Monday, May 28, 2007

How Experts Make Decisions

In Sources of Power: How People Make Decisions, Gary Klein explains how experts make reliable snap decisions over novices.

RPD Model
The recognition-primed decision (RPD) model integrates two processes:

  1. How decision makers size up the situation
  2. How they evaluate the course of action by imagining it

There's 3 basic variations of the RPD:

  1. Simple match - how have I solved this before? This is a typical case. The decision maker recognizes a situation and knows the goals, cues, expectancies, and actions. "if ... then,"
  2. Diagnosing the situation - which situation is this closest to this? This doesn't match a typical case, or it maps onto multiple cases. The decision maker needs to figure out which case is a closest match. "if(???) ... then,"
  3. Evaluating courses of action - what's the best action? This can involve adjusting an action or ruling options out. "if ... then(???)"

While the RPD model sounds obvious, it's different from earlier decision theories.

Rational Choice Strategy
In the rationale choice strategy, you define the evaluation dimensions, weight each one, rate each option on each dimension, multiply the weightings, total up the scores, and determine the best option.

RPD vs. Rational Choice

  • If you can't trust someone to make a big judgement, such as which option is best, why trust all the little judgements that go into the rational choice strategy?
  • There's usually not enough time or information to make rational choice strategy work.
  • The rational choice strategy is not going to ensure that novices make good choices.
  • The rational choice strategy is usually not helpful for experienced decision makers.
  • The rational choice strategy can be useful in working with teams, to calibrate everyone's perspectives on the options.

Key Take Aways

  • RPD quickly evaluates courses of actions by imagining how they'll be carried out, not by formal analysis and comparison.
  • Be skeptical of formal decision making methods, since they're not what most people use in real scenarios.
  • Be sensitive to when you need to compare options and when you don't. When you're new to a situation, you may need to cast a wide net. Other times, you can rely on your expertise and drill down on a smaller set of alternatives.
  • RPD is focused on being poised to act rather than stuck in analysis paralysis until all evaluations have been completed.

My Related Posts

2 comments:

Anonymous said...

One might also include the "cost of failure" into this evaluation. The rational choice strategy can also be useful when the cost of failure is so high as to make the time spent on this process worth it. Think Space Shuttle and Airplane design.

One might use RPD and then use the Rational method further down in the process in relationship to specific engineering trade offs. The closer you get to number and real measurements (as opposed to subjective evaluations and opinions), the more valid a rational method is. At least in my book. But in the absence of real data, you go back to the RPD model using what little data you have.

J.D. Meier said...

Nice distinction. I think the "cost of failure" is a key factor.