Talk:Opinion polling for the 2019 United Kingdom general election/Archive 5

Graphical Update
The graph hasn't been updated for a week (12th according to the note). Could someone do it please? TheMysteriousEditor (talk) 19:04, 18 November 2019 (UTC)
 * I'm puzzled. It looks to me like has updated it on the 11th, 12th, 14th, 15th and 17th of this month. (Thank you, Bellowhead). --Dweller (talk) Become old fashioned! 17:21, 19 November 2019 (UTC)
 * Unpuzzled. See --Dweller (talk) Become old fashioned! 17:23, 19 November 2019 (UTC)
 * Thanks, I've now updated the graph again. Note that previously the text referred to the date on which the polling was completed, rather than when the poll was released. I've now changed this because 1. it made the graph seem more out of date than it was and 2. if a poll is released after I update the graph which completed after the other polls, then it will be missing from the claimed date. Bellowhead678 (talk) 08:24, 20 November 2019 (UTC)

Kantar link?
This week's Kantar has just a link to Britain elects Tweet. 94.174.113.31 (talk) 00:07, 22 November 2019 (UTC)

Brexit Party stand down in 317 seats
Can we not have this in the table? It feels like it has an editorial rather than electoral reason for inclusion, which isn't what the table is for! Thanks VelvetCommuter (talk) 15:52, 11 November 2019 (UTC)
 * I agree. It's been said many times before, put this information in the 2019 in the United Kingdom article and then people can compare if they wish too. At this point, national debates should be the only things in the national polling section. Perhaps something like this Russia report if it got published but maybe not. Jonjonjohny (talk) 16:00, 11 November 2019 (UTC)


 * Agree and I've removed. Nothing to do with polling at all. Andy mmu talk 18:09, 11 November 2019 (UTC)
 * I am uncertain. If there's no candidate, you can't vote for them. It necessarily affects their actual performance in the popular vote, which is what polls are trying to predict. Bondegezou (talk) 20:26, 11 November 2019 (UTC)
 * If RSes invoke the event in relation to a change in polling (e.g. for companies that prompt for local candidates), then I think there's a case to include it. I don't think the case is there yet. Ralbegen (talk) 20:57, 11 November 2019 (UTC)
 * Nobody has mentioned the Unite to Remain pact, and that's at least as significant. Bearing in mind the Lib Dems poll higher, actually had seats in parliament, and this involves other parties too. There's a similar pact in Northern Ireland. We're not trying to keep a running commentary of who is and isn't standing in any given seat. Andy mmu talk 20:59, 11 November 2019 (UTC)
 * I remain uncertain. I take your point, Andymmu, but Unite to Remain affects 60 seats, while the biggest arrangement in Northern Ireland affects 4 seats so far. This is 317 seats.
 * Instead of a row in the table - and I don't like have any rows in the table - perhaps we can look at revising the introductory text (in line with RS commentary) along the lines Ralbegen suggests (e.g., what do companies do in terms of prompts)? Bondegezou (talk) 09:26, 12 November 2019 (UTC)
 * RS discussion of the matter: https://www.thetimes.co.uk/edition/news/tories-gain-14-point-lead-after-farage-withdraws-candidates-ts7vck3nw Bondegezou (talk) 07:10, 13 November 2019 (UTC)
 * In addition to revising the introductory text, do you think it might be worth including a column distinguishing between local party/national party prompts? It's a pretty significant methodological difference—at least a bigger impact on VI figures than GB vs UK. Ralbegen (talk) 10:50, 13 November 2019 (UTC)
 * The sudden change in Brexit Party polling is almost all to do with methodology, so I would support this (although we usually do it with notes and that might be better). We still shouldn't mark them not standing in seats directly in the table though. Andy mmu talk 11:32, 13 November 2019 (UTC)
 * If all post-Nov 11 polls follow the approach of YouGov and don't promt for Brexit for responders in the 317 seats, this is a significant change (in both methodology the "real word"). It is useful to the reader to include this in the table. In terms of contributing to the aggregate for national polls, Unite to Remain isn't nearly as significant, as it covers only one in six seats and the vast majority of standing-down candidates were on track to get low-single-figure totals. --LukeSurlt c 13:13, 13 November 2019 (UTC)
 * Perhaps we could have a 'close of nominations' row in the table, as this would reflect the formal deadline of which candidates there are in each constituency and pollsters who only show parties running in the respondent's constituency will update their prompts accordingly. The inability to vote for a party in a constituency will affect the headline vote shares, as shown in the YouGov poll (11-12 Nov) where Brexit fell from 9% to 4% just by having the constituency candidates prompt - not because of a fall in support (as shown in a supplementary question using their previous methodology). Clyde1998 (talk) 21:13, 17 November 2019 (UTC)
 * I have added such a row. I think it's absurd that a party suddenly announcing it will not run in half of GB (not what was expected) and rescinding its candidacies, causing some pollsters to change their methodology for that party and their polling numbers to halve, during the campaigning period, is not something there can be consensus on including. But if we can't have that, this is the next best thing surely. &mdash;ajf (talk) 19:40, 21 November 2019 (UTC)
 * while it is relevant to the election as a whole this isnt the place for it, this page is for polling data, party politics should really be on the main election page, only timeline events should really be events that affect most parties (e.g. start of the campaign period or by-elections), in my opinion we shouldn't even have the timeline events about new party leaders. &mdash;Popeter45 (talk) 21:52, 21 November 2019 (UTC)
 * Major events that affect polling should be included. The Brexit Party standing down in about half of seats is important, as the drop their support wasn't due to a fall in support rather methodology changes in polls to reflect their decision. Wording it as 'close of nominations' affects all parties, as that's the deadline to have a candidate in a constituency. WP:TECHNICAL notes "every reasonable attempt should be made to ensure that material is presented in the most widely understandable manner possible". Having timeline events within the tables helps less knowledgeable readers understand the data and helps people view the reasons for changes at a glance. Clyde1998 (talk) 03:17, 24 November 2019 (UTC)
 * Seriously &mdash;ajf (talk), this has been debated to death and consensus is NOT to include this in this page, you have now violated the The three-revert rule and if this continues i will open a Dispute resolution against you as a Disruptive user &mdash;Popeter45 (talk) 00:09, 23 November 2019 (UTC)

Brexit Party only standing in 300-odd seats more important than ITV debate
Surely! However did the ITV debate get its own line in the table? IMO the BXP announcement last week is much more line-worthy. Boscaswell  talk  01:23, 21 November 2019 (UTC)
 * Agreed. I've removed the debate line. This is still not a timeline article and we should not have endless commentary in a table of polls. US election polling articles, for comparison, don't include TV debates. Bondegezou (talk) 10:30, 21 November 2019 (UTC)
 * Both these things had had lines in the table, and I have restored both. There is a clear justification for both. The Brexit Party announcement took them out of the running in half the country and had an obvious polling impact, why shouldn't we show that? And TV debates are a typical campaign milestone. If you can present good reasons not to have those, I'm all ears, but I don't think "lack of consensus" from nobody having discussed it yet is a good excuse to remove either. &mdash;ajf (talk) 12:24, 21 November 2019 (UTC)
 * By the way, Opinion polling for the 2015 United Kingdom general election included all the debates, the US is not a good excuse. &mdash;ajf (talk) 12:26, 21 November 2019 (UTC)
 * Strong oppose by adding the statement about Brexit party you are putting that above actions by all other parties leading to bias, why not add timeline elements about how many seats the conservatives, Labour or liberal Democrats are contesting, we should just not include any such events to avoid the entire timeline becoming just who's sitting where instead of its actual role of showing Opinion Polls - Popeter45 12:36, 21 November 2019 (UTC)
 * It's not bias, if Labour announced they weren't standing in half the seats in GB then it would be equally noteworthy. But the other parties haven't done anything anywhere near as significant which would affect not only their polling numbers but polling methodology like this. &mdash;ajf (talk) 12:38, 21 November 2019 (UTC)
 * the Remain Alliance, why not add that them? - Popeter45 12:42, 21 November 2019 (UTC)
 * If you think it's significant enough and it happened during the period polled, I have no objection. &mdash;ajf (talk) 12:43, 21 November 2019 (UTC)
 * All text in a Wikipedia article requires consensus support. Let us respect WP:BRD and keep any contested lines out until we have a consensus for their inclusion. I see no consensus for either at this time. Bondegezou (talk) 12:46, 21 November 2019 (UTC)
 * This seems to be an argument in order to bring back the line about the Brexit Party than not including the TV debates. We've already ruled out the Brexit Party announcement in the sections above. I'd rather not have both than include both as this is not a timeline article. We have 2019 United Kingdom general election for that if you want a timeline. Also, Ajfweb, you've not understood that the status quo is for them not to be in the table and you should argue for them to be there, not have the rest of us have to persuade you otherwise. Andy mmu talk 12:50, 21 November 2019 (UTC)


 * Strong oppose any time line additions apart from changes of leadership and elections and by-elections. The is a link  just above the graph for all events that might influence polling. ~ BOD ~ TALK 23:57, 21 November 2019 (UTC)
 * If we have reliable sources discussing how polling has changed (in reaction to specific events or not), we can have some text discussing that as well. Bondegezou (talk) 12:08, 22 November 2019 (UTC)
 * Strong support: Just to reiterate what I've said elsewhere as this issue is all over this talk section: context to the numbers is important under the WP:TECHNICAL and WP:AUDIENCE guidelines. Having the 'close of nominations' line is a neutral way of noting that from that point, (most) polls will only include parties/candidates who are running in a respondent's constituency - ie. not asking about the Brexit Party where they're not standing. Clyde1998 (talk) 03:50, 24 November 2019 (UTC)

Pending review
Given that this is a fast-moving situation where users would benefit from rapid updates, there are appalling delays in  pending review. The most recent Survation poll was added 5 hours ago, pending review, but the reviewers seem asleep on the job.Cutler (talk) 22:03, 24 November 2019 (UTC)
 * Only unregistered or users that are not autoconfirmed (WP:AUTOCONFIRM) will require their edits to be reviewed. This protection was added because of a couple of fake polls added by unregistered users. The fact that this is a rapidly evolving event is exactly why the protection is required. Andy mmu talk 22:18, 24 November 2019 (UTC)
 * In fact I've just noticed that the Survation poll you reference has not got a valid reference, either deleted or never existed, so this is a good example! Andy mmu talk 22:25, 24 November 2019 (UTC)

ITV Debate not listed in polls table
This has been indirectly negatively discussed above but I would like to make a positive case for listing the ITV Johnson v Corbyn debate in the polls list. In the article Opinion polling for the 2015 United Kingdom general election this was done for the major debates, so precedent is surely on its side. Debates can be significant campaign milestones with polling often commissioned for immediately before and after to see if it has changed opinions. Surely it is reasonable to provide this as context? Of course not all events need to be in the table, but debates are some of the most significant ones. &mdash;ajf (talk) 19:29, 21 November 2019 (UTC)
 * Most polling articles on Wikipedia don't do this, so most precedent is against.
 * If the debate becomes a significant campaign milestone, then a reliable source will say that. Do you have one? If not, looks like WP:OR.
 * This is not a timeline article. Bondegezou (talk) 20:30, 21 November 2019 (UTC)
 * "Most polling articles", are they about the UK?
 * Opinion polling for the 2015 United Kingdom general election
 * Opinion polling for the 2010 United Kingdom general election
 * Opinion polling for the 2016 Scottish Parliament election
 * Opinion polling for the 2014 Scottish independence referendum
 * &mdash;ajf (talk) 19:24, 22 November 2019 (UTC)
 * Most polling articles, not specific to the UK. I cannot see why a UK polling article should be treated differently from other polling articles in this regard. Bondegezou (talk) 10:59, 23 November 2019 (UTC)
 * Let's be clear, the reason they appeared in the 2010 article is because they were brand new and a novelty. Indeed I seem to remember they only got added because of "Cleggmania", and not because they were thought of as significant before the event. Although I wouldn't be against their inclusion in the table now, as at least they are clearly defined events, there is no indication that these debates will trigger large shifts in polling. Also the differences in formats, various long interview programmes that aren't seen as debates for example, and the far more ad hoc nature (and not the appointment to view draw of 2010) of the debates this year lead me to lean away from their inclusion. As with various who's standing or not announcements, or major policy shifts, or other events that might "affect" the polls, we should leave readers to draw their own conclusions, rather than have a handful of Wikipedia editors decide why such and such a party lost/gained support. Andy mmu talk 15:24, 23 November 2019 (UTC)
 * 74% of people stated that the debates are one of the three biggest factors in how they decided to vote in 2017. Even if they don't cause changes in the polls, they're important in the sense that they allow for scrutiny of the leaders in a way that they wouldn't usually be; a lot of people only really pay attention to politics during election periods and the debates are one of the few ways that they see policies challenged. It's not about "Wikipedia editors deciding why such and such a party lost/gained support" rather giving context to the numbers and to help people navigate them. The timeline events give context to the political situation at the time of the poll, I think not having them would just leave a list of numbers that a passing reader wouldn't necessarily understand - going against the WP:TECHNICAL and WP:AUDIENCE guidelines. Obviously every event shouldn't be included, but I think debates are relevant to the context of the campaign. Clyde1998 (talk) 03:46, 24 November 2019 (UTC)
 * The proportion of people saying the debates were a big factor in 2017 is relevant for the 2017 article. It would be WP:OR to presume that carries forwards. "Giving context to the numbers" is a form of WP:EDITORIALIZING. Wikipedia articles do not have the freedom to act in that way. Bondegezou (talk) 09:19, 25 November 2019 (UTC)
 * The section you link to does not really support your point. &mdash;ajf (talk) 22:15, 25 November 2019 (UTC)

Speedy Graph Update
There needs to be more than one editor with the ability to update the graph. It's 20 days out of date right now which is ridiculous considering we're now running up to the actual election. It shouldn't be down to just one editor - though if said editor is available might a request they update it whenever they can? TheMysteriousEditor (talk) 19:56, 1 November 2019 (UTC)
 * Exactly right. This page always comes at or near the top of Google results. It is essential that everything is kept accurate and up to date. It's also disappointing that the confusing/ambiguous "NAT" legend on the graph has not been clarified despite request. If it means Scottish National Party (only) then please change it to "SNP". 2A00:23C5:4B91:AB00:5F3:C50E:265E:36A (talk) 01:55, 2 November 2019 (UTC)
 * You should bear in mind wiki is done by volunteers, if you wish for a graph update, feel free to read https://commons.wikimedia.org/wiki/User_talk:AdamFilinovich#UK_poll_diagram & https://docs.google.com/document/d/1khuyz_552Gt26uIMGhAIV8Q8prwlNubzfnmIQ05mLOI/edit and do it; it isn't down to one editor. The NAT in the graph is for both major Nationalist parties, so certainly shouldn't be changed to SNP. Perhaps a note could be added, though I thought it was obvious to any reader who'd know what SNP & PC stood for. 86.132.18.158 (talk) 15:07, 2 November 2019 (UTC)
 * Ok I just attempted to follow them instructions and it is way too complicated. TheMysteriousEditor (talk) 15:43, 2 November 2019 (UTC)
 * Check out what I did at Nationwide opinion polling for the 2020 Democratic Party presidential primaries. That graph is updated without editing any pictures; you just need to know how to use the "code." --Comment by Selfie City ( talk about my  contributions ) 22:24, 3 November 2019 (UTC)
 * If someone wants to practice this method, they could try it on the Scotland or Wales polling which needs a graph. There aren't too many data points so shouldn't take too long. Bellowhead678 (talk) 22:26, 25 November 2019 (UTC)

Short axis graph


I've created a graph just for the campaign period, and I think we can add it in once it fills out a bit, probably sometime next week. Bellowhead678 (talk) 13:07, 14 November 2019 (UTC)


 * This would be useful in the main article, if kept updated. If not, a line on the main graph at 6 November. Stub Mandrel (talk) 16:33, 20 November 2019 (UTC)
 * Everyone happy for me to add this to the article now? Bellowhead678 (talk) 06:40, 25 November 2019 (UTC)


 * Looks like you have, it certainly helps. Stub Mandrel (talk) 13:38, 26 November 2019 (UTC)

Graph Update
don't suppose you could update the graph please? Also, just a few questions: it says the graph is correct up to 2 November, so why does the data on the graph itself not go up the November line coming up from the x-axis? Unless I'm being dumb. Also, the Brexit Party trend line seems to have a disconnected bit from March 2019. TheMysteriousEditor (talk) 13:24, 10 November 2019 (UTC)
 * I'm away for the weekend, but I'll have a new graph up tomorrow. I'll take out the March 2019 bit and adjust the x-axis. Bellowhead678 (talk) 13:35, 10 November 2019 (UTC)
 * No worries - thanks. Just realised how demanding that previous message sounded lol. TheMysteriousEditor (talk) 17:01, 10 November 2019 (UTC)

- I think the x-axis on the graph is all out by 1 month, as it jumps from Feb 19 to July 19. Should the x-axis start with July 17 instead, or is there a month of data missing?
 * Good spot, now fixed. Bellowhead678 (talk) 06:41, 28 November 2019 (UTC)
 * Cheers!

Seat predictions being deleted?
Why would we delete old seat predictions, per this? Seems contrary to the whole point of this page that shows changes in reliable opinion polls over time. --Dweller (talk) Become old fashioned! 18:37, 24 November 2019 (UTC)
 * Why not add a graph that shows the trends in seat predictions?--12.144.5.2 (talk) 20:57, 24 November 2019 (UTC)
 * There was some discussion about this above at Talk:Opinion_polling_for_the_2019_United_Kingdom_general_election.
 * The answer to 's question is because these aren't opinion polls. We're meant to follow reliable sources and reliable sources don't give as much emphasis to most of these seat predictions, and indeed they get criticised . I've never seen a reliable source doing a graph for seat predictions over time, so we shouldn't be innovating, as the IP editor suggests. If I remember correctly, last election we did not track every seat prediction that came out. Bondegezou (talk) 22:40, 24 November 2019 (UTC)
 * Only including the latest seat prediction from each RS-weighted predictor makes sense. Including different seat predictions at n weeks before the election, as with the main 2017 election article, could give an idea of how things change over time. I think the column-comparison format of this is a nice way of presenting the data, and that should be the format to follow. Tracking every change would give them undue weight, and doesn't really tell the reader anything extra. Following the format of the main 2017 article but applying sensible inclusion criteria, and transcluding or recreating it on 2019 United Kingdom general election, seems the most sensible way of doing it to me. Ralbegen (talk) 22:48, 24 November 2019 (UTC)

Wait, hang on a second. Either they're relevant to this article, or they're not. If they're not, they shouldn't be here at all (I'd oppose this). If they are here, past projections make as much sense as including past opinion polls that fill the entire page. --Dweller (talk) Become old fashioned! 23:40, 24 November 2019 (UTC)
 * Relevance isn't binary: something can have some relevance without being the core topic of the page. I concur with Ralbegen's suggestion and that's what we've done before.
 * This is an article on "Opinion polling for the 2019 United Kingdom general election", so I think it's pretty clear that past opinion polls should be on this article. If you want an article on "Seat predictions for the 2019 United Kingdom general election", you can create one. But long tables of seat predictions that don't attract RS coverage seem inappropriate to me here. Bondegezou (talk) 09:15, 25 November 2019 (UTC)


 * Can we discuss whether we keep the ForecastUK forecasts? It is updated most days, the website is over 10 years old, it was the best all parties forecast of the non MRP forecasts in 2017 and is used in the Elections etc model as a "complex model" Peterould (talk) 13:23, 28 November 2019 (UTC)
 * We need to base inclusion on reliable sources, in my view. Have ForecastUK forecasts been covered in newspapers or academic papers? Ralbegen (talk) 13:27, 28 November 2019 (UTC)
 * Not recently. There was some very good media interest back in 2015 when forecasting using complex models suddenly became a big thing.Peterould (talk) 13:41, 28 November 2019 (UTC)

Seat prediction inclusion criteria
I'm struggling to work out the basis seat predictions are being included on this page. I would have thought that a sensible criterion would be to limit them to sources that are either in a reliable source themselves (like the New Statesman predictions last time), or which are considered credible and noteworthy by reliable sources (like Electoral Calculus, perhaps). Just including the latest prediction from each source seems more straightforward to me, like in the 2017 polling article, rather than keeping a timeline of them. I can't see the case to include Election Maps UK, Complete Politics, 326 Politics, Electomania or Election Polling on the page right now. Ralbegen (talk) 23:22, 6 November 2019 (UTC)
 * Agreed. Bondegezou (talk) 11:50, 7 November 2019 (UTC)
 * Agreed. I think we've had some COI editing by Election Maps, fwiw. --Dweller (talk) Become old fashioned! 12:12, 7 November 2019 (UTC)
 * discussion here. Currently 3:0 for cutting this down. Bondegezou (talk) 12:58, 7 November 2019 (UTC)
 * Thanks for the hint. Since the discussion is more about the quality of sources than about general inclusion yes/no, I will leave it up to the UK natives to find a smart solution. Since the popular vote does not count directly, only the seat prediction carries a relevant information. --Nillurcheier (talk) 13:05, 7 November 2019 (UTC)

I made this edit previously. Are we happy to go with that? I went with only the latest prediction, and only predictions from RS, as defined by Ralbegen. Bondegezou (talk) 13:08, 7 November 2019 (UTC)

Generally I support including past predictions to create history of them, but my problem is that EC does not store previous calculation results (and even if you use their predictor with same input numbers, you get slightly different results). Therefore, if you click on one of older links, it will lead you to their main page which just contains the most recent projections. Wikipedia's main directive is that every piece of information must be traceable to its very origin, which isn't achieved in this case. So we remove old polls or accept that they cannot be found at source. Opinions? — Preceding unsigned comment added by CptBalu87 (talk • contribs) 08:54, 29 November 2019 (UTC)

Seven DataPraxis MRP estimates posing as polls
If you visit the report linked by the source for all the recent Datapraxis polls, they pretty clearly state and indicate that they are not constituency reports but simply modelling. To avoid confusion and given the unequivocal purpose for this section in addition to its urgency, i have removed them directly without consultation. Let me know if you have any issues with this, but just wanted to give everybody a head's-up on here first. A Red Cherry (talk) 19:25, 30 November 2019 (UTC)
 * I have concerns - MRP is modelling. If it's claiming to be constituency polls, it shouldn't, but if it's clear that it's an MRP from a large enough dataset, there's nothing wrong (indeed it's the only real purpose) with then applying that to constituency demographics. 176.254.60.161 (talk) 21:31, 1 December 2019 (UTC)

Can we change name back?
The article has had its name changed by from "Opinion polling for the 2019 United Kingdom general election" to "Polls and predictions for the 2019 United Kingdom general election". I can't see any discussion about this. Can we, under WP:BRD, change it back? Or get some discussion on this? The article does contain some predictions, but no more than every other recent UK general election polling article, but is clearly focused on polling. I don't think that focus should change, as the title change would imply. Bondegezou (talk) 09:06, 2 December 2019 (UTC)


 * I tried undoing but it didn’t work. ‘’’The change should definitely not happen!’’’   Boscaswell   talk  09:14, 2 December 2019 (UTC)


 * I Support reverting this name change as there has been no prior discussion and there is no precedent for this. Also, the seat predictions are product of the polls so they are no more significant. Jonjonjohny (talk) 10:16, 2 December 2019 (UTC)


 * I also support reverting, as per the other arguments. Bellowhead678 (talk) 11:25, 2 December 2019 (UTC)

Aggregate by leave/remain parties
Is there a role in this particular election for a 2nd graph that tracks leave vs remain-or-2nd-referendum supporting parties? Crispin Cooper (talk) 10:30, 31 October 2019 (UTC)
 * How would you classify Labour? Because they don't seem to have decided if they'd campaign for Leave or Remain in a second referendum. --Dweller (talk) Become old fashioned! 10:35, 31 October 2019 (UTC)
 * It could work by having it as leave vs 2nd referendum as Labour supports a 2nd referendum on the deal that they would get. Lancashire2789 (talk) 11:50, 31 October 2019 (UTC)
 * Yes, but the LibDems do not. So perhaps, Leave vs. 2nd ref/ Remain. Jontel (talk) 12:09, 31 October 2019 (UTC)
 * But Labour aren't a 2nd referendum/Remain party. They might be, but they aren't. --Dweller (talk) Become old fashioned! 14:29, 31 October 2019 (UTC)
 * In saying 2nd ref/ Remain, I was combining the polling of parties supporting second referendum or remain positions. Labour support a second referendum. . Jontel (talk) 15:04, 31 October 2019 (UTC)
 * This may be too ambiguous, as the positions of parties are more of a spectrum rather than binary. Additionally, the positions of parties have changed throughout the term of the parliament; Labour, for example, supported leaving with a customs union as recently as the European elections in May and now prefer a second referendum (presumably with a remain option). Clyde1998 (talk) 18:29, 2 November 2019 (UTC)
 * I oppose this suggestion. It's WP:OR and it has no electoral significance under FPTP. It matters not one bit what the combined remain or leave vote is: it has no effect on who actually gets elected. Bondegezou (talk) 08:30, 3 November 2019 (UTC)
 * Hmmm. Not sure how you can say Lab isn't a 2nd Referendum Party when they clearly are.Datawarrior (talk) 16:02, 2 December 2019 (UTC)

Graph Update?
I believe there's now 3 weeks worth of polls to be added to the graph. Any chance of someone doing this? Thanks. TheMysteriousEditor (talk) 12:02, 15 October 2019 (UTC)
 * Done Filinovich (talk) 15:30, 16 October 2019 (UTC)

I believe there's now 8 new polls to be added to the graph - thanks. TheMysteriousEditor (talk) 20:01, 23 October 2019 (UTC)
 * Related to this, I see that one can see when the graph was updated (and by whom) by clicking on it, but it would be helpful to have the update date added to the text when upating occurs. Thanks! Jontel (talk) 20:11, 23 October 2019 (UTC)
 * Additionally, there's now more polls from the last week which haven't been added to the table (and therefore not the graph). The usual editors who are usually on top of these things seem to be away and I don't know enough about where to find them lol TheMysteriousEditor (talk) 20:15, 26 October 2019 (UTC)
 * Why the lag in graphical presentation? It is a full 10 polls out of synch now Daveyb100 (talk) 07:46, 29 October 2019 (UTC)
 * The main graph hasn't been updated for over 2 weeks now - who is responsible for updating it? TheMysteriousEditor (talk) 10:03, 31 October 2019 (UTC)
 * There is updating guidance here [], produced by the person who has been updating it. Jontel (talk) 10:56, 31 October 2019 (UTC)

Why doesn't the chart just update automatically when data is added to the table? — Preceding unsigned comment added by 150.143.87.81 (talk) 17:06, 3 November 2019 (UTC)

Don't know where to put this talk, but I just inserted a graph comparing the polling for the 2017 and 2019 elections. I don't know what people will think of this, but it is interesting. RERTwiki (talk) 14:12, 5 December 2019 (UTC)
 * It's certainly interesting, but as identified by another user who reverted your edit, it seems to violate NOR. Perokema (talk) 18:05, 5 December 2019 (UTC)

Column for relative change to poll by the same company
Every pollster has their own methods, so it would be useful to see the relative change to the same pollster's previous results. For example the latest Panelbase survey shows a 9% lead "only" - a negative change from the survey before it by IPSOS at 12%, but if you compare it to the last poll by Panelbase, it's actually a +1% change. I though about editing this page, but if someone creates this from an Excel or something I think it would be easier for them to do. Thanks. — Preceding unsigned comment added by Andrisi (talk • contribs) 20:52, 7 December 2019 (UTC)


 * Good idea.---Ehrenkater (talk) 22:25, 7 December 2019 (UTC)

"Sub-national polling"
Could this section be re-named? Are Scotland and Wales not nations?
 * Fair enough, I will change it to regional polling. Onetwothreeip (talk) 22:42, 7 December 2019 (UTC)

UKIP/ChangeUK 0%
I hate to nitpick, but on the main graph, the few 0%'s data for UKIP and Change UK need to be added. At the moment, their respective lines just abruptly end prematurely when there's still (limited) data for them to have inputted. I know it's 0%'s, but it's still important. Could whoever manages the graph fix this please - thanks. TheMysteriousEditor (talk) 18:12, 2 December 2019 (UTC)
 * I haven't included them because there is no data for them in most of the polls, but if others want me to include the 0% from the few polls that actually give a result for them then I'm happy to add them. Bellowhead678 (talk) 18:31, 2 December 2019 (UTC)
 * Yeah I think whenever there's a result for them it should be added, regardless of how infrequent or low they are. TheMysteriousEditor (talk) 18:42, 2 December 2019 (UTC)
 * Anyone else have any opinions either way? Bellowhead678 (talk) 08:05, 7 December 2019 (UTC)
 * I think UKIP could be removed from the polling table (and graph) completely starting in December.
 * Are minor parties (like Health Concern, etc.) running in this election? If so, UKIP's 0% is the same percentage many other, minor political parties in the UK must be getting at the moment. We might as well not include UKIP as we include those unless UKIP sees gains. --Comment by Selfie City ( talk about my  contributions ) 11:33, 7 December 2019 (UTC)


 * I think UKIP and Change UK can probably be cleanly removed after the Election next week, which is um very soon. ~ BOD ~ TALK 16:36, 7 December 2019 (UTC)


 * Not a good idea to show very small percentages on graph; they are very imprecise figures, they are not important, and they won't show up on the graphs anyway.---Ehrenkater (talk) 22:25, 7 December 2019 (UTC)
 * But the UKIP and Change UK lines are already existing on the graph. I absolutely agree that Change UK and UKIP should not be included in the tables for the next opinion polls article for the next election, but for now I don't see why certain figures should be omitted after a certain point. If a figure exists, it should be implemented onto the graph. TheMysteriousEditor (talk) 16:57, 8 December 2019 (UTC)

Constituency polling is selective and designed for political objectives against Wikipedia rules
National polling information has adequate supporting evidence from multiple sources of polls, along with sub-nation polling. This is entirely accepted.

What is unacceptable, is "Constituency polling". Only 9 out of 650+ seats have been selected. Only a single source of polling provided and this has been funded directly by a political party - the Liberal Democrats.

The names of the Constituencies are incorrect. For example Bath is now Bath and North East Somerset - see https://en.wikipedia.org/wiki/Bath_and_North_East_Somerset

On the ITV news, the Liberal Democrat leader has already been exposed for presenting false data in this constituency. The selective list presented is a selective number of Liberal Democrat target seats, using a single source Liberal Democrat funded data.

It is clear that the information in the section is a deliberate political attempt by the Liberal Democrats to influence the General Election. Unacceptable normally anyway, it is even more important that such political campaigning bias is not done during a General Election.

It is impractical, given the data and time and lack of independent evidence available for a comprehensive section to be provided at constituency level. It is recommended the section is urgently removed to ensure the Liberal Democrats or others do not try to manipulate independence of Wikipedia for political gain.Aligib213 (talk) 15:35, 6 November 2019 (UTC)
 * If the polling is carried out by a BPC member, I don't see why the funding source should matter. Bellowhead678 (talk) 22:23, 6 November 2019 (UTC)
 * Agree with Bellowhead678: if constituency polls meet our existing criteria, we should include them. The LibDems have been criticised for using Flavible constituency predictions: those don't meet our criteria, so we don't include them. But the polls listed on the article do meet our criteria, so they're in. Bondegezou (talk) 23:03, 6 November 2019 (UTC)
 * PS: the constituency name is correct, see North East Somerset (UK Parliament constituency). Bondegezou (talk) 23:06, 6 November 2019 (UTC)

I see where you're coming from, and I can see why the LibDems are doing this, but I don't think we can ignore them because they're reliable. Hopefully, the other main parties will catch on and do the same, which will balance it out. --Dweller (talk) Become old fashioned! 12:17, 7 November 2019 (UTC)

If there is genuinely reliable constituency poll evidence across all constituencies which reflects the boundary changes, so without bias towards a certain political party that is reasonable. What is unreasonable is allow a single political party, in this case the Liberal Democrats, to try an use Wikipedia to suggest questionable data based on their target seat list. Either provide all fairly, or none fairly. Not only those that suit a certain political party.Aligib213 (talk) 12:28, 7 November 2019 (UTC)
 * There is no evidence that the Liberal Democrats are trying to use Wikipedia to suggest questionable data. I would suggest,, that you review WP:AGF and moderate your comments.
 * We have criteria that ensure we only include reliable polls. We are applying those criteria. We are not including everything the LibDems are saying about seats, only polls that meet our criteria.
 * If there are other constituency polls paid for by other parties, but again meeting our criteria, then, yes, of course, we will include them. If you know of any that are missing, let us know. Bondegezou (talk) 13:03, 7 November 2019 (UTC)
 * I think you are mistaken Bondegezou, it is widely accepted by experts that the Lib-dems (and indeed other parties, when they do this) are only releasing the results that make them look good, and having them listed on wiki will be a (very) small part of their aim in this. However, IIRC, wiki always publishes them all regardless; I don't see a reason to not include them for this election, but it is true that this is manipulation by the lib-dems (all major parties do it). The fake polling in the LD campaign leaflets is a seperate issue, something I think should definitely be included somewhere...maybe here, maybe another page. 86.132.18.158 (talk) 16:37, 7 November 2019 (UTC)
 * Aligib213's specific accusation that the LibDems are trying to use Wikipedia casts doubts on the actions of Wikipedians who added this content to the page. I suggest it is unhelpful to skirt close to violating WP:AGF, as I suggest that does. Bondegezou (talk) 23:07, 7 November 2019 (UTC)

Our standard operating procedure when a poll is released or paid for by a political party is to attribute it to the source (in this case, the Liberal Democrats). , you make the important point that the LibDems are releasing these polls because it makes them look good (and alternatively, not releasing polls that would give the opposite). You have been able to discern that there might be some issues with the poll, (whether that is the small sample size or possible bias) which is exactly why we attribute these polls. If people saw a regular Survation or YouGov poll showing the LibDems ahead, they would look at it in a different light than a poll paid for by the LibDems that shows them ahead. Attribution allows our readers to make their own judgements on the validity of polling. In the end, none of us will know the actual result until election day. Bkissin (talk) 18:52, 7 November 2019 (UTC)
 * How about we keep these constituency polls but come to an agreement on prose at the beginning of the section that outlines the often selective nature of constituency polls being released? Onetwothreeip (talk) 11:25, 8 November 2019 (UTC)
 * If it's written in a neutral way, and backed up by multiple RS, then I'm happy with that. Bellowhead678 (talk) 11:53, 8 November 2019 (UTC)
 * I'm all for prose introducing and contextualising polling data. However, as per Bellowhead678, it has to follow usual Wikipedia rules: we have to draw on RS. Bondegezou (talk) 12:48, 8 November 2019 (UTC)

There is clear evidence that the Liberal Democrats, in particular, use polling data selectively to try to influence constituency polls. It is a deliberate campaign tactic which they use at every election in target and marginal seats right across the UK. Here is a classic example which was debated on ITV with their leader:

https://twitter.com/i/status/1190952807756632069

For even, questioning that the list provided was not fair and balanced, Bondegezou, made the unreasonable threat that I should "moderate your comments". I understand he has experience of writing sections specifically on the Liberal Democrats previously, where as I have no experience of writing sections, by or for, any political party.

If the section is not removed, then I would agree that an agreement on a prose or warning should be provided in a neutral way that the data accuracy and selection is limited and may not accurately representative and could be open to party political motives.Aligib213 (talk) 12:53, 8 November 2019 (UTC)
 * I take the above as clear breach of WP:AGF. Bondegezou (talk) 22:06, 8 November 2019 (UTC)
 * That applies to the actions of other editors, it doesn't require that editors take the propaganda output of political parties as highly likely to be truthful.88.145.150.206 (talk) 19:02, 9 November 2019 (UTC)
 * I agree with Bondegezou, insinuating that because someone has edited a political party page means they cannot be neutral is a breach of WP:AGF. Bellowhead678 (talk) 09:34, 10 November 2019 (UTC)
 * Where are you drawing the line? There are accusations of conflict of interest edits above, so 'assume good faith' isn't 'be naive as a matter of principle'. Suppose you were to follow the links on their profile and determine that the editor in question has been a candidate for elected office for that party. Are you still content? 88.145.150.206 (talk) 18:06, 10 November 2019 (UTC)
 * I suggest you actually read WP:AGF which says to assume good faith unless there is clear evidence to the contrary. Bellowhead678 (talk) 18:09, 10 November 2019 (UTC)
 * If an election candidate for a political party adding polling commissioned by that party, which is regarded as having been used misleadingly on a number of occasions by that party to influence opinion, rather than quantify it, and said editor then rushing to the defence of it without declaring their interest, isn't evidence of a conflict of interest, even if the edits are well-intentioned, I'm not sure what you would accept. For what it's worth, I agree that the polls exist and should be documented, but I also agree with the case made for caveating that information. 88.145.150.206 (talk) 00:11, 11 November 2019 (UTC)
 * I am unaware of any election candidate for a political party adding polling commissioned by that party to this article. Stop making false allegations.
 * Multiple editors have supported keeping the polls in question in this article. Past practice has generally agreed that membership of a political party does not require a COI statement. Editors are allowed to be interested in politics and to support different candidates. WP:COIPOLITICAL says, "Political candidates and their staff should not edit articles about themselves, their supporters, or their opponents." Ergo, I would be concerned if a candidate in this election or their staff were editing without some COI declaration. I see no evidence of that. I suggest you strike all your breaches of WP:AGF, refrain from WP:ADHOM and focus on content.
 * As far as I can see, we have consensus that the LibDem-commissioned polls that meet BPC standards stay in. There is also support for some introductory text to this section, based on WP:RS coverage. Let's work on that content then. Bondegezou (talk) 10:52, 11 November 2019 (UTC)


 * I've added a note about potential publication bias in constituency polls, as a starting point - wasn't unable to find a better source. The Land (talk) 13:40, 9 December 2019 (UTC)

No Direct Data Link
Can I suggest no polls are added until there is an actual BPC Pollster or Ashcroft data link. A link to a media outlet isn't necessarily correct & not all data is available.Stevenxlead (talk) 16:12, 9 December 2019 (UTC)

Lost Polls
What happened to the Survation GMB poll that was the latest this morning? Stevenxlead (talk) 16:14, 9 December 2019 (UTC)

TBC
What does TBC in the opinion polls? Does it mean unclear, is it a geographic area, what is it? — Preceding unsigned comment added by 2602:304:B3E1:3E30:547A:32BB:1BC9:A639 (talk) 19:18, 9 December 2019 (UTC)


 * It stands for 'To Be Confirmed'. It means that the poll has been announced (for example on twitter or a newspaper article) with partial results but the full breakdown is not yet available. Perokema (talk) 21:48, 9 December 2019 (UTC)

Pending review
Why is anyone bothering to edit this article now? If Wikipedia can't come up with a better system than this, whereby a backlog of "unapproved" edits builds up due to tight control of who can do the approvals, then people are going to look elsewhere for the latest information. What a stupid protocol! 31.52.163.214 (talk) 10:16, 10 December 2019 (UTC)
 * today incident of somebody trying to add a fake poll highlights the very reason this system is needed, i know its annoying but vital to avoid rouge bot using this page to share mis-information — Preceding unsigned comment added by 165.225.81.98 (talk) 12:48, 10 December 2019 (UTC)
 * I'm guessing 'rouge bot' is responsible for 'made up' information :P Btljs (talk) 15:24, 10 December 2019 (UTC)
 * What did the fake poll say, out of interest? I'm trying to check something else.  Which company and what Tory lead?  Ganpati23 (talk) 20:14, 10 December 2019 (UTC)
 * https://twitter.com/Britain_Erects/status/1204423225461137408 — Preceding unsigned comment added by Popeter45 (talk • contribs) 23:08, 10 December 2019 (UTC)

YouGov MRP
I've just edited in the 2nd and final YouGov MRP poll. But I'm surprised that it's gone 9am UK time and it was released at 10pm, 11 hours ago, which is very unusual - editors are usually falling over one another to edit in the latest poll. Is there some particular reason why it wasn't there? I explored further and it looks to me as if the 1st YouGov MRP poll isn't included in the main list. Really. That's this one. I can't see it, anyway. Poll size 100,319 interviews. I'd edit it in but wanted to be sure. Boscaswell  talk  09:18, 11 December 2019 (UTC)


 * That's because MRP is not a poll. It's a statistical model and not comparable with traditional polling, even if it gives similar results. It shouldn't be included in the table. Andy mmu talk 11:01, 11 December 2019 (UTC)

Surely the percentage voter intention is effectively a poll, the other research is the use of demographic data to predict constituency results. On their website, YouGov refer to "predicted vote shares in our final poll" https://yougov.co.uk/topics/politics/articles-reports/2019/12/10/final-2019-general-election-mrp-model-small- Saxmund (talk) 11:39, 11 December 2019 (UTC)


 * MRP uses a methodology different from any of the other polls and, in my view, percentage votes should be in a separate table. At the moment we don't have MRP as party percentage anywhere on the page.Cutler (talk) 14:23, 11 December 2019 (UTC)

We now have the MRP polls in the main table. I can live with that if it is the consensus. The two recent polls are identified as MRP. However, some of the earlier results were MRP too and we have lost that information.Cutler (talk) 10:35, 12 December 2019 (UTC)

Graph Update
Since the election is tomorrow, can someone update the graphs please? TheMysteriousEditor (talk) 22:25, 11 December 2019 (UTC)
 * The election is today...the graph is 5 days out of date. TheMysteriousEditor (talk) 09:54, 12 December 2019 (UTC)
 * Anybody? TheMysteriousEditor (talk) 12:09, 12 December 2019 (UTC)
 * I've just updated it after being away for a few days, the last update was three days ago as you can see if you click on the graph and scroll down to see the updates. If this is a particular issue for you and you actually want to help, then for next time you could have a go at implementing Selfie City's suggestion in Talk:Opinion polling for the 2019 United Kingdom general election. Bellowhead678 (talk) 16:14, 12 December 2019 (UTC)

Should exit polls be included here?
The section title is reasonably self-explanatory. Is this a good place in Wikipedia to hold data from the exit poll(s)? If not, where ought they to go? RERTwiki (talk) 11:41, 12 December 2019 (UTC)


 * Good question. Depends. I reckon if that bbc/itv exit poll gives a seat prediction that should be included with the seat predictions. If there's no seat prediction given by bbc I'd put it in the list of polls (very clearly highlighted as different), but if there is, in the seat predictions. In 2015 Ashdowns' millinphagy got some comment too, so see what's newsworthy after the fact.86.173.65.225 (talk) 17:00, 12 December 2019 (UTC)
 * Agreed with 86.173.65.225. Bondegezou (talk) 21:45, 12 December 2019 (UTC)