Tuesday, February 21, 2012

Toronto--Danforth on track for repeat

On Wednesday, the Toronto Star released the details of a new Forum Research poll for the riding of Toronto--Danforth. The by-election to fill Jack Layton's vacant seat is scheduled for March 19. The poll shows that voters intend to cast their ballots almost exactly the way they did on May 2, 2011.
The poll indicates that Craig Scott, the NDP candidate, has the support of 61% of respondents, identical to the 60.8% who voted for Jack Layton in the last election.

Liberal Grant Gordon has 19% support while Conservative Andrew Keyes has 14%. This compares to the 17.6% and 14.3% that their parties each received in May 2011.

Another 4% intend to vote for the Green candidate Adriana Mugnatto-Hamu, the only major returning candidate from the last election. She got 6.5% at that time. Finally, 2% said they will vote for someone else. They will have the option, as another half-dozen names will be on the ballot.

It does not appear that it will be an exciting election night, with no variation in support of more than 1.5 points since last year.

Demographically, Scott gets his best numbers among voters aged 18-34 (67%) and 55-64 (70%). It could be coincidental, but Jack Layton was 61 when he passed away.

Scott also over-performs among women at 68% support.

Grant Gordon does better among men than he does women, but his numbers are generally uniform across all age groups. This is also the case for the Liberal Party itself in most national polls.

Andrew Keyes does best among the oldest cohort of voters (65+), with 22%.

Somewhat oddly, the poll asked respondents how they would vote if either Thomas Mulcair or Justin Trudeau were leaders of their respective parties. Neither case changes much for the NDP, as Scott would receive 58% of decided support in both scenarios. It improves slightly for the Liberals with Mulcair leading the NDP (22%) and somewhat more with Trudeau leading the Liberals (24%). Of course, Trudeau has already said he isn't interested and the NDP leadership race comes to an end after the by-election.

The New Democrats could actually end up with a higher level of support when the votes are counted. The poll shows that 54% of its respondents voted for the NDP in the last election, rather than the actual 61%. The poll could have under-sampled NDP voters. It does not appear likely that people would be reluctant to admit that they voted for Jack Layton. Voters who say they voted for another party were over-represented (6%), but it could be that these people were just not willing to divulge who they voted for.

The poll also says that 2% of respondents did not vote (rather than the actual 35%), but this may simply show that undecideds and people who do not respond to polls also tend to be non-voters. Though it would have been better had respondents lined up more closely to the results of the 2011 election, the population has undoubtedly changed over the last nine months. The population of students, for example, would be quite different in the months of February and May.

In any case, the poll quite clearly shows what everyone expects from the riding - it is a safe NDP seat. My history of the riding makes that clear. The Conservatives may have liked to portray this as a seat for the Liberals to lose, but that was a little far-fetched. And as my article in the Globe showed yesterday, the retention rate for incumbent parties over the last half-century is more than two-thirds in by-elections. Barring a huge upset, Craig Scott will be Toronto--Danforth's next MP.


  1. Do you have any stats on how accurate these riding level polls are? I've seen them done federally in NL and in cases they've been off by a wide margin.

    One thing I've looked at a bit is social media regarding this by-election. I read before that social media can predict outcomes, which I don't think is that accurate. However, from looking at the Twitter hashtage #tordan and looking at Craig's and Grant's Facebook pages it seems to me that there is more discussion about Grant then Craig, which is weird if there is in fact a 40% gap between the two.

  2. That might have more to do with Gordon being more recently nominated in a contest that gained national media attention due to the pro-life stance of his opponent in the nomination contest.

    Riding polls are a mixed bag, and it is difficult to judge them when they are released mid-campaign. Things change in a few weeks. But some of the recent by-election polls done in Quebec were relatively solid, and with the poll showing such a huge edge that is very similar to last year's election result it seems pretty reliable. It is extremely unlikely that it would be inaccurate to the point where there would be a 21-point swing from the NDP to the Liberals (what they need to win).

    But, of course, the by-election is still a month away. Things could change, so we won't really know if this poll was on the money or not.

  3. That the poll results should be roughly similar to the last election is not too surprising. I was expecting something more along the lines of the NDP at 45-50% since Layton was so personally popular in the riding.

    I will be very, very, surprised if the result is anything other than an NDP victory on the 19th.

  4. In St. John's South Mount Pearl in 2008 a poll showed Siobhan Coady with a 10 point lead over Ryan Cleary, but on election night it had all but evaporated and she just squeaked out a win. Last year a poll also showed her with a slight edge on Cleary, however on election night he beat her by almost a 20 point margin. The "orange crush" may have played a roll in that though.

  5. St. John's South-Mt. Pearl has clearly proven over the last decade or so to be a seat in flux, electing members from three different parties (Loyola Hearn for the CPC in 2004 and 2006, Siobhan Coady for the LPC in 2008 and Ryan Cleary for the NDP in 2011). By contrast, the NDP won every federal election in Toronto-Danforth in the same period, and by increasing margins.

    It makes much more sense that riding level polling would be inaccurate in St. John's South where the electorate is in flux than that it would be so in Toronto-Danforth where a clear trend has been established.

  6. By-elections are meaningless.

    I used to think otherwise, that maybe they could hint at trends or directions or produce some political intelligence.

    However, looking at the by-elections before the last election shows just how random they can be.

    Liberals score an upset against the NDP and win a Manitoba seat. They're later wiped out across the country by the NDP.

    Conservatives make steady inroads against the BQ in a string of Quebec by-elections over the years only to see both themselves and the BQ wiped out by the NDP.

  7. Some by-elections are sometimes very meaningful. In retrospect, the outremont byelection was HIGHLY meaningful.

  8. By the numbers

    Current NDP membership and growth since October 2011, according to figures released by the party on Feb. 21:

    * Alberta: 10,249, +13%
    * B.C.: 38,735, +29%
    * Manitoba: 12,056, +17%
    * N.B.: 955, N/A
    * N.L.: 1,030, +415%
    * Nova Scotia: 3,844, +196%
    * Ontario: 36,760, 65%
    * P.E.I.: 268, +98%
    * Quebec: 12,266, +623%
    * Sask: 11,264, +26%
    * 3 territories: 924, N/A
    * Total: 128,351, +53%

  9. The NDP would win this by-election, but I suspect the Liberals would do much better than they did in May 2011.

    I expect the Conservative popular vote to be less than 10%. I doubt many Conservatives in Toronto-Danforth would bother voting in a by-election that does not matter.

  10. There is a new Forum poll on Ontario vote intention: PCs 36%, Libs 32%, NDP 26%

  11. The idea of Jack Layton is much more popular than Jack Layton was himself IMHO.


COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.