Why are we here? I mean: why are we here in a specific Scottish sense? Why is support for the SNP holding up so well? Why has support for independence risen, then fallen again? Some of it, I suspect, will never be satisfactorily explained and will stay in the privacy of the voting booths forever, but quite a lot of it, I think, is down to one particularly curious, interesting, and rather shy group of voters. You could call them Scotland’s Ten Percenters.

The Ten Percenters, for want of a better term, are essentially that middle slice of voters – usually around 10 per cent, sometimes a bit more, sometimes a bit less – who tell pollsters they are undecided or don’t-know. You can see them in every opinion poll, although some of them will drift into other camps now and again. By looking at how these undecideds have behaved in the past, we can take a stab at guessing how they might behave in the future; we can also divide them into little sub-groups, and the results for the SNP are a little bit good, and a little bit bad.

In recent weeks – and who can blame them? – some of the Ten Percenters have obviously been wobbling on what they might do in future votes. Most of the recent polls on independence, for example, have been showing a rise in support for Yes but the most recent one shows it dropping back down to 47%. It would seem like there’s a hardcore of 40-45% that back Yes, it’s the same for No, and there’s a portion of the middle group, the Ten Percenters, who may have been leaning towards Yes in recent weeks but have now moved back to Undecided.

Read more: SNP: The unholy alliance between nationalists and liberals 

In some ways, this is understandable – voters may look at the current political situation and think "if it’s this bad now, what on earth would it be like under independence?" But the curious thing is that all the stuff that’s been going on recently – the divisions in the SNP, the Salmond inquiry debacle, the inefficient rollout of the vaccine in Scotland – would appear to have affected support for independence, but not support for the SNP at May’s elections.

This will confuse some people (it certainly confuses me): why would you think independence is a bad idea but still vote for the party that’s pushing hard for independence? I’m afraid I do not have the answer to that, except that people may judge the Tories and Labour as even worse than the SNP. It does, however, leave us with the big puzzle: the scandals and disarray in the SNP are affecting support for independence but not support for the SNP (even though the scandals and disarray tell us much more about the SNP than they do about independence). This is the bit of the polling that’s promising for the nationalists; it’s the part that’s a little bit good for them.

As for the bit that’s a little bit bad (and the SNP’s own pollsters know this), it relates to who exactly the Ten Percenters are. Some of them will be people who are properly wrestling with the complex questions independence raises (and in an age where so many of us seem to be 100% certain of our opinions, I have a lot of respect for this group). We should also assume that some of these voters, when their wrestling days are over, will decide to vote for Yes.

But let’s look at the remainder of the Ten Percenters. For a start, a good chunk of them aren’t undecideds at all; they’re people who have an opinion but for various reasons prefer to tell pollsters they don’t. This may be because their opinion is unpopular or socially controversial – we saw it in the 1990s, when the polls showed Neil Kinnock ahead of John Major when the truth was that “shy Tories” had told the pollsters they were undecided, and we saw it again when the polls showed Hillary Clinton ahead of Donald Trump. The lesson is that a large chunk of undecideds have actually made up their mind, they’re just not prepared to say so.

The same situation will almost certainly apply to the “undecideds” on independence. Some will be genuinely swithering, and may move from camp to camp (hence the polls), but a lot of them aren’t Don’t-knows at all, they’re Nos but don’t want to say so. We also know from previous referendums that most don’t-knows end up backing the status quo on voting day, and staying in the UK would involve less change and would be the status quo. What this means, once again, is that most of the don’t-knows are actually No voters.

Read more: Tom Devine is right about the union flag, but for all the wrong reasons

To top it all, there’s another aspect of the Ten Per Cent phenomenon that’s bad for the SNP and it’s this: Ten Percenters are more likely to be women than men. There are probably good, complicated social explanations for this, but, as anyone who’s spent more than five minutes in a pub will know, men are more likely to have 100% certain, unshakeable opinions on politics, and it means that, when it comes to opinion polls, it’s women who are more likely to say “don’t-know”. This is not good for the SNP because we know from 2014 that women broke more heavily for No and, if the same applies next time, it means the female part of the Ten Percenters are more likely to go for No again.

All of these factors together make the “don’t-know” issue a big problem for the SNP, even though the forecasts for May remain good. What’s particularly clouding the issue is that the headlines on opinion polls tend to omit the Ten Percenters. Not only does this artificially inflate the Yes vote, it obscures the fact that it’s the “undecideds” – not the certain Nos or the certain Yessers – who will determine the result. And what do we know about the Ten Percenters? A lot of them are really No voters in disguise and the rest of them are likely to go for No in the end anyway. As I say, not good news for the SNP. But it might explain where we are. And where we’re going.

Our columns are a platform for writers to express their opinions. They do not necessarily represent the views of The Herald.