Editor’s note: Jon Margolis is VTDigger’s political columnist.
Denial, as has been noted, is not just a river in Egypt. It’s also an easy out, a convenient way to ignore unwelcome news, and it’s been much on display in Vermont these days.
The current outbreak was prompted by the latest poll from the Castleton Polling Institute. It showed that most Vermonters favor: stronger gun control laws, physician-assisted suicide, a single-payer health care system, and more wind towers producing electricity.
In a sensible world, opponents of all of the above would use the poll results as a guide to help them figure out better ways to convert folks to their point of view.
In our less-than-sensible world, most of them went immediately into the denial mode. The poll is wrong, they say. The sample is unrepresentative. Or the questions were misleading. How can a mere 616 people represent 600,000 Vermonters?
Besides, didn’t this same polling firm blow it big-time last year when it put out a poll showing Attorney General Bill Sorrell 19 points ahead of his Democratic primary opponent? Sorrell barely won.
Not Castleton’s finest hour, but not a sign of ineptitude, either. Rich Clark, the director of the Institute, warned everyone at the time not to read much into that result because too many poll respondents said they planned to vote in the primary.
“It was near or over 50 percent,” he remembered the other day. “The actual turnout was closer to 10 percent.” And as Clark pointed out, there was nothing wrong with Castleton’s polls on last year’s other races. The firm knows what it’s doing.
The temptation to reject inconvenient poll data is by no means unique to Vermont. Back in 2004, many Democrats insisted those polls showing George W. Bush ahead of John Kerry had to be wrong. Last year, statistician Nate Silver, who was studying polling from the key states, was pilloried by conservatives because he kept reporting that President Barack Obama was likely to be re-elected.
In both cases, the critics were wrong; the polls were right.
So is this one. Most Vermonters favor stronger gun control laws, physician-assisted suicide, a single-payer health care system and more wind power. Opponents would be smarter to deal with it than to deny it. They’re in the minority.
That doesn’t mean they’re wrong. The public can be bamboozled, unenlightened or indifferent.
Also malleable. Gerald McMurray of Arlington, a leader of the True Dignity Vermont organization which opposes physician-assisted suicide, noted that last year polls showed a majority of Massachusetts voters favored a similar proposal, which was the subject of a referendum. It lost.
So those pleased by the poll results strutting around claiming that the debate is now over are as wrong as the deniers. Voters can be persuaded to change their minds. Just ask President Dewey.
But attempts to discredit this poll reflect discredit only on the attempters. In an email to supporters, the anti-gun control group vtguns claimed that the “CSC poll is in stark conflict with national polls on gun control.”
It is not. Almost all those polls show comparable majorities in favor of stronger gun control laws.
Polling is a useful but limited tool. It measures the public’s opinion – often barely distinguishable from prejudice – but not its knowledge, and it does precious little to add to that knowledge. Polling doesn’t have much nuance.
But it performs its superficial task – gauging how many folks are pro or con a candidate or a position – quite well. And for those still confused about how (as one commenter to the VTDigger story about the gun control results put it) “620 polled of 600,000” could possibly be representative, herewith a very brief explanation of probability.
On a table in the office of one of Washington’s leading polling firms is an oversized brandy snifter filled with oversize blue and white marbles. Visitors are urged to reach into the bowl and, without looking, pick out 10 marbles.
Usually, they will pick out seven white and three blue marbles, because that’s the proportion in the snifter. Sometimes it will be six and four or eight and two. That’s the margin of error. Very, very, rarely will any other ratio result.
Not never. The basis here is probability, not certainty. But any polling firm that knows how to create a competent random sample and how to phrase questions in a neutral manner (and Castleton does) can accurately measure public opinion.
Accurately does not mean precisely. When the Castleton poll reports that, with a plus-or-minus 3.9 percentage point margin of error, 61 percent of Vermonters favor banning the sale of assault-type weapons, it is effectively saying that could one question every adult in the state, between 57.1 percent and 64.9 percent of them would have supported the ban at a “95 percent confidence level.”
That’s a fancy way of saying “19 out of 20 times.” Meaning every 20th poll could be wrong, an “outlier” in polling jargon.
Could this one have been an outlier on gun control?
No, because it is consistent with other poll results. A Vermont poll taken several years ago showed a majority in favor of banning those weapons. So did several recent nationwide polls, with the margins highest in the Northeast and among people who tend to vote Democratic. Vermont is a Northeastern state in which most people tend to vote Democratic.
Some poll skeptics also doubted that the poll sample accurately represented Vermont’s electorate. Noting that more than half the respondents were college graduates from households with more than $40,000 a year in income, anti-wind activist Annette Smith said, “It does not seem to be representative of Vermont.”
When it comes to income, though, it is.
“The median income for a Vermont household in the period 2009-2011 was nearly $53,000,” noted Michael Moser, a research specialist at the Center for Rural Studies at the University of Vermont.
The college graduate percentage of the sample is higher than the state’s. The Census, Moser reports, concluded that only 34 percent of Vermont adults had a four-year college degree.
But look at the poll results. Those without college degrees were actually more likely to favor “building turbines along the state’s ridgelines” than their bachelor-degreed fellow citizens.
Poll results are not the final word on many issues because the surveys often measure public opinion which is weakly held.
On only one question – about banning the sale of high capacity ammunition magazines – was there a big difference between respondents who had been to college and those who had not. But adjusting that result to account for the disproportionate number of college graduates (“weighting” in poll jargon) would have only whittled a few points off the huge (67 percent to 29 percent) majority in favor of the ban.
Not everyone unhappy with the results argued that the poll was wrong. Jeffrey Wennberg of Vermonters for Health Care Freedom, which opposes the single-payer concept, said, “The notion that around half of Vermonters (52 percent to 30 percent in the poll) are favorably disposed toward single-payer, given the near 100 percent monopoly on messaging (by the pro single-payer faction) we take no issue with that at all.” He added that he was confident public opinion would turn against the proposal once voters understood how much it would cost.
Clark said he had not yet had time to analyze the results by regions within the state. When he does, it might reveal less support for wind development in the Northeast Kingdom. The poll did show slightly less support for wind projects statewide than earlier surveys, and there is ample evidence that public opinion has turned against wind power in the Kingdom, site of two big wind projects and two more in the offing. Several town boards and the regional planning agency have called for a halt in approving new wind projects. So have almost all the region’s legislators, who are not likely to be miscalculating the wishes of their constituents.
But the three NEK counties combined (Caledonia, Essex, Orleans) have less than 10 percent of the state’s population. So even a sharp drop in support for wind power – or anything – in that region is not likely to have much impact on the statewide total.
Poll results are not the final word on many issues because the surveys often measure public opinion which is weakly held. Health care and gun control are central issues which have been on the public agenda for decades, so most people probably have some knowledge and at least general policy preferences. But unless they live near an existing or proposed wind project site or have seen a loved one die a lingering, painful death, most voters have probably not paid much attention to those issues.
That gives opponents the chance to try to change public opinion. The best place for them to start is by not denying that it is what it is.
