In light of these lingering uncertainties, we turn to the central analysis of this chapter. In order to gather pertinent studies, we conducted a search using
Criminal Justice Abstracts to find studies published between 1980 and 2013. We paired search terms related to substance use with search terms related to offending. Search terms for substance use included alcohol, drug, marijuana, cocaine, crack, crystal meth, methamphetamine, heroin, narcotic, and opioid. Outcome terms included homicide, robbery, assault and violent crime paired with the terms theft and burglary. Note that we used Boolean logic operators to pair our search terms so as to maximize the likelihood of finding studies that reported both violent and nonviolent dependent variables. Our searches turned up many thousands of returns, the majority of which were not germane to our effort.
We reviewed the titles and abstracts of the articles, books, and research reports contained in these search results, and the studies selected for our review generally fall into one of two categories. One set of reports tests the relationship between substance use and/or availability and both violent crime and nonviolent crime. A second set of studies compares substance use between violent and nonviolent offenders. We included studies that tested various levels of substance consumption, ranging from casual use to binge consumption to addiction. We did not include studies that measured noncriminal forms of “aggression.” We excluded analyses of drug dealing as an independent or dependent variable on its own from our pool. The majority of the studies report analyses of data from the United States, but some data are from the United Kingdom (Ensor & Godfrey, 1993; Farrington & Lambert, 1994; Lasley, 1986; Richardson & Budd, 2003), Finland (Felson et al., 2008), Sweden (Benson & Holmberg, 1984), Belgium (Colins, Vermeiren, Schuyten, & Broekaert, 2009), New Zealand (Fergusson & Horwood, 2000; Fergusson, Lynskey, & Horwood, 1996), Australia (Baker, 1998; Makkai, 2001), and Canada (Baron & Hartnagel, 1997; Levi et al., 2010).