Research in social embodied cognition has found numerous positive associations between higher temperatures and prosocial behavior, with "warmer is better" effects demonstrating greater altruism, social proximity, and affective attributions. However, recent failures to replicate experimental findings in this area have called some effects into question. By contrast, epidemiological research has frequently observed that higher ambient temperatures are correlated with increases in societal volatility and antisocial behaviors, with "warmer is worse" effects demonstrated by increases in warfare, city-center rioting, and sporting aggression. Using a sample size (N = 611) larger than many previous studies in the social priming domain, we examined whether higher ambient temperature and/or higher prop temperature (i.e., holding a hot/cold therapeutic pack) lead to more prosocial behavior. Results showed that warmer ambient temperatures were weakly associated with more prosocial than selfish choices. However, calculation of Bayes Factors suggests that the evidence is better characterized as inconclusive. As previously published analyses of the same data set used in this study show (Lynott et al., 2014), prop temperature did not lead to increased prosocial responding, which fails to support predictions from particular accounts of social embodiment. We discuss possible reasons for the inconclusive findings and for the broad range of effect sizes observed more generally in the literature.