Talk:Opinion polling for the 2015 United Kingdom general election/Archives/2014/November


Scottish only opinion polls

Where can this information be placed?

This would result in some pretty major changes north of the border. --Crazyseiko (talk) 21:41, 30 October 2014 (UTC)

Already covered at Opinion_polling_for_the_next_United_Kingdom_general_election_in_individual_constituencies#Scotland. Bondegezou (talk) 22:13, 30 October 2014 (UTC)
The problem is that Scotland, Northern Ireland and Wales aren't individual constituencies. Even if they're constituent countries, they aren't constituencies. The significance of the SNP is about as significant as the rise of UKIP, especially since the SNP is likely to win more seats. It should be reflected on this page, relevant to next year's election. 123.2.85.195 (talk) 00:07, 1 November 2014 (UTC)
We've had some discussion about whether the sub-national level polling should be on the Opinion_polling_for_the_next_United_Kingdom_general_election_in_individual_constituencies article or somewhere else, and indeed whether we can re-name that article to something better. That article's Talk page has details: please do weigh in to the discussion there. One possibility is to have a new article for nation-level polling, or indeed for Scotland, Wales, England and Northern Ireland each separately. To an extent, I'm less bothered which article the data is in, as long as it is clearly signposted where it is.
The SNP is included in the detailed polling table in this article, for GB-wide polls. I'm in favour of dumping the first table and just sticking with the detailed table, including the Greens, SNP/PC and BNP. Bondegezou (talk) 09:13, 1 November 2014 (UTC)

Scotland

There are many Scottish Westminster polls (and a few Welsh ones too). Should they be included here (given that the NI's ones are), in the constituency polls page or in a new one? I just find it strange that at the moment they do appear to appear anywhere on Wikipedia... 193.62.42.156 (talk) 14:46, 3 October 2014 (UTC)

I think they should certainly appear somewhere, but like you, I've never been certain where! I suggest we start with a subsection here and if it gets too big, we can create a separate article. I think it's also important to exclude small subsample results of national polls: we should only cover proper polls of Scotland (or Wales). Bondegezou (talk) 15:06, 3 October 2014 (UTC)
Aye to that. Though maybe a separate article would make sense, for NI, Scots, Welsh, English only (there must have been one?). This article is pretty scrolly already. It could be a good idea to include 3-way results there too (westminster, assembly const, assembly region) Iliekinfo (talk) 16:29, 4 October 2014 (UTC)
Absolutely agree that they should be somewhere - I'd tend to agree with [User:Iliekinfo|Iliekinfo] that a separate article for NI, Scots, Welsh, London-wide etc would be good (would polls of groups of marginals also go there?). Agree also with the 3-way results. One more question of the content - there would need to be a decision (again) about which parties to include. I can't see myself contributing, though, as I'm working on the gaps in the incomplete "Detailed poll results" section whenever I have time to add extra content (as against keep up-to-date). DrArsenal 87.112.138.68 (talk) 16:49, 7 October 2014 (UTC)
I see there is now a "Constituent Country Polls" section on the slightly misnamed "Opinion polling for the next United Kingdom general election in individual constituencies" page. Good luck to those of you who want to add such polls. I will find them helpful. Should there be a link from this page? DrArsenal 46.208.137.165 (talk) 15:17, 15 October 2014 (UTC)
Now relocated to this page DrArsenal 46.208.137.165 (talk) 21:44, 5 November 2014 (UTC)

Scotland, Wales and the BNP?

There's been significant changes on this article, and I think there needs more to be addressed. Currently we have a system where the main poll results are displayed in the featured table, which now includes the Green Party. I agree with that, but it only leaves the SNP, Plaid Cymru and the BNP (as well as the larger parties) in the detailed poll results. There are many inherent problems with this system, and I think the solution is quite obvious.

  • The detailed poll results do not address the significance of polling results for the nationalist parties. A doubling of Commons' seats for SNP or PC could come about from what appears to be a very small percentage change.
  • I don't know why BNP is listed here, and it shouldn't be at all. It's polling far too low for us to consider it as a legitimate party in opinion polling and should be considered with the rest of the "others".
  • Most of the data in the second poll table repeats data in the first table and provides no new information.

The only conclusion I can make from this is that there should be two separate sections for opinion polling.

  • The first: much as the current table is now for the whole of Great Britain. Tory, Labour, UKIP, Lib Dems, Green and Others, and the lead. The accompanying graph should probably also include a line for Green and Others.
    • There could be a combined Nationalist vote included in the main table, which would remove most of the Others, simply because it is at the same Lib Dem/Green level.
  • The second: separate polling results for Scotland, Wales and Northern Ireland which show purely results and polling data relevant to the next general election. This shows adequately the position of nationalist parties and the changes of all parties in those specific regions.

The current system, especially now that the Green Party has been added to the main table, is obsolete. The BNP is irrelevant and the data for SNP and PC is useless. We need these changes to accurately convey what the opinion polling is saying. 123.2.85.195 (talk) 02:08, 1 November 2014 (UTC)

The BNP are included because the polling companies (generally) include them. Following WP:RS, that is the most obvious approach to take -- we report what the (reliable) polling companies report. For us editors to make a decision to include some parties reported by the polling companies but not others, it seems to me, would be in violation of basic Wikipedia principles. Bondegezou (talk) 09:16, 1 November 2014 (UTC)
WP:RS is a policy about identifying reliable sources, and we've already acknowledged that the main polling companies are reliable. I'm not sure how following WP:RS in this instance would improve the article. There are many scientific articles, and when they source a study they do not have to present all the information in the study. It is irrelevant and useless, per WP:RELE. They poll less than 1%, and reliable sources routinely ignore the BNP but report on UKIP, the nationalists and the Green Party. It would be a violation of Wikipedia principles not for editors to remove useless information. 123.2.85.195 (talk) 12:45, 1 November 2014 (UTC)
Many reliable sources report Con, Lab, UKIP and LibDem and don't cover the Greens or SNP/PC when discussing polls. Indeed, most reporting of GB-wide polling doesn't mention SNP/PC. If we go by how polls are reported, as far as I can see, we should stick to Con, Lab, UKIP, LD and maybe Green. If we go by what the polling companies report themselves, then we should reflect what they say with the same emphasis they give, as per WP:UNDUE. The polling companies' reports give the same weight to the BNP as they do to the Greens or SNP/PC. (WP:RELE is an essay, not policy nor a guideline, and I don't see how it applies here.)
At the moment, (thankfully) the BNP vote has collapsed and the party is in disarray. However, this article covers all polling since the last general election, and the BNP were doing better back in 2010: they got more votes than the Greens, SNP or PC. An article on polling across this period should clearly cover the BNP if it covers parties who got fewer votes in an election at one end of that period. Wikipedia articles shouldn't get too caught up in short-term, recent developments. Bondegezou (talk) 18:19, 1 November 2014 (UTC)
Reliable sources do not report on the BNP but they do report on Conservative, Labour, Liberal Democrat, UKIP and Green for the whole of Great Britain. Reliable sources further report on Plaid Cymru in Wales, Scottish National in Scotland, and the main parties in Northern Ireland.
I agree with when you say "by how polls are reported [...] we should stick to Con, Lab, UKIP, LD and Green". What I am suggesting is that we keep those five for the GB-wide poll (with "Others") and then publish the poll results for Scotland, Wales and Northern Ireland. This would preserve the GB-wide poll summary for those polling high enough, while the nationalist parties can be seen better as 40% of the Scottish vote instead of 3% of the British vote, which reflects reliable sources. 123.2.85.195 (talk) 22:25, 1 November 2014 (UTC)
Do you have any comment on my second point, that across the full period represented, the BNP were more significant than they are now? Bondegezou (talk) 09:17, 2 November 2014 (UTC)
What I am saying is that the article doesn't, and does not need to be, a pure reflection of the polling results. It's not in the spirit of the article and the encyclopaedia to give all the information possible in the same way that polling companies do. If the BNP become large enough in polling to significantly affect the next election as UKIP and now Green is doing, then they should be added. The best polling result for the BNP was 3% at the beginning of this article's span, and has been down since then. SNP and PC are actually relevant to the electoral outcome and should be compared with their respective Scottish and Welsh opponents, whereas reporting on the BNP's polling results is utterly irrelevant and completely pointless.
The threshold for UKIP was to consistently rival (at the time current) Lib Dem results. If your point is simply that the BNP had more votes last election than other minor parties we report on, I'd like to remind you that the name of the article is for the next general election. 123.2.85.195 (talk) 12:39, 2 November 2014 (UTC)

An article like this isn't merely about predicting the next general election. It has a historical purpose, describing parties' popularity over time. As such, I feel it should reflect all the ups and downs over this period, not merely what's going to matter at the next election. The collapse of the BNP over the period is part of the story, as is the rise of the Greens and the SNP, and of course the decline of the LibDems and rise of UKIP. Were the article merely about the next election, there wouldn't be any point keeping all the results from 2010-2013! Bondegezou (talk) 15:10, 2 November 2014 (UTC)

I agree with you, but I don't think there is rationale to add the BNP on the same level. The BNP doesn't have relevant ups and downs, following WP:UNDUE. It started this parliament with a low vote, too low to be a notable participant, and now it's even lower. UKIP's results were added in from 2010 because they were retrospective when they were beating the Liberal Democrats. If UKIP or Green never rose to past 5% then they would not be notable enough for this article. The SNP is much like the BNP when considering Britain as a whole, which is why the nationalists should be compared in their nations.
The role of the BNP here was not the purpose for this discussion. I think there should be a consensus on separating the article into two sections, one with the results of the five main parties, and one with the main parties of Wales, Scotland and Northern Ireland. 123.2.85.195 (talk) 22:52, 2 November 2014 (UTC)
As mentioned above, polls specifically for the four nations are already covered at Opinion_polling_for_the_next_United_Kingdom_general_election_in_individual_constituencies. I, personally, have no strong view about whether they stay there or are moved here. Bondegezou (talk) 08:47, 3 November 2014 (UTC)
It would be best for the three (not four) national Westminster polling to replace the detailed polling section. They would be replacing SNP and PC results in their respective countries for more accurate and clearer results, leaving only the BNP which I believe the consensus would be against keeping BNP as giving undue weight. This has nothing to do with the individual constituencies page. 123.2.85.195 (talk) 09:31, 3 November 2014 (UTC)
If there is any England-only polling, I don't see why we would exclude it. You say this has nothing to do with the individual constituencies page, but that's where the Scottish and Welsh polling currently is.
There has been a long and wide-ranging discussion on several overlapping issues (which parties to include in what tables in what articles), going back some years. I don't know what consensus is; I do know consensus has been hard to achieve previously! I would suggest an RfC process would be a good way of trying to sum up thinking, put forth specific plans, and determine what consensus is. Bondegezou (talk) 11:06, 3 November 2014 (UTC)
Three, because there isn't any England-only polling. Having these results on this article does not have to change any other article.
I think if there was every option that could possible be considered, the community would be infinitely divided that consensus would be impossible. I think a compromise of GB-wide five-party polling with a section for polls from Scotland and Wales covers the seven political parties, making the detailed poll section redundant. It's not what any person including myself would most want, but it's the option that would cover most reasonable people's idea of the article. 123.2.85.195 (talk) 13:04, 3 November 2014 (UTC)
It is generally considered a bad idea to have blocks of the same material repeated. If you want Scottish/Welsh/N Irish polling in this article, it should be removed from the other article.
I'm not suggesting every possible option is considered (& I agree with you as to why not). I am suggesting that you, or someone, put forth a specific proposal in an RfC format, encourage people to comment on that (including putting links to the RfC on other Talk pages) and then we see what the consensus is.
On a tangential note, Opinion polling for the next Spanish general election is an interesting example of a polling article where a party has surged in the polls and what they've done about it. I note table and graph include parties regularly polling below 5%. Bondegezou (talk) 16:30, 3 November 2014 (UTC)
I wouldn't suggest having every party that is included in the polls on the main table. I think every party regularly polling above 5% should be included in the main table (Conservatives, Labour, Lib Dems, UKIP & Greens). The detailed polling should probably be removed - as the only other parties included are the BNP and the SNP/Plaid alliance. Additionally, there is missing data on the table, which now can't be added to this table as the polls have been deleted from the relevant websites. The BNP and Nationalists should just be included in the other category of the main polling section as the BNP are polling around 1% and the SNP/Plaid alliance is only relevant to Scotland and Wales.
On where the Scotland and Wales polls should go. I put the Scotland polling originally in the constituency article, as I felt that it had more of a place in that article, than the main article. There are currently only 17 polls from Scotland included (plus five from Wales & NI), so they should have their own page at this time. The page could be renamed "localised polling for the next United Kingdom general election" to solve the issue of where they should go (or something along those lines). Clyde1998 (talk) 17:49, 3 November 2014 (UTC)
I'm not convinced that we should try to come up with a cut-off ourselves, like every party regularly polling above 5%. What is that cut-off based on? Not anything in Wikipedia policy, and it's not consistent with polling articles for other countries. Instead, I feel we should follow what reliable sources do, what they report. However, if you go with that cut-off... Last 20 polls listed in the article, as I write, Greens were above 5% in exactly half. And that's a recent phenomenon: go back to August, say, and the Greens are hardly ever above 5%. If that's the cut-off that is chosen, then I think it's questionable whether the Greens should be included (at least yet). Bondegezou (talk) 19:50, 3 November 2014 (UTC)

I agree with Clyde that not every polled party should be included, and that 5% is a good benchmark considering that the party has to be consistently above that mark and not simply reach it once. The detailed polling was basically for the Green Party when they weren't in the main section, and I think bringing the Scotland and Wales polls into this article removes any more need for the detailed polling since it covers SNP/PC much better. 123.2.85.195 (talk) 22:11, 3 November 2014 (UTC)

It looks like we are heading towards broad agreement to keep the Greens in the main table and delete the 'detailed poll results'. As the person who has added the most rows to that table, I couldn't say I mind too much, but I would like the data there of the Greens share earlier in the year added to the main table, rather than just deleted. Could someone with a table editor work on that, please? DrArsenal 46.208.137.165 (talk) 07:48, 4 November 2014 (UTC)
I don't feel like we're heading towards broad agreement. Most people who took part in earlier discussions have not expressed a view now. That's why I feel a more concrete effort to establish consensus, like an RfC, is appropriate. I feel editors wanting a change should put forward concrete proposals and then seek to bring in opinions from multiple editors. I, for one, don't feel there is strong argument for a table including the Greens but excluding SNP, PC and BNP. I see no reason to drop the 'detailed poll results' table. Bondegezou (talk) 12:57, 4 November 2014 (UTC)
I feel that the main table should include Conservatives, Labour, Lib Dems, UKIP and the Greens. The Nationalists shouldn't be included as they only relate to Scotland and Wales - and with the separate tables for these countries, there isn't any need for them being included in the main GB table. Additionally, the samples for Scotland and Wales are far too small (usually around 10% between the two) to draw real conclusions about how the Nats would perform on a national scale. The BNP are polling very consistently below 1%, so shouldn't really be included in the main table. Clyde1998 (talk) 18:37, 4 November 2014 (UTC)
Bondegezou, I'm not sure what you mean by there is no strong argument. The arguments have been made, and if you disagree then that is your opinion. There hasn't been a discussed consensus for the detailed polling or for adding the Greens. Most of the information in the detailed polling is already in the first section (especially since the Greens have been added), and Britain-wide poll results do not help to display PC and SNP results. I suggest we follow the same methods for editing, since the detailed polling table was very good at the time (before Greens added). 123.2.85.195 (talk) 21:11, 4 November 2014 (UTC)
The argument made does not seem strong to me: I am unclear how it is rooted in Wikipedia policy (WP:UNDUE, WP:RECENCY, WP:LOCALCONSENSUS), it does not appear consistent with other Wikipedia aritcles, it entails cut-offs for which no specific reasoning has been given, and the cut-offs suggested don't appear to support the changes suggested (Greens barely meet the "regularly polling above 5%" suggestion and are barely above the SNP). It doesn't seem to me that there is yet a consensus for the changes (see the debate above in "Green Party [2]" section). Bondegezou (talk) 11:35, 5 November 2014 (UTC)
Bondegezou, it is true there isn't perfect consensus. A single editor undid the inclusion of Greens in the main table and added some accusations to the "Green Party [2]" section above. However, that editor has not yet engaged in real discussion. If what that editor says is correct he/she has not, so far as we can tell, ever been involved before in discussion on this talk page. I have made two relatively minor 'undos' and left a message on that user's talk page, in addition to here, to try to get that person's engagement in discussion on this page. Apart from that it appears that everybody participating, yourself included, is arguing for Greens to be included in the main table, and the scope for discussion is precisely what to do about the parties mentioned in this section heading, currently listed in the 'detailed poll results'. I think that is because pretty much everybody discussing accepts that a significant number of reliable sources are now routinely listing Greens' poll results. In addition, it is a reasonably regular occurrence (eg tonight's YouGov again) for the Greens to be level with the Lib Dems, and sometimes Greens are ahead of LDs in poll results, making it inconsistent within the page for one party on 7% to be listed, but not another: to do otherwise may be undue.
As for WP:RECENCY that is not Wikipedia policy. There is no clear policy I can find (despite searching) requiring consistency between articles. Certainly WP:LOCALCONSENSUS is relevant, but what "community consensus on a wider scale" do you feel is being overridden? It isn't at all clear to me that WP:UNDUE is at stake here if Greens are mentioned in the main table and the detailed table is deleted, because SNP and Plaid both get a good degree of prominence in the country-specific tables, and BNP is regularly polling at well under 1/10th the support of the Greens or the nationalists, at such a low level that fairly frequently a sample of 2000 fails to find a single BNP supporter (ie relevant margin of Error consistently greater than level of support, so failure to find a single supporter is happening more frequently than could be explained by the 5% of polls falling outside the confidence interval). You need to make some case why you think including Greens but not SNP/Plaid/BNP in the main table is undue attention/emphasis. Dr Arsenal 46.208.137.165 (talk) 23:25, 5 November 2014 (UTC)
Bondegezou, I'd like to hear how you think keeping the BNP in the polling data is helpful for the article itself, outside of any policy. I don't see media sources reporting on the BNP as much as PC or the SNP. 123.2.85.195 (talk) Onetwothreeip (talk) 06:10, 6 November 2014 (UTC)
Onetwothreeip, hello! I think if you read the thread you will get as much idea from Bondegezou as you could from Bondegezou repeating. I would tend to agree with Bondegezou's stance that BNP earlier in this parliament had enough reporting of their polling in RS and a high enough vote share to warrant inclusion. However, I don't think that is terribly relevant, because nobody is likely to add the data to show that they had a higher vote share earlier in the parliament. Nobody has added as many rows to the detailed polling results table as I have, and I don't think it is worth the effort, or at all likely to happen. Bondegezou argued, and argues, for inclusion of all parties mentioned by the pollsters, but then doesn't implement it (neither here nor at Opinion polling for the next United Kingdom general election in individual constituencies). And as anybody who tries to add detailed polling results for 2010 will find, a lot of the links to the tables for polls conducted in 2010 are now dead (indeed quite a few for earlier in 2014 are already). And what is the point of including BNP in a table because of them polling at levels that we don't actually record? DrArsenal 46.208.137.165 (talk) 10:15, 6 November 2014 (UTC)

Thanks for a considered reply, DrArsenal, and for your continuing editing work on the article.

We had a long debate about these issues quite recently, culminating in an RfC concluding in August ("RfC: Should the Green Party, along with other parties be included in the table of polling results" section above). We then had considerable post-RfC discussion through September and beyond ("Achieving two tables" and "What on earth?" sections above). It seems to me rather soon to ditch all that for a new proposal. You are right that WP:RECENCY is just an essay and not policy, but it is a much cited essay that is worth considering, I think. It expands on WP:NOTNEWS, which is policy. The main driver for new proposals is a very recent rise in the Greens' support that, for all we know, could dissipate just as quickly. This section has seen a long and dense discussion, but few are participating in it. I see the best solution to this scenario to be a new RfC, clearly stating what is being proposed. This will garner more attention and will help crystallise discussion. I am unsure why there is resistance to this idea. (By the way, I don't see why it is relevant that the editor objecting to the inclusion of the Greens in the "Green Party [2]" section above has "not, so far as we can tell, ever been involved before in discussion on this talk page". WP:BITE!) I think a better approach is to drop (or de-emphasise) the main table and just give detailed poll results, including Greens, SNP/PC and BNP. I think this is better because:

  • (a) It better reflects prior discussion, including a previous RfC (as above).
  • (b) We should be guided by the sources we cite. This article reports the activities of polling organisations. Those pollsters, in their reporting and when they prompt for different parties, treat the Greens, SNP/PC and BNP the same. Ergo, we should treat them the same.
  • (c) WP:LOCALCONSENSUS argues we should not make up our own rules just for UK polling articles. Articles like Opinion polling for the next Spanish general election, Next_Italian_general_election#Opinion_polling, Opinion polling for the French Presidential election, 2017 &c. all report parties with low levels of support along side parties with higher levels of support. The arguments put forth here to exclude certain parties do not appear to arise for those articles. If those articles go for an inclusive approach, why don't we here?
  • (d) Wikipedia is an encyclopaedia, not a news source. The purpose of this article is not a rolling list of recent polls to help predict the next general election; rather, it is an encyclopaedic archive of all polls in the period between the last and next general election. As such, decisions on what parties to include in tables should be based on that whole period, not just the last month or two. The BNP were much more significant at the beginning of this period than they are now. While I take your point, DrArsenal, that data from some years back is harder to find and someone has to fill in the table, policy is clear here that that is not an argument against the approach: see WP:IMPERFECT.
  • (e) The argument made for including the Greens is that they are close to, and occasionally overtake, the LibDems. Yet the SNP is close to and often matches the Greens on GB-wide polling (and SNP+PC sometimes overtakes the Greens). I do not see the clear blue water between the Greens and the SNP such that it makes sense to include the Greens and exclude the SNP. I do not accept the argument that the SNP's performance will receive due prominence because there will be a Scotland-only poll. The GB-wide poll will clearly get much more attention from readers. Bondegezou (talk) 13:20, 7 November 2014 (UTC)
Bondegezou, while (as I mention above) I have never been active on talk pages before (iirc not at all, before the RFC on this page had already closed), I did read the whole of this page, so I was aware of the RFC above. I can certainly see your point that it "seems ... rather soon to ditch ... for a new proposal". I have, I think, done more practical editing to implement the outcome of that RFC than anyone else.
I don't think WP:NOTNEWS is relevant. The bit that is newslike is the reporting of each poll as it happens, and if we didn't do that, the page would not exist. It is suitably encyclopaedic to list Greens (and other parties - cf. ipsos-mori) in that especially as we aren't constrained by paper. What has happened over the couple of months since the RFC closed is that failure to list Greens, in particular, has become considerably harder to justify than it was before. I think I would agree with you in that, as I argued above on 2nd Oct (in what I think was my first ever contribution to a talk page), any party that is reported as polling 2% should be included in the table for the whole year. All I have been saying here is that by trying to implement it, I have learnt just how hard to implement it retrospectively is. I don't have any strong opinion between dropping the main table that we currently have or dropping the detailed polling results, but I do want realism.
You "don't see why it is relevant that the editor objecting to the inclusion of the Greens ... has 'not ... ever been involved before in discussion on this talk page'". It is relevant because the if they had been (or acknowledged they had been, if you believe accusations), then we would have more to go on about their reasons. Clearly the editor in question has reasons, and I am not arguing that their view should be discounted, but the editor in question hasn't really justified their position, despite my attempts to get them to discuss. (I'm not sure WP:BITE is so relevant with an editor whose first ever edit was apparently to a talk page more than a year ago, and whose first edit on this page is to make accusations and who says the page should be protected).
Enough for now. May respond more later. DrArsenal 46.208.137.165 (talk) 19:56, 7 November 2014 (UTC)
Sorry my earlier response yesterday was a bit incoherent: that meant it wasn’t a reasonable response to yours, Bondegezou: I ran out of time to knit it all together coherently.
You call for a new RfC. I just don’t know: it seems to me that an RfC is a way of avoiding WP:Localconsensus causing problems. But at the moment, I don’t see how the (attempt at reaching) local consensus is resulting in us making up our own rules. You repeatedly say we should be consistent with other articles, and I repeatedly point out that consistency between articles doesn’t seem to be Wikipedia policy (unless someone can give some evidence to the contrary). Perhaps we are making up (or trying to make up) our own rules, but if we are, I don’t think you have yet put your finger on how. Your point (b) hardly gets us anywhere: afaik some of our reliable sources mention 8 parties routinely, some a slightly different 8, most are now listing 5 or more, but some are still only listing 4, while Ashcroft seems to mention any that are >0.5%, so doesn’t necessarily mention the same number twice running on his ‘published’ page.
I absolutely agree that WP:NOTNEWS and WP:IMPERFECT are relevant in the ways you list in point (d), but I now think that arguing for BNP to be listed on the basis of its higher polling early in the parliament isn’t so much arguing for “work in progress” as arguing for a merely theoretical possibility. When I started editing the “detailed polling results” only two months’ worth of data was included and it had an out-of-date flag. I added more than 5 months’ worth of further data. Since the start of September, the number of rows added by people other than me to that table is in single figures. I might put in the effort to take it back to 1st Jan 2014, but can’t see myself taking it further back and I really don’t think over the next 6 months anybody is going to take it back significantly before that date, or add much at the 2010 end; and past 6 months time, it won’t be “Opinion polling for the next United Kingdom general election” any more, and even less people will feel like editing it. “Opinion polling for the United Kingdom general election, 2010” has had a total of 10 edits in the last two years, most of which I think were by bots or gadgets. So, in theory I think it would be better to include all the parties currently in the “detailed polling results” for the entire parliament, in practice, I don’t think it is going to happen, and we need to think whether the likely incompleteness is going to render what is there rather nonsensical. You could, of course, make the point moot by adding a significant number of months when BNP were polling better for yourself.
The key argument for including the Greens is encyclopaedic coverage of polling for the 2015 UK GE ought to include all parties that are significant. It ought to include parties that grow in significance before they were polling at high levels to show that growth in significance, and it ought to include parties that decline in significance showing that decline. But significance is not just measured by polling levels. I have no doubt that SNP claims to significance are stronger overall than Greens, despite Greens currently polling higher than SNP. I simply don’t know whether SNP will get due prominence by listing Scotland-only polls. I have tended to take the view that if (people who I think are) SNP supporters are happy, then SNP are getting (at least) due prominence. DrArsenal 46.208.137.165 (talk) 08:27, 8 November 2014 (UTC)
Bondegezou, I'm not aware of your reasons to keep the BNP in polling. As I said previously, to which you did not respond, "Bondegezou, I'd like to hear how you think keeping the BNP in the polling data is helpful for the article itself, outside of any policy. I don't see media sources reporting on the BNP as much as PC or the SNP. 123.2.85.195 (talk) Onetwothreeip (talk) 06:10, 6 November 2014 (UTC)".
To your four points, I cannot find them sufficient enough to support the changes you seek.
  • Prior discussion is not relevant, only current consensus is.
  • Polling companies do not treat the main parties, the nationalists and the BNP as the same because they do not treat political parties. They treat people who respond to polling, and do not decide, influence or intend on how polling is reported. This is done by reliable sources (per WP:RELIABLE) who rarely if ever report on the BNP.
  • This is an unusual interpretation of WP:Local consensus, which suggests that a WikiProject's community cannot override generally accepted guidelines and policies, which happens anyway. Wikipedia does not "employ hard and fast rules", and we are entitled to edit an article to make it as suitable for an encyclopedia as possible. The articles that you linked showed unequivocally bad ways of reporting polling, and there are opposites such as Opinion polling for the Australian federal election, 2013 and Opinion polling for the 42nd Canadian federal election.
  • DrArsenal brings up the point, which I concur, that it's not tenable to add the BNP in polling when your basis is because of the first polls, which would then require all reported polls since then to provide criteria for the BNP. I still must disagree that the BNP ever had a high enough share of opinion polling in this parliament to be considered along UKIP and the Greens, let alone polling at a consistent level.
  • I think adding the SNP to main polling is a separate matter, but even if they were included, they would not be receiving due prominence since the likelihood to win more than half of Scotland's seats is not expressed in a percent below 5% (of Britain) like how it is above 40% (of Scotland). The national polling gives better context for the SNP and PC, which stops them from being considered as one party which should definitely not happen at all.

Apart from policy, I simply can't see why you would want to find all the polls you can and find poll results for the BNP as far back as you can find and then continue to add their very low results. It seems to be a useless endeavour that serves no actual purpose for the reader other than a reminder that an irrelevant party is still irrelevant. 123.2.85.195 (talk) 12:17, 8 November 2014 (UTC)

Green Party (2)

I publicly thank User:Chessrat for having inserted the poll numbers of the Green Party of England and Wales in the main table. They deserve to be there. I would add the poll numbers of the SNP and PC to the main table too. --Checco (talk) 14:24, 30 October 2014 (UTC)

Thanks to User:Chessrat for all the work. Given ongoing discussion on the matter, it would have been nice to have reached consensus here first. I too am uncertain about adding the Greens without also adding SNP/PC, and indeed there is an argument that the BNP should be included too, i.e. using the detailed poll table further done the article. Bondegezou (talk) 16:41, 30 October 2014 (UTC)
Thanks. I was wondering whether to wait for official consensus first, but it seemed that almost everyone agreed that the Greens should be included in the main table, while many thought the other parties (SNP/PCY etc.) should be included, but their inclusion is more contentious. So I added the most popular change even before the formal consensus, in the spirit of WP:BOLD. Chessrat (talk) 17:54, 30 October 2014 (UTC)
I think it is about time to include them on the main table. Last time it was discussed, the Greens had reached the LD score a couple of times and were being mentioned in the headline results only I think by Lord Ashcroft. YouGov have now joined him in this and there have been a few more occasions where the Greens have matched or exceeded the LDs. I think we can now say it is being sustained. As the table shows the "main parties" in the UK, originally only the big 3, it seams reasonable to add a new party as and when it shows its ability to match one of the existing "main" parties on a sustainable level. So I wouldn't add any other parties unless they grow to a point where they can match one of the new "big 5".Saxmund (talk) 18:20, 31 October 2014 (UTC)
Saxmund, now Greens are being mentioned routinely in Sun_Politics tweets and on UKPollingReport, as well as the sources you mention. DrArsenal 46.208.137.165 (talk) 22:42, 31 October 2014 (UTC)
I'm getting bored of pointing out that matching one of the 'existing main' is logical nonsense for an argument. It is completely irrelevant whether a party scores higher or lower than the lib-dems, the lib-dems are not a yardstick. It is frustrating that some refuse to understand this. Iliekinfo (talk) 22:41, 1 November 2014 (UTC)
If you're getting bored of pointing it out, it's because you are just expressing an opinion, if you think it's logical nonsense then how about deploying logic and explaining why.Saxmund (talk) 19:47, 5 November 2014 (UTC)
Iliekinfo that has only been mentioned by one of the contributors to this thread, who also mentioned extra RS who were routinely mentioning Greens in headlines. So, few of us are using LibDems as a yardstick. Are you OK with the page with the Greens' column or not? DrArsenal 46.208.137.165 (talk) 18:18, 2 November 2014 (UTC)
I agree with User:Chessrat on everything. Also, I would add the Greens and SNP/PC both in the table and the graph. --Checco (talk) 15:48, 4 November 2014 (UTC)
Ps: At least the Greens should appear in the graph as they are now in the table.
Oh no you don't, that was very sneaky. YOu have not involved any of the main contributors to this section and you have deliberately done this in as discrete way as possible. Edit will be reversed, besides a few days is not a sufficient amount of time to overturn such a consensus. Edit will be reversed. I think this page may need to be protected. 212.159.166.201 (talk) 21:50, 4 November 2014 (UTC)
^Sheffno1gunner? It's against the rules to use a sockpuppet to evade a block, for obvious reasons. 123.2.85.195 (talk) 22:00, 4 November 2014 (UTC)
123.2.85.195 - I don't know who you are suggesting is sockpuppeting - if it is me, please say so. But 212.159.166.201 and 123.2.85.195, can you please note and heed the instructions at the top of the talk page: "Assume good faith". thanks DrArsenal 85.133.27.19 (talk) 22:50, 4 November 2014 (UTC)
I always assume good faith, and I ask you to do the same, 85.133.27.19. 212's talk page shows multiple blocks for sockpuppetting. 123.2.85.195 (talk) 22:55, 4 November 2014 (UTC)
Sorry, 123.2.85.195 - an odd football/politics juxtaposition. _I'm_ new enough around here to not know to look at an IP editor's talk page to find things like that out. DrArsenal 46.208.137.165 (talk) 00:15, 5 November 2014 (UTC)
No worries, you seem to be one of the most contributors around here. 123.2.85.195 (talk) 05:07, 5 November 2014 (UTC)
212.159.166.201. I could not agree that adding the column was done "in as discrete way as possible". There has been renewed discussion on this page, as you can see, since 21st Oct. The first resultant change was by me, adding a note, rather than the column, for which I waited 24 hours after I raised the possibility here, and even then I did it explicitly in the spirit of boldness, it was reverted quickly, and we all allowed it to remain reverted. However, the discussion didn't go away. On 26th Oct because of subsequent developments I argued "The circumstances at the time of the RfC no longer hold. Something needs to change on this page". Nobody defended the status quo, then another editor (Chessrat) added the column on 28th Oct. Rather than leaving the edit summary empty, which would be consistent with "in as discrete way as possible", Chessrat explicitly said in the edit summary what was being done, and that it was in the spirit of boldness. NOBODY reverted that change or came to the talk page to argue that Greens should not be listed in the main table until you deleted the column on 4th Nov. The arguments were all about how to deal with SNP/Plaid and BNP, not about whether to include Greens in the main table. Now, you delete the column without first joining the discussion on this page and waiting to see what happens as a result of the discussion after you join it. Please join the discussion, and in doing so, assume good faith. Do you accept that circumstances have significantly changed, in that considerably more reliable sources are routinely including Green poll results in headlines/initial reports? If you don't, how do you justify that position in the face of the evidence given on this page? If you do, how do you suggest the page should take account of that change? DrArsenal 46.208.137.165 (talk) 09:12, 5 November 2014 (UTC)

I would support Bondegezou's suggestion for having a more detailed chart further down with lesser parties included, at least those for which we have enough data to do so.--ERAGON (talk) 12:40, 8 November 2014 (UTC)

Eragon, welcome to this field, and thanks for your input. I can't quite work out what you mean though. Are you arguing for both a main table that includes Lab/Con/LD/UKIP/Greens/others and a detailed table that includes Lab/Con/LD/UKIP/Greens/SNP/PC/BNP/others (the status quo) or are you arguing just for a table that includes Lab/Con/LD/UKIP/Greens/SNP/PC/BNP/others (which is what Bondegezou is arguing for, afaik)? Meanwhile, your input on topics further down this talk page would be welcome: you will see mention of timelines. DrArsenal 46.208.137.165 (talk) 23:28, 8 November 2014 (UTC)
Sorry, I should have been more clear. I like the current arrangement of things, but I would advocate an additional graph down at the bottom which plots all of the parties we have data for. So there would be two graphs on the article- one at the top with 'the big four', and a detailed graph at the bottom with the minor parties included as well.--ERAGON (talk) 15:21, 9 November 2014 (UTC)
Thanks ERAGON. I think, though, you aren't agreeing with Bondegezou. My understanding is Bondegezou is arguing JUST for the Lab/Con/LD/UKIP/Greens/SNP/PC/BNP/others table. Very few people have said anything about the graphs. Perhaps we will discuss that in time, but we haven't got to a settled position on the tables, and which columns to include yet. DrArsenal 46.208.137.165 (talk) 20:48, 10 November 2014 (UTC)

Synth

It looks to me like indicating which party is in the lead and calculating the margin contravenes WP:SYNTH, unless a reliable source gives it. Populus don't appear to give leads on their website, on their tables, or normally have a client who does this either. My conclusion is that editors are making that calculation themselves, which constitutes original research, and thus Populus (and other pollsters where no reliable source for leads exist) should not have a coloured background for the party in the lead, or a calculated margin of lead displayed. DrArsenal 46.208.137.165 (talk) 17:47, 24 October 2014 (UTC)

I've always been slightly uneasy about the lead columns. They are not used on most articles of polling data. Bondegezou (talk) 19:33, 24 October 2014 (UTC)

Addressed in WP:SYNTHNOT per WP:CALC that obvious calculations are not original research. 123.2.85.195 (talk) 08:41, 29 October 2014 (UTC)

Two points arise. (1) Is it an obvious calculation? While it appears to be on the surface, there are complexities that arise. For example, the error term on the difference from an opinion poll is bigger than the error term for individual levels of support. (2) Why calculate this figure? Is it WP:UNDUE to highlight this figure? It is a figure that has no direct role in UK election results. Bondegezou (talk) 15:23, 29 October 2014 (UTC)
It might not have a direct role in UK election results, but in the UK the RS routinely give the "poll lead" of the Tories over Labour (or vice versa) as a key component in any summary of the results. Pollsters such YouGov publish it on their web page, websites such as ukpollingreport and politicalbetting do this, as do the mainstream media. The figures given always relate to the rounded party shares, despite the level of error that might arise from this. So I think it is an acceptable practice, in the UK at least. For example, over the last few days YouGov has published references to "Update - Labour lead at 1" and "Update - Conservatives and Labour tied". See http://yougov.co.uk/news/categories/politics/. I am sure I could easily find a lot more citations.Saxmund (talk) 18:46, 29 October 2014 (UTC)
I can see your point, Saxmund, where the lead is given. But as I said, it doesn't happen for Populus. DrArsenal 46.208.137.165 (talk) 19:05, 29 October 2014 (UTC)
I think you're being picky here. For example, the poll lad might be given on politicalbetting with a headline "labour leads by 2%" or maybe the UKIP figure is more interesting "UKIP on 20%" so you may have the bizarre situation where we can report a poll lead one week because it has been reported on a RS, and not another week. As has been pointed out by others, it is a simple arithmetic calculation, and it is commonplace in describing British polls.Saxmund (talk) 22:52, 29 October 2014 (UTC)
By contrast, WP:SYNTHNOT seems to make it crystal clear that the trendline in the 'Graphical Summary' is Synth. Whether the lead is, or isn't, is less clear-cut, however, I think highlighting with all three of (1) bold for the leading party's %, (2) the coloured background in that column, and (3) the coloured background again for the 'lead' column is clearly WP:UNDUE. DrArsenal 46.208.137.165 (talk) 15:42, 29 October 2014 (UTC)
If you want to claim that the poll lead is original research, then just about anything becomes original research. WP:SYNTHNOT makes it clear that obvious calculations from sources are not original research. The example given is that if a source says 1 plus 2, then the article can say 3 without violating WP:SYNTH. As long as 38 minus 33 equals 5, we are entitled to say that Labour leads Conservative by 5. This is not supposed to be merely a reflection of reliable sources, this is an encyclopaedia. If you want to get rid of that, you might as well look into removing the "Others" column, the graph with a moving average, the detailed poll results, the political events in the poll data, and much more throughout polling articles. The lead doesn't have a direct role in UK elections, but nor does the poll results themselves! We do not decide on what goes in an article and what doesn't based on fitting Wikipedia policy. We decide based on what is best to improve the article, of which the policies help. The policies on WP:SYNTH are there to prevent editors from making the article worse by their original research. 123.2.85.195 (talk) 22:08, 29 October 2014 (UTC)
One thing which concerns me about the graph is the calculation used to do the smoothing. The caption states that it is a 14-day moving average, but it has quite obviously been smoothed, surely there should be a link to a description of the algorithm used?Saxmund (talk) 22:52, 29 October 2014 (UTC)
It says it's a 15-day moving trend line, but since it's obviously made in Microsoft Excel, I assume the "algorithm" is really just an average trend line for twice a month. Either way, a question for Impru20 but the graph and the trend line should absolutely be kept in some form. 123.2.85.195 (talk) 00:23, 30 October 2014 (UTC)
123.2.85.195 - you have more-or-less convinced me that the calculation of the lead and its display is acceptable. However, nobody has responded to my comment that highlighting with all three of (1) bold for the leading party's %, (2) the coloured background in that column, and (3) the coloured background again for the 'lead' column is clearly WP:UNDUE. There is also implicit highlighting in the graphical summary. When editing the 'Detailed Poll Results' last night, I left out two of the three forms of highlighting: to me just having the 'lead' column is fair enough, but more is repetition and undue emphasis. DrArsenal 85.133.27.19 (talk) 08:19, 30 October 2014 (UTC)

Opinion polling is done in many countries and Wikipedia has articles on opinion polling for many countries. These other articles often do things differently to what we do here on this article. Why? Some other articles don't do a lead calculation; some do. Other articles generally include everyone listed in the polling results. I am concerned that habits have developed on the UK articles for no good reasons. Consider:

That's mildly interesting, but I don't think moves us on, Bondegezou. So far as I can see from searching Wikipedia policies, precedent and consistency of coverage between articles is hardly valued in itself. WP:Bias is relevant in general, but not in this case. What is crystal clear is that WP:UNDUE is massively valued, and I'm pretty sure that if there is consensus that current practice of the three/four forms of highlighting is a form of undue emphasis, we should reduce the amount of emphasis. DrArsenal 46.208.137.165 (talk) 22:14, 30 October 2014 (UTC)
I don't think it is UNDUE. It is used without favour for whichever party is in the lead, and I think it is useful in guiding the human eye to what is otherwise a fairly dry table of data, so really just helps readability.Saxmund (talk) 23:11, 30 October 2014 (UTC)
I'm not arguing that there should be no visual cue as to who is in the lead, just that three or four visual cues is too many for the 1st/2nd place when there are none at all for 2nd/3rd, 3rd/4th or 4th/5th. "Whichever party is in the lead" obscures the reality that only two parties, with a combined support of <70% of the electorate, get 100% of that triple/quadruple emphasis, thus being emphasised at the expense of parties with the support of the other 30%+ of voters. If removing the 'lead' column is not acceptable, then I would say the coloured backgrounds in it are sufficient to guide the human eye as to who is in the lead at any time. Alternatively, if the colouring of the cell displaying the % of the leading party is vital, I would say the 'lead' column should not use colour to indicate who has the lead. Either way, why put the % of the leading party in bold? DrArsenal 46.208.137.165 (talk) 00:03, 31 October 2014 (UTC)
I don't know what this has to do with WP:UNDUE. I don't think it's bias to Labour by having a lead column highlighted in red. The lead column is there to compare not only the parties' poll results, but to compare other leads. It's very helpful to compare Labour's relatively small lead in the past few months with the lead around 10% a few years ago. The decision to modify the article shouldn't be to conform to Wikipedia policies, it should be what actually improves the article for anybody who wants to analyse the data presented, of which the policies are often a guideline. 123.2.85.195 (talk) 00:44, 1 November 2014 (UTC)
123...you helped persuade me that the lead column on its own is more-or-less OK. But the question I am raising now isn't about any one element on its own. The lead column, colour of the leading party's share of the vote, and putting that in bold ALL give emphasis to the same basic fact. Yes, which party is in the lead is of interest, but I hold that the repeated emphasis on it in a single row is to over-emphasise. Historically we may judge one of the parties passing another in the opinion polls as being a really significant event in UK politics, but we don't yet know whether it will be Lab passing tories, UKIp passing lds, greens passing LDs, or perhaps in future SNP passing LDs or Tories passing Lab. To give three lots of emphasis in the table for which of Lab or Tories is in the lead is to give 1st position in the poll ranking undue emphasis at the expense of other positions in the ranking, especially when there is no simple or necessary relationship between that and who will form the next government. While there are problems with the 'graphical summary', at least it shows the 2nd/3rd, and 3rd/4th relationships. DrArsenal 46.208.137.165 (talk) 18:12, 2 November 2014 (UTC)
I don't see why it's a bad thing to have a lead column, to colour the lead and the leading party's results, and to bold the text.
  • The coloured lead is essential in determining which way the lead is. This is helpful for comparing recent polling where Labour leads low, where Labour leads high and where Conservative was leading.
  • The coloured box of the winning party's result instantly shows the highest result, which is much different to the lead. Labour on 34% and Labour leading by 2 have different purposes.
  • The text being bold makes it much easier to read against the coloured background and gives it more attention, which is a good thing.
This is not over-emphasis of the same thing, this is emphasis for related yet different things that the polls can show us. 123.2.85.195 (talk) 23:04, 2 November 2014 (UTC)
You "don't see why it's a bad thing". The reason is that by giving undue emphasis to one valid and interesting element you are implicitly downplaying other valid and interesting elements that don't have any emphasis at all. The "coloured lead" box does not have to be coloured to determine which way the lead is. It could be done textually: I would suggest this would still give it mild emphasis, but less than at present. "coloured box of the winning party's result instantly shows the highest result": the same colour for the lead box is enough of a clue. "The text being bold ... gives it more attention": absolutely. But why, with a coloured box (or two) does it need YET MORE attention, taking attention away from other interesting data in the row - such as the share of the vote the LDs, a governing party, (didn't) get? DrArsenal 46.208.137.165 (talk) 23:47, 2 November 2014 (UTC)
What are the other elements that we can get from the polling results that you believe are diminished? It's not really an issue about WP:UNDUE, you just feel that there shouldn't be some things. That's fine to have that opinion, as long as you can say why it would improve the article rather than simply quoting a policy. 123.2.85.195 (talk) 05:59, 3 November 2014 (UTC)
I'm not complaining about undue inclusion or exclusion. I'm complaining about undue emphasis. The other elements that are diminished are elements you can still 'get', if you pay attention, but your attention is drawn away from them. That is the nature of undue emphasis. These are the share of the 2nd party, the 3rd party, the 4th party, the 5th party, and the relative position of those parties. If one element takes a greater share of your attention due to being bold or having a coloured background, it follows logically that other elements inevitably take a smaller share of attention. It is fair enough for there to be some highlighting of which party is leading (which would be perfectly well achieved by graphical summary alone, but there is a massive difference between that and the repeated highlighting in the same row we have at present. DrArsenal 46.208.137.165 (talk) 20:39, 3 November 2014 (UTC)

I understand why you're complaining, but can you tell us what the problem with diminishing the other elements are? The emphasis is on the winning party, the winning party's share of the votes and the winning party's lead. Everything else is thereby diminished, but what is the problem that you have with this being emphasised and other things not? 123.2.85.195 (talk) 05:46, 4 November 2014 (UTC)

We are in a situation where whichever party it is that is in the lead is getting less than 1/3 of the support of those who intend to vote, but not just 100% of the emphasis, but 100% of the emphasis repeated again and again within a single poll report row. Even if, as in recent weeks, the lead swaps, attention is disproportionately and unnecessarily drawn away from the parties supported by the other 1/3 plus of the voting public. DrArsenal 85.133.27.19 (talk) 23:00, 4 November 2014 (UTC)

Clever, 46.208.137.165 You have made a big change without gaining consensus, you have only made a partial change going back 3 weeks so the formatting is now inconsistent, and the table is to my mind now much more difficult to read at a glance, and hence less informative. And you want to be a major editor on this page but can't even be bothered to set up an account. I think the emphasis should go back to what it was. Complaining that the lead party's figures are bold and coloured is like complaining that Impru20's graph shows undue influence as the leading party is at the top. Saxmund (talk) 09:02, 7 November 2014 (UTC)

Saxmund, as I understand WP policy, it is intentionally not required to set up an account. At the moment I haven't seen any advantage, and just don't want another password to forget. It isn't a question of "can't be bothered": please don't use the talk page to cast unwarranted aspersions on others. As for the substance...
Yes, I knew there wasn't consensus, but you stopped responding a week ago, since which time the discussion had been just me and 123... and the discussion had moved on a bit, so I was trying to initiate a bold, revert, discuss process. That is why I only edited a few weeks, rather than trying to be really consistent in the table: it wasn't worth the effort for an edit I fully expected to be reverted quickly. So, thanks for coming here to discuss, I will not get remotely upset by the revert. I disagree with you, though. Complaining that the lead party's figures are bold and coloured is not like complaining that a graph has the leading party at the top. It is just as possible, as I showed, for the table to display the data without that emphasis. It is simply not possible for the graph to do what it is supposed to do usefully without having the leading party at the top. So, or analogy just doesn't hold. If you are going to disagree with me, as is your right, please come up with arguments that hold water a bit better. Thanks. DrArsenal 46.208.137.165 (talk) 11:22, 7 November 2014 (UTC)
Hi DrArsenal, I didn't realise it was you or I wouldn't have been so snotty, I thought it was just a passing IP editor. Particularly if you post from different IP addresses maybe that is a good reason for having an account, we know who it is. I have always assumed that it was courtesy to have an account name if I am going to regularly edit an article, but YMMV. The reason I haven't posted recently is partly I have been busy on another interest, partly because I see little point repeating the same opinion. I don't see emphasising the leading party is giving undue emphasis, as the leading party can change. Other countries' polling articles might not do this, but in the UK's old two (and a bit) party system it has always been the custom to give the Tory/Labour lead as a key part of reporting the poll. The highlighting helps to guide the eye in what is otherwise 8 columns of undifferentiated numbers and is useful for that alone. I don't see having bold numbers and a coloured background as being two levels of emphasis: I see it as one cell with emphasis. Maybe at this point we need an RFC?
One problem is that the UK political system has evolved quite a bit during the lifetime of this page. Back in 2010 it no doubt seemed entirely reasonable that the table showed the "main three" parties and a Tory/Labour lead, and of course UKIP and now the Greens have caught up. We maybe need to think about displaying the table more like other countries which have mature multi-party systems. But with an election only 5 months away I would suggest we wait until it is over and we have a new "Opinion polling for the next UK General Election" page. We are going to have to make other changes as the GE approaches in any case, I would have thought at least an additional graph to effectively show the large number of polls published in the last few weeks of the campaign. Saxmund (talk) 12:12, 7 November 2014 (UTC)
I know I have been arguing to overturn a widespread and longstanding practice, and it may be an RFC is the only way of getting a sensible debate, but I have never been active on talk pages before (iirc not at all, before the RFC on this page had already closed), so I don't really know. Can anybody point me to where there has been a discussion before about whether to include in the same row all of bold, coloured column and coloured 'lead' column (whether for UK or elsewhere with a multi-party system)? If it is something based on a real, debated decision at some point, it is worth thinking about the arguments put forward at that time. If, on the other hand it is something that is being done because it has always been done, or just unthinkingly copied from US practice, then perhaps it is time for an RFC. I don't know. I don't really understand the criteria for an RFC. Or is there something less than an RFC that could be tried first?
PS, Saxmund, while it is nice to be thought of as better than a 'passing IP editor', can I mention that I have been reminded below not to bite newcomers? Meanwhile, I can certainly understand 1) having plenty of other things to do and 2) not seeing the point in keeping repeating the same thing. DrArsenal 46.208.137.165 (talk) 17:13, 7 November 2014 (UTC)
  • "We are in a situation where whichever party it is that is in the lead is getting less than 1/3 of the support of those who intend to vote, but not just 100% of the emphasis, but 100% of the emphasis repeated again and again within a single poll report row. Even if, as in recent weeks, the lead swaps, attention is disproportionately and unnecessarily drawn away from the parties supported by the other 1/3 plus of the voting public."
Can you tell me what the problem with that is? I understand how the current system works, but why do you want to go back for years of the tables and change it? 123.2.85.195 (talk) 20:44, 9 November 2014 (UTC)
To me the problem is that it doesn't comply with the Neutral Point of View policy - three quotes from that policy: "neutrality means carefully and critically analyzing a variety of reliable sources and then attempting to convey to the reader the information contained in them fairly, proportionately, and as far as possible without bias." "An article should not give undue weight to any aspects of the subject but should strive to treat each aspect with a weight appropriate to the weight of that aspect in the body of reliable sources on the subject." "Neutrality assigns weight to viewpoints in proportion to their prominence." I don't doubt that giving some prominence to the lead in any given poll is compatible with NPOV, but I can't see how the amount given is proportionate. If we have, for years, been violating NPOV, we should change it, even if it needs going back through years of tables. If, OTOH, there has been careful thought and current practice has, therefore, been decided as fitting with the requirement to be proportionate, I would expect someone to be able to find it, and then I can read it, appreciate it, and shut up. DrArsenal 46.208.137.165 (talk) 23:12, 9 November 2014 (UTC)
I still don't see why highlighting the leading party whatever it is violates NPOV, especially as the Tory/Labour lead is a key component of how opinion polls are reported in the UK. It's not as if we're making it up. Or if we highlighted a party because its score was particularly interesting.Saxmund (talk) 23:33, 9 November 2014 (UTC)
DrArsenal, why doesn't it comply with neutral point of view? Who do you think is preserving the current arrangement because of bias to one party or another? I feel that it does not violate my neutrality to say that Labour, or the Conservatives, are leading. Although the leader is treated differently to those who are not the leaders, the leader can always change. If we only treated the lead like we did for Labour, that would certainly be biased, but it appears that we do the same for the Conservatives when they lead. 123.2.85.195 (talk) 00:31, 11 November 2014 (UTC)

PCC by-election

I deleted the line about the 30th Oct Police and Crime Commissioner by-election and it was reverted. In my view it is not sufficiently notable because 1) it was PCC rather than parliamentary, and so is effectively a local election 2) it was not part of a nationwide series of elections, and so as with local election results, should not be reported for just a by-election 3) it doesn't seem to have had an significant effect on polls (I waited to see, rather than deleting earlier) 4) turnout was so ridiculously low. To me, it is a case of a local government by-election, most similar to a mayoral by-election, which, so far as I can tell, haven't ever attracted such a row when they have occurred (indeed, neither have mayoral elections in general, such as in Leicester in 2011 or Bristol in 2012...). DrArsenal 46.208.137.165 (talk) 11:45, 7 November 2014 (UTC)

I suppose my counter would be (1+2) We reported the (national) PCC elections in 2012, and although this is a local election, it covers a number of Westminster constituencies so is a fairly large electorate (3) We don't know that it hasn't had an effect on polls. It could be that Labour or UKIP would be lower without this election having happened - an effect on polls can be to prevent a change that would otherwise would have happened. There may be Westminster by-elections that don't cause a change in the polls and we would include them anyway (4) turnout is low for other elections including local and European parliament elections. But happy to go with consensus. Saxmund (talk) 12:20, 7 November 2014 (UTC)
PCC elections are a new thing, so I don't think we have a developed way of thinking about them. They are local elections, but they do represent large electorates than most local elections. I'm not certain, but I think I'd go with DrArsenal's position that a single PCC by-election should not warrant a row, but that the national PCC elections do warrant a row, just as May's local elections do. Bondegezou (talk) 13:24, 7 November 2014 (UTC)
I don't see a problem with keeping the line for a specific Police and Crime Commissioner. We have covered other non-Westminster events and non-election events, but looking at the table shows that there certainly isn't too many events in the table. It wouldn't hurt to add more events, since there's hardly any already. 123.2.85.195 (talk) 23:08, 7 November 2014 (UTC)
Saxmund, Mayoral (by)-elections also cover a number of Westminster constituencies, but we don’t include them even when they take place away from May local election time. Certainly we don’t know for sure that the PCC election didn’t have an effect. However, we don’t yet know that PCC by-elections even can. By contrast we know from experience that some parliamentary by-elections do have an effect on the polls, meaning that ones that take place and apparently don’t are notable in part for the fact that they didn’t cause an obvious movement. Yes, other low-turnout elections are mentioned (including parliamentary by-elections), so that on its own is probably not a reason to exclude. But the turnout was so low, at under 15%, as to make it meaningless as an expression of the balance of political views of the electorate (and indeed it is further distorted as such an expression by being a PCC election, where only policing and crime were really relevant, rather than a broad range of policies.
123.2.85.195 I’m not averse to the idea of adding more events to the table, if they are warranted, but I would say others (such as the conference season) are more worth including, and I would like the events listed to better reflect the balance of what’s needed to give some hints at the explanations for polling shifts. At the moment, the rise in UKIP support is vastly over-represented among them, in that only one so far this year doesn’t relate pretty directly to the level of UKIP support (either growing or them not quite achieving a breakthrough that would give them a boost). Currently in 2014 Labour are explicitly mentioned 4 times, UKIP (or Farage) 4 times, Tories once explicitly and once as ‘Chancellor’. Clegg mentioned once. Nothing relevant to the SNP/others rise (most obviously there could be a line to do with devolution referendum). Nothing about Green rise. In 2013 There were two mentions of the chancellor and one of PM (but none of Tories as a party), one of Lab and UKIP, one of LDs and one of Depty PM. That was better balanced, even if rather minimalist. DrArsenal 46.208.137.165 (talk) 13:31, 8 November 2014 (UTC)
We should not be looking to add more "explanatory" rows to the polling table. Other polling articles don't do this. Discussion of the rise and fall of parties' popularity can go in articles about those parties, with appropriate reliable source citations. Bondegezou (talk) 14:07, 8 November 2014 (UTC)
Who are you quoting when you say "explanatory"? I think it's a very good addition to the article, and that we should be looking at adding more events. Significant local elections should be included which are generally ones that have a Wikipedia article about them. It's not only there to explain possible changes in polling, but to give a brief history of British politics during this parliament. Whoever thought of it was a genius. 123.2.85.195 (talk) 20:58, 8 November 2014 (UTC)

There are many Wikipedia timeline articles. If you want a timeline of British politics during this parliament, you could create that. This article is a list of poll results. Juxtaposing events within a list of poll results is editorialising/WP:SYNTH. Bondegezou (talk) 21:17, 8 November 2014 (UTC)

So what's the problem with adding events to the article? There are a few articles that are timelines of British politics in the last few years, and this is one of them. 123.2.85.195 (talk) 22:43, 8 November 2014 (UTC)
Bondegezou, I too am wary of Synth in these listings, and I am also VERY wary there being a violation of the neutral point of view policy here. I would suggest that 1) including this PCC by-election exacerbates biases already present in the article, 2) even if it didn't, it wouldn't be warranted, 3) even if it was warranted, a row mentioning the Scottish devolution vote is more important and 4) especially if it remains, further actions need to be taken to restore a neutral point of view in regard to these 'notable event' rows for 2014. DrArsenal 46.208.137.165 (talk) 23:43, 8 November 2014 (UTC)
I don't understand what biases may exist. We try to mention key occurrences in national politics that may (or may not) influence public opinion, with as little "editorial" as possible to maintain NPOV. It helps to punctuate the table, and is not meant as a detailed timeline of British politics. The PCC by-election may fail the notability test, and I agree we should add the Scottish referendum (and also in the Scottish table).Saxmund (talk) 23:29, 9 November 2014 (UTC)
The question is which occurrences to include, that is where bias may lie. Juxtaposing an event in a table of polls implies that event has meaning for interpreting the poll results. Bondegezou (talk) 11:38, 10 November 2014 (UTC)
Absolutely, Bondegezou. Saxmund's understanding of 'bias' seems odd to me (see also the recent discussion above at the end of 'Synth'). I think we have agreement that it would help if we included a row about the Indyref. Perhaps "'No' wins the referendum on Scottish independence"? I don't think that is enough to cure the bias in these rows, but it does help a bit, and it does seem to meet the notability test. DrArsenal 46.208.137.165 (talk) 20:55, 10 November 2014 (UTC)
A little while ago there was a discussion about what constituted a political event that should be marked. Concern was expressed that there was too many of them, and it was possible that those being noted might favour one party or another. I think some attempt was made to codify the events. It is, from memory, something like elections, Government set piece statements (such as the Budget) and broadcast election debates. It is here in the archives. So we have already gone through this discussion. I don't think I started editing until 2012 so can't comment on what thoughts the article's founders may have had.Saxmund (talk) 21:51, 10 November 2014 (UTC)
Thanks, Saxmund. I have read the conclusions of that discussion. I don't think it gives us a clear lead on what to do now, though. Perhaps the 2nd April row should go. By the looks of it Saxmund, you are OK with the PCC by-election row being deleted: I think, on balance, you think it should be kept, but even you seem to accept it is on the edge of not being sufficiently notable. 123... seems to want more such rows in general, but hasn't given any reason to keep the PCC one in particular. With others arguing for less rows, and Indyref having been added, I think it is time to delete the PCC one. DrArsenal 46.208.137.165 (talk) 22:21, 10 November 2014 (UTC)
The 2 April one is in on the basis that set-piece radio or TV debates in national elections would be marked (precedent being 2010 general election), there is another one about a week earlier.Saxmund (talk) 23:00, 10 November 2014 (UTC)
Not sure I agree with you on the Scottish EU result, if it's notable enough to include in the GB section surely it's notable enough to include in the Scottish subsection? Saxmund (talk) 23:16, 10 November 2014 (UTC)
But it is only a single constituency. And as for "notable enough to include in the GB section", I'm not convinced that it is warranted to have separate lines for the European election and the local elections that took place on the same day. IMO is not _separately_ noteworthy that UKIP did well in the local elections that took place at the same time as EuroElections where they topped the poll. DrArsenal 46.208.137.165 (talk) 23:23, 10 November 2014 (UTC)
Two lines because the local and EU election results were counted on different days, and hence reported on different dates. Surely the Scottish EU election is a national election, in Scotland. It might only be one UK constituency, you might as well not report any Scottish elections at all on that basis as they are only approximately only one twelfth of the UK total.Saxmund (talk) 23:29, 10 November 2014 (UTC)

I think that a proper guideline to follow is not to necessarily determine whether an event is deserving of being placed in, but whether there are too many events or not. Apart from not thinking the PCC by-election is significant, there's no good reason to remove it. All the policy that has been quoted is quite tenuous. It just comes down to aesthetic at this point. 123.2.85.195 (talk) 00:26, 11 November 2014 (UTC)

As both have posted here, and neither Saxmund nor 123....came back with an argument to keep the PCC by-election in particular since I said I would delete it, I have deleted.
Saxmund I understood why it had arisen that the election results of the 22nd May elections were reported on two rows, but I am not sure it was justified. I note that the local election results are reported on polling day, even though virtually no results came that day. Reporting it as 22nd May suggests that it was polling day that was significant, not counting (or media report) day. Similarly General Elections and by-elections are listed according to polling day, even though only a couple of constituencies ever complete the count before midnight. DrArsenal 46.208.137.165 (talk) 09:14, 11 November 2014 (UTC)
Events are normally reported the day they happen. I recall I originally added a note about the EU results being out on the Sunday which made it clear why there were two rows, but someone changed it to the current (admittedly briefer) wording. It seemed inaccurate to list it under the 22nd as the results weren't out until Sunday/Monday. In any case, if we put both the local and Euro elections on the same date, the wording will still take up two lines.Saxmund (talk) 20:37, 11 November 2014 (UTC)
The essential reason to combine would be to reduce undue emphasis/prominence for UKIP (as I have outlined above, in 2014 Lab mentioned 3 times -now, UKIP (or Farage) 4 times, Tories once explicitly and once as ‘Chancellor’. Clegg mentioned once. Nothing about Green rise). Undue emphasis on UKIP would be improved by keeping to the standard practice and reporting election results as being on the day of the election. Maintaining a neutral point of view is an absolutely fundamental wikipedia principal, that should be respected. To combine the two with very nearly same essential content as now I would offer "The UK Independence Party comes first in the UK component of the 2014 European Parliament elections, while it and Labour make gains in local elections". It would, incidentally, also give some chance of reducing the amount of text, so it wouldn't always take up as many lines (the current version usually works out as two rows each with just over a line of text on my screen). Ideally I would make more changes to the content to further improve balance, but one thing at a time... DrArsenal 46.208.137.165 (talk) 22:20, 11 November 2014 (UTC)
Certainly in 2009 the same was done, but that doesn't mean it was right to do then, either. Also in 2009, there was a single row for "The Daily Telegraph begins publishing full details of MPs' expenses", even though for that reporting over an extended period was a crucial element to it having so much impact. So, in 2009 we have an example of an event (the 5th June elections) listed on two rows because it was reported on two separate days, and another event (expenses claims being revealed) that is only listed on one row, despite being reported on separate days.
The previous EuroElections (2004) fall within an opinion polling table with precisely no 'notable events' listed. Can anyone think of another country where the EuroElections are combined with a simultaneous election - what is done on their opinion polling table? DrArsenal 46.208.137.165 (talk) 09:14, 11 November 2014 (UTC)
I haven't checked all of them, but by and large, other country's polling tables do not include any notable events other than the actual election in question (e.g. Opinion polling for the German federal election, 2013, Opinion polling for the French Presidential election, 2017, European_Parliament_election,_2014_(Italy)#Opinion_polling, Opinion polling for the next Spanish general election (particularly interesting to see what's included in that one), Opinion polling in the Polish parliamentary election, 2011 (DrArsenal, you might like the colours in that one!), Romanian_legislative_election,_2012#Opinion_polling). As I've said before, why do we do UK opinion polling articles so differently? Bondegezou (talk) 10:20, 11 November 2014 (UTC)
Those Polish colours are a bit lary for me! Anyway I think it's quite a poor table, it needs some spot emphasis to help guide the eye. I would ask why other countries' pages don't include information on current political events, as it helps the reader to make some context of what is otherwise endless tables of figures. I would also ask why the two practices can't exist side by side. WP:NOT states "Long and sprawling lists of statistics may be confusing to readers and reduce the readability and neatness of our articles. In addition, articles should contain sufficient explanatory text to put statistics within the article in their proper context for a general reader." It suggests "using tables to enhance the readability of lengthy data lists". Judicial use of both emphasis and chronological notes in my view helps enhance the readability of the data table while not compromising on NPOV.Saxmund (talk) 20:49, 11 November 2014 (UTC)
I note WP:NOT reads: "In cases where this may be necessary, (e.g. Nationwide opinion polling for the United States presidential election, 2012), consider using tables to enhance the readability of lengthy data lists. Where it is not necessary, as in the main article United States presidential election, 2012, omit excess statistics altogether and summarize any necessary data concisely." In other words, it gives a specific example relevant to our context here and the solution given is that the polling article is a big table of data, and the election article discusses and contextualises. If Nationwide opinion polling for the United States presidential election, 2012 is being held up as the standard to which we should aspire, then let's try to copy it. That means no additional notes in the table. It also supports the current emphasis of the lead party (sorry, DrArsenal). Bondegezou (talk) 10:51, 12 November 2014 (UTC)
Thanks Bondegezou. That is useful. Cutting the 'notable event' rows seems appropriate (even if my proposal to merge the two 22nd May election rows was adopted, the Rochester and Strood by-election will mean that very soon more rows will be about UKIP than any other party again, restoring the degree of undue attention to UKIP in a problematic way. I absolutely concur that all the 'notable event' rows should all be deleted.
As for emphasis of the lead party, the issue isn't as big with a two party system (as the US is, especially for presidential elections). Contrary to what some people seem to think I do not oppose having some emphasis for the lead party in each row. With a two party presidential system, either one candidate/party is clearly ahead, in which case he/it is on course to take all of the prize, or the lead switches, so both gain from the emphasis, meaning it isn't ever undue. On the other hand, with a multi-party parliamentary system, nobody is on course to take all of the prize (ie all of the seats in parliament), and even if the lead switches, all of the emphasis is shared by those who it switches between, meaning smaller parties are consistently under-emphasised (there is nothing wrong with smaller parties getting a smaller share of the attention, which is why some emphasis of the lead is OK) DrArsenal (talk) 13:03, 12 November 2014 (UTC)
On "notable event rows", the reason UKIP may be getting the publicity is because they might happen to be winning by-elections. If you look further down the table, the other parties get them. So no-one is getting undue attention. We have already reduced the number of notable events to bona fide political happenings, as people were wanting an "event" every time, for example, the Leader of the Opposition made a speech. I would also ask why this practice has been OK for the previous four and a half years this page has been in existence (and previous ones) and is only now being questioned. WP policy hasn't changed.
On highlighting the lead - it might have occurred to you that our political system is more like the US, being a first past the post election. A lead of a couple of percent might genuinely mean the difference between the Tory or Labour party having an overall majority. This page started out with the UK having a 2+ party system, if current polling VI figures hold up at the election we could genuinely be moving to a multiparty system. I also think the tables on this page are much more readable than the ones you hold up as shining examples (as well as previous UK ones). It only has 6 months' life left to the GE. Why not leave it as it is and argue about the 2020 polling page? Otherwise changes are going to have to be made going back ro 2010 and the page is going to be in a state of flux going into the GE, when it will become more useful and more people will want to read it.Saxmund (talk) 20:22, 12 November 2014 (UTC)
How could you be saying I haven't made an argument? 123.2.85.195 (talk) 10:31, 13 November 2014 (UTC)
We've had debates about the practice of what events to list for some time. I've argued my position often before and we've had some variation back and forth in editing. This isn't a brand new issue.
123.2.85.195, I'm not certain to whom you're replying. If I've just agreed with you, good! If I haven't, perhaps you can expand? Bondegezou (talk) 12:33, 13 November 2014 (UTC)
I think it was me who 123... was replying to. 123... argued there should be more 'notable event' rows, but neither that none should ever be deleted, nor that "the PCC by-election in particular" should be kept. It was only the in particular that I was saying you hadn't made an argument for, 123... DrArsenal (talk) 22:36, 13 November 2014 (UTC)
I'm saying I'd like to keep the row for the PCC by-election. That's what I've said. 123.2.85.195 (talk) 04:19, 14 November 2014 (UTC)
Yes, I know 123... . BUT I think the only reasons you have given are "I don't see a problem with keeping the line for a specific Police and Crime Commissioner. We have covered other non-Westminster events and non-election events, ... there certainly isn't too many events in the table. It wouldn't hurt to add more" and, yes, "Significant local elections should be included which are generally ones that have a Wikipedia article about them."
Other significant local elections aren't listed (except devolved assemblies), unless they are nationwide. You haven't put forward any criterion other than 'generally [the] ones that have a Wikipedia article about them',. The word 'generally' suggests an openness to exceptions, but no criterion for exceptions is given; and you haven't given any reason why that one non-Westminster event in particular should be kept. As a result of the discussion, we added an different non-Westminster electoral event (indyref). Given that others were arguing for a reduction in the total number of rows, and you were the only person arguing for an increase, it seemed to me to be appropriate to maintain the number at that stage, rather than to allow it to increase, overriding the wishes of a larger number without any wikipedia policy to justify that override.
Meanwhile, discussion has moved on. I would say at the moment it is more likely that all these rows will be deleted than that the PCC one will be reinstated. How do you answer Bondegezou's challenge that we should be following Nationwide opinion polling for the United States presidential election, 2012, which he points out is held up as a model on WP:NOT? DrArsenal (talk) 10:44, 14 November 2014 (UTC)

Saxmund - "the reason UKIP may be getting the publicity is because they might happen to be winning by-elections". Well only one so far. And that one they engineered in order to get the publicity. I think there is an argument that if everything else was equal (and it isn't), winning a by-election caused by an MP standing down and seeking re-election is less noteworthy than winning a by-election with a less artificial cause. "If you look further down the table, the other parties get them. So no-one is getting undueattention." No. Parties that do well in simultaneous elections that are listed on separate rows are getting undue attention.

As I pointed out, the counts took place on different dates so counts/publicity were overnight Thurs/Fri for the locals and Sun/Mon for the Euros. Given that the tables are dated, and the notes have to fit in with the fieldwork dates, to do other than have two lines would be inaccurate. There were two national elections, Labour and UKIP got credit for increasing their vote in the locals, and UKIP for winning the euros. Seems fair to me. You do seem to be unduly concerned because UKIP is getting publicity at the moment. Maybe one compromise would be to mention the event, but remove any mention of who won or did well? Saxmund (talk) 17:41, 18 November 2014 (UTC)

Parties that replace their leader by having a leader replaced by an interim leader and then a longer term leader are getting undue attention (cf SNP leadership change, where Salmond stayed on until replacement could formally be made). The LBC radio debate was between two parties who felt (probably rightly) that the sheer fact of it taking place would be good publicity for them. Lab and Tories didn't take part, SNP, Greens etc weren't invited. Mentioning it adds to the undue attention for UKIP (removing it would perhaps result in lack of attention for the junior coalition partners)

"I would also ask why this practice has been OK for the previous four and a half years this page has been in existence (and previous ones) and is only now being questioned." I've only recently started really paying attention. "WP policy hasn't changed." I would argue it has been problematic all along.


"I also think the tables ... are much more readable than the ones you hold up as shining examples" I haven't held up any as shining examples at all.

The US page and a few others have been mentioned, maybe by others. Personally I find tables of bald data unpunctuated by textual emphasis or notes to be hard to follow. Wikipedia is an encyclopedia, not a data dump. Saxmund (talk) 17:41, 18 November 2014 (UTC)

"It only has 6 months' life left to the GE. Why not leave it as it is and argue about the 2020 polling page? Otherwise changes are going to have to be made going back ro 2010 and the page is going to be in a state of flux going into the GE, when it will become more useful and more people will want to read it." Its precisely because more people will want to read it that it is so important to get it right now with 6 months left before the GE. "Changes going back to 2010" are not that difficult to implement. DrArsenal (talk) 00:12, 18 November 2014 (UTC)

I read a good deal of the discussion. I broadly agree with DrArsenal and, generally speaking, I think that only sufficiently notable events (i.e. those which could have affected the following opinion polling) should be included in the table: by-elections, big debates and nothing more. But, in fact, why not removing all the events altogether? The article is merely a list of opinion polling, after all. --Checco (talk) 12:55, 18 November 2014 (UTC)
In my view the notable events rows help to make the table readable by punctuating it with a few carefully selected political events so the reader can see at a glance what was going on that might change opinion. To say that because the page is called "Opinion polling for..." it can't include a few carefully selected political events is to my mind confusing the map with the territory. Saxmund (talk) 17:26, 18 November 2014 (UTC)
Thanks Saxmund. You are right, I would be vastly less worried if party names were not mentioned (so often). Thanks for the suggestion. I do understand why there have been two separate rows for 22nd May elections, but the precedent (and indeed upheld in the 22nd May row) is to mention on day of poll rather than day of count, and if party names are mentioned, then I would say wiki policies should take precedent over that reasoning. DrArsenal (talk) 20:16, 18 November 2014 (UTC)
Euro elections are unusual in that they are not counted until the Sunday, and in the UK the election is held the previous Thursday. All other polls are counted overnight and/or the next day and news starts to come out immediately. So it seemed reasonable at the time to put in a separate line for the Euro poll as the results were not available until late Sunday/early Monday. Although we usually list the polling date for elections, it is of course the news of the election result that voting intention will react to. One compromise might be to restrict the wording of notes relating to election to simply the fact the election took place, without any description of the result, although personally I don't think that's necessary to avoid UNDUE. I don't think UKIP will have had many mentions in the 2010-2013 tables, so the fact that it is getting a few now doesn't mean it is getting undue attention in the article as a whole. Saxmund (talk) 21:03, 18 November 2014 (UTC)
Absolutely agree that it is the news of election results that VI will react to, and what was done seemed reasonable at the time. But we aren't simply putting rows in in proportion to the number of separate days on which there is media coverage above threshold 'X': if we were, there would be more rows about the demise of Huhne, and the Eastleigh by-election for example, than about the Wythenshawe and Sale East by-election. The elections of 22nd May were worthy only of a single row in the same way that the elections of 2nd May 2013 or 15 Nov 2012 were only worthy of a single row (no matter how many days the counts were spread over).
Meanwhile, mentions (or lack of them) in 2010-2013 are much less important, because attention declines the further down the article you go, and especially if the table in question is hidden when you come to the page for the first time.
I think your suggestion of having rows there without mention of party names is a really helpful, creative suggestion. Would really like to know what others think about it. DrArsenal (talk) 23:37, 18 November 2014 (UTC)

Opinium 17th November

The Company's tables linked to don't include the Voting Intention percentages listed in this article's table. Therefore can't be verified. 137.205.171.45 (talk) 20:03, 19 November 2014 (UTC)

I was thinking about removing this poll, given that Opinium seem not to have a table of VI percentages, but is there any support for such a move? best, Sunil060902 (talk) 20:15, 19 November 2014 (UTC)
The VI figures have been published by Mike Smithson http://www1.politicalbetting.com/index.php/archives/2014/11/18/opinium-poll-that-slipped-out-has-con-ahead-with-the-lds-down-on-5/ and Anthony Wells http://ukpollingreport.co.uk/blog/archives/9069, both of whom would normally be regarded as Reliable Sources. According to Wells the figures can be gained from the crossbreaks. I think it should stay but in this instance we should cite Wells as the source, rather than directly link to the tables (he links to them anyway). Also, I am not sure why the comment about "accidental" publication is necessary. The poll is now published. Saxmund (talk) 21:05, 19 November 2014 (UTC)
But how exactly are the percentages derived? I had a go myself, but didn't get a match for the published VI! Hopefully I'm not missing something obvious! Please reply on my talk page if you deem it off-topic for here. Sunil060902 (talk) 14:09, 20 November 2014 (UTC)
The percentages, in the main, look like a very simple calculation: on pp4-5 if you take the "weighted total" of voters for each party, you have Con 498, Lab 475, LD 71, SNP 63, Plaid 4, Green 66, UKIP 266, BNP 11, Resp 2, Others 5. That totals 1461, which gives Con 34%, Lab 33%, LD 5%, Green 5%, UKIP 18%, Others 6%. UKPollingReport seems to have mis-calculated the Green share, while the Others 5% appears to be a mis-calculation by User:2a02:2c40:400:0:1f9:2ffd:d610:2cd2 on adding the row. DrArsenal (talk) 14:00, 21 November 2014 (UTC)
User:2a02:2c40:400:0:312d:f0bd:3712:f292 and Saxmund have exchanged edits this evening, with Saxmund having added earlier in the day a (non-visible) comment to the page, but not to the discussion here. While I can see some logic of linking to ukpollingreport (especially as explained in the comment there that I have now radically changed), it isn't clear-cut to me that it is a better source than politicalbetting for this poll, at least in part because Anthony Wells appears to have got his maths wrong (unless I missed something, which is likely enough, but if I have, can someone please explain?). DrArsenal (talk) 23:50, 22 November 2014 (UTC)
I linked to ukpollingreport because he explains that you can get the VI from the crossbreaks. This isn't necessarily immediately apparent to the casual reader. The additional politicalbetting link is because Mike Smithson does get his maths right. You could swap them over, of course, but I think both are needed. Saxmund (talk) 11:33, 23 November 2014 (UTC)
If the consensus is that Mike Smithson's maths are less perplexing, wouldn't it be better to have his politicalbetting as the primary source, with the link to ukpollingreport as the explanation? DrArsenal (talk) 11:47, 23 November 2014 (UTC)
Looks like Opinium have now pulled the tables completely from their website... Sunil060902 (talk) 14:02, 23 November 2014 (UTC)

The event bar things.

Do we seriously have nothing better to do than to change the format from "a wins the xyz by-election" to "xyz by-election"? Is that really what we've come to? I bet someone will say it has to be changed to conform with policy. Onetwothreeip (talk) 23:04, 21 November 2014 (UTC)

If you look above, towards the end of the long "PCC by-election" thread, you will see it was a suggestion by user:Saxmund to deal with my worries about those rows breaching WP:UNDUE. I do have plenty of other things to do, both on this page and in the rest of life, and am getting on with them (changing the format was pretty quick and easy, anyway). Do you, Onetwothreeip, think it needs to be debated or changed back, or were you just having a moan because you had got used to the way it used to be, and don't like change in this respect? DrArsenal (talk) 14:49, 22 November 2014 (UTC)
What you're proposing is to offer less information in an attempt to remain impartial. I don't think any neutral reader would've thought that disclosing the winner of a by-election is bias or undue. Onetwothreeip (talk) 21:46, 22 November 2014 (UTC)
I agree with you Onetwothreeip, but I made the suggestion as a compromise, otherwise DrArsenal seemed bent on deleting all the event bars. Saxmund (talk) 21:54, 22 November 2014 (UTC)
DrArsenal and another user have been making minor but very retrospective suggestions for a while, such as removing the lead column and the coloured shading of the lead party data, or even the graph at the top. When I argued that the PCC by-election result should remain, it was largely dismissed with saying that "the discussion is now about whether they should be kept altogether". I feel instead we should be focusing on bold and relevant changes, such as the inclusion of UKIP and the Greens. Onetwothreeip (talk) 22:46, 22 November 2014 (UTC)
Hi, Onetwothreeip - I had guessed that you were the same editor as with the 123... ip address, but confirmation helps us know who we are talking to. Thanks. I felt I answered all the detailed arguments put against the PCC by-election row being deleted - I know we disagreed. I'm sorry I was unable to convince you. You, equally, were unable to convince me, or it seems, other editors who were arguing for less rows. I didn't try to prematurely close discussion, merely reflect what I thought was the balance of opinion as I saw it represented. I'm sorry if you felt I was prematurely trying to move discussion on: we both agree, I think that sometimes issues are too small to endlessly obsess about.
Yes, having event rows without mention of the winning party in the row is slightly less information without the need for a click, but the information is still easily available to users on wikipedia. The question is, what information is appropriate to display on this page? Clearly it isn't as simple as 'the more the better', otherwise we would include (several?) front page headlines for every day, and quickly swamp the data that this page is here for. So we need to make a decision about what is appropriate.
Including UKIP, the Greens, or removing the lead column, coloured shading or the graph at the top are all equally retrospective. All apply equally to the entire Parliament, so far as I can see. I still am not convinced that having a coloured lead column, coloured shading and bold for the percentage of the leading party on the same row are all justified rather than, in reality, undue emphasis, but I have a degree of realism: I can see that fellow editors are not convinced by (or don't understand, which is effectively the same) my argument. DrArsenal (talk) 23:25, 22 November 2014 (UTC)
  • Your edit with the PCC by-election wasn't a constructive addition. You were seeking to remove something that you thought contravened Wikipedia policy. I was not seeking to convince anybody, I was intending to make my perspective clear, that there are certainly not enough event rows to even consider it to be "swamping the data that this page is here for". There have been no event bars from early June to late September.
  • Event rows for elections that don't mention the winning party/s is much less information than it was before. I don't think any neutral reader of this article would think to themselves that simply stating the winner of a by-election is in any way bias or undue.
  • It really doesn't matter whether you think colouring the lead column and related emphasis on the winning poll results is "undue emphasis" or anything else, since nobody is actually going to go through the entire years of data to make a backwards change in the name of conformity. We should be discussing and editing things that actually make a difference, like adding UKIP and the Greens, not about something which literally does not matter to any normal reader. Onetwothreeip (talk) 03:25, 23 November 2014 (UTC)
Onetwothreeip I initially deleted the PCC by-election as part of a bold, revert, discuss process, and when it was reverted, rather than going into edit war or making accusations, I explained my reasoning, and engaged in a discussion long enough to have two undents over 4 days before deleting it again. If you think there are not enough event rows, please suggest some more. The only extra one that was suggested (iirc) was suggested by me. Rather than complaining about a particular deletion of a row, wouldn't it be more 'constructive' to suggest others to add, that you might find are less controversial? I might agree with you about them, since I don't take the position of Bondegezou.
I would agree that there is every chance that uncommitted readers of this article would not think that stating the winner of a by-election is bias or undue. However, that doesn't stop the accumulation of several rows actually resulting in undue emphasis, especially when the by-elections have been artificially created by a political party at least in part in order to gain the publicity boost.
If there was consensus to remove colouring of the lead column, it is easy to do. Using wikEd I have just done so for 2011 in less than 5 minutes (I did it on the least visible year on the page to avoid causing controversy while it remains). So, your argument that "nobody is actually going to go through the entire years of data to make a backwards change in the name of conformity" falls. As for whether it 'literally does not matter to any normal reader', you are no more in a position to judge than I am. Your judgement about what is important to be discussed at the moment is different to mine, but it doesn't mean it is better (or worse). DrArsenal (talk) 11:23, 23 November 2014 (UTC)
I've now reverted my deletion of colouring of the lead column in 2011 because you can look to the page history to see it, and because I forgot to put an edit summary in, in the first place. It isn't worth winding up other editors to prove a point! DrArsenal (talk) 11:55, 23 November 2014 (UTC)
  • Firstly, I'd like to ask you to remain WP:CIVIL. Your edit was not a WP:BOLD edit, your edit was a revert of somebody else's edit. I think it's very disingenuous to present your revert of a contributory edit as a bold edit.
  • It doesn't matter how a by-election is created, it's a by-election all the same. Even if you think UKIP artificially created by-elections, we wouldn't treat that any different to any other political party doing the same, and any other by-election. These by-election results are fundamentally important to this page as they indicate changes in voter support, which is what this page is dedicated to.
  • "Conservatives retain their seat in the Newark by-election" is a very simple sentence and completely not biased. Likewise is "UKIP gains a seat in the Clacton by-election", and no reader would think we are giving bias by stating the winner.
  • I'm not saying it's not possible to change the table, I'm saying that there is really no point. Most of the suggestions you have made are not contributions, despite nobody having a problem since 2010. The fact is that this article does not have problems with bias, undue weight and synthesis. Onetwothreeip (talk) 04:57, 24 November 2014 (UTC)
I'm not a regular contributor or anything, but it looks really odd to say that "'No' wins the Scottish independence referendum", but not "PartyName wins the ConstituencyXYZ by-election". I don't really see the point of obscuring the results of the by-elections, local elections, and European Parliament elections. I mean, saying "An election occurred" doesn't provide the reader with any information other than a link to the election. I think that most readers would much prefer to just have the winner in this table, rather than being forced to travel to another page just to see who won. PaperKooper (talk) 16:14, 25 November 2014 (UTC)
Thanks PaperKooper. It is good to get input from a different editor. I would agree that virtually every reader would much prefer to have the winner indicated. The problem is that nobody has come up with a way of doing that without there being a resultant bias by putting more mentions on these bars of one or two parties than their size or national level of support justifies ...and maintaining a neutral point of view is absolutely fundamental to wikipedia.
I am open to creative ways to solve the problem: I think Bondegezou's interpretation of the way Nationwide opinion polling for the United States presidential election, 2012 is held up as an example on wp:not is reasonable, but would then result in the sub-optimal solution of deleting all the 'event bars'. The question then is what can be included without breach of the neutral point of view policy. DrArsenal (talk) 18:18, 25 November 2014 (UTC)
Hmm but wp:not doesn't state "Neutrality requires that each article or other page in the mainspace fairly represent all significant viewpoints that have been published by reliable sources, in proportion to the number of people who hold that view." It in fact says "in proportion to the prominence of each viewpoint in the published, reliable sources". UKIP is a bit more prominent at the moment, it has won two recent by-elections, won the Euros and did well in the local elections. According to UNDUE the amount it is mentioned has to be broadly proportion to its "prominence" not the number of people supporting it. Saxmund (talk) 19:42, 25 November 2014 (UTC)
I understand the argument and the relevance of it, Saxmund, but UKIP haven't been more prominent than the Tories, who had only been mentioned explicitly in one of the rows this year as against UKIP's four (and would be five by now). Reliable sources as a whole have given the larger party in the Government more prominence this year than a party that has only just had its first MPs elected. DrArsenal (talk) 20:04, 25 November 2014 (UTC)
I am in favour, for reasons stated previously, of minimising the number of event bars in the table. However, I don't object to by-election results being there and I think DrArsenal is worryingly unnecessarily in this case. I think it's OK for there to be a bar saying UKIP won this and another for UKIP won that because they did. Bondegezou (talk) 16:55, 26 November 2014 (UTC)
For the avoidance of doubt, I am also in favour of keeping the event bars at the current fairly minimal level. The 2010 page has far too many, and I think we should keep it much more neutral in the run up to the General Election. I am not sure much more needs to be recorded than the Autumn Financial Statement, the Budget, the dissolution of Parliament, and the leaders' debates, if they take place. Saxmund (talk) 19:04, 27 November 2014 (UTC)
I have refrained until now from replying to Onetwothreeip's of 04:57, 24 Nov. I did post quickly after on Onetwothreeip's talk page asking for explanation of how WP:CIVIL was relevant, but have had none. Having allowed time for things to calm down, I will respond now.
  • I did not claim my edit was a "a WP:BOLD edit", but part of a "bold, revert, discuss process", which clearly allows for reverts.
  • Onetwothreeip claims "These by-election results are fundamentally important to this page as they indicate changes in voter support, which is what this page is dedicated to." I dispute that. To me, this page is fundamentally about the presentation of data tables that are too detailed to include on the 'United Kingdom general election, 2015' page. I have made a suggestion on the talk page of 'United Kingdom general election, 2015' page that there should be a prose description of the main shifts in levels of support of the parties over the life of the parliament and the causes of that, but with only one really substantive response. It is, I think, on that page that the likely explanations of changes in voter support should appear (where a neutral point of view look at the balance of reliable sources indicates a change in voter support that should be described and the availability of a suitable description/s of its cause). ...but I don't mind there also being event bars on this page, provided they don't stop the data being usable for its primary purpose and provided they don't exhibit bias (well, to be honest I wouldn't mind if they exhibited bias in my preferred direction, but that aint going to happen).
  • I had already explained (and have explained since) that I don't have a problem with a party being named in any particular row: my issue is with the cumulative effect of the rows for 2014 as a whole, and particularly the more recent rows.
  • Onetwothreeip asserts "this article does not have problems with bias, undue weight and synthesis." Perhaps it doesn't, but I don't think Onetwothreeip is the final arbiter on this, and as I understand it, it is legitimate to raise for discussion that it might, and to make edits in the hope that one's own concerns might be met in a way that is acceptable to wikipedians as a whole. I'm beginning to think that might not be possible without a 'request for comment' process, but I haven't got my head round them yet. DrArsenal (talk) 23:27, 27 November 2014 (UTC)
Personally, I see no need for the Autumn Financial Statement, the Budget, or the dissolution of Parliament to be noted. Bondegezou (talk) 11:03, 28 November 2014 (UTC)