ABSTRACT
Most of our data about human information processing come from laboratory experiments. It is not difficult to think of most experiments as problem-solving tasks, involving, particularly, a set of instructions and a set of possible solutions. Yet it was quite difficult for me, at one time, to think of many interesting social-psychological phenomena for which an explanation in terms of problem-solving processes was satisfying. Surely, I thought, no two people view their social world exactly alike. Clearly, then, no two subjects in a social psychology experiment are trying to "solve" the same "problem." From these premises, it followed that only problems that would be interpreted by most subjects in the same way were worth studying in the laboratory.
