This project addresses two, related assumptions of the growing literature on racial/ethnic diversity and prosocial behavior. The first is that the effects of diversity are symmetric across group compositions (i.e., across majority-majority and majority-minority contexts). The second is that an aggregate association between diversity and prosocial behavior implies that people are individually less likely to engage in prosocial behavior as a function of diversity. Instead, such an association could result from the aggregation of people responding to outgroup share. Both assumptions stem from an elision between the concepts of diversity, i.e., heterogeneity, and minority/outgroup share, which are (a) analytically distinct, and (b) occasionally orthogonal to each other. We illustrate these assumptions and their implications using data on non-emergency calls in NYC.