The research reported within examines two hypotheses derived from reference group theory regarding the inverse effects of religiosity on the prevalence of premarital, extramarital, and homosexual relations across faith groups. The first hypothesis asserts that the strength of these inverse religiosity effects varies systematically across faith groups, such that as religious proscriptiveness increases, the effect of religiosity increases. The second hypothesis predicts that there will also be systematic variation in these effects across these forms of nonmarital sexual relations such that faith group variation in the effects of personal religiosity will be most evident with regard to premarital sexual relations but relatively invariant with regard to extramarital and homosexual relations. Using logistic regression analyses with data available in the National Opinion Research Center (NORC) 1988-1996 General Social Surveys, we find a mixed pattern of results which provides limited support for these hypotheses. Cochran and Beeghley (1991, pp. 45-6) argue that the five decades of social scientific research on the relationship(s) between religion and sexuality can be organized into "three steadily progressive periods" to which they recommend adding a new, fourth period of research. These are: (1) an early period in which the empirical research was largely descriptive in nature, atheoretical, and employed a rather low level of analytic sophistication; (2) a middle period characterized by improved, yet nascent, theoretical efforts and slightly more advanced statistical analyses (i.e., bivariate correlations); (3) the more recent theory-driven period in which the research designs have emphasized the use of least-squares regression techniques but restricted themselves to simple linear models; and (4) the current period in which the model specification errors of the previous period are being corrected through the utilization of advanced statistical modeling techniques that permit testing of theoretically appropriate, nonlinear functional forms. They conclude that the collective body of research throughout these four periods, which they claim to be based on more than 80 studies, has consistently revealed inverse relationships between religion/religiosity and nonmarital sexuality (attitudinal and/or behavioral).Despite longstanding Judeo-Christian doctrines of asceticism and traditions restricting sexual activity to married heterosexual couples (DeLamater 1981), Cochran and Beeghley caution against any conclusions that these relationships are simply "the empirical validation of common knowledge" or the "discovery