An examination of a reciprocal relationship between religiosity and different forms of delinquency within a theoretical model.

This study of 1,093 adolescents from six public high schools does not support the argument that property crimes, crimes against persons, and use of alcohol and other drugs are behavior manifestations of an interrelated constellation or syndrome of delinquency. A factor analysis clearly shows that the various forms of delinquency studied load on three distinct factors. In addition, whereas the hypothesized theoretical model does explain considerable variation in frequency of alcohol use and of criminal behavior (22 percent and 24 percent, respectively), it does not account for much variance in drug use (6 percent). Whereas there are reciprocal relationships between religiosity and drug use and religiosity and crime, only the feedback effect of religiosity on alcohol use is significant. These latter findings suggest that future studies need to examine reciprocal relationships and that the relationship between alcohol use and religiosity needs to be reexamined conceptually and empirically in future studies.

The literature on the relationship between religiosity and delinquency is a lengthy history of inconsistent empirical findings, contradictory conclusions, and impassioned polemics (e.g., Ellis 19 10; Lombroso 1911; Bonger 1916; Steiner 1924; Barnes and Teeters 1951; Merton 1957; Benson 1960; Schur 1969; Sutherland and Cressey 1978; Burkett 1993). The genre of debate has ebbed and flowed according to the social milieu and theoretical perspectives of different time periods. For example, whereas Durkheim ([1897] 1951) posits that religion is a crucial integrative mechanism for maintaining social order and fostering common beliefs and values, Merton (1957) asks what evidence there is to indicate that persons committed to religion uphold these beliefs and values more than do those who have little or no identification with religion.

Conversely, Stark (1984, p. 273) announced his conversion to the position that "religion has truly potent effects [on delinquency]," after co-authoring (Hirschi and Stark 1969) one of the most influential articles stating that church attendance and belief in supernatural sanctions are not important to delinquency. His conversion appears to be in reaction to the volume of recent research supporting, however modestly, the relevance of religiosity to various forms of adolescent delinquent behavior (e.g., Albrecht, Chadwick, and Alcorn 1977; Burkett 1977; Jessor and Jessor 1977; DeLamater and MacCorquodale 1979; Stark, Doyle, and Kent 1980; Stark, Kent, and Doyle 1982; Elifson, Peterson, and Hadaway 1983; Tittle and Welch 1983; Hadaway, Elifson, and Peterson 1984). Although there is a growing literature of studies with increasing methodological rigor and consistent findings of a low-to-moderate inverse relationship between personal religiosity and delinquency, there remains (see Tittle and Welch 1983) a paucity of research that examines religiosity as an element of a theoretical model, and especially across different forms of adolescent delinquent behavior (see reviews, Burkett 1993; Cochran, Wood, and Arneklev 1994).

Indeed, with very few isolated exceptions (see Burkett 1993; Cochran et al. 1994), the relationship between religion and delinquency has been examined by controlling for selected demographic variables. And yet, religion is a central concept in at least three preeminent theories used to explain crime and delinquency (see reviews, Jensen and Rojek, 1992; Trojanowitz and Morash, 1992; Burkett, 1993; Shoemaker 1996; Benda and Corwyn, forthcoming-a). From the parallel perspectives of functionalism (Parsons 1951) and social control theory (Durkheim [1897] 1951; Weber [1905] 1958) comes the postulate that social order rests on collectively held beliefs in the moral validity of societal norms and values, which are translated initially into individual internalized moral commitments through social bonding to parents (Nye 1958; Reckless 1961; Hirschi 1969) and through involvement in religion (Stark and Glock 1968; Brownfield & Sorenson 1991).

Based on a social learning theory (Akers 1997), Paternoster's (1988) panel study shows that peers often assume greater influence than parents over adolescents, and delinquent peer associations offer norms that undermine conventional beliefs taught in the family and in religion (also see, Patterson and Dishion 1985; Burkett 1993). The purpose of the present study is to test a hypothesized theoretical model containing relationships that include religiosity to see how well it explains different forms of delinquency.

HYPOTHESIZED THEORETICAL MODEL

The hypothesized theoretical model, shown in Figure 1, tested in this study consists of elements from control theory (Hirschi 1969) and a social learning reformulation (Akers 1985, 1997) of differential association (Sutherland and Cressey 1978) theory (see Matsueda 1988; Benda and DiBlasio 1994; Benda, DiBlasio, and Kashner 1994). The model is constructed using theoretical elaboration (Thornberry 1989) instead of attempting to fully integrate (see Akers 1997) two theories that rest on contradictory assumptions about motivation for delinquent behavior (see Agnew 1995). For example, control theory rests on the Hobbesian assumption that delinquency is one expression of natural (or innate) impulses, whereas social learning theory rests on the foundational assumption that all behavior, including delinquency, is learned (and is not innate) through social interaction processes (Akers 1997). A true integration of these theories requires resolution of differences in assumptions (see Hirschi 1979), while theoretical elaboration requires that the propositions of the hypothesized model be consistent (Thornberry 1989).

[Figure 1 ILLUSTRATION OMITTED]

In the present study, a preeminent control theory (Hirschi 1969) is extended, or elaborated upon, through the addition of useful concepts from social learning theory hypothesized to provide a more accurate explanation of adolescent delinquent behavior, Useful is defined by theoretical scope, parsimony, testability, and empirical validity of concepts (see Shoemaker 1996; Akers 1997). The hypothesized model is formulated on the fundamental assumption that delinquency occurs among persons who have weak inner and social restraints on natural impulses, resulting from inadequate bonding, initially to parents and later to society (Hirschi 1969; Nettler 1984). Unlike classical control theory (Kornhauser 1978), however, the assumption underlying the model shown in Figure 1 is that weak restraint, while necessary, is not sufficient motivation for delinquent behavior (see Elliott, Huizinga, and Ageton 1985; DiBlasio and Benda 1990, 1992, 1994; Agnew 1995). Weak inner and social controls simply allow for the expression of a variety of impulses, including delinquency, such as sexual behavior (Benda and DiBlasio 1991). Moreover, accumulating evidence indicates that influences in addition to lack of restraint motivate delinquent behavior (e.g., Elliott et al. 1985; Patterson and Dishion 1985; Agnew 1993; Simons et al. 1994). The most prominent of these influences are social learning processes that occur in peer association (Elliott et al, 1985; Agnew 1995; Akers 1997). The argument undergirding the hypothesized model, however, is that there is an indeterminate, rather than an etiologically specific, relationship between motivation and any particular form of behavior (see McCord 1990). Natural impulses for gratification are channeled into delinquent behavior through social learning processes (Akers 1997) occurring in peer associations that encourage a variety of forms of delinquency (see Elliott et al. 1985; Gottfredson and Hirschi 1990). Hence, a single theoretical model is expected to explain variation in adolescent crime, alcohol, and other drug use in this study.

Other modifications to classical control theory reflected in Figure 1 include specifying certain interrelationships between the four critical elements of bonding identified by Hirschi (1969) and the addition of religiosity, peer associations, excuses, age, and residence to the theoretical model. According to Hirschi, attachment, commitment, involvement, and beliefs are the essential elements of bonding, and each element has equal and independent effects on delinquency (see Shoemaker 1996). However, based on empirical, as well as logical, grounds there appear to be interrelationships between the elements of bonding identified by Hirschi, and any proposed theoretical model should account for the strongest predictors of delinquency (e.g., Marcos, Bahr, and Johnson 1986; Thornberry et al. 1991; Benda and Whiteside 1995). A variety of models, with varying patterns of interrelationships, have been tested in the literature (see reviews, Elliott et al. 1985; Gibbons and Krohn 1991; Akers 1997). The hypothesized model examined in the present study represents a distillation of previous models based on my impulses for gratification are channeled into delinquent behavior through social learning processes (Akers 1997) occurring in peer associations that encourage a variety of forms of delinquency (see Elliott et al. 1985; Gottfredson and Hirschi 1990). Hence, a single theoretical model is expected to explain variation in adolescent crime, alcohol, and other drug use in this study.

Other modifications to classical control theory reflected in Figure 1 include specifying certain interrelationships between the four critical elements of bonding identified by Hirschi (1969) and the addition of religiosity, peer associations, excuses, age, and residence to the theoretical model. According to Hirschi, attachment, commitment, involvement, and beliefs are the essential elements of bonding, and each element has equal and independent effects on delinquency (see …