Wednesday, February 26, 2014

Largest possible Liberal majority (at this time)

During the convention, a "poll" was released; one that was clearly written in such a way to favour the party. The basic question was would you vote for a Liberal government or not. Do the Tories need to be defeated or not. This ignores possibilities such as voting NDP.

Since the poll did not clarify exactly which party people supported when using this method, it left me with only three numbers to work with. The supposed maximum Liberal vote. The proposed minimal Conservative vote. And "Others". As you know, while we might only have 3 parties with status in Parliament, a full 11% of voters, or, 1.6 million people, voted for parties not among these 3 in the last election; well over half of them for the Bloc, and over a third of them for the Greens.

The poll treats Canada's two "Middle Size" parties as though it were one of the "Small Size" parties. To note; the votes cast, combined, for all Independent, Non-Affiliated, Christian Heritage, Marxist-Leninist, Libertarian, Progressive Canadian, Rhino, Pirate, Communist, Canadian Action, Marijuana, Animal Alliance, Western Block, United, and First Peoples National parties, was a grand total of 130,521, or 0.89% of the vote. This number, even increased fourfold, could not match the votes for the Green Party alone.

However; the 1.6 million Canadians that voted for one of the parties not from the big 3 are, combined, more than one million votes behind the Liberals alone.

I accept the argument that in a poll trying to figure out who should be government, the two middle size parties - which realistically have 0 chance or winning the next election (even if something extraordinary happens; it's just too late, too close to the election) - should be disincluded. It does however make seat projections difficult.

What I've decided to do is thus. In provinces where the Greens would not win a single seat, even if their last-election vote was quadrupled in every riding, the Greens are excluded. This leaves them contending only in BC. In Provinces where the Greens, with all of the "Non Big 2" vote, would win more seats than the NDP would with the same vote, I've assigned all of said vote to the Greens. This means BC; where with the numbers given (14% provincewide) the NDP would not win a single seat, but the Greens would.

In Quebec, I tried to find the lowest poll ever for the Bloc, which turned out to be 9.6%. I then presumed a drop of 1/6th (it was a figure I've used in the past for rough estimations that has served me well)  which drops them to 8%. The remaining (26%) was assigned in whole to the NDP. I should note here, doing this only gains the NDP 2 seats (Beauport from the Tories and Shefford from the Liberals) compared to if I'd actually entered the Bloc at 10%.

The end result?
First; the numbers.
Greens - 1
NDP - 17, including 12 from Quebec. (15 and 10 in the case noted above)
CPC - 109, winning only BC (21 of 42 seats) and the Prairies (18 of 28)
LIB - 211, winning 25 of 32 Atlantic seats, 83 of 121 in Ontario, and 17 of 34 in Alberta.

This poll shows what I've been trying to say for quite a while; we have real potential in Alberta. While you might laugh, remember that similar polls show the NDP had potential in Quebec for years before they finally won. The potential is there, we just need to grab at it.

Since I suspect the Quebec figures to be questioned, I'll point out that the NDP retains party status with as little as a 20%N-14%B split in the vote. This would give the NDP 7 seats in the province. While I fully expect critics to point to their drop from near 60 to 7 and say this means the NDP has lost Quebec, I'd point out that 7/12ths of their Caucus would then come from the province. The NDP can remain a "Quebec Based" party with only a half dozen seats in the province; and in this case, they would do just that.



  1. Another Liberal convention, another poll showing them with a massive majority. Ho hum.

  2. A poll taken at a Liberal convention that says the Conservatives should be defeated and thus replaced by the Liberals.


  3. You both should take the time to read the post before commenting - the poll is of course not meant to be taken seriously, but its a thought experiment.

    1. Asking some to read things before they comment on is too much.

  4. This comment has been removed by the author.

  5. This comment has been removed by the author.

  6. I'm confused: CTV reports; "The poll found that 68 per cent of Canadians say “it is time for another federal party to take over and run the country"; and not "would you vote for a Liberal government or not. Do the Tories need to be defeated or not (sic)". Furthermore; "51 per cent of Canadians “agree” with the statement that “the Liberal Party is ready to be Canada’s next government,” while 49 per cent “disagree”.

    Secondly, it is unclear to me the purpose of the exercise or what you are trying to demonstrate. Thirdly it is unclear what criteria or data you use to formulate your results. Fourthly what is the purpose of merging voter support in some provinces and what is the reasoning behind it? I believe this is what you are doing but, your writing is not clear.

    I assume you have taken the 68% who agree with the statement; " “it is time for another federal party to take over and run the country" and allocated them with some modifications to the Liberal vote to garner a maximum potential of Liberal support across the country. I really am confused how you have allocated votes to the NDP and BQ and when and why you merge party support? It is unclear to me what you assume the Conservative vote to be.

    As a thought exercise it may have some merit. However, even as a thought exercise I am unsure what it demonstrates or helps us understand or knowledge it provides. It demonstrates the Liberals could win Vancouver East for example but only if the odds were stacked in their favour and the NDP did not field a candidate. What does that demonstrate or prove? How does it improve our understanding of politics in Canada?

    Does the exercise demonstrate the Maximum potential for Liberal seats is 211? Such a result would tie Mulroney's record for most seats, although Mulroney enjoyed 75% of HoC seats whereas 211 seats in 2015 will only result in a 62% share.

    1. 1. You are correct. The question was phrased in such a way to result in the largest possible number of people that they could trick into saying they'd vote Liberal.

      2. As the title says, to show the largest possible Liberal majority at this time.

      3. I used the data from the poll.

      4. I had to make a decision on weather or not to have the other parties having ANY seats at all. If I split the "3rd party" vote, the result is that no parties, other than the Liberals and Tories, win any seats at all. The result is then 224 for the Liberals and 114 for the Tories.

      I did not use the 68% figure as if I had... actually let me plug those in right now and see.
      Actually I find this rather interesting. I shall make a follow up post.

      The Liberal vote is the number of people who said the liberals are ready, while the Tory vote is the number of people who said the government should not change. It is presumed those who want to change the government, but see the Liberals as unready, will vote for a 3rd party.

      It is meant to demonstrate which ridings are "winnable" by the Liberal Party. Any riding not shown in (pinkish) red on the map is not one we can really win - with the caveats in Quebec. (My follow up post, however, will not have this problem)

      The ideal result of this post; or more accurately, my follow up post, is that those within the campaign running the next Federal election for the Liberal Party of Canada, will see the map, and decide to not send a nickel more than they absolutely have to, to any riding shown in CPC Blue.

    2. Thank you for your reply.

      So you give the Liberals 51% the Tories 32% and third parties 17% is that correct?

    3. Yes. I had thought that was clear. I apologize if it was not.

  7. Why do you write: "The question was phrased in such a way to result in the largest possible number of people that they could trick into saying they'd vote Liberal"? The questions are fairly straightforward and do not resort to trickery but, are meant to produce a binary result.

    Ipsos: "Some people say that the Conservative government under Prime Minister (sic) Stephen Harper has done a good job and deserves to be re‐elected during the next federal election. Other people say that it is time for another federal party to take over and run the country. Which of these statements is closest to your point of view"? and or

    "Thinking specifically about Liberal Leader Justin Trudeau, to what extent do you agree or disagree with the following statement: I think the Liberal party is ready to be Canada's next government: strongly agree; somewhat agree; somewhat disagree; strongly disagree".

    By insinuating trickery was involved you are discrediting your own work! You already rely on disparate data originating from separate questions that you cobble together (which is a questionable method of data collection). Insinuating one question "tricked" respondents should make people question the validity of the poll itself and your use of the data. More importantly you have slandered Ipsos Canada for something they have not done. Perhaps it is simply a poor use of words but, since, your writing in general is much in need of revision I would suggest you revise and edit then re-post. In particular your work would benefit if you included a description of your methodology and summary of the data used. it would also benefit if you outlined your hypothesis and the question(s) you hope to answer.

    1. "not resort to trickery but, are meant to produce a binary result." Binary results, in Canada, are trickery.

      Even in Manitoba, Saskatchewan, and BC; to suggest an election is a choice between two parties without a third option is plain outright trickery.

    2. This comment has been removed by the author.

  8. I will also note that I will not change how, when, where, and why I post. I speak and write the way I do for a reason. I fully intend to be understood by the widest possible audience. I fully intend for people to be drawn to my numbers and my maps and away from my words. I fully intend to continue to use those numbers and maps to make points.

    1. No, a binary response is not the same as trickery. If it was trickery then the data is unreliable, why would you want to use data that is untrustworthy? No matter how good your idea or analysis may be with unreliable data it is essentially worthless. If the data is unreliable instead of putting forth a viable hypothesis to be tested based on data you have produced a map and results that are imaginary-they must be since the data was accrued through "trickery" and therefore unverifiable. It is unfortunate but, is completely of your own making-you are the one who discredit your own results by describing Ipsos' questions as"trickery".

      If you intend to be "understood by the widest possible audience" I would suggest you take the time and revise your writing. Without publishing a methodology it is not possible to ascertain what you are doing with the "data" and without a stated hypothesis it is not clear what you hope to achieve. More importantly your writing is difficult to understand at points, a half hour of editing would greatly expand your audience.

      I am trying to help, whether you accept it or not is your business.