On Producing Random Responses

Abstract

This paper has surveyed studies of response patterning in an attempt to illustrate the various ways in which this fact of behavior has been used in psychology. The view is expressed that response biasing takes on significance as an important dependent variable when we note that human Ss are unable to generate binary choices in a random sequence even when instructed to do so. Production of random sequences may be intimately related to processes of set and attention; success in generating responses randomly indicates ability to maintain an appropriate set for randomness. Techniques for measuring response patterning, or variability, were reviewed, and a new method was described. Data were presented to illustrate the difficulties encountered by students and psychiatric patients when instructed to generate a random sequence of binary choices. Almost nothing is known about the correlates of ability to maintain a set for randomness, yet there are marked individual differences in this ability. Generating random choices is basically a very non-stimulating task, so that differences in arousal may be of particular importance, when studying this particular kind of set.