[hnlmorg]

Good question. There's a few reasons I use that phrase and I agree with you that, from a statistical perspective, it isn't the correct usage of the term:

- There was a lot of misinformation spread (on both sides of the debate). So people were voting for issues that didn't apply.

- There were a lot of protest votes from people who assumed "Remain" would win and who also did want to remain but wanted to voice disagreement with the government and/or concern about unconditional relationships with the EU.

- There options were too vague. "Remain" largely meant keeping the status quo but some read it to mean including countries like Turkey (who were extremely unlikely to ever join the EU anyways). Likewise "Exit" meant different things to different people. Some people wanted a "hard exit" (no EU trade deal). Some people wanted to stay connected to the EU but to have a revised deal. People were voting for the same options but expecting different outcomes.

And as a result of this there had been a high number of people voice regrets about the vote they had cast in post-referendum opinion polls.

Sadly we will never know just how accurately the results reflected peoples true opinions because all calls for a follow up vote had been literally laughed at. However subsequent general elections have demonstrated just how far from settled the debate was.

[Nursie]

I can accept that there were confounding factors which may have affected the result and how enduring a picture of UK attitudes it might be. I wonder if any referendum could be 100% free of those factors?

I just object to the use of the phrase - there is no margin for error here, it forms the full picture of how the population voted, not an estimate.

People use this phraseology, and I've even seen the term "not statistically significant" bandied around as well, to try to say that no conclusions could be drawn, as if it's a scientific paper with a sample in it.

[hnlmorg]

If you acknowledge the results are not a 100% reliable indicator then there is a margin for error. The key difference is whether you look at it in the context of

  "This result is reflective of what the majority _want_"

or

  "This result is reflective of what the majority _voted for_"

The latter isn't always the same as the former in cases where information isn't clear or where the poll options are too vague. Both of which plagued the EU referendum.

I do get your point that "margin for error" is a statistical term and it is not technically being applied correctly here. But the crux of what that term refers to does still apply to the former context above.

As for why the context matters: because every conversion that happened since focused on the former point with MPs even coining the phrase "the will of the people" yet the people's "will" was still unclear.

That all said, I don't think there is any way such a referendum would have worked on a topic as diverse and complex as our relations with the EU.

[Nursie]

> If you acknowledge the results are not a 100% reliable indicator

But they are a 100% reliable indicator of how people would vote, because they did vote that way. And I'm really not sure it's valid to start second guessing that what people actually wanted is different to that, because you then have to second guess every vote and really, where does thaty leave democracy?

> That all said, I don't think there is any way such a referendum would have worked on a topic as diverse and complex as our relations with the EU.

Very much agreed.

[hnlmorg]

> But they are a 100% reliable indicator of how people would vote, because they did vote that way. And I'm really not sure it's valid to start second guessing that what people actually wanted is different to that, because you then have to second guess every vote and really, where does thaty leave democracy?

Normally I'd agree. But as I said, there were concerns going into the vote that the options were too vague and as a result opinion polls post-referendum showed that a considerable number of people voted for options that didn't support their opinion.

This wasn't a typical poll where you vote for a party to entrust or a narrowly defined set of options. This referendum was vague and poorly defined. In cases where that's been a problem in Europe those respective countries have then responded with second referendum after communicating clearer information, reviewing the poll options and taking other measures to ensure they accurately capture public opinion. Instead the UK called a general election and as a result muddied the conversation even further with topics like education and healthcare.