Wednesday, May 4, 2011

How vote-splitting gave the Tories Ontario - and a majority

The Conservative Party can finally lay claim to Ontario – and the majority bragging rights that come with it – but it was the NDP that may have made the decisive blow to the former Liberal fortress.

You can read the rest of the article on The Globe and Mail website here.

I believe a little further explanation is needed for this article. Clearly, if we add the Liberal and New Democratic votes together the Conservatives would have lost many ridings. In an email to me, a professor from the Université de Montréal noted that doing so would have reduced the Tories to only 137 seats. While this might be true, this isn't what I mean by vote-splitting. The Liberals and New Democrats are two separate parties, and their votes can't simply be added together.

What I was looking at was how New Democratic gains contributed to Liberal defeats. The best example of this is the riding of Moncton - Riverview - Dieppe in New Brunswick.

In 2008, Brian Murphy of the Liberals received 17,797 votes. The Conservative candidate had 16,297 votes while the NDP candidate finished third with 7,394 votes (source).

Using Elections Canada's preliminary results, we see that on Monday night Robert Goguen of the Conservatives received 17,408 votes, an increase of 1,111 votes. Brian Murphy dropped 2,553 votes to 15,244, while Shawna Gagné of the NDP gained 6,659 votes for a total of 14,053.

One could argue that by adding the NDP and Liberal votes, the anti-Conservative option would have almost doubled Goguen's total. But that is not the argument my article makes.

The gap between Goguen and Murphy was 2,164 votes, meaning that those who switched over from the Liberals (primarily, we can't know exactly how the vote split) to the New Democrats helped elect Goguen. Only 2,165 of the 6,659 new NDP voters in this riding (32.5%) needed to vote Liberal to avoid voting in a Conservative.

It was the same situation in many ridings in Ontario, where NDP gains were greater than the gaps between Liberal incumbents and Conservative challengers. Certainly, people need to vote for the candidate they support but there may have been some unintended consequences.

This is not to put the blame on the New Democrats, nor does blame need to be assigned. With their level of support, the Conservatives earned their majority. And in fact, in many of the ridings won by the Conservatives from the Liberals, the Liberal candidate was simply beat. The Conservatives made significant gains in many ridings, something that Liberal-to-NDP vote switchers could not have stopped. And it was the failings of the Liberal Party that pushed many of their supporters to the New Democrats and the Conservatives. Nevertheless, this vote shift played an important role in electing a Conservative majority government.

61 comments:

  1. Of course if you were to add 100% of the NDP and Liberal vote together - the Tories would have fewer seats. But I find this a bit of a pointless exercise. We know from past experience that one plus one never equals two. Also, this discussion tends to do nothing but cause a lot of navel gazing among people on the left while ignoring the fact that the Conservative vote went up substantially in Ontario. Maybe people should spend less time on mythical "vote splitting" and more time figuring out why so many more people voted Conservative.

    ReplyDelete
  2. DL,

    Did you read any of the above, or just the headline?

    ReplyDelete
  3. Notably, this part:

    Clearly, if we add the Liberal and New Democratic votes together the Conservatives would have lost many ridings. In an email to me, a professor from the Université de Montréal noted that doing so would have reduced the Tories to only 137 seats. While this might be true, this isn't what I mean by vote-splitting. The Liberals and New Democrats are two separate parties, and their votes can't simply be added together.

    ReplyDelete
  4. The story is likely to be different in the next campaign, and even more difficult for the opposition parties. I recall that the Conservatives presented a bill last parliament to add 30+ seats West of Quebec to bring the population per seat for the West in-line with the East. That is going to give them an even stronger base of seats to work from, so even if some of the vote splitting in Ontario were to disappear, the increased base of seats for the Conservatives will make it harder to prevent a majority.

    ReplyDelete
  5. I suspect McGuinty will be all over the NDP, (and greens) after results of the federal election. His provincial reelection will be fought on the same boundries and he'll have a good case to make if left wing voters are afraid of the Conservatives.

    As for the splitting --the Conservatives did get about 40% which is often the cut off for a majority. Heck 73 seats with same proportion of votes that the aforementioned McGuinty got for 72 doesn't seem weird.

    Yet... one would be foolish to ignore the obvious specifics. The Liberal vote down means more 2nd place Conservatives winning as well as 2nd place NDPers winning. Was there any previously 3rd place NDP candidate in Ontario who won? Certainly they were outnumbered by previously runner up Conservatives winning. I don't know if the switching voters realised that was a guarantee, but it certainly was going to happen.

    ReplyDelete
  6. The other major story line that seems to be ignored is the impact or the incorrect polls and media on the NDP surge in Quebec.

    On May 2 the people of Quebec were basically guaranteed that there would be a CPC MINORITY government where the NDP would be able to keep the gravy train stopping in Quebec.

    The pollsters ( and seat projection sites like 308) should have been within the margin of error on the CPC majority.... CPC to get more than 160 seats (MOE 5) seats 19 times out of 20.... Basically the CPC majority should have been easily predicted.

    If the people of Quebec went to polls knowing that however they voted there was going to be 4.5 years of CPC majority government they would have definitely voted in more CPC MPs and influential cabinet ministers.

    Pollsters who were more than 3 times outside the MOE need to be investigated by Elections Canada. They are not scientifically viable or supportable.

    ReplyDelete
  7. Elections Canada needs to force the pollsters and media to publish the following disclaimer.

    "Polls of this nature are for entertainment purposes only. They have been proven to not be scientifically valid"

    ReplyDelete
  8. That is an exaggeration. The polls were not very far off at the national level, averaging less than three points' worth of error for all of the parties.

    Certainly it was not a good election for them, but averaging 37.3% or so for the Tories instead of 39.6% is far from a complete failure.

    ReplyDelete
  9. It's interesting to see that the vote swings went from all parties to the NDP in Quebec, while swings in the rest of Canada were from the Grits and Greens to the Tories and the NDP in varying degrees. This will require somewhat more sophisticated analysis than just attributing the changes to "vote splits".
    For example, in Bramalea-Gore-Malton, the Conservatives actually lost about 3 points in their share of votes from the previous election, while the swing from the Liberals to the NDP resulted in a Conservative win. This is a phenomenon worth exploring further.

    ReplyDelete
  10. Yes, there are some ridings across the country where the Conservative winner did not have as many votes as the combined Liberal and NDP candidates. However, that analysis is too simplistic. In Enligish speaking Canada voters from suburban and rural seats tended to vote Conservative, while voters from inner cities tended to vote for the left of centre parties. Look at ridings such as Elmwood - Transcona in Winnipeg, Winnipeg South - Centre, along with the ridings in the '905'.

    Conservative support grew from those voters who wanted 3 things;

    1. Stable government, 4 elections in 7 years was too much
    2. Voters in Ontatio remembered what Bob Rae did as Premier and wanted to stop the NDP.
    3. Canadians by in large part are content and are satisfied with their economic situation and saw that there was only 1 party that was reputable in dealing with economy.

    As for candidates winning on the vote split, I did not hear this complaint coming from the left, when there candidates were winning in the 1990's when the PC's and Reform where fighting it out for the right of centre vote.

    ReplyDelete
  11. Canada and Canadians have voted for Mr. Layton and the NDP with 31.0% and this is the part that the LPC has to understand is that the increase in the NDP vote was NOT a vote against Harper anymore than it was a vote against Duceppe and Ignatieff - they were voting for something that was found to be wanting in the CPC, Bloc and LPC.

    Learn that lesson and Canada will start to listen to the LPC again, because adding the other parties never really worked unless you were talking about a coalition and there is not coalition happening now in my view.

    ReplyDelete
  12. With the NDP holding 58 or 77% of the 75 seats in Quebec and only 42.87% of the vote and holding 33.5% of the federal seats with only 31.0% support.

    What are the chances of Mr. Latyon opening up the question and making the system more responsive and relevant now that he benefits from the process.

    My guess is that Mr. Layton will be very very silent around this question and I wonder if the media will address this question with him.

    ReplyDelete
  13. Eric

    A question. If we had Prop Rep what would the result have been??

    As to a lot of this I agree with Eric, by switching a relatively quite small number of voters to the NDP from the Liberals the Tories got an almost free ride!!

    The point about extra seats for the next election is well taken although a big chunk of those additional seats are actually in Ontario so the "bias" may not be as pronounced as expected ?

    ReplyDelete
  14. Where this now gets interesting is because the NDP and Mr. Layton might be more inclined to support a declining voter subsidy as proposed by the CPC.

    The LPC has a massive debt from the election and after being reduced to 33 seats they have NO visible means to raise the money to pay off that debt let alone to start to rebuild the party.


    The NDP has been effective recently in raising money outside of the subsidy and depending on how much more effective the NDP feels that their fund raising can become with the additional 2.0 million new voters and assuming that the NDP holds to their long term plan - the NDP will have the a healthy war chest going into 2015.

    The LPC, GRN and the Bloc on the other hand without the voter subsidy will be broke, with little or no fundraising and would most likely not be able to field a complete campaign in 2015.

    So the question MAY now become - Will the first act of co-operation between the NDP and the CPC might well be to bury the Bloc, GRN and LPC by doing away with the voter subsidy.

    While this would not be in Canadas favour, it would be politicially favour for the NDP to kill off the three parties fiscally that they would be running against in the next election - the fact that it would benefit the CPC would not ever matter to the Official Leader of the Opposition Mr. Layton .

    My bet is that Jack will play ball and help bury the LPC and that there is no back this time for the LPC - unless the LPC supporters do what NDP Trudeau did when he moved into the LPC and change the party from within.

    ReplyDelete
  15. Its all the Liberal Party's fault for insisting on continuing to exist! Why don't they simply fold and instruct their members and followers to join the NDP - then there will be no more vote splitting!

    ReplyDelete
  16. The issue around voting splitting has benefited every party over the past 20 years - the really issue is there a trend to the CPC.

    A quick look at the numbers show the following, that when you back out the Quebec results from the totals, outside of QC the CPC support goes up to 8% over the 2011 election results to 48%, the NDP goes down to 26.5% and the LPC goes up to 20.7%.

    The CPC carried 8 of the 10 provinces and showed increased levels of support over the 2008 election for the Conservatives BC 45.54%, AB 66.78%, SK 56.26, MB 53.49%, NB 43.85%, PEI 40.96%, ON 44.4%, NS 36.72% and carried two of the three territories and achieved a beachhead not only in the GTA but Toronto as well and showed that a majority is possible without Quebec.

    http://www.cbc.ca/news/politics/canadavotes2011/map/fullscreen.html
    http://www.cbc.ca/news/canadavotes/map/2008/#

    ReplyDelete
  17. I wonder what the prospect for voting reform would be. I'm particularly fond of a variant of Bucklin. It would be very interesting to get a sense for how the results would have gone under that sort of a scenario (although it's difficult to say since we'd need to have a sense for second and even third preferences).

    ReplyDelete
  18. Precisely: a substantial number of right-of-centre Liberal voters might abandon the party to support the CPC, while many New Democrat voters may feel left without a political home and either stay at home or seek an alternative.

    As well, the CPC actually grew their support in Ontario, something missed amid all the recent focus on the role of vote-splitting, which, while critical, was not the entire story of the breakthrough in Ontario.

    Interesting analysis: thanks!

    ReplyDelete
  19. Naw, it really was vote-splitting that handed the Tories their majority. With our first-past-the-post system, sure, they earned their majority government. That applies to almost every majority government in history even won with less than 50% of the popular vote does.

    But so what?

    A breakdown of the results from each riding gives a strong indication that the Tories' increase of their vote happened predominately in their holds, and not necessarily in their new ridings (with a few notable exceptions). The thin margins are indicative of this, versus their ridings out in Alberta and Saskatchewan, for example, where the increased percentage of their leads increased considerably.

    They may have captured more ridings in Ontario, but their geographic popular vote, IMHO, has shrunk significantly. Ontario will be receiving the lion's share of those new ridings, which won't necessarily favour the Tories (which they'd have a difficult time gerrymandering) next time around.

    ReplyDelete
  20. Most Liberal Party supporters and voters go to a Conservative Party if there is no viable Liberal
    Party to vote for (I would say around 60% at least). This is a world-wide pattern, and especially
    true in Canada. You see this in western Canada especially.

    Liberals like a party that "talks left and governs right" (the old tradition) but will be happy with
    a party that "talks right and governs right" if need be.

    Liberals forming a coalition with Socialists is more of an historic anomaly.

    The Liberals in the UK are currently propping up the Conservative Government, just as the Liberals here in Canada propped up the Conservative Government of Steven Harper for almost three years.

    This may be the breakup of Canada down the road. The French have clearly and consistently expressed, and at the ballot box as well, a dislike for Conservative English Canadian (and an increasingly American in tone) culture.

    Believe me, I am not crowing over this. I am very sad with the cultural direction English Canada is taking, and the ballot box only represents this.

    ReplyDelete
  21. brusmit;

    Harper does not need Layton's suppport for any changes to political funding. He has a majority.

    Not that Layton would benefit from losing the per vote subsidy and being forced to rely totally on donations when the third palce party already has more of those while he had more votes. We can expect him to stop caring about election reform if he is benefiting from FPTP though.

    DL;

    What are you talking about?

    ReplyDelete
  22. Re: anonymous
    "As for candidates winning on the vote split, I did not hear this complaint coming from the left, when there candidates were winning in the 1990's when the PC's and Reform where fighting it out for the right of centre vote.
    "

    It was certainly observed on the left and in the media, and complained about by those right of centre. I don't think anyone asked you to complain about it in sympathy. (Although if one wants to get into a contest of the aggrieved, even in 1993-2000 there was certainly never a majority of Canadians voting for right wing parties anyway --there's always been left wing vote splitting)

    ReplyDelete
  23. BC Voice - You always seemed like such a Free Speech guy. First, Elections Canada doesn't regulate pollsters, and I don't see why they should. Second, if those pollsters were actually conducting the polls as they described them, and honestly publishing results, then there's no way they did anything wrong. Even if they intended to mislead with false poll numbers, as long as they only divulged true data then they can't be held responsible for any conclusions drawn by the voters.

    ReplyDelete
  24. Eric

    at the top of the page you have the CPC prediction at 36.4 they got 39.6.

    This is 3.2 off which is outside the MOE of any of the polls.

    The last EKOS poll was over 3 times outside the margin of error for the final result.

    You are only supposed to be outside the moe 1 time out of 20. The chances of being 3 times outside the MOE? For all the EKOS polls during the campaign.

    These obviously fail the criteria for scientific accuracy in the field test.These are entertainment, not scientific valid polls.

    You were once interested on polling bias of pollsters versus other pollsters.

    I argued and continue to do so that the bias should be measured between polls and actual results.

    If you want to use EKOS you have to add 5%
    Nanos 3% to their CPC prediction. They have been fairly consistent.



    I did this and predicted the CPC at 164 seats as would have your model.

    ReplyDelete
  25. kevinsutton

    There have been Liberal after Liberal saying that the Liberal party is the Party of the Centre (not the Left).

    Preston Manning could have worked with Finance Minister Paul Martin and conducted the educate the Canadian People on balanced budgets that Martin worked towards. Chretien/Martin were definitely centre right -- 10 year majority government and no child care, no carbon tax or any action on Kyoto.. they slashed CBC budget.

    Chretien and Martin started the shift to the right. The Liberal party only recently moved to the hard left and the people left voting for them would feel more comfortable supporting Harper than Dion. The leftist Liberals already voted NDP this election.

    ReplyDelete
  26. "What are you talking about?"

    Do you need this part explained

    While this would not be in Canadas favour, it would be politicially favour for the NDP to kill off the three parties fiscally that they would be running against in the next election - the fact that it would benefit the CPC would not ever matter to the Official Leader of the Opposition Mr. Layton .

    My bet is that Jack will play ball and help bury the LPC and not kick up a lot noise about it.

    ReplyDelete
  27. All this shows is we need to reform our voting system. Our FPTP system will lead to a polarized 2-party system with the NDP and Conservatives fighting for votes. Is that what Canadians want?

    This election there were numerous efforts to use prediction sites, like ThreeHundredEight, to influence strategic voting. They failed miserably. Of the 10 closest races, Project Democracy got 5 of them wrong, all of which the Conservatives won.

    We need to take a hard look at our voting system, unless you want to have a 2-party system.

    ReplyDelete
  28. "All this shows is we need to reform our voting system. Our FPTP system will lead to a polarized 2-party system with the NDP and Conservatives fighting for votes. Is that what Canadians want?"

    It's certainly what I want. A FPTP system tends to produce majority governments (which is a good thing), and with two dominant parties they should generally fight the election at the centre.

    I want majority governments. I'd rather have a majority lead by Michael Ignatieff than a minority lead by anyone, because minorites are awful governments.

    ReplyDelete
  29. brunsmit

    The part that you do no seem to be understanding is that Mr. Layton has significantly less influence than he has had since 2004.

    He does not get to play ball in killing the political subsidies. His point of view is irrelevant. He can not hold it up for more than a few days in committee and possibly try to filibuster in the HoC.

    The removal of the political subsidies on the CPC schedule is a done deal. The fact that they said they would do this in the campaign gives them the high moral ground but it really is moot.

    ReplyDelete
  30. Ira

    Free speech is one thing, but publishing invalid biased polls and saying they are scientifically valid is another.

    These polls and the media's acceptance of them as scientifically accurate had more of an influence on this election than any other issue. (IMHO)

    Imagine the impact of the CPC election machine (or a 3rd party CPC supporter) funding 4 pollsters that did daily polling that showed the CPC consistently at 44% and the Liberals at 11%?

    That would only require a small fraction of what they spend on attack ads.


    Read the COMPASS poll methodology and see if you can tell where it is biased.... I did and can't.

    ReplyDelete
  31. canadianveggie ... "This election there were numerous efforts to use prediction sites, like ThreeHundredEight, to influence strategic voting. They failed miserably. Of the 10 closest races, Project Democracy got 5 of them wrong, all of which the Conservatives won."

    I fail to see how you are drawing the conclusion that strategic voting did not work. The Conservative's last campaign message before the vote was that if you were a Liberal, the only way to prevent the NDP from getting significant power was to vote Conservative. 5 of 10 of the seats that you highlight voted in the Conservative. Seems that for those who switched from Liberal to Conservative, the strategic voting worked very well.

    Just because something doesn't agree with your ideals doesn't mean that it didn't occur.

    ReplyDelete
  32. The dynamics have changed. Has he really lost influence? He and the NDP are now to be taken seriously, not only by themselves, but by their opponents and by all Canadians. I'm sure over the next four years their behaviour, including policy, will reflect that of a government in waiting.

    South Parkdale Jack

    ReplyDelete
  33. Ira

    While rational choice theory would predict that two dominant parties would move to the centre, people (even politicians, and especially activists) are not that rational. If you look at the actual history of countries with two party systems (the U.S., the U.K. 1930-recently) it's as common for the major parties to take relatively extreme positions as centrist ones. You cannot count on two party systems producing outcomes anywhere near what the average voter would want.

    ReplyDelete
  34. Dan,

    The US example shows, I think, that the parties do move to the centre. The debates between McCain and Obama showed two people who were struggling to disagree with each other, as they held the same positions on nearly every issue. Their differences were merely the details of implementation - so instead, US elections are fought on personality and trust rather than policies, as there are no meaningful policy difference come election day (there are between elections, but that's just posturing).

    BC Voice - The polls are scientifically valid. They correctly show what the answers to that question is. Whether that question is relevantly similar to the question the voters answer when they vote, or whether the results are comparable at all given the unpredictability of voter turnout is a different issue.

    But the polls are scientifically valid. The scope of that applicability, though, is narrower than people seem to think it is. But that's not the fault of the pollsters.

    ReplyDelete
  35. @BC Voice of Reason:

    I don't disagree that people shouldn't put so much faith into polls, but I don't see how projections of a bigger conservative win would have pushed the Quebecois to vote more conservative. In fact, I think the spectre of another (or bigger) conservative win is what pushed us all to vote for the NDP instead of the Bloc.

    The Quebecois have consistently shown that they would rather be out of power than on the right.

    ReplyDelete
  36. Actually popular support gave Conservatives all of their seats in Ontario. I plotted a graph of popular support in Ontario versus number of seats won in Ontario for the past four elections (2000, 2004, 2006, 2008) for both Conservatives and Liberals. The 8 points form a smooth curve, with the Liberal points moving down the curve with each subsequent election, and the Conservative points moving up the same curve. Then after the election I added the new points for 2011. Guess what? They fell on the same curve. On this curve 39% popular support in Ontario corresponds to 51 seats in Ontario, and 44% in Ontario corresponds to 72 seats in Ontario. The former is not far from what the polls predicted for Conservatives, and the latter is what actually occurred (actually off by one seat). This is the difference between a minority and majority government. Hence the number of seats won by the Conservatives and Liberals in Ontario matched their overall levels of popular support in the same way as they did in the four previous elections. In other words, no special unique arrangement of vote splitting is to blame at all.

    From the beginning, the decision minority / majority in this election came down to overall popular support for Conservatives in Ontario. My question is why couldn't the polls do a better job of measuring support for the leading party in Canada's largest province?

    ReplyDelete
  37. Re: brunsmit
    "Do you need this part explained"

    No. That portion of my post was asked of DL. I knew what you were saying. As others have said, Layton has no say in the matter so his opinion isn't relevant anymore.


    Re: BC Voice of Reason; I suppose you are correct about the LPC and their ideological positon in the nineties. My error, I guess. My larger point still stands though. Vote splitting was not ignored in the nineties nor has it never affected left of centre parties before.

    Re: Centrism

    What empowers centrism is the size, strength, and independence of moderate decision makers. Grassroots parties do not have this as a feature. (It's kind of the point) So while people will appeal to the median voter... it will go no further than that. Anything not well known, or unnecessary to woo the moderate voters, or just too important to the base will not be moderated. I don't think this has much do with the number of parties though. It's just their makeup.

    The only exception would be some kind of party that's connected to a grassroots base of non-commital compromise oriented citizens... generally this kind of person is not the kind to be politically involved. Psychologically, political knowledge increases polarization.

    ReplyDelete
  38. Some people above are debating if the NDP policy on election financing will change. Maybe the better question is, will the CPC policy on election financing change?

    Sure, up until now a key objective of the Tories has been to destroy the Liberal Party. But times have changed. If you're Stephen Harper, having witnessed such a surge for the NDP, do you REALLY want to knock out the NDP's chief rival? In other words, would you prefer 1 party, or 3-4 parties now with Parliamentary representation, competing for votes and seats to the left of the CPC?

    ReplyDelete
  39. Dan - how have you identified what the average voter wants and how that compares with what the parties offer in the US and UK?

    ReplyDelete
  40. Some interesting analysis from http://dataj.posterous.com/a-conservative-majority-was-avoidable

    5,425 votes made the difference between the incoming House of Commons and a minority government in which the Conservative Party would have 154 seats (-13), NDP 108 (+6), Liberals 41 (+7), Bloc 4, and Green 1. That's one percent of votes in 13 ridings. In this post, I'll focus on what strategic voting campaigns could have done to prevent a majority.
    To prevent a majority, anti-Harper strategic voting campaigns would have needed to (a) pick four battlegrounds which they did not, (b) campaign harder in seven others, and (c) pick a different party to defeat Harper twice.

    ReplyDelete
  41. I've seen a few articles indicating how NDP votes cost Liberals seats in Ontario. How about an analysis of seats the Liberals cost the NDP? Liberal voters were surely "spoilers" for NDP candidates in Quebec, Manitoba, Saskatchewan, BC, maybe elsewhere. For example, Bal Gosal of the Conservatives beat Jagmeet Singh of the NDP by 538 votes in a riding where the Liberal got 16,402 votes.

    ReplyDelete
  42. Here is my two cents:

    Canada needs a more representational method of electing its MPs (I'm thinking something like MMP or STV, for example) because I just don't think that around 40% of the total votes should give any party 100% of the power. And I am not just against it now because it was the Conservatives that benefited this time, I was against it before as well. Majority governments are dangerous because it allows a party that may or may not represent the opinions of a majority of people to govern without anyone to keep their agendas in check (the Senate certainly doesn't do that).

    Plus, the current system does not allow people who live in other parts of the country to vote for the politicians they like. For example, I think it is a pretty safe bet that most Quebecers voted for the NDP because they wanted to vote for Jack Layton and not because they wanted to vote for their local NDP candidate. FPTP doesn't lend itself well to voting for parties in this manner. Yet, unfortunately, people still do it.

    Like I say, that is just my two cents.

    Brett

    ReplyDelete
  43. For all you Anonymous types out there !!

    Get some cojones !!

    A slight tweak on the filters in my Email program has ensured I NEVER see any entry from Anonymous !!

    If you haven't got the guts to sign your comments I haven't got the interest to read your twaddle !!

    ReplyDelete
  44. Two other factors to consider: first!, the fact that more people overall voted this time then last. In the example above, about 5000 more. While it appears most went to NDP, some surely went to Tories. Second, the Green party vote was way down - where did these votes go?
    Granted in this example these are smaller factors but they should be taken into account.
    Of course, as someone above aludes to, you never really know why a person votes the way they do. Most have very biased or crudely reasoned preferences ("well I normally vote A, but he did something I don't like, and I simply can't vote for B, so I guess it will be C"), so there is no way that most people put enough thought into it to make it worth analyzing. "Vote splitting" is a post election amusement, not an active strategy.
    - Kevin from Parkdale

    ReplyDelete
  45. Mike - It's hard to do systematically, and what the "average voter" thinks (and doesn't think) is a very complicated question. But to take the U.S. in the 1980s, it's quite clear from polling evidence that majorities were opposed to both the very free market economics of the Republicans, and the aggressive promotion of racial equality by the Democrats (e.g., affirmative action). But it didn't matter, because important parts of the party bases favoured those policies. The British parties then were even more polarized.

    Ira, it's hard to argue over perceptions, but I didn't see the agreement you did between Obama and McCain. In any case, I think that its at least as much the election-time centrism that is the posturing, and the between election differences that are real. At this moment in the U.S., while the Democrats are arguably left-centre (in the U.S. context), the Republicans are off in an alternative right-wing universe. In any case, the differences are stark.

    ReplyDelete
  46. One more factor - if the party's nominated candidate changed between the two elections, that adds another variable into the mix.
    - Kevin fromParkdale

    ReplyDelete
  47. If Canada had the same System as in Germany, then we would have the following results:
    * note that Germany elects half the parlaiment via fptp, the second half from party lists so that overall there is proportional representation. So with 308 ridings, there would be 616 representatives. I counted YT+NT+NU as one state. In order to make results comparable, I divided the results by two, so there are "half-seats" now.
    * Germany has a 5% / 3 direct mandates hurdle to get the poportional candidates, so the Greens would merely get the one representative

    con, ndp, lib, bloq, green - total
    BC: 18, 12.5, 5, 0, 0.5 - 36
    AB: 20, 5, 3, 0, 0 - 28
    SK: 8, 4.5, 1.5, 0, 0 - 14
    MB: 8, 3.5, 2.5, 0, 0 - 14
    ON; 49.5, 28.5, 28, 0, 0 - 106
    QC: 13, 33, 11, 18, 0 - 75
    NB: 4.5, 3, 2.5, 0, 0 - 10
    NS: 4, 3.5, 3.5, 0, 0 - 11
    PE: 2, 0.5, 1.5, 0, 0 - 4
    NL: 2, 2.5, 2.5, 0, 0 - 7
    YT+NT+NU: 1, 1, 1, 0, 0 - 3
    total: 130, 97.5, 62, 18, 0.5 - 308

    ReplyDelete
  48. I think they were shifts in the final days that the polls failed to catch but I still dont think any pollster could have come out to say that the bloc will win only 4 seats or that Harper would win by such a large margin. Oh well at least the polls gave us some direction

    ReplyDelete
  49. Dan - Look at how much intervention the Republicans do in the economy. They pick winners and losers through subsidies and tax cuts (arguably even more aggressively than the Democrats do). That's not a right-wing party.

    Those two parties differ on what America's posute should be (beligerent vs. compromising), but that's about it.

    ReplyDelete
  50. Dan ... you always have differences between two parties - otherwise, you'd just have one. It's not enough to claim differences (and references to supposed polls where people disagree with one policy put forth by a party). You have to have credible information to claim that there is a big chasm between parties - otherwise, you come across as someone who just doesn't like the policies of one party so wants to claim they're extreme.

    ReplyDelete
  51. The Conservatives increased their popular vote in Ontario, from 39% in the 2008 election to 44% in 2011. This rise in popularity moved them from 51 seats to 73 seats in the province, and gave them a majority nationally. (The Liberals achieved similar success in Ontario in 2004: 75 seats with 44.7% of the provincial vote.) The pre-election polls largely missed this crucial rise in Ontario, even though the amount of sampling that was done in the province should have been sufficient to find a change of that size. (Québec polls were far more successful.) Nothing about Ontario’s history or political scene points to volatile voting intentions. Conservative support in the province has increased by similar increments in each of the past three elections. If the polling results indicated volatility during the campaign, I would say it makes more sense to look for volatility in the methods used to estimate voter support in the province, rather than in the Ontario voters themselves. Some thoughts about that:

    There was a moment, on the morning of April 26, when threehundredeight.com gave a glimpse of a Tory rise in Ontario. It showed a monotonic increase for the Conservative curve of popular support in Ontario, beginning near the April 12 debate, and continuing to April 24 - the last point on the graph - with an estimate of just under 48%. That value matched the April 24 Nanos poll which was featured in Jane Taber's April 25 G&M article as having found 47.8% Conservative support in Ontario. However, the appearance of a Conservative rise in Ontario was subsequently erased retroactively on threehundredeight.com, starting on the afternoon of April 26, and continuing until the estimated Conservative support in the province was brought back below 40%. I asked how the curves could be revised so dramatically after the fact. Your response convinced me that the apparent rise had been due to high outliers, which were subsequently averaged down by other more realistic polls, which only became available later. Looking back now, I wonder if that one graph, so briefly available for viewing, was my only fuzzy glimpse of what was really going on in Ontario. If so, what was the nature and quality of the subsequent polling information that erased the appearance of the Conservative rise, and pointed so consistently and convincingly - and yet ultimately incorrectly - to another minority government?

    And what about the averaging process? I now see at the bottom of your web page a list of poll results employed, by region. However, I don’t see the Nanos April 24 poll which found 47.8% Conservative support in Ontario. I don’t see anything nearly that high for that date. If you didn’t have the April 24 Nanos Ontario result in your model, how could you ever have presented an Ontario curve ending on April 24 near 48%? Or did you have it in the model, and then at a later time un-have it? Additionally April 26 was the same day that you changed your method in two ways: retired old data more quickly, and removed an unspecified “correction.” I wonder if these changes helped to erase the appearance of a Tory rise in Ontario. All of these things occurred the day after Frank Graves’ bombshell announcement that NDP could win 100 seats. Did that revelation provoke a response in the quantitative industry that influenced data-collection and modeling decisions in an undesirable way? LJ21

    ReplyDelete
  52. LJ21,

    The April 24 poll was replaced by a newer Nanos poll. You can see that the Nanos poll for April 23 (encompassing April 20-23) is included, as is that of April 27 (encompassing April 24-27, the extra days being the non-polled Easter holidays).

    So, the data was definitely included, but just spread over those two rolling polls.

    ReplyDelete
  53. you'd better register the domain threethirtyeight.com soon!

    ReplyDelete
  54. If the Conservatives were smart, they would have told their supporters to lie to pollsters to make it look as if their support was lower than it actually was (I am not suggesting they did this). Since we've seen that the best predictor of a decline in Tory support is high Tory poll numbers, it was in the party's interests to maintain low poll numbers immediately prior to the vote.

    ReplyDelete
  55. I would just like to thank you for your hard work. Your site has been interesting and instructive.

    I'm not a pollster, so I my comments are of limited usefulness. But it seems to me that there is a built-in defect in pollsters methodology when conservative parties around the world get better results than their polling. Maybe it's just harder to get in touch with conservative voters, who are older, more regularly employed, and are less likely to answer pollsters' calls.

    KN

    ReplyDelete
  56. Just to add one thing to this article rather late but specifically to address the issue of vote splitting.

    More Conservative candidates won with a majority than for any other party. Almost 2/3s of Conservative MPs were elected with a majority. 107 of 144 candidates elected by a majority in 2011 were Conservatives.

    To argue that some how the election could have turned out differently at the ballot box is completely unrealistic. If anything, it is the Conservatives that missed out on a number of ridings that without strategic voting they would have won.

    One only needs to look to the 1997 and 2000 elections to see how in that case the Liberals did massively benefit from the vote splitting.

    ReplyDelete
  57. I think made it quite clear what I was arguing, and what you're commenting on wasn't it.

    ReplyDelete
  58. @Bernard
    How many Conservatives won their riding with an absolute majority is a bad metric to measure whether they deserved the seats in this election - it just means that the conservative support is fairly concentrated. The best measure to find out whether the elections whether the party benefited from the voting system is to see how many seat % every vote % got - in a proportional system a party would get 1% of the seats, for 1% of the vote. Consider this election:
    Con: 1.37
    Lib: 0.58
    Bloq: 0.22
    NDP: 1.08
    Green: 0.08

    The greens got screwed over the most, so did the bloq and the libs. The NDP got approximately what they deserved - the cons get much more. The conservatives got more than twice the seats for every vote than the liberals.

    ReplyDelete
  59. I'm so tired of people complaining about a party winning with 40% of the vote and wanting proportional rep.. We have a system that elects MP's in ridings, that these MP's are suppose to serve them/ represent them. if a party were to win 51% of the vote and got that percentage in every seat in the country they would get 100% of the seats. This is because if they win in their riding they serve that riding. National numbers have nothing to do with it. conversely a party could win a majority government with 26% of the vote nationally (51% in half of the ridings in the country) If a political party can't garner the support in a specific region (riding) then they don't deserve the seat.

    ReplyDelete
  60. @00:36
    You can form a majority government with only about 14% of the eligible votes in Canada's three party system - by winning 33.4% of the votes in canada's 155 least populated ridings.

    You are saying that "If a political party can't garner the support in a specific riding then they don't deserve the seat" - but what you really mean is that "if the _people_ don't all rally behind their lesser evil candidate, then they don't deserve to be represented".

    ReplyDelete
  61. """Anonymous said...
    I've seen a few articles indicating how NDP votes cost Liberals seats in Ontario. How about an analysis of seats the Liberals cost the NDP? Liberal voters were surely "spoilers" for NDP candidates in Quebec, Manitoba, Saskatchewan, BC, maybe elsewhere. For example, Bal Gosal of the Conservatives beat Jagmeet Singh of the NDP by 538 votes in a riding where the Liberal got 16,402 votes."""


    Its actually the opposite Jagmeet Singh of the NDP cost Liberals the seat. He Siphoned off 10,000 votes away from Malhi and still lost the seat which Malhi won last time around by 6000 votes. The funny thing here is that the conservatives in this riding recieved 3% less votes as compared to 2008 and still won over the Liberal incumbent Malhi because the NDP took votes away from the Liberals. This riding always had a very strong Liberal base and had no NDP presence whatsovever.

    ReplyDelete

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.