At first glance the message from the opinion polls on the state of the referendum race is clear and consistent.

But when we look at the figures a little more closely some serious uncertainties about exactly where the balance of opinion currently lies begin to emerge.

So far, 16 polls have posed the question that will appear on the referendum ballot paper: 'Should Scotland be an independent country?' In those polls, on average, 33% have said they would vote Yes. Just over half of them were conducted before MSPs went on their holidays and these put the Yes vote at 32%. In more recent polls the figure has edged up to 34%. So minor a movement is hardly enough to set the heather alight.

Indeed, it looks like no more than ?¨a return to where we started. ?¨During the course of last year, a dozen polls posed the slightly different referendum question originally proposed by the Scottish Government: 'Do you agree that Scotland should be an independent country?'. These polls produced an average Yes vote of 34% too.

In short, after more than 18 months of sound and fury it is quite possible that nothing has changed at all.

However, stable though they might be, these figures are potentially rather misleading. Although only around a third of Scots are currently minded to vote Yes, that does not mean that everyone else is determined to vote No.

Some people - typically a little under a fifth - say they do not know how they are going to vote. The average No vote in fact currently stands at 48%, not 66%.

To avoid painting a potentially misleading picture we should leave aside the Don't Knows and the Won't Says, just as pollsters usually do when telling us where the parties stand in the battle for Holyrood or Westminster. When we do that we discover - as our poll of polls shows - that Yes are on average on 41%, with No on 59%.

That still implies that the Yes side has quite a challenge on its hands, but it does mean that the race is rather closer than many a headline often implies.

And it may be even closer than that. One polling company in ?¨particular has stood out from the crowd. In five polls it has conducted so far this year Panelbase has on ?¨average put the Yes vote (once the Don't Knows are excluded) on 46%. ?¨In one instance it even put Yes narrowly ahead, on 51%. True, the 51% result at least was probably skewed by the decision of those ?¨who commissioned the poll, the SNP, to pose a couple of potentially leading questions immediately before ?¨asking people how they would vote in the referendum.

But even if we leave that poll ?¨aside, Panelbase still put the Yes ?¨vote on average at 45%. That implies there could still be quite a lot to ?¨play for.

Two other companies, Angus Reid and ICM, have both produced figures that still give the Yes side some hope, averaging a 41% Yes vote. But everyone else has consistently put the Yes vote much lower - on average at just 35%. If they are right the Yes side would seem to have an impossible task on its hands.

As you can imagine, there is furious debate going on among pollsters about who is right and who is wrong. It has not as yet, however, reached any very firm conclusion.

The wise punter will hedge his bets, aware that the race might perhaps be close enough that the next 12 months could still just make ?¨a difference.

Not that shifting the balance of opinion will necessarily be easy - for either side. Those who say Don't Know and who would thus seem most up for grabs are not particularly numerous - there are certainly fewer than there were before the referendum on the Alternative Vote for Commons elections.

But more importantly, perhaps, people's views about Scottish independence are at least in part rooted in their sense of identity.

The more Scottish and less British someone feels, the more likely they are to want Scotland to leave the UK and become an independent country. And most people are unlikely to change their sense of identity during the course of the next 12 months.

However, the debate is not just about identity. Indeed, if it were then given that most people's sense of Scottish identity is stronger than their feelings of British identity, the Yes side would be well ahead.

But while most people who support independence have a strong Scottish identity, not all of those who have a strong Scottish identity back independence. Indeed, on some readings up to half do not.

What holds them back from voting Yes are doubts about the practical consequences of independence, ?¨not least what it would mean for Scotland economically.

There is thus little doubt that the Yes side will be disappointed the news that,?¨according to TNS BMRB, around twice as many voters think that Scotland's economy will perform less well under independence than believe it will ?¨do better.

The Yes side will have to turn ?¨those figures around if they are ?¨to have much hope of winning ?¨the referendum.

We can certainly anticipate that during the next 12 months both sides will proffer many an alleged fact and figure about the economic costs and benefits of independence.

There will be some key target audiences for these messages. The Yes side will be wanting to overcome ?¨the feelings of uncertainty about ?¨the consequences of independence that seem in part to explain why women (35% Yes) so far remain far less willing to back independence than men (47%).

The No side may well try to reinforce an apparent reluctance among the over-60s to break with a Britain they have known all their lives by reinforcing the doubts they also appear to have about the future ?¨of pensions.

In any event, Scotland's future now rests on which side proves the more effective at getting their campaign messages through.

John Curtice is professor of politics, University of Strathclyde, and chief commentator at whatscotlandthinks.org, where a comprehensive collection of referendum poll data can be explored.