My cannabis poll is better than yours

The inherent bias of polls makes them meaningless

0

Next time you read or watch a news story about a survey or poll, stop reading or turn it off. It doesn’t mean anything. It won’t help you understand anything. You’re better off watching videos of parkouring goats or hawks knocking drones out of the sky.

Polls, which are surveys of public opinion taken from a select sample, have become staples of newspaper, television and online news. We’re inundated by them. Each is subject to its own methodology, bias and interpretation, and I’ll grant that they can offer simple snapshots into peoples’ beliefs or positions at a given time. Today they are much more than that, with hundreds of polls conducted daily on nearly every subject.

Besides that there are probably too many of them out there, one of the problems is that most media outlets — print, television and especially Internet — seem to feed off polls and surveys, generously using them, especially in election coverage. A casual look at the polls of Colorado citizens on their feelings about cannabis really leaves you wondering why media, which often do little more than regurgitate survey press releases and add misleading headlines, won’t just come out and say that polls are often wrong or unreliable.

The classic poll story, of course, is that of Nate Silver, who rated other pollsters according to their biases and methods and wound up predicting correctly every legislative race in 2012.

Who could forget a pitifully out-of touch Karl Rove phoning Fox News on election night after it called the election for Barack Obama because his polls were telling him that Mitt Romney was going to win? If polls are so important, why are so many so wrong so often?

Let’s take a quick look at five surveys done since Colorado’s cannabis legalization began Jan. 1. Never mind that legalization is only in its 10th month, pollsters have still been busy asking residents different questions and media have come up with fascinating ways to interpret the statistics to their own liking.

A Quinnipac University telephone survey released on Feb. 10, fewer than six weeks after retail stores were allowed to open, asked whether Colorado voters felt legalization might hurt the state’s image. Fifty-one percent of voters agreed and 38 percent disagreed, with Republicans opposed 73-18 percent and Democrats supportive by a 57-36 percentage.

A survey done by the Democratic firm Public Policy Polling released on March 19 found 57 percent of state voters supporting legalization and 35 percent opposed. A second Quinnipac poll on April 28 asked whether voters supported the law, and found, by a 54-43 percentage, that they did.

Then the Marist Institute released one on Sept. 7 which found 55 percent support among Colorado residents and 41 percent opposed, with eight percent actively trying to overturn the legislation.

Finally, on Sept. 17 a Suffolk University/USA Today poll asked likely voters whether Colorado should repeal legal cannabis. This one found 50.2 percent saying they disagree with the decision to legalize and 46 percent in support, and 49 percent disapproving of the way the state is managing legalization.

The big thing to note here is who the surveys asked.

“Voters are those who in response to a standard poll question say they are “registered to vote in their precinct or election district,” according to the Gallup Poll FAQ page. “Likely voters are that group of individuals who the company can estimate [my emphasis] are most likely to actually vote.”

That’s an important distinction, and it means that the polls are asking two different subsets of people. The two polls that asked voters and the one that asked state residents each found about the same support for legalization as there had been in 2012. The Suffolk poll of likely voters indicates slightly lower numbers.

Given that most polls offer a few percentage points either way as margins of error, the numbers are actually pretty consistent. That didn’t stop some media outlets from trumpeting the Suffolk poll.

“Colorado voters may be having second thoughts about the legalization of marijuana,” 9News wrote. “A slight majority of voters [50.2 percent] say they do not agree with the decision to legalize recreational marijuana while 46 percent continue to support the deci sion.

Nearly 49 percent do not approve of how the state is managing legalized pot, compared to 42 percent who approve.”

But the Suffolk poll, remember, asked likely voters, not voters, as the story says. This is mixing apples and oranges and needs to be noted when we’re talking polls and public opinion.

The Gazette in Colorado Springs took the Suffolk results a step farther. It quoted a national telephone random sample survey conducted by the conservative Public Religion Research Institute that found national support for legalized marijuana falling from 51 percent in 2013 to 44 percent this year.

“Two recent polls show that support of legalized marijuana has waned,” the story ominously states. “And that there is unhappiness in the way regulations are handled by the state.” Really, all it “shows” is that one telephone survey of people outside Colorado says support has fallen a few percentage points.

Oh, and by the way, the Public Religion Research Institute study contradicts 2014 surveys by Pew Research, NBC News, CBS, CNN, Wall Street Journal and Gallup polls, all of which indicate continued, growing support for legalizing cannabis nationwide.

And see how easy that was to use polls to make my point?

The only reliable figures are that 1,383,139 Colorado voters (55.32 percent) voted for Amendment 64 in 2012. The rest is just shifting sand.

Respond: letters@boulderweekly.com