Religious Deviance

Introduciton

This study tested multi-level models that represented relationships among individual-level and community-level measures of religiosity and measures of deviant behavior that individuals reported they would be likely to commit. Data came from 2,667 adult Catholics surveyed as part of the Notre Dame Study of Catholic Parish Life. Results from multivariate analyses supported the argument that the level of religiosity within a given social context affects projected deviance, although there was no evidence of any substantial interaction between personal- and community- level religiosity. Results also suggested that contextual effects are not necessarily consistent across all types of deviance.

Theoretical Background

A decade ago, social scientists were unsure about the general deviance-inhibiting influences of religion (see Tittle and Welch 1983:653-656, for an early assessment; and Bainbridge 1989 for a contrary view). However, recent research has convinced most that religiosity is a potent generator of conformity (Cochran 1987; Cochran and Akers 1989; and Sloane and Potvin 1986). Nevertheless, there have been disagreements about the complexity of these inhibiting effects. Some (Sloane and Potvin 1986) have contended that religious influences are more or less uniform across conditions, while others have contended that religious effects are dependent upon a number of contextual contingencies (Tittle and Welch 1983). To resolve this controversy, research has begun to focus on multilevel effects (Cochran and Akers 1989; Elifson et al. 1983; Stark et al. 1980, 1982; Tittle and Welch 1983; Watters et al. 1985). This shift away from simple individual-level analysis corresponds with trends in sociology generally (van den Eeden and Huttner 1982; Markovsky 1987), and is especially evident in the study of crime and deviance where contextual analyses are becoming more frequent and prominent as movement toward integrated explanations proceeds (O'Leary 1989; Liska et al. 1989).

Interest in specifying the conditions under which religion inhibits deviant behavior has focused mainly on the "moral community" hypothesis. Two early papers by Stark and his associates (1980,1982) concluded that communities displaying high levels of religious integration ("moral communities") are most effective in promoting conformity to the moral order. However, Cochran and Akers (1989) recently tested the hypothesis and, despite data that seemed to be somewhat supportive, concluded that contextual variations are relatively insubstantial.

The main study reporting evidence directly contrary to the moral community hypothesis (Tittle and Welch 1983) suffered from a related, but somewhat different, problem. It did not use a geographic-based measure of the morality of communities, but instead relied upon "approximated" contexts.

Stark et al. acknowledged some of these problems when they introduced a second study that attempted to represent "the religious character of a boy's immediate surroundings" (1982:15). In this second study, they used data from a national sample of high schools with separate subsamples of approximately 25 boys in each of the 87 schools. By aggregating individual responses to selected religious items, they were able to construct a measure of the moral climate that exists within each of the high schools, differentiating them into two categories: those composed of students who were relatively favorable toward religion ("moral communities") and ...