New poll shows a slightly closer race. Maybe. Perhaps.

Readers of my two recent posts (here and here) on the Corporate Research Associates rolling poll for the Halifax Chronicle-Herald* should consider a new poll, commissioned by the right-wing Sun News Network, that shows a somewhat closer race than CRA’s eyebrow-raising results.


Sun headlined the 48-26-25 split, but it’s worth noting that when Abacus Data, the firm that conducted it, counted only respondents it deemed likely to vote, the numbers edged closer to the CRA results: 51-26-23. The likely voter sample included only 243 respondents, which yields a sampling error of ±6.4 percent, 19 times out of 20, somewhat higher than CRA’s. In short, the two polls are within each other’s margin of error.

Abacus showed the PCs nipping at the NDP’s heels, while CRA had them further back.

Although the Abacus sample was smaller than the seven day rolling average reported by CRA, all its calls were made on Monday and Tuesday, so it has the advantage of greater recency. This could be a sign the race is tightening; the Liberals still enjoy a commanding lead, but perhaps no longer a precedent-smashing one.

Abacus provided more detail than CRA about the questions it asked and the methodology it used.

Finally, Sun News announced that it had, commissioned Abacus to survey up to another 250 voters tonight, Saturday, and Sunday. That compares to the 90 or so voters CRA has been interviewing for the Chronicle-Herald. Unless the Herald ups CRA’s game, Sun-Abacus will have the authoritative poll going forward to election day.

At this point, I would summarize the difference between the two polls as a Liberal blowout (CRA) vs. a solid Liberal majority (Abacus). This could mean that some races thought to be tight in the early going—like Sydney-Whitney Pier, and Victoria-The Lakes—could still go either way. Or not. Because of the small numbers and the two companies’ different approaches to “likely” and “leaning” voters, the differences could be more apparent than real.

That’s enough bet-hedging from me for one night.

* I am pleased to note that the Chronicle-Herald, in a public spirited gesture, appears to be keeping much of its election related material, including the daily poll, outside its new paywall.


Don Mills plumbs the causes of the probable NDP collapse

Assume for the moment that the Corporate Research Associates poll showing Stephen McNeil’s Liberals with a 30-point lead is accurate (which I assume it is), and assume McNeil holds that lead until Tuesday (which remains to be seen).

The next question is, how the heck did this happen?

The NDP made serious mistakes (see here and here) but they did not run a horrible government (see here and here). Not everyone will agree, but there is a reasonable case that Dexter deserves another term, something of a tradition in Nova Scotia, as many have pointed out.

Don-MillsThe curious thing is that one detects little passion in this election at all, certainly no mad fury to throw the bastards out. Election signs are few, and in my circle at least, hardly anyone is talking about the election. The public exudes little warmth for McNeil, and his slender platform is pockmarked by a handful of dreadful ideas.

Without the CRA poll, my main expectation would be a record low turnout. These are not the impressions one expects a week before an electoral tidal wave strikes.

CRA Chairman and CEO Don Mills thinks the NDP’s problems are rooted in the fact it took power just as the worst recession in 80 years settled in to stay.

“These have been the most difficult economic times in most people’s living memory here,” he told Contrarian in a telephone interview Tuesday.

CRA tracks the economic wellbeing of Nova Scotians every quarter, and in each of the last four years, 40 percent of its respondents reported no increase in pay. Inflation, meanwhile, has eaten away at the value of their pay checks. In effect, a large cohort of Nova Scotia’s middle class has seen a 10 percent drop in buying power during the four-plus years of the Dexter Government. This may be no fault of the NDP, but that’s of little consequence to fed-up voters.

“Virtually every household is worse off,” said Mills. “I think they would have very little patience with any government—and it doesn’t matter what government it is, by the way. I think this is why electricity rates are such a sore point.”

To that economic malaise, add the voters’ dashed hopes that the NDP would be different. Take a closer look at Mills’s poll. Dexter was elected in June 2009 with 45 percent of the vote. Within six weeks, his government’s popularity soared to 60 percent. Voters—even those who didn’t vote NDP—must have said to themselves, “Let’s give these new guys a chance. Maybe they really will turn out to be different.”

The new government did little in its first six months, and then the expense scandal hit, with Dexter smack in the middle of it. In retrospect, I believe this created a sense of disillusionment that voters have been nursing ever since.

That Don Mills poll

Don Mills sounds nervous.

Nova Scotia’s best known pollster has been conducting a rolling poll for the Halifax Chronicle-Herald, and over the last week, his numbers have pointed to an historic rout. For the last five days, he has shown Stephen McNeil’s Liberals holding steady between 55 and 57 percent of decided voters—enough to propel him to a lopsided majority.

“We’re under a lot of scrutiny here,” he told Contrarian.

Here’s the latest edition, published Tuesday morning:

To understand how unusual such an outcome would be, I looked at every Nova Scotia election since 1960. Over those 15 provincial votes:

  • The winning party got more than 55 percent of the vote only once: Robert Stanfield’s PCs took 56.2 percent in 1963, in what was essentially a two-party race.
  • Only three times has the winning party won more than 50 percent of the vote: Stanfield did it in ’63 and ’67; John Buchanan got 50.6 percent in 1984.
  • In all three of these contests, the opposition party or parties were crushed. The Liberals won just four seats in 1963 and just six in ’67. When Buchanan got just 50 percent of the vote in the 1984 election, the opposition parties shared 10 seats: six for the Liberals, three for the NDP, and one for independent Paul MacEwan.

John Savage’s Liberals won 49.7 percent of the vote in 1993, and took 40 seats to the PCs’ nine and the NDP’s 3. All of these number reflect the reality that, in a first-past-the-post election system, when one party’s vote percentage goes above 50 percent, the number of seats it can win goes up exponentially.

Here’s the historical record, minus third parties and independents:


[Yellow highlighter indicates a minority government. The complete spreadsheet, which includes third parties and independents, can be downloaded here.]

If CRA’s numbers are anywhere near correct, and I expect they are, then every assumption about this election goes out the window. Seats thought to be in play will fall easily to the Liberals. Some seats assumed to be safe will fall to the Liberals. The premier’s seat, Jamie Baillie’s, and those of prominent cabinet ministers, could be at risk.

Mills views this cautiously. Much could change in a week, but if the current spread holds, he expects both opposition parties will have high single digit seat tallies, “closer to 10 than to zero.” If the total exceeds 55 percent, Mills may be understating this.

“To have [a 30-point spread] in a three-party election is extraordinary,” Mills said. “I’m stunned by it. It’s very hard to explain.”

On which, more later.

Has Don Mills put CRA in a conflict of interest?

A Contrarian reader asks:

Does it not seem to you that there is a major conflict of interest in the Savage-for-Mayor camp? [, the online news service]  lists Don Mills as one of Savage’s top supporters. Since  Mills operates Corporate Research Associates, the major polling firm in the province — one that just recently reported Savage with a big lead — why would one trust anything CRA has to say on the race?

A fair question, and we put it to Mills, who replied:

Corporate Research Associates has been since its inception a non-partisan polling company. It is one of the reasons our polling is so respected by the media. Every time we publish poll results , our reputation is on the line. We adhere to the highest standards in the conduct of our business. We attempt each and every time to ask questions in a fair ad unbiased manner. We publish the questions used in order for the public to judge the quality of our work. I have worked hard all my career to ensure the integrity of our work. Our record speaks for itself in that regard.

I have lots of issues with political polls and the way they are presented, but in this case, I’m inclined to accept Mills’s assurances. He runs the pre-eminent blue-chip polling firm in Atlantic Canada. It’s his bread and butter. He would be foolish indeed to squander that standing by rigging polls in a municipal election whose outcome doesn’t seem all that hard to predict, with or without polling data. Reinforcing this view is the fact Mills made no effort to hide his support for Savage.

Cape Breton seats in play?

The old saw says they should let Cape Bretoners vote the next day, so we get it right. Until now, there has been a widespread assumption that of the 10 nine Cape Breton seats, only Victoria-The Lakes, held by PC Keith Bain, is in play. Today’s CRA poll gives no reason to challenge that assumption. CRA’s Don Mills says the NDP are merely holding their own on the island, where they currently hold but two seats.

This bears watching. If Cape Bretoners get a sniff of a majority NDP government, things could change quickly.

An NDP majority in the making

The latest poll from Don Mills of Corporate Research Associates shows the NDP at 44 percent. More importantly, it shows them in first place in the rural mainland. results-60-06

Some will say the NDP vote is highly concentrated in metro, where they will “waste” votes by winning with unnecessarily huge majorities. Elections are won by seat totals, not vote totals.

Still, 44 percent is well into majority territory. In the last 14 Nova Scotia elections, no party has ever won more than 40 percent of the vote and failed to win a majority. John Hamm won a majority in 1999 with 39% of the vote.

The only partial exception is the 1960 election, when the Liberals won 46% of the vote and exactly half the seats in the legislature. However, one riding was controverted, and another ailing member failed to take his seat, so the Liberals ruled unimpeded until the 1974 election. In any case, the NDP was a small regional rump in Cape Breton at that time, and its vote was highly efficient in winning two seats.

In contrarian‘s view, polls taken between elections mean little, because the public isn’t thinking about casting ballots. Once the writ is dropped, voters turn their mind to the question, and that’s when polls become meaningful. Some dramatic bombshell could still change things, but the strong trend in the NDP’s direction since the election was called shows that voters have finally decided to give Dexter’s New Democrats a chance to govern, for better or for worse (for richer or for poorer, in sickness and in health, etc., etc.)