A Stochastic Description of the Learning Behaviour of an Individual Subject

Abstract

A stochastic process for the description of learning behaviour in a two-choice situation, a special case of a family of processes given in an earlier paper, is introduced. It can be used to describe individual learning data, including both the choice sequence and the latency of the choices within the description. This has the advantage that parameter estimates may be based separately on the latencies and on the choice sequence, and the agreement between these estimates can be taken as an indication of the validity of the stochastic description. The process is fitted to fifty learning trials of an albino rat. Only two parameters are employed in the description of the sequence of choices made by the subject, and a further one to include the response latencies. No generality is claimed for the success of the particular process employed in this one example, which is given only as an illustration of the possible value of the method.