Talk:Opinion polling for the 2017 French presidential election/Archive 2

Archive 1

Archive 2

Archive 3

Embedded links in lieu of inline citations

Hi Rami R. Not sure if I follow your edit sum. If the links in question are being used as supporting citations per WP:ELRC, then they should be formated as inline citations as explained in WP:CS#Avoid embedded links. A properly formatted cittion will provide the reader with more information about the source (WP:CITEHOW), make it easier for others to fix any problems which may occur due to link rot. There are no citation details in the table, only more embedded links. If the reader looks at the March 2017 Table and sees Ifop-Fiducial, Ifop-Fiducial, Ifop-Fiducial and Ifop-Fiducial , how do they discern any information other than Ifop-Fiducial? If the link becomes "broken" or "dead", there is no information about the source for the reader to see. This is sort of an embedded citation, a style which has been deprecated. A inline citation can be added just as quickly as <ref>[url address name]<ref> and at least give others something to work with. A bare url is not desirable, but the article can be tagged with Template:Cleanup bare URLs and a bot or another editor will clean them up.

More over a properly formated citation is not prohibitively difficult to do. For example, <ref>{{cite web|url=http://www.ifop.com/media/pressdocument/955-1-document_file.pdf|title=Rolling 2017 « La présidentielle en temps réel » Ifop-Fiducial pour iTELE / Paris Match et Sud Radio|date=31 January 2017|language=French|publisher=[[Institut français d'opinion publique|Ifop]]-Fiducial}}</ref> would look like Ifop-Fiducial in the article.

References

-- 21:27, 3 March 2017 (UTC)

If it's not difficult, why aren't you doing it yourself? I don't per se oppose switching to refs, I just see it as pointless bureaucracy. Contrary to what you say, a reader does not just see Ifop-Fiducial, he sees Ifop-Fiducial, 26 Feb–1 Mar 2017. That all a proper citation actually is: a source and a date, and these are both presented in the table. A url is really just a convenience at this point. Rami R 08:55, 4 March 2017 (UTC)

I think a proper citation involves a bit more than that. FWIW, I have no problem with doing this myself. It will probably take a bit of time, but it can be done in stages. The citations for the March 2017 table would look as follows:

• Ifop-Fiducial
• Odoxa
• BVA
• OpinionWay
• Elabe
• Ifop-Fiducial
• OpinionWay
• Ifop-Fiducial

References

I don't mind adding these to the article and then working on the rest. -- Marchjuly (talk) 10:54, 4 March 2017 (UTC)

Knock yourself out. Rami R 16:53, 4 March 2017 (UTC)

Way that French Wikipedia does it is just to provide archive links for all of the PDFs. Easy, preserves all links, and doesn't require a ridiculous amount of citation-ing. I'd probably suggest, for least disruption, to use something similar to Opinion polling in the 43rd Canadian federal election (adding a separate "archive" column next to the "poll source" column with archive links reading "PDF" or "HTML" depending on format). Mélencron (talk) 01:18, 7 March 2017 (UTC)

French Wikipedia and English Wikipedia are different projects which may have some different policies and guidelines and embedding external links into articles is something English Wikipedia asks us to avoid. As I posted above there are quite a lot of embedded links, so it's hard to do them all in one big edit, so I was going to do it section at a time. While looking at some of the links in the "Hypothetical polling" section I found quite a few that were dead links such as [1] [2] [3] and I was trying to sort those out when you made your changes. I also found a number of duplicates which could use WP:REFNAME. As to where they citations should be located, I was only adding the citations to where the links were embedded, but if there's a better place to added them, then that's fine. The full citations don't even need to be directly added to the tables as such; They could be added to the "References" section and then use simple syntax to link them to locations in the article. -- Marchjuly (talk) 02:40, 7 March 2017 (UTC)

Poll source	Archive	Fieldwork date	Sample size
				Cheminade SP	Arthaud LO	Poutou NPA	Mélenchon FI	Hamon PS	Macron EM	Fillon LR	Juppé** LR	Dupont-Aignan DLF	Le Pen FN
Ifop-Fiducial	PDF	2–6 Mar 2017	1,381	0%	0.5%	0.5%	11.5%	13.5%	25.5%	19%	–	3%	26.5%
OpinionWay	PDF	3–5 Mar 2017	1,671	–	0%	1%	11%	15%	24%	19%	–	3%	27%
Kantar Sofres	HTML	2–4 Mar 2017	1,027	<0.5%	1%	1%	12%	16%	25%	17%	–	2%	26%
				<0.5%	1%	1%	11%	13%	20%	–	24.5%	2.5%	27%
Ifop-Fiducial	PDF	28 Feb–3 Mar 2017	1,383	0%	0.5%	0.5%	11%	14%	24.5%	20%	–	2.5%	27%
Odoxa	PDF	1–2 Mar 2017	907	–	0.5%	1%	10%	14%	27%	19%	–	3%	25.5%
				–	0.5%	1%	8%	11%	25%	–	26.5%	4%	24%
BVA	PDF	28 Feb–2 Mar 2017	1,413	–	0.5%	1%	11.5%	15.5%	24%	19%	–	2.5%	26%
OpinionWay	PDF	28 Feb–2 Mar 2017	1,654	–	0%	1%	11%	15%	24%	19%	–	3%	27%
Elabe	PDF	28 Feb–2 Mar 2017	1,507	–	1%	1%	12.5%	12.5%	24%	19%	–	3%	27%
Ifop-Fiducial	PDF	27 Feb–2 Mar 2017	1,394	0%	0.5%	0.5%	11%	14.5%	24%	21%	–	2.5%	26%
OpinionWay	PDF	27 Feb–1 Mar 2017	1,639	–	0%	1%	11%	16%	23%	21%	–	3%	25%
Ifop-Fiducial	PDF	26 Feb–1 Mar 2017	1,392	0%	1%	0.5%	11%	14%	24%	21%	–	3%	25.5%
Ifop-Fiducial and OpinionWay polls in this table are rolling unless otherwise denoted by an asterisk (*); see methodology: Ifop-Fiducial, OpinionWay Continuous results can be viewed here: Ifop-Fiducial, OpinionWay. Historical data and demographic breakdowns can be viewed here: Ifop-Fiducial (**) Note: Alain Juppé was floated to replace Fillon as a result of the Penelope Fillon affair (Penelopegate), with some polls including Juppé based on this hypothesis. Juppé announced on 6 March that he would not be a candidate, regardless of what happened with Fillon.

Mockup. Mélencron (talk) 02:27, 7 March 2017 (UTC)

What you have basically done is doubled the number of embedded external links in the article when Wikipedia is asking use to remove them altogether. Once again, the same problem with embedded links/embedded links citations is that there's only a link, but no information about the source. If you want to do it this way, then you should add information about the cited source to the references section as explained in WP:ECITE#In references. Embedded citations is a style which has been deprecated on Wikipedia, but if that's approach to be taken then it should be done correctly. -- Marchjuly (talk) 02:46, 7 March 2017 (UTC)

It's a guideline, not a policy; polling articles have traditionally been designed for greatest convenience to users (which is why in practically every case they link directly to the poll in the table as opposed to providing a citation). In my view, providing citations outside of tables, unlike most cases, actually reduces accessibility to users: rather than being able to access a link within the table as always, they'll have to click to get to the link in the references section – and the citations, for what it's worth, don't actually provide any useful information that's not already in the table. (Citing without providing proper archive URLs, for instance, means that they're as much at risk of link rot as they would be in a table.) The information that's already in the table – the pollster, dates, and sample size – are already sufficient to be able to locate the original source any potential dead links that might arise, whereas the actual titles of the documents aren't at all helpful to locating them, should the links fall victim to link rot. Mélencron (talk) 02:52, 7 March 2017 (UTC)

Directly linking to the polls may be expedient, but it provides no information about the source being cited. I don't think it's prohibitively impossible to properly cite the sources, it just takes a bit of time to do it. Moreover, readers do not have to click on the link in the references section, the link will be visible when they move their cursor over the citation note and they can access it directly from there if they choose. In addition, titles of document can be searched via Google, etc., which does sometime help find archived versions of dead links. Anyway, I've asked for feedback about this from WP:ELN and WP:WPE&R. -- Marchjuly (talk) 03:14, 7 March 2017 (UTC)

"provides no information about the source being cited" – that's literally the table itself, and as for your argument of "titles of document can be searched via Google"... you do realize that almost all of these documents have extremely similar titles? Mélencron (talk) 03:27, 7 March 2017 (UTC)

Mélencron, is the data that Wikipedia should convey in the table, or is all the data in the linked document? --Dirk Beetstra ^{T C} 03:13, 7 March 2017 (UTC)

Not sure I quite get your question. Everything that's notable is in the table (pollster, fieldwork date, sample size, percentage support for each candidate). Mélencron (talk) 03:27, 7 March 2017 (UTC)

You got my question .. so the documents are just references then .. no need to link them inline. --Dirk Beetstra ^{T C} 04:00, 7 March 2017 (UTC)

So, what I mean is that the first two columns need to go, replaced by a first column for the polling organisation, wikilinked, and a new last column called refs, where each cell gets a proper formatted ref, containing the links that are currently in the first column. You say all info is in the table, so that is then sufficient, you do not need access to the documents except for verification purposes. --Dirk Beetstra ^{T C} 04:04, 7 March 2017 (UTC)

Ahh, that's an excellent idea. Thanks for the help. Mélencron (talk) 04:17, 7 March 2017 (UTC)

@Beetstra: Your suggestion seems workable to me, but just want to clarify what you mean by "proper formatted ref". Are you suggesting something like (I'm keeping the syntax basic on purpose) <ref>[url address]</ref> or rather something like [url address]? The former is basically what I was trying to do, but the latter would be more like an embedded citation. Do you have any suggestions on how the embedded links in the note-like section at the bottom of the tables should be handled? I'm not sure adding another column would work here as well here. -- Marchjuly (talk) 04:43, 7 March 2017 (UTC)

I believe they're referring to a ref that would simply look like ^[1], with the proper cite template, in a "Refs" column. Mélencron (talk) 04:54, 7 March 2017 (UTC)

Indeed, a referencing style that is conform the rest of the document. The generally accepted style on Wikipedia is indeed with the <ref>-tags, as is used in the document. I believe that at the moment there is duplication (different tables referring to other information in the same external document), that could then also be minimized. --Dirk Beetstra ^{T C} 05:13, 7 March 2017 (UTC)

If that's the case, then that is something I've seen done on some list articles and it does seem to work fairly well. There are few ways it can be done: (1) add the complete citation to the cell itself like is normally done or (2) add the full citation to the "References" section per WP:LDR and then use WP:REFNAME for the cells. Both would look the same to the reader, though option (2) might be easier to do since there are so many links cited throughout the article. I'm happy to help with this once the tables have been tweaked accordingly. It's quite a bit of gnome work, and dividing it up might be easier than trying to do it all at once. -- Marchjuly (talk) 05:19, 7 March 2017 (UTC)

I'll take care of it, don't worry. (In the process, I'm trying to trace down the original PDF sources as well.) Mélencron (talk) 05:58, 7 March 2017 (UTC)

You may find Template:In use helpful then because it can sometimes help avoid edit conflicts. Just a suggestion. -- Marchjuly (talk) 06:02, 7 March 2017 (UTC)

@Mélencron: that was what I had in mind as well, it makes it really easy later to update a reference to a new version without having to find it in the document (it may only confuse newbies .. but well, that is a tradeof). --Dirk Beetstra ^{T C} 06:11, 7 March 2017 (UTC)

I'm off for now – I'll update the article later. The only poll I can't track down is the Odoxa poll for 21–22 May 2015, so if anyone could trade down an archived copy/slideshare/etc., it'd be greatly appreciated. Mélencron (talk) 06:14, 7 March 2017 (UTC)

I think this might be what your looking for. There are quite a few archived versions, so you can pick the one which best suits fits. I usually try and pick the one closes to the date the link/source was added to the article. -- Marchjuly (talk) 06:25, 7 March 2017 (UTC)

Brought here from the request for input at WP:E&R, embedded links is the standard method of presenting opinion poll lists on Wikipedia. See e.g. Opinion polling for the Spanish general election, 2016, Opinion polling for the Italian general election, 2008, Opinion polling in the Canadian federal election, 2011, Opinion polling for the Polish parliamentary election, 2015. This should be applied here for consistency. Attempting to change a convention like this would probably require a centralised discussion. Number 57 17:57, 7 March 2017 (UTC)

I respectfully disagree - I think that the other pages should also be adapted. This is not in line with what our policies and guidelines prescribe. Consensus is not to have embedded external links, and to have one referencing style. --Dirk Beetstra ^{T C} 05:42, 8 March 2017 (UTC)

Too many names

Why not take Cheminade, Arthaud and Poutou out of the lists? Fair enough, they are polled. But as each scores 1% or less (usually less!).. why not use a little editorial discretion and leave them out of the tables for the sake of clarity? 83.115.123.190 (talk) 22:01, 3 March 2017 (UTC)

That isn't a valid reason. We don't exclude candidates just because they're polling low and removing them doesn't provide "clarity" – it does the opposite, in fact. Mélencron (talk) 22:18, 3 March 2017 (UTC)

I am sympathetic to this idea but I do think we need to be clear and consistent in whatever route we take and of course follow wiki policy reflect our reliable sources. Here are my thoughts:

1. It is quite normal to have an "Others" column in polling tables on wiki for parties/candidates that have such low polling figures but we do this based on the distinction the pollsters themselves make. And yes, if you look at the sources in the table they do clearly highlight the top 5 candidates, as well as just the top 2 (sometimes 3) candidates; Mélenchon, Hamon, Macron, Fillon and Le Pen. Meanwhile Cheminade, Arthaud, Poutou and Dupont-Aignan are just about lost in the data e.g. They are not in many/any of the graphs and they're faces do not appear in the polling articles, whereas the others do. Also within the media, these candidates are scarcely mentioned; so yes, I do see an argument for bundling these 4 candidates into others.

2. If we look at the main article, the infobox has 5 candidates there, to my mind there is an inconsistency here. Personally, I think there are merits to the argument that the main article should just have the 2/3 candidates that look likely to make it to the second round. However, I acknowledge and respect that consensus has been reached on that issue and that the 5 candidates shall remain in the infobox until 3 are eliminated. I do get that these 5 candidates are the ones with a more significant level of support and coverage, so I am inclined to think that we are giving the minor 4 candidates undue prominence in an inconsistent fashion; especially when we consider how the pollsters themselves publish their data.

It would be nice to see some constructive discussion here, I do not think Mélencron is right to shut down 83.115.123.190 as above. There are some valid points here and they should be discussed sensibly.

If it comes to a vote, currently mine is to create an others column and remove the 4 minor candidates individual columns. I will of course listen/read alternative/counter arguments and am open minded to changing that view depending on the information presented. 2407:7000:875B:C370:51AA:A4B9:4E5D:337B (talk) 09:21, 4 March 2017 (UTC)

There is no good reason to remove candidates simply because they're polling low. In this case again, I defer to precedent. Personally, I don't also see why we should treat infoboxes and polling articles similarly – each has their individual precedents. In the case of the main presidential election article, pretty much only the top-polling candidates were included in 2012, while the polling article included all candidates tested in polls (i.e., all candidates who were on the ballot). (I don't really see the IP's reason for why it might provide "clarity," either. The table... is pretty clear to me.) Mélencron (talk) 13:48, 4 March 2017 (UTC)

I also don't see the problem with listing all of the candidates for whom we have data. An infobox is by definition designed to present a summary, an article devoted to the polling for a particular election is by definition designed to go into detail. And while some might come to this article looking to see who's winning, others might come looking at how their preferred but minor candidate is doing, or researching the breadth of candidates on offer, or any other of a multitude of reasons. We shouldn't second-guess that. Unless there is a technical or space reason to not include data (and there's not) or a clarity reason (and I agree with Melencron that there is not that either), why would we omit potentially interesting data for the sake of it? Jdcooper (talk) 14:17, 4 March 2017 (UTC)

In response, I'd like to say that I have not made reference to the point about "adding clarity" as I also do not think this point is valid; I am in agreement with you on this one point but it is completely wrong of you to use this to shut the discussion down by making this point your only focus of discussion when other points have been raised that reference our sources. Both Mélencron and Jdcooper have failed to even discuss, never mind address the point about what our reliable sources are showing and how they represent the poll results, I therefore don't see how the above 2 comments add anything to the discussion.

Indeed, certainly in the case of Cheminade, the polls do not even publish data for him as can be reflected by his gappy polling column. I repeat, it is not unusual or without precedent for there to be an others column that includes minor candidates and if we look at how the sources present the poll results, this is not an unreasonable suggestion. Take [this poll] for instance;

1. Page 10 has a graph of the top 5 candidates (+Bayrou), the 4 minor candidates are not included in that graph.

2. Page 11 shows the faces of the top 5 candidates next to a pie chart. Where are the other 4 faces? They are just not there.

3. On to page 12, we have another graph, top 5 candidates all represented by their own line on the graph, the 4minor candidates are not mentioned.

4. Page 14 and 15 references the likely top 2 candidates for 2nd round voting, with page 16 showing how voters of the top 5 candidates would likely vote in the second round. Again, there is absolutely no mention of the other 4 candidates here.

5. Now here is the clincher. If you look at page 17, again you will see the top 5 candidates represented on the graph, along with a line titled "Aucun de ceux-là" (None of the above) and another line titled "Vous ne savez pas" (Don't know). This graph is actually telling us to bundle the minor candidates into an others column as this is what they have done. Don't know's are given more prominance than these 4 minor candidates.

6. Page 19, again we just see the 5 faces; where are the other 4? Are they given equal prominence in this polling data article? NO, why are we?

7. Page 20, again lists the 5 candidates, plus Nicolas DUPONT-AIGNAN; it seems to be the exception that he has been included here, as opposed to the rule.

8.Page 21, again just the 5 candidates. This is getting a bit repetative but hopefully this is sinking in.

9. Page 22, Again, just the 5 candidates plus a "None of the Above" line and "Don't knows", another example where a reliable source is telling us there should be an "others column".

10. You then have various pages of raw data throughout and often they do include the minor candidates but not always; eg page 32, so even in the raw data sheets, these 4 minor candidates are sometimes bundled in to a "Aucun de ceux-là" column.

Please can you look at this properly and address the points raised. The talk page is here to examine sources in this way, it is not here for lobbying for changes solely based on opinion. Come on guys if you have a credible counter argument, I want to hear it. 2407:7000:875B:C370:85D0:8F8B:A5F3:6427 (talk) 19:31, 4 March 2017 (UTC)

I don't think you get what page 17 is saying. It's Ifop giving survey-takers the option to select between five candidates and "don't know/none of the above" options about who they believe will win. Logically, none of the candidates who are polling low are even going to be included in this question.

And the reason that they don't list non-top five candidates in most of the other graphics? Because while those candidates might be significant for the topline, the sample size isn't large enough to justify a crosstab, so to speak (i.e., that because the sample size of the number of affirmative respondents for those candidate is so small, the data is likely to be extremely noisy... just as an example, this is what the "certainty of choice" graph looks like with all candidates). The reason that candidates outside of the top five aren't included is simply because variations on other questions won't be statistically significant... you might always see Poutou polling around 1%, but if you take a subsample of his supporters, you'd expect to see wild – and completely uninformative – swings in graphs regarding the certainty of choice among his supporters. They're exercising editorial discretion, so to speak. On the other hand, we're only using the topline numbers here, which aren't noisy at all.

tl;dr: why are they only including the top five? Because the data for other candidates, outside of the topline question, is extremely noisy and not informative as to the nature of their support. Mélencron (talk) 22:58, 4 March 2017 (UTC)

I understand perfectly well what page 17 is showing Mélencron. I could make the same argument you have just made about both Mélenchon and Hamon, given that not a single poll has put either candidate remotely close to making it in to the second round. Needless to say if you don't stand a chance of making it through to the second round, then you don't stand a chance of the presidency. However, I suggest we disist from going round in circles on this point.

Lets get back to the point here, we are meant to be reflecting our reliable sources here. Your above response amounts to original research, a violation of Wiki policy; please see Wikipedia:No original research. You are not suppose to come on to talk pages like this and talk about "noisy data" or questioning the sources themselves or seeking to explain why they have presented things the way they have. The point is they have presented them the way they have, they are a reliable source and we are not suppose tobe questioning their methods. If you want to have psephological debate, take it somewhere else; it has no business on a wikipedia talk page.

In asking the question "why are they only including the top five?" and providing your answer, you are acknowledging the fact that this is what the reliable sources are covering. The question of "Why?" and your answer don't actually have any relevance. We need to reflect the coverage of our reliable sources. 2407:7000:875B:C370:F16A:8333:ACFE:E52 (talk) 00:13, 5 March 2017 (UTC)

There is no applicability of WP:OR to talk pages. You know a good way to reflect the coverage of our reliable sources? Well, let's start by including the candidates they polled. I haven't heard any reason from you to exclude candidates for anything else, really, than "I don't like it." (I'm still not convinced you understand page 17: they literally only offer five candidates as options to that question – they aren't grouping the other candidates together.) They aren't polling only 5 candidates. Mélencron (talk) 00:20, 5 March 2017 (UTC)

As a matter of fact, you know that data for all candidates shown in this article are published in equality terms, right? That is, they don't make a difference between the "top five" and the others. Yeah, you make mention of other pages of the poll... yet those do not show vote figures, but other different questions/issues in which the "top five" are listed just to show their voters' opinions (with the others not appearing there because of their samples being too small). You bring Wikipedia:No original research here as an argument in your favour, when it actually goes against you: you're deliberately pretending to ignore the poll's main voting intention figures (where all candidates are shown equally) using other pages of the poll's pdf—related to other issues—to try to build an argument against including other candidates in a Wikipedia article showing vote figures. And you're forced to check around other questions/issues asked in the poll because in the main voting headlines all candidates are actually represented fairly equally. So, who's originally researching here?

Mélencron is right. They're not polling just 5 candidates, nor do they group the remaining ones together. My call is that showing all of them does offer a more representative and faithful picture of what opinion polls actually publish. Impru20 (talk) 00:32, 5 March 2017 (UTC)

It seems clear we need other people to look at this. Mélencron and Impru20 have got your heals dug in so deep on these issues that it is not possible to have a substance based discussion with you. I was very on the fence when 83.115.123.190 brought this up but the more you two conduct yourselves in this manner, the more I see the opposite side.

The reason I looked through the polls is to inform my opinion on this discussion, to see what the polls actually showed, to see how they treat the data. I am not ignoring the topline figures; I just haven't mentioned them as I think Mélencron over does this point when shutting down 83.115.123.190.

Impru20 how can you make out that Mélencron has not been engaging in original research? His psephological piece is written up above for all to see. To then accuse me of doing my own original research is a bit rich. All I have tried to do is actually look at our original sources and site what they show us. I have not sought to explain "why?" and do a normative analsys like Mélencron has; I have sited what the sources themselves show. This is not original research and I would ask you to withdraw that accusation. 2407:7000:875B:C370:F16A:8333:ACFE:E52 (talk) 00:46, 5 March 2017 (UTC)

To various anon IPs, I try not to get involved in the political side of wikipedia, and indeed I am following this page only because I am interested in the content of the updates. However, it is slightly annoying to read how you are treating these editors, who are doing fantastic work updating and improving the encyclopaedia, when all they are doing is explaining to you, quite politely, why and how these articles, and by extension most of the similar articles on wikipedia, are written. And all you are doing is accusing them of unspecified evil intentions. If you do not agree with other editors, please remain civil and calm. I can promise you that articles get improved quicker and easier that way.

On the substantive point, the reliable sources report data on the minor candidates. So we should include that data. Leaving it out constitutes giving undue weight to the main candidates. This article reports on what the polling says, and it's not our call to leave the lower candidates out. Jdcooper (talk) 03:14, 5 March 2017 (UTC)

Jdcooper, I am really disappointed in this comment. It seems you haven't properly read the whole text above because there is no truth to your accusations. I am not the one treating editors badly, I am the one being treated badly. Robust argument about substance, should not be interpreted as personal attacks, especially when I have gone out of my way to say things like Mélencron's contributions are "valid and valued" - how does fit the description of a personal attack? At no point have I attacked anyone's character. This is starting to seem like an extreme form of Safe-space, where by we are not aloud to challenge someones arguments/point as it might cause offence, or in this case actually be interpreted as a personal attack; completely ridiculous of course. And, no before the whining starts, I am not calling anyone ridiculous, that is also not a personal attack, I am calling the situation ridiculous and the fact that we are being discouraged from analysing the substance and presenting a counter argument: That is what merits ridicule here.

For you to site unspecified evil intentions is really unfair here and I would ask you to withdraw your comment. I have been really really clear in saying that I am not assuming bad faith and that I am not making any accusations. It's all there above (and in the section below), read it. I am very unhappy with how this has been handled and that accusations have been leveled against me, instead of the editors actually engaging in the substantive argument. Of the 10 or so points I raised above, there has only been an attempt to address one of them and even then that involved bringing in personal psephological analysis (original research), something that is prohibited on Wikipedia. When I call this out, they attempt to round and turn the tables on me. This is not acceptable and I think it is time that this was reported as I want this to move forward and for empty accusations to stop being thrown around. I have not accused anyone of anything, nor have I attacked anyone, though I feel like there has been an attempt to try and provoke me to do so (again for clarity, that's the way I feel - not an accusation). How do I report these accusations to some form of authority? 2407:7000:875B:C370:829:7DDA:70DD:7AEF (talk) 08:39, 5 March 2017 (UTC)

Alain Juppé

Alain Juppé has been added to the table for March as he faetured as an alternative option in one poll. It seems only one organisation has polled on this and it may be that this is a one off or that he is only polled a handful of times; we will have to wait and see. However, it is important to bare in mind that as things stand; not only is Juppé an undeclared candidate but he is also a defeated candidate. I realise that there is a theoretical route to him re-entering the race but as things stand this is remote and indeed theoretical.

In creating a whole extra column to report this poll, we are giving undue prominence to what is in effect hearsay and media speculation; something which Wikipedia should not involve its self with.

I agree that we should not ignore this poll but I think there is a better way of reporting it. For now at least, I think we should include it's sister poll (the one with Fillon) on its own with a note/asterix next to Fillon's percentage, with a footnote giving a full breakdown of the alternative poll and explanation. It may be the case that we need to do this for 2 or 3 polls: If we need to do this for 5 or 6 polls then perhaps we should include Juppé.

I reserve the right to change my view on this as the facts change but as things stand the fact is that we have one single poll, billed as a hypothetical alternative to another poll. Lets bare in mind that all other hypothetical alternatives (second round polling) have been minimized/reduced prominence within the article; Whereas this one we have given added prominence to, in the form of a whole extra column within current polling based on media speculation. 2407:7000:875B:C370:51AA:A4B9:4E5D:337B (talk) 09:35, 4 March 2017 (UTC)

I think it's still up until 17 March or so that candidates can be changed. Until not formally proclaimed, all candidates are still presumptive, and given the extent of Fillon's scandals it doesn't seem rare that at least one pollster has chosen to ask on Juppé in the event that Fillon withdraws. This is an article on opinion polling and it should list what polls say. Just as up until the end of 2016 there were multiple-scenario polls and none of these were shown separately, that shouldn't be the case here either since candidates are still not official yet. Juppé could still legally be appointed before the deadline. Impru20 (talk) 16:02, 4 March 2017 (UTC)

In any case, I'll be creating a new table once the list of official candidates is published in a couple weeks. Mélencron (talk) 16:19, 4 March 2017 (UTC)

That seems appropiate, indeed. Impru20 (talk) 21:21, 4 March 2017 (UTC)

I'm not sure either of you have got my point here. I haven't got a problem with us including the alternative poll; indeed it would be wrong not to include it. My question relates to how we include it. I maintain that it seems completely disproportionate for an encyclopedic article to give undue prominence by giving Juppe a column based on this one poll. I think there are a number of alternative ways we can represent this. I would value hearing the ideas/proposals and the constructive in put of others. 2407:7000:875B:C370:85D0:8F8B:A5F3:6427 (talk) 19:38, 4 March 2017 (UTC)

No, it isn't disproportionate. In the 2012–2016 period there have been candidates which may have been shown in maybe a few polls only and they've been added nonetheless. What was done in those cases was to limit the timespan of the tables to a few months/weeks only so that these only appeared at their appropiate time. The same will be done here, and once the definitive list of candidates is available, a new table will be created to list all polls into the election itself. There's no issue here. Impru20 (talk) 21:21, 4 March 2017 (UTC)

Your point about 2012 and 2016 is somewhat different, given that this time round we are talking about a candidate who lost the leadership election and has not declared a rival bid.

I do take the point that there is a theoretical path to the presidency for Juppe but there are a number of ifs and buts and it would be unprecedented for this to happen. As things stand we are talking about 1 poll that has come about as a result of media speculation; I do not see how it is proportional to create a whole extra column based on one alternative/sister poll; I really do not. All, I am saying is that we can maybe change the format slightly to highlight this poll without affecting the table for the whole month. I do not see how this is unreasonable.

I also think this comment from Mélencron is very telling: "In any case, I'll be creating a new table once the list of official candidates is published in a couple weeks. Mélencron (talk) 16:19, 4 March 2017 (UTC)" - It seems to assume some ownership over the article. I am not for a minute suggesting that Mélencron's contributions to the article are not valid or valued (they clearly are). However, it makes it difficult to assume good faith when comments like this are made; who says there will need to be a new table before the end of March? What's to say that exactly he same candidates will be running after nominations close? We do not know and there seems no logic in creating a new table for the sake of it just because Mélencron has decided that's what he/she wants to do. I am doing my best to assume good faith here and I want to make it clear that I am not accusing anyone of anything or indeed assuming any bad faith here; I am just going off of what has been said.

I was initially inclined to say lets wait a week and see what happens (see the non committal tone of my first comment) but this hardline No, No, No and pooh-poohing of every point that lays outside of your personally preferred format of the two of you is so unconstructive that I feel this needs to be challenged. Can this talk page not just be about discussing the substance of arguments; you know examining the facts and coming to a consensus; "No No No" and not addressing points that are made is really unhelpful. 2407:7000:875B:C370:F16A:8333:ACFE:E52 (talk) 00:34, 5 March 2017 (UTC)

Again, I'm referring to the creation of a table upon the publication of the final list of candidates on 18 March by the Constitutional Council, which means that a potentially different list of candidates will be considered by pollsters (e.g., possibly Jean Lassalle or others); again, I'm merely deferring to precedent in this case. Mélencron (talk) 00:36, 5 March 2017 (UTC)

Again, it is not yet clear whether or not to assume there will be a need for a new table or not come the 18th of March. It is important to point out that you are indeed comparing an apple with an orange when you say you are referring to precedent: The 2012 article has one table for oct-March was over 6months; here we have had one table per month since November, this is not the same as last time. We should only create a new table if one is called for, not just because nominations close. 2407:7000:875B:C370:F16A:8333:ACFE:E52 (talk) 00:54, 5 March 2017 (UTC)

...the October table was created because of a primary. In this list, the separation of tables in November 2016 was because of a primary; in December 2016 because Hollande bowed out and hence pollsters tested new scenarios; and in January 2017 because of a primary. I anticipate the possibility that that the present table may be changed if Cheminade or Poutou fail to secure the requisite signatures, another candidate manages to secure 500 parrainages, or Juppé decides to run and receives the requisite signatures and provides the necessary notification to the Constitutional Council by 18 March. Mélencron (talk) 01:21, 5 March 2017 (UTC)

It is? Many of the candidates shown throughout 2012 to 2016 were shown in polls even at times when they did not publicly announce their bids. Macron was even listed as a PS candidate. We're not talking about "ifs" or "what ifs" here, but about a poll that has indeed come out and which offers such an scenario. What Mélencron has said is just the precedent here: to create new tables when new sets of candidates are available, and collapse/hide the rest.

Anyway, I think you may actually be more respectful to others when you see that your views are unsupported. WP:NPA is a policy too, and it seems worrying to me that just because someone doesn't agree with you, you suddenly accuse the others of pooh-poohing and say that there is a "need" for this to be challenged. You proposed a change and no one supported you so far; you can't now come and disregard what others say just because they don't agree with you. Nobody's obligued to support your views if we don't agree with them, mind you. Impru20 (talk) 00:51, 5 March 2017 (UTC)

Impru20 you are being unreasonable in saying that I need to be more respectful and for siting WP:NPA policy. I have not made any personal attacks and I have gone out of my way to say things like Mélencron's contributions and edits are valued. I have been absolutely explicit in not making this personal; I have challenged the points, not the character of the person. I maintain that there has been a fair amount of pooh-poohing of a number of points; it is not a personal attack to say that. All this can be seen as above, neither yourself or Mélencron have addressed a number of the points I have raised, you have addressed the points you have raised; this is not engaging with the others argument (something at least I am trying to do). I really do think this needs the input of others; it's important to point out that it seems no one else has seen this discussion and therefore had a chance to engage with it yet. I would kindly ask that you withdraw your accusations about me as I clearly have not made any personal attacks or been disrespectful. I have made a robust challenge on substantive points, whilst maintaining respect of other contributors (read above - its all there). 2407:7000:875B:C370:F16A:8333:ACFE:E52 (talk) 01:07, 5 March 2017 (UTC)

I suggest that we desist from pursuing this discussion further and wait for outside opinions rather than litigating this. As of now, there exists no consensus for these proposed changes. (Might I suggest that you find a few articles to edit rather than merely spending all of your time on Wikipedia complaining about other people's edits? It's a better use of both your and my time.) Mélencron (talk) 01:15, 5 March 2017 (UTC)

The discussion around any substantive point ended when accusations started being thrown around after the psephology lesson, it is clear that without further contributors that this discussion cannot progress. No, I think this needs litigating, I am unhappy that accusations have been made about me when I have not done anything wrong, I want to be able to report this.

Furthermore, you seem to dislike my approach of taking things to a talk page to discuss a change before I just go ahead and do it. If you look at my opening comment, my approach is a constructive and open minded one, I am asking for the views of others; whereas you came on with your initial comment which sought to shut the argument down. I do indeed edit other pages on Wikipedia and I do not seem to have this same problem with other editors, mainly because they do not pooh-pooh your arguments, there is actual engagement usually. This article takes up more of my time but I am not going let that make me give up because of the principle. On talk pages, arguments should not just be pooh-poohed, there should be proper explanation and examination of each side, with constructive discussion (believe me, it is a lot faster than this). There is also the principle that Wikipedia should not be treated like a safe-space, people should not be protected from having their arguments challenged because it might hurt their feeling (obviously respect and civility should be maintained throughout - something which I have done). The accusations that have been made against me are clearly false and I want these lifted on principle. 2407:7000:875B:C370:829:7DDA:70DD:7AEF (talk) 08:57, 5 March 2017 (UTC)

Juppe has just declared that he will definitely not be a candidate in the election regardless of what happens with Fillon. MFlet1 (talk) 12:07, 6 March 2017 (UTC)

Which signals to me that he needs to be placed in "hypothetical" to avoid giving an undue indication that he could be a candidate. Maswimelleu (talk) 12:35, 6 March 2017 (UTC)

We included hypothetical scenarios (e.g. the 4-major-candidates scenario from BVA) in the original table – many of the old ones are already hypothetical scenarios, contingent on candidates winning a primary. Mélencron (talk) 13:06, 6 March 2017 (UTC)

Yes, but as I recall many of the second round polls were hidden in collapsed boxes once the scenarios they displayed were ruled out? I would argue Juppe is now sufficiently unlikely to be a candidate that he warrants going back into a collapsed box as a "hypothetical poll", even though the polls concerned refer to a first round. Maswimelleu (talk) 16:51, 6 March 2017 (UTC)

Now that Juppé has discarded himself out, it is all but certain that a new table should be done with the official list of candidates once the deadline is reached on 17 March. Then, the current table should be collapsed in the "older polls" section in order to hide Juppé scenarios. Impru20 (talk) 18:43, 6 March 2017 (UTC)

I almost didn't see this as it was tucked in before my below comment. Your solution doesn't really address the above points, it also relies on the need for a new table. If we reach a point when Juppe is the only difference in the candidates here then all that serves to do is further highlight that he should not have his own column for the March2017 table. It does seem that there are more contributors in favour of representing the 2 polls Juppe features in a different way, then not.v I think we should start making progress in deciding what that way should be. 2407:7000:875B:C393:B810:1CCB:78C9:B434 (talk) 07:20, 8 March 2017 (UTC)

Welcome to the conversation MFlet1 and Maswimelleu. As I said in my original post I can kind of see both sides on this one. On the one hand, we are now talking about 2 polls but on the other the candidate in question has not only lost a leadership contest but he has now also explicitly ruled himself out for good. I think we do need to be reporting these polls but I do not think we are doing so in the right way; that whole extra column is excessive. Given that these are sister polls to a poll we have included; I am starting to think what we should do is put a clickable icon next to Fillon's percentage for the sister polls that takes you down to a table with the full result that is in the hypothetical polling section (or elsewhere in the article - I am open minded as to where.) This way we are reporting the poll and it is clear to see that there was something different about the regular poll and readers can access the full info easily but we will also not be giving undue weight to media speculation about a candidate who lost a leadership election months ago and has further ruled himself out. Any Thoughts? 2407:7000:875B:C393:AD8F:490A:78DC:DA2A (talk) 17:35, 6 March 2017 (UTC)

Alternative proposal

Poll source	Fieldwork date	Sample size	Jacques Cheminade			Jean-Luc Mélenchon
			Cheminade SP	Arthaud LO	Poutou NPA	Mélenchon FI	Hamon PS	Macron EM	Fillon LR	Dupont-Aignan DLF	Le Pen FN
Ifop-Fiducial	3–7 Mar 2017	1,390	0%	0.5%	0.5%	11.5%	14%	25%	19%	3.5%	26%
Elabe	5–6 Mar 2017	1,000	–	0.5%	0.5%	12%	13.5%	25.5%	19%	3%	26%
OpinionWay	4–6 Mar 2017	1,559	–	0%	0%	10%	16%	25%	20%	3%	26%
Ifop-Fiducial	2–6 Mar 2017	1,381	0%	0.5%	0.5%	11.5%	13.5%	25.5%	19%	3%	26.5%
OpinionWay	3–5 Mar 2017	1,671	–	0%	1%	11%	15%	24%	19%	3%	27%
Ifop-Fiducial*	2–4 Mar 2017	1,822	0%	0.5%	0.5%	12%	14%	25.5%	18.5%	2.5%	26.5%
			0%	0.5%	0.5%	11.5%	13%	23%	**20%	3%	28.5%
Kantar Sofres	2–4 Mar 2017	1,027	<0.5%	1%	1%	12%	16%	25%	17%	2%	26%
			<0.5%	1%	1%	11%	13%	20%	**24.5%	2.5%	27%
Ifop-Fiducial	28 Feb–3 Mar 2017	1,383	0%	0.5%	0.5%	11%	14%	24.5%	20%	2.5%	27%
Odoxa	1–2 Mar 2017	907	–	0.5%	1%	10%	14%	27%	19%	3%	25.5%
			–	0.5%	1%	8%	11%	25%	**26.5%	4%	24%
BVA	28 Feb–2 Mar 2017	1,413	–	0.5%	1%	11.5%	15.5%	24%	19%	2.5%	26%
OpinionWay	28 Feb–2 Mar 2017	1,654	–	0%	1%	11%	15%	24%	19%	3%	27%
Elabe	28 Feb–2 Mar 2017	1,507	–	1%	1%	12.5%	12.5%	24%	19%	3%	27%
Ifop-Fiducial	27 Feb–2 Mar 2017	1,394	0%	0.5%	0.5%	11%	14.5%	24%	21%	2.5%	26%
OpinionWay	27 Feb–1 Mar 2017	1,639	–	0%	1%	11%	16%	23%	21%	3%	25%
Ifop-Fiducial	26 Feb–1 Mar 2017	1,392	0%	1%	0.5%	11%	14%	24%	21%	3%	25.5%
Ifop-Fiducial and OpinionWay polls in this table are rolling unless otherwise denoted by an asterisk (*); see methodology: Ifop-Fiducial, OpinionWay Continuous results can be viewed here: Ifop-Fiducial, OpinionWay. Historical data and demographic breakdowns can be viewed here: Ifop-Fiducial (**) Note: Alain Juppé was floated to replace Fillon as The Republican candidate as a result of the Penelope Fillon affair (Penelopegate), with some polls including Juppé based on this hypothesis. Juppé announced on 6 March that he would not be a candidate, regardless of what happened with Fillon.

This is what I am proposing, open to any suggestions/tweaks that will improve things. 2407:7000:875B:C393:B810:1CCB:78C9:B434 (talk) 08:14, 8 March 2017 (UTC)

This is a terrible suggestion. It's actually worse than not including the hypotheticals at all. It gives the impression that Fillon is leading in the polls when he isn't (and no, that footnote is just not good enough). To be perfectly honest I don't see any problem with what we currently have, despite all that's been said above. Rami R 08:43, 8 March 2017 (UTC)

The footnote does make it pretty clear that he is not in the lead. The ** are pretty clear and if you hover your mouse over the numbers you will see that there is a link to Juppe's name; so I do not see how you can argue this as there are 2 major flags to show this is not a business as usual poll. In both sets of polls there is only one republican candidate and they are all included in the republicans column. I really do not see how this is worse than the status quo that blows three alternative polls and media speculation out of all proportion. Like I say when you look back in 5 years time it will look very odd and even more out of proportion. I welcome any alternative suggestions for displaying this and constructive criticism to try and take us forward. 2407:7000:875B:C393:B810:1CCB:78C9:B434 (talk) 08:51, 8 March 2017 (UTC)

Well, it's worse because it's confusing. The alternative is to leave it as it's currently now, which is pretty clear and straightforward for everyone. Impru20 (talk) 10:48, 8 March 2017 (UTC)

Thanks for your proposal, but I can't sincerely find any advantage it could provide over the current scheme. Further, it gives the impression that it's Fillon leading the polls. The footnote, btw, is the same currently in place. So far, the only change you propose is to merely removing Juppé's column and give his numbers to Fillon, then pretend that with a mere footnote (which, anyway, shouldn't be this one) you explain to people that those numbers are actually Juppé's. This is rather chaotic and unclear, and will surely lead to confusion, so I don't support the idea. Impru20 (talk) 10:48, 8 March 2017 (UTC)

It is really inconsistent to use this argument when if you go to Opinion polling for the next United Kingdom general election, you will see that Plaid Cymru's polling figures regularly appear in the SNP's column with footnotes; they're not even the same party; is this not far more "chaotic and unclear"? You look at that percentage and have no idea how much of that support belongs to Plaid or the SNP, it's just bundled together which suggests it should be in the others column (a seperate argument). Doing this with Juppe for only 3 alternative/sister polls that Fillon does not deature in is by far more appropriate and in no way confusing when compared to the issue in the UK article.

You say you do not see any advantage over the current scheme; well several other's seem to agree that there is a problem with the current scheme that gives too much significance to 3 alternative polls based on 4days of media speculation.

The proposal does not "give" Juppe's number to Fillon, it gives it to the Republican's column; party contests are finished. Aside from the ** and the note, the polling figures have a link to Juppe's name. Alternatively, we could mirror what was done in the UK article with Plaid and put a link to the note, instead of the **; this would further undermine the consistency of your argument. Given that we would be doing this for only 3 instances, compared to countless over the period of 5 years, I would say you argument holds zero weight here.

Having looked through the edit log of the UK polling article, I know that you are a regular contributor to that article; it makes no sense to me that you use this argument as an objection here and not there: Bizarre indeed. In any case, there are still some accusations that you have yet to withdraw which undermines your whole involvement in any of these discussions. 2407:7000:875B:C393:55E1:F169:27C4:D6BB (talk) 17:34, 8 March 2017 (UTC)

While I don't think it's ok to combine Plaid Cymru and SNP, it does need to be said that there two key differences from our situation here: (1) Plaid Cymru is a minor party with around 1% support (contrast to ~20% for the Republicans), and (2) many pollsters report only combined numbers. Since we don't do original research here, we have to report what the pollsters state. These differences mean that the UK situation is not equivalent to the French one. Rami R 18:28, 8 March 2017 (UTC)

Well, the Opinion polling for the next United Kingdom general election article is a way different deal and not really equivalent.

1. "Plaid Cymru's polling figures regularly appear in the SNP's column with footnotes" because some pollsters (specifically, YouGov) automatically combine their data. So, it's obvious you can "have no idea how much of that support belongs to Plaid or the SNP", because those polls don't allow for it. And that's why they're shown together, because their data can't be separated. You want to merge Fillon and Juppé's columns into a single one under the basis that, somehow, having a Juppé column is giving him "undue prominence". As a first point, I can't even figure out where's the slightest similarity between both what happens between SNP and Plaid in the UK and your proposal.

2. What you speak about for the UK is referred to a parliamentary election, not a presidential election. Their mechanics are different, starting from the point that, in parliamentary elections, pollsters poll parties, not candidates; whereas in presidential elections, candidates are polled and NOT their parties.

3. It's not really frequent at all for polls showing different election hypotheses to be published in the UK. It is in France. So, in the UK they don't have this issue. For France, the way to go until now has been to have a different column for each candidate. You obviously don't see this in countries not polling candidates but parties.

So, no. In the UK they've no other choice but to do that because some polls just put SNP and Plaid data together. But even so, it's nowhere as unclear as you propose, since they don't have different election hypotheses that could mislead readers if not shown properly. Also, the fact that you make a comparison with a situation in another country which is not really comparable at all makes your own claim even more chaotic. What do you actually propose? Having what it's done in the UK opinion polling article, even when that's for a parliamentary election? Use for France the UK mechanics, even if they are way different?

Then, the columns are for candidates, not for parties. Yes, under your proposal you'd be giving Juppé's numbers to Fillon's column (even if you say you wouldn't), since the column is for Fillon, not for The Republicans as a party. This is a presidential election, not a parliamentary election. You should make an attempt to understand the difference.

"You say you do not see any advantage over the current scheme; well several other's seem to agree that there is a problem with the current scheme that gives too much significance to 3 alternative polls based on 4days of media speculation" Where's that? The only user stating his disagreement with Juppé being in the table was Maswimelleu, and what he proposed was to have Juppé in a table that was collapsed (which both me and Mélencron agree: we'd just collapse the current table after nominations close and the final set of candidates is available). And that's for Juppé in general, because for your specific proposal we've yet to see someone aside from you agreeing with it (with already two users openly against it–including myself–, which would rise to three if we count Mélencron's stance in favour of the current scheme).

As a result, I stand by my argument, and after seeing you put the UK opinion polling article as an example, the more convinced I am to stand against your proposal. I said it was chaotic and unclear. Now, I must add to all of it that you yourself seem to confuse what the columns do represent (parties or candidates) and the actual differences between different types of elections. Such a confusion is visible in this proposal of yours, and would be translated to others as well.

Btw, yes, I've frequently edited the UK polling article, but I've never spoken either in favour or against the way the currently show SNP and Plaid, mainly because that's the decision they seemingly chose through consensus. So, you'd have to try harder than to use the fact that I edit other articles as some sort of evidence that I support for those what you propose. Firstly, because I haven't participated in such a decision; and secondly, because those are not situations that are minimally comparable. And of course, because your own arguments would be discredited if what you're intending to convey to us is that you propose the UK model here, being a wholly different country/scheme/mechanics/etc and under a wholly different reasoning. Impru20 (talk) 18:40, 8 March 2017 (UTC)

New Table

I have reverted the creation of the new table for March-April for the following reasons:

1. This edit was too bold as this is a current and unresolved discussion point on this talk page.

2. We do not have a settled consensus on the current table, indeed we have 2 separate discussions about it; one to create an others column with the 4 minor candidates; the other to find an alternative way of showing the 2 hypothetical/alternative polls Alain Juppé featured in, instead of Francois Fillon. There was unpleasantness and false accusations directed at my self by a regular editor which is being ignored; to make this edit in the face of that and just assume the discussion is over is not acceptable, especially when other contributions have been made to the discussion.

3. It is too early, we are still in early March, hold your horses and don't get over excited. There is nothing to say for sure that we will need a new table once nominations close. At the last election we didn't have monthly tables, so I understand why that was seen as a natural breaking point but if all the same candidates are going forward, we don't need a new table for the sake of it.

Hopefully we are able to move forward in a constructive way with these discussions and find a solution; regular editors should not just plough on with what they think is right when a discussion has been raised and no solution found. 2407:7000:875B:C393:108B:2995:7121:D36A (talk) 17:20, 7 March 2017 (UTC)

"There is nothing to say for sure that we will need a new table once nominations close" ... well, maybe except for the fact that it's almost certain that Asselineau will appear on the ballot and hasn't appeared in any recent polls? If you look at past years, new tables were created because of primaries or new polling configurations; that's also the case this time around. Mélencron (talk) 17:26, 7 March 2017 (UTC)

If that is the case, then when nominations do actually close and we have the full slate of candidates, then that would be the appropriate time to create the new table: Not now. Wikipedia is not news, it is an encyclopedia. We can't be treating speculation in the media like this: "the fact that it's almost certain"; that statement is an oxymoron by the way: Something can not both be a fact and almost certain, as far as an encyclopedia is concerned it is one or the other.

I restate what I said previously; if it comes about that the candidates in the current table (aside from the Juppe discussion) and there are indeed extra candidates, then yes, it may well be appropriate to add an extra column for them. However, now is too early, lets wait and not get too excited. Also lets remember, we are here to report reliable sources; you yourself concede there is no polling for Asselineau; I am therefore unsure what business he has on a polling page in an encyclopedia (I fully accept this may very well change.) 2407:7000:875B:C393:108B:2995:7121:D36A (talk) 17:35, 7 March 2017 (UTC)

Yeah, I don't have any objection to not adding a new speculative table until, say, Asselineau or others are included in polls. Mélencron (talk) 17:55, 7 March 2017 (UTC)

The sole fact that Juppé won't be on the nominations listing is reason enough for creating a new table, since obviously, the definitive slate of candidates won't include him. It also may include others, but we do know for sure he won't be included.

This brings us to the Juppé discussion. It was proposed that, as an alternative, a separate table should be made including only the Juppé polls (so, meh, it's obvious a separate table including the Juppé polls will be there anyway, yes? It either being the current one—including these as regular polls with the rest of them—or another one). So we know that there will be at least two tables on this anyway, the only divergences being on the polls that those should include. But additionally, the current tables show opinion polls in a very linear and clear way, showing Juppé just at the time the re-emergence of his candidacy was possible due to the Fillon affair. Having a separate table for candidates once nominations officially close would both solve the Juppé issue and be a logical and consistent solution. The other way, you would have to separate the Juppé polls entirely into a separate table, depriving them of the context that could be offered by the comparison with other contemporary polls. Further, if we end up having other candidates appearing in opinion polls, then we would need to add an additional, third table.

Regardless of what happens to Asselineau, it's fairly obvious that a new table will be needed anyway. I don't have any issue that we wait for nominations to actually close before adding that table, though. Impru20 (talk) 17:58, 7 March 2017 (UTC)

Juppé should not be used as a reason to create a new table here; the above discussion is unresolved and there are more contributors in favour of an alternative representation of these alternative polls (polls that have come about purely as a result of media speculation - speculation that has been quashed by the non-candidate himself).

It is incorrect to suggest that creating a separate table for the two Juppé polls was the main/only suggestion for resolving the Juppé issue. Therefore, no, we do not know that there will be an additional table anyway.

Also your solution does not solve the issue of giving undue prominence to Juppé. Indeed the way you phrased your argument actually highlights this. You say "the re-emergence of his candidacy"; this gives added weight to the media speculation and so does including him in the table for the entire month. This is an encyclopedia, we are meant to try and be timeless; in 5 years time, if we look back at the article and see Juppé featuring in the polling table for the whole month it will be giving undue weight to what will be in reality a few days of historical media speculation played out in a couple of alternative polls.

It is wrong to see it as a full gone conclusion that there will need to be a new table before April; the only logical reason I can see is that another candidate from another party is added to the polling data. It is really important to point out that while nominations have not closed, the parties have had full and open candidates and have formally decided who there candidates are. The only way that can change is is if a candidate voluntarily absented themselves or there was a new selection process. So in this sense the candidacy of all the party candidates is formal and final: just as say Jeremy Corbyn is Labour's candidate for Prime Minister at the next UK election and Mmusi Maimane is the DA's candidate for President of South Africa at their next election despite an election not even having been called in either of those countries.

Basically, your whole argument rests on Juppe and therefore undermines its self. If the facts where to change and another candidate was to actually enter the race or if pollsters to start including another replacement candidate for Fillon (who doesn't then rule themselves out) in their polls regularly, then this would indeed be comparable to the early stages of the race and that candidate would warrant having their own column. As things stand your argument does not hold. 2407:7000:875B:C393:B810:1CCB:78C9:B434 (talk) 07:55, 8 March 2017 (UTC)

The thing is that there's no undue prominence for Juppé. You base all your argument in that he is given (somehow), some undue weight, but he actually is treated like every other alternative candidates have been in French opinion polling articles. Other candidates have been used at other times to create separate tables; if that must be done for Juppé, then it should, as it has been done for everybody. Juppé would be treated exactly the same as everyone else in his condition.

Also, when you say this -> "You say "the re-emergence of his candidacy"; this gives added weight to the media speculation and so does including him in the table for the entire month."

Maybe you are aware that Juppé had to actually go out and give a press conference to actually deny that he would stand as candidate, right? That's not just "media speculation". LR officials did propose Juppé as an alternative to Fillon, Fillon had to stage his rally in Trocadero to counter the internal party moves aimed to force him to quit and Juppé himself did come out to explain he would not be candidate regardless of what happened (but not because his hypothetical candidacy was false or speculation, but because he said that he "couldn't unite the party"). I think this is much more evidence that many alternative candidates opinion polls reported at their time (Peillon, Copé, Yade, Lassalle, Asselineau...), since Juppé's possible candidacy has been considered in the media by a greater deal than was those of other candidates back at their days. So I can't really see the "undue prominence", really: having Juppé in the table is just the listing of a seriously considered election hypothesis.

Just as party have had full and open candidates, they could change their candidates before nominations close (that's what the nominations period is there for). The example of Jeremy Corbyn you put is different from what happens here: Corbyn is Labour's official leader, and is so after fulfilling all deadlines and procedures in the party's leadership election (and you'd be also wrong to assume he would be the automatic candidate in next elections. He is listed in the Next United Kingdom general election by virtue of his condition as leader, which he officially has). Candidates for French presidential elections are not fully official until nominations close. The French presidential primaries are not a legally binding procedure that means that a candidate can't be removed by the party before the deadline and there was a genuine attempt at removing Fillon from his candidacy by his own party and replace him for another candidate (in fact, if it wasn't possible, then there would not even be that "media speculation" you speak about).

My argument does not rest on Juppé, but rather, on using for him the same treatment used for other candidate hypothesis (with much less media echoes) at other times. Furthermore, these seemingly unspoken rules you speak about now ("If the facts where to change and another candidate was to actually enter the race or if pollsters to start including another replacement candidate for Fillon (who doesn't then rule themselves out) in their polls regularly") are not what have been applied in the past. There have been candidates who have been added to the tables without actually entering the race nor without even polling at a regular basis, as I already pointed before.

Nonetheless, since you take yourself the free license to think for me and what I actually intend, I would say that it seems to be you who seem to base your argument on removing Juppé from the table for the sake of it. And your own "alternative proposal" shown above is a very clear evidence of this. It's obvious you want to get rid of Juppé for some reason, even if that doesn't meet with currently established practice for French opinion polling articles. And you say "because he would be given undue prominence", but fail to provide a valid reason that explains why Juppé is given undue prominence in this case and not other candidates back at their time. I think strong enough reasonings have been provided to counter your claim, so you'd do good in providing additional reasons that don't clash with current practice in these articles. Impru20 (talk) 10:44, 8 March 2017 (UTC)

Graph

Would it be possible to include when the graph was updated in the caption? Absolutelypuremilk (talk) 19:42, 7 March 2017 (UTC)

Yes, but I don't think we do that for any polling graphs... (if you open the link, it'll show you past versions/the time of the latest update). Mélencron (talk) 19:46, 7 March 2017 (UTC)

We do in the article Opinion polling for the next United Kingdom general election, I am sure we do in others too. 2407:7000:875B:C393:B810:1CCB:78C9:B434 (talk) 08:19, 8 March 2017 (UTC)

Well, I'm sure it does because it's Absolutelypuremilk's work... Mélencron (talk) 13:02, 8 March 2017 (UTC)

Yes, I update the date in the article in Opinion polling for the next United Kingdom general election when I update the graph. Would it be possible to do a similar thing here so that readers can see how old the graph is without having to click on the graph (which many people won't think to do). Absolutelypuremilk (talk) 14:21, 8 March 2017 (UTC)

There are now 11 official candidates

Now that the last batch of nominations has been published, the candidates who have received 500 or more valid nominations number 11: the ones on the current list, minus Alain Juppé, plus Jean Lassalle. Please can someone create a new table to include these and only these candidates. Thanks. Greengreen2 (talk) 14:56, 18 March 2017 (UTC)

Is Lassalle officially a MoDem candidate or is he simply a MoDem member who has chosen to run as an independent? The main page for this election seems to be describing him as an independent. It would be good if we could pick one for consistency. Maswimelleu (talk) 17:44, 18 March 2017 (UTC)

No, Lassalle is not a MoDem candidate. Ron 1987 (talk) 18:46, 18 March 2017 (UTC)

Ex-MoDem, apparently under the banner of Résistons ! – not too clear on this, he only popped up on my radar relatively recently when I was working on the main article. I've tentatively described him as an independent candidate/without a label (SE). Mélencron (talk) 21:49, 18 March 2017 (UTC)

His movement is called Résistons (source, another, and a third): I will change his label to this. Greengreen2 (talk) 10:10, 19 March 2017 (UTC)

On the other hand, a quick show there are far more sources referring to his candidacy as that of an independent than that of Résistons !, as he currently doesn't have any party affiliation since quitting MoDem last year. I can't find any evidence that Résistons is registered as a political party, either. Mélencron (talk) 13:34, 19 March 2017 (UTC)

Macron is also widely called an independent. I don't think "Résistons" is a political party that contests other elections. It's just the name of Lassalle's "movement". Macron's "En Marche" is also a "movement", even it's also a "party". The initials under the candidates' names should refer to their parties or movements, i.e. the brandnames used by their efforts. Political parties before in France have been vehicles for one individual. But leave the reference to "Sans Etiquette" if you wish, even if there's little justification for it. Wikipedia is full of this kind of thing. Greengreen2 (talk) 17:17, 19 March 2017 (UTC)

Describing Macron as an independent is a complete misnomer which I wish English press would stop using – it's only because he doesn't have the support of any of the major parties that he's been described as one. So far as I can tell, on the French Wikipedia he's only described as an independent as well (with no affiliation post-2016, when he left the MoDem). Searching around, it does appear that it does have a registered financing association – but apparently dropped the description of "party" by February 2017, oddly enough. Mélencron (talk) 17:55, 19 March 2017 (UTC)

We should just name the main organisation backing each candidate's campaign, and not worry about whether to call it a party, a movement, or a campaign organisation. Macron's "party" is called "En Marche" so as to have the same initials as he does.Greengreen2 (talk) 23:11, 19 March 2017 (UTC)

Can people take everything they know to the Lassalle article and enter it there? I moved a small amount of text over from the French version, but my grasp of the language isn't strong enough to faithfully translate it. We should explain what "Résistons" actually is and how his presidential bid came about if we're going to have a redirect pointing to his page. Maswimelleu (talk) 12:48, 20 March 2017 (UTC)

Tables rearrangement

On the tables of the primary candidates of this election the polls are arranged from newest to oldest. However on the head-to-head for the second round they are from oldest to newest. I suggest a rearrangement to maintain consistency throughout the page. I also suggest that the newest should come first so that its more user friendly for anyone that comes here just to get the latest polling. — Preceding unsigned comment added by 2.83.241.62 (talk) 21:11, 18 March 2017 (UTC)

This is on my to-do list... I'll see if I can get to it now. (fr.wikipedia similarly already has consensus on this issue, but haven't reordered the tables there, either) Mélencron (talk) 21:50, 18 March 2017 (UTC)

Maybe change Macron's color to pink? (like ffb6c1?)

In polls like IFOP, it has him as a lighter pink, not yellow :) Grahamdiedrich (talk) 15:25, 26 March 2017 (UTC)

Here's what I've got in terms of color usage:

•   Ifop-Fiducial
•   Kantar Sofres
•     Elabe
•   Paris Match (Ifop-Fiducial)
•   OpinionWay
•   BVA
•   Visactu (Ipsos)
•     Odoxa
•   Ipsos
•   franceinfo (Ipsos)
•   Harris

See also Template talk:En Marche!/meta/color Mélencron (talk) 15:47, 26 March 2017 (UTC)

Pink is the color for PS... Can we try to give Macron an unique color? Purple and pink are already fairly covered. Also keep in mind that, in maps, regions won by Macron would look like if won by Hamon (all of these pinkish shades have been traditionally used to depict the PS; in fact, they were used for the PS before Macron's bid. Check the 2012 election article). Yellow would be close to orange, too, but would be more distinct than any purple or pink shade. This, or revert back to the previous version. Impru20 (talk) 15:51, 26 March 2017 (UTC)

"Abstention/protest vote/spoilt vote", but what about "undecided"?

The French version of the article has one column for "abstention" and another for "undecided", in the sense of "hasn't decided to vote for a candidate", which includes "abstention".

I think we should follow suit. For example the "undecided" figure for the latest Ipsos poll is 41%, with "abstention" at 35%. Greengreen2 (talk) 15:37, 28 March 2017 (UTC)

That's not "undecided", that's the percentage unsure of their choice – i.e., it's among those who already indicated a preference in the poll. Mélencron (talk) 15:41, 28 March 2017 (UTC)

A row for important events, such as the validation of nominations and the TV debates?

The French version of the article includes this information in the table too. Again, we should follow suit. (I would implement both this and the above-proposed change if I knew how or could learn how within a reasonably short time.) Greengreen2 (talk) 15:37, 28 March 2017 (UTC)

This was already discussed above and there was no clear consensus to do so. Mélencron (talk) 15:42, 28 March 2017 (UTC)