Abstract
A procedure originally introduced by Sternberg and coworkers was used to examine the effect of selective attention on temporal-order judgment. Two-response and three-response paradigms were employed. As in Sternberg et al's work, a shift of psychometric functions for the two-response paradigm was found that suggested shorter latency for the stimulus to which attention was directed. However, no shift was noted for the three-response paradigm. This discrepancy can be explained by the assumption that the shift obtained for the two-response paradigm results from an artifact caused by a specific task. The artifact occurs because the instruction to attend to a particular stimulus induces a response bias when the subject is forced to guess in situations where he/she cannot recognize the temporal order of stimuli. The artifact disappears with the three-response paradigm where the subject has the additional option to respond “simultaneous”.