Time judgments in global temporal contexts

With three experiments, we examined the effects of global temporal context on time judgments as gauged by constant errors (CEs) and estimates of a preferred period (P). In Experiment 1, participants in seven different conditions listened to sequences of a given rate (with interonset intervals ranging from 200 to 800 msec) and judged the relative duration of a final (comparison) time interval. No P emerged. In Experiments 2 and 3, we embedded the same rates in different global (session) contexts that varied according to (1) mean session rate, (2) standard deviation, (3) range, and (4) number of different rates in a session. Evidence from CEs indicated that P varied primarily as a function of mean session rate and range of tempi. The best predictor of errors involved a measure termed relative range (RR = range/mean session rate). A general algorithm incorporating RR successfully predicts P, and the implications of this algorithm are discussed.