A growing number of survey-based studies have examined individual environmental behavior and support the idea that social norms are an important determinant of the behavior. We depart from the literature by estimating a structural model of the social interactions in an individual's decision to engage in energy-saving practices and account for the methodological issues that are inherent in survey data: simultaneity, common shocks, and nonrandom group selection. Using data from a Japanese household survey, we find that the influence of social norms on individuals' energy-saving practices is small or insignificant and that unobserved individual characteristics are correlated between members in a group. Although based on a specific sample and a particular identification strategy, our results illustrate that reduced-form evidence, of positive correlation among group members, which is abundant in the literature, should be interpreted with caution because it may not reflect causation.