Federal Election Polling Thread

[Frequitude]

Originally Posted by rubecube

These are the Conservatives "would consider" numbers across the various regions (ATL, QC, ON, Prairies, BC)

32.6 28.3 44.5 52.3 33.3

Compared to the Liberals:

67.4 39.7 58.5 45.1 55.2

And NDP:

44.0 41.8 44.3 38.8 46.8

And here is the "would not consider numbers:

CPC

63.9 57.7 52.6 43.1 61.5

LIB

27.7 49.1 36.6 47.6 41.2

NDP

49.5 39.7 51.1 55.1 46.8

It really looks to me that CPC's numbers are being inflated by their support in the prairies, and that's likely also the case with the "preferred PM" numbers.

Interesting numbers. Thanks rube. Fair to say that the CPC has the least downside and Liberals the most upside (though more downside risk).

[CorsiHockeyLeague]

61.5% of BC "would not consider" voting conservative? I find that difficult to believe.

[rubecube]

Originally Posted by CorsiHockeyLeague

61.5% of BC "would not consider" voting conservative? I find that difficult to believe.

Why? Vancouver and Victoria are pretty solidly anti-Harper and they make up the majority of the province.

[crazy_eoj]

Originally Posted by MarchHare

Nanos was pretty accurate overall in the last few federal electiond, predicting support of all five major parties within the MOE of the final pre-election poll.

Final Nanos poll - Actual vote share

2011 Election
CPC 38.7 - 39.6
NDP 30.5 - 30.6
Lib 20.9 - 18.9
Bloc 5.0 - 6.1
Grn 3.7 - 3.9

Except you are not using their own methodology. On Election Sunday they were close but on Saturday they polled BOTH the CPC and the NDP at 33.8%....? Nanos' Final numbers using their three day averages were:

2011 Election
CPC 36.9 - 39.6
NDP 31.6 - 30.6
Lib 21.3 - 18.9
Bloc 5.2 - 6.1
Grn 3.7 - 3.9

It seems they really blew the CPC and Liberal numbers. As they likely are doing again.

[crazy_eoj]

Here's what Nanos' polling looks like on a daily trend:

[burn_this_city]

Thats some garbage data

[flylock shox]

All those blue and red diamonds suggest a lot of people flip flopping between the Libs and Cons.

Then again, there's a lot of red and orange diamonds too.

[edslunch]

Originally Posted by flylock shox

All those blue and red diamonds suggest a lot of people flip flopping between the Libs and Cons.

Then again, there's a lot of red and orange diamonds too.

Actually no, it just shows they have a small sample size - only 400 country wide. That's why they don't publish the daily numbers only the rolling average. My guess is if you took one of the bigger polls that is run over several days , each day would also show a lot of variability due to sample size

[Regorium]

Originally Posted by crazy_eoj

Here's what Nanos' polling looks like on a daily trend:

I admit I involuntarily laughed out loud in my cubicle when I saw this.

3 day moving averages should not have such ridiculous spikes. I don't really get it unless they're basically rotating their methodologies every 3 days.

[Parallex]

Originally Posted by Regorium

3 day moving averages should not have such ridiculous spikes.

That's not the three day rolling average that's the dailies (October 5, October 4, October 3 rather then October 3-5, October 2-4, October 1-3).

[mrkajz44]

That thing looks more like a seismograph than a graph attempting to show trends.

I don't think the "small sample size" holds up as much, so long as their selection process is truely random (and I'd have to think polling companies should have that taken care of pretty well). I really think people are flip flopping on how they feel. I know I'm still not decided yet myself. If I had to answer at the end of each day who I was voting for, it would definitely change back and forth every few days.

[Drak]

Looks like a lie detector test.

[Ashartus]

Today Poll Tracker is actually showing the Liberals with the most predicted seats (127 vs 124 for Conservatives) as well as the highest level of support - I think that's the first time I've seen that since the election was called.

[starseed]

Forum's latest poll has the Liberals jumping 8 points from their poll last week, Conservatives down 3, NDP down 2.

CBC seat projection model is now showing a Liberal minority government based on a number of razor thin victories. Tomorrow is the start of the advanced polling weekend.

http://www.cbc.ca/news2/interactives...015/index.html

[CaptainCrunch]

I don't know, it seems like polls just don't mean anything anymore, and I'm tending to discount them completely.

We saw what happened in Alberta, and BC and Ontario with the polls.

In fact, I believe that positive poll results work against the leading party because people walk into the voting booth and decide that they don't want the party that they were going to vote for in government.

[starseed]

I think provincial politics is very different from federal politics. You have a large 'change' block waiting to see which of the change parties to vote for. Being higher in the polls is very important for those parties before election day. The latest abacus data shows that the change block is seeing the Liberals as more likely to win, which is why we have seen soft NDP supporters move to the Liberals in the last week.

[Ashartus]

One of the key questions is how much bias is in the polls due to survey methods and who that favours. My gut feeling is that most polls will be biased towards older voters, who may be more likely to vote Conservative. However that is also the population historically most likely to actually get out and vote.

If young voters are engaged enough to vote though, or if the "ABC" vote starts to shift more towards the Liberals as a result of the polls, then that could be bad news for the Conservatives.

[FlamesAddiction]

Originally Posted by Drak

Looks like a lie detector test.

Heh, in a strange way, it kind of is one.

[Drak]

Looks like the Liberals have finally pulled ahead

http://www.cbc.ca/news2/interactives...015/index.html

Picked up in Quebec and Ontario. NDP transfers coming in hot and heavy.

[rubecube]

Originally Posted by CaptainCrunch

I don't know, it seems like polls just don't mean anything anymore, and I'm tending to discount them completely.

We saw what happened in Alberta, and BC and Ontario with the polls.

In fact, I believe that positive poll results work against the leading party because people walk into the voting booth and decide that they don't want the party that they were going to vote for in government.

Well the pollsters got this year's election in Alberta correct, and Nanos has been pretty accurate in the last couple of federal elections.