Talk:Opinion polling for the 2022 Italian general election/Archive 1

Archive 1

Archive 2

Archive 3

Archive 4

Coalition polls

Some of the polls (e.g. Tecnè) do not contain specific data on coalition vote nor on smaller parties like NcI and CP which are part of bigger coalitions. Does it make sense, in this case, to report an estimate for the coalition vote? Ritchie92 (talk) 09:39, 6 April 2018 (UTC)

No, it really does not make sense to estimate coalition votes in my opinion. This is mainly because these coalition are informal and are only for the 2018 election which means they are most likely going to change before the next election. JDuggan101 (talk) 10:00, 6 April 2018 (UTC)

I would actually remove the section on coalition polls, entirely. --Checco (talk) 10:46, 6 April 2018 (UTC)

I completely agree with Checco. JDuggan101 (talk) 11:02, 6 April 2018 (UTC)

Some pollsters still give their data for the coalition vote, in this case I think it's meaningful data and it might be even useful. Let's wait for the government formation, when probably the coalitions will have some changes. In the meantime I would add to the Coalition vote table only polls that give explicitly the coalition data . Ritchie92 (talk) 11:37, 6 April 2018 (UTC)

I ask to create a visual chart

I ask to create a visual chart — Preceding unsigned comment added by 62.101.0.30 (talk) 08:51, 18 April 2018 (UTC)

I seconded that. this page should have a graph of the polls. Leftwinguy92 (talk) 12:15, 18 May 2018 (UTC)

I’ve asked Impru20 to do it. -- Nick.mon (talk) 12:27, 18 May 2018 (UTC)

Y Done. :) Impru20^talk 13:06, 18 May 2018 (UTC)

Sortable

It honestly surprises me this even has to be a conversation, but I think the opinion polling charts should be sortable. This would be useful, for example, if I want to see what poll shows a certain party doing the best, or what one has the largest sample size. Edit: If I get no response today, I'm putting it back on. 2 yea 0 nay Alex of Canada (talk) 22:50, 15 June 2018 (UTC)

Agree. At least the polling firm column should be sortable so that all polls from the same polling form can be seen together. Please make sure that the polling dates don't get scrambled when you do sort by firm. The person who initially reversed your edit did not even complete his thought in the comments section. I cannot think of a good reason NOT to have sortable charts.Juve2000 (talk) 23:07, 15 June 2018 (UTC)

One problem was that the table was suddenly much wider, and the party abbreviations weren't centred properly, which all made it look a bit ugly. I don't think the tables are sortable in most similar articles, except for polling company and sample size in a few cases. Also, I think it would be a good idea to wait at least three days or so for some sort of consensus to be reached, and preferably get a different user to make the edit. Speed74 (talk) 08:14, 16 June 2018 (UTC)

I agree with you on the usefulness of the sortable columns. The reason why I undid your edit (I'm sorry but I made a mistake in the comment section, I just realised) was that the sorting algorithm didn't work properly with the numbers in the sample size (somehow it showed me 800 after 1,200 maybe because of the comma), nor with the dates (i.e. it put the dates in order of day and alphabetically, so 1 Mar came after 1 Apr). It obviously works well with the polling firms column. Maybe somebody more expert than me on this can solve this problem. Ritchie92 (talk) 07:41, 20 June 2018 (UTC)

Coalition

Not being a resident of Italy, I do not follow Italian politics closely. However I am puzzled about the Coalition section on this page. Do the shown coalitions still make sense? Lega and M5S have joined in a government, and that is not shown as a coalition. — Preceding unsigned comment added by 212.93.55.132 (talk) 19:22, 18 June 2018 (UTC)

This topic is discussed here. In my opinion, since many pollsters still give results also for the coalition votes, it oculd still be interesting to report these numbers. Secondly, officially the centre-right coalition is still such (especially in local elections) and it's not so obvious that in the future elections (in the near or far future) they won't run together.Ritchie92 (talk) 07:48, 20 June 2018 (UTC)

Opposition parties

Since the opposition parties estimation is just obtained by (100% - gov't parties) I find the column a bit useless, should we remove it? Or at least, to make it significant, we could consider Parliamentary "opposition" only the parties that voted against in the confidence vote. Ritchie92 (talk) 10:28, 24 June 2018 (UTC)

Regarding the recent changes, I think it's useful to have a column showing M5S and LN's combined support at a glance, and the "opposition parties" column may come in handy if support for the government drops below 50%, as in that case the opposition party column would be highlighted in some sort of reddish colour. I agree that adding an extra "lead" column is a bit much, and with CPI being excluded from almost all polls, that column being excluded is also not a bad idea, to keep the number of columns under control. I also agree that we should be cautious about adding too many events into the table. Speed74 (talk) 20:20, 24 June 2018 (UTC)

Well, probably having a column that shows M5S and LN together could be useful, but not all the parties in the "Opposition" column are against the government. Moreover we should consider only parties which have some seats in the Parliament and thus we had to remove PaP, CPI and also the "Others" from the sums; I think it could be a bit difficult. Maybe we could insert some polls about the government's support among Italian citizens. -- Nick.mon (talk) 20:47, 24 June 2018 (UTC)

@Nick.mon: The tag "Others" now feels much better, but I disagree with the presence of another "Lead" column. About your final suggestion, the polls about government support are a bit off-topic and would be part of a different article (or a new section on this article), because in principle support for the government is not linked to the voting intentions in Parliamentary elections. For example, in the case of the US there is a different page for the opinion polling for the Donald Trump administration. Ritchie92 (talk) 22:46, 24 June 2018 (UTC)

Agree. The current scheme has the issue that it just calculates "opposition parties" as 100 minus LN+M5S, which is absurd. Not everyone who is not LN and M5S are opposition parties. What about parties which have no parliamentary representation? And those who abstained and may support the government in some issues throughout its term? Opposition parties are not all allied between themselves, so it would be misleading. Impru20^talk 21:01, 24 June 2018 (UTC)

Io ho inserito il "Lead government vs opposition" come si fa in tutte le altre pagine dei sondaggi: precisamente io ho fatto in questa pagina quello che tutti voi altri avete fatto in UK, Danish, Belgian, Finnish e altre pagine "opinion polling for the next general election" e quindi non togliete la mia modifica perchè essa è giusta; qui invece tutti voi altri state facendo il contrario senza alcun motivo, perciò non annullate la mia modifica: per favore, datemi retta. Quindi siete tutti voi che in questa discussione state sbagliando (altrimenti vorrebbe dire che sono stati tutti gli altri a sbagliare in tutte le altre pagine, il che ovviamente sarebbe assurdo). --151.67.93.14 (talk) 22:57, 24 June 2018 (UTC)

The pages that you cited (except the Belgian one) contain, as this page does, a Lead column for the votes to the single parties and not for the Government coalition vs Others.Ritchie92 (talk) 23:08, 24 June 2018 (UTC)

@151.67.93.14: This is the English Wikipedia, so first of all I would please ask you to write in English here.

It is not true what you say about "other polling articles":

• UK polling articles do not use a "Govt vs Opp" scheme of any kind. Plus, it only includes leadership replacement and elections as events (and even those are hotly disputed at times). Nothing else.
• Danish polling articles do not show "Govt vs Opp" but the Red and Blue coalitions, which would be akin to the Italian centre-left and centre-right coalitions. They show no events.
• Finnish and Belgian do show a "Govt vs Opp" scheme, but only Finnish articles show events (and just as UK, only elections and leadership replacements). These are an exception, though.

Then, a vast majority of opinion polling articles in Wikipedia show no events or "Govt/Opp" at all: Spanish, Greek, German, French, Norwegian, Swedish, Czech, Hungarian, Swiss, Maltese, Romanian, Bulgarian, Polish, Latvian... so, it is not true that this "Govt/Opp" table or the inclusion of events is done in "all other polling articles". The use of these is an exception, actually, not the general rule.

I would also ask you to stop edit warring. You have already broken the three-revert rule. Impru20^talk 23:12, 24 June 2018 (UTC)

Questa modifica dimostra che il mio parere non conta nulla visto che avete deciso voi da soli senza neanche accettare un parere contrario. Complimenti vivissimi e almeno cercate di non rimuovere gli "events" (presenti in Opinion polling for the next United Kingdom general election) per un motivo: quando il M5S si spaccherà in due pezzi (come accaduto in Finlandia con il partito populista), allora vedrete sorprese (e magari un cambio di governo...) e quindi scoprirete che l'Italia è (anzi, sarà) come il Belgio dove la coalizione di governo sarà composta da tanti partiti... aspettate ancora poco tempo e vedrete un governo tra un pezzo del M5S - l'altro pezzo, quello di sinistra, andrà all'opposizione pur di non sostenere la politica contro i migranti voluta da Salvini - e un pezzo del centrodestra, cioè Lega e Fratelli d'Italia con un segmento di FI lasciando l'altro segmento di FI all'opposizione insieme a Noi con l'Italia (e ai partiti di sinistra e centrosinistra). Inoltre la storia d'Italia non esclude la ripetizione di ribaltoni verso metà legislatura (chi era all'opposizione entra dentro il governo e viceversa). --151.67.93.14 (talk) 23:18, 24 June 2018 (UTC)

The political system of Italy is near to Belgium. I don't speak very well, sorry... events in the Italian politics are very important (wait and see... the M5S will broke: the left-wing pro-ONG Roberto Fico is against Luigi Di Maio and his political friendship with the anti-ONG Matteo Salvini. The government will broke like Finland and the next Italian goverment will be very similiar to Belgium as political multi-party fragmentation). Wait and see --151.67.93.14 (talk) 23:24, 24 June 2018 (UTC)

@151.67.93.14: Events at Opinion polling for the next United Kingdom general election are the result of a specific local consensus there, and they are limited to elections and leadership changes. Events are not included in most polling articles in the English Wikipedia, and none of the events you have added to this table would have been added to the UK polling article, so do not try to impose your particular vision of the article's layout to others, because it is obvious that your changes are being hotly contested by other users. Also, from what I can translate from your comment, your whole editing is based on original research: namely, your personal assumptions on how Italian politics work. You have engaged in edit warring already, which means that you could be blocked as a result. Impru20^talk 23:24, 24 June 2018 (UTC)

Many thanks: in polling for the next United Kingdom general election I read "English local and mayoral elections" and now I write "Italian local and mayoral election". 10 June 2018 is first round, 24 June 2018 is second round. I will write also Regional election. It is all ok for you? You write that "they are limited to elections and leadership changes" and I will wrote Maurizio Martina nomination as new Secretary of PD after Matteo Renzi's resignment. It is ok for you? Many thanks for your attention. --151.67.93.14 (talk) 23:34, 24 June 2018 (UTC)

I will repeat it again for you because it seems you have not understood it correctly: events are not encouraged in opinion polling articles. Firstly because their addition can constitute WP:NPOV, as it should not be up to any of us to decide which events are relevant for polling trends and which ones are not. In Opinion polling for the next United Kingdom general election they are used under strict lines as a result of a local consensus, which is reviewed each short time due to disagreements over events in the table. There is not such a consensus here. Events should not be in the table, but if anything, their inclusion should require a previous consensus here. You should remove these from the table. Impru20^talk 23:52, 24 June 2018 (UTC)

Ehm... in Opinion pooling for the next United Kingdom general election I read "Gerard Batten officially becomes leader of the UK Independence Party" and "Vince Cable becomes leader of the Liberal Democrats". What is the difference with Maurizio Martina? Goodnight --151.67.93.14 (talk) 00:43, 25 June 2018 (UTC)

@151.67.93.14: It seems like you haven't understood the replies of Impru20. There needs to be consensus among the users in order for you to make such a change in the page, especially because it seems that not everybody at all agrees with your edits. I suggest you to start another discussion topic here, if you want. Also I don't understand why are you writing a huge amount of useless political details in the event descriptions. This is not a time-line of Italian politics, this is a list of opinion polls. -- Ritchie92 (talk) 06:57, 25 June 2018 (UTC)

Chart doesn't look good

I know it's a kind of trivial issue, but the opinion polling chart does NOT look good. There are random spikes all over it, and it's hard to see where the line actually is. Alex of Canada (talk) 08:29, 22 July 2018 (UTC)

In my opinion it would gain in visibility if the dots were faded compared to the line, like in the swedish poll chart https://en.wikipedia.org/wiki/Swedish_general_election,_2018 but I still find it readable as it is. Thanks to whoever edit it for this work. --Aréat (talk) 09:30, 22 July 2018 (UTC)

next?

In my opinion this is about the last (or previous) election. Wikipedia is not a crystal ball. Klaas `Z4␟` V 07:52, 26 July 2018 (UTC)

It would be better for you to find another place to bring this issue (some Wikipedia page or WikiProject, like Wikipedia:WikiProject Politics), as there are similar articles for dozens of future elections (take a look to Category:Opinion polling for future elections). --Checco (talk) 07:50, 28 July 2018 (UTC)

I don't understand your objection though. This page is about listing opinion polls, which are by definition about future elections. And obviously there is no claim to predict the future at all. Ritchie92 (talk) 15:58, 28 July 2018 (UTC)

It would be a bit odd to call the article "opinion polling for the next election, if there is one", lol, or some other convoluted wording that is technically less Crystal-bally. I agree that it's fine as it is. Speed74 (talk) 16:28, 28 July 2018 (UTC)

What is the likelihood Italy turns into a dictatorship, or is wiped off the map? I don't think there's much question there'll be another election. Alex of Canada (talk) 22:11, 6 August 2018 (UTC)

CPI and PaP

Hi, I have a question about CPI and PaP. The two parties scored poorly in 2018 general election and did not won any seats; moreover they aren't listed in almost every polls, so should we really insert them in the table? -- Nick.mon (talk) 11:39, 24 March 2018 (UTC)

I would list them for now. We will see... --Checco (talk) 08:36, 26 March 2018 (UTC)

I'm of the opinion that they should be listed for completeness's sake. It's a bit weird to see parties suddenly appear/disappear with every new calendar year, in any case. Mélencron (talk) 21:51, 26 March 2018 (UTC)

Somebody got rid of CasaPound who polled as well as both CPI and PaP. can somebody back their polling back. QubecMan (talk) 06:32, 10 August 2018 (UTC)

CPI refers to CasaPound (CasaPound Italy). I believe it was excluded as it was only listed in one or two polls, and it consistently scored less than 1%. Speed74 (talk) 09:41, 10 August 2018 (UTC)

Hi I'v seen the party in a Lot more then two polls and with that being said you said it consistently scored less than 1%. Then what about NcI Us with Italy they poll less the 1% so ether they should go or CPI should be included back. QubecMan (talk) 09:54, 10 August 2018 (UTC)

I think we should decide once and for all on a general rule. I think NcI should be listed since it got more than 1% in the last elections, it is regularly listed in the opinion polls and anyway it's a party which has MPs in the Parliament. CPI has been listed in less than 10 polls overall since March and only in a few of them it had more than 1%. Ritchie92 (talk) 10:16, 10 August 2018 (UTC)

Significant digits in some polls

I would like to have some other users' opinion on this: some pollsters (Piepoli and Noto for example) usually have a very small sample and give their public results with numbers in steps of 0.5 (e.g. 18 and 3.5 and never 17.9 or 3.76 except rarely for very small parties like NcI). Furthermore, when the result is a full number they do not put the ".0" like we do instead in the tables in this page. They do this because for small samples of ~500 people they estimate an error (at 95% confidence) of ~4% and therefore it does not make sense (and it's not correct) to report all decimal digits. Should we also follow this rule and, in these cases, not include the ".0" for rounded numbers? Scientifically speaking, this makes sense because writing "18.0" implies a higher level of accuracy with respect to just "18". --Ritchie92 (talk) 10:36, 25 January 2019 (UTC)

I don't think it is a good idea, personally. Firstly, because removing the .0 would make it look inconsistent with other polls that do show decimal digits. Secondly, because as you point out these polls do show their numbers in steps of 0.5, which means they are basically rounding their numbers to the closest .0/.5 digit. It would look awkward for the same poll to show .5 but omit .0, as it would mean some figures would show a decimal digit but some others not, within the same opinion poll, when they are all basically a result of a similar rounding process. In this sense, .0 would not imply a higher level of accuracy than .5, but just the same. Impru20^talk 11:32, 25 January 2019 (UTC)

@Impru20: Yes I agree with you that this would look stylistically inconsistent with the other data, however my doubt was on whether it's better to be rigorous (mathematically, and also reporting no more significant digits than the ones given by the publications we cite), or have a pretty and consistent table. In the page Opinion polling for the next German federal election for example something like my proposal is implemented: however in Germany it's an easier decision because most pollsters don't give their decimals or they give just the ".5", even for middle-large sample sets. --Ritchie92 (talk) 12:15, 25 January 2019 (UTC)

@Ritchie92: Yeah, that's the issue. It works for Germany because most polls do show their numbers without any decimals at all (though I think it could be debatable there whether the .0 figure should be used in those polls, as such sources do indeed show the .0 digit as well). Nonetheless, I guess that, rigorously and mathematically speaking, it is just a matter of aesthetics; if we consider that the .0/.5 figure comes from rounding, in the end it means the same in numerical terms, so showing a .5 figure in such a poll wouldn't mean a different level of accuracy that using a .0 one or no one at all (i.e. 18.2 would be rounded down to 18.0, whereas 18.3 would be rounded up to 18.5. In the end there's little difference there). Impru20^talk 12:23, 25 January 2019 (UTC)

I am not an expert like you two, but I agree with User:Impru20 (and I would add "0.5" also to German polls). --Checco (talk) 11:47, 28 January 2019 (UTC)

I'd just point out that they don't do the same on the INSA list, and the Bild-Zeitung does the same. Mélencron (talk) 13:39, 28 January 2019 (UTC)

Notes on LeU and PaP

Let's discuss on the notes on LeU and PaP as there is some disagreement on them. User:Ritchie92 argues that "the situation is not the same: LeU dissolved itself as a party (it only exists as a Parl. group, while Art1 and SI still exist as separate parties), PaP is still active as a party/organization (and by the way is not true that PRC and PCI were the majority in PaP); moreover I can't recall of any opinion poll reporting PRC instead of PaP". That is not accurate: LeU was never formally dissolved and is now led by a group of autoconvocati (self-summoned militants). However, this is not my main point. My point is exactly that no opinion polls are reporting PRC instead of PaP! They report PaP instead of PRC, as they mostly report LeU instead of MDP and SI. PaP numbers are exaggerated because they de facto include PRC and PCI. Surely, PRC was the main force behind PRC, at least electorally, and will probably obtain more votes than PaP. Pollsters are not very reactive to party changes, but our readers should be informed that polls are not including PRC and that PaP numbers, as well as LeU numbers, are frankly exaggerated. --Checco (talk) 14:03, 15 February 2019 (UTC)

I can be wrong about LeU being not a party anymore. However LeU was born exactly as the collaboration of MDP and SI (and Possible, I guess), and when these two (three) members left LeU, a few polls started reporting MDP+SI or not reporting at all (and we, for simplicity, still write LeU with a note for this detail). Regarding PaP, sorry but there are many personal opinions and some inaccurate facts in your considerations, which I don't think we should discuss here . Regarding pollings, your sentence "PaP numbers are exaggerated because they de facto include PRC and PCI" is actually also a personal opinion and honestly, this is something we cannot rely on (unless you point us to a source confirming your idea). In the end my view is: if we write "PaP" in the table, no reader is going to assume that the PaP number is instead PRC. If some polls in the future will be reported as "PaP+PRC" or "PRC", we will point it out. --Ritchie92 (talk) 14:46, 15 February 2019 (UTC)

• LeU: It was not dissolved and there are ongoing talks between the autoconvocati, SI and also some MDP members (Laforgia) about reviving it as a party.
• PaP: Clearly to me, PaP wields little support, apart from the fact that most pollsters boost it by not proposing PRC and PCI as options. However, time will tell.
  --Checco (talk) 14:51, 15 February 2019 (UTC)

Reshuffling of left parties

As some might have noticed, there has been a bit of reshuffling of the various leftist parties which finally was applied in the last opinion polls. However most of these last polls are nominally for the European election coalitions. Should we adapt those polls and also add them to this list? If so, how do we treat the data for The Left? --Ritchie92 (talk) 10:33, 3 April 2019 (UTC)

I would add those polls to the list and I like the current situation. Eventually, we could split the table as many times as we want, as I saw other editors doing in several cases, e.g. Opinion polling for the 2019 Israeli legislative election. --Checco (talk) 06:40, 4 April 2019 (UTC)

Ps: Self-praise is not good, but, as it was obvious to me from the start, PaP without PRC counts little, so little that it has chosen not to participate in the election. Eventually, pollsters will stop tracking virtually-irrelevant PaP (and not PRC, now in partnership with SI), as they have stopped tracking the moribund or already-dead LeU.

@Checco: You are a true political scientist, I must say. But for the time being, I would base decisions on the current facts instead of (even very good) predictions. --Ritchie92 (talk) 12:19, 4 April 2019 (UTC)

For the rest of your proposal, I agree. In the future if the situation changes in a definite state I would also like to split the table. --Ritchie92 (talk) 12:20, 4 April 2019 (UTC)

I am surely a political scientist (and a historian) by training, but those were not predictions, they were facts. It was a factual mistake for pollsters to continue to ask voters about LeU, PaP and NcI. However, I totally agree with you and I love patience. --Checco (talk) 06:18, 5 April 2019 (UTC)

European Parliament

I know that European elections are and will be a good poll for parliamentary elections, but to include all the European polls just because they consider the same parties to be weird Braganza (talk) 15:24, 4 April 2019 (UTC)

Then we should remove all EMG, SWG, Noto, recent Piepoli & Euromedia, and Ipsos polls. --Ritchie92 (talk) 15:31, 4 April 2019 (UTC)

I would not. As above, pollsters are not so professional in Italy and, for their sake, the electorate does not differentiate European Parliament elections from general ones. Turnout aside, EP elections are mostly interpreted by voters and parties as mid-term elections. Having months without polls just because they are all on EP elections now would not be a good solution, for sure. Let's keep those polls here! --Checco (talk) 06:21, 5 April 2019 (UTC)

Green Europe

@User:Ritchie92,

Green Europe was listed by Rai 1 and Scenaripolitici.com as part of the center-left coalition.

Termometro Politico, on the other hand, showed other centre-left and other centre-reight, but this time it was not. see here

Braganza (talk) 12:02, 24 April 2019 (UTC)

Yes, that's why this time for Termometro Politico we cannot estimate the CDX and CSX coalitions. About EV, I think there is a distinction between being a centre-left party and being in the centre-left coalition. --Ritchie92 (talk) 14:18, 24 April 2019 (UTC)

EV should clearly be included in the centre-left coalition. --Checco (talk) 19:08, 25 April 2019 (UTC)

Why? Did EV mention it's part of a centre-left coalition? --Ritchie92 (talk) 08:10, 26 April 2019 (UTC)

Its main member is the Federation of the Greens, which is part of the centre-left coalition. Same reasoning for More Europe, even if some of its associate parties (for instance, the Italian Republican Party) are not members of the centre-left coalition. --Checco (talk) 08:21, 26 April 2019 (UTC)

Ok, then the polls listing EV in the centre-left coalition should be added (i.e. Termometro Politico) but not the polls that are not listing all the remaining centre-left (or centre-right) parties, like SWG. --Ritchie92 (talk) 08:27, 26 April 2019 (UTC)

Government Trust numbers

Currently, this page deals only with party preferences, coalition sums and coalition seat totals. However, at least one Italian pollster, "Demos", provides other opinion polls worth mentioning (for an example, see this page):

1. trust of the Conte Cabinet (% of respondents who rate it >5/10),
2. trust of party leaders (same method)
3. predicted length of the cabinet

All of these questions are asked in Spain and were added by Impru20 (talk · contribs) on the Spanish polls. I think we could go on and add them, especially 1. and 2.

Remarks?

Kahlores (talk) 21:22, 21 May 2019 (UTC)

I disagree. I don't see the polls you're referring to in the Spanish page, by the way; in my opinion there's even too much detail in that page, but that's debatable. In general, I don't think polls about the government trust should be in a page regarding the opinion polling for the vote in a general election. This type of polls are better in pages like Opinion polling on the Donald Trump administration where the title is self-explanatory, but I would not make the effort of creating and maintaining a similar page for each Italian government. --Ritchie92 (talk) 09:25, 22 May 2019 (UTC)

I fully agree with User:Ritchie92, even though several pages on opinion polling include the issues proposed by User:Kahlores. Despite that, I think that those polling numbers, especially "trust in government" and "predicted lenght of cabinet", would be quite redundant. --Checco (talk) 13:48, 22 May 2019 (UTC)

Sorry Ritchie92, the Spanish example I wanted to show is on the 2016 article.

Another example—and a better one—is Australia's page (checked): the party preference polls had not once predicted the Coalition ahead of Labor, as it eventually happened last Saturday ; however, the Leadership polling did show that Morrison was preferred to Shorten, which seems to indicate that this type of question is very important, at least for Australia.

Technically indeed, "government trust" isn't election-specific, but it is election-related. We're not going to add government trust polling on the Conte Cabinet article, are we? I've never seen this practice before, and although it would make sense, I don't want to set a precedent.

It matters to know the ups-and-downs of a government in public opinion, especially for an ad hoc coalition such as that one, as it may not translate in party preferences, for instance: supporters of M5S and Lega may come to trust the coalition government less, but still stick to their party; and vice-versa with the supporters of opposition parties.

Kahlores (talk) 18:11, 22 May 2019 (UTC)

I fully understand the meaning of government trust polls. In the Australian case I would argue it's probably more relevant because of their different electoral system based on preferential voting, where the secondary "taste" of voters weighs more. However if I had to choose, contrary to what you say, I would insert a list of government trust polls in the Conte Cabinet page, rather than here, where it just makes the page longer and goes out of the scope of the article. It makes much more sense, also because the election and the legislature are usually not linked to a single government. --Ritchie92 (talk) 07:31, 23 May 2019 (UTC)

Verdi

Some of the latest polls (Termometro Politico and SWG, for example) have included the Greens (Verdi). Should they be added to the table? --5.34.154.217 (talk) 19:20, 10 June 2019 (UTC)

As said before, I would add them! --Checco (talk) 05:21, 15 June 2019 (UTC)

I think the Greens should be added, possibly going back to older polls before the European election. I think it's also time for a new separate table, possibly starting from May 2019: LeU should be substituted with The Left, and we should discuss about NcI and PaP, which have not been polled since they did not run at the European election, but might reappear in the future. --Ritchie92 (talk) 10:26, 15 June 2019 (UTC)

I strongly agree with Checco and Ritchie92.--Facquis (talk) 14:58, 16 June 2019 (UTC)

Yes, I agree, we must add them, as well as we should remove Us with Italy. -- Nick.mon (talk) 18:42, 16 June 2019 (UTC)

Decimal numbers

Is there any particular reason that decimal numbers are used here? This is normally avoided as it gives a false impression of precision, and is generally not necessary. Onetwothreeip (talk) 22:23, 9 August 2019 (UTC)

Please read a discussion on this same page. --Ritchie92 (talk) 08:50, 10 August 2019 (UTC)

Aside from the arguments already explained in the linked discussion, it's simply not true that "this is normally avoided". There are a few selected countries where decimal numbers are indeed "avoided" altogether (UK and Austria which I can think of right now), but they are done so because most sources do not report decimal numbers, and those few that do get usually rounded by other secondary sources. Since Wikipedia must report on what reliable sources report, and Italian opinion polls are indeed reported with decimal numbers, the best course of action is to respect them. It's not up to us to decide whether Italian (or any other country's, for that case) opinion polls are being too precise so for us to somehow correct these numbers here, as that would be a distortion of sources. Impru20^talk 11:37, 10 August 2019 (UTC)

Coalitions Table Issues

A little background - I added a couple recent polls to the coalition table, and Richie reverted me on the basis that the pollsters didn't present the data that way. I reverted again, because I think this is pretty clearly covered by WP:CALC - Wikipedians are free to do simple calculations on data to present it in a helpful manner. The pollsters are providing all the underlying data in both instances - it doesn't matter whether they or we add up the constituent parties.

However, there are a couple issues that arise. Firstly, it seems that Free and Equal (LeU) has disbanded and somewhat reassembled as The Left (Sin). Unless we split the table, either name is going to be somewhat misleading as it will cover both at different points in time. My preference would be to have the current formation in the header with a note that polls prior to whenever referred to LeU.

Secondly, it would seem that the components of the coalitions have varied by pollster. Noto doesn't include the Greens in the Centre-Left coalition; Euromedia and Piepoli do include them. And as a result the "Others" column includes the Greens for Noto and not for Euromedia and Piepoli. My preference would be to align all the polls with the most recent coalitional alignment, those that existed at the European Elections - they are probably the best indication of what the coalitions will look like for the next election, and have the merit of being very close to the coalitions in the last election. Gabrielthursday (talk) 16:50, 21 June 2019 (UTC)

The coalition table is a topic up for discussion, I think. People wanted to eliminate it altogether in older discussions and I insisted to keep it as long as the pollsters gave expliticly their count for coalitions. The reason is that the number and composition (and the very existence) of the coalitions is debatable at this moment (and since a year ago I would say), therefore most pollsters do not care to estimate all small parties that belong to a coalition. More specifics:

1) The centre-right coalition is not a coalition anymore, if not at a regional and local level. Lega is in the government, FI is in the opposition, FdI supports the Lega part of the government, and NcI sort of disappeared. It also looks like SVP is now considered part of what it should be the centre-right coalition (at least according to Euromedia and Piepoli).

2) The centre-left coalition would be composed of the PD, part of LeU, +Eu, the Greens and more small parties (with or without SVP), but like you said, some pollsters exclude the Greens from the coalition. So, also here, the composition is dubious.

3) There were no coalitions at the European election, there is no such a thing. Each party ran separately, so there is no recent reference for the coalitions.

4) About The Left: it is technically a merge of SI (from LeU) and PRC (from PaP). Therefore it's a bit weird to title a column listing LeU as "Sin". My hunch: this "The Left" was only created for the European election and will also soon disappear. My proposal is to keep "LeU" with the footnote for now.

In the end, either we choose to define our coalitions and stick to that for all polls (i.e. we do our calculations from the single-party estimates, therefore Noto data should be changed, and we can add estimates for a lot more pollsters also retroactively like SWG, Index, etc.), or we choose to use each pollster's definition (then for some pollsters we cannot know the parties composing the centre-left coalition unless a total number is given). A mix of these two prescriptions is confusing and inconstistent. --Ritchie92 (talk) 17:57, 21 June 2019 (UTC)

I disagree on keeping LeU instead of "Sin" as long as pollsters keep polling The Left, mainly for two reasons:

1-in Italy pollsters poll the lists that took part in the last elections (or the ones aiming to take part in the next one), instead of the existing parties. Hence, it does not matter if the list will participate in the next general election. Therefore, I think we'd better keep several tables, according to the lists polled in each period.

2-LeU is also disbanded and it won't take part in the next election: A1 will likely merge or run in a joint list with PD, Possible is currently allied with the Greens and SI with the Communist Refoundation, which left PaP after the election, but those alliance will probably change again or disappear before the election, so it would be absolutely senseless to keep LeU. -- TommasoM (talk) 19:56, 21 June 2019 (UTC)

Sure, several tables is also an option, it's actually better for me. --Ritchie92 (talk) 00:55, 22 June 2019 (UTC)

Ritchie, those are all excellent points, and I feel foolish about coalitions in the European Election - a bit of a brain freeze, as I knew they ran separately. I think it makes sense to break the tables when one of the common polling options is significantly realigned, as is the case with LeU/Sin. Having a footnote is also useful, but I think the current party should be in the header, with the past configurations explained in the footnote. Another solution would simply be to put "Far-Left" or something similar in the header, with the changing composition noted in the footnote. This is hardly a major issue, but I think it could be a little clearer for readers.

On the coalitions issue, I think what makes the most sense is to define our coalitions, and have the table be consistent. Coalitions are either relevant or not, and polling companies presenting their information in one way doesn't make it so or not. That would allow for a fuller data set, which is always helpful. Having been quite embarrassingly corrected about the Euro elections, my opinion may not count for much, but I'd suggest sticking with the coalitions from 2018 - which do pretty closely resemble the coalitions that are running in the local and regional elections. If, at some point, one of the parties announces a substantial realignment, we can deal with it then. Gabrielthursday (talk) 07:30, 24 June 2019 (UTC)

This has been quiet for a while. @Ritchie92:, my reading of your earlier comment is that you don't have a preference between the existing approach and a "define-the-coalitions" approach. I prefer the latter, and so I'll try to find time to implement that.

On the LeU/Sin debate, taking TommasoM's view into account, is there an objection to moving to Sin with an appropriate footnote? Gabrielthursday (talk) 20:33, 10 July 2019 (UTC)

I would wait for some other editors' opinion. Frequent editors like User:Braganza, User:Koranion, User:Checco, User:Impru20, User:Nick.mon and others haven't expressed their opinion about this. In case we want to implement our coalition definition, we should agree on a definition first. About Sin, I think we can start a new table where Sin replaces LeU starting from some well-defined point. --Ritchie92 (talk) 23:41, 10 July 2019 (UTC)

I'm in favor of making a new table for LS Braganza (talk) 07:03, 11 July 2019 (UTC)

With respect to the Coalitions, I think the simple answer is use the 2018 coalitions as far as possible (Lega, FdI and FI on the right, PD, +EU, EV on the centre-left, and treat Sin as the successor to LeU on the far-left). It's not perfect, but it reflects what most pollsters who do allocate the parties to coalitions are doing, and will allow us to add data and present the coalition data in a consistent way. Gabrielthursday (talk) 03:14, 12 July 2019 (UTC)

As I said earlier, these coalitions are a delicate matter: the centre-right coalition is falling apart at a national level, and from the generic "centre-left" one does not have a definite status. The EV was not present at the 2018 election; we can consider it centre-left because of the FdV's past, but it's not technically "using the 2018 coalitions as far as possible". Same for the SVP, which was with the centre-left in 2018 but now it's closer to the centre-right coalition (after elections in Trentino-South Tyrol). LeU split in two, one side went back into PD, another formed Sin together with PRC. I would wait for more input from other expert users before fixing our own definition of electoral coalitions. --Ritchie92 (talk) 08:15, 12 July 2019 (UTC)

Well, in my view, coalitions’ table would be inserted only when they will be formed and formalized, and I mean some months before the election. As of July 2019, the situation is very unclear, maybe Lega would run alone or with FdI only, maybe FI will split into two different parties, and what about the centre-left? So, I think we should remove it, but this is only my personal opinion. -- Nick.mon (talk) 09:06, 12 July 2019 (UTC)

Thanks for waiting for my opinion! However, I have no real opinion on the issue.

Side note: As User:Braganza, I reject "Sin" for "The Left" and I support the use of "LS", the real acronym. --Checco (talk) 06:50, 16 July 2019 (UTC)

We should only be reporting shares for coalitions when the reliable sources specifically report that. Onetwothreeip (talk) 09:31, 16 July 2019 (UTC)

It seems clear that there is no consensus on what to do with the coalition table. In that light, I'm not sure that there's much utility in keeping a table that is both internally inconsistent (at least with respect to who is counted in the PD-led coalition) and doesn't include most of the relevant data. If we can't have a consensus to fix the problems, then imho, it's not really worth having - and those who are truly interested can do the math themselves. Nick.mon's position of bringing it back when coalitions are formally announced seems reasonable - and presumably at that point we can include all the pollsters regardless of whether they present their data with the coalition breakdowns or not. Gabrielthursday (talk) 21:03, 18 July 2019 (UTC)

I think it's reasonable to keep the table as it is now, and be careful not to include self-made calculations, but only cite official data when they are explicit. --Ritchie92 (talk) 20:37, 19 July 2019 (UTC)

@Braganza: I thought we decided to only keep coalition data which were explicitly given by the sources and not do self-made calculations. Your edit here is your self-made calculation. The composition of centre-left coalition is arbitrarily picked by you and cannot be inferred by that source. --Ritchie92 (talk) 12:32, 8 August 2019 (UTC)

@Ritchie92: There is an order roughly based on the right-left spectrum (Left, CSX, CDX, M5S&Other) but especially on the last parliamentary election and the alliances (so Left: LS, LeU, Other; CSX: PD, +E, Other, EV; CDX: LN, FI, FdI, NcI)

So it is not arbitrary, after all, it is the same as with the last Ipsos poll!

Braganza (talk) 12:39, 8 August 2019 (UTC)

@Braganza: This is WP:SYNTH, I am afraid. The EV was not in the coalition in last election, since it's a new list. The problem is not even that, but the fact that in the centre-left coalition there might be more parties that Ipsos counts in "Others", so we are underestimating the CSX. Similarly for the CDX. But whatever, I guess this has not been cleared enough on this page, and I am not going to start a war over this. --Ritchie92 (talk) 12:45, 8 August 2019 (UTC)

@Ritchie92: After all, there was earlier the option "Altri centrosinistra" which was just replaced by EV
It is strange that EV is being discussed over and over again, but only now about EV at Ipsos (a total of 6 polls since the formation of EV!)
After all, you said that: "Ok, then the polls listing EV in the centre-left coalition should be added (i.e. Termometro Politico) but not the polls that are not listing all the remaining centre-left (or centre-right) parties, like SWG." Braganza (talk) 13:03, 8 August 2019 (UTC)

@Braganza: That is exactly my point: the source from which the last Ipsos poll is taken does not list EV in a coalition. Assuming that from a previous source technically constitutes WP:SYNTH. --Ritchie92 (talk) 13:32, 8 August 2019 (UTC)

@Ritchie92: here [1] (this was also posted by Ipsos Italia on Twitter): Lega, M5S, PD-+E-EV and FI-FdI Braganza (talk) 19:37, 11 August 2019 (UTC)

@Braganza: This is obviously a different source from the one given in the table. I know what you mean but this still constitutes SYNTH. And anyway it's not only that poll, there are tens of polls in that table which are self-calculated, one should check all of them. --Ritchie92 (talk) 20:06, 11 August 2019 (UTC)

Gpf Polls

Am I the only one thinking that Gpf polls are extremely unrealiable and should be included in the list? They always tend to overrepresent M5S and underrepresent Northern League by 2/3 points each. --Thorin III (talk) 08:52, 26 August 2019 (UTC)

How can you prove that? --Ritchie92 (talk) 08:57, 26 August 2019 (UTC)

It is quite obvious that each pollster tends to overestimate several parties/lists and underestimate others, but in this page we track every poll conducted by every pollster. Moreover, it is not up to us to decide wether a pollster is reliable or not.TommasoM (talk) 09:41, 26 August 2019 (UTC)