How do partner reactions affect task set selection: Specific imitation or abstract tasks?

Task set selection is facilitated when people expect a partner to perform the same task, suggesting that the features of the partner's performance are represented. However, it is unclear how similar the partner's reactions must be to promote compatibility effects: does a partner have to imitate subjects' specific actions or is it enough to perform the same task while responding to different stimuli with different actions? This present study investigated this question in a joint picture-word interference paradigm. Subjects either named pictures or read words, and a partner responded by performing the same or the competing task. In Experiment 1, the partner used the same picture-word combinations as the subject and thus compatible trials implied a complete imitation. Compatibility benefits were observed. In Experiment 2, the partner performed the same or the competing task on different stimuli, producing different actions. Compatibility effects were absent. To test whether this indicates that an overlap in abstract task features is insufficient or resulted from excessive task difficulty, Experiment 3 replicated Experiment 2 with a smaller stimulus set. Compatibility benefits were found. Taken together, the results suggest that a partner's abstract task can be represented and affect task set selection processes even without an overlap in stimulus-response mappings.