A great paper, available here: www.snowpit.com.
"Even though people are capable of making decisions in a thorough and methodical way, it appears that most of the time they don’t. A growing body of research suggests that people unconsciously use simple rules of thumb, or heuristics, to navigate the routine complexities of modern life. In this paper, I examine evidence that four of these heuristics – familiarity, social proof, commitment and scarcity – have influenced the decisions of avalanche victims."
If you're interested in this kind of area, read stuff by Daniel Kahnemann and Amos Tversky. But this is a good start.
Thursday, 10 May 2007
The "Helpful Model"
I've worked in a number of investment banks, and the people there are very keen to "monetize their skills" as a hedge fund manager once put it to me.
In an environment where everyone's chasing the year-end bonus, it's easy for a self-reinforcing negative dynamic to develop. Team A views team B as out to get them. They act to defend themselves againt team B's (real or imagined) actions. Team B sees the actions of team A, and takes a corresponding action. Everyone ends up with an entrenched negative view of each other.
One alternative, referenced in Gerald Weinberg's Quality Software Management books, is called the "helpful model". The default assumption is that even if appearances might lead you to suspect otherwise, everyone is trying to be helpful. Of course, someone might reveal themselves to be out to stitch you up. But you start by assuming that everyone is trying to do the right thing, at least according to their definition of what "the right thing" is, and analyse things from that viewpoint.
A lot of people think that the "helpful model" is still unrealistic hippy nonsense. If that's your reaction, then try a quick thought experiment: figure out what your default way of analyzing people's actions and motives is, and then work out what is going to happen in your organisation if everyone thinks like you.
If you want to get all game-theoretic about it, you could view the "helpful model" as corresponding to "nice strategies" such as tit-for-tat in iterated prisoner's dilemma simulations. A "nice strategy" is one that starts off by being cooperative. What you do when it's clear that your game partner is not cooperating is a different question. "Nice" strategies co-operate well. Strategies that start by not cooperating ("defecting" in the terms of the prisoner's dilemma) don't interact so positively.
In an environment where everyone's chasing the year-end bonus, it's easy for a self-reinforcing negative dynamic to develop. Team A views team B as out to get them. They act to defend themselves againt team B's (real or imagined) actions. Team B sees the actions of team A, and takes a corresponding action. Everyone ends up with an entrenched negative view of each other.
One alternative, referenced in Gerald Weinberg's Quality Software Management books, is called the "helpful model". The default assumption is that even if appearances might lead you to suspect otherwise, everyone is trying to be helpful. Of course, someone might reveal themselves to be out to stitch you up. But you start by assuming that everyone is trying to do the right thing, at least according to their definition of what "the right thing" is, and analyse things from that viewpoint.
A lot of people think that the "helpful model" is still unrealistic hippy nonsense. If that's your reaction, then try a quick thought experiment: figure out what your default way of analyzing people's actions and motives is, and then work out what is going to happen in your organisation if everyone thinks like you.
If you want to get all game-theoretic about it, you could view the "helpful model" as corresponding to "nice strategies" such as tit-for-tat in iterated prisoner's dilemma simulations. A "nice strategy" is one that starts off by being cooperative. What you do when it's clear that your game partner is not cooperating is a different question. "Nice" strategies co-operate well. Strategies that start by not cooperating ("defecting" in the terms of the prisoner's dilemma) don't interact so positively.
Subscribe to:
Posts (Atom)