Join 3,556 readers in helping fund MetaFilter (Hide)


You give me two notes and I'll give you one shiny coin.
May 16, 2013 9:13 PM   Subscribe

Behavioral Economics for Kids [pdf] is a free ebook from the Ivey School of Business that illustrates (to adults, really) the basic principles of behavioral economics, including the Endowment Effect (we value what we have more than what it is worth), Hyperbolic Discounting (the time we wait for rewards influences value in non-linear ways), the dishonesty of honest people, and Base Rate Neglect (why we make bad assumptions based on inherent biases). Though the findings are well-established, the labeling is subject to change, as many social psychologists argue, this is not behavioral economics, it is well-established psychology.
posted by blahblahblah (30 comments total) 78 users marked this as a favorite

 
Metafilter: prefers earning $100 while another person earns $0, to earning $200 when the other person gets $300.
posted by michaelh at 9:27 PM on May 16, 2013


This is totally going on the reading list for my next decision making class.
posted by mixing at 9:33 PM on May 16, 2013


Why would something being "well-established psychology" preclude it from being "behavioral economics"?
posted by yoink at 9:33 PM on May 16, 2013 [4 favorites]


People seem to prefer "either/or" to "both/and."
posted by Rustic Etruscan at 9:38 PM on May 16, 2013 [1 favorite]


Why would something being "well-established psychology" preclude it from being "behavioral economics"?

As an academic in the area, it is a bit insidery as to why this matters to some people. Essentially, economics is often viewed as an ascendent discipline in the social sciences, especially as micro-economics has moved away from pure rational actors to embrace the many biases that influence how decisions are made. Some psychologists (and sociologists) view behavioral economics as merely repackaging or rediscovering well-known principles in their fields. This annoys them, because they view economics as getting credit for these concepts, and they believe that it leads to a perceived undermining of the importance of their fields relative to economics, in policy circles and within universities.

The degree to which you care about this depends on your answers to two questions: 1) Are you an academic? 2) Are you a sociologist or social psychologist? If you answer "no" to both of these questions, then you can leave the disciplinary fight alone.

Besides, the trend has become to be more integrative across disciplines among younger academics. I was trained as an economic sociologist, and I write papers with psychologists and economists (though usually not at the same time). So, as I said, this is mostly a literally academic issue, but I wanted to acknowledge it in the FPP. Also, the linked piece, by Nobel winner Daniel Kahneman, explains the issues about credit and cooperation among disciplines quite well.
posted by blahblahblah at 9:43 PM on May 16, 2013 [7 favorites]


Why would something being "well-established psychology" preclude it from being "behavioral economics"?

Daniel Kahneman gives the answer to this question in TFA: the work came out of psychology, and it's not great from our perspective to see our work being relabelled as belonging to another discipline just because someone else needs it. After a while, one gets the sense that "psychology" is a term that means "everything that psychologists do that isn't valuable to someone else". Oddly enough, this gets us kind of annoyed.
posted by mixing at 9:46 PM on May 16, 2013 [1 favorite]


Or, what blahblahblah said, I guess.
posted by mixing at 9:51 PM on May 16, 2013


I guess it's not that verboten to derail this towards talking about Kahneman, since he basically pioneered this stuff.

I got recommended Daniel Kahnemann's book as a fundamental way to look at people differently by a math person, by a rich startup person, and some other startup person.

His work seems to be quite popular in general across the startup set, I don't know why.

Perhaps it's his frequent collaborator, Amos Tversky. He did a lot of the measure theory stuff that influence the mathematical thinking (about ordering and our perception) that underlies this stuff. Measure theory is the study of ways to assign numbers to a set. Trivial for a set like, say, {1, 2, 3}, but what happens if your set is something like {x | x is under the curve (1 / x)}?

And all that abstruseness leads to stuff like his work on measuring probability and utility in the same experiment, and multidimensional representation, which leads to subjective probability, which leads to asking hard questions about what the hell utility is, actually, when we all come down to it. So the mathematical underpinnings are pretty cool, and it's something of same mathematical underpinnings as the recent coolness in machine learning (an influence diagram, after all, is just a bayes net (belief net) with decisions stuck on it, and Geoffrey Hinton did not call his spiffy new thing "deep belief net" for not reason). If you're willing to look at it long and sideways, and stuff your thoughts into some weird holes, this is an interesting look at some psychology/economics which informs AI.

Perhaps because it gives you a nice sense of how you can convince people legitimately. Of course, all the libertarian paternalism "reword this tax form to get 3% more people to respond properly so that you get a billion dollars more in taxes" stuff is based on that work, and I suppose it's quite the salve for people who want to go into business on a semi-ethical basis.

Perhaps I'm taking this assumption from out of a hat, and it's simply an intellectual phenomenon, which it is: stranger acts of beanplating have happened.
posted by curuinor at 10:19 PM on May 16, 2013 [1 favorite]


Oh, man, that hyperbolic discounting article is bad. Disregard just, like, all of it.
posted by logicpunk at 10:27 PM on May 16, 2013 [3 favorites]


I'm fascinated by that base rate neglect stuff. One thing that puzzles me, in the two examples they assume that the probabilities match going both directions, and I'm not convinced that makes sense.

For example, in the taxicab problem.. What if the witness only incorrectly identifies a green cab as blue 10% of the time, but incorrectly identifies a blue cab as green 30% of the time?
(which, if the test used equal numbers of cabs of each colour, would add up to 20% error rate in aggregate.)

Then the wiki list of probabilities would become:Interesting that the witness is still much less reliable than a naive interpretation would assume. Not nearly as bad as a naive application of the theory assumes though..

Am I missing something?
posted by Chuckles at 11:04 PM on May 16, 2013 [1 favorite]


No, you're not missing anything, and it touches on a bit of a worry with the base rate neglect phenomenon itself.

The classic taxi cab scenario assumes that the witness is unbiased: Pr( says green | is blue ) = Pr( says blue | is green ). I've never been completely convinced that this is a fair assumption to make of real world witnesses. In most real-world judgments from memory, people do rely (in part) on their own knowledge of base rates. So it's almost certain that, if the witness is actually a resident in the city where the accident took place, he or she will already have incorporated the base rate into the judgment. For instance, if you witness an accident in New York (or at least "mixing's mental model of New York in which NY taxis really are as yellow as they appear on TV") and can't quite remember the taxi, odds are you'll say "yellow" when asked by an investigator. Later on, when the witness is tested in "controlled" circumstances, the base rates used in the test differ from those that hold in the real world, and so the witness adjusts his or her criterion for saying "yellow" (this happens unconsciously, but there is a mountain of evidence showing that people do this in standard memory experiments). So the "controlled test" of the witness' ability (in which base rates are equal) acts to hide the fact that the witness has intuitively relied on the real world base rates when making the original judgment.

Given this, it's not at all obvious what the decision maker should do. He/she needs to know something about how closely the controlled test matches the real world scenario that the witness experienced, and moreover he/she needs to know whether or not the witness' memory was influenced by the exact same base rate that the scenario presents. Because if the witness was relying on his or her knowledge of real world taxi colours, then the decision maker is not allowed to use that base rate, and thus base rate neglect becomes rational!

All of this is a way of coming around to one of the main ideas about why base rate neglect happens, namely that people don't intuitively perceive it to be relevant to the decision. When shown the supposedly-compelling logic of Bayes' rule they're forced to agree that they were wrong (especially since that would entail arguing with Amos Tversky about maths), but it's actually not clear whether they were. I suspect that the truth lies in the middle here... people do seem to gloss over the base rate information when reading these scenarios (they do better when the base rate is something learned from personal experience rather than read in a book), and so there's a sense in which people are doing the "wrong" thing by ignoring it, but I think it's very very likely that a naive application of Bayes' rule here in real life would actually give rise to judgments that double-count the base rate.

(Yes, I have spent too much time thinking about the taxi cab problem. I know that no-one else really cares. I can't help myself though. It's so much fun)
posted by mixing at 11:26 PM on May 16, 2013 [7 favorites]


Why would something being "well-established psychology" preclude it from being "behavioral economics"?
Because "behavioral economists" are acting like they've discovered something when really all they are is ignorant of what's already been well studied in other fields.
posted by delmoi at 11:27 PM on May 16, 2013 [2 favorites]


Let’s say you have just met with your doctor who has informed you that you have tested positive for a typically fatal disease. To make things worse, this test is accurate 95% of the time. ... If the prevalence of the disease is 1 in 1000, the likelihood that you actually have the disease (based on this test) is less than 2%.
You walk into a doctors office after noticing that you have purple spots. The doctor gives you a test, which is 95% accurate, that confirms that you have purple spot disease. Since only 1:1000 people in the general population get purple spot disease, there is a less than 2% chance that you have purple spot disease.
(based on this test)
That's a cute way to completely absolve yourself from shitty assumptions, poor extrapolation, sample errors, and a complete disregard for the diagnostic practices of medical professionals and/or reality.
posted by clearly at 1:08 AM on May 17, 2013 [1 favorite]


I do think it's interesting that so many of these things are economists basically talking about how the fundamental axioms of their field don't actually describe the real world.
posted by hattifattener at 1:25 AM on May 17, 2013


...prefers earning $100 while another person earns $0, to earning $200 when the other person gets $300.

This actually makes economic sense for a smallish group. If someone else is getting $300, they can drive prices up that I may not be able to afford with my $200, whereas if I'm the one getting paid the most I'm guaranteed that I can buy anything for sale.
posted by DU at 1:54 AM on May 17, 2013 [1 favorite]


Metafilter: prefers earning $100 while another person earns $0, to earning $200 when the other person gets $300.

Nah you're thinking of conservatives trying to enjoy breakfast
posted by yoHighness at 2:18 AM on May 17, 2013


Does anyone else feel like they just read a textbook written in Family Circus strips?
posted by whatzit at 3:21 AM on May 17, 2013


I just finished A Beginner's Guide to Irrational Behavior on Coursera, taught by Dan Ariely at Duke. I guess he kind of popularized Behavioral Economics with books like Predictably Irrational. The video lectures were pretty fun and engaging, and there were (comfortably boring) research papers to dive into as well.

I believe anyone interested can still sign up for free and watch/read (although not having deadlines for quizzes may make it hard for you to feel engaged).
posted by tychotesla at 4:48 AM on May 17, 2013 [2 favorites]


I do think it's interesting that so many of these things are economists basically talking about how the fundamental axioms of their field don't actually describe the real world.

Which is why I have absolutely zero beef with calling it "behavioral economics". If their taking undue credit for these "discoveries" is a necessary precondition to them waking up to how actual people behave, then let it happen. It's worth it.
posted by Jpfed at 5:33 AM on May 17, 2013


Very interesting, I'll read all this stuff when I get back from behaving irrationally in the woods in the immediate future!
posted by jeffburdges at 5:37 AM on May 17, 2013


The best part of behavioural economics is that it keeps winning fake nobel prizes.
posted by srboisvert at 6:15 AM on May 17, 2013


I think behavioural economics is a foundational field of study for anyone interested in creating change -- everyone from managers to politicians to writers -- and so I want to love this unequivocally, but for the love of god people HIRE A COPY EDITOR.

"Past loses"? In what amounts to display text? Honestly?
posted by gompa at 6:20 AM on May 17, 2013


You guys almost have my "Econ post on Metafilter" Bingo Card filled out. I'm hopeful we can get this wrapped up by Noon east coast time so I can go to lunch. Its Baby day in the investment bank caff.
posted by JPD at 6:45 AM on May 17, 2013 [5 favorites]


Something something bank malfeasance?
posted by Potomac Avenue at 7:07 AM on May 17, 2013


Hilariously, the "dishonesty of honest people" link says:

"In another experiment, Ariely tested how “moral flexibility” was related to the level of creativity required in different jobs by visiting an ad agency and studying the capacity for dishonesty in representatives of its various departments:

[T]he level of moral flexibility was highly related to the level of creativity required in their department and by their job. Designers and copy-writers were at the top of the moral flexibility scale, and the accountants ranked at the bottom. It seems that when ‘creativity’ is in our job description, we are more likely to say ‘Go for it’ when it comes to dishonest behavior."


Accountants have moral scruples? Tell it to the big-4 UK accountants and their massive tax avoidance schemes, or their PFI scams, or the revolving-door between them and the government/the treasury, add in civil service consultants and secondments and you will see acountants are as morally backrupt as the banks and the government. What a pile of tosh.
posted by marienbad at 7:44 AM on May 17, 2013 [1 favorite]


Accountants have moral scruples? [...] What a pile of tosh.

If there's one thing to be learned from Behavioral Economics or Psychology it's that the context you're placed in will have great impact on how you act.

There are also many other factors at work, and the quote you quoted talks about just one in particular: the effect thinking of yourself as creative has on your ability to rationalize the kind of cheating where you cheat just a little.
posted by tychotesla at 8:43 AM on May 17, 2013


Because "behavioral economists" are acting like they've discovered something when really all they are is ignorant of what's already been well studied in other fields.

But that's a question of "who came up with this first?" not "can these concepts belong to more than one field at once?" What struck me as bizarre in the wording of the FPP was the idea that a concept can only belong to either behavioral economics or to psychology. It's like saying "the study of symbolism isn't part of literary criticism, it's part of anthropology!"
posted by yoink at 9:58 AM on May 17, 2013 [3 favorites]


yoink: "What struck me as bizarre in the wording of the FPP was the idea that a concept can only belong to either behavioral economics or to psychology. It's like saying "the study of symbolism isn't part of literary criticism, it's part of anthropology!""

Obviously you're right, it's both. But I think it helps to understand why this is such a sore point.

Like most psychologists in the field I'm actually really happy to see economists taking this stuff seriously, and I like working with them when I get the chance to do so. It's much better than the status quo ante where most economists really didn't seem to want a bar of all this psychology rubbish. In that sense, the rise of "behavioural economics" as a recognised label is a great thing, because it provides a "respectable" name for those economists who want to take individual human behaviour seriously.

But the academic politics behind the names really does matter to those of us who do this work within psychology departments. We've been doing decision making research for a long time, yet public discussions of the topic do tend to have a bias towards the new sexy "behavioural economics" name. And the new name does make it sound like this research is an example of "economics self-correcting" (e.g.) rather than "economists and psychologists starting to work together" or even "economists finally listening to what the psychologists have been banging on about". And the framing of public discussions matter: it's why I spend time trying to correct horrible misrepresentations of IQ tests on MeFi, for instance. Because people who make funding decisions about psychological research are influenced by this sort of thing: they read blogs and newspapers all the time, and the more senior they are the less likely they are to also be reading the relevant empirical literature as well.

So yes, there is a political aspect to the choice of words, and I was really cheered to see blahblahblah acknowledging it in the FPP. Right now, some of my work gets called "behavioural economics" (if it has a decision making component), while other parts get called "cognitive science" (if it has maths) or even "machine learning" (if it has too much maths). Similar things occur to other colleagues. This has an upside: it gives me a lot of flexibility when recruiting students. But it's important to not forget that all of this stuff is still bread and butter psychology. Given that psychology is getting a lot of flack at the moment because of the very public airing of our fears about replicability and occasional revelations of outright fraud in the discipline, we're pretty keen to make sure people remember the good stuff too, and not relegate the word "psychology" to the rump discipline that consists solely of frauds, charlatans, and those honest researchers who don't have the option of just adopting a new name for themselves.
posted by mixing at 5:40 PM on May 17, 2013 [1 favorite]


the rump discipline that consists solely of frauds, charlatans, and those honest researchers who don't have the option of just adopting a new name for themselves.

This actually sounds like a pretty good description of economics.
posted by anewc2 at 4:29 AM on May 18, 2013


You're part of the problem now.
posted by LogicalDash at 2:56 PM on May 21, 2013


« Older Electronic band Pigeon perform a medley of DAFT PU...  |  Faith the Dog has only her hin... Newer »


This thread has been archived and is closed to new comments