Ontario: McGuinty doomed! Or possibly saved! - Macleans.ca
 

Ontario: McGuinty doomed! Or possibly saved!


 

Toronto Star:

The poll also found that the Progressive Conservatives, led by Tim Hudak, sit at 40 per cent, ahead of Premier Dalton McGuinty’s Liberals at 28 per cent. Andrea Horwath’s New Democrats are at 23 per cent…

Globe:

Progressive Conservative support came in around 30 per cent… with Dalton McGuinty’s team sitting at 39.9 per cent, according to the latest Nanos Research numbers. Support for the Ontario NDP, meanwhile, has seen incremental increases… to 24.7 per cent.


 

Ontario: McGuinty doomed! Or possibly saved!

  1. Heh. This is only gonna make McGuinty’s job more difficult. He governs by polling, and if the polling is going to be this confusing, he’ll have no idea what to do.

  2. My instinct tells me the reality is somewhere in the middle.  I’d guess the Forum poll might be more accurate, I find it hard to believe McGuinty is doing better than the election.  Both these polls may be the 1 in 20 that fal outside of 3% accuracy.

  3. The Tories are $6.000,000 in the hole from the last election (they spent $14 mill). Don’t count on any snap elections.

  4. See here’s the problem with Ontario politics – Canadians don’t care about their own politics. Half of them are glad we got rid of George W. Bush and elected Obama, while the other half voted for Sanjaya. And of course Canadians pay even less attention to their provincial politics (outside of Quebec), even though we are perhaps the most fiscally decentralized country in the OECD. 

    What you get in that environment, is the same thing you got in newly democratizing Eastern European countries. Inconsistent polling driven by weak preferences and low information. I know university graduates that weren’t aware of who the Prime Minister was. Ever since I’ve followed Ontario politics, I’ve seen the same thing. The incumbent government has low poll numbers between elections. Happened to Rae, happened to Harris, happened to Eves and it’s happening to McGuinty. But that picture didn’t always hold up come election time. 

    The premier happens to be the one person some people know enough about to blame stuff on. It’s kind of like an election between Captain Picard and Riker. Picard is a guy with definable characteristics, whereas Riker is basically some guy with a beard. You’re going to vote based on how you feel about Picard, not Riker. 

    My guess is that the short-term economy has a lot more to do with who wins and loses. McGuinty stayed in power (despite Ontario’s continuing long-term economic decline) by the skin of his teeth, thanks to the same recovery that won Harper his majority. As that picture darkens somewhat, his chances will worsen. If I were Horwath or Hudak, though, I’d wait for the Toronto housing market to collapse before pulling the plug. 

  5. Polls haven’t been reliable for a long time now, so partisans just read whatever they want into them.

  6. There are three kinds of lies:  lies, damn lies, and statistics.
    – Samuel Clemens (attributing Benjamin Disraeli)

  7. Given that the Forum poll was asking about a controversial issue, that many people are opposed to (according to the poll!) it would not be surprising that it found worse results for the sitting government. What’s important is if they told people who answered “this is a survey about the OLG’s plan to open new casinos” and also if they asked the casino questions before asking about vote preferences. 

    First, opening the survey by saying the topic would have caused some self-selection on the topic (those who are passionate, i.e. probably those strongly opposed would likely have a higher probability of participating).

    Second, both introducing the topic of the survey as the casino’s, but more importantly asking about people’s views on them fist, would make their opinions on the casino much more salient in considering their opinion on the government itself. Since the issue is one the people tend to oppose disproportionately relative to opposition to the government itself (and assuming that people draw some connection between the casinos and the government of the day) this would likely skew people’s responses.

    It isn’t that polling is hopelessly useless, it can be incredibly powerful. However it is also incredibly easy to introduce significant bias, which may well have occurred in this case.