Studies consistently find that individuals who engage in delinquency also have peers who engage in delinquency, and this relationship has typically been explained as the result of normative social influence, or conformity to norms operating within an individualʼs peer group. However, there is little evidence that adolescents strongly endorse delinquency and recent research indicates that the effect of peer behavior may actually reflect measurement error. I address these seemingly paradoxical findings by discussing the importance of actorsʼ perceptions of normative properties and research documenting systematic errors that occur when estimating the properties of norms. I develop a model of social influence based on the misperception of behavioral regularities, a particular normative property. I test the model using 1,046 respondents from two waves of the Netherlands Institute for the Study of Crime and Law Enforcement (NSCR) School Project. The data include self-reports by school friends of their own delinquent behavior, as well as respondentsʼ estimates of peerʼs behavior, making them uniquely equipped to calculate how much respondents misperceive the behavior of their school friends. Findings indicate that misperception of peer behavior has a robust effect on the delinquency of an individual, net of individual and situational characteristics, which has not been demonstrated in prior literature. The implications of the findings for current and future research on social influence are discussed.