Talk:2015 Canadian federal election/Archive 1

Archive 1

Archive 2

Misleading

Citing a source improperly is worse than no source. The date given is not in the source. Please find a source, without doing calculations. It is false to suggest that "war and insurrection" are the only things that can change the date. The constitution, and the Canada Elections Act, make very clear the governor general retains his absolute authority to dissolve parliament whenever he chooses. Also, the 4 year fixed term is in the Election Act, not the constitution. So, it can be changed or abolished. On a lessor not, it's a mistake to refer to "Bill C-16" without qualification. Bill numbers are re-used. Only a few are so famous/historic, that they continue to be referred/known by their number, without qualification long after their introduction. --Rob (talk) 19:24, 3 May 2011 (UTC)

A Masters tournament date can be changed as well; however, that does not prevent us from assuming that it will not be changed and writing down the current date on its article. The governor-general is mostly an honorific role and, while he can change the date, they rarely do -- if ever -- for no good reason. _EricLeb⁰¹ (Page | Talk) 00:41, 4 May 2011 (UTC)

Actually, Governor Generals always set a date, normally on the advice of the PM. That's how our constitution works. Every single parliament in Canadian history has been dissolved by the Governor General before the legal deadline to do so (except in time of war, when it went past the normal deadline). Throughout history, every election date (post 1867) only become known after the governor general dissolved the preceding parliament. The Canada Election Act is very clear that "Nothing in this section affects the powers of the Governor General, including the power to dissolve Parliament at the Governor General’s discretion". The only people with reliable advance knowledge (e.g. before the writ is dropped) are the PM, GG, and their inner circles. That's quite different than a Masters tournament, or any regularly scheduled event. We should not assume a pattern will repeat until the pattern has actually happened. --Rob (talk) 07:18, 4 May 2011 (UTC)

The GG may have always set a date, but only because we didn't have fixed election dates until Bill C-16. This is the first majority government since that bill, so we are assuming that the date will not change until the GG changes it. To assume that he will change it is WP:CRYSTAL. _EricLeb⁰¹ (Page | Talk) 23:39, 4 May 2011 (UTC)

We shouldn't assume anything without a proper source. Currently, this article has only two citations. One provides no date, and says the GG retains full discretion (the law itself). The other (a parliament web page about fixed dates) does give a specific fixed date for the next election, but that date is in 2012. Neither mentions 2015. It would be really great if people could get good sources (3rd party reliable) *before* they made articles, and then write something based on what those sources say. It would save us all so much time. If you read my original point, you'd see I'm not insisting on my own interpretation of the situation, but rather am just saying that whatever it is we say must match up with sources. --Rob (talk) 07:21, 5 May 2011 (UTC)

How many do you want? 117Avenue (talk) 07:39, 5 May 2011 (UTC)

Interesting how the sources you added indicate more flexibility than what the article suggests. For example one says "However, under the Parliament of Canada Act the prime minister is still free to request an election at any time and Prime Minister Stephen Harper did so in September 2008." That and another suggests there could be an unwanted conflict between Ontario and federal election dates, which opens the way for a change in one. And, I see you still have the original two sources, which indicate the full discretion of the GG remains, and suggests 2012 as a date. The article used to conflict with two sources. Now it conflicts with four out of six. You've missed the whole purpose of sources. What you want to do is write an article based on your personal understanding of the topic, and add a bunch of citations, regardless of what the sources actually say. However, it should be the other way around. First, find the best sources possible. Then, write the content of the article based on those sources. Don't try to force sources on to existing inaccurate/incomplete content. --Rob (talk) 07:53, 5 May 2011 (UTC)

I'm sorry, I am still confused on how there is conflict. Nowhere in this article does it says that the election will take place on October 19, 2015. It only says that it is scheduled for the third Monday of October in the fourth calendar year following polling day for the last general election, which is supported by all the references. 117Avenue (talk) 16:25, 5 May 2011 (UTC)

Polling graph

I am thinking that once we have a couple more polls that it would be useful to have a graph to follow though to the the next election, such as this one from last time. I would be will to make that happen and stick around to update it between now and the next election. Objections? - Ahunt (talk) 19:00, 26 May 2011 (UTC)

Thanks for volunteering again. But I think we need quite a few more polls. Looking at the last one, I think it wouldn't have been viable to start until three months after the election. 117Avenue (talk) 06:01, 28 May 2011 (UTC)

Oh I agree - a graph made up now would be pretty boring-looking. I think we need at least ten or so polls before it is started. - Ahunt (talk) 11:10, 28 May 2011 (UTC)

Well we have ten results, so I started in on the graph and will add it to the article. - Ahunt (talk) 12:07, 3 September 2011 (UTC)

It still doesn't seem like there is enough polls. But there's no reason to remove it now that it's been uploaded. 117Avenue (talk) 22:45, 3 September 2011 (UTC)

It will get better over time as we get more data. Heck Wikipedia has so much influence maybe the pollsters will will see the graph and decide to do more polling! - Ahunt (talk) 23:20, 3 September 2011 (UTC)

This is a bit of an obsessive-compulsive comment, but I'm wondering if there are thoughts on the listed order of polls with identical field dates insofar as how it affects the polling graph. Take for example the recent Angus Reid and Innovative Research polls, both conducted on March 20-21. My reflex was to list the Innovative poll after Angus Reid only because it was discovered later, but I note that graphing the Innovative poll first would have resulted in a cleaner trend line for the NDP, Liberals and Green Party, by reducing the somewhat unsightly "mini-spiking" that resulted from the current order. Again, sorry for this rather nitpicky comment and I realize polls with identical field dates are generally rare, but I'm almost itching to reverse the order of those two polls for this aesthetic reason. Overall great job on the polling graph! Undermedia (talk) 18:28, 4 April 2012 (UTC)

I am glad that you like the graph, I try to keep it up to date! It isn't very sophisticated, as it is just done on an OpenOffice Calc spreadsheet. The problem of which order to put same date polls in is an interesting one. Once we have a year or two's data the individual spikes will be less obvious as it gets more visually compressed. I do understand your concerns though, the different methodologies used by different companies results in a lot of jumping up and down in the line. I could suggest that because we use the poll's closing date to arrange them that if they have the same closing date then the one with an earlier opening date should go first. For instance if poll "A" was done 28 March - 3 April and Poll "B" was done 29 March - 3 April, then Poll "A" would go first. Otherwise I suggest you just go ahead and arrange them as you think best and once I see that you have done that in the table I'll upload a new version of the graph to match. - Ahunt (talk) 19:58, 4 April 2012 (UTC)

Totally agree with everything you said. I would normally look at the opening date to decide the order of two polls with the same closing date, but these two polls are exceptional in that they even have the same opening date. My OCD got the best of me so I've switched them around, but if this happens again in the future I'll now consider the repercussions on the graph *beforehand* so as not to cause undue graph re-editing.Undermedia (talk) 01:58, 5 April 2012 (UTC)

No problem, there is a new Harris-Decima poll added this morning as well so I'll add that, reorder the other two and post a new graph. Wikipedia is all about collaboration, so there will always be little adjustments to make like this, it is just part of the process. The table and the graph are important as I understand that national media journalists do refer to them both for information. No one else seems to be aggregating this information in one place like we do, so we have to get it right. - Ahunt (talk) 12:03, 5 April 2012 (UTC)

 Done - Ahunt (talk) 13:09, 5 April 2012 (UTC)

Made a slight correction to the date of the Ipsos Reid poll from back in November 2011; you may opt to tweak to graph to reflect this. Undermedia (talk) 14:37, 22 May 2012 (UTC)

Thanks for sleuthing that error out. Naturally we want all the info here to be as accurate as possible. I have changed the date on the spreadsheet and it will be incorporated into the Commons graph when the next update (poll) is added. If you find any other errors in the table, please do fix them and note them in the edit summary and I will adjust the graph as required. Accuracy is our watchword. - Ahunt (talk) 14:48, 22 May 2012 (UTC)

Governor General's power

The Governor General, representing the monarch, has the power to call an election at any time. This power has not changed at all, and the fact that it hasn't happened without the advise of the Prime Minister in the past, doesn't mean it can't still happen. It is even stated in the act "Nothing in this section affects the powers of the Governor General, including the power to dissolve Parliament at the Governor General’s discretion." I see no reason not to mention this in this article. 117Avenue (talk) 05:22, 29 July 2011 (UTC)

Having reviewed this I agree that the wording should remain as you left it - the power to dissolve parliament and call an election belongs to the GG. - Ahunt (talk) 12:37, 29 July 2011 (UTC)

That would be great, except that since King-Bing (1926), the Governor General has not dissolved parliament without the advice of the prime minister. Furthermore, it is unlikely, and would likely lead to a constitutional crisis, if the GG dissolved parliament without the advice of the PM. Even in the recent case of the coalition government crisis of 2008, the GG prorogued government - the exact request of the sitting prime minister. The wording I proposed does not negate the power of the GG, it merely recognizes a parliamentary convention that has existed for over 80 years. This convention did not exist in the earlier part of our history, thus to word it otherwise would infer that Canada still has the parliamentary conventions of the pre-confederation era. — Preceding unsigned comment added by Brandemon (talk • contribs) 18:46, 29 July 2011 (UTC)

Yes, I agree, that the country would have to be in some sort of crisis state for this to happen, but our country has seen crisis and war before, and it still could happen. It was only four months ago when Canada became the first country in the commonwealth to have a government in contempt. The wording I propose states that it is still the Governor General who calls the election, as said in the referenced act, and that it is currently conventionally done on the advice of the incumbent prime minister. 117Avenue (talk) 23:47, 29 July 2011 (UTC)

Very well. It's a minor point really. I'm just a stickler, I guess. — Preceding unsigned comment added by Brandemon (talk • contribs) 09:39, 30 July 2011 (UTC)

Leaders

Why are Stephen Harper and Elizabeth May included in the infobox? I thought it had been decided on other election articles that we shouldn't assume who the leaders are going to be when the election is finally called. Who knows what will happen to the leaders over the course of the next few years, Jack Layton had been included in the infobox and we now know he won't be leader. Newfoundlander&Labradorian (talk) 13:30, 30 August 2011 (UTC)

I've been saying the same thing, but no one seems to agree with me. 117Avenue (talk) 01:36, 31 August 2011 (UTC)

The opposite was decided. As with elections elsewhere, the current leaders are listed until it becomes a certainty they won't lead their parties at the next election. -Rrius (talk) 08:55, 31 August 2011 (UTC)

I think it seems fairly intuitive that we should list the current leaders and update the article if they change before the election is held. - Ahunt (talk) 11:52, 31 August 2011 (UTC)

I agree with Ahunt and Rrius. Think about the 2011 election article. We originally had Dion as Liberal leader, and then Ignatieff. That was partially a different situation, mind you, because there was the threat of an election around every corner. Another example is the Next United Kingdom general election. The election was held in 2010, and a lot could change (and has changed) since then. However, whoever has been leader of the party (the major parties, that is) has been mentioned in the infobox (if you look through the edit history, you'll see that originally Gordon Brown was Labour leader, then Harriet Harman and finally Ed Miliband. Wikipedia is pretty fluid, I think we can keep up and make changes to leadership standings when they arise. Bkissin (talk) 23:49, 31 August 2011 (UTC)

I would also like to raise the question, how long do we leave a leader empty? Members of the Liberal Party have elected to not hold their leadership election until 2013. What if something triggers a snap election before the Liberals can organize? This is why I don't like adding something too open to change to an encyclopaedia, both statements "Harper is the leader of the Conservatives", and "the Liberals will hold a leadership election before the federal election", can be proven wrong. 117Avenue (talk) 04:46, 1 September 2011 (UTC)

I think the article should continuously reflect the facts as they are "now" and thus list interim leaders, noting them as such and replacing them as new leaders are elected. That way we are ready for a snap election on this page too! - Ahunt (talk) 11:57, 1 September 2011 (UTC)

The consensus at Talk:Bloc Québécois#Party leader versus party president versus parliamentary leader is that the BQ currently don't have a leader. We shouldn't contradict that. 117Avenue (talk) 22:25, 3 September 2011 (UTC)

Talk about being late, jeepers. I agree with those, who say we shouldn't be showing party leaders in the infobox. GoodDay (talk) 06:05, 9 November 2011 (UTC)

Yes, most people vote for a party, not a leader. 117Avenue (talk) 14:28, 9 November 2011 (UTC)

I don't think we should include the leaders. If we do include them, we should only include the leaders who will still be leader on the predicted election day (barring their resignation or death). That is, we shouldn't include the leaders of parties that have a party convention between now and election day. —Arctic Gnome (talk • contribs) 22:17, 7 December 2011 (UTC)

This has been talked about a lot and members feel that leaders should be included. Seeing an election could technically be called tomorrow we can't just assume that because there are upcoming leadership conventions that the interim leader will not be the leader in the next election. I don't see the sense of having any pictures there till possibly 1 year to 6 months before the election is expected to be called but others do not feel that way. Newfoundlander&Labradorian (talk) 02:01, 8 December 2011 (UTC)

Are we now seeing support to remove the image, leader, and since date, from the infobox, until 2015? 117Avenue (talk) 15:59, 16 January 2012 (UTC)

Well I for one support keeping the current leaders and just updating it as more information becomes available. - Ahunt (talk) 16:05, 16 January 2012 (UTC)

I don't really care either way anymore. An election could potentially be called tomorrow and all these people could serve as leaders in the election, obviously there's a 99.9% chance that will not happen. I'm fine with waiting till the writ for the election drops to fill-in the leaders info. Newfoundlander&Labradorian (talk) 02:30, 17 January 2012 (UTC)

This is an important detail. While we don't expect an election before 2015, there could be an election at any time. If an election were unexpectedly called in late 2012, Bob Rae would likely lead the Liberals into it. In Canada, elections are always imminent. --Llewdor (talk) 18:31, 31 January 2012 (UTC)

And if an election is called, the leaders can be added. That doesn't mean we need them now. 117Avenue (talk) 23:23, 31 January 2012 (UTC)

It's accurate now. If the leaders change, we can change them. Removing them now means that when an election is called the article will be incomplete until we fix it. This article describes a future event; of course its content is in flux. Leaving out crucial information like the party leaders is a cure far worse than the disease. --Llewdor (talk) 18:37, 3 February 2012 (UTC)

I agree fully with Llewdor. Wikipedia is not made of stone. If leadership changes, then we will account for that. But for now, it is accurate, so let's not throw out the baby with the bath water. Bkissin (talk) 22:18, 3 February 2012 (UTC)

Party leaders are not crucial information, others have said this above. This article will never be out of date, whenever there is news I see this page updated immediately. And Bkissin, are you hearing yourself? "But for now, it is accurate," yes, but you both admit it can quickly become outdated, if it isn't updated. WP:CRYSTAL doesn't say throw out the bath water, or spread a disease, it just says to be careful about what is said, so that we aren't inaccurately predicting the future. (To keep the analogy going) let's keep the bath water, but make sure we put the right baby in, at the right time. 117Avenue (talk) 03:35, 4 February 2012 (UTC)

2013

From what I understood of the prior consensus, is that leaders are listed in the infobox, as they will be on election day, if the election was called today. If an election was called today, we don't know who the leader of the Liberal Party would be in five or six weeks. Listing Rae is a CRYSTAL violation. 117Avenue (talk) 04:54, 13 March 2013 (UTC)

Removing him and claiming that the winner of the leadership election would be the leader is a CRYSTAL violation because we don't know that the party would continue with a leadership election in the middle of an actual election. Frankly, that sounds pretty far-fetched. -Rrius (talk) 05:13, 13 March 2013 (UTC)

I agree, which is why I didn't say that. I said "we don't know who the leader of the Liberal Party would be in five or six weeks." 117Avenue (talk) 02:39, 14 March 2013 (UTC)

We also don't know who the leader of the Conservatives or NDP will be in 6 weeks. :) Newfoundlander&Labradorian (talk) 12:23, 14 March 2013 (UTC)

Yup. 117Avenue (talk) 04:31, 15 March 2013 (UTC)

You may not have said it here, but that was what your edit did. Since you claim both are CRYSTAL violations and you don't have consensus to drop party leaders from the infobox, we have to fall back on the reason you proffered for your edit removing this one. Your grounds for removing Rae, as stated in your edit summary, was that we know Rae won't be the leader if an election is called. That isn't true. If an election were called today, it is hard to see how the contest could continue. So if we are to apply the test you suggest—who would be the leader if the election were called today—your edit should not stand. I don't really think that should be the test. In a fixed-term, majority Parliament (even of the loosey-goosey Canadian form of fixed terms), I think we should use the current leader unless they have announced a resignation effective before the election or an interim leader who's set to be replaced by a convention before the next election. -Rrius (talk) 05:53, 15 March 2013 (UTC)

Which would be a CRYSTAL violation. There shouldn't be any leaders until the election is called, because we just don't know. 117Avenue (talk) 00:08, 16 March 2013 (UTC)

Agreed, but we've been through this before. Newfoundlander&Labradorian (talk) 00:10, 16 March 2013 (UTC)

I'm going to be the bad guy, and start this discussion again. Including any leaders in the infobox, is speculation. This article is about a future event, not the current state of government in Canada. 117Avenue (talk) 01:53, 10 June 2013 (UTC)

I am a supporter of removing leaders till closer to the election, how close is the question. The dropping of the writ? Newfoundlander&Labradorian (talk) 02:09, 10 June 2013 (UTC)

I'm not. We should be assuming that each leader will stay at the helm for the next fifty years until indicated otherwise. Interim leaders included, seeing as the writ could have been dropped on the advice of the Prime Minister at any time Rae was acting. Contrary to what 117 says, it's not CRYSTAL to assume this; rather, it is CRYSTAL to assume otherwise. _EricLeb⁰¹ (Page | Talk) 03:23, 10 June 2013 (UTC)

You aren't speculating when you say Paillé will live to be 113? Exclusion is not an assumption of resignation. Not saying something, isn't looking into a crystal ball. As you are aware of WP:BLPREMOVE, WP:RS, WP:V, etc., if we can't verify it, we don't say it, likewise with WP:FUTURE, if we can't predict it, we don't say it. 117Avenue (talk) 04:54, 10 June 2013 (UTC)

No, because the default position here is that Paillé is the current leader of the BQ and that's all we know for certain. The point of that hyperbole was to clarify that argument. Now, while FUTURE seems like an enticing reason to take down the images, we have to give our readers some credit: none of them will actually believe that we are guaranteeing that they stay on for two more years. This is also why we are displaying the date of the election in the article, even though it could be called earlier, and why we generally post dates of events (such as the Sochi Olympics), even though they may be postponed, and why we generally create articles for marathons, even though they might not finish. Prince Charles will become King of the United Kingdom until indicated otherwise, and that isn't a false statement. Essentially, the encyclopedic value of this information substantially trumps the off-chance that a reader thinks our prediction is set in stone. _EricLeb⁰¹ (Page | Talk) 06:03, 12 June 2013 (UTC)

In order to comply with FUTURE, all future events must be prefaced with "expected" or "scheduled". "Prince Charles will become King of the United Kingdom until indicated otherwise" is a sentence that violates FUTURE, it must be prefaced with "expected", or you can simply say he is the next in line for the throne. 117Avenue (talk) 02:57, 13 June 2013 (UTC)

So why don't we find a way to specify that instead of blatantly removing important and convenient information from the infobox? Better yet, why don't we add a note reminding readers that the leadership information is "current as of this date"? Or much more simply, why should we sacrifice conveying important information over an unnecessary policy? (By unnecessary, I'm coming back to the fact that the reader should be given some credit.) _EricLeb⁰¹ (Page | Talk) 02:58, 5 July 2013 (UTC)

Do you mean like having a section titled "Current standings" with a list of major parties and their leaders? 117Avenue (talk) 03:20, 5 July 2013 (UTC)

No, more like having a small note added to the existing infobox stating that the leadership information is current as of the last edit. Also, looking at the actual infobox, there is a parameter for whether or not this is an "ongoing" election. Putting it that way, the whole run-up to the election could be considered the "election cycle", which is what the articles are currently covering. Under that definition, having current leaders displayed would make sense and not violate FUTURE. _EricLeb⁰¹ (Page | Talk) 03:31, 5 July 2013 (UTC)

I think you are forgetting that the infobox is a summary of the article, not the article itself. What's leading up to the election can be in the prose of the article. The infobox should only show what is known. 117Avenue (talk) 05:39, 5 July 2013 (UTC)

Since this issue has come up several times across multiple articles (see this for instance) I think we need to take this to a suitable place to get a consensus on how to deal with it on a wider basis so all articles will have the same standards applied. Would Wikipedia talk:WikiProject Governments of Canada be the place to take this for wider agreement? - Ahunt (talk) 13:26, 10 June 2013 (UTC)

Agreed. We may want to bring in other discussions from other Westminster parliamentary systems such as the Next United Kingdom general election, Australian editors, as well as New Zealand pages, which also uses a similar style on Wiki as to the one used by Canadian poli wiki. Perhaps a more appropriate place to gain consensus for Wiki as a whole can be found at Wikipedia talk:WikiProject Elections and Referendums. Krazytea(talk) 19:54, 10 June 2013 (UTC)

I should add that it is my impression to include leader data in the infobox. I am not actually 100% sure that WP:Crystal is even violated by the written sense of the data which seems to imply that Wiki is not in the business of creating events that are not yet in cycle such as the given examples as ones that are positive such as the 2016 U.S. Presidential election (which does include leaders because there are obviously no elected leaders in place with regards to the primaries yet) and the 2020 Summer Olympics. Events such as 2028 Summer Olympics are not appropriate under WP:Crystal. But there doesn't seem to be any information that says too much about the information included in such articles.

Given that in the Canadian parliamentary system that party leadership elections/conventions determine the leadership of the party in the next election this is really not different from a primary situation in the U.S. From this perspective situations such as Jack Layton's death are highly unusual events. Leaders are essentially elected (except when appointed to interim positions) that they will be the leader of the next election and are generally replaced after losing elections when the party is ready for a new leader and new path into the next election. I guess this essentially comes down to a debate on what is a leader of a party in Canadian politics, are they there to lead the party into the next election or is it a fragile position. I would argue that once elected these leaders are in fact the leader for the next election (give that we do not know the exact date, an election can take place at any time). More importantly it provides important information particularly for readers from outside Canada who have little information as to the political situation going into the next election. Krazytea(talk) 03:54, 11 June 2013 (UTC)

We're not removing their names from the article, the leaders are still mentioned in the prose, and the party standings table. But this is not an article on the current state of government, readers looking for that information can go to several other articles. It isn't our job to predict the stability of politicians, and the government; for the same reason that this article isn't at Canadian federal election, 2015, we can't say that the election will be early either. 117Avenue (talk) 04:05, 11 June 2013 (UTC)

So that's it, can't convince you that you are violating a policy? 117Avenue (talk) 02:33, 25 June 2013 (UTC)

Krazytea makes an interesting point above that makes me think of this hypothetical situation. If Wikipedia were to have existed in 1972, would it have been a CRYSTAL violation to have listed Thomas Eagleton as the Democratic running mate on the United States presidential election, 1972 article? Of course it would not have been. He had been elected by his party's delegates to be their vice-presidential candidate and it would have been very reasonable to assume he would have been the candidate in the ensuing election. Similarly, party leaders are elected in Canada with the explicit purpose of leading their parties into elections and to form governments when their parties command the confidence of the house. Once a party leader has been elected, we have to assume that they will lead their party into the next election until it is proved otherwise, just as we would do in an American election article after a candidate becomes the presumptive nominee (which, by the way, is what is done; Wikipedia does not wait for the formal nomination and that passes the test as not being a CRYSTAL violation). - Nbpolitico (talk) 13:03, 5 July 2013 (UTC)

Party leaders must routinely go through leadership reviews, it isn't seen if the party members wants their leader to go into an election until the year before the election is expected. 117Avenue (talk) 05:32, 7 July 2013 (UTC)

Please provide a source for this assertion. My understanding is that all or most of the major parties in Canada only require a leadership review at the first meeting after an election lost, with either doesn't apply or has already occurred for all of the leaders of Canada's major parties. - Nbpolitico (talk) 18:13, 7 July 2013 (UTC)

You're right, they're typically after elections, I didn't realize they were scheduled that way. Thanks, 117Avenue (talk) 02:41, 8 July 2013 (UTC)

I'd like to repeat the point that the article already covers the entire run-up to the actual 2015 election, therefore it isn't exactly ill-conceived to feature the leaders present during said run-up. As long as we aren't explicitly mentioning that the leaders are expected to be on the ballot, it should be assumed that – as is the case with the rest of the article – we are dealing with the ongoing "election campaign" from Election 2011 to Election 2015. _EricLeb⁰¹ (Page | Talk) 04:28, 6 July 2013 (UTC)

The article may cover the run-up to the election, as apart of the history of the election, but the infobox is about the election itself. 117Avenue (talk) 05:32, 7 July 2013 (UTC)

Then, as I mentioned several times, let's make it about the run-up by adding a note to it. _EricLeb⁰¹ (Page | Talk) 21:48, 7 July 2013 (UTC)

We don't add additional disclaimers to articles, if we don't know something, we don't say it, this is what WP:CRYSTAL outlines. 117Avenue (talk) 02:41, 8 July 2013 (UTC)

Ugh, those leaders are leading now. How is that a prediction? Read what I said once more... _EricLeb⁰¹ (Page | Talk) 06:00, 8 July 2013 (UTC)

We've gone to repeating ourselves. This article is about the election, not the current state of government, go to 41st Canadian Parliament. 117Avenue (talk) 02:52, 10 July 2013 (UTC)

But the bulk of the article deals with the current state of government... if we want to be proper, 41st Canadian Parliament should be linked to from the Timeline section on this page. _EricLeb⁰¹ (Page | Talk) 03:06, 14 July 2013 (UTC)

Good idea. 117Avenue (talk) 05:42, 19 July 2013 (UTC)

Margins of error for polls

I'm growing a bit leery of lack of consistency in the margins of error reported by the various pollsters. Many of them appear to give the margin of error for the total sample, even though most often (if not always) the published results reflect only the decided/leaning respondents. It seems Nanos is the only pollster that explicitly reports the margin of error for *both* the total sample and the subsample of decided/leaning respondents upon which the results are based. Whenever I've posted Nanos results here, I've always gone with the latter margin of error because it is in a way "truer", however it's clearly inconsistent with the pollsters who give the margin of error for the total sample. Then there are other cases where it simply isn't specified whether the margin of error is for the total sample or for the subsample of decided respondents. Finally, you have Abacus Data who in recent polls has included the following passage in their methodology: "Since the online survey was not a random, probability based sample, a margin of error cannot be calculated. The margin of error for a survey of 1,004 respondents using a probability sample is +/- 3.1%, 19 times out of 20." Any clever ideas out there for addressing this issue? Dare I suggest that if consistency can't be assured, the margin of error column be removed from the table altogether? — Preceding unsigned comment added by Undermedia (talk • contribs) 20:27, 26 January 2012 (UTC)

It is an interesting problem, but then we have seen that none of the polling data is really comparable between polling companies either. All of their methods vary and all of their results vary and, when compared to the election results, none are really very accurate anyway. I sympathize with your concept that the margin of error numbers are not comparable and perhaps we should then omit them, but given the fact that nothing in the table is comparable I think it would be best to leave it as is and let the readers read the sources if they want the fine detail. As you may recall in the last election campaign we had a graph that drew lines for each company, just comparing their own results to their past results. That is really the only way to do it, but in the period in between elections it really produces too sparse a data set. While I agree you have an excellent point I think we should just accept the differences, note what we can and leave it. - Ahunt (talk) 23:49, 26 January 2012 (UTC)

Clearly there's some ongoing wrangling over whether the "pseudo-margin of error" for Abacus Data polls ought to be listed in the table or not. Personally, I could lean either way; what I like about including it is that it at least gives a sense of the sample size. I've been thinking we could consider replacing the MOE column with a straightforward Sample Size column, therefore avoiding this dilemma as well as the various other inconsistencies in reported MOE among the different pollsters outlined above. I would suggest reporting the total sample size (i.e. before elimination of undecided voters) as this number tends to be mentioned in just about all polls. Any support for/thoughts about this? Undermedia (talk) 15:33, 24 May 2012 (UTC)

Or we could mark the Abacus polls as "+/-3.1 (eqiv)" or something similar. - Ahunt (talk) 15:35, 24 May 2012 (UTC)

That's a fair suggestion. However, another thing I've noticed as a follower of ThreeHundredEight.com (which often seems to obtain more detailed info on polls than what is reported by the media) is that Abacus may not be the only pollster using non-random samples. For example, check out the first graph in the article, http://www.threehundredeight.blogspot.ca/2012/05/federal-ndp-closes-gap-in-ontario.html, showing the latest Ipsos Reid poll. Notice at the bottom it says "Random sample of comparable size would have MOE of +/- 2.0%", suggesting that this poll doesn't have a true margin of error either, a detail that wasn't mentioned however in the media article (and unfortunately original Ipsos reports don't tend to be freely available). Also, there's still the issue of certain pollsters reporting the MOE for the total sample and others for the reduced sample of decided voters only, which still slightly irks me. It's all in all a tricky issue for sure. Undermedia (talk) 15:59, 24 May 2012 (UTC)

Well we can only report what the refs indicate! - Ahunt (talk) 16:02, 24 May 2012 (UTC)

True. Shall we go with your above idea then? I'm sensing you're not crazy about the sample size column (we could alternatively add it while keeping the MOE column as well)? Undermedia (talk) 16:11, 24 May 2012 (UTC)

I am just not sure that a sample size will be as useful or as familiar to readers as a margin of error. Sure I think using "(equiv)" is probably as good as we can get. Do you want to do that, then? - Ahunt (talk) 16:38, 24 May 2012 (UTC)

Done! The 3 oldest Abacus poll reports don't actually state that a true MOE couldn't be calculated due to lack of a random probability-based sample, as is the case with the more recent ones, so I didn't put "equiv" for those. Undermedia (talk) 17:00, 24 May 2012 (UTC)

Looks good. At least we have a process for future ones now. - Ahunt (talk) 18:23, 24 May 2012 (UTC)

Assigning any margin of error of any sort to non-representative polling is misleading and statistically invalid Use of "equivalent to" does not work and is specifically not allowed in the US under the standards of practice of the AAPOR. We in Canada have no body giving any guidance to how pollsters should report the results. Keeping in an equiv statement makes it look like there can be comparisons made across very different methodologies. We already have enough problems in Canada with low quality public opinion polling, somewhere the line has to be drawn and people need to aware that there differences between a random public opinion poll and an opt in online panel. Bernard (talk) 18:54, 28 September 2012 (UTC)

I fully sympathize with your concerns, although as discussed above the main goal of the MOE column in this article is to give readers some sort of idea what the sample size/statistical accuracy of the polls are. At one point I suggested replacing the MOE column simply with a sample size column, but it was argued that MOE is easier to directly relate to the polling results. Another concern I have is that whereas Abacus is "honest" enough to explicitly state in its methodology overview that a MOE cannot technically be calculated, I've noticed other pollsters who also use online panels — e.g. Angus Reid, Ipsos Reid (sometimes), Léger Marketing — do not seem to be so forthcoming about this technicality, and media articles reporting on their polls tend to state a MOE for them without specifying whether it's a "real" one or an "equivalent" one. Should all polls conducted online be stripped of their MOE then? Undermedia (talk) 20:15, 28 September 2012 (UTC)

I also understand your concerns, but as we debated out above it is better to indicate this information to give the readers some guidance as to the equivalent margin of error than to leave it out. There is no doubt from the last election that the polling did not accurately reflect the results, but most of that seems to be from other factors that the pollsters did not account for, like not surveying households that don't have landline telephones, than whether samples are truly random or not. Because participation in these polls is always voluntary and many people either don't have phones, hang up on poll calls or just screen calls and don't answer, truly no polls these days are random samples. The best we can do is give the readers the best information to allow at least theoretical accuracy to be compared. - Ahunt (talk) 22:54, 28 September 2012 (UTC)

In order to clean up the table a little bit with regard to MOE, how would you folks feel about adding a couple of footnotes at the bottom? First, I'm thinking a superscript "1" could be added next to the column title "Margin of Error" linking to a first footnote explaining that the shown margins of error are from the info provided by the pollsters or the media articles reporting on the polls and that they represent the MOE of the total sample of respondents before distribution of undecideds. This brings me back to a point I raised some time ago and that still irks me, which is that for Nanos polls we've been listing the MOE for the reduced sample after elimination of undecideds (as Nanos is nice enough to provide this info in its reports), whereas I'm quite sure that for all other pollsters only the MOE for the original total sample is given (this is certainly the impression one gets from reading the reports). For the sake of consistency, I would therefore suggest changing the MOE for all Nanos polls to the total sample MOE (the info is easily retrievable from the linked reports), and throwing in the footnote to set the record straight from now on. Second, rather than adding "equiv" next to the MOE of Abacus polls, which is somewhat unsightly and unclear, how about adding a subtle superscript "2" that links to second footnote explaining that the survey was not done with a random probability-based sample and that the shown MOE is what it would have been had the sample been random probability-based? This would make it clearer for readers as well as potential editors, and hopefully help to address the concerns of User:‎Bern99. Incidentally, should other pollsters that use online panels reveal that their MOE's are not "real", we could easily point to this same footnote. Undermedia (talk) 14:46, 5 October 2012 (UTC)

Actually, nevermind my comment about Nanos being the only pollster that gives the MOE for the sample after distribution of undecideds as I just realized Environics appears to do the same. Since it thus seems impossible to confidently achieve consistency on this, I'm going to advocate leaving all the MOE's as they are and explaining in the first footnote that the MOE is based on the info provided by the pollsters or media and, when specified, based on the sample size after distribution of undecideds. Undermedia (talk) 14:59, 5 October 2012 (UTC)

I like your suggestion of adding the footnotes - it gives better technical accuracy, without sacrificing the data that we have to work with!

Done! I tried to word the footnotes as accurately and explicitly as I could, but feel free to tweak them if you think they can be improved. Undermedia (talk) 15:38, 6 October 2012 (UTC)

It looks good! - Ahunt (talk) 16:10, 6 October 2012 (UTC)

I'm not an expert, but I'm increasingly getting the impression that all polls conducted online—because they draw respondents from a pre-existing panel which is itself a non-random sample of the overall population—do not technically have a calculable margin of error, even if their reports or media articles don't clearly state this. As a reader of ThreeHundredEight.com where author E. Grenier (regular writer for Globe and Mail, Huffington Post and Hill Times) commonly obtains more detailed information directly from the pollsters than what is reported in media articles or pollster releases, I've noticed that all online polls are presented with the note: "Random sample of comparable size would have MOE of +/- ...%". These include all Abacus Data, Léger Marketing and Angus Reid polls as well as a few internet/telephone "hybrid" polls by Ipsos Reid. I'm therefore tempted to systematically slap footnote #2 on all polls known to have been conducted online, and modify the footnote to explain that all online polls are necessarily based on non-random samples. Any support for this? Undermedia (talk) 16:36, 6 October 2012 (UTC)

That sounds logical and may go some distance to explain why the pooling results in recent elections have not been very accurate. I am just concerned that we ought to be reasonably certain before adding the footnotes to those polls, least we be accused of WP:OR. - Ahunt (talk) 17:02, 6 October 2012 (UTC)

Good point. I therefore propose citing this article by the Marketing Research and Intelligence Association, a not-for-profit organization which basically sets out the rules and standards for Canadian pollsters. Under "Online Polls" it states: "Online survey participants are self-selected, meaning that a margin of sampling error cannot be calculated or quoted for online panel research studies. Using margin of sampling error to describe the accuracy of online polling is misleading, and prohibited under MRIA’s Code of Conduct." I will go ahead with the edit. Undermedia (talk) 18:01, 6 October 2012 (UTC)

Good find! - Ahunt (talk) 18:36, 6 October 2012 (UTC)

I just want to say that I'm generally not convinced that this apparent witch-hunt against online polls being led by some folks is in fact justified. I'm fully aware of the technical criticisms that have been leveled against online polls by various traditional pollsters, academics and other pundits in the field, however, the reality remains that online polling is becoming increasingly prevalent while repeatedly demonstrating its capacity to yield accurate results. For instance, Angus Reid has proven itself to be formidably accurate, outperforming all other pollsters in its final projections for last two federal elections, the 2008 Québec election, the 2009 BC election and the 2011 Manitoba election. Meanwhile, Léger Marketing had the closest final projection for last month's Québec election. Finally, Abacus Data had the second-closest projection for the 2011 Ontario election. IMO, we've reached the point where the record of online polls can no longer be discounted. I'm down with including a "warning" in the opinion polls table on this page to the effect that online polls don't really have a margin of error, but I don't think any further notes aiming to discredit online polls are warranted. Undermedia (talk) 22:16, 17 October 2012 (UTC)

Well the first ref cited says "Virtually all surveys taken seriously by social scientists, policy makers, and the informed media use some form of random or probability sampling, the methods of which are well grounded in statistical theory and the theory of probability." The second ref cited says "Online research polls using convenience samples have, for a number of years already, produced results that have proven to be reliable predictors of outcomes, including election results. Online survey participants are self-selected, meaning that a margin of sampling error cannot be calculated or quoted for online panel research studies." The on-line polls are a form of probability or representational sampling and, as you noted, the results speak for themselves, as they seem to be as accurate, or inaccurate, as any other methods. If you want to split hairs you could point out that none of the existing "random sample polls" actually use random samples, as participants can opt out by not answering the phone or hanging up, thus biasing the data. I think you will soon see the end of random phone surveys as they no longer work in a era where most people screen calls. - Ahunt (talk) 23:31, 17 October 2012 (UTC)

adding a "338" to the "Current standings" section

The Current standings section of the page shows a table. This table is actually created by the transcluded template {{Canadian federal election, 2015}}. This table is not transcluded anywhere else except here.

I would like to alter this table to visually illustrate the fact that there will be 338 members elected in the next federal election, up from the current 308.

• Here is the current version (status quo)

1. Here is idea number 1, call it the full-column version
2. Here is idea number 2, call it the unobtrusive version

However, 117Avenue (talk · contribs) opposes either modification and prefers the status quo. What do you folks think?

Perhaps each person replying could rank their preferences in order, and perhaps also discuss, or suggest a fourth option. -- P.T. Aufrette (talk) 05:44, 4 February 2012 (UTC)

• my preference: 1. unobtrusive 2. full-column 3. status quo. It's useful to show a "338" in the table, in an unobtrusive and non-distracting way. The "338" will be added there eventually, so why not now instead of on election day? -- P.T. Aufrette (talk) 05:52, 4 February 2012 (UTC)

Has the 30 seat extension been made law yet? Newfoundlander&Labradorian (talk) 14:41, 4 February 2012 (UTC)

The law has been passed, yes, it's a done deal (the "Fair Representation Act" became law in December). There will definitely be 338 seats in the next election, though new riding boundaries and so forth have not yet been drawn up. -- P.T. Aufrette (talk) 02:02, 5 February 2012 (UTC)

I thought it was generally accepted not to expand the table with a results column, and changes, until the actual election. This was done on the previous election, [1][2]. This kind of ties into the Leaders section above. The 338 is already mentioned in the article four times, so not having it in the table, isn't losing any information. We don't need to be getting ahead of ourselves. 117Avenue (talk) 23:29, 4 February 2012 (UTC)

Agreed. What possible purpose could it serve to have an empty column for nearly three-and-a-half years? Once we actually have an election, the "current" column will become the "dissolution" column, and we will spawn a candidates column. Then, when we hit polling day, the table will be replaced with the standard results table. Given that, I ask again why we would add this column when it will never be filled? -Rrius (talk) 11:34, 10 February 2012 (UTC)

What about the "unobtrusive version" (idea number 2), linked above? It adds a bit of useful information which takes up very little space. -- P.T. Aufrette (talk) 03:32, 14 February 2012 (UTC)

It has the same problem. The third column adds nothing, and will never be filled in. The only bit of information the unobtrusive version adds is that there will be 30 more seats in the Commons in the next Parliament, which is not something the table needs to tell us. If we get notional seat totals from a reliable source, then we can re-evaluate having a third column. -Rrius (talk) 03:52, 14 February 2012 (UTC)

Sorry, I don't understand that last sentence... "notional seat totals" of what exactly? -- P.T. Aufrette (talk) 04:12, 14 February 2012 (UTC)

In some places, I don't know about Canada, when a seat redistribution occurs, someone studies the new seats and tries to estimate who would win each of the seats. Each seat is referred to as notionally belonging to that party, and this yields notional seat totals for each party. -Rrius (talk) 04:21, 14 February 2012 (UTC)

OK, I have read the section below and have responded there. Regarding the "338", this is not a "third column" anymore, it is almost a footnote of sorts, only alongside rather than beneath. -- P.T. Aufrette (talk) 05:35, 14 February 2012 (UTC)

Such a footnote is not really necessary since the only point of the table is to show what happened at the last election and what the state of the parties is now. Putting a new cell to the right of "308" is even more pointless. Despite what you said, there is a whole column above the "338" note, meaning there really is a third column that is pointless and ugly. The increase in seats is already noted elsewhere in the article, and since the table has nothing to do with the number of seats to be contested at the election, there is no reason for the increase to be noted there. Also, I have responded to you in the section below. Basically, you misunderstand the point of notional calculations: it is about what would have happened at the last election had there been 338 seats, not predicting who will win next time. Polls don't come into it at all. -Rrius (talk) 19:54, 14 February 2012 (UTC)

The title of this article is the "42nd Canadian federal election", not the "Current party standings in the House of Commons". Although the table is a template, it is transcluded nowhere else, so it is effectively part of this article. It is, in effect, a table about the next election, so it seems strange to exclude from it the one bit of absolutely certain information we already have about that election. But let's drop the subject. -- P.T. Aufrette (talk) 00:33, 17 February 2012 (UTC)

That information is already in the article, and adding to the table provides no tangible benefit. Even if it were to be added, it would make a hell of a lot more sense to add it as a footnote placed after "308" rather than adding a whole blank column just to cram the information into the bottom row. In the end, you proposed a change that doesn't need to be made. The bottom row is for totals, and the territory above the proposed "338" cell will remain empty because a completely different table will be used when we have results. So I go back to what 117Avenue said, we have the fact of the seat increase in four other places in the article, so we aren't losing anything by not having it in the table. I would go further and say it adds nothing to include it since the table has precisely nothing at this point to do with the seats that will be occupied after the next election. It is about party strength, which is a valuable thing to have in an election article. (I would point to the fact that virtually every past legislative election article includes the number of seats held before and after, and ever future legislative election article has the present or last election seat totals or both. It would be no good to point to the infobox because it doesn't reflect independents. -Rrius (talk) 01:48, 17 February 2012 (UTC)

Notional seats?

After UK boundary reviews, well-respected reports come out basically saying for each seat that if the last election had been held with these boundaries, that party would have won it. Does anyone draw up similar notional results for Canada's federal boundary reviews? -Rrius (talk) 11:38, 10 February 2012 (UTC)

I don't think we'll know until the media outlets react to the maps later this year. 117Avenue (talk) 03:05, 11 February 2012 (UTC)

In the UK, this a regular thing, so I'm asking if this is something we would normally expect to see at some point. -Rrius (talk) 03:12, 11 February 2012 (UTC)

That sounds quite farfetched for Canada. For a significant fraction of seats, it's not even possible to accurately predict the winner a mere month before election day, even with long-frozen riding boundaries. Even on election day itself, you would do well to accurately predict seat totals for each political party with an error of less than plus or minus 10. For the most part that seems to be because a large percentage of the electorate itself doesn't really tune in and give a lot of thought to it until the actual election call, which is little more than a month before election day. The NDP surge of 2011, for instance, is merely an extreme example of this: it was not reflected in polls until after the election campaign began, because it hadn't happened yet. Published polls tend to be on a national or provincial level; any riding-by-riding polls, to the extent they even exist, would be privately commissioned by individual political parties and the results closely guarded. These notional seat totals might perhaps make more sense for the United States, where 1) there are only two parties 2) district boundaries are often gerrymandered to guarantee a clearcut win by one or the other party 3) election occur at fixed, predictable times, which means the electorate has often made up their minds a year or more in advance, unlike Canada's short 35-day election campaigns 4) electoral districts for the Electoral College or the House of Representatives contain 600,000 voters on average instead of 100,000 on average for Canadian ridings, so there is less randomness and unpredictability from statistical fluctuations. But for Canada, these notional seat totals would be little more than somebody's educated guess, with any two guesses unlikely to agree. I am curious how the UK handles such discrepancies, or is there only one organization that calculates these notional seat totals? -- P.T. Aufrette (talk) 05:36, 14 February 2012 (UTC)

I think you misunderstand. It is not about predicting what will happen at the next election; rather, it is an attempt to estimate who would have won the seat at the last election had it been contested under the new seats. I'm not sure exactly what goes into it, but I imagine it is some combination of the last election's results and the partisanship of the local government wards that are affected by the change. -Rrius (talk) 19:48, 14 February 2012 (UTC)

Also, you keep acting as though I am talking about something that it is done in the US, but if you go back and read what I wrote, I never mentioned the US, where no one does this. What I mentioned were notional seat studies in Britain, which is like Canada in each of the four ways you try to point out as somehow debunking the possibility of notional calculations. In fact, items three and four are even more pronounced in Britain. -Rrius (talk) 19:58, 14 February 2012 (UTC)

No, I realized you were talking about the UK. If such notional seat studies are done at all in Canada, they must be fairly obscure. Since there is little or no predictive value, it would be of questionable relevance for this article; it could perhaps go into Canadian federal election, 2011, but you could just as well do a "what-if" in the opposite direction to see what the 2011 results would have been with the 301-member or 295-member riding boundaries of years past. It is not particularly useful information. -- P.T. Aufrette (talk) 00:51, 17 February 2012 (UTC)

You still don't seem to grasp the point. Its predictive value is in providing a place to make predictions from. It is certainly not something that would go at the 2011 election article because its only value is for the next election. You can't really speak intelligently about how many seats any given party needs to gain or what kind of swing is necessary for a majority without translating the 2011 results into the 2015 seats. That is what the notional seat studies do for you. There is no benefit in trying to translate backward, but there is tremendous benefit in being able to turn what happened in 2011 into something meaningful for the new seats. -Rrius (talk) 01:39, 17 February 2012 (UTC)

It would not be to difficult at all to see which party would win the new seats if they had existed in the previous election. You only need to look at individuals polls from the former ridings that the new ridings were created from. For the most part the polls won't change that much, you can currently find riding maps that show how each poll voted. Newfoundlander&Labradorian (talk) 01:42, 17 February 2012 (UTC)

Except that you can't assume a marginal NDP seat (based on the 2011 election) is uniform in its partisanship. Anyway, if you are adding seats, you probably are splitting 30 seats in half anyway. A new riding might take pieces of two, three, or even four different seats, some of which might belong to different parties and have wildly different shares for each party. So it isn't as easy as you make out. Anyway, since you guys don't seem to know about the concept, it probably isn't done for Canada, so can we please stop talking about it? My point in asking about this was to see if it is done, not to discuss whether it has merit. Most people who follow British elections would tell you it does have value, which is why it was used at United Kingdom general election, 2010. So if this useful tool doesn't exist for Canada, it doesn't exist, and I won't bother keeping an eye out after the boundary review is finished. -Rrius (talk) 02:04, 17 February 2012 (UTC)

Thanks for telling us what we are allowed and not allowed to talk about. Canada hasn't added seats to parliament in a while, and I don't know if we've ever added as many as 30 at one time, so it's a bit hard to know what the media is going to do or has done in the past. As well it wouldn't make a difference how many times one riding was broken up if you know where the polls are located you can still get an estimate of what support each party would have in the new ridings. Newfoundlander&Labradorian (talk) 02:46, 17 February 2012 (UTC)

First, I didn't tell you what you are and are not allowed to talk about. I asked if we could stop since my question (which is what started this discussion) has apparently been answered. Talking about whether calculating notional seats is worthwhile is far enough from talking about improvements to the article that I'm not sure I see the value in continuing. Second, it is not easy for a normal person to get access to which polling areas were moved between ridings. Even if they did, the average person wouldn't know what to make of it. That is where having someone actually put the information together comes in. Incidentally, the work is done by academicians, not the media. Finally, it doesn't really matter how long ago Canada last added seats. This is done when boundaries are redrawn, regardless of whether the number of seats has changed. -Rrius (talk) 05:13, 17 February 2012 (UTC)

To answer your original question, the concept of notional seats has not been applied to Canada in any formal sense. I've seen such projections on amateur election prediction websites and the such but nothing is officially drawn up as in the UK. Vale of Glamorgan (talk) 05:31, 17 February 2012 (UTC)

Actually that's not true. Elections Canada used to perform a "transposition of results" after each redistribution. See here for the official Elections Canada transposition after the last boundaries were drawn in 2003: [3]. The reason this was done is because the party that placed first would nominate the returning officer for the riding. That is no longer the case, so I am not sure if Elections Canada will be doing a transposition for the new boundaries that were just finalized in June. Worth monitoring. - Nbpolitico (talk) 17:02, 6 July 2013 (UTC)

After a quick read of the Canada Elections Act, it seems that there will be a transposition of results. Section 33 requires the returning officer to get names "from the registered parties whose candidates finished first and second in the last election" and section 41 says Elections Canada "shall transpose the results from the previous general election to the polling divisions that are in the new electoral district in order to determine which registered parties' candidates have the right to provide the returning officer with lists of persons to be appointed". This would follow the official publishing of the representation order which is expected in September 2013. - Nbpolitico (talk) 17:11, 6 July 2013 (UTC)

Incomplete polling results

An editor recently added a poll that reported only the results for the Conservatives, NDP and Liberals. I removed it as incomplete, but the removal was reverted in favour of a discussion. In past election articles we had a consensus to not include incomplete polls (unless the full data could be found and added) for two reasons: 1. Because they do not report full results they do not adequately contribute to the data or the article. 2. They cannot be added to the graphing of the polling results and therefore the chart and the graph diverge, when they are intended to be displaying the same data. Because of these reasons I propose that incomplete polling results once again be omitted. - Ahunt (talk) 14:28, 13 February 2012 (UTC)

Seconded! Newfoundlander&Labradorian (talk) 16:13, 13 February 2012 (UTC)

Agreed, and your reversion should have stood; it is pretty basic that we allow the status quo to prevail pending discussion. -Rrius (talk) 03:55, 14 February 2012 (UTC)

Another editor has re-instated the polls results, but has found a ref for all the party results, which is great! I have updated the graph as a result. So this specific poll can remain, but I think we can say now that we have a consensus here that only polls that have complete results should be included. - Ahunt (talk) 01:21, 15 February 2012 (UTC)

There were three different editors who re-instated the poll results, and I was one of them. I'll explain my reasoning: For one, the consensus that existed during previous elections is not necessarily transferable to this one, which is why we're having this discussion. We now find ourselves in political conditions where the fourth and fifth parties are flirting with lows that haven't been seen for some time. If this keeps up, we may see a return to polling of only the top three or four parties. I don't yet think this is the case, and it may not happen, but I'm pointing it out because it's not unforeseeable for conditions to return where polling companies only focus on three or four parties. And we must remember that all polls are incomplete. The threshold to track small parties is fluid and ever-changing depending on the current context. This makes any consensus inherently unstable over time. If we begin seeing a return to three- or four-party polling, a solution to the graph issue would be to have dotted lines appear between "gaps" in polling for smaller parties. This is standard practice in statistics, including that of many polling companies. Some other points dealing more with the specificity of this particular poll: I knew the Forum poll had results for the BQ and Greens that would soon be disclosed but harder to find in the mainstream press (indeed, the new source link is a blog of a trusted and reliable researcher who has private access to the full poll results). Knowing that this information would come within a matter of days, I think it is prudent to post polls from mainstream sources that at least have the results of the three main parties, especially when there is reason to believe more complete results are coming shortly from an obscure but reliable source, as they did. This prompts editors to continue searching for the full results, rather than inadvertently omitting polls because editors forget to follow up. I have seen these omissions happen before, which is why I posted the poll before it could be forgotten. We should recognize that this page is tracking events in progress, so there should be room to post the best information at the time, rather than none at all. In the end, the short-term "cost" of having a soon-to-be completed poll up on the page might be worth the benefit of having a more complete polling record in the end when the events before the next election are over. If we want to ensure that incomplete polls don't conflict with the standards of the graph, we can either use the dotted line technique, or we can have a grace period of, say, a week, after which the poll is taken down if the full results aren't found. It took four days for the full results to be found for this poll. Another option is to set up a chart here, on the talk page, that keeps track of incomplete polls so that people can periodically check new sources for more complete results. G(A)IA (talk) 23:43, 16 February 2012 (UTC)

I disagree about using incomplete polls, but I'd rather address the other question. When you wish to include text and it is reverted, the answer is not to add it back and demand a discussion. Rather, the answer is to start the discussion. The way consensus works at Wikipedia is that we do things the way we do them until there is consensus to change it. The onus was on you, not Ahunt, to try to persuade editors of your position. Incidentally, a better alternative for achieving what you wish is to add the incomplete poll, but put it in comment tags so it doesn't display. That way, information whose value we can't be sure of isn't displayed, but it also isn't lost. -Rrius (talk) 00:12, 17 February 2012 (UTC)

We don't half write sentences in wikipedia so why give half a poll? Newfoundlander&Labradorian (talk) 02:03, 17 February 2012 (UTC)

All polls are "incomplete," so the framing of the issue is problematic to begin with. The threshold at which smaller parties are cut off is variable, but some are always cut out, so all polls are "incomplete" to some degree. The poll that was repeatedly re-instated accounted for 90% of voting intentions and was far off from being "half a poll," so please spare the spurious argument--I provided fair reasons that you're welcome to disagree with, but those reasons shouldn't be diminished or misrepresented for the sake of cheap rhetoric. Also, if we really "do things the way we do them until there is consensus to change it," then please note that articles covering Canadian federal elections as recently as 2006 include polls that don't account for the fourth or fifth parties--there's even one that doesn't have the third party. Plenty of current articles tracking provincial polls also do the same thing. So the claim that your preferred way of doing things is *the* way is more than stretching the truth. There's been more than one way that has reached consensus, and we're equally entitled to defend the different ways things have been done legitimately on Wikipedia. Ahunt's original reason for taking down the poll didn't cite a previous consensus--his reason was that readers "can't make use of [the poll]." Many articles show that readers find such polls useful, and that's a significant reason why they're included. I appreciate Ahunt's additional reasons in this discussion, but people shouldn't be faulted for restoring useful information from a poll that was 90% complete, especially given the original reason for deletion. Polls like this that require days to find more complete results have been forgotten and thus omitted in the past. This issue of omission hasn't been fully addressed in this discussion. I'm actually making a case for a more complete set of polls. We're all arguing for "completion" from different angles. The suggestion about comment tags is appreciated and might be a way forward. G(A)IA (talk) 03:22, 17 February 2012 (UTC)

I'm not proposing that we give half a poll. If you comment something out, it doesn't display in the article. The benefit is that the incomplete poll is still there in the code to be filled in if someone can find the data. If it is not there at all, fewer people will know the info is needed. -Rrius (talk) 21:55, 27 February 2012 (UTC)

Despite predictions that we will see incomplete polls and therefore should keep them I haven't seen one yet (the one in question here was found in a more complete version). So I don't see any good arguments for keeping them presented here. Commenting them out until they can be completed is fine with me. I can't graph incomplete polls with data missing with the software I am using to create the graphs. - Ahunt (talk) 00:59, 28 February 2012 (UTC)

With regard to the incomplete Robbins SCE poll 28 March 2012 (field dates actually 21-26 March), it may be worth noting that there used to be a bunch of polls by that firm listed on the 40th British Columbia general election page which were eventually removed for the following reason: "Robbins SCE Research is not recognized by anyone as a valid polling firm". I must say their website does look awfully nonprofessional compared to those of all the other "established" polling firms, the contact information simply lists "Glen P. Robbins", his phone number and email address, and if they didn't bother reporting BQ and Green support in their own original release, I somehow doubt any other source will. Undermedia (talk) 01:12, 30 March 2012 (UTC)

The data gives a generally similar result to the Forum poll of the same time, but I agree that here are problems with the poll report itself. The poll was a very simple three-question survey and that looks okay as far as it goes, but the narrative that follows it draws all kinds of conclusions that cannot be made from the three questions asked. The poll really looks like an excuse to write an editorial. I had to comment it out as the poll did not even report the results for the Greens or Bloc, so it cannot be graphed properly. I would support just removing the commented-out entry on this basis. - Ahunt (talk) 11:12, 30 March 2012 (UTC)

Infobox pictures

• 
  Current picture
• 
  Original picture, my preference

This is in regard to what infoxbox picture we should use for Stephen Harper, and if we should keep it how it is now, or change it back to what it's been up until a couple of days ago. Before changing anything, I think their should be a consensus. Personally, I prefer the second option, which is what we've used for the previous elections he has been leader for. Although the current picture is a more recent picture of Harper, when included in the infobox, it does not look good, mostly due to the white background. I feel that the original picture makes the infobox look better, is still relatively recent, and is a better fit for the article. Thoughts? Tholden28 (talk) 19:27, 24 March 2012 (UTC)

The first picture is two years newer and shows how he looks now. I support using that one, regardless of the background. - Ahunt (talk) 20:15, 24 March 2012 (UTC)

Support the addition of the new picture Krazytea(talk) 21:50, 24 March 2012 (UTC)

Original picture. It isn't distant, its square on the face, it isn't an action shot, and it doesn't have an ugly grey sky. 117Avenue (talk) 03:41, 25 March 2012 (UTC)

• 
  Option 3

I would actually like to suggest a compromise picture. This picture is fairly recent (only a couple months old), it's a frontal shot, and fits into the infobox. It's also fairly close up, and although it is an action shot, I don't think it entirely matters. Tholden28 (talk) 22:55, 2 April 2012 (UTC)

• Original. Current is too low res, and its white background is a problem. Option 3 is problematic because of its open-mouth "pose". The original is far from perfect, he's oddly hunched in it, but it is better than the other two. -Rrius (talk) 02:53, 3 April 2012 (UTC)

I think we need a picture of Daniel Paille. this might be a good one to add = http://www.ledevoir.com/images_galerie/d_109125_87166/image.jpg. — Preceding unsigned comment added by 207.81.36.201 (talk) 21:47, 9 April 2012 (UTC)

That picture would be a good one, but do you know what the license is on it? Tholden28 (talk) 22:10, 9 April 2012 (UTC)

Owned by a newspaper, unusable. 117Avenue (talk) 04:14, 10 April 2012 (UTC)

We could remove the images all together. 117Avenue (talk) 01:48, 10 June 2013 (UTC)

Another new and incomplete poll

Harris Decima has a new poll out as reported in the National Post, but the numbers reported lack the Bloc and Greens. We need to find the complete report to add this, as it isn't up on the Harris Decima website yet (the latest report there is the 30 April polling data). - Ahunt (talk) 12:38, 10 May 2012 (UTC)

User:‎Undermedia - thanks for finding that! I have updated the graph to match the new data. - Ahunt (talk) 15:35, 10 May 2012 (UTC)

Robbins Sce polls

We recently rejected a poll by Robbins Sce Research because it lacked results for two federal parties, although we also noted a few other dubious things about it (see Incomplete polling results above). I have just removed another one which, despite showing results for all parties, still seems dubious. Please see the poll at: http://www.robbinssceresearch.com/polls/poll_905.html. I have never seen a poll by this company reported by any major media outlet in Canada, nor does the linked report mention that the poll was commissioned by/done for any such outlet. As mentioned previously, the report itself and the Robbins Sce website in general come across as strikingly unprofessional compared to all the other pollsters regularly referenced on this page. The commentary is poorly written and contains typos as well as evident partisan undertones (just one example: "No one is happy when the price of crude oil drops and Big Oil decides to increase the price at the pump. Even fewer people are happy when their government refuse to take a stand against apparently deceptive tactics to justify this exorbitant price regime." Who on earth is this commentator Peter Kelly anyway? The mayor of Halifax?). While it is common for opinionated editorialists to comment on polls in media articles, I don't believe it is considered acceptable practice to include such commentary directly in pollsters' reports. The methodology does not state whether the poll was done by phone or online. There appear to be fairly major proportional imbalances in the number of people polled among the provinces (e.g. BC - 921, ON - 742, QC - 142) and I'm not even sure what to make of the statement: "Not all research conducted within Canada". In my view, all of these issues add up to render this poll unreliable and unfit to be listed on this page. Undermedia (talk) 14:54, 27 May 2012 (UTC)

Your objections seem reasonable to me. It certainly looks like a one-man company trying to establish a name for himself and as as you note the methodologies used seem questionable. I think a case has to be made by the editor who wants to include the poll for doing so. - Ahunt (talk) 15:51, 27 May 2012 (UTC)

Ya it seems fairly sketchy to me as well. There are definitely some push polling type questions in there which is not that unusual but I do not see too much legitimacy here at this point. Krazytea(talk) 20:21, 27 May 2012 (UTC)

Bev Oda resigned

http://www.thestar.com/news/canada/politics/article/1220735--international-development-minister-bev-oda-resigns-as-conservative-mp-for-durham

So I'm lowering the Conservatives seat from 164 to 163 in the info box. Bodo3 (talk) 16:59, 3 July 2012 (UTC)

She didn't resign yet. Newfoundlander&Labradorian (talk) 17:33, 3 July 2012 (UTC)

The cited article above says "Her resignation takes effect on July 31", but it seems to be a done deal. - Ahunt (talk) 17:34, 3 July 2012 (UTC)

Her resignation letter says she will finish July 31, therefore she's still a minister and MP. Newfoundlander&Labradorian (talk) 21:45, 3 July 2012 (UTC)

Polling

I added the 2011 election results to the top of the polling graph to make a comparison easier without having to scroll down to bottom and up again. This will only increase as a problem with possibly three more year's of polling. The format is used in most similar articles eg the next elections in Israel and Germany. Rsloch (talk) 14:01, 21 November 2012 (UTC)

I removed it because it is at the top and the bottom, is out of chronological order and isn't reflected that way on the graph. We have never had anyone remark that they need that last election result next to the current polling results, probably because the results of the last election are not really relevant to the current polling results - the last election results are essentially old data and don't have much bearing on current public opinion. There are quite a number of other editors actively watching and contributing to this page, so let's see what the consensus is. - Ahunt (talk) 14:12, 21 November 2012 (UTC)

I second Ahunt. Doesn't really make sense in terms of chronology. Furthermore, the current presentation is consistent with the polling tables featured in the various Canadian provincial election pages, e.g. ON, BC, NS, AB, etc. Undermedia (talk) 19:19, 21 November 2012 (UTC)

I agree with Rsloch, putting the results of the previous election at the top is helpful. They act as a benchmark to compare the recent polling data to. That's why we include both the past and current election results in the infobox. As long as they're differentiated somehow (background colour/italics) I don't see a problem with disturbing the chronological order. TDL (talk) 19:23, 21 November 2012 (UTC)

It is not a comparison table, it is a trend table, and should remain in chronological order. The old results can be found in the infobox, and the relevant article. Also see http://www.youtube.com/watch?v=eM8dphYKigI, polls don't mean anything anymore. 117Avenue (talk) 06:20, 22 November 2012 (UTC)

And the difference would be? It's awfully hard to see a trend without comparing data points. Unfortunately, the table is so long that I only see 4 rows on my mobile device which makes it impossible to separate the trend from the noise. Having the baseline values visible at the top allows readers to see the first order, long-timescale trend of the data. TDL (talk) 07:48, 22 November 2012 (UTC)

That would be what the graph is for. 117Avenue (talk) 03:35, 23 November 2012 (UTC)

You've concluded that polls don't mean anything anymore based on Rick Mercer's rant? Haha, suit yourself I guess! Anyway, the proponents of this change are going to need more supporters, as the status quo is winning at this point. Undermedia (talk) 16:20, 22 November 2012 (UTC)

I've concluded polls don't mean anything anymore after they were predicting the Wildrose would win a majority in April, and then the party didn't come close. Normally I don't agree with Mercer's rants, but this one I did. 117Avenue (talk) 03:35, 23 November 2012 (UTC)

The purpose is to develop a consensus not 'win'. Both sides have presented arguments and I suggest we look for common ground. Is there any way that the results of the last election could be added to the top of the table that would satisfy those opposed? Rsloch (talk) 18:00, 22 November 2012 (UTC)

Sorry it was just a figure of speech. I'm not in fact staunchly against the idea, but unfortunately I can't think of a suitable compromise given the opposing arguments. Can anyone? Undermedia (talk) 18:08, 22 November 2012 (UTC)

Well it occurs to me that one compromise method would be to create a separate one-line table right above the current table with the 2011 election results in that. That way it would be visible, but clearly separate, to eliminate the chronological confusion. If anyone wants to see that I can put it in and if no one likes it it can easily be removed. - Ahunt (talk) 23:54, 22 November 2012 (UTC)

How it is done for the Israel election might be a helpful template. Rsloch (talk) 13:39, 23 November 2012 (UTC)

Check out the change I just made. Like the Israel page it's got the double line separating the election results from the opinion polls, plus the election results are in italics and have a yellow background. Can we live with this? Undermedia (talk) 14:41, 23 November 2012 (UTC)

Looks good to me. Another option would be to put the results of the last election in the same cells as the party name at the very top, like this:

Polling Firm (Results of last election)	Last date of Polling (May 2, 2011)	Link (HTML)	Conservative (39.6)	New Democratic (30.6)	Liberal (18.9)	Bloc Québécois (6.0)	Green (3.9)	Margin of Error[1]

This would clearly separate the election results from polling data. TDL (talk) 17:26, 23 November 2012 (UTC)

Personally I think that's gonna look a bit messier but let's see what the others have to say. Undermedia (talk) 18:56, 23 November 2012 (UTC)

I don't like any of these attempts. Even with the double lines, the top row version is unnecessarily confusing. "Last election? Must be bottom-posted. Wait, no, fuck. Ugh." That is the likely reaction, even for people who have visited the page before. The heading version is unpleasant to look at, and it is not clear a reader would notice the purpose. If we are going to decide that people can't hold numbers in their heads long enough to scroll down, then perhaps someone could make one of those tables where only so many lines display at once. Then the last line (the election line) could be sticky (i.e., always displayed). I haven't got a clue how to make one, and I hope that explanation was clear enough because I haven't got a clue where I've seen one before (only that I have), so I can't link to one. -20:42, 23 November 2012 (UTC)

I tend to agree with 117Avenue that if the primary concern is the ability to discern the trend, the graph is much better for that in the first the place than looking at any combination of rows in the table. Nevertheless I attempted a compromise, but if we can't build more consensus around it I guess it will have to be undone. Undermedia (talk) 21:28, 23 November 2012 (UTC)

While I don't think it is ideal I can live with Undermedia's current version as the least-worst compromise. - Ahunt (talk) 00:27, 24 November 2012 (UTC)

Sorting doesn't work any more. 117Avenue (talk) 04:50, 24 November 2012 (UTC)

 Fixed TDL (talk) 05:30, 24 November 2012 (UTC)

So... I'm not sure what's supposed to happen at this point, but following all these deliberations I'm counting 2 people who are fully in favour of the change, 2 who are fully against it, and 2 more (including myself) who are by default against it but willing to toy around with some potential compromises, which I think we've pretty much exhausted now. Where exactly does this leave us according to the wiki rulebook? Undermedia (talk) 14:53, 24 November 2012 (UTC)

Let me propose this then to get a final consensus. We have a proposal of how to display this information at the top of the table in the article right now. Who objects to this compromise? If no one voices an objection then we have a consensus to leave it as it is. - Ahunt (talk) 15:06, 24 November 2012 (UTC)

But that's not right. It looks terrible, and people have already objected to it. The answer to Undermedia's question is that there was no consensus for change. -Rrius (talk) 20:35, 9 December 2012 (UTC)

If consensus is needed then I would like to add my support. I think it is indeed a very nice comparison to current polling numbers, and at the very least, is more useful to the average reader than having those same numbers at the bottom. Even better, I'd recommend that the election results at the bottom remain, but remove the double lines and highlighting. Treat it as "another poll". If aesthetics are the issue, that can be very easily fixed – I'd much prefer just having a "double" border beneath the row. So the only real complaint is chronology, which I don't personally see as something that impedes comprehension if we label the top row correctly. "Results of last election" would be very precise and would take up just a bit more place than "Innovative Research Group". Either way, separating them visually would be a sufficient move. _EricLeb⁰¹ (Page | Talk) 23:43, 9 December 2012 (UTC)

First, the last election is not "another poll"; highlighting the difference is essential. Having the election twice is problematic because there is no convention on whether polls are top posted or bottom posted, so people will see the election at the top and not think it least recent end of the results, then see it at the bottom, then have to figure out which is end is the recent one. The last election is expected at the chronological beginning, not the end. As I said above, if people here think readers are too brainless to scroll and compare, then make the heading and election results sticky so that when you scroll down through the table the results will always be right there. -Rrius (talk) 03:32, 10 December 2012 (UTC)

We've been treating it as just "another poll" since the article's creation. We highlighted the difference with italics. If you want we can even add a background color. If the issue is not knowing which way order is, we can find a way to make people visually ignore that first row's date, by centring the date cell, or maybe changing the text color. _EricLeb⁰¹ (Page | Talk) 18:49, 10 December 2012 (UTC)

Would it be possible to see an example of what the sticky heading would look like? Undermedia (talk) 15:59, 10 December 2012 (UTC)

I'm guessing a very basic one would look like:

Polling Firm	Last date of Polling	Link	Conservative	New Democratic	Liberal	Bloc Québécois	Green	Margin of Error[1]
!a	!a		-9e99	-9e99	-9e99	-9e99	-9e99
!b2011 Election (May 2, 2011)		HTML	-99839.6	-99830.6	-99818.9	-9986.0	-9983.9
Leger Marketing	December 6, 2012	PDF	35	30	18	7	9	±2.5 pp[2]
Forum Research	November 19, 2012	PDF	33	28	28	6	4	±2.0 pp
Nanos Research	November 15, 2012	PDF	33.8	27.2	29.0	4.9	3.7	±3.5 pp
Abacus Data	November 11, 2012	PDF	36	29	22	7	6	±3.0 pp[2]
~y2011 Election (May 2, 2011)		HTML	99839.6	99830.6	99818.9	9986.0	9983.9
~z		9e99	9e99	9e99	9e99	9e99

First and last rows are unaffected by sort. Except when sorting by date; I have no idea how to set a sort key for that second column I merged (idea courtesy of that Israeli one)... _EricLeb⁰¹ (Page | Talk) 18:51, 10 December 2012 (UTC)

Does that have a "sticky" heading though as Rrius alluded to? That's what I'd like to see. Otherwise, it looks exactly the same as my very first attempt at a consensus change on the main article page, which was subsequently edited a few times by other users before getting removed altogether... so the above example is really just going in a circle back to square one. :S Undermedia (talk) 19:20, 10 December 2012 (UTC)

If by sticky, you mean it always stays at the top, then yeah, that's what it does. If by sticky you mean follows your screen as you scroll down then no, that can't be done with Wiki markup alone. I don't know why you would want that anyway because it would be more irritating than useful. _EricLeb⁰¹ (Page | Talk) 19:52, 10 December 2012 (UTC)

I meant the latter because I think that's what Rrius was suggesting. Not sure what else to add at this point since your above example has basically been tried already and rejected. Undermedia (talk) 13:23, 12 December 2012 (UTC)

My example was not to be evaluated; it was just to demonstrate how a stickies would work. I gave ideas and I'm trying to continue discussion on this because I do want that feature put on again. _EricLeb⁰¹ (Page | Talk) 17:38, 12 December 2012 (UTC)

19 November 2012 Forum poll

I am not seeing the Bloc or Green percentages quoted in the ref cited - where did those numbers come from? - Ahunt (talk) 12:56, 22 November 2012 (UTC)

I expect any time now Forum will release the full report on their website. I'll keep checking. Undermedia (talk) 16:14, 22 November 2012 (UTC)

Yeah I checked earlier and it wasn't there. If it turns up we can add it in and then i will update the graph accordingly. - Ahunt (talk) 16:17, 22 November 2012 (UTC)

3 April 2013 Ipsos poll

I just added this poll to the table, though its validity might be debatable because the article insinuates that Ipsos went ahead and assumed that Trudeau will win the LPC leadership and consequently identified him as leader of the party when prompting respondents to indicate their voting intention. I.e. the questionnaire probably went something like: "If an election were held today, which party would you vote for: the Conservative party led by S. Harper, the NDP led by T. Mulcair, the Liberal party led by J. Trudeau, etc.". Meanwhile other pollsters, such as Forum and Léger, have lately been asking the voting intention question twice in their polls, a first time with no leaders mentioned and a second time in the form of something like "If an election were held today and J. Trudeau were leader of the Liberals, which party would you vote for?". For these polls we've been entering the results to the first question (no leaders mentioned) in the table. So the possible point of debate is whether the Ipsos data is "consistent" with the data we've been entering for the other polls? I'm leaning towards yes because (a) I think it is quite reasonable to assume that Trudeau will be the leader of Liberals at this point and (b) whereas I've always considered the other pollsters' follow-up question emphasizing Trudeau as LPC leader while neglecting to mention any of the other party leaders to border on push-polling, if Ipsos' question indeed mentioned the leader of each party, that seems fair to me. Anyone disagree? Undermedia (talk) 01:59, 5 April 2013 (UTC)

It's a hypothetical poll, like others conducted, and shouldn't be added. Newfoundlander&Labradorian (talk) 03:21, 5 April 2013 (UTC)

It is kind of an awkward one huh. While all things point to a Trudeau victory this poll does not actually have him while he is officially leader. The problem for me is that there are a lot of other polling firms such as Forum and Le Devoir who also had hypotheticals with Trudeau as leader. Do we then have to add those? I could see it getting quite messy with many editing conflicts. Krazytea(talk) 05:47, 5 April 2013 (UTC)

I get that it's a hypothetical poll... even though everyone and their dog has heard at this point that Trudeau is all but certain be LPC leader in a matter of days. But to me what sets it apart from the other ones is the way they ask the question, as I explained above. If the question is "If an election were held today and Trudeau were leader of the LPC, which party would you vote for?" (without ever mentioning the other party leaders, as Forum and Léger have been doing), I consider that a push-poll; I suspect a respondent who's not very politically informed and can't name the party leaders off the top of their head on any given day might hear that question and tend to think "Oh yeah, I've heard of Justin Trudeau, I'd totally vote for the Liberals if he were leader". But now imagine the same respondent being asked a question in which all the party leaders are mentioned, as in the Ipsos poll. In this situation they're at least equitably "reminded" of all the party leaders, and thus are given the opportunity to instead think, for example, "Oh yeah, that Mulcair fellow, he seems decent, I'd vote for him"... Or Elizabeth May, although I suspect the effect would be diminished for Harper because he's probably the only leader that just about everyone knows without needing to be prompted. Undermedia (talk) 14:16, 5 April 2013 (UTC)

UPDATE: Éric Grenier over at threehundredeight.com, also a regular writer for Globe and Mail and Huffington Post, is saying that Darrell Bricker of Ipsos has informed him firsthand that this was not in fact a hypothetical poll and that Trudeau's name was not mentioned in the questionnaire. The false impression that it was results from a miscommunication between Ipsos and the media. See the following post and my subsequent query in the comment section. Undermedia (talk) 15:14, 5 April 2013 (UTC)

UPDATE 2: Grenier's exchange with Bricker is viewable on Twitter. I think this constitutes sufficient proof that there was a miscommunication with the media and the poll was not in fact hypothetical. Undermedia (talk) 15:44, 5 April 2013 (UTC)

Agree it can be added. Just wish Ipsos would publicly release polls so people know what they are actually saying. Krazytea(talk) 16:07, 5 April 2013 (UTC)