Federal Election Polls

[Derek 2.0]

who is Stephen McMurtry and what did you say about internal polling before?

That women wearing face coverings and not keeping terrorists in Canada could translate into a bigger majority, in part thanks to Quebec, then the party has been expecting.

[angrypenguin]

They are behind......EKOS should be releasing a poll, later today or tomorrow morning, confirming/aligning with today's FORUM poll for the Star, both of which are under representing, to a degree, our internal polling.

Agreed. I think the EKOS poll will be released today.

Any idea why the spread in the Nanos poll is consistently smaller than FORUM/EKOS?

[Derek 2.0]

Agreed. I think the EKOS poll will be released today.

Any idea why the spread in the Nanos poll is consistently smaller than FORUM/EKOS?

No idea, historically, Nanos polls the Tories 2-3 points behind the actual numbers.

[angrypenguin]

No idea, historically, Nanos polls the Tories 2-3 points behind the actual numbers.

I noticed this wrt the 2011 #'s. Nanos was definitely a few points behind. EKOS was closer, but then again the data I can glean off of EKOS is much better. (age, education etc). I don't get any of that with Nanos. I use the Nanos data to track voting intentions by regions though. And the blues are definitely on their way up in Quebec

[ScottM]

Any idea why the spread in the Nanos poll is consistently smaller than FORUM/EKOS?

The Liberals are really the only party the two are widely separated on. The gaps between the numbers on everyone else are within the margins of error for the polls. Forum is at odds with everyone else on Liberal numbers, though, so I think the question is who was a little high, the Tories or the NDP? If it's the Tories, their numbers move closer to the Nanos. If it's the NDP, their slide has continued. That said, I think it worth mentioning that a look at Wikipedia polling articles shows that the governing party has been underestimated by polling averages in the last four campaigns. If I had to make a gut instinct guess, I'd say the NDP is too high, but again, that's a pure guess.

[ScottM]

And if you change it back to average, Tory high average, NDP high average etc?

I gave only the Liberal numbers because they were central to the discussion at hand. But here's the link to not only that map, but individual riding projections all in one. It's updated every day, so things may be a little different from yesterday: http://www.threehundredeight.com/p/canada.html The most recent polling was good for the Tories, bad for the Grits, so there may be some noticeable difference.

FWIW, here's the polling tracker too: http://www.cbc.ca/news2/interactives/poll-tracker/2015/index.html The Conservative high number of seats is up to 159, so the chance of a majority is creeping closer and closer to falling within the 75% range between the high and low estimates. Again, it's only as good as the polling that goes in, but it's interesting, nonetheless.

[cybercoma]

Forum is at odds with everyone else on Liberal numbers

Not anymore. The numbers are corroborated by Angus Reid and Derek claims EKOS, but I've yet to see that one published.

[angrypenguin]

Not anymore. The numbers are corroborated by Angus Reid and Derek claims EKOS, but I've yet to see that one published.

You mean the EKOS ones from last week?

It's on their front page.

Everyone went...EKOS numbers have to be wrong, now the FORUM numbers also indicate a clear Conservative lead. Nanos #'s also indicate this, as we all know the Conservative #'s for Nanos are usually behind by a few points.

Edited October 1, 2015 by angrypenguin

[cybercoma]

You mean the EKOS ones from last week?

It's on their front page.

Everyone went...EKOS numbers have to be wrong, now the FORUM numbers also indicate a clear Conservative lead. Nanos #'s also indicate this, as we all know the Conservative #'s for Nanos are usually behind by a few points.

Nanos numbers do not indicate that. Nanos numbers has the Conservatives and Liberals in a statistical tie over the last 4 days.

[angrypenguin]

Nanos numbers do not indicate that. Nanos numbers has the Conservatives and Liberals in a statistical tie over the last 4 days.

When I looked at the Nanos #'s for the 2011 election and the 2008 election, the Conservative % have always been below the others. If I account for the 1.1% gap as per last night's #'s, and I add say 3% to the Conservative #, it would bring that in line with EKOS and Forum for who's "clearly" (I'm using this word lightly, but the newspapers aren't) winning.

That's what I was getting at. But I see your point.

Edited October 1, 2015 by angrypenguin

[ScottM]

Not anymore. The numbers are corroborated by Angus Reid and Derek claims EKOS, but I've yet to see that one published.

Indeed. I hadn't seen that Angus Reid numbers were out until I checked after reading your post. The Conservatives have opened up a significant lead in Ontario in that poll. I notice that they also have the Bloc back to 2011 levels of support. I think Quebec will really be a mess come the 19th. Anyway, I wonder if we'll see Nanos move in the same direction in a couple of days.

[angrypenguin]

This greatly saddens me. I am reading the Angus #'s right now and I read this. :(

• Three-in-ten (31%) say they’ll make up their mind in the last week of the campaign. This equates to one-in-five (18%) eligible voters overall
• Another nearly two-thirds (32%) won’t finally make up their minds until just before or on voting day. This equates to another one-fifth (19%) of the overall electorate, as seen in the graph below:

[cybercoma]

When I looked at the Nanos #'s for the 2011 election and the 2008 election, the Conservative % have always been below the others. If I account for the 1.1% gap as per last night's #'s, and I add say 3% to the Conservative #, it would bring that in line with EKOS and Forum for who's "clearly" (I'm using this word lightly, but the newspapers aren't) winning.

That's what I was getting at. But I see your point.

The Nanos numbers for the 2011 election were within the margin of error for all of the parties. You can't just make up numbers and assume they're true.

As for those EKOS numbers, there've been 13 polls since then that have all had the Liberals above 30. Of the last 10 polls, 3 of them have the Liberals below 30 with only the most recent 2 outside of the margin of error.

[angrypenguin]

The Nanos numbers for the 2011 election were within the margin of error for all of the parties. You can't just make up numbers and assume they're true.

As for those EKOS numbers, there've been 13 polls since then that have all had the Liberals above 30. Of the last 10 polls, 3 of them have the Liberals below 30 with only the most recent 2 outside of the margin of error.

Yes, while it's the case that they are within the margin of errors, the only thing I'm looking at (and care for) is who is ahead. Each poll has their own methodology, and if I stare at Forum/Ekos/Angus, they all indicate a clear lead for the Conservatives. The odd ball? Nanos. They're all over the place when it comes to NDP/Liberal #'s, but I don't really care. So for me, the question I ask myself every morning is whether or not the Conservatives are ahead. And based on now 4 polls, they are.

[cybercoma]

This greatly saddens me. I am reading the Angus #'s right now and I read this. :(

• Three-in-ten (31%) say they’ll make up their mind in the last week of the campaign. This equates to one-in-five (18%) eligible voters overall
• Another nearly two-thirds (32%) won’t finally make up their minds until just before or on voting day. This equates to another one-fifth (19%) of the overall electorate, as seen in the graph below:

Yeah, the polls are showing a lot of indecision at the moment. I think that's because there is 3-way movement at the moment. Red-tories who were clamouring for change during the Duffy trial, recession talk, and Tory candidates pissing in people's coffee mugs, are moving back to the Tories over the anti-Muslim rhetoric we're seeing from the Crosby-led campaign now. The soft NDP support, who were hesitant to vote ABC are now switching to the Liberals on the threat of another Tory government. There's too much uncertainty right now and people are considering voting in ways that they didn't initially intend.

[angrypenguin]

Yeah, the polls are showing a lot of indecision at the moment. I think that's because there is 3-way movement at the moment. Red-tories who were clamouring for change during the Duffy trial, recession talk, and Tory candidates pissing in people's coffee mugs, are moving back to the Tories over the anti-Muslim rhetoric we're seeing from the Crosby-led campaign now. The soft NDP support, who were hesitant to vote ABC are now switching to the Liberals on the threat of another Tory government. There's too much uncertainty right now and people are considering voting in ways that they didn't initially intend.

To be fair, every party has their scandal. Yesterday a Liberal candidate resigned. I'm not too sure about the "switching" though. The Angus poll anyways indicates the Liberals and the NDP are tied.

I wish people would do their research and make up their mind. (emphasis on "do their research"). The "game day" decision....sigh, what a waste of a vote.

[angrypenguin]

How anyone can vote Liberal is beyond me.

This pretty much sums it up.

"Ontario now qualifies for equalization payments, confirming its shift from “have” to “have-not” status"

http://business.financialpost.com/fp-comment/wynnes-losing-performance-how-ontarios-growth-killing-policies-are-sinking-the-economy

[Derek 2.0]

I gave only the Liberal numbers

---SNIP---

but it's interesting, nonetheless.

Thanks, I've seen both sites, for ease of use, I preferred the old threehundredeight site, but I don't blame the sites creator for getting in bed with CBC.

Edited October 1, 2015 by Charles Anthony
[---SNIP---]

[cybercoma]

Yes, while it's the case that they are within the margin of errors, the only thing I'm looking at (and care for) is who is ahead. Each poll has their own methodology, and if I stare at Forum/Ekos/Angus, they all indicate a clear lead for the Conservatives. The odd ball? Nanos. They're all over the place when it comes to NDP/Liberal #'s, but I don't really care. So for me, the question I ask myself every morning is whether or not the Conservatives are ahead. And based on now 4 polls, they are.

The margin of error is important because it let's you know that 5% of the time the results could fall outside that +/- range. More importantly, you can't be sure from the reports where the "true" number lies other than to say there's a 95% probability of it falling between +/- 3% (for most surveys). When two parties are within that margin of error, there's no way of knowing which one is ahead. One random sample may have one party ahead by a few percent, the next may have the other. That's exactly what has happened with Nanos over the last week or so. And that's why we say they're "statistically tied" regardless of what the headlines read or what that particular sample showed. In the overall population, the best we can say is that they're tied. There is no significant difference between the two parties.

Also, the polling methods may be slightly different (IVR, online panels, phone, etc), but their methodologies are the same. It's important to note that margins of error are not supposed to be reported for online surveys (Angus and Abacus). The sample that they draw is not random from the population. Therefore, they need to do some "monkeying" with the numbers by weighting them according to census demographics. The weighted versions are reported and then they give a margin of error for a "similar size survey" because their association's code of ethics says they are not to report a margin of error for a sample that is not random.

Nanos you say is all over the place with Liberal and NDP numbers, but that's also not the case. Let's look at the last 5 polls for Nanos, going from earliest to most recent.

LPC: 31.6, 31.4, 32.2, 31.7

NDP: 26.9, 27.0, 26.3, 26.1

Nanos is quite consistent and their numbers are in line with what other polling firms are reporting. Over the last week, the Liberals have been between 31 and 32.5, while the NDP has been between 26 and 27 (except the current Forum poll which puts them at 28). So they're not "all over the place" as you say.

To your last point as to whether or not the Conservatives are ahead, it depends on the day. These are snapshots. They're most likely ahead today, given that they're outside the margin of error in the Angus Reid and Forum polls. The EKOS poll is out of date at this point and was not supported by numerous polls that followed it. It was almost certainly an anomaly at that time and it's too old to say anything about where the parties stand now.

[Derek 2.0]

Not anymore. The numbers are corroborated by Angus Reid and Derek claims EKOS, but I've yet to see that one published.

I was unaware Angus Reid was releasing a poll today, from looking at it, it appears its findings should align (perhaps a degree under) what I'm to believe EKOS (soon to be released) numbers should show, namely:

And here in BC:

Both confirming my (mocked) assertion from last night, in which I stated the Liberals are not a major factor in BC, and that if the NDP bled support to the Liberals, net result, Tory wins (Perhaps I should have kept my mouth shut ooooops)

[ScottM]

Thanks, I've seen both sites, for ease of use, I preferred the old threehundredeight site, but I don't blame the sites creator for getting in bed with CBC.

Yeah, I miss the percentage ranges as well, but the way I look at it, if it brings in more funds/audience, go for it. I love the work he does, and if it helps him out, I'm all for it.

[Keepitsimple]

It seems that every time the polls start to favour the Conservatives, there a lull in posting..........

[eyeball]

...and an increase in smugness.

[angrypenguin]

Why didnt EKOS release data today?

GRR.

Edited October 2, 2015 by angrypenguin

[angrypenguin]

I was just looking at poll #'s. In 2008, Angus Reid was the closest to the results (T-1 day). In 2011, Nanos was the closest.

Weird. (in some ways)