The Vancouver Scrum

SES Research has released its final tracking poll of Indecision Canada; once again, they're showing 34% Liberal, 30% Conservative, 20% NDP, 12% Bloc, 3% Green -- the trend has been towards these numbers for the past week. SES hasn't hazarded a seat projection based on those numbers. That much is up to you, but Andrew Coyne would like you to place your bets, inspired as he was by the most recent Ipsos poll and projection, showing the Tories and Conservatives in a statistical tie, but with the Conservatives getting the advantage on seats.

Looking at the comments on Coyne's Conservative Echo Chamber (term scalped from Jim Elve), most are based on Conservative partisanship and hope, but one tidbit from a Quebec poster caught my eye.

The anglo pollsters have put some unrealistic figures for the Bloc here (the most ludicrous being the 64-68 figure I've read in the most recent Ipsos-Reid). Nobody here seriously believes this crap coming from Toronto. Let me explain why:

Since there is a correlation between the linguistic origin of voters and their voting patterns in Quebec. Historically, the Bloc never wins a riding if the proportion of non-francophones is greater than 25% (17 of Quebec's 75 are in this situation).

The Bloc got their best performance in 1993 when they've elected 54 MPs with 49.7% of the vote. Most polls put the Bloc at 50%. Therefore, the Bloc won't elect more than 58 MPs, and they will lose some predominantly francophone ones (Beauce and Brome-Missisquoi are prime candidates, unless the Libs and Tories really split the vote).

Most posters are convinced that the Conservatives have a more efficient Ontario vote distribution than do the Liberals, based on the fact that the Liberals win Toronto ridings by massive margins, while the old Reform/Alliance party was close-but-no-cigar in the towns and rural areas. Of course, the Conservatives have their own inefficiencies, especially in places like Alberta -- the Conservatives can go from 50% to 70% of the popular vote in Alberta without winning more than an extra seat (and they can only win two more in Alberta anyway.) They also point to fondly-held beliefs in order to boost their Conservative projections, such as that people won't tell pollsters that they'd actually vote for the "evil" Conservatives, that Conservative supporters are too busy and successful to bother with telephone surveyors, that Conservative supporters are angrier, more motivated, and more likely to vote (but wouldn't that have been true for the Alliance in 2000?), and that the Conservatives' greater appeal to older voters will help Conservative support (shurely the pollsters also account for that when making projections?), and so on.

Meanwhile, Ipsos-Reid has put out another BC-wide poll, showing the Conservatives leading the NDP 38%-26%, with the Liberals getting the support of 25%; the Green vote has contracted down to 7% after being as high as 13% earlier in the campaign. (n=800, +/- 3.5%, 19/20) Based on those numbers, Ipsos predicts 24-28 Conservative seats in BC, with 4-8 Liberals and 2-6 New Democrats. I've previously disputed Ipsos' projections -- the Conservatives have a large surplus of support in some areas such as the Fraser Valley and much of the Interior. That extra support pads the popular vote, wins no more seats, and means that Conservative support is actually lower in other areas than the top-line numbers indicate. By contrast, the NDP's support never reaches the 70-plus percent pull of the Conservatives in Chilliwack; most of their wins, regardless of the year, are of the "just enough" variety. Now, the Reform Party took 24 of 32 seats in 1993 while pulling only 36.4% of the popular vote, although it won many of those seats by only a couple of percentage points, and that was in a year that the NDP ran a disastrously bad campaign. A resurgent NDP may pull areas it lost to Reform in 1993 when the left-of-right vote was split.

But wait, there's more!

Two parties have inflicted internal polls on an unsuspecting British Columbia public this week. First it was the NDP. Trying to counteract the Liberals' "a vote for the NDP is a vote for Stephen Harper" message, the Dippers put out some numbers from house pollster Strategic Communications showing the NDP in the lead in Vancouver Centre, Vancouver-Kingsway, Burnaby-Douglas, and in contention in Nanaimo-Cowichan -- with the Liberals (including star candidates David Emerson and Bill Cunningham) trailing the NDP badly in all 4 ridings.

The Liberals have done the same in a couple of tight races. From Public Eye:

The poll, which was conducted by Synovate (the market research arm of the Aegis Group plc) between June 23 and 24, showed the Liberals at 23.7 percent in Saanich-Gulf Isalnds, the Conservatives at 21.3, the New Democrats at 20.7 and the Greens at 12.3. Meanwhile, in North Vancouver, the Conservatives are at 29.7 percent, the Liberals at 28, the New Democrats at 15 and the Greens at 11.7

Being as these are internal party polls, selectively released to the public, the usual caveats apply.