under conditions of increasing effort, but not such that you suddenly have really huge pupils that are letting in too much light. We're talking .1 mm to .5 mm increases in pupil diameter in the lab studies. It would also be a poor way to run an experiment, though, due to changing light conditions.
Nathan will set you right in Chapter 11, "Anchors".
manicurists really are unlikely to be built like a linebacker
Ah, but what is that based on? It's not an actual constraint of the job like "basketball players are unlikely to be five foot six" or "postmen are unlikely to use wheelchairs". Enormous people could just as easily be manicurists as could tiny people. What you mean is that in your experience manicurists tend to be small in build.
But an important point is that "most manicurists are short and most rugby players are tall" does not translate into "a tall person is more likely to be a rugby player than to be a manicurist". It's the same thing as that puzzle about false positive rates on tests for a rare disease.
I think if an enormous white man became a manicurist, he would become a small Korean woman.
Enormous people could just as easily be manicurists as could tiny people.
Probably not, in practice, in that it's a customer service job, which generally involves a certain amount of hiring for the image the service wants to present. I would expect that a giant broken-nosed thug, much as he might be skilled at nail art, would have a hard time getting a manicurist job.
Sure, in my experience manicurists are small in build. The question is not just is my folk sense of frequency in the population is right or wrong, but conditional on information do we figure out the correct conditional probabilities.
Isn't the glucose finding a famous example of failed replication?
You could have at least gotten your nose set by a professional.
42: I hope so, because I hate people who eat little snacks all the time and don't want to be one.
8: My eyes are almost always dilated. I have circadian rhythm problems and have to be careful about wearing sunglasses in summer if I want to sleep at night. I kind of wonder if that's related to the pupil thing.
Do you do difficult mental arithmetic problems all the time?
If I ever for some reason have occasion to do some screenwriting for teevee, I'm totally going to write in a giant broken-nosed thug character who has an uncommon talent for nail art. If it's a superhero show, maybe he even can turn himself into a little Korean woman.
42: The glucose finding w/respect to Baumeister's willpower stuff has defo been called into question (to put it gently) but to my knowledge this decision fatigue finding hasn't and I'm not sure that even if Baumeister's program is ultimately determined to be entirely bullshit that the entire notion of a bloodsugar/cognitive fatigue relationship can be discounted.
49: No. But I am tired a lot. Given the fact that my eyes are almost always dilated (remarked upon by optometrists and my PCP once), I wonder what they would look like while I'm doing difficult arithmetic problems.
My absolutely favourite glucose experiment is the one that has people judge the slope of a hill by eye -- after a shot of Ribena they estimate the hill to be shallower. But they are also asked to estimate it by feel, so to say, adjusting a lever so that it has the same slope as the hill they can see. And that test is completely unaffected by glucose.
This is more about embodied cognition than s1 and s2, and I can't off the top of my head remember whether the pre- or post-glucose estimates were more accurate. I think it was the post-glucose ones.
If I ever for some reason have occasion to do some screenwriting for teevee, I'm totally going to write in a giant broken-nosed thug character who has an uncommon talent for nail art.
Definitely - it has a Pratchett feel to it. After all, the screens are already full of tiny women who can outfight enormous thugs. No reason the reverse should not also be true. I knew a manicurist once who (though admittedly female and not large) ended up in Iraq with the Parachute Regiment.
it has a Pratchett feel to it.
A bit like Mr Tulip.
52 should be wearing a big sign: "Replication crisis! Take me now!"
I developed my own theory for many of the experimentstate described in the book -- AASQ--GASA. Ask a stupid question -- get a stupid answer.
57: My phone is determined to make me seem even stupider than. I am.
I bet your phone doesn't even have blood sugar.
Is your phone running System 1 or System 2?
I had one thought while I was doing my write-up. (Yes, exactly one thought. I don't multitask.) The hot thing in AI these days is deep learning, which seems to be good at discovering patterns when given massive amounts of data. The exact thing it is good at seems pretty similar to the tasks performed by System 1 -- image recognition, voice recognition, etc. It's bad at System 2 tasks. For example, at playing most Atari 2600 games it is better than human, but it's bad at games like Pac-Man, which require a certain amount of tactical planning.
Possibly it's just afraid of ghosts.
63 is interesting. It's good to know that, with computer learning, we can make incorrect snap judgments faster than ever.
That's a real danger. People were able to train Microsoft's chat bot into spouting white supremacist bullshit within a couple of days.
Honestly, plenty of people take much less time to train to repeat that stuff.