Talk:2016 United Kingdom European Union membership referendum

Petition for a new referendum

YouGov: "British public opposes a second referendum by almost 2 to 1"

"The latest research from YouGov/Channel 5 Shows ... that most British people (58%) oppose holding a second referendum. This includes not only 91% of Leave voters, but also 29% of Remain voters. 11% don't know.Even in an extreme situation, such as the break up of the United Kingdom, most people (51%) still oppose holding a second referendum. In the event of Scottish independence, only 30% of people would support holding a second referendum."

--77.10.21.166 (talk) 23:14, 1 July 2016 (UTC)

Let's wait for the press to pick it up (probably tomorrow) and we'll insert it somewhere into the section. This is Paul (talk) 23:42, 1 July 2016 (UTC)

Referring to 51% of an unspecified polling sample as "most people" seems a strange choice of words. Makes me wonder about the quality of this poll.Charles (talk) 08:34, 2 July 2016 (UTC)

"Most" as a quantifier: One uses most to refer to the majority of a group of things or people or the largest part of something. --77.10.28.236 (talk) 13:35, 2 July 2016 (UTC)

I think this is acceptable usage for the organization producing the survey, because if the sample is designed to be representative, they generally assume that statements about the sample apply to the population (depending on the confidence interval and statistical significance). I don't think the usage would be appropriate for Wikipedia, because most tends to imply a (very?) significant majority (significant as used in common parlance), and would therefore be misleading.. --Boson (talk) 14:02, 2 July 2016 (UTC)

"shows ... most British people"? What a ridiculous claim. What was the sample size and how was it selected? To find out what "most British people" want, you'd have to ask them, wouldn't you... using, I don't know, maybe a national referendum? And even then millions might not vote, and you'd be left wondering is if it was a fair representation, wouldn't you? 217.38.93.131 (talk) 14:07, 2 July 2016 (UTC)

"The latest research from ..." You do understand the concept of survey methodology, don't you? --77.10.28.236 (talk) 15:25, 2 July 2016 (UTC)

So what was the precise methodology employed here? A great deal of political opinion polling is regarded as unreliable. If this was a particularly robust study, that could be made clear? Martinevans123 (talk) 15:29, 2 July 2016 (UTC)

It is not for us to question the methodology of this survey result. This YouGov survey must be added to the article, otherwise please also remove the YouGov results that are used in the opinion polling section, as well as in the Opinion polling for the United Kingdom European Union membership referendum article. --89.204.130.21 (talk) 16:33, 2 July 2016 (UTC)

We actually question methods all the time, even in areas such as science. This is part of responsible editing. However, I take your point about content from the same organisation being included in some sections but not others. DrChrissy ^(talk) 16:47, 2 July 2016 (UTC)

I would imagine that sample sizes vary widely, even for polls taken by the same organisation. I'm not very convinced by the argument that if an organisation has one poll included in an article, it should therefore have all its others reported too, in whatever section they best fit. The full results page here shows that the sample size was "1760 GB Adults". There seem to be figures by region, although they total 1759! I don't see any explanation of "weighted sample", so I assume there was no weighting. Martinevans123 (talk) 17:50, 2 July 2016 (UTC)

Come on, folks ... seriously?! You don't want to add this YouGov survey to the article, because it does not reflect your personal opinion?
A wp:rs that is used all over the wp?
"I disapprove of what you say, but I will defend to the death your right to say it"
Evelyn Beatrice Hall (often misattributed to Voltaire) --82.113.99.217 (talk) 17:57, 2 July 2016 (UTC)

Yes, but wasn't it Evelyn Beatrice Hall who originally attributed it (or the thought behind it) to him? --Boson (talk) 18:11, 2 July 2016 (UTC)

This might be of interest. Yougov write "we publish a margin of error of 3% on most polls, which we encourage media to make reference to", so we should perhaps mention that. --Boson (talk) 18:05, 2 July 2016 (UTC)

I'm sure S. G. Tallentyre would be proud of you. So, YouGov tell us that "only the online polls revealed the true state of the race". Was this one on-line too? Martinevans123 (talk) 18:11, 2 July 2016 (UTC)

@Martinevans - I was actually imagining the opposite to you, i.e. an organisation such as YouGov would use a similar number of respondents. I have no evidence to support this. I would have thought there is a minimum number of respondents to be considered representative of a population and going above this would simply be a case of diminishing returns. DrChrissy ^(talk) 18:16, 2 July 2016 (UTC)

Could someone explain to me what is meant by "we publish a margin of error of 3% on most polls"? Are they saying, "our results may be 3% wrong in either direction"? i.e. three out of every hundred people we ask may have given us the wrong answer? How do they test for that? (by the way, the margin of difference in the actual referendum result was 3.78%) Martinevans123 (talk) 18:25, 2 July 2016 (UTC)

I take this to mean that (if you assume that the sample was truly random) there is a 95% chance of the sample results correctly predicting the population results within 3% either way. That would probably hold true for a sample size of 1000. Assuming that the sample is half male and half female, that would probably mean that the margin of error for predictions about males or females is 4%. Since we can assume that the sample is not truly random, we can assume that the margin of error (at 95% probability) is higher. --Boson (talk) 19:33, 2 July 2016 (UTC) PS: If you used a sample of 2000, the margin of error would probably be more like 2%. If the respondents gave false answers, I don't think you can say anything about the margin of "error". --Boson (talk) 19:37, 2 July 2016 (UTC)

To what does that "error" refer? That a response given was inaccurate or that a response, correctly given, was reported wrongly? I don't see "YouGove" make any mention of a 95% confidence value. Martinevans123 (talk) 20:04, 2 July 2016 (UTC)

No, there is no error in the sense of a reply or report being wrong. "Margin of error" refers to the extent to which the result for the whole population can deviate from the result for the sample, based on statistics, usually assuming that the sample is random. The figure is not very meaningful without the confidence level, which is conventionally 95%. Yougov explain the margin of error: "It is, of course, important to note that all polls have a margin of error of around plus or minus three points. This means that 95% of the time we can be confident that the real figure is within three points of what polls show ...".--Boson (talk) 21:07, 2 July 2016 (UTC)

Thanks for the clarification. But there is error, isn't there. A poll respondent may not answer truthfully, for whatever reason. I've not looked to see in the relevant article(s), but do you know if we present those studies that support this "three points" claim? I think the public perception (certainly mine) is that polls may be much more inaccurate than this. Martinevans123 (talk) 09:53, 5 July 2016 (UTC)

Margin of error is a specific statistical concept that describes the uncertainty in the result due to sampling error. It does not cover other possible sources of error, of which there are many. Thus it is an underestimate of the true uncertainty in the result. It is, however, the only bit that we can calculate explicitly and it is a standard term. I think we should thus include it, but with internal links to the relevant Wikipedia articles on what it all means. Bondegezou (talk) 10:07, 5 July 2016 (UTC)

So there is no need for any specific study that proves that a given poll exhibits normal sampling error? Every poll will show a three point margin? I thought that depended on the size of the sample - a 100% sample would show no sampling error. Or Are YouGov saying that for all their polls they choose a sample size that will mean a 3% sampling error? I'm not sure that's correct. Martinevans123 (talk) 10:27, 5 July 2016 (UTC)

Each poll with have its own "margin of error" based (normally) on the sample size. Polling companies often pick a sample size to give a 3% "margin of error", so we see lots of polls with that "margin of error", but may not. The polling company's published report of the poll will give the exact figures. Some Wikipedia articles on polling results, e.g. Statewide opinion polling for the United States presidential election, 2016, give margins of error for each poll. Presuming the poll has been carried out to certain standards, the "margin of error" will be correct and other sources of error should be minimised. However, recent elections have shown that polls can get it wrong and that they're not successfully minimising other sources of error as much as they should: this probably reflects problems in sampling, although that debate is ongoing. Bondegezou (talk) 10:35, 5 July 2016 (UTC)

slightly off-topic discussion about sample sizes and statistical distributions
The following discussion has been closed. Please do not modify it.
1760 doesn't look like a very large sample to me. Not when the population of referendum votes was 33,577,342. In fact that’s a sample size of 0.0000524%. Martinevans123 (talk) 10:50, 5 July 2016 (UTC) You'll have to trust me on the maths, but the size of the population doesn't matter much (if the sampling is done randomly). Indeed, we assume the population is infinite when we do the "margin of error" calculations. 1760 is a pretty big poll. Bondegezou (talk) 10:58, 5 July 2016 (UTC) If that's "pretty big", what is the smallest sample size that would ensure a 3% margin of error (for this population, or indeed for an infinite one). Looking forward to your statistical proof. Sorry if this is considered "off topic". Martinevans123 (talk) 11:04, 5 July 2016 (UTC) A sample size of about 1000 delivers a margin of error of about plus or minus 3%. The requisite maths is in the margin of error article. A larger sample size gives a smaller margin of error, following an inverse square root rule. We are getting off topic if we're discussing your personal understanding of statistical theory: we follow reliable sources, the margin of error article is replete with citations and I am unclear at this point what editing changes you wish to see here. Bondegezou (talk) 11:46, 5 July 2016 (UTC) I am amazed that "a sample size of about 1000" is large enough to ensure that the margin of error for any population whatsoever including an infinite one is acceptable. It seems somehow counter-intuitive. But, as you suggest, the article does say: "A random sample of size 1600 will give a margin of error of 0.98/40, or 0.0245—just under 2.5%." So that's just the way statistical theory works, I guess. Perhaps a link to "margin of error" would be useful somewhere? Not just for my benefit, but for that of the general reader. I must apologize for my ignorance, but would a normal distribution equate to a 50/50 split of yes and no votes and does the actual result mean that the population distribution is actually 51.89/48.11 (notwithstanding non-random error)? Does this make any difference? Would that help a reader here at all? Thanks for your clear replies. Martinevans123 (talk) 12:30, 5 July 2016 (UTC) Martin, I think these data would not be analysed using a normal distribution. The normal distribution requires a continuous variable such as age or shoe-size. Here we are faced with a choice of only Leave or Remain, so I think it would be analysed using a binomial distribution and the associated statistics. DrChrissy (talk) 16:44, 5 July 2016 (UTC) But people can vote only once, so not a Hypergeometric distribution instead? Martinevans123 (talk) 18:05, 5 July 2016 (UTC) I suspect Bondegezou has now given up with me in disgust. Where shall we hat this as off topic? I thought that this was a distribution I had not heard of before, but our article says the Fisher's exact test is based on this. Whatever, the non/replacement is clearly critical and you are correct to make the point. I'm happy to have this hatted if you are. DrChrissy (talk) 18:33, 5 July 2016 (UTC) Hoorah! I always use a fisher's exact test, it's so much safer. I have now asked a question over a WP:RD/MA, for my sins, where I expect to totally be baffled by an expert answer. Martinevans123 (talk) 18:56, 5 July 2016 (UTC)

DrChrissy ^(talk) 23:26, 5 July 2016 (UTC)

So let's chuck a few more stats into this discussion. This report from The Independent tells us about an Opinion poll suggesting that 7% of those who voted to leave the EU (an estimated 1.1 million) now regret their decision, while this article from the Evening Standard cites a BMG poll that claims it is 13%. However, the latter poll says that most people are against a second referendum. I suspect there are a few polls floating around at present, all of them giving differing results. If we were to mention the YouGov poll I would suggest including the others as well. This is Paul (talk) 18:41, 2 July 2016 (UTC)

You know something is wrong, when this YouGov poll would have been included in the article no questions asked if it suggested "most British people support holding second referendum". Antiochus the Great (talk) 20:08, 2 July 2016 (UTC)

Nice logical proof there, Antiochus the Great. Comforting that you know. Martinevans123 (talk) 20:12, 2 July 2016 (UTC)

The "Independent" opinion poll has already been (indirectly) added to the article: "Labour MP Geraint Davies has also suggested that a second referendum would focus on the terms of an exit plan, with a default of remaining in the EU if it were rejected. Citing a poll published in the week after the referendum that indicated as many as 1.1 million people who voted to leave the EU now regretted their decision, he tabled an early day motion calling for an exit package referendum, feeling it would "pull [Britain's] future out of the faire"
The article section "Petition for a new referendum" is highly biased. There is a lot of pro-REMAIN POV-pushing going on here. --77.10.28.236 (talk) 22:53, 2 July 2016 (UTC)

We do have at least two high-profile figures who rule it out, as well as David Cameron himself. You could always find some views of people from the Leave camp then post them here for consideration (but nothing from The Sun or The Daily Mail please). I should add here that I have been looking myself, but to me it seems as though post-referendum, the Leave side have been far less vocal on this particular subject. This is Paul (talk) 23:08, 2 July 2016 (UTC)

And nothing from UKIP Daily either please. Martinevans123 (talk) 23:17, 2 July 2016 (UTC)

Or their best friends. This is Paul (talk) 23:19, 2 July 2016 (UTC)

Well, there is the YouGov opinion poll. That should do. --77.10.28.236 (talk) 23:34, 2 July 2016 (UTC)

So wait for the media to run with it. One wonders why they haven't; it ought to be music to the ears of some news outlets. This is Paul (talk) 23:44, 2 July 2016 (UTC)

There is NO reason to wait for the media to run with it. Otherwise please also remove the YouGov results that are used in the opinion polling section, as well as in the Opinion polling for the United Kingdom European Union membership referendum article. --77.10.28.236 (talk) 23:56, 2 July 2016 (UTC)

Sorry, but I would argue there is a need to wait. That section is about opinion polling, while this particular section is a commentary on whether or not there should be a second referendum. My own personal concern is that to add it rather than a source discussing it (and which places it into context) could look like original research. This is Paul (talk) 00:03, 3 July 2016 (UTC)

There's this story, YouGov: Britons don’t want a second referendum from New Europe, if that's any good. Incidentally, that article also asks whether the survey is credible. I know the ip will jump on it straight away (or maybe dismiss it as biased towards remain), so let's have the views of everyone else. This is Paul (talk) 00:10, 3 July 2016 (UTC)

Ah, ok, but writing "YouGov also found that euroscepticism correlates with people of lower income and that "higher social grades are more clearly in favour of remaining in the EU", but notes that euroscepticism also has strongholds in "the more wealthy, Tory shires" in the Opinion polling section, using a YouGov reference is not?! This is really getting ridiculous. This kind of biased editing is not doing any favour to the REMAIN campaign, nor will it help to keep the UK in the EU. All it does, is to create an unbalanced, biased article. --77.10.114.96 (talk) 00:19, 3 July 2016 (UTC)

City A.M. is now reporting this as well (see Sorry, petitioners: Most Britons don't support the idea of a second EU membership referendum, so we could use that. BTW, I'm not aware of this page being used as a campaign tool to keep Britain in the European Union. I think you probably need to calm down and take a stiff brandy. This is Paul (talk) 12:46, 3 July 2016 (UTC)

(But none of that French rubbish.) Martinevans123 (talk) 12:55, 3 July 2016 (UTC)

• This section seems to verge into undue territory, with far too many quotes, which ought to be very rare in Wikipedia articles. There is now almost more quotes referring to this petition than quotes referring to the discussion prior to the actual referendum, so I think this is also part of WP:Recentism. I would suggest we remove the Blair quote that was inserted just now, and in general this section should be shortened rather than expanded. Iselilja (talk) 16:25, 3 July 2016 (UTC)

But what if the outcome has proved to be much more controversial than the run-up? The result was a great surprise to many, so that would hardly be surprising would it? Martinevans123 (talk) 17:31, 3 July 2016 (UTC)

Since Blair is a former prime minister I think his thoughts on the topic are important, although perhaps we can summarise them. There are actually two issues here–the petition itself, and a debate over whether or not there should be a second referendum. The two are connected, but not mutually exclusive. It may be worth splitting the section into two sections, since not every view expressed specifically references the petition. This may also help to address any concerns of bias. I'd be happy to draft something, although I probably won't have time to do that today. But maybe later on in the week This is Paul (talk) 17:40, 3 July 2016 (UTC)

I think that's a very fair and reasonable suggestion, Paul. Martinevans123 (talk) 17:44, 3 July 2016 (UTC)

Great, I should be able to do it on Tuesday as I'm not around much tomorrow. It doesn't look like it'll be too difficult actually, as the first half talks about the petition, before the second part opens up into a wider debate about whether or not there should be another referendum. I also trimmed the Blair comment a bit to take out the Brexiteers lied to us stuff. This is Paul (talk) 19:04, 3 July 2016 (UTC)

Splitting out Reactions to the result section

This article is getting very long, and some of the reactions section is quite secondary - the Conservative leadership election was certainly caused by the referendum, but Boris being knifed by Gove is quite tangential to the referendum itself. Therefore, I will split it out to Aftermath of the United Kingdom European Union membership referendum, 2016, unless anyone has any objections. Smurrayinchester 11:28, 4 July 2016 (UTC)

I agree it is getting way too long. Is the intention to create 2 articles, one pre-referendum and one post-referendum? DrChrissy ^(talk) 18:23, 4 July 2016 (UTC)

A good idea. No objections, thanks. Martinevans123 (talk) 19:44, 4 July 2016 (UTC)

Unfortunately, we've not got a messy situation where material is repeated across both articles. We should drastically cut down aftermath coverage here and point people to the separate article. Bondegezou (talk) 09:37, 5 July 2016 (UTC)

Yes, I agree. I think you mean "now". Martinevans123 (talk) 09:42, 5 July 2016 (UTC)

I did indeed mean "now"! I've done some more chopping of the article. Bondegezou (talk) 09:44, 5 July 2016 (UTC)

Mishcon de Reya

Mishcon de Reya is taking pre-emptive legal action against the government, following the EU referendum result, to try to ensure article 50 is not triggered without an Act of Parliament. Many sources for this including The Guardian, The Daily Telegraph, BBC News. This seems a significant development. But not sure if it belongs in this article or in United Kingdom withdrawal from the European Union. Martinevans123 (talk) 12:21, 4 July 2016 (UTC)

...Or perhaps in the "Aftermath..." article? We need a little rationalisation here. What about 3 articles, i.e. pre-referendum, post-referendum and United Kingdom withdrawal from the European Union which only applies once article 50 is invoked. Technically, we still might not be leaving the EU. DrChrissy ^(talk) 18:30, 4 July 2016 (UTC)

Yes, of course. Martinevans123 (talk) 19:45, 4 July 2016 (UTC)

NPOV???

I don't think the non-legally binding thing in the beginning is OK. It should be placed somewhere else, maybe in the explanation below. I don't see any other referendum page saying that the referendum is non-legally binding or is binding. In the Alternative Vote Referendum, I don't see a starter saying that that referendum was legally binding (correct me if I'm wrong).
49.200.244.109 (talk) 12:06, 7 July 2016 (UTC)

I disagree. This is very important, and it should therefore appear in the lede.Zigzig20s (talk) 12:22, 7 July 2016 (UTC)

You mean lead. So, maybe you can edit the article pages for the other referendums like AV and ensure that the rule is followed everywhere? It looks odd on a single article. If you want it to stay here I suggest we take it as precedent and reform the other referendum articles along with it.

No, xe actually meant lede.--143.159.32.60 (talk) 10:13, 8 July 2016 (UTC)

49.200.244.109 (talk) 12:29, 7 July 2016 (UTC)

Should we allow popular sentiment to get into the article? Maybe put the legally binding note at the end of the lede.

--188.165.30.149 (talk) 12:34, 7 July 2016 (UTC)

I don't have a strong opinion on this, but I do think it would be best not to mention this fact in the intro if all the other British referendum articles also don't include this fact in their intros. --1990'sguy (talk) 16:32, 7 July 2016 (UTC)

I disagree. I think it's very important to keep the information in the lead. I have zero interest in editing articles about other referenda, but the IP addresses above could do that instead.Zigzig20s (talk) 16:52, 7 July 2016 (UTC)

I agree content about it being non-binding should be in the lead/lede (either word can be used). It is a vital piece of information. I was not aware of this until after the referendum and I think many voters were the same - I'll see if I can find an RS for this. DrChrissy ^(talk) 17:06, 7 July 2016 (UTC)

I find the idea that it's "POV" to state a fact that the referendum isn't legally binding frankly bizarre. The other British referendums may not have legally binding, but mostly opted to preserve the status quo, with the votes for change long implemented Dtellett (talk) 20:10, 7 July 2016 (UTC)

Yes, I was having that thought myself about the heading of this thread. The OP has made very, very few edits and could be considered as new and making an acceptable mistake due their lack of experience. Does anyone know if it is acceptable to make a good faith change of a thread heading? DrChrissy ^(talk) 20:22, 7 July 2016 (UTC)

Issues: A number of voters didn't receive their voting times on time

I have noticed that there is no mention of complaints from a number of voters who didn't receive their voting times on time, or received it too late, or their votes were disposed by the Spanish postal service, since the return envelope sent to voters by the Electoral Board had a free postage stamp, but the stam was valid for the UK only. the Spanish post disposed the letters as sent with no postage fee. More here: http://www.theolivepress.es/spain-news/2016/06/25/thousands-of-expats-had-problems-sending-and-receiving-postal-vote-in-eu-referendum/ — Preceding unsigned comment added by Tony.bangor (talk • contribs) 21:35, 7 July 2016 (UTC) Expats also plan a lawsuit against the Electoral Commission for denying them the right to vote: http://www.msn.com/en-gb/news/uknews/brexit-british-expats-considering-lawsuit-over-missing-postal-votes-after-hundreds-denied-say-in-eu-referendum/ar-BBu43MP?li=AAdeCd7&ocid=spartanntp --143.159.32.60 (talk) 18:04, 8 July 2016 (UTC)