What a difference a week — or a poll — makes.

On December 6, Corporate Research Associates released its quarterly snapshot of the state of play in Nova Scotia politics, and it seemed to be more of the same-old, same-old, ho-hum, nothing-to-see-here-folks, move-along news story.

CRA concluded the governing Nova Scotia Liberals remained our “preferred party” with 56 per cent support — the same as it had polled in August — while the Tories (20 per cent), NDP (19 per cent) and Greens (four per cent) couldn’t scrape together enough random voters combined to come close to matching those Liberal numbers.

Stephen McNeil. Photo: Halifax Examiner

Although Premier Stephen McNeil had begun to lose some of his Teflon coating, CRA reported he still claimed the backing of 38 per cent of voters, a figure that was almost double that of PC leader Jamie Baillie (20 per cent), and left NDP leader Gary Burrill (11 per cent) and Green Thomas Trappenberg (four per cent) eating his electoral dust.

That was then.

Exactly one week later, the Angus Reid Institute released its own survey of Canadian premier popularity, asking, “Do you approve or disapprove of the performance of each of the following premiers?”

Its results showed Stephen McNeil’s popularity had plummeted further and faster than any of his fellow premiers in the three months since Reid’s previous survey. His 38 per cent approval rating then had fallen to just 31 per cent now. (If anyone is checking — and no doubt the Liberal brain trust is — that’s barely two percentage points higher than former premier Darrell Dexter scored in the last such Reid poll before he and his New Democrats tumbled off the power cliff and went down to ignominious defeat in 2013.)

What happened?

Short answer: teachers happened.

The CRA poll was conducted between November 3 and 29, while Reid didn’t begin its own week of political pulse-taking until December 5.

In between — on Saturday, December 3 — Education Minister Karen Casey announced she was closing all the province’s schools in response to a decision that week by Nova Scotia’s teachers to begin to work to rule, which, in turn, had come in response to the government’s refusal to bargain with them in good faith.

Two days later, on the same day Reid’s pollsters began polling, Stephen McNeil’s government recalled, then un-recalled the legislature to introduce, then not introduce legislation to force the teachers back to work under a previously rejected contract, and Education Minister Casey herself appeared in front of reporters to wave her wand and declare that all was now well (though nothing had actually changed), and schools would re-open the next day… and then promptly disappeared again from view, along with her boss, the premier.

So those specific results may not be surprising. But what do they mean, if anything, in the larger scheme of political things?

Well, it is important to begin any discussion of political polls with the late Canadian prime minister John Diefenbaker’s response to a 1957 poll, which showed he had no chance of winning the upcoming federal election that would, in the end, make him prime minister: “I’ve always been fond of dogs,” Diefenbaker told reporters, “and they are the one animal that knows the proper treatment to give to poles.”

We should add, by way of another caveat, that while the steady-as-she-goes CRA poll was based on a sample of 801 adult Nova Scotians (accurate to within plus or minus 3.5 per cent 19 times out of 20, as the pollsters say), the Reid survey was based on a less robust sample size of just 300 Nova Scotians, creating a margin of error of 5.6 percentage points on your 19-times-out-of-20 scale.

And finally — and most importantly — polls are inevitably simple snapshots in time. And times change quickly. As if we needed reminding. Consider the last few weeks in Nova Scotia. Consider Donald Trump.

But the most intriguing  — and meaningful — numbers may be found in the Reid trend lines. From a post-election high of 66 per cent approval in June 2014, Stephen McNeil has been on a mostly downhill, 35 per cent tumble through voter-alienating controversies like the film tax credit, Pharmacare, the Yarmouth ferry, public servants, teachers… to the reality that, today, just over three in 10 voters approve of the job he’s doing.

My own unscientific, un-poll-tested view is that McNeil’s broad support in the polls during much of his years in office was always more shallow than deep, more reluctant than heartfelt, more apparent than real. Voters were still recovering from a bitter break-up with the NDP. They’d never warmed to Jamie Baillie’s courting. In the long interregnum between majority government elections, they were simply parking their votes and keeping score.

Can new NDP leader Gary Burrill bring back the progressive urban NDP vote that once allowed the party to win metro’s major riding bloc all for itself? Can Jamie Baillie, who has finally re-discovered there is a progressive in Progressive Conservative, claim the middle ground in a polarizing province?

The next provincial election campaign, whenever it comes, will likely be far more interesting — and possibly consequential to the outcome — than the last two pre-determined, throw-the-rascals-out, we-have-an-alternative campaigns.

And, for political junkies, that may make 2017 worth waiting for.

Happy New Year!

Stephen Kimber

Stephen Kimber is an award-winning writer, editor, broadcaster, and educator. A journalist for more than 50 years whose work has appeared in most Canadian newspapers and magazines, he is the author of...

Join the Conversation


Only subscribers to the Halifax Examiner may comment on articles. We moderate all comments. Be respectful; whenever possible, provide links to credible documentary evidence to back up your factual claims. Please read our Commenting Policy.
  1. Some time in the early 20th Century, the New York Sun and the New York Post had an extended editorial turf. At one point the Post accused the Sun of “yellow dog journalism.” The Sun responded:

    “The Post accuses us of ‘yellow dog journalism.’ Our attitude toward the Post is that of any dog toward any post.”

    1. Hahahaha. Thank you for this at the end of the angry comments above.

      Really, there’s absolutely no reason to doubt that McNeil’s approval has suffered. There are almost 20,000 people who have joined the Nova Scotia Parents for Teachers Facebook Page, and many people also showed their support in person during demonstrations. Social media lit up in a way that was not favourable to the Liberal strategy. I’ve got advanced courses in statistics but I don’t need them here, one only needs to look around to know that any polls showing a decline in approval are likely correct. I’m about 99.9% confident in saying that.

  2. Good time to be a pollster in NS. Two phone calls to the house yesterday…two different polls. Both asking variations on approval of the Libs vs others and the NSTU follies.

  3. How can you say this is an “extreme left echo chamber” when all of our middle-of-the-road parties are referenced, where no one party is supported in the article, and where the article itself points out the unreliability of polls but that they remain dominant in political discourse such that reporting on them as a phenomenon is entirely relevant.

    1. Kimber in particular and the Exaiminer as a whole have yet to convince me of their non-membership to the extreme left echo chamber. Given the heavy angle on this and, well, EVERY SINGLE ARTICLE it is hard to really claim a balanced view or perspective.

      The seeming willingness to compare surveys that are wildly off in number as it truths comes to mind.
      On the ARIpoll, it clearly shows 38%, yet CRA’s significantly more scientific poll shows over 50%, but that is before December 6th and it is juxtaposed with the ARI report at 31%. It states later that it had a support level of 38%, which is from the ARI poll and I guess some sub-segment of hard-line supporters in the CRA poll, but they are not the same number, they do not represent the same thing, and seem to be used to connect the two surveys, which have little to no connection at all, to the point where one is scientifically significant and the other is not.

      I guess if you are looking for the Lizard-King McNeil to look bad, you only ever seeing him looking bad. Just wait until they get a picture of him eating a baby. I saw it. He lights them on fire with his breathe first. Then he eats babies. McNeil is a Lizard-person who eats babies. There, I FINALLY said it.

      Well, at least that’s what every Kimber and most Examiner articles seem to be leading to, anyway.

      1. I agree that comparing the numbers from the two polls – with different samples, methods, contexts, and questions – is questionable. Among other concerns, the ARI poll compares McNeil to other premiers, the CRA poll compares him to provincial political party leaders, implicitly or explicitly, the bias is the same.

        The chart comparing the current ARI numbers to past numbers is a more robust piece of quantitative data, and it is consistent with the narrative that McNeil is facing declining popularity, and certainly the cause is pure speculation. Acknowledging that 31 is less than 66 is not particularly partisan.

        With all of that being said, I think Tim’s views on many issues are pretty clear, that many of them would be classified as left-of-center, and it is not surprising that readers who identify with those views are more likely to subscribe (and thus comment). To me, one of the features of the Halifax Examiner is informed, educated, engaged debate on these views from people who don’t share them or hold them to a lesser or greater degree.

        I have no information on Premier McNeil’s genetic lineage, nor on whether or not he eats babies, nor on, if he does eat babies, the method of their preparation. 🙂

  4. Overly simplistic, almost not worthy of debate.

    Angus Reid uses a 2% ME with a total sample size of 5400 Canadians with a Confidence of 95% (or so I am told, they seem to lack figures, even from their own methodology document. Hmm). In order to have an accurate (I mean statistically valid sample) it would require over a 1000 participates in their NS study.

    They have 300.

    For that to be a valid sample size, it would require a confidence interval of like.. 50% ish, significantly below the 95% Confidence of most polls and the 2% Margin of Error advertised.

    Please, research for yourself: https://www.surveymonkey.com/mp/sample-size-calculator/

    That strips all the fancy math out.

    Every since the Raise of Trump, I have paid more attention to polls. This polling… is bullshit. And I think any self-respecting Journalist would have been able to read the sampling number from the report to figure that out. But I forgot, this isn’t for self-respect. This is the extreme left echo-chamber.

    I am sorry, I forgot. I guess I’ll just leave those actual facts here.

    1. The regional MOE for the Reid poll is 5.6% for Nova Scotia, which requires a sample of 307 people to achieve 95% confidence. That is clearly stated in both the poll and Kimber’s assessment of it.

    2. Yeah… the margins of error given in the article and in the poll itself are statistically correct and line up with the usual standards for polls. Perhaps the article was corrected after this comment, but usually there is an accompanying note. As of right now, it is not claimed anywhere here that the margin of error was 2%.

      1. A lot of what I did was back of napkin math and I believe I read that number from the holistic numbers posted from the survey itself.

        But, usually, you work toward limiting the Margin of Error, between 2 to 5%, and you work out the calculation of how many people you should sample in order to get within that range then you survey more people. Which is why you see CRA samples 800 people, I believe it is to get a Error of around 3%. For 3% is around 700 for 5% is it almost 400.

        Instead it appears they got 300 people willing to do the survey, and adjusted the Error in their “afterthought” section to compensate. I.E. Poorly developed survey.

        Regardless, I still find it stupefying the level of bias and lack of context in general on the Examiner and with Kimber’s articles. Literally, the only context or balance that was brought in all the articles I have read here from Kimber so far was when he said these numbers are a snap-shot in time. That’s it. Only context in many articles. I do have to say I turn-off after the third or fourth completely biased statement and stop reading, which I assume is the goal of all newspaper article writers.

        This article is the most “fair”, given the subject matter and how it was handled, yet it still comes out of a heavy anti-Liberal anti-McNeil vibe. The Angus poll? Doesn’t track the opposition leaders for THEIR numbers. If it did, he didn’t post them. (I can’t find them either).

        It’s just, I am getting sick of literally everyday seeing an incredibly biased article after incredibly biased article but only for one party about issues that tend to involve several parties. I am not even talking political parties, it just seems that the Lizard-people Liberals are the reason that Death and Disease happen and they stay up at night to make babies cry. That’s what I get EVERY FUCKING DAY. It would be nice to mix-it-up a bit, maybe?

        I somewhat thought this was Kimber getting it out of his system to move on to maybe more context oriented pieces, like this with statistics. Then it takes a heavy bias.

        It just would be nice to actually have a good newspaper in Nova Scotia. This is turning into a Labour newspaper. Which, as a Union member, I wouldn’t mind reading but I would just accept that whatever I read was completely biased in a way and be open-minded about it. But a biased Labour booklet is really not the standard we should have for a newspaper. At least, not in my opinion.