Abstract/Summary

Researchers studying curiosity and interest note a lack of consensus in whether and how these important motivations for learning are distinct. Empirical attempts to distinguish them are impeded by this lack of conceptual clarity. Following a recent proposal that curiosity and interest are folk concepts, we sought to determine a non-expert consensus view on their distinction using machine learning methods. In Study 1, we demonstrate that there is a consensus in how they are distinguished, by training a Naïve Bayes classification algorithm to distinguish between free-text definitions of curiosity and interest (n = 396 definitions) and using cross-validation to test the classifier on two sets of data (main n = 196; additional n = 218). In Study 2, we demonstrate that the non-expert consensus is shared by experts and can plausibly underscore future empirical work, as the classifier accurately distinguished definitions provided by experts who study curiosity and interest (n = 92). Our results suggest a shared consensus on the distinction between curiosity and interest, providing a basis for much-needed conceptual clarity facilitating future empirical work. This consensus distinguishes curiosity as more active information seeking directed towards specific and previously unknown information. In contrast, interest is more pleasurable, in-depth, less momentary information seeking towards information in domains where people already have knowledge. However, we note that there are similarities between the concepts, as they are both motivating, involve feelings of wanting, and relate to knowledge-acquisition.