Religion and Crime

Social scientists in the 1960s questioned the relevance of examining the effects of religion on crime. Their position was consistent with the view of many academics and intellectuals that religion in American society was not a major source of operative norms, values, and beliefs (Glock and Stark). Hirschi and Stark's widely cited article, "Hellfire and Delinquency," led to conclusions supporting that perspective.

Elaborating Hirschi's social control theory, Hirschi and Stark forged theoretical links between religiosity and internalized intrapersonal sources of conformity to the normative sociolegal order. They predicted that church attendance would have a direct negative effect on crime and would affect intervening variables that were associated with criminal/delinquent activities: amoral attitudes ("to get ahead, you have to do some things that are not right"; "suckers deserve to be taken advantage of"); disdain for the law ("it is all right to get around the law if you can get away with it"); rejection of a positive view of the police ("I have a lot of respect for the police"). They also predicted that church attendance would be associated with beliefs in supernatural sanctions that could deter delinquency: "there is a life beyond death"; "the devil actually exists." Hirschi and Stark assessed their hypotheses with data from a survey (Richmond Youth Study) of over four thousand students from a working-class town in the San Francisco Bay Area. Self-reports of delinquency (property crimes and violent offenses) were validated using official or police data.

The results of the empirical tests of the predicted effects can still be surprising to many people. Church attendance had no significant direct effect on delinquent activities nor did it affect the social control variables (endorsement of conventional moral principles and positive attitudes toward the law and the police). Church attendance was positively associated with beliefs about super-natural sanctions, but no significant relationships were found between these beliefs and criminal activity.

Where conventional religiosity had no effect on the measured indicators of social control, respondents who claimed a high level of respect for the law and the police were, indeed, much less likely to be delinquent than chiselers and those with contempt for the police. Likewise, those who believed in fair and just treatment of others were much less likely to be involved in delinquent activities than those who, for example, agreed that "suckers deserve to be taken advantage of."

Given its theoretically based argument, the analysis of data gathered from a large, carefully drawn sample of respondents, and the cross-validation of self-reported delinquency with police records, "Hellfire and Delinquency" seemed to be the final word about the relationship between religion, crime, and delinquency in American society. Translated into general theoretical language, Hirschi and Stark's results led to the conclusion that religion in an advanced industrial society such as the United States was differentiated or decoupled from the forces of social control that were thought to affect crime and delinquency. The strong tie between religion and social control that philosophers and sociological theorists from Feuerbach, Marx, Nietzsche, and Freud to Durkheim and Parsons had assumed to be the case was missing according to Hirschi and Stark. But was it?

Over the next decade, a number of studies examined the effect of religion on crime and delinquency. Some of Rhodes and Reiss's findings were consistent with those reported by Hirschi and Stark. There were no significant differences in delinquency rates across denominational lines in either Hirschi and Stark or Rhodes and Reiss. On the other hand, although it was not highlighted by them, the data in Rhodes and Reiss did show a negative relationship between church attendance and delinquency thus contravening Hirschi and Stark.

Burkett and White found that while church attendance had no effect on delinquent acts such as property crimes and assault, it did tend to curb status offenses (drinking and smoking). Albrecht, Chadwick, and Alcorn reported negative correlations among church attendance, religious beliefs, and delinquency, as did Higgins and Albrecht. In each case, relationships between religiosity and delinquency were stronger for victimless offenses than they were for crimes with victims.

Major studies of the link between religion and crime in the 1970s concluded with Jensen and Erickson's analysis of the Arizona Community Tolerance Study and their reanalysis of the Richmond Youth Study. The Arizona study gathered data from over three thousand high school students attending schools in metropolitan and small town contexts in southern Arizona. Church attendance had a significant impact on thirteen of eighteen delinquency items. The effects of involvement in church-related youth activities and beliefs were less pronounced. Among serious offenses, church attendance had the most impact. Among less serious offenses, both church attendance and youth activities had the most impact.

Jensen and Erickson also found that religious affiliation (Roman Catholic/Protestant/Mormon) had an effect on delinquency. Mormon youths, in particular, were more likely than Protestants or Roman Catholics to report low levels of smoking, drinking, and drunkenness. These denominational differences were more pronounced in the small towns sampled by Jensen and Erickson than in the metropolitan context.

Where church attendance and denomination were controlled, Jensen and Erickson (1979) suggested that the pattern of effects on delinquency could be explained by variation across denominations in the salience of offenses as moral deviations. Serious offenses, they argued, may be interdicted by all denominations, but minor offenses such as smoking may only be negatively sanctioned by groups that value personal asceticism. Finally, in comparing predictors of crime and delinquency, Jensen and Erickson found that measures of religion were of moderate importance, being more predictive than participation in secular activities (including sports and school clubs) but less important as predictors than various measures of attitudes toward the law, including respect for the police.

Jensen and Erickson also reanalyzed the Richmond Youth Study data. They found that church attendance was significantly related to drinking, smoking, and truancy. (Drinking and smoking had been omitted from Hirschi and Stark's delinquency index.) They also found differences between more and less "ascetic" denominations: Personal moral strictness tended to interact with church attendance to curb delinquency.

In the early 1980s, Stark, Kent, and Doyle developed a contextual explanation regarding the effects of religion on crime and delinquency. Stark and colleagues distinguished so-called moral communities from secular communities, arguing that religion in the former is a more prominent feature of daily life than in secular communities. Moral communities are characterized by relatively high rates of church attendance and church membership. Religiously based community activities for youth and adults are relatively common in moral communities compared with secular communities.

Stark and colleagues compared the effects of church attendance on delinquency in a "moral community" with a high rate of church membership (in Provo, Utah) with the effects of church attendance in a so-called secular community (Seattle, Washington, which had an exceptionally low rate of church membership compared with other cities in the United States). They report that the correlation between church attendance and delinquency was stronger (gamma =.45) in Provo than in Seattle (gamma =.13).

Instead of conceptualizing the effects of religion on crime and delinquency in a strictly individualistic manner, it is conceivable that the social context may mediate the effects of religion on individual behavior. However, in subsequent research by Cochran and Akers, the moral communities hypothesis received only limited support. Only when offenses are morally ambiguous (as, for example, marijuana use seems to be) does religion have a substantial effect according to Cochran and Akers. They also conclude that the effects of religion are limited primarily to instances where religious norms condemn certain acts (such as the Mormon church proscription regarding alcohol). Linden also concludes that the deterrent effects of religion seem limited to morally ambiguous offenses such as marijuana use.

In the 1980s, some additional articles contributed to the empirical research literature on religion and crime (see, e.g., Elifson, Petersen, and Hadaway; Tittle and Welch; and Peek, Curry, and Chalfant). Tittle and Welch concluded that individual religiosity deterred deviance most in relatively secular communities or areas with a relatively high percentage of religious nonaffiliates. In a subsequent analysis, Welch, Tittle, and Petee concede that their data have several limitations (such as measuring projected or potential, rather than actual, deviant behavior in a data set including only adult Catholics). Stack and Kanavy report that areas with relatively high proportions of Catholics have lower rates of sexual assault. Hadaway, Elifson, and Petersen conclude that church attendance and religious salience have a significant deterrent effect on drug use, including alcohol and marijuana. Peek, Curry, and Chalfant conclude that there is evidence of a decline in the deterrent effect of religiosity on delinquency as the time span between the measures of delinquency and religiosity increases.

In the 1990s, a series of studies has demonstrated that measures of religion are significantly correlated with measures of drinking and drug use (Cochran et al.; Brownfield and Sorenson; Stark; Benda and Corwyn, 1997; Lee et al.). For example, Cochran and others examine the influence of religious stability and homogamy among a sample of Protestant respondents. Their data were taken from the national General Social Surveys conducted between 1977 and 1989. Although Cochran and colleagues report that some measures of religion are not significantly related to alcohol use (e.g., frequency of prayer, belief in an afterlife, and level of confidence in the clergy), they did find significant negative correlations between drinking and church attendance, strength of religious identification, and holding a literal interpretation of scripture. Cochran and others find that religious stability and homogamy in marriage tend to increase the negative relationship between personal religiosity and alcohol use.

Brownfield and Sorenson, in an analysis of the Seattle Youth Study, examined the effects of religion on several types of drug use, including alcohol, marijuana, amphetamines, barbiturates, and cocaine. Using three measures of religion (religious affiliation, church attendance, and religiosity), Brownfield and Sorenson find that all three are significant, inverse correlates of all types of drug use measured.

Bainbridge concludes that individual religiousness can deter larceny and similar crimes within so-called moral or religious communities, while the correlations between religion and drug offenses are not significantly affected by religious context. Benson also concludes that the deterrent effects of religion on drug use are persistent (though occasionally modest) across a variety of settings or contexts. Benson points out that there is a need for more theoretical work on precisely how religion affects substance use.

Benda and Corwyn (1997) analyzed data collected from two southern states to assess the relationship between religion and drug use among adolescents. They conclude that church attendance is inversely correlated with drinking, controlling for demographic factors such as race, social class, age, and family structure. Lee and others conducted a large-scale survey of Seventh-Day Adventist youth in grades six to twelve. More than seven thousand completed questionnaires were obtained from a stratified random sample of the United States and Canada. Lee and colleagues report that youth who did not follow worship practices were most likely to report higher levels of alcohol and drug use.