The stimulus-response compatibility (SRC) effect is obtained when performance is better on compatible mappings than on incompatible mappings, when comparing pure tasks. When the mappings are mixed in a task, the SRC effect is often eliminated and/or reversed. Prior studies have not examined how practice can influence the SRC effect in mixed mapping tasks. It was predicted that practice would make task components more automatic, decreasing the need to suppress the direct route. However, with all practice conditions in this study, the SRC effect was eliminated. No performance advantage was found for the groups that practiced with the component tasks compared to the whole-task. For the group that practiced with the whole-task, or mixed mapping, there was a reduction in overall reaction time. Results support the suppression and uncertainty accounts, and implications of these findings for training paradigms and future research on tasks involving mixed mappings are discussed.