A Probability Model for Conformity

Abstract

Asch (1, 2, 3) has developed an experimental procedure which provides an excellent means of testing current conceptions of the nature, genesis, and correlates of conformity. design requires an individual to make a sequence of judgments of unambiguous stimuli in the presence of confederates instructed to give incorrect responses. procedure produces conformity under highly controlled conditions; it meets the requirements of reliable measurement and at the same time preserves a flavor of life often missing in laboratory experiments. For these reasons, the Asch situation is suitable for testing, rather than merely illustrating, a wide variety of propositions. Most efficient use of his design, however, requires further analysis of the process involved in the experimental situation. present research is an effort to analyze this process by developing a mathematical model which would describe this type of experiment. Mathematical models can serve to formalize existing theory or to describe sets of data. Thus Simon's model (9) for Homans' The Human Group, (7) derives new consequences from the assumptions of the original theory. On the other hand, Blumen, Kogan, and McCarthy (4), in their work on labor mobility as a probability process, summarize large masses of data with a small number of parameters. present research applies probability theory to describing time changes in conformity behavior. While the model does describe data from a particular experiment, it also formalizes some assumptions as to the nature of the process which occurs in this situation. It differs from Simon's work, however, in that the assumptions are not part of any known theory of conformity. A number of issues arise in connection with broad application of the Asch design. Since the model organizes and tests some assumptions about the process, it offers one way to examine some of these issues. For example, Asch's results show considerable individual variation in conformity behavior and it is necessary to explain this variation. Many investigators would assume that enduring personality differences account for the observed variation in behavior. That is, individuals differ before the experiment; hence they behave differ-