Nearly all the self‐talk cues studied so far have been self‐statements. However, the findings of Senay, Albarracin, and Noguchi suggest that interrogative self‐talk produces better task performance than declarative one. Two of the experiments reported here were meant to replicate that study, but the expected differences were not confirmed. Experiment 3 showed that if a self‐posed question about future behavior was answered positively, task performance was better than in groups exposed either to the self‐statement ‘I will do it’ or to a negative answer following the question. However, these differences occurred only in those who self‐reported the awareness of the impact of self‐talk on their thought processes. This effect and the possible reasons why between‐group differences were not found in Experiments 1 and 2 are discussed. An alternative explanation for the results of Experiment 3 is also proposed beside that stressing the impact of internal answer. Copyright © 2014 John Wiley & Sons, Ltd.