In this article we analyze behavioral data to advance knowledge on how to assess similarities of events and spatial relations characterized by qualitative spatial calculi. We have collected a large amount of behavioral data evaluating topological relations specified in the Region Connection Calculus and Intersection Models. Several suggestions have been made in the literature on how to use associated conceptual neighborhood graphs to assess the similarities between events and static spatial relations specified within these frameworks. However, to the best of our knowledge, there are few (to none) approaches that use behavioral data to formally assess similarities. This article is contributing to this endeavor of using behavioral data as a basis for similarities (and associated weights) by (a) discussing a number of approaches that allow for transforming behavioral data into numeric values; (b) applying these approaches to nine data sets we collected in the last couple of years on conceptualizing spatio-temporal information using RCC/IM as a baseline; and (c) discussing potential weighting schemes but also revealing essential avenues for future research.