Jump to content

Talk:Opinion polling for the 2017 United Kingdom general election/Archive 2

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3Archive 4

Donald Trump

Having just noticed this reversion I think it's worth pointing out here that Donald Trump's election should not be included in the timeline of an opinion poll relating to British politics. DT is (thankfully IMO) not one of our politicians, so including his election here seems unnecessary. This is Paul (talk) 19:04, 12 November 2016 (UTC)

Agreed. Bondegezou (talk) 23:56, 12 November 2016 (UTC)

Results of by-elections

An IP has added the results of the by-elections which has taken place this Parliament. I think that this adds too much detail and clutters the table with too much irrelevant information. Thoughts? Absolutelypuremilk (talk) 11:19, 10 December 2016 (UTC)

It's actually pretty relevant, and can easily be descirbed objectively. Should be kept to a minimum though, ie "Lib-Dem gain from Con", "Con Hold", "Lab Hold". I'd support inclusion, but no strong feelings. 79.74.10.202 (talk) 16:24, 12 December 2016 (UTC)
Having thought about this some more, I'm firmly of the opinion that we should either give the result (in absolutely neutral terms, as per my above comment) or not mention the by-elections at all. I don't have strong feelings as to which, but one or the other. We list party elections as "joe blogs is elected leader of turnip party", not just "turnip party leader elections". Additional comments on this would be helpful: it's been a month, no sign of consensus. 79.74.24.175 (talk) 18:30, 13 January 2017 (UTC)
Most polling articles on Wikipedia don't list by-elections, or party leader changes. I'd favour excluding everything. If we do have them, a simple "Con hold", "LDem gain from UKIP" or whatever seems appropriate to me. Bondegezou (talk) 19:05, 13 January 2017 (UTC)

Suggestion: 'Best PM' polling?

I've noticed that in some other 'opinion polls for next [X] election' articles they have polls showing opinion on the question of 'best Prime Minister'. (See Opinion polling for the New Zealand general election, 2014, Opinion polling for the Spanish general election, 2016 and National opinion polling for the Australian federal election, 2016 as examples).

These 'Best PM' polls in the UK have been cited in reliable newspaper articles (https://www.theguardian.com/politics/2016/nov/19/tories-trusted-by-twice-as-many-voters-with-economy-than-labour-poll) and they frequently provoke discussion, so I was wondering what folks thought of adding them to this article as a separate section at the bottom? So in this style:

Date(s)
conducted
Polling organisation/client Sample size Theresa May Jeremy Corbyn Don't Know Lead
3–4 Jan YouGov/The Times 1,740 47% 14% 39% 33%

FriendlyDataNerd (talk) 20:54, 11 January 2017 (UTC)

These polls are often inconsistent in the way they're asked, who's asked about, how/if they're reported & frequency (though frequency isn't a huge problem, might be a bit misleading to list them, in some circumstances...probably doesn't matter). Also this article is already huge. If this is to be done, I'd suggest they got a separate page, but I'm against it. Getting borderline notability. 79.74.24.175 (talk) 18:24, 13 January 2017 (UTC)
WhatUser:FriendlyDataNerd suggests is not new or different; Exactly this is done in articles for other countries like this one, Opinion polling in the 43rd Canadian federal election. I am puzzled that we have not included this important additional detail. I think that readers who arrive at this page would want to also access leader polling.
The argument of user 79.74.24.175 is not a good one. There are no more inconsistencies with leader polling than with party polling. I also don't think that size of article is important, so long as the reader can easily navigate the page. So I urge FDN to be BOLD and go ahead with adding leader data. Graemp (talk) 18:53, 13 January 2017 (UTC)
Article length is recognised to be an issue by policy, so we absolutely have to consider it. I agree on that point: a new article would be appropriate.
I also share the view that polling of this nature is done in a variety of ways: it is much more varied than standard party polling. But I think the solution to that is presentational: I'd have a table for each polling firm using a different wording rather than mixing them all up. Bondegezou (talk) 19:02, 13 January 2017 (UTC)
Thanks for the perspectives folks - it's appreciated. Length of article is definitely a concern which I understand. I personally feel that as the other articles have included 'best PM' polls in their main articles, it makes sense to keep that format for the UK, but I'm not wedded to it. As regards separate pollsters having separate tables, aye, that makes sense too - the vast majority of polling on Best PM comes from YouGov and is all done in the same way, but there are a few Opinium polls that include a third option, plus Ashcroft polls that have Tim Farron.
If folks are of the mind to have a separate article, might I propose the creation of an article called "Opinion polling on United Kingdom Party Leaders"? We could link to it at the beginning and end of this article. Thanks again for the responses. FriendlyDataNerd (talk) 19:59, 13 January 2017 (UTC)
Given that this article is currently well under 100,000 bytes in length, I think we are no where near any worries about it becoming too long with the adding of a leader section. Therefore I see no problems with adding the leaders section here. I would argue that if anything, the sub-national polling should be split off into a separate article. I'm not convinced about the need to break up the leader tables by different polling companies. I think it would be better for the reader to list them all together chronologically as with the party tables. Perhaps we should start this way and if we then find it creates problems, consider alternatives.Graemp (talk) 07:09, 14 January 2017 (UTC)
I should say, this is what I envisioned it looking like all as one table. Having looked at the pollsters they all ask essentially the same question, just giving different options. -
===May vs Corbyn===
==== 2017 ====
Date(s)
conducted
Polling organisation/client Sample size Theresa May Jeremy Corbyn None of These Not Sure Lead
3–4 Jan YouGov/The Times 1,740 47% 14% 39% 33%
==== 2016 ====
Date(s)
conducted
Polling organisation/client Sample size Theresa May Jeremy Corbyn None of These Not Sure Lead
18–19 Dec YouGov/The Times 1,595 44% 16% 41% 28%
13–16 Dec Opinium/The Observer 2,000 42% 16% 28% 13% 26%
4–5 Dec YouGov/The Times 1,667 49% 16% 35% 33%
28–29 Nov YouGov/The Times 1,624 45% 18% 37% 27%
15–18 Nov Opinium/The Observer 2,005 45% 17% 25% 13% 28%
14-15 Nov YouGov/The Times 1,717 48% 18% 34% 30%
1–4 Nov Opinium/The Observer 2,001 45% 16% 25% 13% 29%
31 Oct-1 Nov YouGov/The Times 1,655 47% 17% 36% 30%
24-25 Oct YouGov/The Times 1,655 48% 16% 36% 32%
11-12 Oct YouGov/The Times 1,669 51% 18% 31% 33%
13-14 Sep YouGov/The Times 1,732 50% 18% 33% 32%
30-31 Aug YouGov/The Times 1,687 52% 21% 27% 31%
22-23 Aug YouGov/The Times 1,660 50% 19% 30% 31%
16-17 Aug YouGov/The Times 1,677 51% 19% 30% 32%
8-9 Aug YouGov/The Times 1,692 52% 18% 29% 34%
1-2 Aug YouGov/The Times 1,722 52% 18% 30% 34%
25-26 Jul YouGov/The Times 1,680 52% 18% 30% 34%
13 Jul Theresa May becomes the Prime Minister of the United Kingdom
23 Jun The UK votes to leave the European Union; David Cameron announces he will resign as Prime Minister
5 May UK elections, 2016 including the Ogmore and Sheffield Brightside and Hillsborough by-elections
11-12 Apr YouGov/The Times[1] 1,693 23% 30% 46% 7%
FriendlyDataNerd (talk) 11:13, 14 January 2017 (UTC)

Adding this in sounds like a good idea to me, many thanks to FriendlyDataNerd for putting together the mock-up table. Absolutelypuremilk (talk) 11:34, 14 January 2017 (UTC)

This looks good. I would however suggest that a column be added for Tim Farron, even though most pollsters have not been prompting on him. I have seen some PM polling that has included him. I expect that we will see him included by pollsters more often from now on. The other thing that I have seen is Leader approval polling, which is not the same as PM polling. Approval polling usually includes far more prompted options. Leader approval polling would be better in another article, linked from the PM polling. Graemp (talk) 11:49, 14 January 2017 (UTC)
Awesome. There seems to be an agreement that we should have this, but I'll wait to see if anyone has strong objections to putting the Best PM polling in the main article before proceeding. As regards Farron, as far as I can find, only Lord Ashcroft included him in the polling and it seemed superfluous to include a whole new column of mostly N/A aside from one line - especially because his question was so radically different in that it didn't include don't knows and wasn't comparable. my feeling would be it should be included in an extra '3 Option' category just below the main one, and if there are more Farron polls, I can easily slide them all into the main table now I've got the hang of things. FriendlyDataNerd (talk) 15:14, 14 January 2017 (UTC)
I have an objection. This isn't what the page is about. It is about opinion polling for the next general election. We don't get the choice to vote for a Prime Minister, we vote for a local MP belonging to a party. The parties then put up the cabinet, including the Prime Minister. Lots of things are polled for, including how much the population trusts the parties with the NHS, the economy, crime, immigration, and including the question about which party leader we think would make a better PM. Ultimately, all of these things may or may not affect the eventual vote in the general election. Either you put all of them on the page because they all might mean something, or you stick with what the page is supposed to be about, and just have the polling on voting intentions for the next general election.Awoma (talk) 22:01, 16 January 2017 (UTC)
True, but those questions can be asked in different ways - e.g. who do you trust more with the NHS vs who would improve healthcare more, whereas the best PM question is pretty standard across all pollsters. And what this page is "supposed to be about" is not fixed - it is determined by editors in collaboration. Absolutelypuremilk (talk) 23:32, 16 January 2017 (UTC)
Indeed, Absolutelypuremilk is right. And we can't divorce this page from the practical reality of British elections - party leaders are front and centre in campaigns, they're treated as semi-Presidential candidates and the party leader can make or break electoral campaigns. As I said above, in multiple similar pages, 'best PM' is included as an important piece of data. Hence why I suggested it as an option to see if there was any support for including it - and there was. FriendlyDataNerd (talk) 15:29, 18 January 2017 (UTC)
"Party leaders are front and centre" is an assumption which I'm not sure is backed up by the data. Nigel Farage used to win polls for the best pm question, but this never meant UKIP had any chance of winning, because the only polling question with any relation to the current article title is the question on voting intention. If people want to change the title to "Contempoary British Political Opinion Polling" then go for it.Awoma (talk) 14:31, 19 January 2017 (UTC)

UK Supreme Court Ruling

I have removed the following note from the 2017 table:

"The Supreme Court rule that Parliament must be given a vote on whether or not to trigger Article 50 and begin the process of leaving the European Union."

This does not properly reflect the case that was being heard, nor does it reflect the outcome in an encyclopedic manner. It reads like a news headline or the opening sentence of a journalistic article. Wikipedia is not news, it is an encyclopedia. Please consider the following points:

1. In short The UK Supreme Court were hearing an appeal by the UK Government to determine whether they could use the Royal Perogative to trigger Article 50. The need for this essentially came about as the European Union Referendum Act 2015 failed to give legal clarity in the event that a Leave vote prevailed and because a barrister (Gina Miller) decided to bring the case forward in her own name to say that this was unlawful (she won); I point his out because Parliament was not a party to this hearing, so the court could not therefore possibly rule in Parliament's favour or "give" it anything.
2. This case and the motives behind it are suppose to have nothing to do with the referendum result (so we're told) and so we must treat it as such, therefore the "whether or not" part of the sentence has no relevance. As the Secretary of State and indeed many on the remain side keep pointing out that "we already passed the point of no return on 23rd June"; whether Article 50 will be enacted is not in question here, nor was it in the court proceedings (Judges were at great pains to point this out and that it wasn't for them to say) it is a question of process(how?).
3. What the UK Supreme Court actually established is how the Government triggers Article 50 and they have said that needs to be done via an Act of Parliament and therefore be voted on by MPs and Peers alike.
4. The Supreme Court also heard another case simultaneously; whether or not the devolved nations would have a veto on the UK Government triggering Article 50. They ruled that they did not as the UK constitution makes it clear that the UK Parliament is responsible for such matters. I think this furthers the need to use the phrase "An Act of Parliament" as it covers the point about MPs and Peers "getting a vote" and that the devolved legislatures do not.
The question for us is: How do we put this into a one liner? And does it have relevance in the table. I would lean towards yes it does because of the constitutional significance of the case alone but there are a number of other factors that loosely back this up, for instance this is something that many members of the public have had strong opinions on and this whole question has shaped the national debate and in turn affected party polling.

May I suggest the following sentence;

"The UK Supreme Court conclude that the legal mechanism for beginning negotiations and Withdrawal from the European Union: Article 50 is to be enacted via an Act of Parliament."

What are people's thought's on the above sentence? Please, if anyone else has an alternative or an improvement, please suggest away but it is difficult to be encyclopedic and succinct. Thanks 2407:7000:8757:FD36:F851:DF0A:A60F:8FFE (talk) 04:03, 25 January 2017 (UTC)
I'm against any non-election notes in the table. There's a (small) discussion about whether to even include by-elections, we don't need this stuff. 92.10.149.1 (talk) 05:07, 25 January 2017 (UTC)
That's an opinion; fair enough, so far myself and the original poster are in favour. My point is that it is of constitutional significance. Personally, I agree that excessive notes are a nuisance but I think it is a question of where we draw the line. I take and respect your point though. 2407:7000:8757:FD36:F851:DF0A:A60F:8FFE (talk) 10:22, 25 January 2017 (UTC)

I also think that this is not needed - but if it is to be included then the current one is better as it is shorter and more understandable. Absolutelypuremilk (talk) 10:44, 25 January 2017 (UTC)

I don't know if you have read my argument or indeed the judgement but the original form of the note does not properly reflect the judgement. There is very little difference between the length of my note and the previous note, so I don't see your criticism here. I also don't take the point of being more understandable; the whole problem with the old note is that portrays a spun understanding of the judgement, mine does not. 2407:7000:8757:FD36:F851:DF0A:A60F:8FFE (talk) 18:59, 25 January 2017 (UTC)
This is absolutely not needed. This is not a timeline of political events. It is a table of poll results. We have previously agreed to limit notes to by-elections and party leader changes. Inserting anything else is WP:SYNTHy without reliable source evidence that the event impacted on polling.
Take a look at the following articles and see how many notes they use in their polling tables: Opinion polling for the French presidential election, 2017, Opinion polling for the next Italian general election, Opinion polling for the next Spanish general election, Nationwide opinion polling for the United States presidential election, 2016, Opinion polling for the German federal election, 2017, Icelandic parliamentary election, 2016, Next Icelandic parliamentary election. Absolutely none. I can give a dozen more of these.
Wikipedia presents reliably sourced material. We do not editorialise. If you find some citations discussing how the polls have shifted in response to events, write some prose using that material, but let the table of polling results be a table of polling results. Bondegezou (talk) 14:36, 25 January 2017 (UTC)
Like I say, I have no firm view either way on whether it should be included or not; I understand that it is a nuisance to be putting notes in on a wide range of things. The constitutional significance of the judgement is what was swinging it for me. My main point was the content of the note, if there was to be one. 2407:7000:8757:FD36:F851:DF0A:A60F:8FFE (talk) 18:59, 25 January 2017 (UTC)

Feb 2016 & Dec 2016

is missing on the graph...

Thanks, the next update will include them. Absolutelypuremilk (talk) 18:55, 8 February 2017 (UTC)

2016 Labour leadership challenge

I added two lines to the spreadsheet concerning the launch and conclusion of the Labour Party (UK) leadership election, 2016 as I feel these events should be included (despite this period already being awash with other major events) since they coincide with dramatic increase in Conservative poll lead which might otherwise look to be solely due to the election of Teresa May to Prime Minister and Conservative leader, although this may be contributory or indeed the main cause. Another event to possibly include would be the mass shadow cabinet resignations on the 26th-29th June 2016 however I felt this might be overkill. — Preceding unsigned comment added by 143.167.63.235 (talk) 13:16, 1 March 2017 (UTC)

I disagree with including leadership challenges here - we already have Corbyn being re-elected and I haven't seen any evidence that the leadership challenge affected his ratings. Absolutelypuremilk (talk) 16:04, 1 March 2017 (UTC)
Other polling articles on Wikipedia do not include any events at all. I feel we should do the same, or keep them to the minimum. This is not a timeline article, and inserting events into a list of polls is WP:SYNTHy, implying those events had some impact on polling. We shouldn't be doing that, least not without reliable sources saying there's a link. Bondegezou (talk) 16:47, 1 March 2017 (UTC)
I won't make further edits other than poll additions but I find your lines of reasoning troubling: If the events are not there to indicate what may have happened to have caused poll movement, why are they there at all? For example, it seems strange to list the three leadership changes at UKIP and a further one with the Greens in the second half of 2016 without any mention of Corbyn's re-election or leadership challenge. My concern is that implication of these events having a causal link with polling is inevitable so either the events need to be consistent or need to be removed. I would certainly contend that the resignation of Diane James is much less important than either the leadership challenge or Corbyn's re-election, for instance.
I would be happy to see every event removed from the table. That's how polling tables are done on numerous non-UK Wiki articles. It is only the UK articles that ever felt this was appropriate. Current consensus, however, is for other elections and new leaders. I agree there are problems with that, but adding to that list increases the problems. It doesn't decrease them. Bondegezou (talk) 13:34, 2 March 2017 (UTC)

Feb 2017

Feb 2017 is missing from the graph, it goes straight from January to March, can you correct please. — Preceding unsigned comment added by 86.183.221.109 (talk) 09:50, 13 March 2017 (UTC) Thanks, done. Absolutelypuremilk (talk) 12:02, 13 March 2017 (UTC)

Notes in table

This is a subject which repeatedly comes up & is discussed, in the interest of heading towards a clearer consensus, this is my view of most to least appropriate to have in the table (other than polls). Where the line should be drawn is not something I have a strong opinion on (though I'd nominate it between 5&6).

1. The General Election Results (both 2015 and 2020 when it comes).
2. Change of government party/coalition (theoretical, but could happen via by-elections or defections)
3. Change of Prime Minister. (should also be new leader for governing party)
4. By Election results (this page is essentially polling for the house of commons, and these are elections to it).
5. National (or nationalish) elections and referendums results. (eg local council elections, EU referendum)
6. Regional non-Commons election results (London assembly, NIish, Welsh, Scottish)
7. Change in leader of opposition.
8. Change in party leader for main (table named) parties (one note on the day leader takes over).
9. further detail/dates on campaign/election procedures (eg "Labour party leadership election campaign begins, leader X resigns").
10. Notable Events (killing of Jo Cox).
11. Non-event news/statistics (eg, economic data).

92.10.152.126 (talk) 17:12, 4 March 2017 (UTC)

I'm not certain that order is quite right, but thanks for laying things out clearly.
My preference remains fewer notes in general: other countries' polling articles don't use them at all and they are WP:SYNTHy. Most polling articles do 1 and nothing else.
My understanding of consensus here lately has been yes to 1-5, 7 & 8, but no to 9-11 and unclear on 6. (But consensus can change.) Bondegezou (talk) 09:50, 5 March 2017 (UTC)


NI Assembly elections

Should the NI Assembly elections be included in the table as a note? The opinion polls listed are almost all GB-only polls and Northern Ireland has a completely different political system with parties that do not stand in GB. It appears completely irrelevant. Saxmund (talk) 17:36, 4 March 2017 (UTC)

I've changed the heading hierarchy as this is closely related to the above discussion. I note the Conservatives and UKIP stand in NI, although they had almost no impact in the latest elections. Bondegezou (talk) 09:52, 5 March 2017 (UTC)


So surely the resignation of Diane James October 2016 (as distinct from Paul Nuttall becoming UKIP leader) should be removed as not being noteworthy enough as per number 8? (I'm not making further edits other than polling data). Also I'll point out that Number 9 didn't go over well when I added it (see above). However, I'm in favour of the events as they give important context to the polling data. I believe the argument put forward that the page is not meant to be a timeline is spurious. By including historical polling information, over time, makes this page a de facto timeline, else only the most recent polling data should be shown which I think all here would find unsatisfactory. — Preceding unsigned comment added by 143.167.63.235 (talk) 11:13, 7 March 2017 (UTC)
My concern is that any claim that events give important context to the polling data should be backed up by reliable source citation. It should not be up to Wikipedia editors to make those decisions: that way lies WP:OR. Bondegezou (talk) 11:50, 7 March 2017 (UTC)
I think it'd be possible to find citations that many things give context. Budgets for example...the Omnishambles is thought to be one of the most poll-affecting things of the last parlialment, which I'm sure could be cited to ukpollingreport & others. I think we should draw the line higher than that (and it is up to wiki editors where the due weighting lies) or the table'll be a lot fuller. I don't think "other countries do/don't it" is a particularly strong point. I'd personally exclude the NI&Scottish&Welsh&London elections, but I'm not that bothered about it. 79.74.2.33 (talk) 03:14, 15 March 2017 (UTC)

Terrorist attacks at Westminster

Wondering if the killings at Westminster ought to feature on the polls? VelvetCommuter (talk) 01:00, 23 March 2017 (UTC)

I don't think so, I think we should keep it limited to things that affect the parties themselves. Absolutelypuremilk (talk) 09:40, 23 March 2017 (UTC)
I also think that the terrorist attacks should not be included. 137.222.248.20 (talk) 12:44, 23 March 2017 (UTC)
Agreed, I don't feel that there's any political significance in this case. Mélencron (talk) 13:15, 23 March 2017 (UTC)
Sorry for having edited the article including the attack. I had not seen this section. 5.34.154.217 (talk) 14:57, 23 March 2017 (UTC)

I've removed it. It's irrelevant. And there's no consensus for including it.Garageland66 (talk) 08:35, 24 March 2017 (UTC)

Some sort of error in the bottom result in this section/only result in next section. Seems to be the same poll twice with different figures. --Dweller (talk) Become old fashioned! 14:52, 16 March 2017 (UTC)

It's against Cameron, as opposed to May. Mélencron (talk) 15:22, 16 March 2017 (UTC)
Hmm. I think that needs to be explained. --Dweller (talk) Become old fashioned! 15:30, 16 March 2017 (UTC)
I think it is: "Some opinion pollsters have asked voters which party leader they would prefer as Prime Minister – Theresa May (Conservative Party) or Jeremy Corbyn (Labour Party). Most pollsters also include 'Not Sure' as an option, while others also provide 'None of these' as an option. Polls published prior to Theresa May’s election as Conservative Party Leader offered then-Prime Minister David Cameron as the Conservative option." Mélencron (talk) 15:33, 16 March 2017 (UTC)
I don't see that text anywhere near the confusing lines, either as prose or as a footnote. --Dweller (talk) Become old fashioned! 09:34, 17 March 2017 (UTC)
I was in two minds about whether to include that poll for May. On the one hand, it's only hypothetical and she wasn't yet Prime Minister. On the other hand, it's a fact that YouGov did ask the same question that they went on to ask once she became PM, and picking and choosing which one to include wasn't something I wanted to do by myself....I've added a clearer footnote FriendlyDataNerdV2 (talk) 18:43, 24 March 2017 (UTC)

For those wanting to add all sorts of dates and events to the chart...

I refer you to the most recent, brief 'discussion', which was here.

Consensus can change on Wikipedia, but to ascertain consensus, you need to discuss not edit-war. --Dweller (talk) Become old fashioned! 10:08, 5 April 2017 (UTC)

Requested move 19 April 2017

The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review. No further edits should be made to this section.

The result of the move request was: Closing as the article has already been moved to the correct title Number 57 08:51, 20 April 2017 (UTC)


Opinion polling for the next United Kingdom general electionOpinion polling for the 2017 United Kingdom general election – Now that the date for the election has been confirmed by Parliament, I feel that the name should now be changed to the aforementioned name. The proposed name redirects to this page, which is why I can't change it myself. Clyde1998 (talk) 18:36, 19 April 2017 (UTC)

The appropriate title in this case should conform with the title of the election article (i.e., move to Opinion polling for the United Kingdom general election, 2017). This is an uncontroversial move which shouldn't require an RM request. Mélencron (talk) 20:07, 19 April 2017 (UTC)
Agree per Mélencron. Charles Essie (talk) 20:14, 19 April 2017 (UTC)
Agreed I moved it, but feel free to revert me. I don't think this RM should remain open, though. -Rrius (talk) 20:17, 19 April 2017 (UTC)
The title should be "Opinion polling for the 2017 United Kingdom general election", not "Opinion polling for the United Kingdom general election, 2017". As well as the former being fairly obviously the more natural way to say it, it also makes it clear that the election is in 2017, not just the polling. 109.145.3.120 (talk) 04:29, 20 April 2017 (UTC)
Yes, but the current title (Opinion polling for the United Kingdom general election, 2017) is consistent with the titles of other articles about polling in the run-up to elections. MFlet1 (talk) 08:25, 20 April 2017 (UTC)
Support move per clarity and consistency with other articles on opinion polling. --Checco (talk) 06:06, 20 April 2017 (UTC)

The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.

@User:Number 57: Uhm no, the article was not moved to the proposed name. I support the article's moving from the current name ("Opinion polling for the United Kingdom general election, 2017") to the proposed one ("Opinion polling for the 2017 United Kingdom general election"). As there is some confusion on the issue, I am asking to you to re-open the debate in order to understand what users really prefer. Many thanks, --Checco (talk) 10:10, 20 April 2017 (UTC)

@Checco: As pointed out by Mélencron and MFlet1, the current title is correct as it matches United Kingdom general election, 2017 and is also consistent with other opinion polling articles, e.g. Opinion polling for the Dutch general election, 2017, Opinion polling for the German federal election, 2017, Opinion polling for the Spanish general election, 2016 (see Category:Opinion polling for elections). I'm not sure there's any confusion? Cheers, Number 57 10:58, 20 April 2017 (UTC)
Ok, fair enough, even though I prefer the other option. Indeed, why was proposed the opposite move? --Checco (talk) 12:20, 20 April 2017 (UTC)

The current way we put in the detailed data for the opinion polls is liable to wp:linkrot.

Would anyone have an issue if instead of using something like this

[https://www.icmunlimited.com/wp-content/uploads/2017/04/2017_guardian_campaign_poll3.pdf ICM/''Guardian'']

we use something like this

[[ICM]]/[[Guardian]]<ref>{{cite web|title=The Guardian – April (3) 2017|url=https://www.icmunlimited.com/wp-content/uploads/2017/04/2017_guardian_campaign_poll3.pdf|website=https://www.icmunlimited.com/|accessdate=25 April 2017|format=PDF}}</ref>

this way the bots can get the pages archived (which won't happen otherwise) => Spudgfsh (Text Me!) 19:21, 25 April 2017 (UTC)

Date to accompany graph

I have previously been updating the graph with "includes all polls starting before or on xx April", but now that we are in election season and there are multiple polls starting on the same day, which are then released a long time apart, this doesn't really make sense any more. Can anyone suggest a better way to do it? I thought of saying "all polls released on or before xx" but then its sometimes hard to tell when exactly the polls are released. Absolutelypuremilk (talk) 07:50, 27 April 2017 (UTC)

Missing ORB poll 19-20 April 2017

Here's an ORB poll from 19-20th April that is missing from the table - http://www.orb-international.com/perch/resources/orb-telegraph-poll-april-22nd-2017.pdf — Preceding unsigned comment added by 2601:643:8004:AB20:48DE:27A9:AD32:A134 (talk) 20:28, 29 April 2017 (UTC)

Constituency Predictions

Electoral Calculus have compiled data from recent polls to cross referenced with political and demographic indicators to give a full seat prediction. This is obviously a working document, with the data set here: http://www.electoralcalculus.co.uk/electdata_pred.txt Is this something we should include on this page as seat numbers matter more than popular vote in UK elections? Or is it not concrete enough and too clunky? In 2015 we had a page for constituency opinion polls. CarlDurose (talk) 10:32, 25 April 2017 (UTC)

I've always been interested in including this (as per Opinion polling for the next Spanish general election). But the problem has always been that there are a lot of different constituency predictors, and there's no one-size-fits-all way of entering poll data to get predictions. It essentially crosses into the area of original content.
I would support including seat predictions if there's a reliable source that has made such predictions. We can certainly have a Wales section as ITV and Prof. Roger Scully have started predicting seat outcomes: http://www.itv.com/news/wales/2017-04-24/shock-poll-suggests-tories-ahead-in-wales/ FriendlyDataNerdV2 (talk) 16:46, 25 April 2017 (UTC)
There have been such things included in the past (see the United Kingdom general election, 2015 article) but most of them were so far off that they stopped doing it. Even Nate Silver (fivethiryeight) decided not to this time around. The only one I know is doing anything consistently is electoral calculus. => Spudgfsh (Text Me!) 17:06, 25 April 2017 (UTC)
There's been the 1st constituency poll of the election, in Kensington, London - http://www2.politicalbetting.com/index.php/archives/2017/04/30/new-polling-suggests-that-con-london-strongholds-could-be-vulnerable-to-stop-brexit-candidates/ Time for a "Opinion polling in United Kingdom constituencies, 2015-17 article"? FriendlyDataNerdV2 (talk) 14:34, 30 April 2017 (UTC)
I'd add a section here for now and if it gets too large then create a separate article. Lord Ashcroft did most of the previous ones but doesn't appear to be doing any this time so I don't think it'd overwhelm the article and having an article for the few that are being done seems OTT.=> Spudgfsh (Text Me!) 14:41, 30 April 2017 (UTC)
Good plan, will do FriendlyDataNerdV2 (talk) 14:48, 30 April 2017 (UTC)

Logos on column headings

I saw the party logos added to the column headings by JackWilfred and then reverted by DrArsenal. I wouldn't say I had a strong opinion, but personally I thought they looked quite good. I also read in the page history the rationale for reverting them. Yes, some logos are bolder than others. But then, some party colours are bolder than others. The images loaded quite quickly on my machine. What do others think?

Date(s)
conducted
Polling organisation/client Sample size UKIP SNP Others Lead
Con Lab UKIP Lib Dem SNP Green

--Wavehunter (talk) 17:53, 28 April 2016 (UTC)

To me the Logos were a positive enhancement of the page. Practically everybody's internet connection can easily cope with these very basic graphics. The fact that some stand out more than others is not for us to judge, the parties they represent are happy to stand by them.-- BOD -- 18:46, 28 April 2016 (UTC)
I have probably overstepped the mark but I have reverted back to versions with logos. I agree with his rationale about logos, and I will look to change some logos to just the central icon, for example just the bird for the Liberal Democrats. However, we don't live in the days of dial-up internet, a bunch of thumbnail sized images will make no real different to page load time. JackWilfred (talk) 20:14, 28 April 2016 (UTC)
British parties' logos are copyrighted, and thus can't be freely used (check WP:NFCC#10c). This is in contrast to Spanish parties' logos, most of which are so simple they don't meet the threshold of originality and, thus, don't require a fair use rationale. It's not a load time issue. Despite I'm all up for using party logos, since they do indeed make an important improvement to article's quality, they can't be used randomly in many cases, as policy does not allow it. Only the SNP logo could be used, since it does not meet the threshold of originality. Impru20 (talk) 20:39, 28 April 2016 (UTC)
I don't think the logos are needed - it distracts from the figures. Absolutelypuremilk (talk) 21:37, 28 April 2016 (UTC)
They don't necessarily do it, though I acknowledge that it could in this article, since font size used for the tables is very small. Impru20 (talk) 21:40, 28 April 2016 (UTC)
By analogy with MOS:FLAG, I'd say leave them out. Bondegezou (talk) 22:59, 28 April 2016 (UTC)
MOS:LOGO is perfectly clear "The insertion of logos as icons into articles is strongly discouraged: While illustration of a logo may be appropriate at the main article on the topic to which the logo pertains, use of logos as icons is not useful to our readers, and often presents legal problems."DrArsenal (talk) 23:05, 30 April 2016 (UTC)
What about pictures of the party leaders above? Necessary or not?User:TomPumpkin69 (talk) 21:05, 24 April 2017 (UTC)
Party leader's photos will spoil the page, logos I think will add something just as a newspaper or TV show show logos when talking about opinion poll ratings. Regarding copyright, they already exist on individual pages so surely that's not a problem? Paulharding150 (talk) 10:27, 26 April 2017 (UTC)
If we reject party leader photos (and I agree) can we remove them from the Preferred PM section? It's stretching out the boxes and distracting from the numbers themselves. I designed the tables for the Preferred PM section and I did not design them with party leader photos in mind FriendlyDataNerdV2 (talk) 11:03, 26 April 2017 (UTC)
If it is really copyrighted then why is the logo on the wiki page for the labour party? Russell's teapot (talk) 07:58, 3 May 2017 (UTC)

Yep taking out the party leader photos is fine with me. Absolutelypuremilk (talk) 11:27, 26 April 2017 (UTC)

Switch LibDems and UKIP

Lib Dems and UKIP should be switched on the tables. This is due to the polling showing the Lib Dems with higher polling since the general election was called, much more accurate polling. This combined with the fact that the Lib Dems are the 3rd largest national party by parliamentary seats, (besides SNP, but of course, they have far less of the vote and are benefitted by FPTP). The Lib Dems have historically been the 3rd party of the UK after the Conservatives and Labour. They are also the 3rd largest party in terms of the number of local council seats. Finally, they have over 100,000 party members, far more than UKIP. I feel this information should be reflected in the tables.

Ie. the tables would display the Conservatives on the farthest left, the Labour Party 2nd from left, the Lib Dems 3rd from left, UKIP 4th from left, and then the rest as normal.

The order of the table is on votes at the last election. If the Lib Dems win more votes than UKIP at this election, then the order will be swapped in the article about polling for the next election. Absolutelypuremilk (talk) 16:11, 28 April 2017 (UTC)
I think it's more complicated than that. Talk:Opinion polling for the United Kingdom general election, 2015/Archives/2013/January is the archive of the discussion of when UKIP got added to the equivalent polling article before the last general election. Many of the arguments then for more prominent display of UKIP's performance now work the other way in favour of the LibDems. That said, it's early days. I don't feel any change is required yet. Bondegezou (talk) 16:26, 28 April 2017 (UTC)
I'd suggest that order changes be done at year break or election breaks only, not third way through the table. The GE coming up, no brainer to do it then...when I think nation wide parties ordered by vote share makes sense. I know that wasn't exactly how it worked originally, but that was an entirely new addition for a party exploding from barely-nothing, not re-ordering (we didn't swap lab&con around last parliament, for eg). 79.74.27.36 (talk) 03:34, 29 April 2017 (UTC)
I don't think the order of parties should be changed. It's based on performance in the previous election; there isn't a good reason to change that – as you noted, the positions of parties change over time, and therefore the ordering would change within the same article, which I don't consider a favorable outcome. This is the most neutral possible solution (the other would be to refer to the order given by the pollsters, but those are inconsistent). Mélencron (talk) 03:38, 29 April 2017 (UTC)
If we introduce a new table for this election, then by all means move the LDs to third in the table. If not just leave them at fouth. I don't think its going to mislead too many people even with ukip fighting just a handful of seats this time. - Galloglass 07:36, 29 April 2017 (UTC)

I agree with the voices arguing to keep the tables in the order set by the last election and to follow suit next time, promoting LD over UKIP then if they do indeed perform better. --Dweller (talk) Become old fashioned! 10:10, 3 May 2017 (UTC)

If the Lib Dems outperform UKIP (in terms of votes) at the election, I would support changing the order; but until they do, I find the suggestion at changing the order a little bit pointless. Brough87 (talk) 10:38, 5 May 2017 (UTC)

YouGov regional figures

These are appear to be collated figures from regional breakdowns of national polls rather than proper polls of each region in question. While this sort of information is probably the only we're likely to get for the various English regions, presenting them as proper polls is highly misleading. — Preceding unsigned comment added by 95.147.72.238 (talk) 22:12, 15 May 2017 (UTC)

95.147.72.238 You make a good point. Can you suggest some introductory text? --Dweller (talk) Become old fashioned! 09:23, 16 May 2017 (UTC)
Information here: https://yougov.co.uk/news/2017/05/15/voting-intention-regional-breakdown-apr-24-may-5/
While they might be regional samples of a national poll, it's a national poll with 14,395 respondents and as the focus is on the regions we can assume they have been properly weighted. Happy to add some intro text but I'm not clear on why these are different to other polls - they're weighted and sampled the same way FriendlyDataNerdV2 (talk) 13:44, 16 May 2017 (UTC)

Wrong figure for SNP

Just a small thing, but as I'm unable to edit the page I thought I'd point it out here - the SNP are incorrectly listed as having 3% in the latest GB-wide Panelbase poll. It should be 5%. — Preceding unsigned comment added by 82.11.84.165 (talk) 23:46, 17 May 2017 (UTC)

The others figure is wrong also. It should be "*", indicating less than 1%. — Preceding unsigned comment added by 2601:643:8004:AB20:11C4:B4A6:E28E:CE4F (talk) 01:11, 18 May 2017 (UTC)
Fixed. Mélencron 01:32, 18 May 2017 (UTC)

Missing poll from graph

I'd add that there is currently only one point showing above 32.5% for Labour but 2 recent polls at 33% and 34%. I've been waiting for the plot to update to check if the 33% point remained missing and that still seems to be the case. 89.241.28.240 (talk) 12:38, 18 May 2017 (UTC)

Thanks for pointing this out, I have added this poll to the graph now. If you see any more issues with the graph then please feel free to bring them up. Absolutelypuremilk (talk) 10:20, 19 May 2017 (UTC)

Moving average graph

The graph is based on the last 10 polls, therefore the point at which the average line ends should be midway in time between the first and last of the 10 last polls. At present the graph appears to end on the same date as the last poll, leading to an inability to see how the latest 5 polls will influence the next calculation of the moving average. - Ânes-pur-sàng (talk) 09:23, 18 May 2017 (UTC)

I don't understand this objection. The lag would be at the start of the graph, not the end. --Dweller (talk) Become old fashioned! 09:48, 19 May 2017 (UTC)
The moving average is what the average of the last ten polls is at a certain point in time, so I think it should go at that point in time (i.e. at the last of the ten polls). This is how I have calculated the graph. Absolutelypuremilk (talk) 10:22, 19 May 2017 (UTC)

Lord Ashcroft polling

He's back (see also his blog). Not sure how we'd handle this, but think it should be in there. --Dweller (talk) Become old fashioned! 14:02, 15 May 2017 (UTC)

Link doesn't work for me – I assume it's the massive constituency breakdown? I think it's merely a model (i.e., "results" in one constituency influenced by other similar ones) and don't think anything other than the toplines (which were only for preferred PM) should be included as such. Mélencron 20:05, 15 May 2017 (UTC)
Blog is here. Link to the breakdown is halfway through, but try this. It won't fit our list of polls, but is perhaps the first of what will be a whole load of pollsters' predictions of the result: Con 406-415, Lab 152-164, LD 8-14, SNP 45-48, PC 4-5, Greens 1, UKIP 0, Others 19. Con majority 162-180. I suppose it's arguable that it doesn't go on this page as it's not opinion polls, but it's interpreting data culled from opinion polls. Any other views? --Dweller (talk) Become old fashioned! 09:27, 16 May 2017 (UTC)
Last time round, we had a separate table for various pollsters' and psephologists' predictions. Bondegezou (talk) 13:44, 16 May 2017 (UTC)
Let's do it again. --Dweller (talk) Become old fashioned! 14:21, 16 May 2017 (UTC)
Shall we put the prediction table in this article or the GE2017 one? In 2015 it was in the main one FriendlyDataNerdV2 (talk) 20:17, 16 May 2017 (UTC)
Probably in the main one, as most of the predictions are not polls. Mélencron 20:23, 16 May 2017 (UTC)
Sounds sensible. To help readers, we could add a pointer in this article to the predictions in that one. --Dweller (talk) Become old fashioned! 09:27, 17 May 2017 (UTC)

89.241.28.240 (talk) 15:25, 19 May 2017 (UTC) (Posting for information only.)The tables for the 40k sample done for Ashcroft's predictions are up but they aren't weighted by demographics. His prediction model appears sound but the polling oversampled Conservative and Lib Dem 2015 voters, Remainers and women, and undersampled Labour 2015 voters so it will be very misleading without weights, I'm not suggesting it be counted as a poll, just in case the link was of use/interest.

Position of most recent election result

This concerns a wider discussion of a template, so this might not be the best place to post. I'm relatively new to editing, so apologies if this isn't the appropriate forum.

Is there a particular thought (aside from chronological reasons) for posting the most recent election results at the bottom of the table? It seems that they would be better placed at the top, to compare an election result with the most recent polls, rather than endlessly scrolling between the very bottom and the very top.

It could mean adding a separate table for the results, rather than leaving them in the top row of the polls, for ease of editing. — Mouldyfox (talk) 11:46, 12 May 2017 (UTC)

Some other articles post them at the top and the bottom. I'm in favour of that. Bondegezou (talk) 12:20, 12 May 2017 (UTC)
Should I go ahead and change it, or wait for more of a consensus? Not sure of the etiquette here. —Mouldyfox (talk) 14:57, 12 May 2017 (UTC)
Feel free to change it, as per WP:BOLD. If other people don't like it, they can always change it back and/or we can discuss it more here. Bondegezou (talk) 15:41, 12 May 2017 (UTC)
It feels weird to have election results, out of date order, in a table of opinion polls. I wouldn't have it at the top, but it's okay at the bottom. Robertshippey (talk) 16:28, 12 May 2017 (UTC)
Looking at the article for the 2015 Canadian election, we put the results of the previous election at the bottom, and, once the 2015 results were in, we put those at the top. Makes more sense I think. Dbrodbeck (talk) 17:56, 12 May 2017 (UTC)

I would prefer to have the old results at the bottom. Absolutelypuremilk (talk) 18:49, 12 May 2017 (UTC)

I'm not sure why people feel that way, other than it feels weird. As I've said, I think it makes sense to be able to directly compare solid results against opinion polls without constant scrolling, especially in the run-up to an election when polls become near-daily and the list bloats, but I'll leave it at that. — Mouldyfox (talk) 18:54, 12 May 2017 (UTC)

I had a look at what some other articles on forthcoming elections do:

However, I have seen putting the previous election results at the top as well in the past! Bondegezou (talk) 18:55, 12 May 2017 (UTC)

I can see how it'd be useful, but it's a bit weird too...they'd surely not stay there after the election. Perhaps a note of them in the graphical summary section, just below the graph, would make sense and be viewed each time you checked most recent polls. 79.74.27.100 (talk) 22:25, 12 May 2017 (UTC)
Agreed, they wouldn't be needed after the election, I thought it would serve as a useful comparison until the new election. Perhaps it would be better in a separate table above the opinion polls. — Mouldyfox (talk) 08:38, 13 May 2017 (UTC)

89.241.28.240 (talk) 17:16, 19 May 2017 (UTC) One of the best things about the graphic is that it starts from the election result (a lot of the moving averages don't and it's horribly misleading given the size of the 2015 polling miss) so perhaps the easiest solution is to state that explicitly with the graphic?

Sub-national polling for England

Where has the table with the polling for England gone? - - 2A00:23C4:C2FA:5400:CDA1:3EFB:C1DF:C871 (talk) 12:59, 17 May 2017 (UTC)

In the history it says: "On the basis that there have not been any England-specific polls (only sub-samples) I have removed this England section until an England poll is specifically published." — MouldyFox (talk) 13:03, 17 May 2017 (UTC)
Doesn't sound convincing. Are we going to remove all of that as well? --2A00:23C4:C2FA:5400:147E:31F5:3678:96E1 (talk) 02:30, 20 May 2017 (UTC)
The thing is, no-one had updated for months, it wasn't informative and none of them were actually polls in their own right. I'm going to sit down and comb through all polls since 2015 for the English sub samples but it's a big task and in the meantime that table isn't actually informing anyone of current polls. If someone wants to take on that task I'm ok with it going back up, but this isn't like the Welsh and Scottish sections - we don't include Scottish sub samples in the Scottish section....
As for the 2015 one, that was updated frequently and contains informative information. The English section here was about 70% events and virtually no polling data. FriendlyDataNerdV2 (talk) 13:21, 20 May 2017 (UTC)

Preferred Prime Minister polling

I desperately want to believe that this section is relevant to the article, but UK elections are not a vote for a Prime Minister, they are a vote for a candidate in local constituencies. This is evidenced by the fact that two of the last three Prime Ministers have assumed office having never faced a general election where the public could either endorse or reject them for this role, whether directly or indirectly.

In addition, the column showing the poll leads is highly inaccurate, particularly in the May V Corbyn table, where Ms May's lead is exaggerated by showing only how many points she is ahead of or behind Mr Corbyn and not the second most popular option - which on many occasions is 'neither of these'.

This is not a two option question, and it is not a race between two candidates for the job of Prime Minister.

I move that this entire section be removed, or failing that, that the poll leads are corrected to show the lead over the second most popular option and not just the other candidate.

iMarc89 (talk) 03:09, 21 May 2017 (UTC)

There have been quite a few of them, and it is about the election, it makes no sense to remove it. And, yes, I know that there is no direct vote for the PM etc. Do you have a policy based reason to remove it? Dbrodbeck (talk) 03:26, 21 May 2017 (UTC)
Psephologists have argued that preferred Prime Minister is significant and can be a better predictor of voting than the normal polling question. Ergo, I think we should keep it in. I'll look for some background papers. Bondegezou (talk) 16:20, 21 May 2017 (UTC)
Quite. My argument for including them was/is that multiple sources cite them as relevant to the election, that UK political campaigns have seen leaders play an important role (see: the leadership debates) and that pollsters are asking the question so often and so uniformly that ignoring it would seem to making a value judgement on what is and isn't relevant to the election.
I appreciate that that is also an argument for including the 'Approval/Disapproval' polls on party leaders, but as the questions are so radically different from pollster to pollster, are not asked as frequently and are not cited as frequently, I'd be fine not including them.
As for "This is not a two option question, and it is not a race between two candidates for the job of Prime Minister" - that's not the Wiki making that judgement, that's the pollsters when they only include May and Corbyn in their questions. There are a handful of multi-option polls and they are included. But for the most part, polls assume this election to indeed be "a race between two candidates for the job of Prime Minister". Now that might be unfair, but it's not us making that call.
As for the poll lead - I took my lead on that from pollsters and newspapers themselves who report the leads as May OVER Corbyn. Example: http://www.standard.co.uk/news/politics/theresa-may-would-make-a-better-prime-minister-than-jeremy-corbyn-new-poll-suggests-a3511831.html << "The current prime minister had a 37-point lead on Labour leader Jeremy Corbyn" FriendlyDataNerdV2 (talk) 18:10, 21 May 2017 (UTC)

Semi-protected edit request on 22 May 2017 - Suggested correction on the recent ICM poll 19-21 May

Shouldn't the Tories have 48% not 47% and UKIP 3% not 4% in the ICM poll (see page 11 of the report for the 19-21st May)? Markroknight (talk) 15:58, 22 May 2017 (UTC)

I'm not sure which report you are looking at, page 9 has the voting intention figures which match those in this article. Absolutelypuremilk (talk) 16:06, 22 May 2017 (UTC)

Add vertical gridlines to plot?

92.7.53.114 (talk) 11:18, 23 May 2017 (UTC)I realise this is a style-guide issue that affects more than just this page. Vertical gridlines to show where each new month starts makes plots like this considerably easier to read, especially when shifts after particular events are of interest, as they are here. In general, I'd argue that horizontal and vertical gridlines should be the default unless there is a good reason for omitting one or both.

Manchester Bombing

Hi there - I included the 2017 Manchester Bombing not because it is simply another event in UK current events, but because it has affected the election directly, with campaigning suspended. The murder of Jo Cox is similar in the sense that it affected the contemporary election campaign. VelvetCommuter (talk) 11:19, 24 May 2017 (UTC)

I agree, the fact that campaigning is suspended means that this directly affects the election. Absolutelypuremilk (talk) 11:35, 24 May 2017 (UTC)
Lots of things affect the election, but consensus has been that we restrict such additional lines on the polling table, as discussed previously. This is a table of polling results: it is not a timeline of events. And it would be original research to presume any such event directly impacts on polling.
Look at the following articles and look how they use additional lines on their polling tables: French_legislative_election,_2017#Opinion_polls, European_Parliament_election,_2009_(United_Kingdom)#Opinion_polls, Opinion polling for the Spanish general election, 2016, Opinion polling for the Dutch general election, 2017. I could list two dozen more. Why should we do things differently here? Bondegezou (talk) 12:37, 24 May 2017 (UTC)
I think it should be included, as well major television debates and "Major parties publish their manifestos". I think a few of these major events are helpful to understanding in such a large table - larger than most of the other examples you highlight. Could you link to any previous discussion of this topic? Rwendland (talk) 12:46, 24 May 2017 (UTC)
I agree - we should include the major TV debates, the publication of manifestos, and the suspension of campaigning following the bombing. This is broadly in line with the Opinion polling for the United Kingdom general election, 2015 page, which includes a note for each TV debate. And, as Dhawk790 points out, the Opinion polling for the United Kingdom European Union membership referendum page includes the suspension of campaigning following the Jo Cox murder. — Zcbeaton (talk) 13:45, 24 May 2017 (UTC)
I agree that it should be included. There is at least a precedent in how polls are shown for EU referundem, which included the Jo Cox murder: Opinion polling for the United Kingdom European Union membership referendum. Dhawk790 (talk) 13:00, 24 May 2017 (UTC)

I tend to agree that this event should be included. If it's significant enough to appear on the front page, it should be good enough for this page as well. I'm also not convinced that there's a consensus to exclude such events; Though it has been discussed before, the prior discussions I've found have cannot be said to have established a clear consensus. Has there been an RFC on the issue? Rami R 13:51, 24 May 2017 (UTC)

We haven't had an RFC. We have discussed the matter several times and, while discussion has gone back and forth somewhat, we've generally reached a consensus on a limited list of things to include. Happy to see an RFC.
All non-UK opinion polling articles I've seen list very few extra rows: normally none whatsoever, occasionally other elections (e.g. local elections). Why are we doing something different? I can't recall anyone answering that question. Most UK opinion polling articles have ended up with a limited list of events (other elections, party leader changes), although there are some exceptions. I see no precedent for including the publication of manifestos. Jo Cox's murder does provide some precedent for including the Manchester bombing, although that event was more explicitly linked to the election at hand.
Just because something is a significant event doesn't mean it should be included in a list of polling. I'm all for a timeline article, but a timeline article should be a separate thing. Inserting events into a table of polling is editorialising. It implies the events have had an impact on polls. We cannot make those judgements: that would fall foul of WP:OR. Bondegezou (talk) 14:10, 24 May 2017 (UTC)
I agree that manifestos shouldn't be included. However, while this bombing was not to do with the election (as far as we know), it did result in the suspension of campaigning, so for that reason I think it should be included. Absolutelypuremilk (talk) 14:30, 24 May 2017 (UTC)

I think I understand the crux of the dispute is whether the events are indisputably linked to the election, or whether it would be editorialising to link them. In the absence of a page listing the events during this period, however, I think it is necessary to give some context in the election. Manifestos maybe less important, as one can assume policy announcements in the build up to the GE, likewise with public debates (though I think they have featured before?). Directly linked to polling seems to be the suspension of campaigning. The reason for suspension is less important, but the fact campaigns are halted seems necessary? VelvetCommuter (talk) 16:04, 24 May 2017 (UTC)

Some editors want the polling table to become a timeline of all sorts of events. I can see that would be interesting, but that becomes WP:SYNTHy and interpretative. That's something for a blog or an analysis article, not a Wikipedia table of data. Personally, I'd rather we did what nearly all the other opinion polling articles on Wikipedia (en. or otherwise) do: which is nothing but polls data.
If there is some useful interpretation of the ups and downs, in RSs, then I'd love to see that summarised in text separate to the poll table. Bondegezou (talk) 16:24, 24 May 2017 (UTC)
I don't want it to be a timeline of all sorts of events, but the pendulum has swung ridiculously far in the virtually no events at all direction. To include the Manchester bombing and not the manifestos is astonishing. Correct me if I'm wrong, but this General Election is supposed to be about policies, right? Please don't get all "I know a Wikipedia rule that agrees with me" ...with me. As it is now, it's just plain silly. Common sense are good bywords, don't you think? Boscaswell talk 07:25, 26 May 2017 (UTC)

Can we include don't knows in the polling table?

Please. — Preceding unsigned comment added by Meznaric (talkcontribs) 23:51, 25 May 2017 (UTC)

British polls are not reported with Don't Knows and we take our lead from how pollsters and sources report them. The only polls that include don't knows are questions like Best PM/best leader on a certain issue. FriendlyDataNerdV2 (talk) 22:34, 28 May 2017 (UTC)

Reintroduce line for release of manifestos?

A crucial time point, or collection of time points, rather. Blurred I know by the Labour 'leak', but crucial nevertheless. And yes, I know that it's discussed above. But there hasn't been a consensus. Boscaswell talk 07:13, 26 May 2017 (UTC)

I'm in favour of this. It was done for the 2010 election polling. It seems fairly clear that (unusually for recent UK elections) the manifesto releases have actually been crucial moments in this campaign that have affected the polling significantly. Jw2036 (talk) 14:12, 27 May 2017 (UTC)
Also in favour, by the same logic. JMaxchill (talk) 16:53, 27 May 2017 (UTC)
No, this has been discussed several times over many years with unanimous consensus against. There are already excessive notes in table. Consensus can change, but it hasn't...users arriving to 'vote' on this issue is not how consensus works. 79.74.27.100 (talk) 22:33, 27 May 2017 (UTC)
But it is how democracy works. I'm sorry, but is glaringly obvious to anyone following the election, to all pundits, that the release of the manifestos has been pivotal to voting intentions. And yet there is still resistance to release of manifestos being noted? Excuse me? Boscaswell talk 06:13, 28 May 2017 (UTC)
The consensus on this article amongst registered editors seems to be to include them. Frankly it's utterly illogical not to. Jw2036 (talk) 09:51, 28 May 2017 (UTC)
If you want to discuss how polling has shifted through the campaign and its events, I think that would be a great thing to do. Prepare some text, citing reliable sources, and stick it in the main article. This is a table of polling numbers. It's not meant to be interpretative.
Look at any of these articles:
All these elections involved a variety of significant events that impacted on voting intentions. Nearly all of them have zero annotations. The Greek and Polish ones list other national elections and the Greek one has a nice shading to differentiate exit polls, but neither has, say, changes of party leader. The Australian one has lines for changes of party leader and the beginning of the election campaign, so about the same as what we're currently doing here. None of those include manifesto launches.
Why should we do something so different from all the other opinion polling articles or sections? Yes, let's have text to discuss the polling ups and downs. No, let's not mix a table of numbers with presumptions about what were significant events. Bondegezou (talk) 14:23, 28 May 2017 (UTC)
In the interests of fairness, I note that the .fr page for the Presidential election polling does use more event lines: debates and the start of the campaign, but also some other events (kinda akin to party leader change?). But still no manifesto launches covered. Bondegezou (talk) 14:27, 28 May 2017 (UTC)

@Bondegezou "...presumptions about...significant events" Do you disagree with the contention that the manifesto launches were significant? Particularly in view of the fact that Theresa May had to make a major announcement about the so-called dementia tax, and the unprecedented in my memory change in poll numbers over the course of the election, and the astonishing turnaround in voting intentions of older voters, which is being 'blamed' in the media on the manifestos. How about some common sense here, eh?  :-) Boscaswell talk 18:49, 28 May 2017 (UTC)

@Boscaswell: I think the main election article should cover the matter in the text. It seems common sense to me to copy standard practice on other Wikipedia articles with the table of data here. Bondegezou (talk) 10:13, 29 May 2017 (UTC)
@Bondegezou: would you care to answer my question? Which was "Do you disagree with the contention that the manifesto launches were significant?" I'd contend that they were highly significant, btw. As have most pundits. Boscaswell talk 13:35, 29 May 2017 (UTC)
@Boscaswell: I have sought to answer your question. My aim is to improve the article through a workable consensus, not to have a battle of rhetoric.
I don't think the LibDem, UKIP or SNP manifestos had any particular impact on the polls, and nor do any pundits as far as I've seen. I think the Labour manifesto probably did, but it's hard to put a date to that because it leaked. The row over the 'dementia tax' was significant: there's more to say on that than merely the date of the Tory manifesto launch. Those are all things I think should be discussed in the campaign section of the election article. Prose gives one the flexibility to discuss such matters rather than reducing them to a single date. TL;DR: They were significant, so they should be covered, but they should be covered in prose in the election article, not in a table of poll results.
When James Comey announced he was examining a new set of Clinton emails, it had a huge impact on the US Presidential race. Pundits galore said that. Many have said it was the one event that won it for Trump. What do you think? Was it significant? Because I note it's not mentioned, as nor is any other event, in the polling article for the US Presidential election. Bondegezou (talk) 22:13, 29 May 2017 (UTC)

Debates/Interviews

Hi all, just wondering what the rationale is behind mentioning the non-debate with Corbyn and May on Channel 4 and Sky News but not the debates of various other party leaders? VelvetCommuter (talk) 23:42, 29 May 2017 (UTC)

  • I'm not arguing for or against the inclusion of omission of either, just so we're clear before I begin. I'm surmising that perhaps a 'debate' between the leaders of the two biggest political parties - with over 75% of public support between them - could be considered more relevant to these polls than another 'debate' between representatives of 3 minor UK-wide parties and one major regional party from each of Scotland and Wales - whose combined population share is around 15%.

The change in poll results, if any, from the ITV debate would be so minimal as to be imperceivable. If we had seen support drain from the Tories to UKIP or the Liberals, perhaps it should have been included as a reference point. iMarc89 (talk) 05:35, 30 May 2017 (UTC)

Across multiple countries, some polling articles list debates, most don't. I (as you may have guessed from earlier comments) would leave them all out. If we are to include them, we do need clarity on which go in and which don't. In answer to iMarc89, I agree there's no evidence that the earlier debates in this campaign shifted polling, but so far there's no evidence last night's (not that it was actually a debate) did either.
As per WP:BRD, I have removed the line for now. Let's seek some agreement on what to include and what not on the question of debates a.s.a.p. Bondegezou (talk) 08:46, 30 May 2017 (UTC)
United_Kingdom_general_election,_2017#Television_debates lists 11 'debates', 6 UK-wide of various formats, some for Scotland (2)/Wales (2)/NI (1), plus a set for the English regions. Which do people think should be included, and where (main polling table, Scottish/Welsh/NI polling tables, other regional polling tables)? (I vote none anywhere, as per US polling articles.) Bondegezou (talk) 08:56, 30 May 2017 (UTC)
I would say none also. Had there been a proper debate with the main party leaders then maybe that should have been included but there has not been one, so nothing really worth inclusion. - Galloglass 09:27, 30 May 2017 (UTC)
If we were to include last nights double interview, then maybe we would have to include every interview with a major party leader.-- BOD -- 11:30, 30 May 2017 (UTC)

I would also say only include them if they are actually a proper debate between the main party leaders. Absolutelypuremilk (talk) 11:42, 30 May 2017 (UTC)

Anything that is likely to affect the polls should be included, thus enabling a judgement to be made as to whether or not the event did have an effect. The Paxman interviews are pivotal in this respect, so of course they should be included. Other events involving just the minor parties shouldn't necessarily be included. If particular events are not mentioned the reader may end up wondering what, if anything, caused a marked swing in the polls. MidnightBlue (Talk) 19:43, 30 May 2017 (UTC)

Yougov projections

Is the latest controversial yougov projection relevent? http://www.independent.co.uk/news/uk/politics/election-2017-polls-latest-yougov-hung-parliament-prediction-why-conservatives-labour-win-seats-a7764401.html Russell's teapot (talk) 10:06, 31 May 2017 (UTC)

I think that this is very relevant. It's a daily opinion poll of 7,000 voters/potential voters and YouGov claim it was very accurate for the referendum. Horseshoe123 11:07, 31 May 2017 (UTC) — Preceding unsigned comment added by Horseshoe123 (talkcontribs)

They may "claim" so but they were wrong on the Scot Ref prediction, wrong on the EU Ref prediction and wrong on the 2015 GE projection together with all the other pollsters. So let's take it with a pinch of salt, or a bucket as some commentators have said. Let's wait until there were others at least. 90.215.121.192 (talk) 13:18, 31 May 2017 (UTC)
Also, if it is included the whole margin of error range must be included, which is something like 275-310-350 for the Tories and same range for the rest. As it stands, it is nothing but a modelling, in the same way climate change models are. Now whilst we may agree that the underlying theory in both CC and polling is correct, the range given by these models has shown some to be completely badly calibrated and wrong. 90.215.121.192 (talk) 13:22, 31 May 2017 (UTC)
I think all polls should be included, they are quickly averaged out by other polls. You are just as likely to get oddball polls favouring each party in turn and as has been mentioned above, people tend to take polls with a pinch of salt nowadays. It is not our job to omit polls if they have been prepared by a professional firm and they have asked a reasonable number of people. - Ânes-pur-sàng (talk) 13:34, 31 May 2017 (UTC)
Except it's not a "poll" though run by the BPC methodology. It's a "model" based on YouGov's internal algorithm based on 7000 people across ALL 650 constituencies. That's why the margin of error is so huge. A weighted poll of 1000 across the country is fine for general preference of party but a seat projection poll needs at least 500 people per constituency to be accurate in the same way, i.e around 3 plus or minus. What's 500 x 650? It clearly is way above 7000 people. 90.215.121.192 (talk) 13:39, 31 May 2017 (UTC)
It's based on MRP, which helps to provide individual constituency projections based on voting patterns among similar demographics/regions. It is, however, based on actual interviews as a standard poll; if those numbers are released, they should be probably included. Mélencron 13:43, 31 May 2017 (UTC)
"We have produced a new election model based on Multilevel Regression and Post-stratification analysis". Clearly and unambiguously NOT A POLL in any sense of the word. It's just their personal punt on the election. You might as well ask your psychic. 90.215.121.192 (talk) 15:03, 31 May 2017 (UTC)
The point of MRP is that it's based on a large number of actual interviews and uses demographic information to project breakdowns for subgroups. You have to conduct a voting intention poll to construct such a model. YouGov has published those headline VI numbers. Mélencron 15:16, 31 May 2017 (UTC)
Absolutely, by all means include the actual poll that they did. I believe it was 42/38 CON/LAB. Just not the projection punt because of it. They are plenty of internet swingometers for people to play around with it on their own time, just not on Wiki. 90.215.121.192 (talk) 15:20, 31 May 2017 (UTC)
...which is why the projection itself isn't included on this article. The main election article already includes projections, and the polling articles never have in the first place. Mélencron 15:21, 31 May 2017 (UTC)
The sample size was NOT 50000. Stop putting in misinformation. And putting in the updated model based on the old sample is clearly wrong as well. 90.215.121.192 (talk) 15:27, 31 May 2017 (UTC)
It's clearly mentioned on the page. Mélencron 15:29, 31 May 2017 (UTC)
They will be using the OLD SAMPLE daily until the election to model every day. THEY ARE NOT REINTERVIEWING 50,000 PEOPLE EVERY DAY. Stop being facetious. 90.215.121.192 (talk) 15:30, 31 May 2017 (UTC)
Which is why the fieldwork is listed over, yep, seven days. In no way does it say or imply that they're interviewing 50,000 people every day; I don't know where you're getting that idea from and I don't know why you're taking issue with the way it's listed on the article (it doesn't imply they're reinterviewing the same set of individuals every day). Mélencron 15:35, 31 May 2017 (UTC)
That means the daily poll today is based on OLD SAMPLE and so will each one until election day on the MODEL. Which is completely at odds with the other polls included. But if you want to skew the table to your preference then be my guest. Seems to be all the rage here. I don't know why I bother. 90.215.121.192 (talk) 15:59, 31 May 2017 (UTC)

Semi-protected edit request on 31 May 2017

J.pB321 (talk) 20:14, 31 May 2017 (UTC)
Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. You may request confirmed rights on your account at Requests for permissions, but you would need a strong justification for such a request. Requests for removing the protection on this article should be directed to the administrator who protected it (preferred) or at requests for unprotection. —KuyaBriBriTalk 20:28, 31 May 2017 (UTC)
  1. ^ Hypothetical poll