Jump to content

Talk:2014 Ontario general election

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Ontario Liberal Party Leader

[edit]

Wondering whether the infobox should list the Ontario Liberal Party leader as TBD? Dalton McGuinty is still the leader of the Ontario Liberal Party and will remain until the new leader is chosen at the convention in January. And since WP:NOTCRYSTALBALL, if an election were to be called before January 26, 2013, the leader of the Ontario Liberal Party would still be Dalton McGuinty. Mr. No Funny Nickname (talk) 13:37, 1 November 2012 (UTC)[reply]

Or how about NOTCRYSTALBALL, in that the article was predicting McGuinty would take the party into the next election? My opinion is no leader until the election is expected to be called. But I've lost that argument. 117Avenue (talk) 02:47, 2 November 2012 (UTC)[reply]
McGuinty should remain in the infobox. He could easily change his mind & remain as party leader. GoodDay (talk) 14:50, 2 November 2012 (UTC)[reply]

Why is Green Party in list of party standings?

[edit]

It did not win any seats in the last election. Neither did Libertarian or Reform. So why is Green listed, but the other FRINGE PARTIES are not? — Preceding unsigned comment added by 208.72.124.147 (talk) 18:59, 27 January 2013 (UTC)[reply]

I would argue that the Greens are in a league slightly above the assortment of so-called fringe parties because they: ran candidates in all 107 ridings last election, unlike any of the latter; got more than twice as many votes as all of the latter combined; consistently poll at or above 5%, whereas the latter consistently total about 1%; and are listed by all of the pollsters, unlike any of the latter. Undermedia (talk) 21:22, 27 January 2013 (UTC)[reply]

I would respond that the inclusion of Greens is a POLITICAL CHOICE that belies the assumption that Wikipedia purports to be politically neutral. The only proper criterion is election to public office. For the same reason, the Canadian TV networks have historically refused to allow parties that have not been elected to public office to participate in the main debate involving party leaders. The fringe parties are relegated to a separate debate (which no one watches) involving only fringe parties. For the same reason, Greens should be excluded from the list of standing parties until they have elected at least one member. — Preceding unsigned comment added by 199.127.109.99 (talk) 22:21, 27 January 2013

I think a better question is, is why is the Green party not included in the infobox? I believe that the current consensus for Canadian election infobox inclusion is a party that contested the election, and won at least one seat, had at least one seat prior to the election, or won at least 5% of the popular vote. For future elections this means a party that plans to contest the election, and has at least one seat or is consistently polling at least 5%. The Green party hasn't polled below 5 since June. 117Avenue (talk) 04:55, 28 January 2013 (UTC)[reply]
The independents only got 0.21% of the vote in the last election. Why no comments about them? 117Avenue (talk) 05:03, 28 January 2013 (UTC)[reply]
Agreed with 117Avenue. In fact, it probably even makes more sense to include them in the infobox than in the standings of the legislative assembly. Is there any pre-existing consensus for the latter? A quick survey of the other upcoming provincial election pages shows a lack of consistency: they're included in ON, PEI and NS but not in any of the other provinces. It might be relevant to note that the provinces in which they've performed best over the years are BC (2001 - 12.4%; 2005 - 9.2%; 2009 - 8.2%; though never any seats) followed by Ontario (8.0% in 2007, though they dropped significantly in 2011). Undermedia (talk) 13:23, 28 January 2013 (UTC)[reply]
I haven't met anyone passionate about the standings table until now, which is why there isn't a consistency there. Some provinces only have four or five parties, so listing all of them doesn't make a huge table, other's it is just simpler to keep it short. 117Avenue (talk) 06:54, 29 January 2013 (UTC)[reply]
From what I remember about this debate with the Ontario general election, 2011. The consensus seemed to be any party that had at least 1 seat or had received at least 5% of the popular vote would be included in the infobox. I believe this is why the Green Party of Ontario is listed in the infobox for the 2007 Election, where they got 8.02% of the vote, but not the 2011 election where they only received 2.94%.Mr. No Funny Nickname (talk) 14:16, 29 January 2013 (UTC)[reply]
And until the results of the election, the latest polls should be used. 117Avenue (talk) 05:07, 8 February 2013 (UTC)[reply]
What has not been discussed is the practical visual effect of adding the Green Party of Ontario. It means that the Ontario Liberal Party Leader, and the Progressive Conservative Party leaders are presented together, above the leader of the Ontario New Democratic Party. On many mobile devices, laptops, the OLP and PC leaders are the only ones seen when the page is loaded, as if they are the main electoral choices with the other two leaders being secondary. The NDP leader is paired with a party that has never held a seat in the Legislature, and received less than 3% in the last Ontario election. The difference between the Libertarian Party of Ontario and the Green Party of Ontario in the 2011 election in popular vote was only 2.49%, whereas the difference between the NDP and Green vote was 19.82%. It is absurd to count public polls that have margin or errors that are plus or minus more than half the Green's perceived support as evidence for their inclusion. The results of the last election where more than 4.3 million people voted should count for more than a survey of 600-1000 people where some regions are represented by under 50 respondents. — Preceding unsigned comment added by 23.91.139.236 (talk) 03:14, 3 June 2014 (UTC)[reply]
Inclusion above visual appeal. 117Avenue (talk) 03:29, 3 June 2014 (UTC)[reply]
Perhaps you should re-read the statement above. This is not a cosmetic issue. It is gives preference to two political parties, and is misleading to pair the NDP leader with the Green leader as if both are a secondary choice. If your evidence is public polls, and they explicitly say their error rate is equal to half or more of the Green's perceived support it cannot be used as the standard for their inclusion. — Preceding unsigned comment added by 206.248.171.220 (talk) 14:43, 3 June 2014 (UTC)[reply]

Libertarians?

[edit]
From assessing the arguments made below, the result should be that the Libertarians should be listed in a separate column. The best argument put forward was 117Avenue's comments about this merely being a list of candidates, meaning that whether the party was minor or major is irrelevant (as opposed to a discussion around a results infobox) and that whether the party was running a full slate of candidates (or close to it) is the more pertinent issue here (which several other editors alluded to). Furthermore, no-one came up with a policy/guideline-based reason for excluding them. Number 57 12:10, 21 June 2014 (UTC)[reply]

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Should the Libertarians be included in the candidate tables as a separate column? Me-123567-Me (talk) 14:51, 11 May 2014 (UTC)[reply]

  • Oppose - In the 2011 election we saw the Libertarian Party, its offshoot the Freedom Party, the Family Coalition Party (which got 2.7% in 1990 despite running in only 68 of 130 ridings, and may do well again now that the Liberals have a pro-choice policy they are enforcing), the new Socialist Party (which will run more candidates this time, and will demand a column if the Libertarians have one), the assertive Kevin Clarke's "People's Party," and nine other more or less serious parties plus four others I've never heard of. Too many columns. General media practice is to show Green Party poll results, but all other are in "others" because they register too low to have statistically significant showings. If the Libertarians or any other party start registering over 1% in the polls, I would give them a column. Until then, there is no reason to prefer them over the other micro-parties.Wilfred Day (talk) 20:25, 11 May 2014 (UTC)[reply]
  • Support - As long as they run in most ridings. Otherwise the "Other" column will get too crowded. In the Quebec election, some of the regions had columns for some of the smaller parties, like the Conservatives- if the ran in every riding in a particular region. -- Earl Andrew - talk 20:43, 11 May 2014 (UTC)[reply]
  • Qualified Support - Agree with Earl Andrew: whatever makes the presentation neater. If the party runs near a full slate, and the Other column is reasonably well populated, adding a column makes the presentation neater. If the Other column is almost empty, or if more than one fringe party runs a full slate, I would not think a separate column would make the presentation neater. I would like this to be a presentation issue, not a political issue. Tunborough (talk) 01:30, 12 May 2014 (UTC)[reply]
  • Oppose - I agree that it makes the presentation neater and cleaner looking and I'd like to support the idea of adding a column for them (or anyone) that runs a full slate of candidates; but, that opens up a real can of worms.  When decisions like this come down to what I think, or what someone else thinks, on a case-by-case basis, it's bound to come to an argument somewhere along the line and someone will be displeased that so-and-so is benefitting from some perceived unfair advantage.  It is for that reason that we come up with rules ahead of time to deal with these situations.  The rule appears to be already established, if they won a seat in the last election or if they won at least 5% of the popular vote in the last election or are consistently polling at 5% or more they are identified separately; otherwise they are lumped together as "Other" until they achieve one of those milestones.  If someone knows of other quantitative measures to base the decision on, those should be discussed; otherwise, if those accurately summarize the precedent, it should be followed on this issue.  As for neatness of the table/article, adding another separate column under "Other", for the party affiliation, would reduce the wrapping & clean it up. — Who R you? Talk 02:34, 12 May 2014 (UTC)[reply]
  • Support - Relegating fringe parties to "other" categories on pages meant to inform simply means that those fringe parties will remain so. It's suppression at its finest, and it comes at no surprise that people with close ties to the Green Party are trying to push the other parties out. This is wrong, folks. Just plain wrong. Livefree.sw (talk) 03:02, 12 May 2014 (UTC) — Livefree.sw (talk • contribs) has made few or no other edits outside this topic. [reply]
  • Support - As long as the information is legitimate and accurate, why omit anything? Is that what wikipedia is about? This said "5%" rule seems to be coming off someone's "memory" and has no real basis for being implemented. Until I see a genuine reason why we should be omitting information, Libertarians being included has my full support. Personal research shows that their membership has more than doubled in the last year alone, individuals affiliated with the Green Party (or any other party) shouldn't be able to dictate which information is included and which isn't - if all the information provided is honest and accurate. Brainstewie (talk) 03:28, 12 May 2014 (UTC) — Brainstewie (talk • contribs) has made few or no other edits outside this topic. [reply]
Nothing is being omitted from the article. The only question here is whether the Libertarians warrant their own standalone column or should be listed under "others"; there is no possibility of them being entirely excluded from the table. Bearcat (talk) 07:52, 12 May 2014 (UTC)[reply]
I support the motion, but that doesn't stop me from pointing out an obvious case of new users being brought in just for this very discussion. Please stop it - your !votes will only get discounted. Redverton (talk) 17:07, 12 May 2014 (UTC)[reply]
Nothing is being "withheld" or "suppressed" from the article. The only question here is whether the Libertarians warrant their own standalone column or should be listed under "others"; there is no possibility of them being entirely excluded from the table. Bearcat (talk) 07:52, 12 May 2014 (UTC)[reply]
  • Support - The Libertarians are running a full slate and so merit their own column. The Green Party doesn't have any seats, and using a % of popular vote as a margin is not a progressive method (will we raise the % next year if the Libertarians gain popularity but still fall behind the Green?). I believe "Other" should remain for fringe parties unable to yield a full slate. Let's leave the omitting and bias to the media and let Wikipedia continue to be a fountain of information for those who seek it. Wikipedia isn't about politics, it's about sharing information. 206.188.67.90 (talk) 03:56, 12 May 2014 (UTC)[reply]
Nothing is being omitted from the article. The only question here is whether the Libertarians warrant their own standalone column or should be listed under "others"; there is no possibility of them being entirely excluded from the table. Bearcat (talk) 07:52, 12 May 2014 (UTC)[reply]
  • Support - Not long ago, the Green party was also considered a "fringe party". They won no seats in the past election. Yet there they are, and there they should remain. If one takes the Libertarian voting rate from the last election, extrapolate it to factor in the fact that they are running in more than double the ridings than they did in the last election, and also factoring in our soar in membership numbers over the last year alone, I feel we should be given our own category. As the election unrolls, the Other column will begin to fill even more, and it will cause layout problems. When it comes down to it, this should not be a political issue. We all have different opinions, and therefore none of us can be considered "unbiased". I believe we should be allowed our own category as long as it does not interrupt the flow of the article or present any inaccurate information. After all, this should be about content, not politics.Mburnison (talk) 04:01, 12 May 2014 (UTC)[reply]
  • Qualified Support for the sake of neatness of the Other column. If they are fielding a candidate in each riding, then why not. But, only if that's the criteria we are applying. By the same token, why are we including Greens? The question should be less about which parties to include, but rather what general criteria we apply to determine the inclusion. --Truther2012 (talk) 13:46, 12 May 2014 (UTC)[reply]
  • Provisional Support At first, my instinct was to oppose. The Libertarians are quite clearly not on the same level as the main four parties, and putting them in their own separate column would invite creating columns for other parties. However, I see that so far they're the only minor party to put forward a near full slate of candidates. To move them to the 'Other' column under that scenario would just clutter. I'm minded to support them having their own column, provided they remain the only minor party to have a full slate of candidates. If another minor party should do so, I can't see being able to justify yet another column, and so the Libertarians should then be moved to 'Other'. Redverton (talk) 17:01, 12 May 2014 (UTC)[reply]
  • Support. The candidates list is not about major and minor parties, it is about listing the candidates, and tables are the cleanest way of doing that. If a party is running candidates in most of the ridings in the section, it should have a column. In Alberta general election, 2012 the parties listed varied, some have the Alberta Party as the fifth party, some have the Evergreen Party as the fifth party, and some don't have a fifth party. It is just about making the information clean and easy to read. 117Avenue (talk) 02:07, 13 May 2014 (UTC)[reply]
    • @117Avenue: - Seems like an WP:OTHERSTUFFEXISTS argument. Plus, where would you have us stop listing candidates? Should we put them ALL in? NickCT (talk) 21:00, 13 May 2014 (UTC)[reply]
      • I'm not saying OTHERSTUFFEXISTS at all. I am saying that table columns exist to make information easier to read. When a row gets too tall, a column should be added. I suggest if a party is overwhelming the "other" column, it should have a column of its own. And the decision should be made on a table by table basis. If you're looking for me to say a number, I say most of the ridings in that section. 117Avenue (talk) 02:20, 14 May 2014 (UTC)[reply]
  • Comment I think we should wait until candidate registration closes to see the actual numbers. Until then all we can do is write in hypotheticals. -- Kndimov (talk) 03:04, 13 May 2014 (UTC)[reply]

Threaded discussion

[edit]

So I saw someone had added all the Libertarian candidates. I removed them as I hadn't seen any consensus about this. I got a nasty complaint on my talk page from the Libertarian who added it. So since he won't do it, I thought I'd start a discussion here about weather or not they should be on the candidates list outside of the 'other' category? Me-123567-Me (talk) 04:13, 11 May 2014 (UTC)[reply]

All registered parties and candidates should be included, regardless of polling or other arbitrary measurements. Keeps it fair and simple. Livefree.sw (talk) 04:22, 11 May 2014 (UTC)[reply]

We are running a full-slate in the upcoming election, and articles such as these are supposed to be free of bias and as livefree said "arbitrary measurements". I don't see why anyone should get to decide which parties should be removed and which ones should be included, especially when he or she appears to make up the rules as he or she goes along.Mburnison (talk) 04:37, 11 May 2014 (UTC)[reply]

I agree with Livefree.sw, all available info should be included. I feel it's irresponsible to remove legitimate information without consulting first. Brainstewie (talk) 04:29, 11 May 2014 (UTC)[reply]

I believe it's fair to allow the Libertarian party our own column, since we are running a full-slate in the upcoming election, and my original edit did not in any way alter or disrupt the format of the page. I made sure of that. What harm is there in allowing the Libertarian party, or any other party running full-slate, to help inform people. I would fight to make sure the Greens get their column. It's about doing the right thing.Mburnison (talk) 04:40, 11 May 2014 (UTC)[reply]

Many existing examples of Libertarians listed in past elections... York West, Ajax—Pickering, Vancouver East to name a few examples. EncyclopediaUpdaticus (talk) 04:47, 11 May 2014 (UTC)[reply]

But did they have their own candidate column in the main election article? Such as Ontario general election, 2011? Me-123567-Me (talk) 04:57, 11 May 2014 (UTC)[reply]
I am learning so much about Wikipedia formatting from you, so thanks for that! As to whether or not a fringe party deserves their own column in the riding listings, we should either make columns for them all or none. The Green party, for example, has not been elected and therefor does not deserve a column, if we are to also exempt the Libertarians, Communists, et cetera from having their own. Livefree.sw (talk) 12:59, 11 May 2014 (UTC)[reply]
We always include data for all candidates in any election, regardless of mainstreamness or fringiness. The only question in the context of the main election results table for the entire election is whether the party has its own column or has its candidates listed in the "others" column instead — but no candidates are ever excluded entirely for any reason. Bearcat (talk) 16:45, 11 May 2014 (UTC)[reply]

Parties grow and change as time goes on. Past elections should not be the judge. The issue here is the deletion of material, not whether or not we should have our own column. I believe we should, but either way our information should not have been deleted.Mburnison (talk) 05:06, 11 May 2014 (UTC)[reply]

Our standard here is that every candidate for every party gets some kind of notation in the election table, with the only relevant question being whether their party has its own standalone column or gets included in the "others" column instead. It was definitely offside for Me-123567-Me to just delete the information entirely — all registered candidates get mentioned in the table with no exceptions. The only valid question here is whether they warrant their own separate column or belong in "others" instead; there's no argument to be made for "they just don't get to be named in the table at all". Bearcat (talk) 16:45, 11 May 2014 (UTC)[reply]
In general I believe what has always been included (in no particular order) is Conservative, NDP, Liberal, and Green. This is because those have typically been the most mainstream parties (probably the wrong word, but the best I could think of). I would normally ask Bearcat this sort of thing, as he has a lot of historic knowledge of this subject, in particular about why things are the way they are. I will post a note at WT:CWNB linking to this discussion.--kelapstick(bainuu) 14:34, 11 May 2014 (UTC)[reply]
Indeed, the issue is that because it's not realistically feasible for us to always give every party in an election its own separate column in the table — the end result of that would be columns so narrow that the candidates' names would have to be read vertically, because the columns would be only one or two characters in width. So we've always had to have some standard for deciding which parties got columns and which ones got restricted to "others". For obvious reasons, the Conservatives, Liberals and NDP always have their own columns going back historically; at least in Ontario, the Greens only graduated from "others" to a column of their own as of 2003.
The standard, for the record, is not purely "mainstream vs. fringe" — you could, for example, formulate an argument that even today the Greens (who have never actually elected an MPP and are still generally in the single digits percentagewise in the popular vote) are still too "fringe" to warrant their own column. (I wouldn't agree with that argument, but it is still possible to argue that.) But the standard hasn't historically been purely "size of candidate slate", either — there have been parties in the past (e.g. the Family Coalition Party of Ontario ran candidates in over 75 per cent of the province's ridings in 2007) which have run slates large enough that you could formulate an argument that they should graduate to "own column" on the basis of how many candidates they ran. But I don't think many people really think that the FCP's role in the 2007 election was prominent enough to really warrant that.
Rather, the past standard has always been one of "relevance" — which is not exactly the same thing as "mainstream vs. fringe", but isn't completely divorced from that either. You could certainly still argue about the mainstreamness or fringeness of the Greens — but I don't think anyone can seriously doubt that they do play a much more relevant and prominent role in election coverage than the Libertarians have to date. It's not impossible for a party to break out from the "fringe" pack and become relevant enough to warrant its own standalone column in the table — the Greens did it, obviously — but a party has generally had to accomplish more than just running a certain arbitrary number of candidates to get there. Bearcat (talk) 16:45, 11 May 2014 (UTC)[reply]

I agree that Wikipedia should do its best to be inclusive and remove bias, but in this case I think it will be difficult to do this. On the page Ontario general election, 2011 (candidates) only the candidates from the Liberal, PC, NDP and Green parties got their own column. The rest were all in the fifth column labelled "Other". If we give the Libertarians their own column then we shall have to do so with the Family Coalition, Freedom, Communist, Northern Ontario Heritage, Special Needs, Reform, Confederation of Regions, Socialist, People's Political Party, Vegan Environmental, Republican, Equal Parenting, Human Rights, Canadians' Choice and Pauper parties. As well we shall need another column after all that for the Independents. That makes 21 columns and is, quite frankly, ridiculous. The model should be Liberal, PC, NDP, Green and Others.

If you are wondering why we draw the line at the Green party and not anywhere else, the answer is quite simple: Only the Libs, PCs, New Democrats and Greens have fielded a full slate of candidates for the past 3 elections and are most likely going to do so again. None of the other fringe parties come close. The Libertarians fielded only 51 candidates in the last general election, less than half the number of Green party candidates. Furthermore, in the last general election the Green Party won 126,021 votes, while ALL of the other fringe parties and independents PUT TOGETHER won only 48,431 votes. The Green party more than doubled their combined vote. The Green party is in an interesting situation because it is clearly more popular than the other fringe parties, but does not have representation, which makes placing it difficult. -- Kndimov (talk) 14:43, 11 May 2014 (UTC)[reply]

On the other hand, if they are actually running a full slate as they claim, a separate column for them would keep things neater, and you could limit it to only parties running a full slate. Heck, it's done by region, so you could have Libertarians in one region where they have a candidate in every riding, and not in another where they don't. Me-123567-Me (talk) 15:05, 11 May 2014 (UTC)[reply]
I despise Libertarians, but I agree that if they run a full slate (or close to it), they need to have their own column. -- Earl Andrew - talk 16:38, 11 May 2014 (UTC)[reply]

I assure you we are running a full-slate. I have proof, but I would rather keep these party documents private. Seeing as how the election was just called, even some of the "major" parties haven't released the names of all of their candidates, I ask that I be able to keep these party documents private until the names of the rest of our candidates are unrolled in the next few days.Mburnison (talk) 03:44, 12 May 2014 (UTC)[reply]

Please keep in mind, that the moment you make a statement as "We are running a full slate" or "I despise X", you should really abstain from editing the article as it is a clear non-NPOV and will only lead to more disputes. Please provide all supporting evidence on this page, and NPOV editors will be more than happy to include it in the article.--Truther2012 (talk) 13:52, 12 May 2014 (UTC)[reply]

Just going to put the idea out there, in the survey discussion it was brought up from a presentation perspective that six parties (plus an "other" column) can fit on to the page without looking too ridiculous. Rather than having a debatable "5% in current polling" (which polls are we going to use?) - why not have a threshold as the top five or six parties based on counts from the last election? It's more definitive and less up for debate than "current polls." That way if the Greens continue to trend down, and the "fringe" parties continue to trend up, it will garner less of a debate as to who should be shown where and why. It will also be helpful in the event that these fringe parties start attracting formerly disinterested voters and we have 10-12 with 5% or more. Brainstewie (talk) 00:05, 13 May 2014 (UTC)[reply]

Obviously we cannot include all parties or there would too many columns and they would not all fit on the page. And most readers are only interested in the main three parties anyway. I suggest we follow what mainstream sources do, which is suggested by WP:WEIGHT. Polling companies report only four parties and relegate the others to "other." Mainstream news covers the campaigns of only the three parties on a regular basis, and the Greens receive some coverage too. The other parties are ignored. So while we could argue whether there should be three or four parties, there is no policy-based reason to have more. TFD (talk) 20:03, 16 May 2014 (UTC)[reply]

Someone appears to have re-added the Libertarians in their own columns. Was there a consensus on that? Me-123567-Me (talk) 21:33, 19 May 2014 (UTC)[reply]

They've been there since 06:02, 11 May 2014 (UTC). 117Avenue (talk) 00:19, 20 May 2014 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Campaign polls: "all voters" vs. "likely voters"

[edit]

Despite having used "likely voters" polling results on several previous Canadian provincial election Wiki pages without any controversy, there is evidently some disagreement this time around, so I figure it's worth discussing. The background context is that given the difficulties Canadian pollsters have experienced in accurately calling a few recent elections with historically low voter turnouts, several of them have started reporting a parallel set of results among "likely voters" in addition to their regular results among all eligible decided voters. Incidentally, this has been standard practice among U.S. pollsters for many years, as voter turnout in U.S. elections has historically been quite low (e.g. all presidential elections <63% since 1960 and as low as 49% in 1996). Canadian pollsters who have adopted this approach now include EKOS, Abacus Data, Ipsos Reid and Angus Reid. Thus, it has become the convention to show these "likely voters" results, when available, on Wiki election pages, including the recent BC (EKOS polls), NS (Abacus Data polls) and QC (EKOS, Ipsos Reid and Angus Reid polls] elections. The "likely voters" results in these elections have proven to come closer to the actual results than their corresponding "all voters" results, or at worst equally close, so it appears to be an overall sound practice. It is also worth noting that the renowned U.S. poll aggregator FiveThirtyEight bases its election projections on "likely voters" polling results, as well as Canadian spinoff ThreeHundredEight. Taking after the latter, I think it makes sense to stick to "all voters" results during pre-campaign polling, since in the absence of an impending election it is of greater relevance to show the mood of the whole population; but switch to "likely voters" (when available) during campaigns when it becomes more relevant to predict the actual election results. So I'm going to switch the 9 May Ipsos Reid poll back to the "likely voters" results, as has been done for the last three Canadian provincial election campaigns, and invite anyone with objections or comments to post them here and we'll work it out. Cheers. -Undermedia (talk) 18:58, 13 May 2014 (UTC)[reply]

Firstly, the single paragraph in the middle of the Ipsos (no longer Ipsos Reid) article, which you repeatedly choose to cherry-pick, is not truly likely voters; it is poll respondents that answered “nothing short of an unforeseen emergency would stop me from getting to the voting booth and casting my vote” (a category that Ipsos says 52% fell into, which is more than the 48% voter turnout of 2011).  It flies in the face of WP:NPOV for a Wiki-editor to pick "other information" reported in the middle of a press release in place of the report's heading and primary reported information.  If there were any doubt (which there really isn't in this case), then the issue must get decided by deferring to what mainstream media report; and even CTV (BellMedia being about as Tory slanted as they get) reports the Ipsos poll in question as PC-37%, Liberal-31%, NDP-28% and ignores the "likely voters" numbers you're touting.  If media outlets were reporting it the way you're trying to, you might have an argument; but at this point they don't seem to be!
Secondly, and more importantly from a WP point of view, you've reverted other peoples edits on this here, here, here, here, here, and most recently here all in the last 4 days, and possibly more before those (I didn't look).  Needless to say, you're beyond WP:3RR.
It's obvious that you feel strongly about this, as is common in political discussions; but at this point you seem to be fighting this particular battle alone.  I'm going to, once again, revert your edit and hope that you leave this issue as it stands or get some others to agree with you here on the talk page and get one of them to change the article.  I suggest that one possibility would be to identify both the overall poll results and (below in smaller print or something) list the "among likely voters" numbers as a footnote to the official poll results; but even that would really only seem reasonable if you can find some situations of mainstream media reporting similar information.
If you're looking to affect voter opinion, WP isn't the place to expend your efforts; if you're looking to try to improve the (unbiased) information provided here, then there's lots of areas your efforts can prove useful, given all the areas of constantly changing information surrounding Ontario's 2014 General Election, let alone the millions of other things going on all over the world every day.  Cheers  :) — Who R you? Talk 05:12, 14 May 2014 (UTC)[reply]
I'm afraid you're significantly overanalyzing my motives here, as the only thing I'm being stubborn about is consistency, and as stated above these exact same "likely voters" (call them what you will) results were used without objection for the Ipsos Reid (yes, still Ipsos Reid, see the PDF press release, or just scroll down to the Methodology, Contact and About sections of the main link) polls conducted during the recent Quebec election, both on the English and French Wiki pages. The full CTV article does in fact mention these results, though granted midway through the article under the heading "Tory voters eager". Meanwhile, I had been reverting edits by (mostly) anonymous users who were not heeding my calls for discussion over the matter, however I appreciate that you've taken the time and effort to debate.
FYI, I've been diligently contributing to and maintaining opinion polls sections on several Canadian election pages (polls + graphs) for over 2 years (notably here, here, here, here, as well as the current page), where you will find that I've by far contributed more edits than any other user and consequently earned recognition for my work, and I strongly reject your insinuation that I'm operating under some sort of political motive. I'll invite a few of the other regular contributors to these pages to weigh in on this discussion, and refrain from reverting this evidently contentious edit any further in the meantime. Cheers. -Undermedia (talk) 17:05, 14 May 2014 (UTC)[reply]
I have worked along with User:Undermedia on a number of polling articles and I have never seen any hint of bias or attempt to promote a political agenda on his or her part, apart from a bias towards accuracy and consistency.
In looking at this issue and its history over several recent elections the "likely to vote" numbers seem to have produced results that are closer to the actual election day results and thus are better data for this purpose. For that reason alone I support using them when available. - Ahunt (talk) 17:47, 14 May 2014 (UTC)[reply]

The opinion polls have been out in past elections, so I don't think they are reliable anymore. I think that we (as a population not just Wikipedia), should not be putting much weight into them. However, the media continues to report on opinion polls, so we (on Wikipedia) must mention them. I don't want to put much of my concentration into this either, so I haven't read this entire discussion to form an opinion. I think I've said in the past that articles should use what is being reported on by the media, to avoid WP:OR. 117Avenue (talk) 03:36, 15 May 2014 (UTC)[reply]

Yes I am not sure best how to proceed here. One of the problems with consistency is that generally the most commonly reported numbers are the general opinion as I understand the matter. Two of the ways we could proceed are either
I. Add a column noting that these are the "most likely voting numbers" or however you want to label it and the other as "general opinion". This can result in updating the current chart where you may have to list one or more numbers for each poll which is kind of problematic.
The second approach here is to create a second list, similar to the Best PM list and report it as the most likely to vote.
Just a few thoughts. Krazytea(talk) 16:57, 15 May 2014 (UTC)[reply]
How about something like this, with a footnote explaining what the numbers in parentheses mean:
 
Polling firm Last date of polling Link Liberal PC NDP Green Type of poll Sample size
Ipsos Reid May 14, 2014 HTML 30 (31) 39 (43) 24 (22) 7 (4) Online 801
Forum Research May 12, 2014 PDF 38 35 21 5 IVR 996
Ipsos Reid May 9, 2014 HTML 31 (28) 37 (42) 28 (27) 4 (2) Online 821
Innovative Research May 6, 2014 PDF 38.8 32.8 19.9 6.5 Online 1,000
Innovative Research May 6, 2014 PDF 38.7 32.8 23.7 4.3 Telephone 500
Oraclepoll Research May 5, 2014 PDF 31 42 25 3 Telephone 1,000
Forum Research May 3, 2014 PDF 33 38 22 6 IVR 1,845
EKOS May 1, 2014 PDF 34.7 31.6 22.2 9.4 IVR 1,576
Election 2011 October 6, 2011 HTML 37.62 35.43 22.73 2.94
-Undermedia (talk) 19:39, 15 May 2014 (UTC)[reply]
The problem with showing both numbers (in the same table with Likely Voter numbers in parentheses) is that it seems to make the table go from informative to confusing (just my gut reaction looking at it the first time).  Krazytea's suggestion of a separate column for Likely Voter numbers was also my next thought, but (from the looks of it), only Ipsos Reid is reporting Likely Voter numbers so the column would be almost entirely blank for the rest of the polling companies.  In looking at the Quebec election polls you provided (thank you), particularly the April 4th Angus Reid poll, it's the exact opposite situation with Angus Reid providing only Likely Voter figures and those Likely Voter numbers being reported in their headline.
In this case, particularly given that only 2 polls (both Ipsos) seem to exist (thus far) offering Likely Voter numbers to cite, what do you think about 2 tables; the existing General Opinion Poll numbers (as they exist in the article), immediately followed by another table of Polls of Likely Voters?
  General Opinion polls
Polling firm Last date of polling Link Liberal PC NDP Green Type of poll Sample size
Ipsos Reid May 14, 2014 HTML 30 39 24 Online 801
Forum Research May 12, 2014 PDF 38 35 21 5 IVR 996
Ipsos Reid May 9, 2014 HTML 31 37 28 Online 821
Innovative Research May 6, 2014 PDF 38.8 32.8 19.9 6.5 Online 1,000
Innovative Research May 6, 2014 PDF 38.7 32.8 23.7 4.3 Telephone 500
Oraclepoll Research May 5, 2014 PDF 31 42 25 3 Telephone 1,000
Forum Research May 3, 2014 PDF 33 38 22 6 IVR 1,845
EKOS May 1, 2014 PDF 34.7 31.6 22.2 9.4 IVR 1,576
Election 2011 October 6, 2011 HTML 37.62 35.43 22.73 2.94
Note: I removed the Green Party from the second Likely Voters table strictly because Ipsos Reid isn't reporting their numbers — Ipsos refers to "Other parties, including Mike Schreiner’s Green Party, would receive…" which adds yet another level of complexity to our process (Oh, joy!).  If we had discernible numbers for Green Party, they should be reported in their own column based on the  polling 5%+  or  at least one elected  test discussed in other threads above.
ALSO: I came up with different numbers for the May 14th Ipsos Likely Voter figures.  These numbers must be verified before this is copied into the article.
I hid the Likely Voters table primarily because it only contained 2 polls and isn't really the standard yet (at least in Ontario, a claim I make only because only 2 polls seem to report the numbers).  It could also just be included without hiding, or perhaps (in future when Likely Voter results are more standard, proven more reliable, and reported by all pollsters) the General Opinion table could be hidden and only Likely Voter shown by default.  Something I consider relatively unimportant at this stage, with the primary question being, ‘What makes the article the best for readers?’
Undermedia — Given your obvious understanding of Likely Voter polling, particularly citable (non-OR proof that such numbers are proving to be more representative of election results and are being more widely used and reported (even if the citations for that happen to be USA or otherwise related)), maybe you could add a section on Likely Voters to the Opinion Poll article to provide readers some insight on the shift towards reporting, and potentially improved reliability of, Opinion Poll#Likely Voters figures.  (And then add the redirect from Likely Voters to Opinion Poll#Likely Voters.)
Your thoughts? or WP:BOLD it! — Cheers — Who R you? Talk 05:59, 16 May 2014 (UTC)[reply]
Ah yes, I had forgotten about how that Angus Reid poll in the QC election reported only likely voters. And the other twist is that one or more other pollsters are probably applying some form of "likely voters" weighting to their results without explicitly saying so, as is apparently the case with Forum Research according to ThreeHundredEight. Overall, the problem is a major lack of standards and transparency among Canadian pollsters as well as media reporting on polling at the moment.
As or the table, I personally favour a one-table presentation since it puts all the information in the same place and is more straightforward in the sense that it makes it obvious that each parallel set of results indeed stem from a single poll. I concede it's perhaps not the prettiest presentation though, so further suggestions from others are welcome. If EKOS or Abacus Data poll some more during this campaign, which I suspect they will, there's a strong chance they'll include some likely voters results.
Finally, regarding your suggestion that I add to the opinion poll article, I'll definitely look into doing that once I find a little more spare time! Cheers. -Undermedia (talk) 13:15, 16 May 2014 (UTC)[reply]

I would favour Undermedia's original suggestion to report percentages for all respondents before the campaign starts, and "likely voters" (where available) during the campaign, since this presumably represents the pollster's best estimate of the actual result of the election. Whether or not it is a good estimate isn't our problem; we need only report what the polling firm thinks is a good estimate. Where "likely voters" isn't available, I would support a footnote to distinguish which sort of number we're using. Tunborough (talk) 14:57, 16 May 2014 (UTC)[reply]

A poll of likely voters is not predictive of final results. For example, say both party A and B are tied, but each Party A voter is 60% likely to vote while each party B voter is 40% likely to vote. So party A has 100% of likely voters. But the probable outcome is 60% for party A and 40% for party B. But that would require the pollsters to quantify the likelihood of voting and crunch the numbers, which they have not done. So all the poll tells us is that PC support should be higher than the poll of all voters, but does not tell us by how much. So it is no substitute for the poll of all voters. It could be too that PC voters are more likely to say they will vote, even if they are not. Obviously there are not anywhere near 50% of voters who will vote barring an "unforeseen emergency." TFD (talk) 20:28, 16 May 2014 (UTC)[reply]

The likely voters results (for pollsters that calculated them) have generally performed overall better than their corresponding all voters results in recent provincial elections. Three examples. In the 2014 Quebec election, the final Ipsos poll (conducted a week before ballot day) had PLQ 37 PQ 28 CAQ 19 QS 13 among all voters and PLQ 40 PQ 28 CAQ 18 QS 12 among likely voters, whereas the actual results were PLQ 41.5 PQ 25.4 CAQ 23.1 QS 7.6. In the 2013 Nova Scotia election, the final Abacus poll had Lib 46 PC 28 NDP 24 GP 1 among all voters and Lib 46 PC 27 NDP 26 GP 1 among likely voters, whereas the actual results were Lib 45.7 PC 26.3 NDP 26.8 GP 0.9. In the 2013 BC election, even though it was badly missed by everyone, the final EKOS poll had Lib 33.4 NDP 38.0 GP 16.2 Con 10.0 among all voters and Lib 34.5 NDP 40.5 GP 13.0 Con 9.3 among likely voters, whereas the actual results were Lib 44.1 NDP 39.7 GP 8.1 Con 4.8. *BONUS* example: EKOS' likely results also performed slightly better overall than their all voters results in the 2011 Ontario election.
In any case, it looks like opinions are rather evenly divided so far, so I suggest we focus further discussion on settling on a way to show both sets of results. As explained above, I favour a one-table layout with alternative numbers clearly explained in a footnote. Anyone else? -Undermedia (talk) 13:05, 17 May 2014 (UTC)[reply]
There is no statistical difference in the two sets of polls for the three provinces. Where the polls of all voters were wrong (NS), the likely voters poll was wrong too. And comparing three sets of polls is itself statistically insignificant. BTW you missed the Alberta election where an Environics poll of likely voters showed the PCs leading, but it was way out too. TFD (talk) 16:14, 17 May 2014 (UTC)[reply]
If you add up the cumulative error per party (compared to the actual results; a standard way of measuring the performance of polls), in all four cases the likely voters results had less total error than the all voters results. Perhaps not statistically significant on an election by election basis, but overall indicative of a trend favouring likely voters results. Also, as mentioned above, some polls now only include likely voters results (Angus Reid in the QC election) while others are de facto showing them without explicitly saying so (Forum Research). In the end though, it's not so much a matter of whether we believe the likely voters results are more accurate or not, as it is a simple matter of whether we include on Wikipedia this new information that, agree with it or not, an increasing number of pollsters are media outlets are opting to publish. So again, I would suggest that we're now beyond debating the merits and accuracy of likely voter polling models, as a total of three users have now come out in favour of showing them and two more are open to it; the question now is how to show them, even if only as a footnote to the top line results. -Undermedia (talk) 16:45, 17 May 2014 (UTC)[reply]
A sample size of three is too low for accuracy and you have not calculation the confidence level. Also while you are right that some polls contain only likely voters, they determine likelihood in a different way, for example did they vote last time. TFD (talk) 17:37, 17 May 2014 (UTC)[reply]

Here's another suggestion for the presentation of both sets of results, with the likely voters results as a footnote below the table:

Polling firm Last date of polling Link Liberal PC NDP Green Type of poll Sample size
EKOS May 15, 2014 PDF 37.1 30.3 20.9 7.3 IVR 1,111
Ipsos Reid[1] May 14, 2014 HTML 30 39 24 7 Online 801
Forum Research May 12, 2014 PDF 38 35 21 5 IVR 996
Ipsos Reid[1] May 9, 2014 HTML 31 37 28 4 Online 821
Innovative Research May 6, 2014 PDF 38.8 32.8 19.9 6.5 Online 1,000
Innovative Research May 6, 2014 PDF 38.7 32.8 23.7 4.3 Telephone 500
Oraclepoll Research May 5, 2014 PDF 31 42 25 3 Telephone 1,000
Forum Research May 3, 2014 PDF 33 38 22 6 IVR 1,845
EKOS May 1, 2014 PDF 34.7 31.6 22.2 9.4 IVR 1,576
Election 2011 October 6, 2011 PDF 37.7 35.4 22.7 2.9

1 Results shown for Ipsos Reid polls are among all eligible voters. Published results among "likely voters" are as follows: May 14 - Lib 31%, PC 43%, NDP 22%; May 9 - Lib 28%, PC 42%, NDP 27%.

-Undermedia (talk) 15:01, 18 May 2014 (UTC)[reply]

When we present stats, which are primary sources, we should explain their relevance. The relevance of polls of voters is obvious - it is a projection of what the results would be if an election were held today. But what is the relevance of a poll of likely voters? It merely shows that the polls may be skewed but fails to quantify it, or provide an alternative projection. Similarly we could say that people earning over $100K per year or having university degrees are more likely to vote and providing polls for them. The best approach would be to have an explanation of how pollsters think the raw polls may be skewed. And why should we accept that a statement by voters that they intend to vote - which we know is highly inaccurate - is more reliable than weighting demographic groups based on how likely we think they are likely to vote based on previous voting? TFD (talk) 17:49, 18 May 2014 (UTC)[reply]
I'm seeing that we could probably debate this endlessly, so I'll restate the bottom line, which is that I'm simply proposing to include information which, like it or not, has been published by both the pollster and the media outlet that commissioned the poll as part of an increasing trend in Canadian polling and one that has been well established in the U.S. for quite some time. Look at the CTV News article on the poll: the "likely voters" results are clearly emphasized immediately after the "all voters" results. Your suggestion to provide further explanation surrounding "likely voter" models and projections is a good one, which I reckon would be best suited for the general Opinion Poll article, as proposed by Who R you?, and fully intend to get to work on as soon as I find the time. In the meantime though, please note that so far you're the only user who is steadfastly opposed to showing this information, compared to 3 in favour (myself, Ahunt and Tunborough) and 2 who are open to it (Krazytea and Who R you?), so unless others join you in opposition, I will soon move to incorporate it into the main article. -Undermedia (talk) 18:32, 18 May 2014 (UTC)[reply]
the article does not put the two polls on an equal footing and its chart does not either. I am sure you can find data about voting intentions among 35 to 44 years olds, people who earn between 50K and 60K, etc., but we don't add them to the list of poll results. Your theory is that there is a stronger relation between polls of people who say they will vote and election results than between polls of intended voters and actual results, but the pollsters do not claim that polls of people who say they will vote are intended to project what actual votes will be. TFD (talk) 19:44, 18 May 2014 (UTC)[reply]
Personally, I'm perfectly satisfied with Undermedia's suggestion above to footnote Likely Voter results, or even (particularly if at least ⅓ of the polls have the information available) to add 4 more columns to the right of the table for Likely Voters.  My concerns were that Likely Voter numbers can make for incomparable poll results (i.e. when Likely #s were first added for Ipsos, it appeared that a huge swing had occurred when really it was poll reporting accounting for much of the variance) and that including the numbers in parenthesis made the table seem confusing.
TFD — I think there's a big difference between selectively reporting only 35–40yos earning $50–60K, and segregating those identifying a strong intent to vote; but I do agree with you that it would be very useful (for editors & readers) to see some type of backup to support that polling of self-identified likely voters is more representative of actual election day results.  I note that Undermedia has graciously accepted my challenge/request to (time permitting) provide some Wiki article documentation supporting that likely voter figures reported by pollsters are more representative of election day results; but until we have that, I think we can all accept that likely voter numbers are often reported by mainstream media and that our reporting them, as long as they are clearly differentiated from general poll results and aren't likely to lead to confusion, means just one more piece of information for readers to have at their disposal.  Cheers — Who R you? Talk 22:18, 19 May 2014 (UTC)[reply]
In fact there is no difference. If the polls of 35-40 yr olds earning 50-60K came closer to the actual outcome than the polls of all voters, you could make same the argument that we should use them instead. Similarly it sunspot activity were a better predictor we could argue for using them. If you want to alter the polls based on likely voters, then you need to provide weight, not just totally ignore voters who are less likely to vote. Many of the certain voters will not vote, while perhaps 40% of the unlikely voters will vote. In fact some polls do provide greater weight to likely voters, which is why the 2008 and 2012 U.S. polls made the race appear closer than it was. The pollsters understated the number of young people, African Americans and people who would vote. TFD (talk) 14:28, 21 May 2014 (UTC)[reply]
I really think it might be a bad idea to discard completely the general opinion polls in favour of likely voters. You might even want to report two numbers for certain polls. I think the footnotes are too small a reference to indicate why there is a disparity in votes. You might want to create a new column indicating the poll type, similar to online or phone polling.
 
Polling firm Last date of polling Link Liberal PC NDP Green Polling method Type of poll Sample size
Ipsos Reid May 14, 2014 HTML 30 39 24 7 Online Likely voters 801
Forum Research May 12, 2014 PDF 38 35 21 5 IVR General public 996
Ipsos Reid May 9, 2014 HTML 31 37 28 4 Online Likely voters 821
Innovative Research May 6, 2014 PDF 38.8 32.8 19.9 6.5 Online Likely voters 1,000
Innovative Research May 6, 2014 PDF 38.7 32.8 23.7 4.3 Telephone Likely voters 500
Oraclepoll Research May 5, 2014 PDF 31 42 25 3 Telephone Likely voters 1,000
Forum Research May 3, 2014 PDF 33 38 22 6 IVR General public 1,845
EKOS May 1, 2014 PDF 34.7 31.6 22.2 9.4 IVR General public 1,576
Election 2011 October 6, 2011 HTML 37.62 35.43 22.73 2.94
Krazytea(talk) 16:52, 21 May 2014 (UTC)[reply]
TFD, the key difference between the likely voter models and all the other arbitrary predictors you cite is that the former are expressly intended to better predict the outcome of elections, having been created in direct response to the failure of conventional "general population" data to predict the outcome of a string of recent elections. All the pollsters who now publish likely voter results alongside their general results have at one point issued a report or commentary rationalizing their decision to start doing so or explaining the need to focus more on voters who are most likely to actually show up at the ballot box. EKOS first discussed it in detail immediately after their big miss in the 2011 federal election, while Abacus Data, Angus Reid (see Notes on Methodology on page 5) and Ipsos Reid each discussed it at some point following the BC election. -Undermedia (talk) 23:35, 21 May 2014 (UTC)[reply]

Polling Statistics – Margin of Error

[edit]

We have a minor difference of opinion on these changes which, other than the very appropriate corrections to the Forum Poll, I reverted.  I don't disagree that 19 times out of 20 is a pretty standard level of accuracy in polling today; but there was a time, in the not too distant past, that 95 times out of 100 (as it was known then) was considered a poor polling margin with 98 or 99 times out of 100 being considered not only more desirable, but realistically obtainable.  This is also connected to the Margin of Error note on the Innovative Research online poll — Innovative correctly points out that (all) online polls are inherently flawed as they're skewed and not a random sample group… the statement from Innovative (added via the note) simply provides a sourced comment for readers to understand that all online polls are unrepresentative of the population as a whole (I could argue that IVR & even telephone polls, with their sample sets confined to those that have a landline telephone (an ever decreasing cross-section of the population) and/or don't immediately hang-up on computer generated calls, are potentially unrepresentative as well – but I know of no 3rd-party source on that).  It also alludes to why none of the other online polls provide that information (it's because they aren't considered statistically accurate – at least in the sense that they aren't based on a random sample group).  The 19 times out of 20 informs the reader of the statistical margin of error in the particular polls, informs them that it might not be what they expect, and informs them that online polls may be subject to a different set of considerations.  For those reasons, I've restored that column of information in order to determine if others agree.
Then there's the matter of "pp"… firstly, I note that the article that supports the use of the acronym "pp" doesn't site any sources (and might be appropriate for a deletion discussion)… in fact I've never heard of "pp" other than here on Wikipedia (not that my having heard of something, or not, is particularly meaningful).  Also see this discussion on removing it elsewhere.  But more than that, in this case, I'd argue that ±% isn't actually percentage point, since it isn't really a change in percentages typically represented by percentage point (although I do understand it being referenced in this context) — case in point, if an article said "50% ±40%", it's arguably understood that it's 10%–90% (not 30%–70%) being presented, just like 54% ±2.4% is universally understood to mean 51.6%–56.4%… and anyone that doesn't understand that arguably wouldn't understand what "50% ±40pp" means (or the basic concepts of statistical deviation or Margin of error, even if they read the linked article).  If there are any published 3rd-party references to reporting of polling numbers using ±pp (or ± percentage points) I'll withdraw my objection on that.  But, I am unable to think of any middle ground compromise on it.  — Who R you? Talk 19:38, 22 May 2014 (UTC)[reply]

Honestly, I don't have any terribly strong opinions on any of this so I'll live with whatever way it's presented. I've previously discussed opinion polling margins of error on Wikipedia, for example here. My main reservations with them is the inconsistent way in which they're reported/presented (sometimes for the total sample, sometimes for the reduced sample of decided voters only, sometimes "confidence intervals" as Ipsos Reid, etc.), and the simplistic and dubious way in which they tend to be interpreted (the given MoE is actually the maximum error that would apply only to a party whose support approaches 50%, whereas the MoE actually decreases for increasingly "less popular" options). The debate over whether online polls should be allowed to have margins of error—which inevitably leads to online poll bashing—also sort of drives me nuts since in reality NO poll is truly random: telephone-based polls may dial random numbers but they certainly never end up reaching a random sample of actual respondents within the general population. A lot of people are increasingly difficult to reach by phone, and even among those who are reached, there are demographic biases in the likelihood that they will actually agree to participate in the survey. You could actually argue that certain segments of the population are likely better sampled online than by telephone, but also vice versa. It all turns into a bit of a pissing contest with self-proclaimed "random" telephone polls trying to cling onto the illusion of holding the high ground. If online polls really were garbage, it's sure hard to square with their ever-increasing adoption by many veteran pollsters (e.g. Ipsos, Léger and even Nanos) and the corresponding progressive disappearance of conventional telephone polls. Finally, as an ecologist who has done LOTS of stats, I'm fully confident that the use of % when pp should be used is a grossly pervasive error (even by the pollsters themselves), though ultimately so pervasive that you're probably right that readers will tend to be more confused by seeing pp and probably tend to interpret % in the sense of pp, even though it's technically incorrect to use % in this situation. I'm just a bit of a purist when it comes to these things, but again I'll live with whatever way it's presented. Cheers. -Undermedia (talk) 21:04, 22 May 2014 (UTC)[reply]

Threehundredeight.com

[edit]

What's the consensus here is http://www.threehundredeight.com/p/ontario.html a WP:RS? It seems pretty cohesive in aggregating and presenting data and can easily qualify as a secondary if not tertiary source. Seeing how in this type of elections polls are not as important as seat projections, this could be useful source. What do you think? --Truther2012 (talk) 17:37, 24 May 2014 (UTC)[reply]

I've used 308 before as a source for individual polls, which the site regularly reports on in detail, but only when no other link directly from the pollster or a news article was available, or a previously accessible such link became broken or locked. 308 is technically a blog, i.e. a "self-published source", which is generally not acceptable as a source for WP articles, however exceptions can be made "when produced by an established expert on the subject matter, whose work in the relevant field has previously been published by reliable third-party publications". I would argue that 308 author E. Grenier qualifies as such an expert, as he regularly writes polling-related articles for the Globe and Mail (latest here), Huffington Post (here), Hill Times (here) and l'Actualité (here). I don't know about referencing 308 for seat projections though. Let's see what others think. -Undermedia (talk) 22:56, 24 May 2014 (UTC)[reply]
I agree he meets WP:SPS as a reliable source despite being self published there on that website. - Ahunt (talk) 23:13, 24 May 2014 (UTC)[reply]
We should not post his projections because they're usually terrible. tooclosetocall is much better, as is electionprediction.org (as is my site, but my mention of it would be against WP:POV). -- Earl Andrew - talk 18:45, 31 May 2014 (UTC)[reply]
OR if we do mention his site, we should list them all. Election Almanac has a good list: http://www.electionalmanac.com/ea/ontario-seat-projections/ -- Earl Andrew - talk 18:47, 31 May 2014 (UTC)[reply]
Yeah, I was referring to using 308 as an occasional source for results of individual polls, but I don't think we should report seat projections from any of the "poll aggregators". Too many different ones to choose from, too much uncertainty around them, and although I've seen 308's polling averages (the %'s) quoted in news articles, I don't think I've ever seen references to the seat projections. -Undermedia (talk) 20:19, 31 May 2014 (UTC)[reply]

Innovative Research: Online polls not random

[edit]

If the online poll isn't random, why is it being included in the table? Innovative Research already conducted a poll by telephone that exact same day, one that has a defined margin of error, so having this second poll doesn't make much sense to me. -- Kndimov (talk) 21:14, 25 May 2014 (UTC)[reply]

Please check out my post above in response to Who R you?. First and foremost there's no such thing as a truly random poll, and that Innovative online poll was more than likely conducted in the exact same way as the online polls by Ipsos Reid and Abacus Data, except that Ipsos chooses to give a "confidence interval" rather than a margin of error and if you read the fine print of Abacus Data's methodology, they in fact report what the margin of error "would be" had the sample been truly random (Léger Marketing and Angus Reid do the same). Innovative is just a little old-school I suppose in refraining from providing a margin of error altogether. This is actually another reason why I personally favour not showing the margin of error at all: the inconsistency with which it's reported seems to inevitably cause debates over the legitimacy of this poll or that, or this pollster or that. Plus the margin of error is calculated directly from the sample size, so it's redundant IMO. -Undermedia (talk) 22:58, 25 May 2014 (UTC)[reply]
I fully agree, with you (Undermedia), that the margin of error is redundant in the sense that it's a mathematical calculation with input variables sample and population size; but few would know how to calculate it from those numbers and even fewer could do it without looking up the formula first.  Just as the ±2% is at least somewhat informative to the layman reader, I'd submit the 19 times out of 20 can be as well.  I might not know exactly what ±3.6% 19 times out of 20 means, but I'm sure I'll know it's less accurate that ±1% 19 times out of 20, and I'll know that it's more likely to be within the specified range when compared to ±3.6% 3 times out of 5.  Even if it sparks discussion (as here) about whether online polls should be included in WP articles or referenced by media, I'd say it's better to cause more discussion (hopefully based on increased knowledge) by telling people more about what they're reading, than to avoid people questioning things by telling them less.  I don't disagree with many of your points regarding phone polls, etc. (in fact some of those exact points were touched upon in the methodology section of one of the recent polls); I think it's simply that online polls are in a league of their own.
Telephone polls (excl. IVR) have some advantage at least in terms of the phone number having an accurate physical location (to permit weighting of polling results for region, urban/rural, etc.) and the human-being on the pollster's side of the call can occasionally weed out the obvious cranks (the 12-year-old that gets the call and decides to answer claiming they're a 40-year-old, etc.).  Compare with the online poll — I'm personally in an online polling group, registered as "Anonymous", I sometime access via anonymization proxies (altering my IP), and, until the last couple days, I had forgotten to update my profile's Postal Code after moving some months ago — all pretty good examples of why online polls are inherently more error prone regarding location (which pollsters use when trying to extrapolate sample sets into representations of the entire population).
Then there's the sub-basement-dweller aspect of a disproportionate portion of online respondents; you're more likely to find someone online providing incorrect information than you are in other real-world situations – I submit such intentional misinformation is more prevalent in online polling compared with telephone or public polling.  And lastly, online polls suffer from the inherent and unavoidable bias that their sample sets only include people that have internet connections (and almost exclusively high-speed internet connections).  Plus many polls are first-come/first-served with a limited number of participants accepted; which further skews the sample set towards those that never leave their computer screen, have their email program always open, and read it immediately when it beeps.  If I wanted to convince someone that online polls were a good method of determining election results, I'd likely start by having an in-person exit poll that asked “Can you access the internet at your residence?”, “How much personal time do you spend on the internet?”, “Do you ever answer online polls?”, & "Who did you vote for?”… but I suspect I'd find it would indicate that older generation (particularly males) can't/don't answers online polls (and they tend to vote PC more than the average), and you'll find that younger generation (particularly females) do answer online polls (and they tend to vote NDP or Green more than the average).
So why do pollster's use online polls?  Because they're cheap!  Pollsters generally pay online participants (or at least allow them to accumulate credit in an account, with a promise of future payment) in the range of $1 – $2 per 15 – 20 minute polls; you can't pay for someone to sit on the telephone and make calls all day for that amount of money, and you can't pay someone to stand outside and interview people for that kind of money; so if you're a professional polling company in the business of making a profit, you use online polls whenever possible.
And then there's Kndimov's original question, Why does Wikipedia report online polls if we know they aren't truly random sample groups and they likely are somewhat less accurate than other types of polls?  Mostly because they exist and are reported by other independent, 3rd-party media.  They may not be completely accurate, but neither is any poll; online might not be as accurate as door-to-door polling, but online polls exist and nobody goes door-to-door to poll people anymore; online polls occur often and might show a change in attitude before the next other kind of poll is released; online might be inaccurate, but it's more accurate than tea-leaves (which would be WP:OR anyways).  For the most part, if it comes down to weighing how accurate something is, as long as we accept by consensus that it's at least reasonably accurate, we report the information to the readers (with as much detail as we can provide about why it is or isn't particularly accurate), and then we let the reader(s) decide for themselves whether or not they want to give that particularly polling methodology any credence.  If someone reads the WP article and learns that all polling methods are not the same, and thinks about that the next time they read a headline or watch a news segment about new polling numbers, then we've probably done something worthwhile.  — Who R you? Talk 02:06, 26 May 2014 (UTC)[reply]

RfC on presentation of Canadian Election results

[edit]

Two questions have arisen about what to include & how to format and calculate information in election results tables for Canadian (and right now particularly Ontario's) elections.  Input from interested parties would be appreciated regarding Calculation of Swing and Use of “Election results” tables in politician articles.  Thanks  — Who R you? Talk 19:54, 3 June 2014 (UTC)[reply]

  • Comment Ok you can't calculate swing. Doing so to me would be origical research. If you have a reliable source that has done this then by all means use it. As for which to use I'd say use all notable ones. I don't see the actual problem with showing the election results.Serialjoepsycho (talk) 02:54, 23 June 2014 (UTC)[reply]

Article's Media endorsements section

[edit]

Recognizing that it's directly contrary to WP:MOSLINK, I'm relying on WP:IAR to improve the article… I trust others will agree it's an improvement (or will revert it accordingly).  A list of itty-bitty footnotes, directly above the References section just doesn't read well and isn't useful.  The point of the section is to provide the reader with more information, not to increase the size of the References list; accordingly, the list of Media endorsements should actually list the endorsements within that section.  Converted <ref>'s to plain (externally linked) {{cite …}}'s in that section.  — Who R you? Talk 15:43, 8 June 2014 (UTC)[reply]

You are correct, it is a direct violation of the WP:EL. There is definitely a better way to present the endorsements than footnotes, but sticking links in the body of the article ain't it! That's the whole point of Refs is to provide proper sourcing for a statement made, i.e. "Group A endorsed Party B". So, I disagree. And if I get a moment will re-format.--TRUTHER 02:46, 9 June 2014 (UTC)
I think the current format, with the tables, honestly looks uglier than both previous versions. I think Who R you?'s version is probably better because it is more user-friendly, though it could be made to look a bit prettier. Esn (talk) 17:31, 10 June 2014 (UTC)[reply]
@Truther2012: Actually, WP:EL says “…should not normally…” and “Some acceptable links include … that could not be added to the article for reasons such as copyright…”.  Indeed, WP:ELYES supports external links “…relevant to an encyclopedic understanding of the subject and cannot be integrated into the Wikipedia article due to copyright issues…” (I grant that that sentence explicitly references “neutral” material; but there's no such thing as a neutral endorsement; so we satisfy WP:NPOV by including all the unique RS endorsements we find.).  Without direct and unedited quotation of what the endorsement said, it can't be included — someone can't restate my endorsement of "I think … should win" as "… said they support …"; the selection of words is vital to the precise meaning of the endorsement; there is no rewording, that wouldn't violate copyright law if not cited as a direct quotation or be a question of POV because it's been so significantly altered, that can properly and accurately convey (using the first example) "Kathleen Wynne has earned a fresh mandate in Ontario: Editorial". Toronto Star. 2014-06-06.  The only other option would be to list it as: “The Toronto Star endorsed the Liberal party on June 6, 2014, in an editorial entitled "Kathleen Wynne has earned a fresh mandate in Ontario: Editorial"[1234]”, or in a table format providing the same information.  But then what is the purpose of making it more difficult for the reader to follow the link by forcing them to hover the cursor over the little tiny [#] footnote in order to display the popup so that the user can aim the cursor again to click on the link; the whole point of WP:IAR is, when a rule doesn't make the encyclopedia better/more usable, ignore that rule (and in this case I believe WP:ELYES suggests it's appropriate to use external links regardless).  You're worried (as I interpret your user page) about color selection factors affecting the color blind (Excellent, someone should be); but how about some consideration for those that find it a challenge to get the cursor in the ⅛"×⅛" footnote marker; or those that have a hard time reading the smaller text of the footnote?  If you've got a better suggestion, great; however, if it's just there must be a better way but I'll leave it to someone else to figure out what that is, then how about we go with inline cites until that better way gets thought of; and personally I don't think the current table (with the identical footnote issue as before) is the solution.
@Esn: Thanks, I agree with you regarding the new table format.  Cheers — Who R you? Talk 18:56, 10 June 2014 (UTC)[reply]
I do like the colours you added, but I decided to be bold and revert it to something simpler, for the reasons I gave in my edit summary: If there are no NDP endorsements, there should be no NDP heading. Simple bullet points allow more room for extra text to explain the nuances of the endorsement ("minority"). Also, identical editorials can be listed under same bulletin (I do think it's noteworthy to mention which papers printed an editorial, even if it's identical. This could get out of hand if there are more than a few papers printing identical editorials, but so far it's not a problem.) Esn (talk) 19:41, 10 June 2014 (UTC)[reply]

So, the one thing I did like about your version was the colour-coding. I guess one option for keeping colours in the headings is something like this:

Liberal
Progressive Conservative
None of the above

I think it would be cleaner if there was a way to just have a small coloured bullet point or square beside the heading, but I'm not sure that's possible. Esn (talk) 20:02, 10 June 2014 (UTC)[reply]

@Esn: We could use "█ Liberal", "█ Progressive Conservative", "█ New Democratic", etc.; but to my knowledge, the only support for colour in the Wikipedia standards is in tables and photos; but I might be wrong on that.  You might already have noticed, I changed your "None of the above" caption; that was my first thought as well, but then I remembered that "None of the above" (NOTA) is actually a registered party name this time around, on Election Ontario's official list of registered parties; so on the off chance that some editorial staff decides to support Greg Vezina's party, and in order to avoid confusion, we can't really use that (even though it fits sooooo well); I changed the title to "Explicitly not endorsing any party".  As for the list format, back to the original list is probably better than the table; but it still has the same issue that I thought was the problem originally; it's difficult for readers to know what the endorsement means and difficult for them to follow the links.  Oh well, c'est la vie. Cheers — Who R you? Talk 21:09, 10 June 2014 (UTC)[reply]
Who R you?, I agree with you that WP:IAR and WP:ELYES are useful at times to present information when no better way exists. However, simply sticking a link in the prose cannot possibly qualify as "better". If your concern is to over-populate the Ref section, then you should not worry, as the high number of Refs is simply a sign of well-researched and well-supported article. Personally, I find information more-readable when it is well laid out in a table, if the consensus here is against it, so be it. I am a bit confused by your comment about "⅛"×⅛" footnote marker", as I am not proposing to shrink text in any shape or form.--Truther2012 (talk) 13:52, 11 June 2014 (UTC)[reply]
@Truther2012: — By the “⅛"×⅛" footnote marker”, I'm referring to the “[##]” actual link to the footnote, which the user must click or hover over in order to find the link so they can follow it to see what the endorsement said; it's like a carnival game, hit the target, win (some of) the information you're interested it.  I simply find that the information provided in this section is meaningless; it might just as well say:
  • Supporting Liberals – 3 [38], [39], [40]
  • Supporting Progressive Conservatives - 6 (9) ([41], [42], [43] – Same editorial by Metroland Media Group), [44], [45], [46], ([47], [48] – Identical editorial by Canoe Sun Media), [49]
  • Supporting New Democrats – 1 [50]
since that would actually provide just as much meaningful information as the current 13 lines do.  Nor is it that I care about how big the References section is; rather it's that the meaningful information, the "Wynne has earned a fresh mandate in Ontario: Editorial" – Toronto Star vs "In this provincial election: (sigh) The Liberals – Much to our chagrin, the only endorsement we can make." – Torontoist, the "Ontario election part 4: For a Conservative minority" – The Globe and Mail vs "The Star's View: Hudak has the right plan to restore Ontario's economic health" – Windsor Star that is missing, or at least difficult for the reader to access, in the current format.  If it's better, I'll suggest changing it to Sudbury Star – "The NDP - for now"[50] format; but we have a system for quoting and linking to external sources that is much easier for both editors and readers to use and understand; it's the {{cite}} template set; and that just seems more appropriate to me, rather than typing everything in proses format then adding the footnote containing the identical information written in the prose just with a more difficult link to follow.  Generally speaking, the footnotes are more for us, the editors, in terms of our being able to verify the information provided, and secondarily for the reader to follow should they wish to verify or gain more in depth knowledge of what we've typically summarized; this case is the exact opposite, most readers interested in who media outlets endorsed are going to want to read the articles to learn what/why the endorsement is what it is; I doubt too many readers are coming to WP to read “Oh, The Toronto Star endorsed Liberal; I now know all I need to know on that subject”.  I merely seek the format that provides the most useful, easily accessible information.  Cheers — Who R you? Talk 15:33, 11 June 2014 (UTC)[reply]
@Truther2012: — Regarding the Personal endorsements and the removal of de facto, not only is there presently a de facto Mayor of Toronto, but (depending on council's wording and voting this week), there may soon be, as a matter of law by declaration of city council, a "Deputy-Deputy Mayor"… "Toronto to get 'deputy-deputy mayor'". Toronto Sun. 2014-6-10. {{cite web}}: Check date values in: |date= (help).  Cheers — Who R you? Talk 16:02, 11 June 2014 (UTC)[reply]

Public figure endorsements

[edit]

I find this section a bit odd and I am wondering what the sense of it is. There's currently three names there and each have endorsed the Liberals, but all of them are well known Liberals anyways. Norm Kelly is a former Liberal MP, Hazel McCallion is a known Liberal, and Carl Zehr has run for the party in the past. Unlike with the newspaper endorsements, none of these are that surprising. Tom Muclair was out campaigning with Andrea Horwath and Ed Broadbent announced he supported her. There were also federal Conservatives helping members of the PC Party. Newfoundlander&Labradorian (talk) 19:53, 12 June 2014 (UTC)[reply]

Well, I think mentioning endorsements by city mayors at least is noteworthy, especially since in their title as mayor they're not officially part of any party (unlike in some other provinces/countries), so aren't expected to automatically support "their" party. David Miller was an NDP-er but tended to stay neutral in provincial and federal contests, campaigning for particular policies rather than a particular party. Rob Ford did not endorse anyone in this campaign, and neither did any of the other mayoral front-runners (not even Olivia Chow). I'm not entirely sure about how "public figure" should be defined. Perhaps it can include provincial union leaders, major company/business leaders, international/out-of-province leaders. Maybe something along the lines of the "supporters" sections in Ontario Liberal Party leadership election, 2013. I wouldn't necessarily include people like Tom Mulcair whose official job duty is to be a leader of another branch of that same party (the NDP is the same party provincially and federally, unlike, for example the federal and BC Liberals). Anyway, just throwing a few thoughts out there. Esn (talk) 13:44, 13 June 2014 (UTC)[reply]
I think endorsements, other than media endorsements, are best covered in the body of the article provided they are significant. This election featured endorsements by mayors, which is unusual. Miller btw had been an NDP city councilor, but resigned from the party when he became mayor. TFD (talk) 19:59, 13 June 2014 (UTC)[reply]
Not true, Miller only allowed his NDP membership to lapse in his last term. Esn (talk) 02:16, 14 June 2014 (UTC)[reply]
Just because a mayor doesn't run under a party banner doesn't make them any less a member of a certain party. McCallion for instance was at the Liberal leadership convention last year so she must be a member of the party. I don't know about the others. Just seems a bit odd to me. Newfoundlander&Labradorian (talk) 14:37, 15 June 2014 (UTC)[reply]
The difference is that they do not sit in a party caucus and are free to vote however they like. City councilors tend to form voting blocks. So in the Toronto council, Liberal Norm Kelly was on the right, supporting the Conservative mayor, while Adam Vaughan, another Liberal was on the left and Josh Matlow another Liberal is in the Centre. In the 1960s/70s, the two blocks were Reformers and Old Guard, which also cut across party lines. Mammoliti was a New Democrat during his first years as a councilor, before becoming a Liberal. But he always voted with the Right. TFD (talk) 15:19, 15 June 2014 (UTC)[reply]

A detail

[edit]

The article currently includes a timeline saying that on June 12, "Kathleen Wynne wins a majority, claiming 58 ridings in the Ontario election". The cited source does not show the number of seats (there's only a bar chart), but the total reported on election night was 59 ridings for the Liberals. It was corrected to 58 when the official tabulation on June 13 revealed that the winner first announced for Thornhill had been wrong.

If the timeline is going to quote a source for "Wynne wins a majority, claiming" a specific number of ridings, then it should be one that says that. And the number of ridings should be what was believed to be correct on June 12. And then there should be another item dated June 13 and saying that the official tabulation changed the winner in Thornhill and giving the corrected total. (I don't know if the correction was announced the same day or only on June 14, today as I write, the day I heard about it. None of the articles I found say when it was announced. I have added something about it in the article on the Thornhill riding itself.)

Alternatively, perhaps this level of detail is excessive and the remark about the number of ridings should just be deleted. That would make sense to me.

--69.158.92.137 (talk) 19:46, 14 June 2014 (UTC)[reply]

The +- party seats in the top right box

[edit]

Hi,

Currently the number of seats the parties have gained and lost in the top right box don't add up and they also contradict the figures given in the first section. I changed them to reconcile them with the figures given in the first section. This change was reversed as being "incorrect information". I'm not an expert on Canadian politics so I'm not going to change them again but it does seem that my information was correct.

After all the results in 2011 as given on Wikipedia with proper sourcing etc... were:

- Lib 53 - PC 37 - NDP 17

and the results in 2014 as given on Wikipedia with proper sourcing etc... were:

- Lib 58 - PC 28 - NDP 21

Therefore

- Lib = 58-53 = +5 (not +10 as is currently shown) - PC = 28-37 = -9 - NDP = 21-17 = +4 (not +-0 as is currently shown)

Furthermore the +- results as they are now don't even add up to 0, which they surely should? — Preceding unsigned comment added by 85.159.94.156 (talk) 11:46, 6 November 2014 (UTC)[reply]

Seats prior to the election are used. 117Avenue (talk) 05:42, 7 November 2014 (UTC)[reply]
[edit]

Hello fellow Wikipedians,

I have just modified 2 external links on Ontario general election, 2014. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 17:49, 7 November 2017 (UTC)[reply]

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:2014_Ontario_general_election&oldid=1196668372"