We wish to emphasize several findings from our analysis. First, while we concur with the many reviewers who conclude that alcohol use is associated with violence, we also found that alcohol use is consistently associated with nonviolent offending. Several other authors have questioned the conventional wisdom on this point (Felson et al., 2008; Felson & Staff, 2010; White & Gorman, 2000; White et al., 2002). We expected that alcohol use would be the best differential predictor of violence in this chapter, due to the mountain of evidence and clear consensus on an alcohol-violence relationship, but it soon became clear that not all measures of alcohol use perform the same way. While mere alcohol “use” equally predicts violent and nonviolent offending, as does heavy use, alcohol intoxication at the time of offense has a more consistent association with violent offending than with nonviolent offending.
Our findings were divided on studies of “drug use” as an undifferentiated category. “Drug use” is more consistently correlated with levels of nonviolent crime than violent crime, but in offender studies, and there are quite a few, violent offenders are more likely to report drug use and abuse. Because we can probably attribute relations between measures of “drug use” and nonviolent offending to crime committed in the service of acquiring drugs, rather than caused by drug intoxication, we are more dismissive of these correlations in favor of the analyses were offenders were compared. Thus, in spite of compelling reasoning that a mixed measure of drug use would combine drugs that have opposing psychopharmacological influences on violent behavior, we have to include undifferentiated “drug use” as a differential predictor of violence in the existing literature.
Many scholars have concluded that there is little to no evidence of a relationship between marijuana and violence (Boles & Miotto, 2003; Haggard-Grann, Hallqvist, Langstrom, & Moller, 2006; White, 1991; White & Gorman, 2000). We found that marijuana use is more consistently associated with nonviolent crime than violent crime but that marijuana has been positively associated with violence at a greater than chance level in the set of studies reviewed here. In addition, a series of offender studies suggests that violent offenders are more likely to report marijuana use than nonviolent offenders. In one important study, marijuana use prevailed over other drug use as a positive predictor of violence. These studies reveal that it is inappropriate to dismiss marijuana as a drug that is “only” related to minor property offending; it appears to be linked to violence with some frequency, though our instincts tell us that analyses specifically designed to test the differential etiology hypothesis are needed, especially in order to establish a causal relationship between marijuana use and violence.
Based upon previous research and common assumptions, cocaine use was probably the best prospect for a differential predictor of violence among the illegal drugs, but our findings do not suggest that this is so. Our confidence on this point is not strong, however, and we will recommend future research in Chapter 12. We come to the same conclusion as others have about heroin and opiate use. Associations between opiate use and nonviolent crime were more consistent than they were for violent crime.
In summary, our review indicates, unsurprisingly, that some substances, consumed in certain amounts in certain contexts, are differentially related to violence. Overall, though, we were surprised to find that a thorough review of dual-dependent variables challenges the common assumptions that alcohol use is unrelated to property offending and that illicit marijuana or opiate use is unrelated to violence; all substances with psychoactive properties appear to be related to both violence and nonviolent offending to one degree or another.