According to ideomotor theory, when people perform a movement and observe its subsequent effect, they acquire a bidirectional action-effect association. If at a later point they want to produce the effect, its anticipation activates and allows executing the corresponding action. In ideomotor induction tasks, several task characteristics determine whether participants use the experimentally induced action-effect associations to pre-activate the corresponding actions. Here, we assess the impact of the verbal instruction, the task relevance of the effect stimuli and the presentation of post-response effects on the expression of action-effect associations. The results show that an instruction stressing the stimulus-effect correspondence prompts participants to utilize the presented effects more than an instruction stressing the stimulus-response correspondence. Furthermore, the induced action-effect associations were only expressed when the effects were relevant for the task and when post-response effects were presented in the test phase. These findings show the importance of the particular task construction for the expression of the experimentally manipulated action-effect knowledge.