April 18th: new AR and Leger polls showing NDP surge

We've got two new (large) polls, from Angus-Reid and Leger. Both show a big lead for the Tories, including in Ontario. But they mostly show a surge from the NDP which is actually tied with the Liberals (!?) at 25% in the AR poll, and a couple of points behind in Leger. I'll write later today if there is a chance that the NDP could finish second (but I kept this post simulation-less for those of you who hate this kind of posts).

Most of the numbers from these two polls are confirmed by the latest daily tracking from Nanos, except that as usual, the NDP is a little bit lower in Nanos. However, as I was writing a couple of days ago, the NDP is definitely up in Quebec. I still don't believe that a party would move from 12.5% to 24% and being second, AND not gain more than 1-2 seats. As I've explained, this is such a big change that the model is probably wrong for that. I'll spend the day trying to see if I can fix that, using another methodology.

For the Tories, the national votes intentions haven't changed much since the beginning of the campaign. They are always at 38-39%. The only thing is that they move in Atlantic and in Ontario. In the latter, they seem to have regained the lead the enjoyed previously. And the Liberals seem to be falling down quite quickly if you look at Nanos.

For the Grits, if they end up giving a majority to Harper, I think we'll have to wonder if attacking the NDP and the left-wing voters was the right decision. I mean, Ignatieff's campaign and platform are clearly aimed at these voters, not some soft right-wing currently voting Conservatives. While it was effective at first, a bad debate performance (and a strong one from Layton) have reversed the trend. Also, I wonder if the NDP started attacking the Liberals because of that, i.e: some kind of retialiation. We'll wait and see, but I'm not completely sold on the Liberals strategy so far.

The NDP at 25% and tied with the Liberals? Crazy, although we've seen that in the past (maybe not as consistently in so many polls though). What we've never seen though is the NDP so high in Quebec. One bad news for this party is that a lot of its growth seems to come from the Prairies and not from Ontario. Gains are thus much more limited. As I've said, I'll post later today some numbers for this party only.

By the way, using only the last three polls (AR, Leger and Nanos), as you can see I would get 40 seats for the NDP and that is with the probably underestimated results from Quebec. The "low" numbers for the Conservatives are exclusively because of the Atlantic (for this region, I used an average of other polls as well as the AR one put the Tories so low, it was giving weird results). It is funny how this region is actually a key for a Conservative majority. Indeed, if Harper doesn't win 15 seats+ there, then he HAS to break the GTA in order to secure enough gains. Because he won't get them in Quebec and he's pretty much maxed out in the West.

If you are wondering why the Bloc is so high, well, this is for two reason. 1) the NDP seats are probably underestimating as we've never seen that in the past. 2) The NDP is stealing votes from the Liberals and Conservatives as well, therefore giving the Bloc back some seats.

That's all for now. Come back later for the NDP-focus post where I'll look if the NDP can finish second and if the rise of the NDP really helps getting a Tories majority.