Musings —10.02.2014 05:12 PM—
If you are likely to like what likely voters like, then you are likely to be wrong. Get it?
No? That’s alright – nobody else really understands what “likely voters” are, either. But pollsters are pontificating about “likely voters,” yet again, and all of us should be wary.
Case in point: this week, an outfit calling itself Angus Reid Global declared that, among “likely voters,” the Harper Conservatives and Trudeau Liberals are now tied.
Said the pollster: “Among those who are likely voters, the Liberals and Conservatives are tied, with both parties earning the support of 33 per cent.”
Really? That’s not what the past year-and-a-half of polls have said. In all but a few cases, in fact, Justin Trudeau’s Liberals have been shown to be ahead, or far ahead. Why is it so different with a poll looking at so-called “likely voters”?
In their appended page about methodology, the Angus Reid people said they applied “a weighting structure that further adjusts our sample to reflect known variations in voter turnout – specifically across age groups – while also filtering based on respondents’ own identified reported past voting patterns and habits.”
What does that mean? Who knows. But it, um, likely means the pollster changed the results they got to (a) reflect the fact that older Canadians tend to vote Conservative, and vote more often and (b) reflect the fact that if a Canadian has voted one way in the past, they are likely to do so in the future.
If that sounds, er, like a reasonable way of peering at mountains of polling data, you’re not alone. This writer, and many others, got burned in Ontario’s election, big time, when we started to believe in this “likely voter” category hooey.
Right until election night on June 12, it made sense to us that “likely voters” were the demographic that pundits and politicos need to pay the most attention to – and, as such, the Ontario Liberals and Ontario PCs were therefore tied in voter intention. According to “likely voters,” Ontarians wanted change, and they were prepared to consider Tim Hudak’s PCs as the best vehicle for delivering it.
On election night, however, some of us were over at the Sun News Network, readying to pontificate about the results. Before the polls closed, I ran into pollster David Coletto, and asked him this: “Um, have you pollster guys worked out what this ‘likely voter’ category is, perchance?”
Said David: “No.”
The rest is history. The “likely voters” weren’t nearly as “likely” as we’d been told, were they? A (very likeable) Kathleen Wynne and her Ontario Liberal election team won a majority when (a rather dislikable) PC campaign team – a team that snatched defeat from the jaws of victory with their 100,000 pink slips craziness.
Do we in the media, and sundry pollsters, learn from past mistakes? Ha! Surely you jest! Here we are, three months and a bit later, and we’re back to believing balderdash and baloney about “likely voters.”
Once bitten, twice shy. Fool me once, yadda yadda.
Just forget about the polls, folks, which get things wrong more than they seem to get things right, these days. Justin Trudeau is ahead, full stop.
That’s what my gut is telling me, and I shouldn’t have ever stopped going with it. Hasn’t failed me yet.
You should do, um, likewise. You’re likely to be closer to the likely outcome.