The Simon effect refers to the finding that reaction times are faster when stimulus and response locations correspond than when they do not in tasks where stimulus location is defined as irrelevant. The authors examined the Simon effect for situations in which location-irrelevant trials were intermixed with trials for which stimulus location was relevant. Compatible mapping of the location-relevant trials enhanced the Simon effect relative to an unmixed condition, whereas incompatible mapping reversed the Simon effect. The reversal with incompatible mapping remained evident when task uncertainty was removed by use of a precue and was larger than the reversed effect produced by making incongruent trials more frequent than congruent trials. This result suggests that both attentional biases and task-defined associations contribute to the reversal of the Simon effect.