SVG
Commentary
Hudson.org

Support for Recreational Marijuana: Weights and Measures

 A marijuana advocate who goes by the name 'The Holy Hemptress' demonstrates on October 25, 2011 in San Francisco, California. (Justin Sullivan/Getty Images)
Caption
A marijuana advocate who goes by the name 'The Holy Hemptress' demonstrates on October 25, 2011 in San Francisco, California. (Justin Sullivan/Getty Images)

Quinnipiac University released today a poll covering, among other Colorado issues, measures of support for the state’s new recreational marijuana law, which began implementation on January 1 of last year. (The call-up phone poll was taken of 1,049 self-identified “registered voters,” with a margin of sample error of +/- 3 percent.)

Because there has been considerable negative discussion about the law, including Governor John Hickenlooper calling it “reckless,” the results were somewhat surprising, showing undiminished support at fully 58 percent of Coloradans, with those opposed coming in at only 38 percent.

Some of the surprise, however, fades a bit once the polling sample is unpacked, and the effect of sample weighting is taken into account.

First there is the model Quinnipiac used for distributing party identification. In this sample, while the distribution of Democrats (allocated 27 percent) and Republicans (25 percent) is pretty similar, there is a clear preponderance of Independents (39 percent). These three total only 91 percent.

Perhaps this distribution is a reflection of Colorado voting patterns (if true, it might prove a surprise not only to the other 9 percent but to former Senator Mark Udall, who lost in November his Senate reelection to the Republican upstart, Cory Gardner, 48.5 percent to 46.0 percent—even Quinnipiac’s pre-election poll had Gardner up 45-43). Nevertheless, the lean towards self-identified Independents in the sample tends to load it in the direction of marijuana support.

In the cross-tabs, one finds that 74 percent of Democrats support legal, recreational marijuana, contrasted with only 36 percent of Republicans.

Independents, however, were substantially closer to Democratic sentiment, with 62 percent of them supporting marijuana. So, the more self-identified Independents in the sample, the more likely that the total will be supportive of Colorado’s law.

But if there is a thumb on the scale, it might be found once one turns to the distribution by age. Everyone knows that support for legal marijuana is generational; the younger, the more likely to both support the measures, and to use the stuff. Older Americans are generally more concerned, and less supportive, even if (perhaps, because?) they lived through the ‘60s.

Thus, how Quinnipiac handles the age distribution in their sample tells us a lot about how the overall results came out.

Quinnipiac subdivides the registered voters by three categories: 18-34; 35-54; and 55+. In the actual poll, raw numbers were distributed in the following way: of the total, 18- to 34-year-olds were only 10 percent of the poll respondents. Those 35-54 were 25 percent of poll respondents. But the largest category was those 55+, who were fully 61 percent of the sample.

Moreover, because support for marijuana is strongest among those 18-34 and 35-54 (at 82 and 58 percent support in the respective age groups) and weakest among those 55+ (only 46 percent support), one would expect the outcome of this poll to tend towards lower support for legalized marijuana.

But that would be the outcome from the raw numbers. After weighting, the distribution takes a decided turn.

According to the demographic model used by Quinnipiac, the answers from the 18- to 34-year-olds, rather than being 10 percent of the overall poll, were elevated to 21 percent, an 11 percentage point increase.

For those 34-54, also likely to be supportive of the measure, the weighting took them from 25 percent to 34 percent of the sample, a 9 percentage point increase.

And as for those 55+, the least likely group to be supportive, weighting by Quinnipiac took their contribution to the overall result from 61 percent of the sample to only 40 percent, a rather striking 21 percentage point demotion.

It’s hard not to get the suspicion that weighting of the sample, though a standard practice in the political polling industry, has in this case loaded the dice in favor of marijuana somewhat more than usual.

Finally, there is the question as to whether the liberalized marijuana laws are having an effect. Reassuringly, the Quinnipiac poll tells us that only 19 percent of Coloradans admitted to using marijuana since recreational sales began.

What one finds in the cross-tabs, however, is more troubling. Fully 37 percent of the 18- to 34-year-olds answered in the affirmative (with only 10 percent of Republicans so answering, compared to 26 percent of Democrats), indicating a potential striking increase in marijuana “experimentation” by young people since the law took effect.

That cannot be good news. Someone should take a poll.