The survey, conducted by Environics between Friday and Tuesday, found that 38 per cent of Canadians would vote for the Conservative party if an election were held immediately.
By comparison, 28 per cent would vote for the Liberal party, 19 for the NDP, eight for the Bloc Québécois and seven for the Green party.
The size of this poll is a full 2.5 times greater than a normal national poll, so the regional MOE's are much lower than most.
Those numbers are fairly standard, relative to others, and this is the biggest sample apart from CROP.
The results for British Columbia show the Conservatives ahead, with the NDP and Liberals a few points behind. Again, that seems reasonable, relative to others.
Same for the "prairie" results, and to a less degree the Atlantic Canada numbers. Where my eyes get wide is the Ontario numbers, and it these results that give the supposed large Conservative lead:
Last week's SC poll found similar results, Conservatives up, although only by 6%. So, calling the SC poll an outlier becomes more problematic, it now finds a similar result. However, these numbers are in complete contrast to three other polls- Decima, Libs up 11%(a full 18% in the last week of polling), NANOS Libs up 13%, Ipsos Libs up 12%. In addition, the NANOS and Decima polls both showed the Conservatives at their LOWEST percentage in Ontario since the last election. For that reason, it is very hard to reconcile those results with this apparent surge in Ontario. Comparing NANOS and Environics, you see a full 22% difference, absolutely staggering, particularly for a region that has low margin of error.
There is one other item in this poll that raises my eyebrows. No surprise that the Conservatives lead on a host of issues, we have seen this before. What is telling, Environics actually puts the Libs and Cons in a statistical tie on who is best suited to deal with the environment. I haven't seen one poll that shows this dynamic, in fact most give the Liberals a healthy lead. Is this anomaly a signal that something is wrong with the respondent demographics, a clue that the poll has representative issues? I'll buy Harper smoking Dion on leadership, the economy, leading on Afghanistan, etc, but a tie on the environment is suspect at best.
Who to believe? First, my own sense of Ontario. I just don't buy the Cons ahead here, all my instincts see real problems. Even if you take the supposed battleground in Guelph, the Conservatives were poised for a very disappointing result, and third was a distinct possibility. If the blue wave was cresting, one would think we would get a hint in a key riding, instead by all appearance the Cons were set to go belly up. Aside from that, it's hard to point to any reasons to support a strong move to the Conservatives, the opposite seems more likely.
So, in the final analysis, pick your poll, or set of polls. Myself, I'm sticking with track record, and given the laser-like accuracy of the past, I feel comfortable believing the data that suggests the Liberals up in Ontario, by a good margin. Call it spin, I call it seeing conflicting signals, falling back on the guys with the accuracy pedigree. Seems entirely logical from here.