Four experiments were performed to assess the effects of task differences on duration judgments. Experiments 1 and 2 used the method of reproduction in prospective, within-subjects designs; their results supported previous research on the effects of task difficulty. Both experiments, using tasks that varied along somewhat different dimensions, found that subjects provided reproduction values that varied inversely with task difficulty. That is, while subjects tended to underreproduce across all tasks, the more difficult the task performed during the target interval, the greater the extent of the underreproduction. Experiments 3 and 4 used a modification of the reproduction method by placing demands upon the subjects during both the target interval and the reproduction phase of each trial; they demonstrated that the greater the degree of contrast between demands made by the task performed during the target interval and those made during reproduction, the less accurate the duration reproduction. The results are discussed in terms of the contextual and resource allocation models of duration estimation.The contextual model of temporal experience (Block, 1989(Block, , 1990) proposes a variety offactors that interact to influence psychological time. They include: (1) the particular kind of temporal behavior required of the observer (e.g., duration estimation, rhythm production); (2) the contents and qualities of the time period (e.g., its duration, the number ofevents); (3) the characteristics of the experiencer (e.g., personality factors, temporal outlook); and (4) the activities performed within the time period (e.g., problem solving, types ofperceptual and/or cognitive requirements). The model has been a valuable heuristic, identifying a variety of relationships and interactions between time-relevant variables that it should be useful to explore, as well as assisting in resolving controversy that existed within the literature. This can be illustrated by an examination of research concerning both the effects of task demands on subjects' duration judgments and the way in which such effects are moderated by different instruction-based temporal outlooks (i.e., prospective versus retrospective paradigms).A number ofstudies have shown that subjects' duration judgments for a given interval of time are affected by the