Template talk:Brexit/Post-referendum opinion polling (2016-2020): Remain-Leave

Sample size

Documents like this quote three values for sample size: "Weighted Sample" (1631), "Unweighted Sample" (1633) and "Sample Size" (1653). Which is it correct to use? -- AlphaMikeOmega (talk) 20:28, 15 August 2018 (UTC)

Good question.

• "Sample size" (1653) is the number of people answering the survey as a whole.
• "Unweighted sample" (1633) is the number of people answering this particular question.
• "Weighted sample" (1631) is the result of separating the numbers for each response, weighting them, and then adding them back together.

"Sample size" appears to be the least suitable. "Weighted sample" appears in the source as the first value, and is in bold, so I am currently using that number. There could be a good argument for using either weighted sample ("the number the percentages are based on") or unweighted sample ("the number of 'real' people answering the question"). Ross UK (talk) 03:07, 18 August 2018 (UTC)

"Do you think that Britain should leave or remain in the European Union?"

The page Brexit shows polls' results from asking "how people would vote in a second referendum", whilst polls such as this ask "Do you think that Britain should leave or remain in the European Union?" These are similar questions, but it might be worth distinguishing between them: one can imagine a respondent who -

1. would vote to remain;
2. believes that more people support leaving than remaining;
3. supports following the wishes of the larger group;
4. would, therefore, think that Britain should leave despite (1).

Personally, I think we should distinguish, and so create a separate table. -- AlphaMikeOmega (talk) 18:27, 19 August 2018 (UTC)

We are now up to three tables on different issues. The second table is a combination of at least three different questions ("How would you vote in a referendum?", "How would you vote in a second referendum?", and "Should the UK remain/leave?" (asked after the first referendum)). These are similar topics that might properly be dealt with in a single table, thus avoiding the clutter of four or five tables instead of three.

I agree that it may be important to note the distinction that you point out. It may be better to deal with it in the introductory paragraph to the second table and/or in the 'Notes' column. – Ross UK (talk) 13:14, 20 August 2018 (UTC)

Thanks for the reply.

Of the three different questions you mentioned, I think the first two can be considered as asking the same thing, since "a referendum" at this stage would necessarily be a "second referendum", so respondents should answer in the same way. Therefore, only one more table should be needed (unless we find other questions in the mix which might require further tables if we decide that adding tables is the way to solve such problems).

Also, I imagine that responses to the polls in Template:Brexit/Post-referendum_opinion_polling:_Right-Wrong would match how people say they would vote in a referendum more closely than would their response as to whether Britain should leave the EU. Since the former has its own table, I think there is precedent enough for giving the latter a new table.

Having a larger number of smaller tables need not look cluttered, and may in fact look less so.

For now, I will start to add notes as you suggest. However, I plan to produce a new table if there are a few polls which ask this question. – AlphaMikeOmega (talk) 23:08, 20 August 2018 (UTC)

OK, thanks for the update. The simple Remain/Leave polling question is the same as the first referendum question itself, which is why it seems to fit with a "second referendum on the same question" description.

Another option would be to change the title and description of the second table from "second referendum vote" to "post-referendum opinion on Remain/Leave". – Ross UK (talk) 17:43, 21 August 2018 (UTC)

Missing polls

Let's list here the polls which are still missing from this table: —AlphaMikeOmega (talk) 23:01, 18 May 2019 (UTC)

• YouGov's (monthly?) Eurotrack polls from September-December 2018 —AlphaMikeOmega (talk) 23:01, 18 May 2019 (UTC)

Update: These may well all be added now.—AlphaMikeOmega (talk) 23:40, 19 September 2019 (UTC)

Permanently(?) missing

• Listed here:

• 23/05/17 Panelbase — seems not to contain this question on its PDF.—AlphaMikeOmega (talk) 16:37, 18 September 2019 (UTC); edited by AlphaMikeOmega (talk) 17:39, 18 September 2019 (UTC)
• 01/11/17 NatCen — I have not found the data tables.—AlphaMikeOmega (talk) 17:39, 18 September 2019 (UTC)
• 13/07/18 Kantar TNS — This poll seems not to exist: presumably it's a mistype of the poll a month later.—AlphaMikeOmega (talk) 17:50, 19 September 2019 (UTC)

Added polls

To avoid repeat work, we can write here which groups of polls are already included at the time of writing.—AlphaMikeOmega (talk) 23:12, 18 May 2019 (UTC)

• The 76 77 Eurotrack polls currently listed at What UK Thinks.—AlphaMikeOmega (talk) 23:12, 18 May 2019 (UTC)

Updated—AlphaMikeOmega (talk) 20:04, 17 September 2019 (UTC)

• The 77 78 polls currently listed here are all added either to the table or to #Permanently(?) missing.—AlphaMikeOmega (talk) 10:10, 19 September 2019 (UTC)

Updated—AlphaMikeOmega (talk) 19:41, 14 October 2019 (UTC)

• The 66 67 polls currently here.—AlphaMikeOmega (talk) 17:04, 19 September 2019 (UTC)

Updated—AlphaMikeOmega (talk) 15:45, 30 September 2019 (UTC)

• The 131 post-referendum polls currently listed here are all added either to the table or to #Permanently(?) missing.—AlphaMikeOmega (talk) 18:04, 19 September 2019 (UTC)
• The 1 poll here.—AlphaMikeOmega (talk) 01:42, 19 October 2019 (UTC)

ComRes Poll (Oct 19)

Does this really belong here? "Regardless of the way you voted in the 2016 referendum, do you support or oppose the UK abiding by the referendum result and leaving the EU?". If anything, it seems to me that this question maps more closely to the right/wrong one: https://en.wiki.x.io/wiki/Template:Brexit/Post-referendum_opinion_polling:_Right-Wrong — Preceding unsigned comment added by 212.219.101.97 (talk) 15:42, 21 October 2019 (UTC)

Long-term, I think it should be split off into a new table, along with other polls that do not explicitly ask how respondents would vote (although I acknowledge that whilst most polls not explicitly asking about a referendum find results comparable to polls that do, this particular poll – thanks to its unusual wording – is the only major outlier serving as evidence that such a distinction is worth making). Short-term, I believe the poll, by asking about membership of the EU rather than about the 2016 vote, fits better into this table than any other, so until a new table is made, my opinion is that it should remain.—AlphaMikeOmega (talk) 22:11, 23 October 2019 (UTC)

Are you sure? To my mind "do you support or oppose the UK abiding by the referendum result and leaving the EU?" (ComRes) seems a lot closer to "do you think Britain was right or wrong to vote to leave the EU?" (YouGov) than it does to "If there was a referendum on the United Kingdom's membership of the European Union, how would you vote?" (YouGov). The former questions refer to the past referendum, whereas the latter question refers to a possible future referendum. TimAngus (talk) 10:27, 24 October 2019 (UTC)

I see what you mean, but I can't say I agree. As a lesser point, the table (currently) includes polls asking "Should the United Kingdom remain a member of the European Union, or leave the European Union?", which is closer to the ComRes question than the "If there was a referendum..." question is. My main objection, though, is that the ComRes question is (as I see it) roughly equivalent to "Do you think the UK should leave the EU given the referendum result?", where respondents are not judging a decision (as in "Do you think Britain was right or wrong to vote to leave the EU?") but directly advocating a policy/action (as in the other questions in the Remain/Leave table). This is the distinction I used (focusing on "leaving the EU"), whilst I guess you focused on "abiding by the referendum result". I reckon most respondents will have seen this question in terms of what policy they would pursue, as opposed to whether the vote to leave was wrong.—AlphaMikeOmega (talk) 00:52, 27 October 2019 (UTC)

Hmm, I see where you're coming from. I guess it's arguable both ways; another reason to put it in a separate table, as you say. I think we can at least agree that the question itself is ambiguous, so maybe it shouldn't be included at all. TimAngus (talk) 11:44, 28 October 2019 (UTC)