In the Marquette Law School Poll conducted earlier this month, fifty-nine percent of registered Wisconsin voters agreed that marijuana “should be fully legalized and regulated like alcohol.” Only thirty-nine percent disagreed.

Support for legalization in Wisconsin follows the recent decisions to legalize marijuana in Colorado and Washington in 2012, and in Oregon and Alaska in 2014. Nationally, support for legalization has grown steadily since the early 1990s and finally crossed the fifty-percent threshold in 2013. (On the local level, the Public Policy Forum published a thoughtful assessment of the costs of marijuana enforcement in Milwaukee earlier this year.)

In the Law School Poll, respondents were asked which arguments for legalization they found most convincing.

Most resonant with voters was the argument that “we spend too many government resources on arresting and incarcerating marijuana offenders; it would be better to spend those resources on other things.” Fifty percent of respondents said this argument was “very convincing,” while eighty-one percent said it was at least “somewhat convincing.”

At least half of respondents also found three other arguments very convincing or somewhat convincing:

• “If marijuana were legalized, then it could be taxed, which could help to support important government services like education” (seventy-two percent)
• “Marijuana is not more dangerous than alcohol or tobacco, so the law should treat all three substances basically the same” (sixty-five percent)
• “The government has no business telling adults what they can and cannot put into their own bodies” (fifty-five percent)

Conducted from July 7 to 10, the Law School Poll included responses from 801 registered voters reached by cellphones and landlines, with a margin of error of +/- 4.1 percent. I collaborated on the design of the drug policy questions with Poll Director Charles Franklin and Professor Darren Wheelock of Marquette’s Social and Cultural Sciences Department.

We found more mixed results when we asked about drugs other than marijuana. We wanted to ascertain whether Wisconsin voters favor treatment over prosecution, or vice versa, for four different drugs: marijuana, crack cocaine, heroin, and crystal meth.

A little background is necessary to explain the results. In the September 2015 Law School Poll, Darren and I designed a set of questions to test for “acquiescence bias,” which is the tendency sometimes found in polling research for some respondents to agree with whatever they are asked, regardless of the content of the proposal. In order to test for this, we randomly divided respondents into two groups and asked each group questions that were the mirror image of questions we asked the other group. We did indeed find evidence of acquiescence. For instance, we found that sixty-two percent in one group agreed that mandatory minimum sentences should be increased, while nearly as large a majority (fifty-eight percent) in the other group agreed that mandatory minimum sentences should be reduced.

In order to control for acquiescence in responses to our drug policy questions, we asked half of the respondents whether they agreed with prioritizing prosecution over treatment, and the other half whether they agreed with the opposite. Once again, the results demonstrate that it can matter a lot how a survey question is framed.

Here are the specifics:

• Marijuana: thirty percent agree (“strongly agree” or “somewhat agree”) that prosecution should be emphasized, and sixty-nine percent agree that treatment should be emphasized
• Heroin: fifty-two percent favor prosecution, and sixty-eight percent favor treatment
• Crack cocaine: fifty-one percent favor prosecution, and sixty-nine percent favor treatment
• Crystal meth: fifty-five percent favor prosecution, and sixty-nine percent favor treatment

In other words, when it comes to the hard drugs (heroin, crack, and meth), a majority favors prosecution over treatment, while a majority also favors treatment over prosecution. From a strictly logical standpoint, these results are incompatible with one another. Acquiescence bias presumably explains some or all of the discrepancy.

Nonetheless, there may be a couple of notable takeaways from these results. First, there do not appear to be strong acquiescence effects when it comes to marijuana. Even when framed in a way that implicitly encouraged respondents to take a hard line, only thirty percent favored prosecution over treatment. This result tends to support the significance of our finding that a solid majority favors legalization.

Second, the majorities favoring treatment for the hard drugs were much larger than the majorities favoring prosecution. This suggests that if a more neutral way of asking could be devised, most respondents would probably support treatment, albeit by slimmer margins than the landslide-level figures we found when we implicitly encouraged pro-treatment responses.

A final set of drug-related questions asked respondents how serious drug problems were where they lived. Heroin clearly emerged as the greatest concern. Nearly two-thirds (sixty-three percent) agreed that “heroin is a major problem in the area where I live.” Here are the corresponding numbers for the other drugs:

• Crystal meth—fifty-four percent
• Crack cocaine—forty-three percent
• Marijuana—forty-one percent

Cross posted at Life Sentences.