Talk:Opinion polling for the 2023 Polish parliamentary election/Archive 1

Archive 1

Suggestion

@Ed88 (talk · contribs)

As the number of parties represented in Parliament and opinion polls grows, it is hard to find space. With a bit of dexterity I've managed to tuck 15 parties, 5 coalitions, and three other columns on both sides.

Here are two versions, one with the two tables (by coalition and by party) and the other in one table.

Kahlores (talk) 21:39, 15 November 2019 (UTC)

References

@Kahlores (talk · contribs)

Your suggestions sound very interesting! Most Wikipedia articles containing opinion polling follow the rule that parties are sorted by decreasing support in the previous election or by decreasing number of seats gained in the previous election. Norway and Sweden are the only notable exceptions I found, with parties sorted from left-wing to right-wing. That's why I believe that a reader expects rather decreasing support sorting and I would prefer to keep it. The other problem is that left-wing to right-wing sorting may prove to be controversial in some cases due to lack of objective criterion for measuring leftism or rightism...

I would suggest to keep split tables and the order of decreasing support of electoral alliances and decreasing number of gained seats of parties:

Electoral alliances

Polling Firm/Link	Fieldwork Period	Sample Size	Law and Justice	Civic Coalition	The Left	Polish Coalition	Confederation	Others / Don't know	Lead
Kantar Public / "WO"	5-12 Nov 2019	1,502	38	27	14	7	9	5	11
Pollster / SE.pl	5-6 Nov 2019	1,099	46.4	25.0	12.7	8.8	5.7	1.4	21.4
Social Changes / wPolityce.pl	1-5 Nov 2019	1,033	43	24	17	8	7	1	19
Indicator / TVP	31 Oct-4 Nov 2019	1,000	44.0	26.1	14.4	8.8	6.2	0.5	17.9
Social Changes / wPolityce.pl	25-30 Oct 2019	1,040	41.1	24.9	15.9	9.8	8.3	0.0	16.2
IPSOS / OKO	21-23 Oct 2019	1,004	41	23	15	9	7	5	18
IBRiS / "Rz"	18-19 Oct 2019	1,100	41.3	21.7	12.3	7.1	6.5	11.1	19.6
Kantar Public	11-16 Oct 2019	986	39	27	14	7	7	6	12
Parliamentary election	13 Oct 2019		43.6	27.4	12.6	8.6	6.8	1.0	16.2

Parties

Polling Firm/ Link	Fieldwork Period	Sample Size	United Right			Civic Coalition				The Left			Polish Coalition		Confederation
			PiS	P	SP	PO	.N	IP	Z	SLD	w	LR	PSL	K'15	KORWiN	RN	KKP	Others	Don't know	Lead
Social Changes / wPolityce.pl	1-5 Nov 2019	1,033	38			20	2	–	–	10	8	2	6	5	4	2	–	3		18
Social Changes / wPolityce.pl	25-30 Oct 2019	1,040	37.7			21.5	2.5	–	–	7.6	7.4	3.7	6.8	5.7	4.3	2.2	–	0.6		16.2
IPSOS / OKO	21-23 Oct 2019	1,004	35.7	1.2	3.3	17.5	2.5	1.4	1.1	4.9	5.5	4.1	6.4	1.9	3.1	1.9	1.3		8.2	18.2

Regards, Ed88 (talk) 14:41, 20 November 2019 (UTC)

Thanks Ed,

There is an ongoing debate about the use of the left-right arrangement. The reason why it was used for Sweden and Norway is because that's how publishers of these countries (such as SVT or NRK) lay out the spectrum.

Indeed, if there is no reference for Poland, it will be hard to defend such an arrangement. There are some references to arrange the 5 coalitions (let alone their name, "Left", "United Right"), and some parties within coalitions are referenced to be more 'centrist' than the others. However, there are exceptions, especially on the fringes.

A mixed system could thus be used: coalitions from left to right, parties in decreasing order. It depends whether the left-right arrangement is widespread in Poland, or not. Then, do as you may!

Kahlores (talk) 16:59, 20 November 2019 (UTC)

Opposition?

Why isn't the Confederation party counted as a part of the Opposition figure? The Wikipedia Article for the Sejm counts them as "Opposition". https://en.wikipedia.org/wiki/Sejm — Preceding unsigned comment added by 109.95.144.216 (talk) 23:58, 13 August 2020 (UTC)

Do we really need this "opposition" column? In my opinion, it makes the table less readable and contains duplicate data. Such columns are not standard on analogous Wikipedia pages for other countries, especially when the government majority is included in just one other column. Regards, Ed88 (talk) 14:45, 14 August 2020 (UTC)

Rafał Trzaskowski's New Solidarity

For the benefit of legibility, we should probably "split" the Civic Coalition into two columns, one for KO and the other for NS. This would avoid looking at two different places of the table, as NS is obviously taking most of their supporters from KO.

That said, there are some subtleties that only Polish Wikipedians here may be aware of.

1. Will NS be a member party of KO? Or another standalone party? Or another coalition that would include PO?
2. Why does Rafał Trzaskowski create a new party, and not try instead to take over PO/KO?
3. Who is set to join him? from which parties?

Whichever logic NS follows, the table should represent that logic.

Kahlores (talk) 02:10, 15 September 2020 (UTC)

@Kahlores: first of all, some polls have that two "parties" separated, some have together - you have a mess with that. So if a poll have NS and KO separated - make that separated for eg. NS 20% KO 6%, but when poll have that "parties" together make that together coz electorate of that both "parties" is pretty much the same one.
Why did I make parties word in a " "  ? - NS isn't really a political party - but polls recognize New Solidarity like a party... Trzaskowski's New Solidarity is more an extension/ support of already existed KO/PO - like Trzaskowski said: "I am making NS to these people who don't like/ don't believe any more in political party". So like RT and politicians from PO/KO already said couple times: NS is NOT going to run in elections but it's going to support already existed PO/KO. Yes, I know - it's a "little" messy, but we still need to wait for a more clear information - probably in a ~month (about October 15 there will be official start of that movement.) If I still didn't explain something - feel free to ask :) Natanieluz (talk) 10:18, 15 September 2020 (UTC)

Thanks Natanieluz, the situation is clearer to me now. KO politicians and RT said won't run. . . but what are these polls for, then? And why not run if NS' presence is a boost to KO? Probably the politicians themselves don't know exactly what to do, and that is why polls are being conducted.

Anyways, I believe we should start the split on 7-12 August, when the first poll about Trzaskowski's new "thing" was conducted. Kahlores (talk) 14:01, 15 September 2020 (UTC)

@Kahlores: Yea, we still need to wait for a more info about that new movement, but for now we have polls like that... And yes, we can start with that split on 7-12 August, give me a sec. Natanieluz (talk) 14:27, 15 September 2020 (UTC)

@Natanieluz: Six weeks later, and since we spoke not a single new poll has provided separate estimates for Trzaskowski's Wspólna Polska. Was it a dud? Or, more likely, does he play the long game, building up his movement on the sidelines, as an extension of KO?

In any case, it seems like only KO is surveyed now. I suggest we fuse both tables back together, and instead of creating two separate columns (despite as you said that "NS is NOT going to run"), we should have one visible column split in two for the 6 special polls, with an explanatory footnote. Kahlores (talk) 20:40, 30 October 2020 (UTC)

@Kahlores: Yea, I thought polls would give two separate % - one for KO, and one for that movement - but since we don't have that, we can do as you write. Natanieluz (talk) 20:47, 30 October 2020 (UTC)

CBOS polls

To editor Natanieluz:, I've been surprised to see that you removed CBOS poll from that table with a note "goverment controled "polls", its a propaganda not a poll". This is absolutely not true. CBOS has been conducting their polls for almost 40 years, the organisation is supervised by the prime minister but it is still relatively independent, whichever government is in charge. It is headed by the same director since 2008 and she is definitely not engaged politically. The methodology used in CBOS polls is sometimes seen as archaic but the reason behind it is the need to keep the comparability of data across years and decades and they are struggling to do so even during the pandemic. The polls usually favour ruling party but delivering propaganda is not their goal ([1]). I would ask you to restore CBOS poll data and keep it. Regards, Ed88 (talk) 20:12, 11 October 2020 (UTC)

Hey @Ed88: I complete disagree, CBOS is a government controlled now more than ever before, since PiS came into rule - the CBOS it's not more than a gov propaganda (like I said), their "polls" are not wrong by a 1 or 2 % but a dozen or more - even I can make a more accurate one. For e.g. in poll for 2020 presidential election they were wrong by more than 22 percentage points [2] and they were considered the least accurate surveying (26th place out of 26), do we really recognize the propaganda tool as a poll? Sure we can have polls from CBOS - but then we also need more private and smaller company surveys because the're more accurate - they're polls not like CBOS-like things... Natanieluz (talk) 20:29, 11 October 2020 (UTC)

@Natanieluz and Ed88:. In this case, what you could do is find the list of the 26 pollsters and provide it in another section, up for the readers to read and assess. I've found such a ranking on the Swedish article for their 2014 election. One link is enough.

You could also provide details about the methodology (by phone, internet, or in person) and questionnaire (the wording is very influential, and some answers aren't suggested) as is done on this article or that one.

That said, I am not in favor of removing CBOS polls, because, aside from the last poll in October, which was a clear outlier, they are not always so far from other polls, as we can see. In any case, fraud cannot last. Kahlores (talk) 01:29, 18 October 2020 (UTC)

@Kahlores: Sure, we can make that list (ranking) for all polling stations, as of CBOS- maybe a good way will be waiting for another, newer poll and see if still they have several dozen percent "recoil" compared to other polls? Natanieluz (talk) 10:48, 18 October 2020 (UTC)

That would be editorializing. If we start removing outliers, where will we stop? Kahlores (talk) 05:16, 20 October 2020 (UTC)

So we went back to the starting point in our discussion... Natanieluz (talk) 11:21, 20 October 2020 (UTC)

@Natanieluz and Kahlores: we have another CBOS poll - [3]. It gives us exactly the same information as other polls - the support of the ruling party is declining and is even lower than the average of its support in the recent polls conducted by other pollsters. I cannot judge whether this is also propaganda or not, but the CBOS research is certainly quoted by the entire Polish press, regardless of its political profile, and it would be strange to omit it in Wikipedia. I am not against adding any additional information about methodology of each poll. Ed88 (talk) 18:02, 30 October 2020 (UTC)

@Ed88: You're right, this (newest) CBOS poll is similar to others, of course we should add that poll. Natanieluz (talk) 19:22, 30 October 2020 (UTC)

We should put the previous back as well. It is not up to us to cherry pick which poll is correct and which is not, for that would be propaganda (although an inane one). Let the readers judge for themselves. The polling companies that produce results that are completely off the charts are giving themselves a disservice anyway. Kahlores (talk) 20:21, 30 October 2020 (UTC)

@Kahlores: Ok, maybe you are right. I will add that previous poll. Natanieluz (talk) 20:42, 30 October 2020 (UTC)

One of the oddities of the CBOS polls is the high proportion of non-respondents (22% "don't know"). That number has more than doubled since the first polls in late 2019. The difference between then and now is, of course, that the opposition landscape became much more chaotic. Does CBOS ask a follow-up question to those who first answer they don't know? In case they don't, we have an explanation for @Natanieluz:. Kahlores (talk) 18:05, 18 June 2021 (UTC)

Suggestion: removing Agreement and Kukiz'15 from the poll table

Hey!
The percentages of these parties oscillate within the limits of statistical error, some polls do not include them at all... Considering that the election threshold in Poland is 5%, and the fact that other insignificant parties such as Agrounia do not have their positions in the table at all: my conclusion is to remove them from the table. What do you think? Natanieluz (talk) 08:51, 16 August 2021 (UTC)

I see that someone even added AGROunia... Do we really need those small, insignificant parties to be added to the poll table? Natanieluz (talk) 10:29, 20 August 2021 (UTC)

solutions:

1. The sections for "Parties" and for Coalitions (2019, 2020...) could be merged.
   • All we need is to use colspan to describe coalition numbers. To make it easier, I moved Porozumienie's column in the "Parties" section to the right of the "United Right" alliance. Problem: the table isn't sortable, because of colspan, we'd need to separate coalition totals from party totals.
   • We could also replace the 17-party table by 17 or more tables specific to each party.
   • We could also add internal links in the Coalition columns each time a pollster asked the more precise question (which coalition? + which party?)
2. In any case, 2021 has to be cleared up, without removing data.

Kahlores (talk) 00:58, 4 October 2021 (UTC)

Proposition: removing government own or related "polls"

My suggestion is to remove CBOS, Social Changes and Estymator "polls" from poll table.
I've already suggested removing CBOS "polls" back in October 2020, but no decision was made then. Not only adding fake polls "destroys" the real view of the political situation in Poland but also disrupt average party support on graphical summary; there not only have extremely bias towards government (PiS) but are own by government people or politicians.
CBOS - whole discussion was already made in October 2020 (if more info needed, I will be glad to explain one's again, or in more pleasant way).

Social Changes for wPolityce - just go to this page and see what I'm talking about - 100% content glorifying the government and attacking the opposition. But that's not the most important issue here - the owner of wPolityce is Law and Justice senator Grzegorz Bierecki. WPolityce direct media owner - Fratria (own by Bierecki like a said) have 31% of their revenues in 2020 were supplements from a state-owned company (spółki skarbu państwa in polish). Also, according to research made by Nahema Marchal, Bence Kollanyi, Lisa-Maria Neudert and Philip N. Howard from University of Oxford in 2019 wPolityce was recognized as the fourth most active Polish producer of the so-called junk news, or "misleading, deceptive or incorrect information allegedly supposed to be the real news about politics, economy or culture.

Update1: I have found information confirming the relationship between social changes and the Ministry of Justice - Social Changes is a poll company which carries out opinion polls, among others for wpolityce.pl. The center is managed by Marek Grabowski, Grabowski is also the president of the conservative Mummy and Dad Foundation, which implemented a PLN 1.5 million social campaign commissioned by the Ministry of Justice, which carries out media campaigns commissioned by the government.

Estymator for DoRzeczy.pl - exactly the same as with wPolityce. DoRzeczy media owner Tomasz Sakiewicz (Platforma Mediowa Point Group S.A (PMPG) - a man very closely associated with the government which organizes annual awards to be won by... members of the government. 25% of their annual revenue comes from government's state-own company's (spółki skarbu państwa in polish). Like said on Do Rzeczy Wikipedia page "Weekly often promotes the PIS party narrative."

This picture shows well the ownership relationships between the people of PiS and the media I'm talking about,

Since no one give an opinion about my proposal, I will ask the editors of this article for their opinion @Ed88, Kahlores, Leo0502, and K2101: What do you think? Natanieluz (talk) 20:54, 12 January 2022 (UTC)

I would strongly oppose removing any polls published in the media from the list. It is true that you can analyze the ownership structure of individual media and the content they publish, and state on this basis that they are one-sided. However, this does not only apply to government media. Similar comments can be made to the anti-government media. In Poland there are no media that are independent and not involved in the political conflict. So if we remove the CBOS, wPolityce, DoRzeczy polls, others will demand to remove the Wyborcza, TVN, OKO.press polls as well. Very soon we will have an empty list. Reading the information AND the source of this information, the reader should be able to properly interpret the received information. Regards, Ed88 (talk) 21:25, 12 January 2022 (UTC)

@Ed88: Even if someone considers e.g. Wyborcza, TVN, oko.press and their polls to be "anti-government media" and their polls to be bias - none of the those media is not in any way connected with a given political party. The situation is completely different when it comes to CBOS, wPolityce, DoRzeczy. There are owned (directly or indirectly) by the PiS politicians and people. Natanieluz (talk) 20:36, 13 January 2022 (UTC)

I remember we had a discussion about CBOS polls in mid-October 2020 (see above), and just afterwards, the anti-abortion ruling was passed and protests erupted, with PiS dropping 10 points including in the CBOS polls (40.4 to 31.3 and then 29.8), on par with every other pollster. If the alleged pro-government bias doesn't hide 10-point drops, then this is not the obvious bias worth removing.

What you can do, however, is help readers know more about these potential biases. For instance, you can translate the polls' specific phrasing, as was done I think for the Swedish and French opinion polls. Otherwise, as Ed88 said, private media companies aren't neutral either. Ownership is detailed on each pollster article.

Kahlores (talk) 08:28, 13 January 2022 (UTC)

@Kahlores: "Otherwise, as Ed88 said, private media companies aren't neutral either." - I don't really agree with that, but that's not the point - you see - there isn't any other private media company in Poland (apart from the indicated pro-government ones) that is associated with any political party. Of course, some of them favour the opposition more, but that's not the point - their owner is not a member of any political party.

"Ownership is detailed on each pollster article." - sure, but we all know well that 99% of people reading these polls will not even see ownership information on their pages. Natanieluz (talk) 20:36, 13 January 2022 (UTC)

Removing polls biased towards government spark another question - why don't we remove polls biased towards others parties or opposition in general - Kantar, IPSOS, IBSP? How do we know who is biased and when? Until election we don't know real numbers and cannot tell if someone is biased, uses incorrect adjustments or asks people in wrong proportions. Do they do mistakes on purpose or by accident (or by "accident")?

I treat opinion polling pages on wiki as an archive of all polls and their exact results and we're not meant to decide who's worthy to be included. However we can add relevant comments with proofs so readers take results of certain polls with a grain of salt. Leo0502 (talk) 14:34, 13 January 2022 (UTC)

@Leo0502: "... why don't we remove polls biased towards others parties or opposition in general - Kantar, IPSOS, IBSP" - because that's not the point here. We are talking about ownership issues and their impact on the "percentages" in the polls, not just by being more government or opposition biased too. Natanieluz (talk) 20:36, 13 January 2022 (UTC)

Proposition no.2

@Ed88, Kahlores, Leo0502, and K2101: Because you prefer relevant information (comments) with proofs as Leo0502 stated rather than removing selected polls from the table, my proposition is to do something similar to pollsters information available on Opinion polling for the next Slovenian parliamentary election or Opinion polling for the next Austrian legislative election (more about quality criteria than ownership, but I believe you see my point) - they created a list with pollsters information. What do you think? Natanieluz (talk) 20:36, 13 January 2022 (UTC)

Agree. Slovenia's list looks good Leo0502 (talk) 21:49, 13 January 2022 (UTC)

If you wish to write an informative introduction, it's absolutely fine! Ed88 (talk) 15:12, 14 January 2022 (UTC)

 Done (feel free to reopen it if necessary) Natanieluz (talk) 20:57, 19 January 2022 (UTC)

Attempt

There are multiple research companies conducting opinion polling that operate in Poland, some of them are consider to be directly or indirectly connected to the ruling party:

• CBOS - directly government controlled polling company, by law CBOS is a foundation supervised by the Prime Minister.

This section is an excerpt from Centre for Public Opinion Research § Criticism.[edit]

CBOS has been accused of performing "magic" on behalf of their clients by both wPolityce.pl magazine, and Gazeta Wyborcza. For example, results of a survey conducted by CBOS prior to presidential elections of 2015 had shown rapidly growing support for Civic Platform (Platforma Obywatelska) even though only ten days earlier, in a different survey conducted by CBOS on behalf of TVN24 (a different client), the Law and Justice (Prawo i Sprawiedliwość, PiS) was ahead of PO in terms of popularity. Agnieszka Kublik writing for Gazeta Wyborcza noted that the poll was a failure not only because of how respondents were being approached by CBOS, but also because of its perceived political affiliations.

After the parliamentary election 2019, Europe Elects showed that CBOS had the largest accumulative error among all Polish pollsters, favouring especially the government party PiS.

• Social Changes for wPolityce - directly own and controlled polling company by PiS senator Grzegorz Bierecki. Also, according to research made by Nahema Marchal, Bence Kollanyi, Lisa-Maria Neudert and Philip N. Howard from University of Oxford in 2019 wPolityce was recognized as the fourth most active Polish producer of the so-called junk news, or "misleading, deceptive or incorrect information allegedly supposed to be the real news about politics, economy or culture".

• Estymator for Do Rzeczy.pl Tied to a person close to the government, indirectly controlled by PiS through government subsidies, 25% of their annual revenue comes from government's state-own company's (spółki skarbu państwa in polish). Do Rzeczy weekly often promotes the PIS party narrative.

---

This is WP:Editorializing, and one could say a completely one-sided view. Pollsters are all tied to their commissionners (media organizations) which are all connected, indirectly or not, to the political and economic spheres. You cannot describe the connections of some and not the others. Since you won't describe all of them, it follows that this section should be left blank.

Kahlores (talk) 00:13, 21 January 2022 (UTC)

References

References

References

New columns for smaller parties

We have AgroUnia appearing in 8 polls this year, once in a while in previous years, it's possible we'll see them more often as party is now registered. If that happens, when should we add a column for AU (and other new parties)? How much polls is enough for new column? Leo0502 (talk) 21:44, 14 April 2022 (UTC)