Looking at a mound of opinion poll figures about September’s independence referendum, John Curtice, Scotland’s leading opinion poll expert, displays a dry sense of humour. “Where are we? The answer is we are not quite sure,” he says.

A few things are clear: the gap in the numbers between those who will vote Yes or No has narrowed over the last two years. However, the half-dozen companies that are consistently polling are coming up with significantly different numbers.

For instance, polls from Panelbase – which are paid for by the pro-independence campaign group, Yes Scotland, but which are published widely in Scotland – have consistently found the strongest support for independence.

Since January, it has, on average found that 46 per cent of voters are ready to vote Yes, compared with 45 per cent found by ICM, 44 per cent by Survation, 41 per cent from TNS/BMRB, 40 per cent by YouGov and 38 per cent by Ipsos/Mori.

Last September, Panelbase banned new members from joining the list of internet respondents it tracks to come up with its numbers, following concerns that it had been skewed by Scottish National Party members signing up in droves.

The controversy erupted after Panelbase reported that the Yes side had nudged ahead, by 44 per cent to 43 per cent in the campaign for the first time – just days after YouGov had forecast a 59 per cent-26 per cent No result and TNS-BMRB gave the No side a 47 per cent-25 per cent lead.

Who is wrong

Since then, Panelbase has used different panels of voters, but has found similar results to its original figures – while the less optimistic numbers for the Yes side found by TNS, YouGov and Ipsos have continued apace.

“We don’t know who is right and who is wrong. Most polls give the ‘don’t knows’ at about 15 per cent,” says Curtis, who is the professor of politics at Strathclyde University in Glasgow. “That’s not particularly high; House of Commons polls six weeks out from an election are much higher, but people in Scotland have had 40 years to think about this,” he says.

However, longer-term research may offer a better guide. The Scottish Social Attitudes Survey has tracked opinion since the Scottish Parliament was reborn in 1999. Opinions have oscillated over that time, but within relatively narrow bands.

Between 45 and 50 per cent want more devolution, especially over taxation, while the numbers supporting full independence from the rest of the United Kingdom "has consistently been around 30 per cent per cent", according to a review by the House of Commons Library.

“The data suggests that public attitudes have remained relatively stable over the time,” it said, in a paper published earlier this month, adding that just one in 10 wants Holyrood’s powers to stay as they are, or to go backwards.

Curiously, however, more detailed research displays some contradictions. Most Scots agree that most decisions, including those over tax and welfare, should be taken in Edinburgh, not in London, bar those dealing with foreign policy and defence.

“But underneath the bonnet, they have not looked into the implications of that,” says Curtice. “Should there be a difference in the basic rate of tax between Scotland and the rest? There’s a 50-50 divide on that.

“Should public services be funded only by Scottish revenues? Only 30-40 per cent want that; 60 per cent want access to UK funding,” he says.

Two years of debate have not necessarily brought clarity, or knowledge according to research by the University of Edinburgh and the University of Stirling. Research examining voters' understanding of the Scottish government's White Paper, showed that nearly half did not know if Alex Salmond was proposing that an independent Scotland should continue to use sterling. He is. Nearly two-thirds were not sure if Scotland would still send MPs to London. It would not, once a separation deal was reached.

“Sixty per cent don’t know whether Scotland will go back to imperial measures after independence,” says Prof Ailsa Henderson of the University of Edinburgh, adding that people intending to vote Yes are more likely to get the answers right.

Dual identity

Most Scots have some sense of dual identity – both Scot and British – but the strength varies according to the individual. However, the more Scottish a voter feels the more likely he or she is to vote for independence, though this statement is subject to some qualifications.

Older people, women and, most strikingly of all, young people aged between 16 and 24 – the very ones that the Scottish National Party were confident would deliver a Yes majority – are less likely to favour Salmond’s dream.

However, the numbers of 18-year-olds ready to vote Yes on September 18th has gone up: by eight percentage points, says Dr Jan Eichhorn of the University of Edinburgh.

“But there is a still a majority voting against,” he said.

Interestingly, younger voters were shown to be surprisingly traditional when it comes to judging the trustworthiness of the information they receive, depending on where it comes from: TV, radio and newspapers are still seen as the most important sources.