Talk:Opinion polling for the 2018 Italian general election/Archive 1

Archive 1

Collapsible v. collapsed

@User:Impru20: Can you explain this? To be clear, "But have you checked how incredibly long the article is without the collapside option? How's it that it is not needed? Of course it is" is definitely not an explanation. Your arbitrary preference is not more relevant than mine. Editors of much more widely visited articles such as Nationwide opinion polling for the United States presidential election, 2016, Nationwide opinion polling for the Republican Party 2016 presidential primaries and Nationwide opinion polling for the Democratic Party 2016 presidential primaries do not share your concerns. --Checco (talk) 12:18, 21 September 2015 (UTC)

The reasons for keeping the "collapsible collapsed" option in the other article was that without it, the article was extremely long. That reason has not disappeared here, so it does not make sense for you to remove it. The article is way too long without that option, and users do not need to have to scroll all the way down the article to check a specific poll in 2013 or 2014. With the collapsible option they can just go to the specific table and do it. Only the current year table is left uncollapsed because of it being showing the most recent polls.

Uh... have you checked the fact that the Nationwide opinion polling for the United States presidential election, 2016 has collapsible collapsed tables for 2013 and 2014? Also, can you explain where editors of those article do not share my concerns?

In Talk:Nationwide opinion polling for the Republican Party 2016 presidential primaries an user just commented on the issue of collapsing the tables due to the article getting too long. Also, it's something that has been done too on the Nationwide opinion polling for the Republican Party 2012 presidential primaries. So it would seem other people do share my concerns.

So then, where's the arbitrary preference? Impru20 (talk) 18:39, 21 September 2015 (UTC)

You have a few points here. I have to admit. However, I don't see any problems with articles, comprising lists, becoming too long and I still think that is just your personal and arbitrary preference. I don't want to loose much time with such an issue, which is definitely of little importance, thus, while I always disagreed and I will forever disagree with you on it, let's go with it, unless other users agree with me. --Checco (talk) 07:20, 22 September 2015 (UTC)

Polls: dates and sources

Hello Nick.mon, I'm answering here to your question because the issue is of general interest. You can easily find the official dates of opinion polls at the official website of the Italian government (this was clearly explained at Opinion polling for the Italian general election, 2013—we should probably explain it also in this article) and, for SWG pols, at the bottom of their graphics. This said, it is OK to use ScenariPolitici.com or TermomentoPolitico.it as links to poll results (be aware that, regarding EMG's polls, only TermometroPolitico.it usually publishes the data on run-offs). --Checco (talk) 06:32, 6 October 2015 (UTC)

Ps: Polls are usually taken on the same days every week. Here is the latest EMG poll with run-offs from TermometroPolitico.it.

Ok thank you very much!! -- Nick.mon (talk) 07:24, 6 October 2015 (UTC)

@User:Nick.mon: It looks like two EMG polls (referred to here) are not included in the list. Have a nice Sunday! --Checco (talk) 06:21, 3 April 2016 (UTC)

Thank you Checco, I didn't notice that polls. Good evening! -- Nick.mon (talk) 20:56, 3 April 2016 (UTC)

Update needed

@User:Nick.mon: What has happended to this article? It is no longer up to date. I understand that you are now quite interested by other issues, notably including the constitutional referendum, but this article needs constant update, otherwise it is pointless. --Checco (talk) 08:55, 23 November 2016 (UTC)

Hi Checco, you are right but I'm the only one who update both this page and the one of the constitutional referendum; I will be glad if you or Impru20, could help me on this page. Anyway I will try to update this article as soon as I can :) -- Nick.mon (talk) 14:14, 23 November 2016 (UTC)

I know that you are the only one, but, if you remember, I was long sceptical on the list itself (see Talk:Next Italian general election#Polls' section needs constant update) because of the few users willing to contribute to it (I was not particularly willing to contribute at the time and I have even fewer time now). I hope that you and other users, possibly User:Impru20, will be able to continue updating the list, otherwise let's abandon the article. --Checco (talk) 05:46, 24 November 2016 (UTC)

Yes I remember very well that you were so sceptical about this page. In few days I hope to update the page. -- Nick.mon (talk) 14:51, 24 November 2016 (UTC)

It looks like some or several polls of the last two monts are not included in the list. Can you, Nick, check it? --Checco (talk) 14:24, 8 December 2016 (UTC)

Ok I will start now! By now will update the first round, and then maybe, I will update the second round, even if I think that it will be removed. -- Nick.mon (talk) 16:31, 8 December 2016 (UTC)

Lorien Poll

Hi. I edited the article a couple of days ago adding a poll which hadn't been included with this source: http://www.electograph.com/p/electomonitor.html#Italy I only wanted to know if I shouldn't edit the page again with a similar source, because someone has deleted that poll although Lorien ones had been included until November 2016.5.34.154.217 (talk) 16:10, 28 February 2017 (UTC)

I have just found another two sources which confirm this poll:

http://www.termometropolitico.it/1246853_sondaggi-elettorali-centrosinistra-unito-raccoglierebbe-33-dei-voti-secondo-lorien.html

http://www.centrometeoitaliano.it/attualita/sondaggi-elettorali-oggi-27-febbraio-2017-nuovo-partito-politico-di-sinistra-al-9-48415/

I'm going to edit the article again. Please, ignore these two messages if there isn't any problem after this new edition.5.34.154.217 (talk) 16:15, 28 February 2017 (UTC)

Sorry I deleted it accidentally :) -- Nick.mon (talk) 07:56, 1 March 2017 (UTC)

No problem. :D 5.34.154.217 (talk) 16:46, 1 March 2017 (UTC)

I would also add the DP's results before the launch of the party (there are a few cases, like this). And what about the Progressive Camp? And what about the Union of the Centre, which is no longer in alliance with th New Centre-Right? Finally, the latest Piepoli poll links to a dead link. --Checco (talk) 08:07, 1 March 2017 (UTC)

I agree about adding DP's results before the foundation of the party. The Progressive Camp appears in only one poll, I believe, maybe we could wait some days and see what will happens. About NCD–UDC we could leave only NCD. -- Nick.mon (talk) 08:41, 1 March 2017 (UTC)

OK, I agree with you on everything: DP (include all results), CP (let's wait) and NCD–UDC (replace AP with NCD). --Checco (talk) 08:51, 1 March 2017 (UTC)

As three opinion polls (two SWG and one Piepoli) have so far included the CP, I decided to include it in the template. Please revert my edit if you think it is still better to wait. Moreover, I am not sure about what the CP will be (pressure group? party? coalition?). --Checco (talk) 23:10, 2 March 2017 (UTC)

Scenari Politici Poll (only left-wing parties)

Hi. I only wanted to comment that there is a new poll (here) which only includes different scenarios of the new left-wing parties (MDP and CP) and SI. Should we create a new section in the article for this kind of polls? 5.34.154.217 (talk) 20:04, 13 March 2017 (UTC)

I frankly don't know, but I would not include it in the article, probably. --Checco (talk) 05:14, 14 March 2017 (UTC)

Hi again. As there is a new IPR Marketing poll which has included many alternative scenarios depending on the potential left-wing list's (MDP+SI+CP) leader, I think we could create the new section I mentioned some time ago if there are more polls that makes these scenarios. Demopolis also has done another scenario including a right-wing coalition (LN+FdI) and that left-wing one (as you can see in the article). What do you think? 5.34.154.217 (talk) 11:03, 8 June 2017 (UTC)

Margin of error and/or sample size

Since we are collecting the data, I think it is also important to collect overall margin of error of the survey and/or the sample size. These are declared for each poll commissioned and published on http://www.sondaggipoliticoelettorali.it/GestioneSondaggio.aspx. Knowing the population size it is possible to derive the margin of error of the survey from the sample size (and vice versa), so to publish both is redundant but probably more transparent for the reader. If I have to vote to publish only one (for compactness), I would vote to publish the margin of error, which can help in evaluating the significance of the survey results.

Why should we publish the margin of error? Because the difference between parties and coalition is often within the margin of error. For example, the average support for the Democratic Party in the October-November period has been 26.2% and for the Five Star Movement 27%. A survey with 800 interviewees has a margin of error of +/- 3.5%. --FrBailo (talk) 02:00, 21 November 2017 (UTC)

Sample size alone should be sufficient, given that margin of error is directly related to it (and that it's what's used on most polling articles). I'm not sure how easy it would be to find them retroactively, however – the Icelandic election article took me maybe 15 or 20 minutes to fix, but that was only with a few dozen polls; this list is massive. Mélencron (talk) 03:01, 21 November 2017 (UTC)

But the margin of error is probably more useful to the reader: it helps in understanding the confidence given by the survey figures. Find them is not a problem: they are all clearly indicated on the website I mentioned. --FrBailo (talk) 19:42, 21 November 2017 (UTC)

It's often misinterpreted and never used on election polling articles, while sample size is. I don't see a compelling reason to do so differently here. In any case, it might be useful to seek consensus for this change first as it's a fairly notable one, given it hasn't been used on previous Italian election articles. Mélencron (talk) 19:52, 21 November 2017 (UTC)

Why misinterpreted? I think it is actually easy to understand, it is expressed in the same unit of the other figures. The sample size can help the reader to compare across surveys (intuitively a bigger sample indicates a better survey), so it is of relative more than of absolute importance. Instead, the margin of error is both of absolute and relative importance. It can be used to compare different surveys but also to correctly interpret the results. For example, in the 17-19 November survey by EMG, the lead is 2.4 and the margin of error is 2.5. Information on the margin of error clearly add to the interpretation of the results more than the information on the sample size (800). --FrBailo (talk) 00:16, 22 November 2017 (UTC)

FrBalio, if you want to calculate the margin of error, the formula is 1.96 * sqrt(p*(1-p)/n). Thus for the example you gave above it would be closer to 3.1% rather than 3.5%: 1.96 * sqrt(.27*(1-.27)/800) = 0.031--Gciriani (talk) 13:21, 22 November 2017 (UTC)

Chart Moving Average

I was checking the average displayed by the chart, which has a caption of 60-day moving average. At first glance it seems more like a 30-day moving average, rather than a 60-day one. Am I missing something?--Gciriani (talk) 13:25, 22 November 2017 (UTC)

You're right. I corrected it. --FrBailo (talk) 20:02, 23 November 2017 (UTC)

Non representative polls

If there are no objections, I will exclude the SWG poll of 13-15 November. This is a poll conducted on an extremely small sample of 221 voters, which targeted people living at the periphery of metropolitan areas. For a comparison, the polling error was 5.7% for the PD vs 2.2% for the same party in the SWG poll of 27-29 November with 1500 voters. There may be more polls that were not representative. As a minimum we should add an asterisk and a note saying Non representative.--Gciriani (talk) 13:07, 2 December 2017 (UTC)

That'd constitute OR, however. Mélencron (talk) 15:48, 2 December 2017 (UTC)

Mélencron I do not understand your comment. Could you please clarify?--Gciriani (talk) 18:26, 2 December 2017 (UTC)

It'd constitute original research to include such a note. If no reporting on the poll, or the poll itself, makes the claim that it's an unrepresentative poll, then such a note shouldn't be included on the Wikipedia article. The obvious way to indicate this would be to include the sample size of polls in this article to show that this one has a particularly small sample and therefore likely to be less accurate – but as I've mentioned before, that'd take a lot of work at this point with hundreds of polls since 2013. Mélencron (talk) 18:37, 2 December 2017 (UTC)

I recall there were standards for opinion polls in Wikipedia. I don't remember in what article I saw them though. Perhaps it was pollsters belonging to an official pollster association.--Gciriani (talk) 03:34, 5 December 2017 (UTC)

Yeah, I've made that argument (unsuccessfully) in the past as well when it came to French polling. The discussion that comes to mind for me right now is this one, though. Mélencron (talk) 03:51, 5 December 2017 (UTC)

New table

I saw that some users changed the table regarding 2018 polls. In my view we could keep the current one, which is more coherent with the previous years, but maybe it's a bit difficult to understand the coalitions. So what do you think about this:

Date	Polling firm	Centre-left				Centre-right				M5S	LeU	Others	Lead
		PD	+E	I	CP	FI	LN	FdI	NcI

Thanks. -- Nick.mon (talk) 17:11, 9 January 2018 (UTC)

Great, good idea, approved!

And I would add PaP, CasaPound, FN-FT and others, provided that they are cited at least in one poll. There will be time to do that.

Cheers, --Checco (talk) 18:01, 9 January 2018 (UTC)

Yes, if they'll appear in the next opinion polls I will do it, but by now I don't see any poll with them... -- Nick.mon (talk) 18:20, 9 January 2018 (UTC)

That's a very good idea for the table. It looks very more comprehensive for users who doesn't know too much about Italian politics and their coalitions. Regards. --Sfs90 (talk) 18:53, 9 January 2018 (UTC)

I think we should do it as before, because parts of centre-right and centre-left are among the others, the camps are listed below again and the parties of the camps stand directly above it. Braganza (talk) 21:20, 9 January 2018 (UTC)

In fact I think that if we'll use "my" solution, we can remove the table regarding parties and coalitions. However if the majority wants to keep the previous version we can do it, but I think that the solution that I've proposed is quite clear and simple. -- Nick.mon (talk) 11:38, 10 January 2018 (UTC)

I support Nick.mon's solution, but not the removal of the table regarding parties and coalitions. The latter is part of the explanation on how the parties have evolved and it is quite useful for outsiders. --Checco (talk) 13:21, 10 January 2018 (UTC)

Sample size

With the new table for 2018, and the election approaching, I think it's time and it's important to insert the sample size, as it is done for most other election polls in Wikipedia--Gciriani (talk) 23:07, 9 January 2018 (UTC)

I added the sample size twice, and both time it was reversed, without an explanation the first time, and with the comment we never insert it, the second time. It's not clear who we are, since this is Wikipedia, and there are plenty of example where this or equivalent information is added. See Next UK general election, UK 2017, UK 2015, US nationwide 2016, US statewide, 2016. Therefore without entering in a reversal war, please give valid reasons before reversing a well established piece of information.--Gciriani (talk) 13:37, 10 January 2018 (UTC)

Sorry I was the one who reverted your edits... in my view we can keep the table without the sample size, but if you think that it can be useful, ok let's add it :) -- Nick.mon (talk) 15:53, 10 January 2018 (UTC)