EU referendum: Why there is no exit poll
Unlike at general elections, there will be no official exit poll for the EU referendum.
Differently to opinion polls, which ask people how they intend to vote, these polls they ask how a voter actually did vote.
They rely on people being camped outside of polling stations asking real voters to cast paper ballots and as they are so labour intensive, they cost a lot of money.
Usually, ITV, BBC and Sky put some money in together and pay for it. At a general election the fieldwork is conducted, then the data is analysed and seat numbers projected.
But the problem arises from the fact that the exit polls, such as during the general election, go back to the same polling stations election after election and essentially measure change in votes from the previous time.
Read more: What the polls aren't telling you
That is, the projections are made by looking at changes in support at those polling stations since the exit poll five years prior.
As Anthony Wells, research director at YouGov, has written: “[Analysts] will look at patterns of change (e.g, are there bigger or smaller changes in different regions, or where there are different parties in contention) and use that to project the swing across different types of seat.
“While the overall swing can be used to come up with national shares of the vote, that’s very much a by-product, at its heart the exit poll is all about change since the last election.”
However, because exit polls are designed to measure change, this is a much more difficult task when it comes to a referendum, as there is nothing to compare it against. That's why there wasn't one for the Scottish independence referendum or the electoral system referendum.
You could, say, try and compare it against the 1975 referendum on the EU, but a lot has changed (not to mention the youngest voter from then would be 60-odd now).
Read more: Davidson more likely to succeed Cameron after Wembley TV debate
Wells says that someone could still in theory conduct an exit poll, but nobody would know if it is accurate or not.
“Instead of looking at swing, one could try and sample from a random selection of polling stations and work out national shares of the vote. This used to be how exit polls were done in this country before the current method was developed,” he said.
“Don’t fall into the trap of thinking it would necessarily be as accurate as the recent BBC exit polls though: back in 1992 the exit poll was conducted this way, and was almost as inaccurate as that year’s pre-election polls.”
YouGov will however publish a poll showing how people have voted in the referendum shortly after polling stations closes, based on responses from a pre-selected group of people viewed as representative of the population on how they actually voted in the referendum.
That could give a good indication, given its fairly accurate prediction of the 2014 Scottish independence referendum. Back then, it predicted 54 per cent to No and 46 per cent to Yes, just outside of the 55-45 result that resulted in Scotland remaining within the UK.
So, while we may have an idea, we’ll just have to wait for the actual count to see what the final result is.