In a Sunday commentary in the Star Tribune and later in an interview on Minnesota Public Radio, Hennepin County Sheriff Rich Stanek made the classic mistake this week of confusing correlation with causation.
While arguing against the legalization of marijuana, Stanek noted that “approximately 54 percent of males arrested for violent crime test positive for marijuana in Hennepin County.”
That observation, he stressed in both the commentary and the interview, points to “a direct connection between marijuana and violent crime.”
Well, it may show a correlation between the two, but it doesn’t show a causation. Just because half of the men arrested for violent crime in Hennepin County test positive for cannabis (a drug that can apparently linger in the body days or even weeks after it is used, by the way), doesn’t mean that the cannabis caused their criminal activity.
It could also mean that men who commit violent crimes are, for whatever reason, more likely to smoke marijuana.
“A” may be associated with “B.” But that doesn’t mean “A” caused “B.”
Research is inconclusive
Actually, the relationship between marijuana and violent crime is still very much up in the air, scientifically speaking. That’s the clear message from a long but painstakingly thorough paper on the topic that was published in the Journal of Drug Education in 2011.
In the paper, Southern Utah University sociologist Michael Ostrowky reviews the leading theories and key research on the relationship between marijuana use and aggressive/violent behavior.
He concludes that the findings are all over the place and thus inconclusive.
“Taken together, the results of some studies suggest that marijuana use and violence are positively associated, some research has found no association, and other studies even reveal that marijuana use can reduce aggressive behavior,” he writes. “These conflicting findings are not overly surprising, considering that marijuana has been classified at different times by different investigators as a depressant, a stimulant, a hallucinogen, and a narcotic.”
Here’s a sampling of some of the conflicting studies discussed by Ostrowsky in his article, starting with those that found no significant relationship between marijuana use and violent behavior:
Utilizing longitudinal data from a community cohort of African Americans, Green, Doherty, Stuart, and Ensminger (2010) did not find an association between heavy adolescent marijuana use and violent crime.
Pedersen and Skardhamar (2010) examined data from the Young in Norway Longitudinal Study and found no evidence that use of cannabis is associated with an increased risk of subsequent non-drug-specific criminal charges, such as violence.
In a qualitative, cross-sectional study of Canadian clients attending substance abuse treatment programs, Erickson, Macdonald, and Hathaway (2009) discovered that alcohol, cocaine, and crack were the substances most implicated in the client’s descriptions of their drug-related violent incidents. In fact, the authors report “it is noteworthy how seldom cannabis was mentioned.”
Using cross-sectional internet survey data, Denson and Earleywine (2008) found no relationship between marijuana use and aggression once other factors were taken into account. In fact their data suggest that marijuana use is not related to aggressive behavior even among frequent, long-time users.
Here are Ostrowsky’s summaries of a couple of the studies that found marijuana use was associated with decreased aggression or that reported mixed findings:
In a four-wave analysis, White and Hansell (1998) observed that heavier marijuana use in early to mid-adolescence predicted decreased aggressive behavior in later adolescence and young adulthood.
In a cross-sectional analysis, Arendt et al. (2007) studied 119 marijuana-dependent subjects and discovered that subjects who reported problems controlling their violent behavior more often used marijuana to reduce their aggression, but while intoxicated they more often reacted with aggression. However, even among this group aggression was one of the least likely reactions during marijuana intoxication.
And here is his summary of some of the studies that found a positive association between marijuana use and violent behavior:
In a study of 85 patients in the first episode of psychosis, Harris et al. (2010) report that regular cannabis use was associated with serious aggression.
Using cross-sectional survey data on adolescents, Walton et al. (2009) discovered that teens reporting peer violence and dating violence were more likely to smoke marijuana.
Brady, Tschann, Pasch, Flores, and Ozer (2008) examined longitudinal interview data on Mexican-American and European-American adolescents, and they found that adolescents who had used marijuana at age 15 were more likely to report violence perpetration at age 19.
Analyzing cross-sectional data from Dutch adolescents, Monshouwer et al. (2006) report that cannabis use was associated with aggressive behavior.
Note that none of these latter studies was designed in a way that could prove marijuana use caused aggressive or criminal behavior.
Research with prisoners
In his paper, Ostrowsky also reviews the research that specifically involved prison populations. These findings, he says, are less equivocal than the general studies.
“The majority of research on high risk and incarcerated samples has not found a relationship between marijuana use and violent behavior,” Ostrowsky concludes.
Sheriff Stanek may wish to take note.
As Ostrowsky points out, a major problem with many of the studies that found an association between marijuana and aggression was that they failed to control for confounding variables — most notably, alcohol and hard drug use.
“Future researchers need to rule out spuriousness before they proclaim that marijuana use ‘causes’ violent behavior,” he writes.
Ostrowsky concludes that “[i]t is evident from the inconsistent findings in the literature that the exact nature of the relation [between marijuana use and aggressive/violent behavior] remains unclear.”
Any statement, therefore, that declares an unequivocable “direct connection” between the two is inaccurate.