Template talk:Same-sex marriage opinion polls worldwide

December 2015
User:Jedi Friend, why is this in the template namespace, where do you plan to transclude such a large table? --Joy &#91;shallot&#93; (talk) 08:58, 6 December 2015 (UTC)

I'm planning to add it to the Same-sex marriage page later on. Before that, I'm looking for other polls (of countries which aren't included yet). Jedi Friend (talk) 15:01, 6 December 2015 (UTC)


 * And where else? The point of templates is to be able to add the same content in multiple locations. If it's only ever going to be added to one place, you might as well not abuse the template namespace and use your user namespace. --Joy &#91;shallot&#93; (talk) 14:21, 8 December 2015 (UTC)

No South Africa?
Was such poll really never conducted in SA? --Aréat (talk) 01:09, 20 April 2018 (UTC)

Green and Red NPOV
I think the green and red colors on the percentages should be removed due to WP:NPOV and the related aesthetics and visual communication of these colors. These colors are frequently used for green=good, and red=bad. I recognize that they are also used for green=go, and red=stop, and can see how that makes some sense here. However, I believe the moral connotation outweighs the potential benefit this second meaning might communicate. The bold red is quite glaring and distracting in articles where this table isn't the most important point on the page. The green is harder to distinguish from black, and therefore the red inappropriately dominates visual attention. The boldface print already indicates which view holds a majority, and therefore I find these colors superfluous. Daask (talk) 16:29, 16 May 2018 (UTC)

There's also the problem of color-blindness. I wouldn't object to a less glaring color than red. — kwami (talk) 17:33, 18 October 2018 (UTC)

Should referendums be included?
I know this template is about polls and not referendum results, but I think it would be way more useful to also include them since they reflect reality better. Take for example Taiwan: The last poll says that 54% of taiwanese supports same-sex marriage while in reality it is only the 27% of the population who does so. Isn't it misleading to still have that 54% for Taiwan? --Baronedimare (talk) 19:38, 25 November 2018 (UTC)

Mexican states?
If we show the Mexican states, then shouldn’t we also show the U.S. states? And why are only about a dozen included, when the poll covered all 31? Panda2018 0 (talk) 17:21, 10 June 2019 (UTC)
 * IMO the Mexican states should be in their own section (appropriately titled Mexican states) and the table should be . Same for the US states, since we have precise receipts for each one of them as well. This is to avoid having a table with secondary information taking up too much space.  ×°˜`°× ηαη¢у  ×°˜`°×  19:12, 11 June 2019 (UTC)
 * About your second question, it's because I didn't have time to add them all at once. Help would be appreciated, also in finding polls about the US states and other sub-national entities.--Baronedimare (talk) 11:43, 12 June 2019 (UTC)
 * The US state-by-state opinion polls from 2017--the most recent ones--can be found here. I strongly suggest to create two hidden table sections for both Mexico and the US.  ×°˜`°× ηαη¢у  ×°˜`°×  22:24, 12 June 2019 (UTC)

Mexican states are more relevant that US (or Brazilian or Canadian or Australian) states because legalization is ongoing in Mexico but faits accomplis in the others. — kwami (talk) 05:38, 16 June 2020 (UTC)

Not using ILGA numbers
Thought we should state overtly that we're not using ILGA online polling numbers from 2016 as unreliable. I'm not advocating that position, just recording the existing consenus -- see Talk:Same-sex marriage/Archive 24. — kwami (talk) 02:51, 30 January 2020 (UTC)

No Flag for Northern Ireland
Per the discussion on the SSM page, I have changed the flag of NI to "no flag" to respect our policies on the Northern Ireland flags issue. Discussion there seemed to be in favour of either using either the Ulster Banner or "no flag" to avoid confusion that SSM has been legalized nationwide in the UK etc. This would also make the SSM article consistent throughout. There is no perfect solution to this question, hopefully this is something everyone can live with.--Darryl Kerrigan (talk) 00:53, 13 February 2020 (UTC)

Failed verification
Liechtenstein and Albania, added by the same person, have failed verification and I deleted them. Checking other countries added by same editor. — kwami (talk) 06:00, 21 July 2020 (UTC)

Costa Rica and New Zealand were off by a couple % points, though basically correct. CR probably because forgot to replace figure from AmericaBarometer, NZ was probably oversight while skimming source. But still, so far 4/4 have been wrong. — kwami (talk) 06:37, 21 July 2020 (UTC)

El Salvador, can't verify, drops 5% from AmericaBarometer (19% > 14%, or 15% not counting 'neither'), new poll added % con and 'neither'. Replacing w old poll because everything else by this editor is at least somewhat wrong.

Gibraltar failed and deleted. It wasn't even a poll, but "written consultations". — kwami (talk) 07:22, 21 July 2020 (UTC)

Italy corrected false claim that 0% said 'neither/none'. Mozambique I corrected some time ago (the 'in favor' and 'neither' figures were missing, 'against' was correct).

Nigeria fails. Poll was not about marriage, but about 'banning same-sex relations'. Doesn't say what's covered by that. — kwami (talk) 07:30, 21 July 2020 (UTC)

The link for Uganda & Rep. Congo doesn't do anything. I'm deleting under the assumption it's no better than the other research. — kwami (talk) 07:34, 21 July 2020 (UTC)
 * I'm sorry. I took most of them from this wiki page https://es.wikipedia.org/wiki/Anexo:Opini%C3%B3n_p%C3%BAblica_sobre_el_matrimonio_entre_personas_del_mismo_sexo_por_pa%C3%ADs
 * I should have checked them better before adding them. Please, check also the other similar templates https://en.wikipedia.org/wiki/Category:Same-sex_unions-related_templates--Baronedimare (talk) 19:56, 21 July 2020 (UTC)

Ah, that explains it. No worries. I've done that a few times myself.

But sorry, I really don't have time to check all of those. — kwami (talk) 05:25, 22 July 2020 (UTC)

Where is Mexico state data from? Need for interpretation.
Source credits INEGI, but I'm not finding it. Source suggests that the pro count is the remainder from the anti count, but that could be a misreading on their part. Could be there's also a 'neither' count. (E.g., for the country, 40% are opposed. Does that mean 60% are in favor?) We need the original data to be sure. — kwami (talk) 10:18, 21 July 2020 (UTC)
 * It should be somewhere here: https://www.inegi.org.mx/programas/enadis/2017/default.html#Documentacion but I can't find exactly where--Baronedimare (talk) 20:48, 21 July 2020 (UTC)

Yeah, I saw that. Closest is attitudes toward SS couples living together, but the numbers don't match. — kwami (talk) 05:56, 22 July 2020 (UTC)

Okay, question D-6 is ¿Con qué firmeza aprueba o desaprueba que las parejas del mismo sexo puedan tener el derecho a casarse?, with a 1-10 scale for answers. Current data set is at datasets.americasbarometer.org/database/index.php as a 1-100 weighted mean (margin). The freeware program '[R (software)|]]' can display the results. With the analysis updated a little from the 2017 publication, they are, Mex 51.09% ±1.14 Guat 23.21% ±1.06 Salv 18.59% ±0.91 Hond 21.43% ±0.85 Nic 24.52% ±0.98 CR 34.99% ±1.18 Pan 21.67% ±1.06 Col 33.98% ±1.34 Ecu 32.72% ±0.93 Bol 34.74% ±0.97 Per 38.44% ±0.87 Par 26.10% ±0.92 Chi 58.61% ±1.16 Uru 74.61% ±1.07 Braz 53.46% ±1.32 Ven 38.99% ±1.16 Arg 64.89% ±1.25 DR 27.48% ±1.03 Hai 4.86% ±0.346 Jam 15.78% ±0.95 Guy 20.6% ±1.29 Gren 11.82% ±1.43 StLuc 10.71% ±0.90 Dom 10.36% ±1.08 StVin 3.75% ±0.56 StKit 8.55% ±1.02 USA 60.90% ±1.35 Can 75.69% ±0.88 — kwami (talk) 22:13, 4 August 2020 (UTC)


 * It looks like those figs are not the %age that approve of SSM: that if a minority of a country strongly disapproves, and a majority approve but are largely indifferent, the result could be < 50%. E.g., if 1/3 rate their approval as 0 out of 10, and 2/3 rate it 6 out of 10, the result would be 40% even though 67% approve. So these figures aren't comparable to other surveys. — kwami (talk) 21:59, 23 December 2021 (UTC)
 * That is why I only included the "against" percentage, leaving the "for" entry blank. I don't know who changed it. Btw, should we move this thread to the new template?--Baronedimare (talk) 12:35, 25 December 2021 (UTC)

Subnational entities?
User:Baronedimare you say there's "no consensus" for the inclusion subnational entities, which means per WP:ONUS that it should be removed. This template is transcluded into major articles like same-sex marriage where polls of subnational entities are definitely WP:UNDUE. You could move the information to a separate template if desired. (t &#183; c)  buidhe  12:18, 23 December 2021 (UTC)
 * I created the template about sub-national entities and it can be found on Template:Same-sex marriage opinion polls sub-nations. What do you think? User:Kwamikagami--Baronedimare (talk) 12:38, 23 December 2021 (UTC)
 * Yes, I think it's better to have them separate like this, so editors can choose how much info to include. — kwami (talk) 21:49, 23 December 2021 (UTC)
 * Template was deleted because no-one was using it, and no-one voted to keep. I think that answers the question of whether the info should be included. — kwami (talk) 21:26, 8 January 2023 (UTC)

Gallup USA 2022
Can't find the full results. Assume similar to year before, with ~1% unsure, but can't verify. — kwami (talk) 21:27, 8 January 2023 (UTC)

Decimals in templates
Hey,. Kyodo News uses decimals in their polling numbers.


 * The 64% means 64.0%.
 * The 24.9% isn't a typo.

Thanks. KlayCax (talk) 07:49, 15 February 2023 (UTC)


 * If the 64% means 64.0%, then we should correct it to 64.0%. But all figures in the table are rounded off to the nearest 1%. No reason to make this an exception. Reverting.
 * Also, they only interviewed 3,000 people. No way that gives them a precision to the nearest 0.1%. So that's spurious precision even if we weren't rounding off elsewhere. — kwami (talk) 07:57, 15 February 2023 (UTC)


 * Are there other polls in the table rounded? The current one for Taiwan uses "64.4%" - for instance.
 * And yes: the 64% does mean 64.0% there. The poll states that "64.0%" are in favor of legal recognition and that "24.9%" are against. I don't think any poll (of a polity) is good enough — regardless of circumstances — to definitely gauge support beyond 4-5%. But I've usually seen decimals quotes as in within templates such as this. KlayCax (talk) 08:20, 15 February 2023 (UTC)
 * Taiwan was an oversight. Yes, it's common to quote improbable precision, but that's sloppy and best avoided. Ideally we would have the sigmas for all the polls, but our sources don't always provide them. (And even when they do, those sigmas are often underestimates.) — kwami (talk) 08:26, 15 February 2023 (UTC)

Breitbart is exclusively publishing polls from a reliable pollster
Recently, the deprecated source Breitbart has been routinely releasing exclusive surveys from the well-respected pollster Meganalisis. To my knowledge, Meganalisis has generally been considered to be a reliable source and is frequently quoted on a multitude of Wikipedia articles. Yet they're now regularly publishing exclusive surveys from a source that has been judged deprecated by Wikipedia editors. (Including on same-sex marriage, abortion, and several other issues.) I wanted to use the data to update this template... I'm deeply uncertain about how I should proceed in this.

It appears that these sorts of exclusives between them and Breitbart are now a regular thing. Generally, I'd state that Breitbart links are inherently unreliable and should almost never be used per consensus... but this feels like a time it might apply. The pollster who provided the data is considered to be a reliable source from my understanding, is frequently quoted for data on Venezuela on here, and is frequently cited. Is there a general agreement on what to do in situations like this?

If a reliable source starts posting information on a deprecated source... What are editors supposed to do? How are we supposed to cite this?

Use their Twitter directly? I don't know if there's an agreement on what to do here. Thanks. KlayCax (talk) 01:35, 25 March 2023 (UTC)


 * I would say citing it directly is better than citing the poll secondhand from Breitbart as Wikipedia editors don't trust Britebart to tell the truth. However, that only applies if the polling agency itself is considered a reliable source.  do you have more insight into that question? (t &#183; c)  buidhe  05:51, 25 March 2023 (UTC)
 * Somewhere I have saved something about the reliability of Meganalysis ... let me go digging for it. I'll be back :) Sandy Georgia (Talk)  13:24, 25 March 2023 (UTC)
 * Oops, no, what I have is a different company:
 * Datanalisis, "a respected pollster in Venezuela."
 * Will have to do more on this. Sandy Georgia (Talk)  13:28, 25 March 2023 (UTC)
 * do you have anything to hand on the reliability of Meganalisis? Sandy Georgia (Talk)  13:31, 25 March 2023 (UTC)
 * Thanks for the ping. Do you have a link to the polls you're referring to? Many thanks beforehand. --NoonIcarus (talk) 14:21, 25 March 2023 (UTC)
 * Nevermind, just saw it below in the same Twitter thread. I will try to offer a thorough response later today. --NoonIcarus (talk) 14:23, 25 March 2023 (UTC)
 * if CATI (define acronym?) is somehow related to CANTV, that is a concern ... what does CATI stand for ? Sandy Georgia (Talk)  15:09, 25 March 2023 (UTC)
 * "CATI, LA ASISTENTE VIRTUAL DE CANTV, CUMPLE DOS AÑOS BRINDANDO ATENCIÓN A LOS USUARIOS" ... if it's related to CANTV, how broad is the sample ?? Sandy Georgia (Talk)  15:11, 25 March 2023 (UTC)
 * Reading in their website, it is an acronym that comes from English: Computer-assisted telephone interviewing. Knowing how anti-government Meganálisis is, it's very unlikely they're related to CANTV. --NoonIcarus (talk) 15:15, 25 March 2023 (UTC)
 * I wanted to point out to this other poll:, conducted by Equilibrium CenDE, apparently offer way more reasonable numbers than Meganálisis: Around 53% against the legalization of abortion and around 40% in support of it. --NoonIcarus (talk) 15:21, 25 March 2023 (UTC)
 * I think that Meganalisis is probably a source for polls that should be attributed, though I suppose that goes for most. Meganalisis don't seem to share their survey sample, which is of relative concern, and I think it has been widely cited on Wikipedia as 'better' in their methods of collection than other companies, and as a balance following insertion of pro-government stats. Based on only my reading of the situation, I assume their association with Breitbart would be due to Breitbart having been anti-Maduro, with no concern for anything else. Indeed, Breitbart's journalistic practices would probably be above-average for Venezuela, so if considering Meganalisis 'legit', the association wouldn't be too much of a black mark, for me. Different polling companies are going to record different numbers, but if there are other polls of the same scope and Meganalisis appears to be an outlier, that would reinforce the concerns about their sample. Users may want to discuss if lack of transparency in that regard is enough to strike Meganalisis - which is more of a concern where it would be the only survey reference. Kingsif (talk) 22:26, 26 March 2023 (UTC)


 * I have to admit that I'm surprised of the relation of Meganálisis with Breitbart. I would guess that some Venezuelan outlets remain to be aware of its unreliability. As an example, I posted about an instance where El Nacional republished a Breitbart hoax (and still remains up today) at WP:VENRS. On the other hand, other reliable outlets and factcheckers are already aware of Breitbart's issues.


 * Regardless: while Meganálisis is an independent pollster and can be trusted to give an idea of the direction of public opinion, my experience is that their numbers can be a little exaggerated, and the best way to read their data would be to compare it along with one or two more independent pollsters. The way I see it, Breitbart is probably just cherrypicking the results that just so happens to allign the most with their editorial stance, although polls about same-sex marriage in Venezuela are scarce. --NoonIcarus (talk) 22:35, 26 March 2023 (UTC)

Excessive polls
No way we need 6(!) separate polls of the US. Can someone who knows more about polling than I do please remove at least four of these leaving the polls with the best methodology? (t &#183; c)  buidhe  05:55, 25 March 2023 (UTC)

Recent sources
I reverted a bunch of edits based on recent sources. They look good at first, but I checked two countries -- Bahamas and Trinidad -- and both failed verification. I reverted to be safe, but it might just be a matter of correcting the numbers we copied. Unless I simply missed the data? Do we have a page number? — kwami (talk) 02:24, 6 June 2024 (UTC)

AmericasBarometer wrong and needs update
There is 2023 approval data for most countries in the Americas from AmericasBarometer on page 65 of https://www.vanderbilt.edu/lapop/ab2023/AB2023-Pulse-of-Democracy-final-20231205.pdf. I believe all countries in the Americas not listed there have 2016 data (not 2017 as stated now). This includes Barbados which is missing at the moment. Belize has numbers in the data portal mentioned below for 2023 but not in the report.

Note the current numbers are not a percentage of approval. The survey responses are an approval from 1-10 (strongly disapprove to strongly approve) and then these are used to create a weighted average as a measure of tolerance which is displayed now (see FAQ: How are averages and other statistics determined? How should they be interpreted?). The raw data can be found on https://www.vanderbilt.edu/lapop/interactive-data.php for all countries and years under the variable "Tolerance:Approval of Same-Sex Couples' Right to Marry (d6)".

I think the approval in the report is calculated as those who answered 10,9 or 8. In my mind, a split by 1-4 = disapproval, 5-6 = neutral, 7-10 = approval makes the most sense. These should add to 100.0% give or take rounding errors Trackscarracing (talk) 01:16, 27 June 2024 (UTC)