EU Referendum online polling 'the gold standard' following accurate prediction
So that was an electoral event and a half. Drama, debate, endless punditry, massive media coverage - the EU referendum had everything. Plus of course, there were seemingly endless polls.
Nearly all the polling was of Great Britain, with LucidTalk carrying out monthly Northern Ireland tracker polls.
So how did we do with the polling in Northern Ireland?
Well LucidTalk's last pre-EU referendum poll, which was held a week before the vote (15th to 17th June - 1,075 NI representative sample), predicted a Northern Ireland 'Remain' victory with that actual poll forecasting a result of 52% Remain to 38% Leave with 10% undecideds. See Table One.
This translated into a final forecast of 57% Remain to 43% Leave - if we exclude the undecideds (or we assume in the end the undecideds split 50/50 in the actual vote).
In fact the final Northern Ireland result was 56% Remain, 44% Leave - so LucidTalk, with their last pre EU Referendum poll, forecast the result to within 1% of error.
LucidTalk has been running these regular monthly online Northern Ireland EU referendum tracker polls with our representative opinion panel since last December.
The benefit of this regular tracker polling is that trends can be determined, and there was a steady trend towards Leave in all the polls since January - see Table 2 above.
It was never going to be the case that Leave were ever going to win in Northern Ireland, but all the LucidTalk monthly polls tracked the trends accurately and steadied at around 56-57% Remain (excluding undecideds) in both the May and mid-June polls i.e. the last two polls before the actual referendum.
This undoubtedly proves that online internet polling, is now the gold standard for accurately ascertaining Northern Ireland opinion.
In addition, it's been proved from the national UK referendum result that robust, professional, and targeted online polling, is going to be the general gold standard for polling in the future.
Remember the majority of the national UK online polls consistently predicted a Brexit, albeit within the context of a close result - and the UK result was close.
Indeed the last two polls late on the night before the referendum (from Opinium and TNS: Internet - online) forecast a 51% to 49% Brexit win.
Unfortunately it seems pure telephone polling is now on its way out - it's just too difficult, and too costly now, to obtain representative samples.
This is not the fault of the pollsters. The telephone channel has been overly affected by the PPI sales calls, and 'we'll fix your computer' scams etc. It's hard to gain the trust of respondents via the phone when they're expecting you to ask for their credit-card details at any stage.
Mind you polling is still needed for all electoral events, and particularly referendums, as voters need to know roughly were the pendulum is in terms of opinion so they can make a reasoned decision.
But polls are approximate. All polls are meant to do is give you a feel for opinion at a point in time - a snapshot, so that further debate and consideration can then take place as part of the wider campaign. As well known UK pollster Peter Kellner said polls were like "looking at a grainy black-and-white photo of the Grand National half way through the race".
We also now hear on the media the idea of a second EU referendum due to the fact that a lot of people (apparently) may have changed their mind since the referendum result.
This second referendum idea could be justified if there had been no polling.
With no polling there could have been complaints afterwards from people (e.g. 'Leave' voters), who perhaps only wanted to send a warning to Westminster, saying things like 'if only I had known the way it was going I wouldn't have voted Leave....', and maybe 'someone told me Remain was going to win easily so I voted Leave to stop Remain winning by too much - I never really wanted Leave to actually win', and so on.
As such, in terms of the EU referendum I have no time for those (well they exist according to the media) who say they wouldn't have voted Leave if they had known Leave were actually going to win.
Everyone should've known from the polls that it was going to be a very close result either way - which it was.
The trend towards Remain in Northern Ireland (as shown by the polls - Table 2) also reflected the trends UK wide.
With the Northern Ireland demographics the way they are, Remain should have scored at least 60-65% in the referendum - why they didn't, and why 30,000 - 50,000 voters switched from Remain to Leave in the last months of the campaign is a big question.
But maybe that's a debate for another day.
Bill White is Managing Director of Belfast polling and market research company LucidTalk. You can follow LucidTalk on Twitter: @LucidTalk.