“…Second, Bayesian approaches are compatible with large uncertainties, which in practice are inevitable when studying interactions. Interactions, in turn, are important because statistically significant but unreplicable results can be seen as arising from varying treatment effects and situationdependent phenomena (consider, just for one example, the wide variation in estimates of the effects of stereotype threat under different experimental conditions; Fryer, Levitt, & List, 2008;Ganley et al, 2013). Finally, hierarchical Bayesian analysis can handle structured data and multiple comparisons, allowing researchers to escape from the paradigm of the single data comparison or single p value being conclusive.…”