Tuesday, December 26, 2006

On the union of problem-solving and decision-making

Simon writes: "[Some of] these activities--fixing agendas, setting goals, and designing actions are usually called problem-solving. The last--evaluating and choosing, is usually called decision-making."

So the idea is that there are 2 different things, and that theories of decision-making should separate them. Should they? At first thought it seems very natural; one is concerned with what we want to do, the second is concerned with how. But human psychology gets in the way.

One example comes from the vagueness of expert advice.

[Chess novice makes a move]
Chess expert: "you don't want to make that move!".
Chess novice: "why not?"
Chess expert: "That's not the kind of move you make in this kind of position."

What kind of advice is that? Why is it so vague? Because it comes from an expert. Experts have immense difficulty perceiving what they know. It is very hard to verbalize knowledge that has been stored deep in memory.

Let's think about a situation where we are all experts. Some stranger approaches you in the street and you just feel that they do not have good intentions. It has happened to me in Paris, New York, and of course in Rio. It's safe to assume everyone has gone through this many times.

Now, get a phone, call a child, and try to explain them how to know when to reject people that approach them. Can you put it in words? It's very, very, hard to put in words how to know which approaching people you should reject and turn your backs and move on, and which you should listen to. That's why people educate children to never talk to strangers. Children can't go wrong this way. But we, adults, know when someone is trouble. Have you ever been approached by someone and just felt "I REALLY don't want to talk to this person"?

I bet the answer is yes. But can you write down how you came to that judgement? Did you "fix agendas", "set goals", and "designed actions", which were later "evaluated" until one was selected? Probably not. Probably you saw something you did not like. Then you felt weird about that person and started looking for more cues, until a moment comes when you decide--"I'm getting out of here".

This leads us to Gary Klein's recognition-primed decision model. A firefighter, a nurse, a radar operator, they all act instantly, as if they had not (i) fixed an agenda, (ii) set goals, (iii) designed actions (problem-solving part), then evaluated the possible actions and finally selected one of them (decision-making). No, that's not what Klein found out. People with expertise would find themselves immersed into a situation; and with the perception of subtle cues they would instantly have a diagnosis and a course of action. No agenda-setting, no goals defined, no set of possible actions, no evaluation of these actions. Just perception and action.

Now, what is perception?