In improvisational theatre (improv) dialog, characters, and story are created spontaneously by the actors on stage. Impro trainers posit that compelling stories in improv are the result of social creativity (as opposed to individual creativity), a process of social interaction in which every actor contributes small impulses to commonly create the story without individual intent. Hence, improv techniques which aim to foster this social creative process have gained increasing popularity and spread into fields beyond the performing arts (e.g., business trainings), claiming to improve, among other things, participants’ social interaction skills. However, these claims have barely been investigated empirically. Here, we compare improv-players with matched controls in a controlled adaptation of an improv paradigm to study how they handle expectation violations in a social interaction situation. This paradigm allows to insert experimentally intended expectation violations into a predetermined dialogue measure their response times to these violations. In Experiment 1, we investigated the suitability of the paradigm and in the pre-registered Experiment 2, we compared the performance of improv-players and controls. The results showed that controls seem to have more difficulties in dealing with unexpected than expected utterances than improv-players. Furthermore, we find that improv-players receive higher scores on creativity and tolerance of uncertainty. We conclude that our results present first humble evidence that the successful handling of expectation violations in social interaction situations is indeed pronounced in improv players and might hence contribute to the social creativity observed on the improv stage.