Serial-choice reaction time: inadequacies of the information hypothesis.

The results of an experiment on serial-choice reaction time, specifically designed as a critical test of the Information Hypothesis, lead to rejection of the hypothesis; information is found to be neither a necessary nor a sufficient condition to account for the data. Where previously information had been interpreted as a determinant of reaction time, it was usually confounded with the probability of nonrepetition of a signal. Thus, to the extent that this confounding is present in previous experiments, the inference attributing an increase in reaction time to an increase in information is logically invalid.