What noise matters? Experimental evidence for stochastic deviations in social norms

Social norms interventions are increasingly used in applied policy-making such as to end female genital mutilation or open defecation. Amongst the reasons for why policy-makers rely on social norms intervention tools is that the relevant underlying theory, which is concerned with dynamic-stochastic deviations processes, has matured substantially over the past three decades. In this article, turning to behavioral evidence, we investigate which of the models proposed in this strand of theory are supported regarding individual human behavior. Specifically, we identify which assumptions regarding individual-level adjustment behavior in terms of optimal play ('best response') and deviations thereof ('noise') are supported. In doing so, we move the focus of research from the theoretical and modeling question 'How Noise Matters?' (Blume, 2003) to the behavioral and empirical question 'What Noise Matters?'.