Small procedural differences matter: Conceptual and direct replication attempts of the communication-intervention effect on infants’ false-belief ascriptions

Abstract

The current replication crisis in experimental psychology has called into question findings regarding infants’ early false-belief understanding. It is, however, debated whether non-replications might be due to procedural differences between replication attempts and original studies. The current set of studies aimed to shed light on this question by trying to replicate a violation-of-expectation study by Song, Onishi, Baillargeon and Fisher (2008). This task seemed especially important since it addresses not only the question whether or not infants hold false-belief assumptions but also whether they update these assumptions given informative (but not uninformative) verbal input. Studies 1a and 1b failed to replicate the original findings conceptually. Study 2 – which included 10-second-delays as did the original study – also failed to fully replicate the original pattern of results. More specifically, informative as well as uninformative statements elicited the communication-intervention effect. The fact that this effect appeared only when using the exact same procedure as the original study suggests that subtle features of a test procedure might influence infants’ performance in an implicit false belief task. However, the failure to fully replicate the original pattern of results emphasizes that even direct replications do not necessarily result in a successful replication.