Time perception,vigilance and decision theory

A statistical decision hypothesis is outlined to account for a progressive lengthening of judgments which occurs when a series of reproductions of a time interval is made without interpolation of a standard. Three experiments are reported in which it is found that this serial effect can be manipulated by hypothetical payoff conditions, that the effect still occurs when subjects are told to count while making their judgments, and that two successive judgments are sufficient to demonstrate the effect if enough subjects are used. The phenomenon is related to recent studies of vigilance in which statistical decision theory has been applied. It is suggested that serial reproductions of time intervals may perhaps be a more sensitive index of the factors underlying vigilance than the traditional measures involving long sessions with weak, sporadic signals.