Tversky and Kahneman

[fnord_too]

I have long meant to read their work, and just started in on a few articles today (I still don't have the "Prospect Theory" article, but some others.) I would highly recomend reading their stuff. It is on decision making and perceived risk, and is just great.

[David Sklansky]

Yes. And of course the the idea is to not be like the people they describe.

[MuckerFish]

any links? Sounds interesting. Thanks

[fnord_too]

Everything I have I have gotten through libraries. (Most libraries pay for magazine content services. My sister sent me the 1979 initial article last night.)

You can probably get a good idea of the nature of their work by googling. In adition to each of their names, try:
Behavioural Economics, Behavioral Finance, Econometrics, and Prospect Theory.

The gist of it is that people make dubious decisions based on perception and the way they think about the world. For instance, if a couple has tickets to a basketball game and there is a blizzard raging, whether they bought the tickets or got them for free tends to have a large impact on the decision to drive to the game, even though at the time that decision is made how they got the tickets is irrelevant.

Another trap people fall into is judging by experience whether praise or scorn is better at getting results. People will think scorn is, because if they berate someone who had an abnormally bad performance, their next performance tends to be better, but if they praise someone who had an abnormally good performance, their next performance tends to be worse. These tendancies hold empirically, but there is a serious flaw in the reasoning.

[PokerProdigy]

[ QUOTE ]
Another trap people fall into is judging by experience whether praise or scorn is better at getting results. People will think scorn is, because if they berate someone who had an abnormally bad performance, their next performance tends to be better, but if they praise someone who had an abnormally good performance, their next performance tends to be worse. These tendancies hold empirically, but there is a serious flaw in the reasoning.

[/ QUOTE ]

Yeah, that definitely is flawed reasoning. In fact, there is a name for that flawed reasoning in statistics and it is called statistical regression (aka regression to the mean). I learned about it in my psychology research methods class. For example, if a SAT prep course claims that their course increased students previous scores by X%, it may have had nothing to do with the course, but rather everyone that took the course already had low scores to begin with and therefore there scores would have probably increased even if they did NOT take the SAT prep course.

[DCWildcat]

[ QUOTE ]
Another trap people fall into is judging by experience whether praise or scorn is better at getting results. People will think scorn is, because if they berate someone who had an abnormally bad performance, their next performance tends to be better, but if they praise someone who had an abnormally good performance, their next performance tends to be worse. These tendancies hold empirically, but there is a serious flaw in the reasoning.

[/ QUOTE ]

This is, is my mind, one of humankind's heretofore greatest failures.

In the long term, positive reinforcement is always a better reinforcer for any behavior. Unfortunately, punishment yields better short-term results.

Taken one step further, punishment is almost always undesirable; in fact, most psychologists teach potential parents only to punish their children if the children have a behavior that may be life-threatening (where short-term results are obviously important...you don't want to have to correct a child 5 times before he understands to not stick the fork in the socket).

[Grisgra]

It's been over ten years since I studied them, but I loved these guys. Truly groundbreaking work. Another of their important findings was that humans are incredibly shitty at dealing with low-risk events, especially those that, if they occur, have large (i.e., dangerous or deadly) effects.

[Carnite]

The idea here is however one of mathematics, it's called statistical regression to the mean. If you observe the short term changes of someone at the extreme end of the performance spectrum, either end, they're most likely to be seen going in the direction of the mean, i.e. good performers getting worse, bad performers getting better. The idea here is that it takes some fairly extreme circumstances to get them to the peak and thus are likely to drop off regardless of what you do.

[PokerProdigy]

[ QUOTE ]
It's been over ten years since I studied them, but I loved these guys. Truly groundbreaking work. Another of their important findings was that humans are incredibly shitty at dealing with low-risk events, especially those that, if they occur, have large (i.e., dangerous or deadly) effects.

[/ QUOTE ]

Could you give an example of this low-risk situations because it sounds fascinating. Also, what should I read that talks about this topic.

[PokerProdigy]

[ QUOTE ]
[ QUOTE ]
Another trap people fall into is judging by experience whether praise or scorn is better at getting results. People will think scorn is, because if they berate someone who had an abnormally bad performance, their next performance tends to be better, but if they praise someone who had an abnormally good performance, their next performance tends to be worse. These tendancies hold empirically, but there is a serious flaw in the reasoning.

[/ QUOTE ]

This is, is my mind, one of humankind's heretofore greatest failures.

In the long term, positive reinforcement is always a better reinforcer for any behavior. Unfortunately, punishment yields better short-term results.

Taken one step further, punishment is almost always undesirable; in fact, most psychologists teach potential parents only to punish their children if the children have a behavior that may be life-threatening (where short-term results are obviously important...you don't want to have to correct a child 5 times before he understands to not stick the fork in the socket).

[/ QUOTE ]

Good points, and I have often heard similar discussions about reward vs. punishment in the psychological community.