Talk:Opinion polling for the next United Kingdom general election

Graph too smooth?

The graph by @DimensionalFusion: is very good, but I think it has the issue that the trend lines are so smooth that they don't really show actual trends well. Compare with File:UK opinion polling 2010-2015.svg where there is more movement in the trend line over shorter periods of time- in the graph on this page by comparison, it covers a period of 7 months so far but the lines are all extremely smooth and don't really accurately follow the trends. The most recent graph update's addition of detail to the Lib Dem and Green lines is good, but the other parties could perhaps do with more accurate trend lines, e.g. the main increase in Reform vote share was the period from mid-November (averaging around 18%) to mid-January (averaging around 24%), with the rise having been slower since mid-January, but the trend line does not really show this.

Also again like the 2010-2015 graph example, I think it would make more sense if the beginning of the trend lines in July started from the actual election result- having the Conservative line start as low as 20% based on a single poll in the start of July rather than the actual election result, and the Labour line as high as 36% again thanks to the same single poll, is a bit misleading. Chessrat ^{(talk, contributions)} 12:51, 26 February 2025 (UTC)

About the script:

@Chessrat: I believe it was originally created for German federal elections (which use Proportional Representation, making smaller parties much more relevant) so it has the option to use different "span" sizes, which are essentially how smooth the line is (because of the aforementioned Proportional Representation thing, different parties can have different span sizes).

In the most recent update I noticed that despite the Lib Dems getting 16% on two seperate polls (I looked into it – their highest since June 2019), the trend line was essentially the same as it was before – so I changed the Greens' and Lib Dems' span sizes to be 0.8 whilst keeping Labour, Reform, and Conservatives' at a span size 4 as they are bigger parties and changes tend to matter less. So currently I have the span sizes set to 4, 4, 4, 0.8, 0.8

By default the value in the script is 0.05 (which is crazy small), and smaller span values tend to give more credence to outlier polls on the graph, whereas larger values (like the present 4) tend to miss short-term changes.

Anyway, should all parties have the same span size or should it be different? If so, what should the new values be?

About your second point:

The lines do start from the last election result – in the script, the previous election result is counted as a "poll" too which is why Labour's line starts so high: it's about halfway in-between the actual result and the July poll. Incidentally, this could also be fixed by setting a lower span size! DimensionalFusion (talk · she/her) 13:11, 26 February 2025 (UTC)

Realised there was already a discussion. All parties should have same span, so either revert span change or apply it to all parties. I think two polls at 16% hardly counts as a trend so ofc there is almost no change. Perhaps, a middle ground between the new span and old span can be used on all parties. The Lib Dems and Greens now look wobbly despite so far there being little change. (Update as of 13:28) @DimensionalFusion, I see that you decreased the span for the Lib Dems and Greens on the graph. But I believe this is a mistake. I do not see a reason for it. According to documentation, "parties which have fewer polling data" need "a separate higher 'span' should be used". In other words, parties with more polling data should have a lower span than those with more. But Labour, the Conservatives, Reform, Lib Dems and the Greens all have an equal amount of polling data. Every pollster mentions all of them. Which would mean that they should all have the same span. Now, the Greens and Lib Dems both look very wobbly despite consistent polling. —Profzed! 13:22, 26 February 2025 (UTC)

I would say all parties should have the same span size for consistency. I do think a more "wobbly" line is preferable for showing small changes over time- see e.g. File:UK opinion polls 2017-2019.svg in which short-term changes can be clearly seen- all previous articles have much less smooth lines like that. What I would be inclined to do is roughly: setting the span for all parties at something like 0.3, and also weight the last general election result to function as the equivalent of 5 polls with that result (because the actual election result is much more important than the two July WeThink polls for long-term trends). It's possible that that would lead to too much of a focus on outliers but I think it's at least worth trying.

Regarding @Profzed's comment that there is "little change" in Lib Dem/Green polling... the Lib Dems were polling around 13% in October, dropped to 12% by December, and have nowadays been having more and more poll results around 14%. I think this is a meaningful change that should be reflected in the trend lines. Chessrat ^{(talk, contributions)} 13:51, 26 February 2025 (UTC)

Also, note that small changes for the big parties are still quite meaningful- one of Lab/Con/Ref being a few percent above the rest would likely have quite a meaningful change to an actual election result if an election were held. Chessrat ^{(talk, contributions)} 13:53, 26 February 2025 (UTC)

In seperate polls in just Feb so far, Reform UK has been anywhere from 29% to 24%. Also in Feb, the Cons have varied from 25% to 18% – if the Lib Dems or the Greens were increasing by 7 points in the same month they'd be on about 19% and 15% respectively. My point is that small changes matter but within big parties they matter less DimensionalFusion (talk · she/her) 14:06, 26 February 2025 (UTC)

@Chessrat There isn't really a difference between 12% and 13%, whilst YouGov have always polled Lib Dems at 14% with the exception of the latest poll. (I think there was only something like one non-YouGov poll at 14%.) I do agree though it was probably too straight previously but I feel like it is too wobbly now. Do you know which span was used in the 2017-2019 graph? Because it appears less wobbly there. If not, we could ping the creator. (And if it is was 0.3, maybe I'm just being weird.) —Profzed! 16:01, 26 February 2025 (UTC)

I don't know, but as far as I can tell from the documentation on the gitlab page, span values should be different depending on the length of time covered by the graph- i.e. the longer that the period is, the lower the span value should be. Chessrat ^{(talk, contributions)} 16:16, 26 February 2025 (UTC)

@Chessrat I see. That'd make sense, given the additional polling data. —Profzed! 16:17, 26 February 2025 (UTC)

Appearances can be deceiving! The 2017–2019 and the 2015–2017 graphs only feel like that because of the shorter timespans – both were only 2 years whereas 2010–2015 and 2019–2024 were both 5. This graph assumes 5 years whilst it's being updated – when the election is called, the graph will thin down DimensionalFusion (talk · she/her) 16:17, 26 February 2025 (UTC)

I agree - span should be consistent. There is already too much assumption being put into this (whichever span is chosen is completely arbitrary) and then altering span based on how popular a party is just adds more to that.

However, (partaking in the arbitrary discussion_ I do think 0.3 is wayyy too granular. My personal opinion is that a middle ground should be sought between the 4 currently given to major parties and the 0.3 proposed for smaller parties. Perhaps 2? This is, broadly, a significant drop from 4 at the moment and will allow little changes to be seen in a way they can't at the moment (7 months on from the election and the trend lines are still effectively quadratic) but would avoid showing small, margin of error differences and can still act as successful trendline.

I also believe the idea of including the 2024 results - and possibly giving them increased weight - is quite strong and something of which I am broadly supportive although I think such a change may need a decent consensus behind it because it changes what the early parts of the graph represents, taking it from showing opinion polling over time to opinion over time (a subtle but real difference). Kirky03 (talk) 16:16, 26 February 2025 (UTC)

I've updated the graph to use a span of 0.3 to see what it looks like DimensionalFusion (talk · she/her) 16:18, 26 February 2025 (UTC)

@DimensionalFusion Looking at it, I think 0.3 is defintely a no-go. —Profzed! 16:23, 26 February 2025 (UTC)

@Kirky03 Since 4 was the previous span and 0.3 was @Chessrat's suggested span, what would you think about 2.15 as a middle ground? —Profzed! 16:21, 26 February 2025 (UTC)

@Profzed: I've now updated the graph to use a span size of 1, which is a nice round number to decide on lower or higher DimensionalFusion (talk · she/her) 16:29, 26 February 2025 (UTC)

@DimensionalFusion That is definitely a lot better. I think I'd go slightly higher, but it'd be interesting to see everyone else's opinions as well. —Profzed! 16:42, 26 February 2025 (UTC)

Looking at the versions with 0.3 and 1, my feeling is that 0.3 is probably slightly too wiggly, but 1 is far too smooth and misleading. Reform mostly stayed static in the polls from August til the end of November and has been mostly static for the last month, and the three biggest parties have all had minimal movement in the last month. The 1 version of the graph fails to show any of this and just shows a steady rise for Reform and steady fall for Labour over the last seven months, which isn't accurate. I would probably go for something like 0.5 or 0.6- 1 is way too high in my opinion. Chessrat ^{(talk, contributions)} 17:01, 26 February 2025 (UTC)

@Chessrat Actually, that is a very good point. —Profzed! 17:32, 26 February 2025 (UTC)

Alrighty @Profzed: @Chessrat: so I'll update the current graph to use a span of 0.5 and use 0.5 for all parties in future graph updates. If it needs to be upped to something like 0.6 then we can discuss that later! DimensionalFusion (talk · she/her) 20:10, 26 February 2025 (UTC)

Looks good to me. Thank you for doing the work on this! Chessrat ^{(talk, contributions)} 20:18, 26 February 2025 (UTC)

👍 —Profzed! 20:54, 26 February 2025 (UTC)

I might raise this point as well - should similar changes not be made to the Scotland graph?

I think the selected span of 0.5 should not be used here (due to the much lower density of polls) but perhaps a span lower than the current 4 but higher than 0.5?

I'm not convinced either way but it might be worth considering Kirky03 (talk) 23:37, 26 February 2025 (UTC)

Probably, but you'd have to find whoever updates that graph and pester them DimensionalFusion (talk · she/her) 08:44, 27 February 2025 (UTC)

Hi, the 0.56 span we selected is miles better than the 4 is used to be but I've started to get concerned that it is too low, mainly because the Conservative and Reform lines have been fluctuating around the same number, all within the margin of error. If a party has an improvement for a week within the standard margin of error that most pollsters use and then goes back to where it was a week later, I'm not sure the graph should be picking that up as my understanding is it should be capturing the trend rather than every granular movement.

I did some experimenting and a very small increase from 0.56 to 0.7 seems to follow the same trend (and be reactive to short term changes if they are significant enough) but flattens out the wiggly lines. I've just updated the graph but before I did so, I uploaded the same graph with a 0.7 span so if you go onto details, you can view how to graph would look with 0.7 instead.

I understand we reached an agreement before so if people are happy enough with how it is, I'm not going to push on this but I figured I'd raise it considering the recent trends.

Tagging the people involved: @Chessrat @DimensionalFusion @Profzed Kirky03 (talk) 10:12, 30 April 2025 (UTC)

Also @Bondegezou Kirky03 (talk) 10:36, 30 April 2025 (UTC)

Thanks. Going slightly smoother seems good to me. Bondegezou (talk) 10:41, 30 April 2025 (UTC)

I have a slight preference for keeping it at 0.56 as there has been a meaningful drop for the Conservatives in the polls lately (7 of the last 9 polls have them on 20% or 21%), but I wouldn't strongly object to 0.7. Chessrat ^{(talk, contributions)} 12:14, 30 April 2025 (UTC)

@Kirky03 I'd be fine with increasing it slightly, perhaps to 0.6, but 0.7 smoothens out Labour's boost over Ukraine stuff. —Profzed! 15:52, 1 May 2025 (UTC)

Since there hadn't been a huge amount of enthusiasm to change, I felt there wasn't enough reason to change status quo and decided to wait and see if anyone else contributed. I see that @DimensionalFusion has today experimented with a span of 1. Firstly, what do people think of this, and secondly are you DimensionalFusion happy to meet somewhere in the middle? Kirky03 (talk) 17:46, 14 May 2025 (UTC)

Agreed that the change to 1 needs to be reverted– it's incredibly misleading to imply the recent Reform surge was gradual rather than what it actually was, a sudden jump after the local elections Chessrat ^{(talk, contributions)} 19:10, 14 May 2025 (UTC)

Hiya Chessrat and Kirky03, I've reverted the graph to using the previous span of 0.56. I felt that the graph was too squiggly and with 3 months of extra data a larger span would work DimensionalFusion (talk · she/her) 20:18, 14 May 2025 (UTC)

I agree 0.56 is too squiggly. How do you feel about a middle ground?

There's the 0.7 I suggested or Profzed said maybe an only slight increase to 0.6, although that doesn't fix the squiggles much. Kirky03 (talk) 20:51, 14 May 2025 (UTC)

It looks like somebody put the graph back to the old way yesterday, but someone else turned it back to the plate of spaghetti today.

I prefer the old "smooth" graph. It's less than a year since the last general election and the next one could be three or four years away, so the less granular trend lines are better. As it is now, we see multiple crossover points between parties just weeks apart, when what I personally want is to pinpoint a clear moment when one overtook the other. James Tweedie (talk) 13:03, 15 May 2025 (UTC)

Hi. I am late to the discussion. I suggest the new span, whatever is currently in use at the time of this message, is too small. The wiggles in the lines now shown are just artefacts, they don't represent anything real. The graph is now not smooth enough. Bondegezou (talk) 13:28, 27 February 2025 (UTC)

Agreed. CR (talk) 13:48, 27 February 2025 (UTC)

You can check the file history of the graph if you'd like – there's lots of options there of different spans. Which one looks best to you DimensionalFusion (talk · she/her) 14:04, 27 February 2025 (UTC)

@Bondegezou I think the issue is finding a span without these swiggles but which is not inaccurate. So 1 is definitely too high as @Chessrat pointed out. Maybe @DimensionalFusion could try some spans in between and either come to a conclusion herself or perhaps update the graph several times so we can all discuss them. Since it's now at 0.5, we could test 0.6, 0.7 and 0.8. And assuming the span can be to 2 d.p., we choose an in between value if need be after testing those. —Profzed! 14:17, 27 February 2025 (UTC)

I've updated the graph with the latest poll and to use a span of 0.6 DimensionalFusion (talk · she/her) 16:20, 27 February 2025 (UTC)

@DimensionalFusion, thank you for testing 0.6 out. I had a look and I think it is generally nicer, but it re-introduces the error with Reform increasing during February. Could you test out values between 0.5 and 0.6 and upload the highest one without this issue? Thanks. —Profzed! 16:57, 27 February 2025 (UTC)

At 2 decimal points, we are getting so precise we are at risk of overfit Kirky03 (talk) 17:48, 27 February 2025 (UTC)

@Kirky03 I suppose. If we stick to 1 d.p., I think we should use 0.5 then. But I still think we could something in between like 0.55 assuming it does not have the aforementioned issue (but maybe we don't go super precise like my previous reply suggested). The parameters can always be changed though; this isn't a prediction model which has to "lock in". —Profzed! 18:22, 27 February 2025 (UTC)

I don't agree- all of the wiggles in the graph all have statistical meaning to them and are backed up by large numbers of polls. They're also similar to what's seen in past graphs, for instance the November 2024 Conservative peak is pretty similar looking to the May 2013 UKIP peak in File:UK opinion polling 2010-2015.svg. I think the current graph is perfect and I don't see a need to change it. Chessrat ^{(talk, contributions)} 14:30, 27 February 2025 (UTC)

@Chessrat I think a bit more smoothing would be nicer. Look at Greens Oct–Jan. And I think the Lib Dems went up a bit too readily as well. It should definitely go up but the amount it went from two polls is quite a lot. Reform dropped a bit Nov, which could probably just be smoothened out since support didn't really drop. So I think a minor increase in span could be good. —Profzed! 14:41, 27 February 2025 (UTC)

Not sure what the problem with the Green line is? It's closer to 7% at the start of October when most polls had them on 7%, and slowly goes up as they get more 8%, 9%, and even 10% polls. The Lib Dem line may be very slightly too much but I don't think it's enough to require much of a change. Maybe 0.5 to 0.6 would be okay but I don't think there's a big issue really. Chessrat ^{(talk, contributions)} 14:50, 27 February 2025 (UTC)

@Chessrat 0.6 seems reasonable. It just needs a minor increase. I just thought it might be good to have several values to compare. —Profzed! 14:59, 27 February 2025 (UTC)

I'd be okay with that. I tried installing the software for generating graphs myself but ended up with enough error messages that I gave up! Chessrat ^{(talk, contributions)} 15:01, 27 February 2025 (UTC)

Yeah, I think the 0.6 version is good. Chessrat ^{(talk, contributions)} 23:53, 27 February 2025 (UTC)

@Chessrat I did notice though that Reform is now increasing in February again according to the graph. I was thinking something in between like 0.55 maybe. —Profzed! 07:51, 28 February 2025 (UTC)

I've update the polling graph with the latest poll and changed the span to 0.55 (as well as fixing the legend so I didn't have to manually add it in every time, plus weighting the start data more) DimensionalFusion (talk · she/her) 10:25, 28 February 2025 (UTC)

@DimensionalFusion Thank you for all your work! —Profzed! 11:10, 28 February 2025 (UTC)

Thanks DimensionalFusion for doing all the iterations on this. I think the current 0.55 is OK. I'd personally go for something smoother on the basis of the known confidence intervals in these figures, but not worth arguing over! Bondegezou (talk) 10:51, 28 February 2025 (UTC)

Not that it is worth strongly arguing over, but I tried out a change to the graph layout (see page history on File:Opinion polling graph for the next United Kingdom general election (post-2024).svg) involving:

• 1) Changing span to 0.15- shows off changes more clearly;
• 2) Duplicating the 2024 general election result 15 times to ensure that the graph starts from the point of the election result;
• 3) Moving the key so that it is next to the lines and thereby easier to immediately find
• 4) Adding an "Others" line showing the combined voting intention for the parties outside of the big five.

If anyone agrees with me that any of those changes may be good, might be worth discussing further. Chessrat ^{(talk, contributions)} 15:38, 29 May 2025 (UTC)

1) While interesting, I don’t think it’s very informative as the lines wiggle all over the place, and so it is not possible to discern trends from noise. May be helpful to show people over-analysing 1-point changes from week to week!

2) I am of two minds. It’s well-known that general elections don’t necessarily match opinion polls: if in 2029 we want to compare parties to how they were polling just after the 2024 election, it would be better to compare to the aggregate rather than the general election result. But equally, a general election is the most accurate poll there is.

3) I prefer it as it is. This will become messy as soon as the lines get close, and I think the colours are fairly well-known.

4) I like the idea of adding “Others”, although I think that requires discussion of when “Others” becomes its own thing. Reform UK would have been considered “Other” once. Sammy zammy1 (talk) 22:00, 29 May 2025 (UTC)

I personally want the span increased rather than decreased from this point, I think it can be too squiggly and is too reactive to margin of error changes.

As for the 2nd point, I can get behind this. At the moment, the election is repeated three times and I can support increasing that number to ensure it is a bit more fixed.

The key is a nice change although I'm concerned about what happens when polls are neck and neck like last month. It would potentially be difficult to determine which label belongs to which line and that puts me off the idea.

Others line is decent and I can possibly support it. Are you calculating that by adding up the PC+SNP+Other columns or by subtracting the main parties from 100? Kirky03 (talk) 09:22, 30 May 2025 (UTC)

@Chessrat

1. Too squiggly
2. The graph should show polling results. I think the election result should just be there for reference, with it counting as one poll just to have it coming from around the right direction.
3. Looks far worse.
4. No strong feelings.

—Profzed! 14:20, 2 June 2025 (UTC)

Preferred prime minister proposal

This discussion has been copied and moved to Talk:Leadership approval opinion polling for the next United Kingdom general election#Preferred prime minister proposal. --2.101.240.129 (talk) 23:01, 29 June 2025 (UTC)

Freshwater Strategy

Freshwater Strategy is now applying to join BPC and applying their rules, so is it time to add them back? Billytanghh (talk) 13:52, 17 April 2025 (UTC)

there had been multiple detailed discussions on pollsters to include a few weeks ago, my recollection was that as long as pollster had disclosure of methodology and published data tables they should be included provided they match Wiki for reliable sources. It is really for an editor to provide those details on there inclusion Pugpa2 (talk) 20:07, 17 April 2025 (UTC)

If they already follow the rules of the BPC, they are good enough for most editors of this article, and if they're in the process of joining then very soon they'll be good enough for everyone so I struggle to see why not. Kirky03 (talk) 22:09, 17 April 2025 (UTC)

Yes, should also backdate inclusion of polls that were abruptly deleted by some editors. 152.37.116.156 (talk) 18:27, 23 April 2025 (UTC)

Yep, its all down to reliable sources as per wiki standards Pugpa2 (talk) 11:55, 27 April 2025 (UTC)

BPC membership really isn't relevant– see above RfC. I'll readd the polls that were removed once I have some spare time. Chessrat ^{(talk, contributions)} 13:33, 26 April 2025 (UTC)

Freshwater Strategy announced its membership of the BPC on 23 June 2025. Assuming that all entries prior to 23 June will be left italicised with the non-BPC footnote, should it be made clear somewhere that non-BPC membership applies at the time of polling? Or would that be obvious? Freshwater is now included in the "Most recent" table (and presumably in the 30-day average). What about the Graph (in particular regarding past results)? This begs the question that as there was agreement via the RfC to include non-BPC pollsters as long as they satisfy WP:RS, etc., shouldn't they have been included in the Most recent table and graph anyway? That logic would still apply to Ashcroft, i.e. shouldn't Ashcroft be included in those? --92.14.26.105 (talk) 20:05, 9 July 2025 (UTC)

I have given a possible answer to my first question in a new topic #Proposal for new section on BPC membership below. 92.14.26.105 (talk) 21:01, 9 July 2025 (UTC)

Rounding

Could editors please adhere to long establish practice across many areas re rounding % to nearest whole number, point 5 or more is rounded up less than point 5 is rounded down. It would appear some of our editors don't understand this basic convention. Pugpa2 (talk) 04:53, 1 May 2025 (UTC)

The pollsters do the rounding, not you. If the pollster has reported something as 24-24-24 in their headline figures, then that is the figure. It is not up to you to change what polls say, that is vandalism of this article and original research, not sourced from reliable sources. This is not the first time you have tried to substantially modify the polling tables to match your own personal views (e.g. trying to remove all YouGov polling because you didn't understand their methodology). Please stop. FriendlyDataNerdV2 (talk) 08:21, 1 May 2025 (UTC)

The data tables are the very best source we can have and should always be used in preference to any other source. Pugpa2 (talk) 10:07, 1 May 2025 (UTC)

It appears the issue is the cited source. MoreInCommon reported it as 24-24-24 but the data tables cited don't and report the unrounded figure which makes 24-24-24 seem a bit random. Kirky03 (talk) 09:08, 1 May 2025 (UTC)

Aye it is about sources and which are the most reliable, me I would say the data tables are the gold standard and should always be preferred to any other source. Pugpa2 (talk) 10:04, 1 May 2025 (UTC)

@Pugpa2 You are double rounding, I believe. I don't know the exact original figure, but for example, 24.45 can be rounded to 24.5 which can be rounded to 25, but 24.45 should be rounded to 24, not 25. —Profzed! 15:54, 1 May 2025 (UTC)

Again, there is no established consensus to say that Wikipedia should disregard what pollsters themselves have reported and instead allow editors to add in numbers that they have calculated themselves. There is absolutely no "long establish practice" of doing that, and I would ask you again to please stop vandalising this page by changing polls from what the pollster actually reported. It's not ok. More in Common conducted the poll, they say it's 24-24-24. That is the end of it, they are the reliable source, not you. FriendlyDataNerdV2 (talk) 16:43, 1 May 2025 (UTC)

@FriendlyDataNerdV2 To clarify, this is the stance I was taking. I was simply pointing out that there is nothing strange about this apparent discrepancy. —Profzed! 17:00, 1 May 2025 (UTC)

I agree there is nothing strange about this discrepancy but believe the data tables should be used as prime source of information Pugpa2 (talk) 18:19, 1 May 2025 (UTC)

And again there is long established practice of rounding across not only Wikipedia but the wider world. You do your self no credit by trying to create false comparisons, data tables are not equivalent to opinion pieces, and rounding these numbers to fit wiki articles like this is very well established practice across many polling wiki's. This is not vandalism but best practice. Your argument relies on taking an article with dubious headline figures and asserting that is more reliable than actual data tables, that is a nonsense Pugpa2 (talk) 17:11, 1 May 2025 (UTC)

I am aware of the idea of double rounding however I am not aware of anything to support the assertion this has happened in this case. Data tables have always been the most reliable of sources and until someone provides a better source then we should go with those. Pugpa2 (talk) 17:05, 1 May 2025 (UTC)

More in Common says the correctly rounded figures are 24-24-24 - you are double rounding. Nicholas13t (talk) 17:25, 1 May 2025 (UTC)

Can you provide a source for that assertion please Pugpa2 (talk) 18:00, 1 May 2025 (UTC)

I literally linked the primary source. Nicholas13t (talk) 19:57, 1 May 2025 (UTC)

The executive director of the pollster also posted the same figures. Nicholas13t (talk) 20:08, 1 May 2025 (UTC)

Regardless, just looking back at the data tables, you insist on rounding, it is correct that the cell for Labour in sheet "Totals" cell C4 displays 24.5%, but the actual number in that cell is 24.4844111030622%. Nicholas13t (talk) 20:28, 1 May 2025 (UTC)

Ahhh, that's helpful, thank you. Pugpa2 (talk) 04:20, 2 May 2025 (UTC)

" an article with dubious headline figures" - what on earth are you talking about? I'm citing the pollster themselves, not an "opinion piece"! https://www.moreincommon.org.uk/latest-insights/more-in-common-s-april-mrp/ If you think that the pollster's summary of a poll is an "opinion piece" I would suggest that you do not understand how polling works and should probably stop imposing your own flawed interpretation of polling onto a Wikipedia page that is relied upon as an important resource. Your sources are deeply vague; "the wider world", "best practice". my source is the pollster who conducted the poll. This is not a difficult discussion. You are quite simply incorrect FriendlyDataNerdV2 (talk) 17:28, 1 May 2025 (UTC)

My source is the data tables produced by the pollster, you continue to avoid that matter, my figures are the pollsters official numbers.

Rounding is a well established process they even have it as part of GCSE https://www.bbc.co.uk/bitesize/articles/z7nqs82#znpyqyc Pugpa2 (talk) 18:03, 1 May 2025 (UTC)

The document you refer to in its first paragraph has a link to the data right there before anything else, why do you find it so difficult to actual read the whole document and especially the data that the pollsters provides https://www.moreincommon.org.uk/latest-insights/more-in-common-s-april-mrp/ Pugpa2 (talk) 18:09, 1 May 2025 (UTC)

@Pugpa2 When the pollster's headline figure is 24% and the datatable says 24.5%, claiming they are on 25% is clearly double rounding. MiC have the actual data, they are the reliable source, let's trust their ability to round. —Profzed! 18:16, 1 May 2025 (UTC)

Thank you another editor pointed out that in the actual data tables they are rounded to 3 significant numbers where is when you go into the cell you see the complete figure showing about 10 significant numbers. Pugpa2 (talk) 04:22, 2 May 2025 (UTC)

Edit Requests and Suggestions

Regular - and other - editors of this article, please note. I am just a ordinary punter who frequents this page out of interest and I am getting sick to the back teeth with all the chopping and changing in recent weeks and months. The latest example is the sudden appearance of a so-called "aggregate" row in the spurious Most Recent Polling table. What does that even mean? It appears to be a simplistic rounded average of the headline VI of each pollster's latest poll (within a 30-day limit).

As I am not that au fait with Wikipedia rules, I have no idea whether this latest addition is compliant or not, but it is certainly arbitrary. Personally, I would like to see the entire table removed, but if you insist on keeping it, I also notice that the non-BPC members that are included in the main tables are not included in the Most recent table and hence not in the "average" or whatever you want to call it. Again, if you insist on keeping the new row, then why not a 30-day "aggregate" of *all* polls for comparison (albeit the differences are likely to be negligible)? Either way, at the very least an explanation of what is actually being calculated should be provided.

I would also suggest that it is now time to transfer the Approval Polling tables to a separate article, as per previous electoral cycles. While on the subject, I notice that the Leadership Approval table still has direct hyperlinks in the pollster column whereas these have been replaced in the Party Approval table. The new Northern / Midlands table has a similar link - is it the default or something? Again, if you insist on making changes, please try to be consistent.

As far as I can make out, several of these unnecessary changes have been introduced by first-time or at least infrequent contributors, intent on adding their own latest big idea with little or no discussion, let alone "consensus".

Postscript: having just written the above, I see that yet another new argument has broken out about rounding, involving some of the "usual suspects" among you. Stop it. Just stop it. 92.14.29.28 (talk) 23:00, 1 May 2025 (UTC)

Personally I am against this rolling aggregation section, as I'd argue it is does teether on the edge of what is permitted as per the rules on self-published works, but I am not eager to commit to any substantial changes without overwhelming consensus, given how much unnecessary edit warring has occurred on this page so far this year. Nicholas13t (talk) 23:41, 1 May 2025 (UTC)

Agreed. "Most recent by pollster" is maybe a bit superfluous on its own but has its uses, but averaging between them by any means over an arbitrary period is entirely OR [1]: if people want to show aggregates then they need to be in a separate section (maybe even a separate article, given how long this one gets after five years) and be the aggregates produced by or reported on by a RS, rather than made up in-house by whatever method people think "looks right" today.

[1] Yes, this would technically apply to the averaging lines on the graph, too, but when parties and polls are so close together that the dots overlap as much as they are I can see the case for having something, so long as not too much time is spent on making it "look right". In the most recent table, all the raw data is already available and clearly separated so there's no need to add anything extra. 2A02:C7C:DA05:1A00:692:26FF:FE59:D09B (talk) 11:10, 6 May 2025 (UTC)

Again I have to agree about the aggregate/pollster section but think we have more pressing matters to address first. In particular the original poster of this topic about the time to split this article as per previous incarnations of it. Broadly this seems straightforward where approval/hypothetical goes to separate article. However in the review of the 2024 article there was an RfC that decided splitting off sub national polling etc to a separate article as well. If so could we consider calling it something like Nations, Regions and constituencies. Pugpa2 (talk) 16:53, 6 May 2025 (UTC)

The earlier edit warring on this article has left behind a few inconsistencies.

It was eventually agreed to include certain non-BPC pollsters, but as I pointed out above, they have not been included in the Most recent polling table and hence not in the so-called "30 day aggregate" (moving average). It looks as though they are not included in the Graph either.

I also mentioned that there are still direct hyperlinks in the Leadership approval table and in the recently added sub-national Northern / Mildands table.

Regarding the separation of the Approval tables and possibly the Sub-National tables, I have had a look at previous articles back to 2015 and find the following:

"Leadership approval opinion polling for the 20yy United Kingdom general election"

1) there are separate articles from 2015 to 2024; these did not include Party approval polling - indeed this only appears in the current article i.e. "Opinion polling for the next United Kingdom general election".

2) 2019 and 2024 included Preferred Prime Minister polling (and also Prime Minister and Chancellor of the Exchequer pairings).

3) 2017: Preferred Prime Minister polling was included in the main polling article i.e. "Opinion polling for the 2017 United Kingdom general election". There was no corresponding section or article prior to this.

"Sub-national opinion polling for the 20yy United Kingdom general election" (separate articles for 2015 and 2024 only):

1) 2024 included Constituency polling.

2) 2015 Constituency polling was in a (third) separate article: "Opinion polling in United Kingdom constituencies (2010–2015)". This included polling for some by-elections.

The link to this is from the main article, under "See also".

3) 2017 and 2019: Sub-national polling was included in the main article. 92.14.27.51 (talk) 18:31, 14 May 2025 (UTC)

Thanks for that I would add one other thing, in reviewing the 2024 article an RfC was made re separating Sub National into a separate article, the view reached was that is should have its own article.

So far I am not hearing any voices raised that would suggest maintaining this one big article, Would suggest the split happens following previous most recent articles.

1 This article continues as main article on Polling for UK GE and includes current headings and sub headings of

Graphical Summary

National Polls results

Seat Projections

2 Seperate article for

Approval Polling

Hypothetical scenarios'

3 Seperate article for Sub national polls to be renamed as Nations, Regions and constituencies

Sub National Poll results

Hopefully we can reach a quick agreement on this which will allow us to move on and hopefully tackle many of the residual anomalies in this article. Or do we have to go for an RfC even for what is consistent with long established practice? Pugpa2 (talk) 09:43, 16 May 2025 (UTC)

Hopefully not an RFC, I'm not sure I have this in me again.

I'm on board with something similar to e.g. 2019 with national poll results, seat projections and sub-national poll results all on one article and then a seperate article for approval polling.

Really, it is split as anything relating to voting intention on any level is on this article, and everything relating to approval polling on the new one.

I personally don't like the idea of spliting away sub-national and think 2019 is the way to go. Kirky03 (talk) 11:12, 16 May 2025 (UTC)

I agree about an RfC but would disagree that 2019 is the model, I think the 2024 is a better model and its RfC about splitting subnations, but lets see what others may think. Lets be positive we agree about splitting the article. Pugpa2 (talk) 11:24, 16 May 2025 (UTC)

Yee, at the very least we can agree that approval should be split. This gets brought up occasionally and it always seems people are supportive but it never actually happens.

Maybe we can start by someone splitting off approval polling as is usually done and then we can iron out the specifics about the things that are less consistent, such as creating a 3rd article for sub-national. Kirky03 (talk) 13:56, 16 May 2025 (UTC)

Just some very minor points. The previous articles for sub-national polling (2015 and 2024) were titled "Sub-national opinion polling for the 20yy United Kingdom general election"; obviously the current one, if it's decided to have it, would replace the year with "next". It can be renamed if you wish, up to you, but wouldn't it be preferable to keep to the same format, thus creating a similarly titled series? Also, the 2024 and 2015 main articles had sub-headings for sub-national polls with links to the separate articles. This hasn't usually been done for Approval polling though (links for those are in the box at top right of the page and also under the "See also" section).

Preferred Prime Minister polling was included in the 2024 and 2019 Approval articles (though not 2017). I can see why, as both concern leadership, but a preference going forward, which is essentially hypothetical, is not quite the same as approval, which rates current/past performance. Perhaps a compromise would be to include Preferred PM in the Approval article, but keep the sub-heading in the main article with an appropriate link.*[1]

Regarding Hypothetical polling, might it depend on the question? e.g. some polls ask about VI, others about preferred PM or party leader. So, each of the separate articles might need their own Hypothetical section; a solution may be to provide links between them (e.g. "see also ..." under the relevant heading).

• [1] I wrote this before noticing that Preferred PM has been moved to the Approval section. Fair enough. Is that a prelude to creating the separate article?

2.101.242.194 (talk) 21:53, 28 May 2025 (UTC)

There has been broad agreement on creating a separate article for Approval etc, I think the delay is down the concern of kicking off, again, the "edit warring" that has plagued and hamstrung this article. That is before we consider if their present format is actually fit for purpose, to my mind they are bloated, confusing and not helpful to the readers, it would be better to revert to the simpler clearer layout of previous articles. Agreement on that is unlikely. Pugpa2 (talk) 04:25, 29 May 2025 (UTC)

Local Election on timeline

Not sure whether the local election needs adding to the timeline alongside the polls? Get by-election but don’t think local elections in the last parliament were called out like this - also technically local elections happen every week, sometimes dozens of wards - so don’t think it’s worthwhile for its own call-out NewGuy2024 (talk) 08:20, 2 May 2025 (UTC)

Agreed. Bondegezou (talk) 08:46, 2 May 2025 (UTC)

All regular local election days were included in the 2024 article. Also, there's a significant difference between mid-term council by-elections and the wholesale local elections that happen on the first Thursday of every May, where thousands of council seats and several devolved offices are up for grabs, as opposed to (at maximum) a couple dozen council seats that will be re-contested within a couple years on the local election cycle that authority (or seat) is assigned to anyway. Nicholas13t (talk) 09:56, 2 May 2025 (UTC)

Add ITV Cymru Wales/ Yougov Poll 6th May 2025

New poll showing Plaid ahead

https://www.itv.com/news/wales/2025-05-06/labour-support-collapses-as-plaid-cymru-and-reform-battle-it-out-itv-poll 2A02:C7C:803E:C200:F4B2:468D:63DF:4E2F (talk) 14:05, 14 May 2025 (UTC)

This appears to be the same poll from the end of April currently on the page. Only the Westminster figures are used on this page, the voting intention for the Senedd (where PC is ahead) is on a separate page. Kirky03 (talk) 20:54, 14 May 2025 (UTC)

This Page is Practically Useless.

What is the point of this page? It supposed to be an encyclopedia. The entry is miles of raw data which is not summarised or well presented. Which is completely pointless. Jayarava (talk) 14:46, 14 May 2025 (UTC)

this is the same as any other future election polling or record of polling for any election world wide, this page is one of hundreds. They all serve the valid of purpose of outlining the state of polling in countries in a neutral and balanced environment where many countries might not have polling aggregates easily available Pathfinder2023 (talk) 18:28, 14 May 2025 (UTC)

Something is wrong with the chart

Thanks to all of you who edit this page. It's one of my favourite pages on Wikipedia. Today the chart of average voting intention has changed enormously. For example, Reform UK are shown now having an average of 27%. All of the last seven polls have them at 28% or more! Please could you have a look at it and sort it out. Many thanks. Planicosta57 (talk) 17:14, 14 May 2025 (UTC)

It looks like one of the people who updates the graph has been experimenting with a larger span which produces a more smooth graph. This might not be permanent Kirky03 (talk) 17:48, 14 May 2025 (UTC)

Thanks for explaining that. I don't think it makes the chart useful to have such a lag behind the levels that each party has now. Perhaps we could have a choice of two spans, like Politico's poll of polls for the UK (Smooth and Kalman): https://www.politico.eu/europe-poll-of-polls/united-kingdom/ Planicosta57 (talk) 18:43, 14 May 2025 (UTC)

Graph Span Update

I'm starting this up as a new discussion solely because the last one is now quite packed and it is hard to follow and I want to give everyone an even chance to dispute this if they wish.

A couple weeks ago, I proposed a slight increase of the graph span from 0.56 to 0.7 (this will make it slightly smoother but nowhere near as smooth as the span of 4 we originally had). There was some discussion but with opinions seemingly evenly split and unenthusiastic, I did not feel there was enough reason to change the status quo nor to open up a whole new detailed conversation over something we already had a convention for. However, recently, DimensionalFusion has experimented with a span of 1 and another person joined the conversation to argue for a higher span, so in today's update I have used 0.7.

I would like to alert @DimensionalFusion of this as the other person updating the graph for consistencies sake.

If anyone, especially someone who was not involved in previous conversations would like to argue against this, please do so here. Kirky03 (talk) 17:54, 15 May 2025 (UTC)

Personally, I quite like the 0.7 but maybe going slightly lower to 0.65 or so could allow for slightly less gradual trends especially with the rapid movement of polls atm. I also think Plaid Cymru and the SNP should be added onto the chart like with Politico's Poll of Polls, but I imagine that has been discussed before. Additionally, there is the possibility of more parties from Rupert Lowe and the Collective / "Corbyns" party, so I wonder if they would count if the SNP / Plaid don't. Pathfinder2023 (talk) 19:14, 15 May 2025 (UTC)

I think 0.7 makes sense. 1 the other day was too large as it made sharp trends look very gradual, and so was obscuring information. I liked 0.56 because it showed each fluctuation, for example, the brief rise for Labour following Starmer’s Ukraine summit. However on second thoughts, that rise was well within the margin of error, and one might argue was potentially exaggerated by the graph’s smoothing function. Having this as “at least a stall in the decline” seems most appropriate.

0.7 therefore feels like a good halfway point between the extremes. It also results in ~3 polls for each weekly poll, which seems sufficient to verify whether a trend is in or out of the MOE. Sammy zammy1 (talk) 23:15, 15 May 2025 (UTC)

0.56 is a perfectly acceptable span, and i feel is most consistent with the smoothness of previous graphs. Of course this should be lowered over time as need when more points are added. 0.7 feels a bit too smooth too me and misses some nuance. 1 is way to smooth, but i think we've established that. As far as im concerned this is the discretion of the author of the graph, and i have no problem with delegating that responsibility for the forseable future. EnglishPoliticalPerson (talk) 03:37, 16 May 2025 (UTC)

I think this should be left to the discretion of the author of the graph and only changed with prior notification. I would suggest that visually one of things wrong with graph is massive amount of empty space to the right of the data points, don't see the point of that wasted space, yes I understand it represents the maximum period before next election but think data would be clearer if we only had say a 6 month period to the right rather than over 4 years. Pugpa2 (talk) 11:13, 16 May 2025 (UTC)

To be fair, i prefer the full extent graph but thats obviously something that could be discussed. EnglishPoliticalPerson (talk) 14:23, 16 May 2025 (UTC)

Preferred prime minister proposal (RFC)

This discussion has been copied and moved to Talk:Leadership approval opinion polling for the next United Kingdom general election#Preferred prime minister proposal (RFC).--2.101.240.129 (talk) 23:06, 29 June 2025 (UTC)

Electoral Calculus Seat Projection

An Electoral Calculus MRP Seat Projection has been added to the MRP table. This is not based on a poll but an aggregated "poll of polls". This contrasts with the two existing entries for Electoral Calculus / Find Out Now which are based on custom polls, with the fieldwork presumably done by FON. The "current" seat projection on the Electoral Calculus website was updated 26 April "based on opinion polls from 11 Apr 2025 to 25 Apr 2025, sampling 11,432 people". There does not appear to be any previous or archived projection, suggesting that the numbers get updated periodically, replacing the previous version. Should this be removed from the MRP table? As EC don't retain previous versions on their website, it wouldn't really warrant a separate section and the link to the source would effectively be broken after each update. 2.101.242.194 (talk) 19:46, 27 May 2025 (UTC)

The monthly Electoral Calculus based on poll of polls has not been included until now and probably should continue not to be.

Seperate polls by Electoral Calculus that actually involved doing polling themselves should continhe to be included. Kirky03 (talk) 21:29, 27 May 2025 (UTC)

I agree it is simply the poll of polls put through their GE predictor and should not be treated as MRP. User4072 (talk) 23:00, 27 May 2025 (UTC)

Thank you for removing it :) Kirky03 (talk) 09:38, 28 May 2025 (UTC)

Opening New article for Approval Voting

I have attempted to do this but don't seem to have enough permissions to create a new article, I my be missing something so if someone could talk me through it I would be happy to do so, alternatively please open new article and move existing information over. Pugpa2 (talk) 11:30, 7 June 2025 (UTC)

I think I have done the ground work for moving leadership approval to its own article, however unable to complete as apparently there already exists an article with the Title Leadership approval opinion polling for the next United Kingdom general election which redirects back to the sub section in this article, I assume there is a redirect, any how I am stymied. I have created an alternative article with slightly modified title Leadership approval opinion polling for the next UK general election Need advice as to how to get this task done, any advice appreciated. Pugpa2 (talk) 16:21, 25 June 2025 (UTC)

I'm afraid I can't help you with your main problem (others more experienced will come to the rescue, no doubt), but I would like to make a few suggestions, subject to general agreement, of course.

1) Article title: as it now includes Party Approval polling perhaps it could simply be "Approval opinion polling for the next United Kingdom general election". As it also includes Preferred PM polling I was going to suggest something like "Approval and preference opinion polling for the next United Kingdom general election", but that might be a bit unwieldy.

2) Under the Leadership approval sub-heading the explanatory text referring to BPC membership needs amending, as the table now includes Freshwater Strategy whose membership is currently pending. Or perhaps all the text should be at the top ot the article anyway.

3) Hyperlinks: Leadership approval table still has direct links, whereas these have been replaced by references in the Preferred prime minister and Party approval tables.

4) Hypothetical preferred Prime Minister polling: I think this sub-section should also be moved, but not the other Hypothetical scenarios which are for Voting Intention. 2.101.247.160 (talk) 20:58, 25 June 2025 (UTC)

Freshwater

semi joking - but looking at the complete out of kilter Freshwater polls - surely we can’t put these guys alongside reputable pollsters anymore? NewGuy2024 (talk) 15:11, 13 June 2025 (UTC)

Cannot see any issue with Freshwater polls nor anything different from the last time this topic was discussed. On the semi-joke, let's treat it as it is and, please, remember that Wikipedia (and, particularly, article talk pages) is not a forum. Impru20^talk 16:21, 13 June 2025 (UTC)

Agreed. It could be that Freshwater having Reform further ahead is actually correct, and the other pollsters are wrong. It's worth actively trying not to ban pollsters just because they have outliers, focus on the methodology. Eastwood Park and strabane (talk) 12:53, 15 June 2025 (UTC)

We had that discussion and agreement reached about sources etc, however I take your point, my particular interest is in Scotland, here the sub set has RuK ahead of SNP by 15% and having twice as much as Labour, that's not credible. I also note they do not specify what question they asked to get voting intentions, nor in which order Qs were presented. Pugpa2 (talk) 22:52, 13 June 2025 (UTC)

The questions are in the TOC sheet in order. See rows 9 and 10. The Methodology sheet includes the following: "Questions & Presentation: All results are shown in full in this spreadsheet, in order, and using the wording that was put to respondents." 2.101.247.225 (talk) 02:28, 14 June 2025 (UTC)

1) "my particular interest is in Scotland, here the sub set has RuK ahead of SNP by 15% and having twice as much as Labour, that's not credible." Well, Freshwater's Scotland weighted subsample is 60. This would equal to a MoE of about ±12%, clearly insufficient to reach any conclusion on that particular area, because the poll actually does not mean to do so (it is a nationwide poll). It has been generally agreed that sub-national samples of nationwide polls are too small and not reliable enough to determine actual party support in those areas due to the large MoE; that is why those are not included in the Sub-national poll results section. Plus, just to compare, the recent "Find Out Now" poll had Reform in the lead by 4 points in Scotland and also over twice the support of Labour.

2) "I also note they do not specify what question they asked to get voting intentions, nor in which order Qs were presented." They actually do. That is very clearly specified in the sheet that can be found in the source, which has both the order and wording of questions and the methodology description. Impru20^talk 18:30, 14 June 2025 (UTC)

Thank you , I had missed that, at least in part because of the peculiar wording of Question.

I also note that VI Q is the 8th or 9th Q asked and those preceding might be considered as shaping Qs, even the VI Q is potentially leading in its wording, I believe many other pollsters, when asking multiple Qs randomise there order.

Anyhow, it is what it is and there because of agreement, that does not mean that pointing out its idiosyncrasies is not valid.

Accept completely that sub sets are to be taken with extreme caution, however given that there have bee 78 UK polls this year with a Scottish subset and you can put your finger on the only 2 to show a RuK lead is interesting, just as the size of the subset being 61 of 1260, is roughly about half the size that it would need to be to represent Scotland. Pugpa2 (talk) 08:38, 15 June 2025 (UTC)

Did you also miss the bit above about Wikipedia not being a forum? That is, not the place to "point out idiosyncrasies" or discuss the accuracy of polls, or pollster bias to this or that political party, methodoligical differences, etc. except insofar as they are relevant to the one overriding reason for inclusion, that they satisfy Wikipedia criteria for a reliable source. Having said that, the apparent divergence of this poll from others conducted around the same time may be due, amongst other things, to the LTV filter of 9 or 10; this and other adjustments means that the final weighted sample for VI is 697 rather than 1260. The poll methodology is described in the Methodology sheet in the source (in case you missed it). 2.101.245.90 (talk) 20:02, 16 June 2025 (UTC)

Sample Size for YouGov/Times systematically incorrect

This is a YouGov data inconsistency which has made it onto Wikipedia.

YouGov post a CSV dataset betneath their graphical tracker at https://yougov.co.uk/topics/politics/trackers/voting-intention usually some time before they publish the full survey results at pdfs matching the pattern https://d3nkl3psvxxpe9.cloudfront.net/documents/VotingIntention_MRP_250616_w.pdf (change date approrpiately). The former is just a summary CSV for a trend line, the latter contains the full data including sample breakdown by region, weighting, etc.

Here's the problem. They disagree on sample size. I believe the later pdf because the figure is consistent with the breakdowns etc, and can't just be a simple typo without making none of the numbers add up. Also the CSV calls the row "Base" and "Unweighted base": who knows what that means, but it's not the words "sample size", as used in the pdf.

What seems to happen regularly on Wikipedia is that someone comes along and spots the CSV is ready (but not the PDF) and uplaods the data, using the incorrect "sample size" in the CSV. Then someone comes along and replaces the reference with the (more detailed) PDF, as they should. But the sample size stays the same because they don't thoroughly check the reference. The date is often left incorrect, too.

I've tried fixing this twice on different days, but I get trapped in editing conflict hell with people updating and tweaking the table while I'm carefully verifying what's going on against the csv and pdf. Could someone with more patience than me please go through the YouGov surveys and fix the sample sizes and/or survey dates? Thanks!

51.6.11.172 (talk) 10:39, 17 June 2025 (UTC)

I see what you say. I think the inconsistency comes because the CSV shows the sample excluding "Would not vote", "Don't know" and "Refused", whereas the PDF shows the whole sample. If you consider that these amount to 26% of the sample in the linked PDF and that the whole sample is 2,255, you can calculate that 74% of that is 1,668.7, whereas the CSV shows 1,665 (the difference here being easily justified by the rounding of the values, etc.). I don't think this actually poses any problem; for consistency purposes, we should always add the full sample size (as opposed to the re-calculation). Impru20^talk 10:58, 17 June 2025 (UTC)

Savanta London Polls

Not an editor, so could someone please add the following.

1) new poll, recently reported in The Standard: https://www.standard.co.uk/news/politics/labour-london-reform-conservatives-greens-liberal-democrats-poll-b1234351.html and by Election Maps UK (apparently originally on X, but I can't see it there). Westminster Voting Intention (London): Lab: 32% Con: 21%; Ref: 15%; LD: 13%; Green: 13%; Others: 6% (this may be by deduction from 100%; could be checked against tables when they become available). Savanta survey, commissioned by Mile End Institute, Queen Mary University of London; 29 Apr - 21 May 2025; sample = 1,003.

2) previous poll, 30 Oct – 11 Nov 2024, can be amended with additional information. Same client, i.e. Mile End Institute; sample = 1,004. Looking at the tables (see following), "Others" was shown as 4% not 3% (the latter may have been calculated by deduction from 100%). Link to tables: https://www.qmul.ac.uk/mei/media/mei/tgc-media/filesx2fpublications/P042423_Wtd_Tables_20241202_V2_Private.xlsx. Link to tables is on the Mile End Institute website blog at https://www.qmul.ac.uk/mei/news-and-opinion/ dated 4 Dec 2024. Hopefully tables for the latest poll will be posted on the same site some time soon. 2.101.240.3 (talk) 08:29, 26 June 2025 (UTC)

I've now managed to complete all the above (another user did some to start with). If anyone is interested all findings of the Polling London project can be viewed at https://www.qmul.ac.uk/mei/research-projects/polling-london/. For tables scroll down to "Data Tables". 2.101.240.129 (talk) 18:02, 1 July 2025 (UTC)

Split up Preferred Prime Minister section

This discussion has been copied and moved to Talk:Leadership approval opinion polling for the next United Kingdom general election#Split up Preferred Prime Minister section.--2.101.240.129 (talk) 09:04, 30 June 2025 (UTC)

Issues arising and some changes associated with move of approval polling to new article

I have replaced the url following the "next=" parameter in the call to the UK general election opinion polling template with "yes". The parameter simply requires any non-whitespace character to yield a truth-value of "true" ("yes" being used by convention). The url was redundant in any case, as links in the sidebar are expanded from the template code and would not be overriden by changing the value passed to the parameter. In the case of the Leadership approval link (the one for "next" does not require a "la.." parameter), the target page has "United Kingdom" in the title; this page initially redirected back to the relevant section in the main article, then to the main article when the section was removed. I have now changed it to redirect to the corresponding approval article with "UK" in the title. Because of the redirection, it is not treated as a self link, so does not appear as non-clickable bold text in the sidebar. More in hope than expectation, I intend this as a temporary measure, until the redirection page and its existing Talk page can somehow be deleted and the new approval page title can be renamed with "United Kingdom". This will require any existing links to the new "UK" page to be altered. 2.101.240.129 (talk) 08:37, 30 June 2025 (UTC)

That was fast! Well, probably coincidence - I see that Adumbrativus has made those changes to title, etc. literally just a few minutes ago. So please ignore the second half of above. 2.101.240.129 (talk) 08:55, 30 June 2025 (UTC)

I had requested the title change on the 28th on Wikipedia:Requested moves/Technical requests, so I'm guessing an admin finally reached my request. Eastwood Park and strabane (talk) 12:39, 30 June 2025 (UTC)

Thank you, I had been down that rabbit hole and got completely lost, well beyond my skill level Pugpa2 (talk) 11:46, 30 June 2025 (UTC)

Hypothetical or Not

I notice that the creation of a new linked article for Leadership and the moving of all related info from this article to the new one, this has included what had become known at the Hypothetical Polls (until early this year it was known as 'other polling'), however 2 editors appear to be reluctant to allow this to happen and keep reverting to include some of these 'hypothetical polls' to continue to be listed on this article. I take the view that 'hypothetic applies to any and all situations where "Of, relating to, or based on a hypothesis." it does not matter what the hypothesis is or about, by the terms of its existence it is hypothetical and has no place in this polling article. Therefore it should be removed by all means include it in the new article or create a new one for it, but they clearly do not belong in this Polling article. I would welcome the views of other editors Pugpa2 (talk) 11:44, 30 June 2025 (UTC)

You have been noted by three users (not two, I include myself here, as well as @Eastwood Park and strabane and @2.101.240.129) how you were wrong in your interpretation (see [1] [2] [3]), yet you have engaged in a persistent edit warring behaviour (to the point of being at the limits of breaching WP:3RR) over a failure to listen to what you were being told ([4] [5] [6]).

The linked article is for Leadership approval opinion polling for the next United Kingdom general election. Thus, its scope covers leadership polls. The polling you kept removing from here relates to voting intention, so that is not "leadership polling" of any kind and, thus, does not belong to an article that covers leadership polls. It does not matter whether these are "hypothetical" or "non-hypothetical" polling: that is not the scope of that article. In other words: hypothetical leadership polling belongs in Leadership approval opinion polling for the next United Kingdom general election, which covers leadership rating polls (as it happens for Leadership approval opinion polling for the 2024 United Kingdom general election, and example you kept citing for the wrong reasons, as no voting intention polls for the 2024 election are shown there). Hypothetical voting intention polling belongs in this article, which covers voting intention polls. There is absolutely no issue with that. Impru20^talk 12:42, 30 June 2025 (UTC)

May I suggest you actually read my comment and respond, try reading and comprehension. My point is that hypothetical by there very definition have no place in this article, none . But let's see what other editors say Pugpa2 (talk) 12:47, 30 June 2025 (UTC)

Impru20's logic is indeed correct. Hypothetical or not, voting intention pollings belongs here and leadership polling belongs on its own page. Moving hypothetical voting intention polling to the leadership polling page would be very strange. —Profzed! 13:20, 30 June 2025 (UTC)

May I suggest you read your own edit summaries (which I linked)? When you said "2 editors appear to be reluctant to allow this to happen and keep reverting to include some of these 'hypothetical polls' to continue to be listed on this article", that's a blatant misinterpretation of what happened and an attempt to mislead people into thinking that the issue was about 'hypothetical polls' and not on you waging an edit war on something you were clearly wrong. None has reverted you because you thought that 'hypothetical polls' should have their own article, but because you argued that all 'hypothetical polls' should be at Leadership approval opinion polling for the next United Kingdom general election, even if those were not polls on leadership ratings.

Anyway, answering your "new'" point: No, there is no problem in 'hypothetical polls' being in this article unless there are so many of them than that by itself justifies a split. So far, the section is too small (and it is not inherently obvious that there will ever be enough 'hypothetical polls') to justify a stand-alone article as of currently. It could in the future, not now. Impru20^talk 13:30, 30 June 2025 (UTC)

I think the important thing here is that in the last election, all the hypothetical polls were leadership related polls, so when the article was split, they were all moved over. This time there are some that fit better with normal voting intention polls, so some should stay here. Eastwood Park and strabane (talk) 14:20, 30 June 2025 (UTC)

Perhaps a compromise would be to insert a See also hatnote under each Section heading linking to the correponding section in the other article. 2.101.240.129 (talk) 19:56, 30 June 2025 (UTC)

... so I have gone ahead and done this. 2.101.240.129 (talk) 21:34, 30 June 2025 (UTC)

Proposal for new section on BPC membership

As it has been (and in some minds may still be) a contentious issue I would like to propose a new mini sub-section immediately following the lead section. This has been prompted by the recent confirmation of Freshwater Strategy's membership of the BPC and what I see as some outstanding anomalies (see my post in Talk topic Freshwater Strategy above). The first sentence is taken from the current lead section. Other editors may of course change and improve the wording:

British Polling Council membership

Most of the polling companies listed are members of the British Polling Council (BPC) and abide by its disclosure rules. Pollsters which are not members of the BPC may be included if they satisfy Wikipedia reliable source criteria. Pollsters which were not members at the time of polling are italicised in the lists. Freshwater Strategy announced its membership of the BPC on 23 June 2025.

An identical section can also be added to the Leadership approval opinion polling for the next United Kingdom general election article. 92.14.26.105 (talk) 20:58, 9 July 2025 (UTC)

This clearly is a nonsense. Polling companies join and leave the BPC every year, to highlight one company in this manner is not only unnecessary but once again adds to the initial paragraph and complicates, for not very sound reasons.

The readers of this article are perfectly able to follow what is going on and to make sense of what is largely a very simply laid out article, 'here are the VI polling in a table', we have used notes to highlight any anomalies. There was a long discussion about what to include, we don't need to revisit that. Pugpa2 (talk) 09:53, 10 July 2025 (UTC)

"This clearly is a nonsense." That's pretty rich coming from you. I doubt that anyone has posted more "nonsense" on this page, by a country mile. I was merely trying to be helpful to the casual/average/ordinary/occasional/naive reader, to whom it may or may not be immediately obvious "what is going on" (whatever that means), when they encounter typographical variations and obscure footnotes, which may or may not be read and which may or may not apply at any one time. My post was not particularly about Freshwater Strategy, which was mentioned only by way of explaining their change of status (and hence typography, etc.) to aforesaid reader. The rest of the paragraph is about membership (or non-membership) in general and the fact that it has absolutely no bearing whatsoever on whether a poll is included in the article. I am not attempting to "revisit" the issue, because as far as I can see the question of a change of status during the currency of the article and how that may appear on the page, was not discussed. In any case there is no ban on revisiting an issue, especially if its apparent "resolution" leaves the article in an unsatisfactory state with inconsistencies and unforseen consequences. Granted the proposition is a simple one, which was satisfactorily reflected in the article before January, i.e. before Freshwater commenced polling and some on here decided that the British Polling Council was invested with a significance it does not possess. FYI I don't give a flying duck whether or not a company is a member of this rather pointless organisation. I suspect that pollsters seek membership as part of a Brand management strategy, to impress potential clients, thus conferring a rather spurious aura of prestige and respectability. As for "highlighting", it wasn't me that decided to accommodate a couple of exceptions to a rule that never was in the first place, with a messy compromise of italics and footnotes. And if you think it's a lot of fuss about something very minor, I would agree, but I see that another user recently had their edit reverted after italicising FS, being unaware that their status had changed. So it matters to some around here. To my mind it would be better if the BPC were simply not mentioned, with members and non-menbers treated equally, but that idea would never be accepted on this site. Anyway, forget the proposal, I do not intend to pursue it any further. However, please explain precisely what it is that you find so objectionable about these two sentences:

Pollsters that are not members of the BPC may be included if they satisfy Wikipedia reliable source criteria. Pollsters that were not members at the time of polling are italicised in the lists.

These facts are not explained elsewhere and must be inferred from a detailed examination of the table. In contrast, the 33 words above more or less summarise the outcome of the earlier debate. Please also explain why the one remaining non-member, Ashcroft, should not be added to the "Most recent polling ..." table and its 30-day average. This omission is such an obvious inconsistency that I am surprised it has not been rectified before now (and dare I say makes no sense, i.e. "nonsense"). It is difficult to tell from the graph whether non-members were/are included or not, so I would welcome an assurance from someone that they are. 92.14.26.105 (talk) 05:40, 11 July 2025 (UTC)

Sorry, but I am getting tired of this issue coming out over, and over, and over, and over, and over (and over!) again just because some people don't like Freshwater and don't know how to depict it so that it somehow is highlighted as somehow different from the others. There have been countless discussions and a specific RfC on the BPC issue back when Freshwater wasn't a member of BPC, yet now that it's a member it somehow is still a problem for some people. Just drop the stick, everyone. Impru20^talk 10:21, 10 July 2025 (UTC)

Please lie down and have a rest. I do not have a stick, but some appear to have grasped the wrong end of one. The best way of ensuring the topic is put to bed is to deal with its inconsistencies, which arise out of the earlier debate. I neither like nor dislike Freshwater Strategy, or any other pollster. I can see very well how it is depicted (and "highlighted"). One week with italics and footnote, the next week not. Of course it has been depicted as "different from the others", that was the outcome of the fractious debate on the matter and was down to those who participated in it, which I did not. Indeed, that was the whole point of my proposal, which applied to (non-)membership generally and which simply sought to clarify the situation for any unsuspecting reader, perhaps less experienced than you. Oh how I wish that all pollsters were treated equally on this site - I believe they were before January, until someone kicked off an unnecessary and heated argument about membership of some relatively inconsequential organisation. Some of your remarks would be better directed to Pugpa2, who as recently as the middle of June was questioning the credibility and veracity of a Freshwater poll, referring to their supposed "leading questions" with "peculiar wording" and "idiosyncrasies" and implying that the poll was politically biased, all of which was based on a completely mistaken reading of the data (and quite likely Pugpa2's own prejudices). Anyway, forget Freshwater and forget the additional sub-heading/section. I will put to you the same question that I put to Pugpa2 above: please do explain precisely what objections you have to these two sentences:

Pollsters that are not members of the BPC may be included if they satisfy Wikipedia reliable source criteria. Pollsters that were not members at the time of polling are italicised in the lists. (33 words).

As I said above, this is not explained elsewhere and must be inferred from the table, while these two sentences more or less summarise the outcome of the earlier debate. While I do not intend to pursue these proposals any further, I would like to know the reasons for their rejection. Finally, perhaps you would like to address a couple of the issues I raised above, e.g. the omission of non-BPC members from the "Most recent polling ..." table and some assurance that all their polls are included in the graph? 92.14.26.105 (talk) 06:06, 11 July 2025 (UTC)

There is no inconsistency. People complained about Freshwater not being a BPC member, a RfC was carried out on whether non-BPC members should be included and consensus was overwhelming in favour of including non-BPC members as long as they were supported by reliable sources (which is what is done everywhere else in Wikipedia). That alone justified Freshwater's presence in the tables. Yes, Pugpa2 complained some times (and I did indeed rebuke him) Thing is: Freshwater is now a BPC member, and yet that still poses a problem for some people, who just will not accept Freshwater's presence because they don't like it. Please get along with it, move on and stop wasting everyone's time with such a perennial debate.

On some specifics:

• Pollsters that are not members of the BPC may be included if they satisfy Wikipedia reliable source criteria. 1) The very first line in the article states that "Most of the polling companies listed are members of the British Polling Council (BPC) and abide by its disclosure rules", so this is already accounted for (though BPC membership is never acknowledged as a requisite for inclusion, because it isn't); and 2) Everything in Wikipedia must satisfy Wikipedia reliable source criteria. Adding a whole sentence or paragraph to highlight what is one of WP's core policies is entirely redundant. We don't do that in thousands of other articles, and the article is already large enough to keep adding unnecessary information.
• On the "Most recent polling ..." thing, I am against that section altogether since it's pointless, only serves to duplicate information and may result (as you hint) in some inconsistencies if it's not kept up to date alongside the main tables. I would be entirely fine if we just get rid of that section. Impru20^talk 07:52, 11 July 2025 (UTC)

@Impru20 entirely agree, the discussion has been had and consensus reached, I say that as one who argued against there inclusion but was won over by the points made ( and the volume of support) there is no need for further discussion or explanation. I also agree that "Most recent Polling..." is entirely unnecessary and would be happy to see its exclusion, perhaps it is time to revisit that topic as I think it was basically kicked into the long grass during a very difficult period for the article. How do we move that forward? Pugpa2 (talk) 09:32, 11 July 2025 (UTC)

References

JC party

Should we include a potential Jeremy Corbyn-led party in the opinion polls? I think it's well above the threshold for inclusion in the hypotheticals at least, getting 15% and tying with Labour. If so, at what point should it be included?

• When pollsters stop polling it as a hypothetical and start including it as a regular party? (done previously)
• When the party officially starts?

DimensionalFusion (talk · she/her) 09:54, 15 July 2025 (UTC)

Hello @DimensionalFusion !

That scenario has already been implemented here, in the hypothetical polls section. I guess it will stay there until Corbyn officially launches his party, indeed.

Have a nice day ! Tacomniscient (talk) 11:06, 15 July 2025 (UTC)

Agree with the above comment - the answer should be the latter, in my opinion. It's unclear whether the party in question actually exists yet, so I think we need to wait for that. It doesn't actually have a name yet as far as I know. Eastwood Park and strabane (talk) 11:07, 15 July 2025 (UTC)

I'll reply to my own comment to mention that when this party does materialise I think it should be placed between the Greens and the SNP on the table. All the major parties, then the newly founded one, then the regional parties. Eastwood Park and strabane (talk) 11:09, 15 July 2025 (UTC)

@Eastwood Park and strabane

I'm not against, but it makes me remind that the sorting of the parties is different from the wiki page dedicated to the 2019 and 2024 UK pollings.

In these pages, the order was : major parties, then regional ones, then smaller parties, then new ones.

So, I wonder if there's a reason for such a change, and if not, shouldn't we go back to the previous way of sorting ?

Have a nice day ! Tacomniscient (talk) 11:25, 15 July 2025 (UTC)

I believe the 2024 election's page might have been simply based on vote share at 2019's election, but 2019's appears to be in seat count order, then new parties by order of their establishment, except UKIP, which is placed after Brexit for some reason. Very random in my opinion. Edit: Actually it makes even less sense than this as the SNP had more seats than the Liberal Democrats. Eastwood Park and strabane (talk) 12:32, 15 July 2025 (UTC)

I think they would need to be at the end as a party that "received" 0% in the last election Kirky03 (talk) 19:23, 15 July 2025 (UTC)

I think we had a long discussion about this earlier in the year, included in that discussion were various options including National, Regional, number of MPs and current polling, my recollection was that we ordered the parties by their vote share at last election, this seems a sensible, transparent and easy way to do it.. Any new ones would be added after those, assuming of course that they are prompted for on a regular basis by pollsters. Pugpa2 (talk) 16:19, 20 July 2025 (UTC)

Polls missing from graph?

Think I've noticed something odd with the graph, it seems some polls aren't included? Find out now 10th July is easy to see missing because of it's outlayer Reform value. It seems the update to the graph on the 11th July added the Opinium 11th July poll, but not the Find out now 10th July, the 14th July graph update also looks weird to me, maybe my eyese are going weird loking at dots, but it looks like it's added values of ref/lab as 28/24...possibly accidentally combining the two 14th July polls?

90.194.100.223 (talk) 09:23, 17 July 2025 (UTC)

@DimensionalFusion —Profzed! 11:37, 17 July 2025 (UTC)

Hi, the data used to generate the graph is available at commons:User:DimensionalFusion/Opinion poll chart DimensionalFusion (talk · she/her) 12:27, 17 July 2025 (UTC)

Poll ages graph

Given the government has announced votes at 16, should 16+ polls start to be included in the graph instead of 18+ polls? DimensionalFusion (talk · she/her) 16:33, 18 July 2025 (UTC)

The difference between the two will be minuscule so I'd say it's fine including either (or if the pollster publishes both like Focaldata did, use whichever figure is the headline figure, which is the 16+ one for Focaldata). Chessrat ^{(talk, contributions)} 20:25, 18 July 2025 (UTC)

I think whichever are the headline figures should be published. I think including both the published polls or xtabs in-/excluding 16-17-year-olds for each poll would be unproductive, especially as the difference will be 0pts for most parties. Now, for marking whether polls include an 18+ or 16+ sample will depend on how quickly pollsters alter to include this age group - personally I think it would be best to mark polls with an 18+ only sample following 17 Jul 25 with a note if most pollsters switch to 16+ soon. If almost every poll is 18+ for the remainder of 2025, I don't think there should be an additional column, but if several pollsters holdout until it becomes law, then vice versa. Additionally, I think there is an argument to be had that polls that *only* query 16-17-year-olds' voting intention should be included separately in the sub-national polling section. Nicholas13t (talk) 22:17, 18 July 2025 (UTC)

I personally like the current table, but I think I agree with the approach of using the headline figures in the graph. Pollsters will begin to shift over soon I suspect. I imagine the ages column will not be necessary for 2026. Eastwood Park and strabane (talk) 22:29, 18 July 2025 (UTC)

At this moment 16/17 year olds do not have a vote, yes the government has announced an intention but at the present it remains just that , an intention. At best at this point in time it is a hypothetical question and should be treated as such and not included in the main table, to do so mixes established polling with a possible change that may or may not come to pass.

By all means lets open a 16/17 year olds in the hypothetical section at least until such time as the legislation is passed. Pugpa2 (talk) 22:59, 18 July 2025 (UTC)

Such a major change to the article main table should seek consensus before making such a large scale change, I would intend to remove these changes until we have the discussion Pugpa2 (talk) 23:04, 18 July 2025 (UTC)

I agree Kirky03 (talk) 23:18, 18 July 2025 (UTC)

Focaldata published the 16+ figures as their main headline poll (https://x.com/focaldataHQ/status/1946226852042846617) rather than as a hypothetical, with the 18+ figures visible only in the actual crosstables, so it would be completely misleading to only include the 18+ data in the table. Chessrat ^{(talk, contributions)} 23:48, 18 July 2025 (UTC)

It doesn’t matter if the main headline was 16/17 included, it is still hypothetical because they aren't yet franchised. Kirky03 (talk) 21:22, 19 July 2025 (UTC)

@Chessrat Here https://www.focaldata.com/blog/westminster-voting-intention-reform-leads-labour-but-left-set-to-benefit-from-votes-at-16 they publish both side-by-side. The 18+ data is presented with the same degree of confidence, so while it is a subset of the 16+ data it is also a valid opinion poll in its own right Saxmund (talk) 22:26, 20 July 2025 (UTC)

16/17 should be included in hypothetical until legislation is actually passed just like the hypothetical polling about if Kemi was leader has remained in hypothetical even though she has since become leader.

At the moment, all legislation has not passed and is subject to a vote, etc so is purely hypothetical Kirky03 (talk) 23:17, 18 July 2025 (UTC)

Additionally, I think that the "Government announces intention to reduce voting age" segment should be removed from the main table and "Voting age lowered" should be included when/if the legislation is passed which it probably will eventually.

Until then, intention to change the voting system is not significant enough by itself in my opinion. Kirky03 (talk) 23:20, 18 July 2025 (UTC)

I think if it's decided to keep the ages shown then the note explaining why we're suddenly including ages should stay. If it's decided to move the 16-17 year old voters to hypothetical then we should remove the note also. Personally I'd lean towards keeping things the way they are, with the ages shown. At least one pollster has already switched to listing 16+ as a headline figure. Eastwood Park and strabane (talk) 00:18, 19 July 2025 (UTC)

@DimensionalFusion A 16+ plus poll is hypothetical, at the moment the voting age is 18 and there is no certainty it will be changed. The usual opinion poll question is "how would you vote if there was a General Election tomorrow?" So we should continue to use 18+ polls until the voting age changes. It is not correct to say we only use "headline figures" as the figures in the tables match the headline and it is often necessary to look at the tables to get minor party results. Saxmund (talk) 22:00, 20 July 2025 (UTC)

I've added a hypothetical section for 16/17 year olds. A lot of the discussion here has not seemed to be arguing against the inclusion of the 16/17 poll but rather that it is currently hypothetical and considering we have a hypothetical section, I've just gone for it. Hopefully this can satisfy everyone as it ensures the poll is included using the headline figures but does not create issues with the consistency of the main table.

I'm just going to notify everyone who has been involved here so they can discuss if they have any complaints.

@DimensionalFusion @Chessrat @Nicholas13t @Eastwood Park and strabane @Pugpa2 @Saxmund Kirky03 (talk) 12:15, 21 July 2025 (UTC)

That works for now. I think if more pollsters start regularly reporting 16+ as headline figures we will have to revisit it. Chessrat ^{(talk, contributions)} 14:08, 24 July 2025 (UTC)

Everything gets revisited when circumstances change, if votes for 16/17 years becomes law, then that would become the franchise we would record in main tables. Pugpa2 (talk) 14:35, 24 July 2025 (UTC)

Headline voting intention data

Several editors have been removing the headline voting intention of the recent Focaldata poll (1,307 voters; Reform on 26%) from the main table of polls. I would hope we would all agree that regardless of your opinion on the table layout, not reporting the headline voting intention in the table of polls at all shouldn't happen. Chessrat ^{(talk, contributions)} 12:59, 20 July 2025 (UTC)

It is the headline figure of a poll that is not voting intention of the current franchise and therefore should not be included on this page.

If you believe non-headline figures shouldn't be included then fair enough and maybe the entire poll should be removed but we cannot include a poll that is not voting intention of the current electorate in my opinion and seemingly that's the opinion of the people reverting your changes. Kirky03 (talk) 14:36, 20 July 2025 (UTC)

It is a hypothetical for 16/17 year olds that do not have a vote at present, that part has no place in this table. Please read comments that have explained this in detail. Let's not get back to the state this article experienced early in the year where topics kept popping up on very closely related matters. Pugpa2 (talk) 14:47, 20 July 2025 (UTC)

There is no point in the history of any of these articles that the main table has failed to include the headline voting intention figure and only included a subsample.

If you want the main table to only include subsamples for some polls rather than headline voting intention, the onus is on you to gain consensus for your proposed change before enacting it- or at the very least leave out the poll entirely until the discussion is over. Chessrat ^{(talk, contributions)} 17:49, 20 July 2025 (UTC)

In that case, the poll should be removed entirely. The 16/17 figure is hypothetical and so is the headline figure of an irrelevant poll.

Would you be okay with the poll being removed entirely while we discuss and work out what to do? Kirky03 (talk) 18:15, 20 July 2025 (UTC)

Including the 16/17 year olds would introduce a hypothetical poll to the main table, now that is something that has never happened. To describe this a subset is in danger of being disingenuous. However if you that strongly then by all means remove the poll entirely until discussion is over. Pugpa2 (talk) 18:16, 20 July 2025 (UTC)

Add creation of Your Party to the main table

I've added the creation of Your Party to the main table, however an IP editor (80.3.181.202) has repeatedly reverted this without commenting why. I wanted to gain consensus here that this event should be in the table, rather than edit warring. Eastwood Park and strabane (talk) 11:06, 25 July 2025 (UTC)

At the moment those polls are hypotheticals, the question given is not standard voting intention but "how would you vote if a new left corbyn party existed". We should wait until normal VI polls contain the Your Party / the chosen name after the conference in every poll. Pathfinder2023 (talk) 12:14, 25 July 2025 (UTC)

To be pedantic, all polls ask hypothetical questions, this one differs because one of the options was a hypothetical party that still does not officially exist so I altered the wording slightly. Orange sticker (talk) 13:08, 25 July 2025 (UTC)

It feels slightly premature currently but I'm sure their time will come. There's a handful of pollsters who poll weekly and over the next few weeks, we may see a couple of those start asking about YP at which point they will probably have hit the same level as Plaid and should then be added.

Until multiple pollsters are prompting about them in their regular polling, I'd say we just need to wait. Kirky03 (talk) 13:48, 25 July 2025 (UTC)

Hypothetical polls shouldn't be included in the main table. I would say the party should only be included once multiple pollsters start reporting its level of supports in their regular polls. This isn't the case yet. It would violate NPOV to include the party while it is not showing up in regular polls. Gust Justice (talk) 16:55, 26 July 2025 (UTC)

I had actually intended this to be a discussion just about specifically adding the event of the creation of the party to the table, rather than adding a new column, but this discussion is clear. I'll wait to add it until the party is being prompted for by a significant portion of the pollsters. Eastwood Park and strabane (talk) 19:24, 26 July 2025 (UTC)

Oh, I am very sorry, I misunderstood. I'm fairly neutral on this, I'm gonna put it as a big maybe? It might be better to put it when the party is officially formed but maybe now is good. Kirky03 (talk) 22:00, 26 July 2025 (UTC)

Personally my opinion has actually changed on this I think. Until the new party has a column on the table I don't think including its creation as an event on the table makes sense. I think these "events" should be limited to parties on the table Eastwood Park and strabane (talk) 10:49, 27 July 2025 (UTC)

Definitely not. There is a (factually, strict) consensus on which events merit to get included in the tables, which is basically limited to relevant elections/by-elections and leadership changes. This one would fail to meet either of these conditions, so it is a big no. Opinion polling tables are not listings or timelines of events. Impru20^talk 10:57, 27 July 2025 (UTC)

If the party begins being prompted for in larger numbers, and gets its own column, I think it would only make sense to put the founding on the list. It's like a leadership election but much more significant. Eastwood Park and strabane (talk) 11:06, 27 July 2025 (UTC)

No, you are not getting the point: opinion polling tables are not timelines of events. We do not add events based on whether we consider them as significant or not, because that is not the table's job to do. We only have some token events as a somewhat balance position between those who wished to add any events (because they thought that these had an impact in opinion polls) and those who just wished to get rid of all of them. By displaying events, we are deliberately telling readers than that event is somehow important for opinion polling, which is basically WP:OR (specially considering that each person's perception of what is and what is not important may vary, or that opinion polling may be influenced by a chain of events going through a long period of time, or that such changes may be temporary and thus ultimately unnoteworthy, etc.). Concurrently, by cherry-picking ourselves which events have an impact and which haven't, we would be breaching WP:NPOV: so, either we add all significant events (which would make the table monstrously unmanageable) or we add none.

If you wish to consult events, you already have a line just above the table explaining that "For political events during the year, see 2025 in United Kingdom politics and government and 2025 in the United Kingdom." That's the place for adding such events. If the party gets reported in headline numbers, just add a column for it with the figures. An event row does not make any sense at all. Impru20^talk 12:41, 27 July 2025 (UTC)

You're completely right that deciding which events are significant and which aren't is subjective and thus is bordering on original research and is at extreme risk of a lack of neutrality. This holds true for the riots last summer which (from memory) had been added at one point and was subsequently removed.

However, in the case of the creation of a new party, the idea that that may have an impact on voting intention polling is, dare I say, common sense rather than a matter of opinion.

Think of it not as a political event but a change in leadership. It is the change of leadership for the new party from no one to ZS/JC in which case it would be eligible. Kirky03 (talk) 14:05, 27 July 2025 (UTC)

It is not a change in leadership, so let's not confuse people with that. Also, I do not support including any events at all, but leadership changes were the minimum agreed when consensus was reached; party creation wasn't.

The addition of a column for the party once it gets mentioned in headline figures would already signal its creation. Creating a specific event row for it is meaningless, and will also be confusing since readers will get the hint that every new party that gets created should merit a mention. Impru20^talk 15:14, 27 July 2025 (UTC)

I would still argue party creation is a much more significant "event" to include on the table, compared to individual leadership changes within parties. It's much more significant context as to why a party is being included now. If a party leadership change is included, of course a party being founded should be, it's a kind of leadership change but even more significant. Eastwood Park and strabane (talk) 16:21, 27 July 2025 (UTC)

You have the column once the party gets into headline figures (and they are significant enough). What's the game changer feature that adding an event row for the party would add to the table? Plus, this assumes that this new party's results will be relevant well before it has fought any election.

"If a party leadership change is included, of course a party being founded should be" Well, this is not an argument for me: under current consensus, that is one of the few token events that are allowed into the tables because some people liked them, but I will gladly see leadership change events removed altogether under the same grounds. So, if that is what you are suggesting, you must know that you will find in me your greatest ally if you push for it. Impru20^talk 17:26, 27 July 2025 (UTC)

I agree. There's perhaps not much of a precedent for this being included but it's not exactly like it happens every day. We'll have to judge it when it happens like you said. If most pollsters are prompting for them and they're polling decently high, it would be a useful bit of context that makes the table more readable.

Guidelines and precedent are useful and shouldn't be ignored but also should not be followed so strictly that they worsen the quality of the article.

Let's circle back to this when/if it reaches a point where it may merit inclusion. Kirky03 (talk) 14:00, 27 July 2025 (UTC)

I would say maybe do it if and when the party shows up in polls from multiple pollsters. This isn't without precendent when you look at other opinion polls pages. But at this point it's premature to do so. Gust Justice (talk) 14:37, 27 July 2025 (UTC)

As Impru20 has said, these tables are not some kind of list of important events. They are a table of poll results. When(/if) Your Party, or whatever name they go with, polls at a significant level, it should be added as a column.

When the Sahra Wagenknecht Alliance (BSW) was created in Germany, it was not added as an event at Opinion polling for the 2025 German federal election, but when they started polling significantly, they were added as a column. Bondegezou (talk) 21:07, 27 July 2025 (UTC)

If leadership changes are being included (which regardless of people's individual opinions on it, they are being) then the creation of a party significant enough to get its own column should also be. These are definitely in the same vein. Kirky03 (talk) 07:09, 28 July 2025 (UTC)

Well, established consensus is not "people's individual opinions". Leadership changes are included because it was part of a broader compromise solution, creation of parties wasn't (particularly when the addition of a column for that party would already account for it). As previously said, if you want us to revisit that consensus and get rid of leadership changes and other event rows as well, I am all for it (and I guess Bondegezou will, too). And surely, if we are going to use compromise solutions as leverage for enforcing even more event rows, then we may have to revisit that consensus indeed. Impru20^talk 09:51, 28 July 2025 (UTC)

The German election page doesn't have any leadership changes either though, I think if we have party leadership changes included, then it only makes sense we should have party creation included (for parties that have their own column), because I'd argue a party being created is actually more significant than a leadership change. I'd personally oppose removing the events altogether, I think the current balance (although with party creation included) is a good balance.

I'd note that the creation of Change UK and the Brexit Party, in 2019, were included as separate events in the Opinion polling for the 2019 United Kingdom general election article, so there is some precedent for this. Eastwood Park and strabane (talk) 09:57, 28 July 2025 (UTC)

I think the Opinion polling for the 2019 United Kingdom general election article has gone too far and would favour stripping those and several other rows from it, but that's a broader discussion. I would also note that the row for the Brexit Party is for when they were registered with the Electoral Commission, which Your Party has not yet done. Picking a date for Your Party is difficult (which in itself argues against putting a row in). Do we go with Sultana's initial announcement earlier in the month (after which there was hypothetical polling)? Do we go with the Your Party website launch more recently? Do we go with when they register with the Electoral Commission or launch the final name of the party and other structures? There isn't a single date, so leave it out the table and expand the text that accompanies the table.

I would also just like to affirm that Impru20's guess is right! I would strip party leadership changes from this and earlier tables too. Most opinion polling articles don't have them.

Maybe we need a broader discussion at the Elections WikiProject? Bondegezou (talk) 10:09, 28 July 2025 (UTC)

I have started a broader discussion at Wikipedia_talk:WikiProject_Elections_and_Referendums#Should_there_be_additional_rows_about_events_in_polling_tables? Input there welcome. Bondegezou (talk) 10:36, 28 July 2025 (UTC)

I agree with your points about dates which is why this is a bit premature as the date that would make most sense is date of registration which is yet to occur.

I do also sympathise because you both are against even leadership changes and so are understandably resistant to any additional events but I must stress that including this would not be broadening the scope of what is allowable as this easily falls under similar logic for leadership changes. It is a change in the political makeup of the country.

I also understand Germany have not included such events but there has been such inclusion with UK polling articles which is more relevant here as different countries will have subtle differences in system which will affect their articles priorities.

Anyway, I will not go on any longer as I'll just end up in circles but that's my thinking. Kirky03 (talk) 11:18, 28 July 2025 (UTC)

I have actually checked this "precedent" and, as far as I can see:

1) The event row for the Brexit Party was added without any actual consensus ([7], nor was anywhere explained how was this compliant with the established consensus.

2) The Independents/Change UK row was added by a user claiming a "talk page consensus" ([8]), but this was because the user mistook a consensus for the column with a consensus for the row.

But the 2019 case what does actually prove is why adding event rows for new parties is not a good idea: the Brexit Party ended up with a bare 2% in the election, and Change UK did not event get polled into the election itself despite the previous claim that it was a super-important event (in fact, it got a bare 0.03%). Looks more like these rows should be removed, rather than being used as an example of anything. Impru20^talk 13:16, 28 July 2025 (UTC)

Too soon to add them into the main table - the party doesn't officially exist right now so it's more hypothetical for what might happen. — Czello ^(music) 12:43, 28 July 2025 (UTC)

Lets keep it simple, if a party begins to be regularly prompted for by pollsters then we include it in table, we do not and should not be following blindly the news cycle narrative. Since beginning of May 2025 10 new parties have been registered with the Electoral Commission, the creation of a new party is not a big issue, dozens come into existence every year, should we be capturing that information? Of course not so why would we treat this new proposed party any diffrently?

KISS Keep It Simple and Straightforward in our case driven solely by polling data provide by pollsters in their data tables. Pugpa2 (talk) 11:30, 30 July 2025 (UTC)

Local Regression Graphical Presentation Not Providing An Accurate Up To Date Representation of the Current Polls

The graph currently shows a notable down tick in the polling fortunes of Reform to the benefit of both Labour and the Tories whereas recent polls over the past two weeks indicate the complete opposite and whilst what is shown in the graph might have been true 2 weeks ago it is now no longer the case. For example the latest 'Find Out Now' poll (which came closest to predicting the national projection at the time of the May elections and therefore is arguably the current' gold standard') has just reported Reform's equal highest polling figure and biggest lead and Labour's and Tories lowest polling figure for 5 years. If you run a simple 7 poll running average for each party you will see that Reform are currently on an average of 29%, Labour an average of 22% and the Tories 18% which net over the last couple of weeks implies changes of Reform +2 Labour -2 Conservative No Change but the local regression graph here suggests something completely different is happening.

That the change in direction of the polls has not yet been picked up by the graph tells me that at least one of the controlling variables (eg timespan of the samples under consideration) is not sensitive enough to pick up the immediacy of voter opinion. Anyway I would appreciate the thoughts of those involved on how to make the graph more accurate because its clear the graph is not representative given the contradiction the graph currently offers. 90.215.252.146 (talk) 17:13, 25 July 2025 (UTC)

The graph will always have a bit of a delay to it just due to it being a rolling average of sorts as you've mentioned. If the recent trend continues another week, what you'll probably find is it undoes a bit of the drop it currently reports and then straightens out slightly.

Honestly, I've felt that it can be too sensitive and sometimes reacts to some very short term margin of error changes but you're right that at the moment, it seems to be a bit slow.

That said, while the average has gone up, some pollsters haven't shown much of an improvement and polling is still somewhat lower for Reform than it's peak hence the drop. Either way, it's a waiting game, the graph won't take more than a month to react to any changes which was a problem when the page was first created (the "span" which dictates how reactive it is was 4 at first and has been reduced dramatically to 0.56 to find a nicer balance). Kirky03 (talk) 19:16, 25 July 2025 (UTC)

These sorts of smoothed curves often struggle at the ends. One approach might be to not run the curve to the end. The line could stop a week or fortnight before the latest poll. That doesn't solve the problem, but may prevent misinterpretation. Bondegezou (talk) 22:07, 25 July 2025 (UTC)

Your PARTY - Jeremy Corbyn

It is polling already at 10%, we need an updated graph, and added to the page man. Gmat605 (talk) 12:39, 28 July 2025 (UTC)

Hypothetical polling. The party doesn't officially exist yet. — Czello ^(music) 12:41, 28 July 2025 (UTC)

It has been polled in about 2 or 3 hypothetical pollings and is not yet at the point where it merits its own column. Kirky03 (talk) 15:02, 28 July 2025 (UTC)

I'd wait until there's a poll released including YP at the very earliest. Eastwood Park and strabane (talk) 16:11, 28 July 2025 (UTC)

Yes, there's no point adding it to the main table until they appear in a poll that goes in the main table. And we should probably wait until we've had a few with them in. Bondegezou (talk) 20:49, 29 July 2025 (UTC)

KISS Keep It Simple and Straightforward in our case driven solely by polling data provide by pollsters in their data tables.

1 This party does not yet exist, according to the Electoral Commission, until it does it has no place in the main article.

2 Only when it is then prompted on a regular basis by pollsters should we include it in the main article Pugpa2 (talk) 11:41, 30 July 2025 (UTC)