The purpose of this study is four-fold: (1) to test delinquency theories in social settings that vary by their degree of religiousness; (2) to determine whether delinquency causal processes vary according to the nature of religious ecology; (3) to assess whether variation exists in the rates and types of adolescent offenses committed in these settings; and (4) whether these offenses are a response to unique influences in each context. Religious ecology is measured by tapping a dimension of school religious characteristics, including a school's level of religiousness and a school's religious group composition. Adolescent boys who are exposed to the confines of schools that are predominantly irreligious or disproportionately low in orthodoxy are significantly more likely to engage in delinquency than boys from more moral or highly orthodox schools. Experiences in fundamentalist reference groups also protects youngsters against engaging in substance use episodes, including harmful drugs such as cocaine. These patterns are independent of demographic characteristics such as race, gender, socioeconomic status, family size and community size. To account for religious ecological differences in problem behavior, strain and control theories of delinquency are tested. These testing procedures reveal little support for processes advocated by strain theorists. Not only is structural and interactionist induced strain not correlated with delinquency and substance use, but discrepancies between cultural expectations and perceived realization of these goals do not lead to psychosocial frustration and tension, as implied in many strain models. Control models more aptly account for delinquency and substance use variation in various religious climates, but the strength of religious, school, and family effects varies with the type of offense and the measure of religious ecology. If we measure religious ecology by the nature of denominational composition, religiosity has a uniform effect on delinquency. However, religiosity's effect in settings that vary by religious level is to more strongly inhibit chronic offending in secular disorganized communities. Involvement in delinquency and substance use is probabilistically less likely in moral and highly orthodox settings because religion's social expression is stronger, the broken home phenomenon is weaker and potentially harmful school behaviors and attitudes are unrelated to delinquency in these settings.