Tag Archives: election polls

Polling the pollsters


The urgent news of the day for New Brunswick is in. It’s now time to close the shutters, batten down the hatches, head to bed with a nice cup of steamed hemlock. For, the pollsters have spoken.

Those who moil for meaning in the nether reaches of the online world – the Internet, blogosphere, social media – have tendered their latest profile of public opinion about Canada’s most youthful premier.

Now we know, and now we may sleep, comfortable in the knowledge that half of this province’s grown-ups think Brian Gallant is just swell; the other half isn’t so sure.

Says Corporate Research Associates of Halifax: “The New Brunswick Liberal Party continues to be preferred, with just under one-half of New Brunswick decided voters supporting (45 per cent, down from 55 per cent in November 2015).

“Meanwhile, one-quarter back the PC Party of New Brunswick (27 per cent, compared with 25 per cent), while two in ten residents support the New Democratic Party (18 per cent, up from 12 per cent). Green Party support is stable (eight per cent, compared with seven per cent), while two percent of voters back the People’s Alliance (compared with one per cent).

“The number of residents who are undecided rests at 29 per cent (compared with 25 per cent), while seven per cent refuse to state a preference (compared with nine per cent), and five per cent support none of the parties or do not plan to vote (compared with three per cent).”

Of course, I’m reasonably certain that, should I turn the tables on the polling industry, itself, public responses would track along predictable lines.

Question: How much do you trust polling data?

Answer: About as much as I trust politicians.

Question: How much do you like being bugged by pollsters while eating supper or beating a deadline?

Answer: About as much as I like answering the door on a sweet, sultry Sunday afternoon.

According to writer Nate Cohen in the New York Times in January, “The polling industry has been hit hard by high-profile misfires in recent years. Exactly why the polls err often remains a mystery. Potential sources for error abound: The initial samples could be biased, the likely-voter models may not reflect the actual electorate, or voters could make last-minute decisions that make even an accurate poll wrong on Election Day.”

Mr. Cohen also references a Pew Research report that declares: “Polls have failed to accurately predict winning candidates in several recent elections, including the 2015 race for governor in Kentucky, several 2014 U.S. races for Senate and governor, the 2015 British general election, the 2015 Scottish referendum on independence, and the 2015 referendum in Greece on acceptance of the European Union’s terms for a bailout. In the 2012 U.S. presidential election, many surveys underestimated the share of the vote that Barack Obama would receive. Errors in modeling the likely electorate are suspected of contributing to many of these polling failures.”

Or could the problem simply be the intellectual triangle pollsters, politicians and the press have managed to forge over the past few decades? After all, these are the only “audiences” who seem to benefit from periodic public opinion surveys.

We, the great polled, couldn’t care less; except, of course, enough of us are more than willing to offer an opinion when gently pressed to do so.

Is Brian Gallant the greatest thing since sliced bread? Sure. Nope. Doesn’t matter. You’ve answered the question, done your civic duty. Now go to bed as images of real and important matters fail to dance in your heads.

Tagged , , ,

Shooting the messenger at election time

Leaves of grass for NB's labour market

Few in these jaundiced times espouse an unshakeable faith in much of anything. But those handful who do believe in the primacy, if not permanence, of polling numbers might be disappointed in events presently unfolding in Scotland.

There, the Scottish National Party, under the spirited leadership of Alex Salmond is rallying it supporters of independence before next week’s historic referendum, the outcome of which could redraw the map of Great Britain both figuratively and literally.

According to Griff Witte, writing this week in the Washington Post, “The once-unthinkable prospect that Britain could be ripped apart this month with a vote for Scottish independence became bracingly real Monday after the campaign to keep the three-century-old union together was accused of panicking amid polls showing the referendum in a dead heat.”

Indeed, “Just 10 days before the vote, the new surveys depicted a dramatically tightening race after months in which the ‘no’ side appeared to hold a comfortable lead. Although both sides have questioned the accuracy of the Internet-based polls, the pro-independence camp immediately claimed the momentum.”

In fact, until last month, Scottish naysayers (those who wish to remain in the United Kingdom) accounted for between 60 and 70 per cent of intended voters. The ‘yes’ forces, in contrast, had trouble breaking above 40 per cent. Now, it seems, those in favour of Scottish independence are nudging the 52 per cent mark.

This is why those of us who know a little something about statistics, approach all numbers meant to startle, scare or otherwise provoke only warily.

Still, election polls are notorious, not so much for their inaccuracy but for their unreliability from one day to the next.

For this reason, they’re also the source of some of the most heated debates, sometimes eclipsing all other, more relevant, issues, as candidates desperately fear being trampled by the herd mentality on voting day.

Indeed, when the circumstances are ripe, even some pollsters will wade into the fray. Witness, for example, Corporate Research Associates chairman and CEO Don Mills last week instructing his lawyers to fire off a stern missive to New Brunswick Progressive Conservative Leader David Alward regarding some unfortunate wording the latter deployed during one of his many stump stops around the province.

“CRA has been great over the years at playing games,” Alward had told an audience of supporters, following the release of its latest polling data showing the Liberals ahead of the PCs in popular support (49 versus 29 per cent). 

“You only have to go back to the last election when in the weekend leading up to the voting, they were saying it was too close to call or even that we were behind. In reality it was a 42 to 13 landslide.”

In a statement, Mr. Mills retorted: “Through hard work and diligence, CRA has built its reputation as a non-partisan public opinion polling company since its founding in 1978. Comments attributed to Mr. Alward impugn that reputation and imply bias in our work.”

If they do, it wouldn’t be the first time a frustrated politician has shot from the hip at political pollsters.

“Gov. Chris Christie wasted little time in taking aim at pollsters during his latest town hall event just as a recent poll found the governor’s job approval rating is plummeting amid the ongoing George Washington Bridge controversy,” reported PolitickerNJ last winer.

“The governor started the event discussing the weather, telling residents on another snowy day in the state that there are people in two professions who continue to get paid despite getting it wrong time after time. Meteorologists? Of course, he said. But according to New Jersey’s governor, there’s another group of workers in the same pool: Pollsters. ‘They don’t ever have to have it right,’ Christie said to laughs from the crowd.”

At best, political polling is an accurate snapshot of people’s opinions and attitudes at the time of asking. They can, and do, suggest longer-term trends. But the reliability of those trends is in direct proportion to the number of people who will never change their mind – who will, with an unshakeable faith in their own world view, vote as they say they will regardless of sound facts and arguments that militate for alternatives.

Fortunately, the world doesn’t work that way. Just ask the Scots.

Tagged , , , , , , , ,
%d bloggers like this: