Dr. Scott Claessens

Cross-national analyses are used to test hypotheses about cultural variation in prosocial and antisocial behaviours. While potentially fruitful for understanding global patterns in social behaviour, such analyses often do not fully account for the various ways in which nations are connected, for example through spatial proximity or shared cultural ancestry. This non-independence of nations can violate statistical assumptions and may increase false positive rates. In this roundtable discussion, I will highlight the extent of spatial and cultural autocorrelation in existing cross-cultural databases of prosocial and antisocial behaviours. I will then demonstrate how accounting for these patterns of non-independence using global geographic and linguistic distance matrices can substantially influence the results of statistical models. Finally, I will discuss how greater attention to historical causal processes may help to refine interpretations of cross-cultural correlations in prosocial and antisocial behaviours.