Everybody knew that all of the models of group decision making says it goes through like four or five stages. You define the problem, then you understand the problem, then you generate some options, then you choose one, then you implement it- as though it went in those blocks. But I've worked with groups all my life, and I knew that they were very chaotic, and they weren't nearly that organized. What I did was to say, “Well, let's see if we can actually map the procedures that it goes through.” What I did was to identify markers, using usually coding systems of the interaction. Sometimes I would have the people write various stages of the group, but usually, it was just from the outside. What I did was to devise markers that would let me identify, for example, when they were defining the problem, when they were analyzing the problem, when they were defining criteria. And what you would have is, a group might do a problem definition, criteria solution, go back to the problem, go back to the solution, redefine the criteria, orient themselves again, and things like that. What I tried to do was to define types, and what I found was about six or seven basic types of groups. And only about a third of them went through the stages in that sequence that was taught to us in our classes, as being the way groups make decisions. Two-thirds of them took other routes.