Ontario 2014 (June 12th)

[Linus Van Pelt]

What are you all actually doing, mathematically, to get these projections?

[Linus Van Pelt]

What are you all actually doing, mathematically, to get these projections?

I use a ratio projection on a regional basis, but altering results in ridings that have had by-elections (taking into account polling averages around the by-elections). Unlike Brendan, I will also be doing some riding level tweaks to get more believable results. Tweaks will be based on math as well though.

But what is the actual formula?

Or are you deliberately keeping the details secret, so that it can't be replicated, like a for-profit pollster?

[Linus Van Pelt]

I see; thanks.

[Linus Van Pelt]

What are you all actually doing, mathematically, to get these projections?

I use a ratio projection on a regional basis, but altering results in ridings that have had by-elections (taking into account polling averages around the by-elections). Unlike Brendan, I will also be doing some riding level tweaks to get more believable results. Tweaks will be based on math as well though.

But what is the actual formula?

Or are you deliberately keeping the details secret, so that it can't be replicated, like a for-profit pollster?

No, I should be transparent for conflict of interest reasons (as I may or may not have access to internal polls that I cannot report on in my model)

In the ratio model, if a party doubles their support in a region, then they double their support in each of the ridings. It definitely has its flaws (which is why the Greens can often be shown to "win" seats), but I have seen it used in the polling industry (not by Forum Research obviously). I first started using it when I was running a Canadian election game in the Individual politics sub forum here.

Wait, sorry: upon further reflection, I actually still don't get it.

Suppose a party had 20% at the last election in a region and 60% in one riding in the region. Then according to the bold, if the party's support doubles to 40% in the region, they double in the riding to 120%? That can't be right.

[Linus Van Pelt]

Is it correct to assume that although pollsters are weighting by the regions that they publish, they are not further weighting within these regions? So a southwest sample could have, say, Windsor overrepresented one time and then rural areas overrepresented the next time?

[Linus Van Pelt]

Not all the relevant issues here are straightforwardly to do with left or right.

[Linus Van Pelt]

As I have stated before, Horwath's opposition to funding new transit with gas taxes or road tolls does not play well in much of Toronto. The Liberals' strength in the city is not purely about strategic voting.

[Linus Van Pelt]

Forum Research has got nothing on Today's Chanakya.

[Linus Van Pelt]

I can't remember ever seeing pollsters diverge like this.

[Linus Van Pelt]

Star looks at Brampton-Springdale.

Is the NDP actually competitive in Brampton-Springdale, as the article suggests? I can believe it, given their advances in the riding to the east, but on the other hand, I can also see some reporter just giving equal space to all three candidates without much local knowledge. This is the sort of thing that would not be picked up by regional polling.

[Linus Van Pelt]

How do they get people for these online panels?

[Linus Van Pelt]

So, after a bit of reading the fine print on the web sites of Ipsos and Abacus, I think I understand the answer: they are online panels of people who sign up to fill out surveys, mostly having to do with product testing for the purpose of corporate marketing research, in exchange for a sort of rewards/points program. Abacus uses a panel from a company called Research Now that you cannot voluntarily sign up for; they have to invite you, presumably having obtained your e-mail from online retailers. Ipsos, as you say Earl, can be voluntarily joined. But the advertising to join the panel suggests that they expect most people to be in it mainly for the rewards program. It's not just a bunch of political nerds who want to fill out polls for fun.

Now, this might be an explanation for what's going on, if, even after all the other demographic re-weighting, people who join these rewards programs are more amenable than the general population to a sort of household-budget-oriented retail style of politics. Or, as Sam Spade would say, it might not. We'll see.

[Linus Van Pelt]

When you were on the Ipsos panel, were a lot of the surveys non-political stuff about which kind of knickknacks you prefer? And did you get some kind of points?

[Linus Van Pelt]

Admittedly I'm an ON novice, but only name I recognize on that letter is Rebick. Wonder if Dippers here know any of the others.

Michele Landsberg was a Star columnist for many years, and perhaps a bit ironically, was a rare pro-NDP voice at that otherwise Liberal paper. She's also Stephen Lewis's wife. Of course they're separate individuals, but it makes you wonder a bit whether Lewis is thinking the same thing but doesn't want to make a bigger stir by criticizing his successor.

Cathy Crowe is active in homelessness/housing issues in Toronto and is moderately well known locally.

Winnie Ng I recognize as the 1993 federal candidate in Trinity-Spadina, but that's only because I can remember this as a kid in this riding. I don't think she's particularly well-known, and I don't know what she's been up to since. I don't recognize the others.

I am not a great fan of Horwath, and some of my criticisms could plausibly be described as from the left, since they relate in part to her aversion to personal taxes to raise revenue. The line in the party's platform where they claim they can save hundreds of millions of dollars a year by appointing a new cabinet post called "Minister of Savings and Accountability", without telling us where this money will come from, is one of the biggest lame-gimmick joke excuses for a policy I have seen in a while. But still, she has been on the receiving end of some pretty weird criticism. An incrementalist approach and some populist rhetoric doesn't make one "right-wing".

[Linus Van Pelt]

It appears from the crosstabs that the two main reasons for Ipsos's likely/eligible split are: (1) seniors expressing more confidence that they will vote than youngs, and (2) men expressing more confidence that they will vote than women.

The first of these is almost certainly right, but the second seems a bit dubious; it might just be men being more likely to express confident opinions in general.

[Linus Van Pelt]

Given the record of these things, a few of these will likely be way off.

[Linus Van Pelt]

"Urban progressives" are not a totally unified category. In some areas of the city, there has been a significant influx of high-income professionals. Other areas are more fashionable with a sort of high-education-but-modest-income demographic that seeks out more affordable city neighbourhoods while being still very averse to the suburbs. The fact that the NDP can now get a weaker result in Trinity-Spadina than in the two ridings to its west is connected to movement of the latter group.

I suspect that the Conservative swing downtown may be linked to an increase in the condo population. Some of these finance or banking people are very right-wing economically. The swing appears to be negative in Davenport and very small in the two east-end ridings. Granted, it's hard to tell with these low swings.

[Linus Van Pelt]

Excellent stuff as usual, Krago.

[Linus Van Pelt]

Wouldn't you normally expect spoilt ballots to be more frequent in working class and minority areas?

Declined ballots are different from spoiled ballots. A declined ballot is where you deliberately go to the polling station, take your ballot from the official, and then immediately hand it back unmarked rather than going into the booth.

[Linus Van Pelt]

Wouldn't you normally expect spoilt ballots to be more frequent in working class and minority areas?

Declined ballots are different from spoiled ballots. A declined ballot is where you deliberately go to the polling station, take your ballot from the official, and then immediately hand it back unmarked rather than going into the booth.

That's... odd.

Yes, it is. It doesn't exist at the federal level, and I don't know the history of it.

[Linus Van Pelt]

Poll-by-poll maps are now available for all ridings at election-atlas.ca. Nice work as usual, 506!

http://election-atlas.ca/ont/