The argument over Scotland's future grows steadily more intense.

Irrespective of the daily dose of indignation, no sensible person would have it any other way. Opinions are changing, decisions are being made. The polls, those strange exercises in drive-by social science, continue to narrow.

Opinion surveys are odd affairs, nevertheless. Most of us see headline numbers and fail to pay much attention to what lies beneath. Psephologists can entertain you for hours with explanations of weighting, random variation, sample sizes, or the wording of questions. Finding out what, if anything, people think about a topic isn't easy. The rest of us, journalists not least, just want to know who's up and who's down.

The argument isn't trivial. In the argument over a country's future, it matters. Politicians, wedded to their lore, like to think they detect signs of intangible qualities such as "momentum" in headline numbers. The political types who fancy themselves as astute (or cynical) will even shape policies according to "public opinion" as expressed by a thousand-odd carefully selected individuals.

This is fine as far (not very far) as it goes. But what happens when, as in the case of immigration, it becomes obvious that the public is deeply ignorant? How does it work when you offer "more powers" to a parliament, produce three contradictory versions of the offer, fail to explain why there are three, or "guarantee" that you will each support one party's proposal, when public knowledge of existing powers is rudimentary? Does this lead, in either example, to honest policy-making?

Caveats aside, the NatCen Social Research British Social Attitudes survey (BSA) is a marvel. It has a way of overturning received opinion. It also provides a record of the process by which opinion forms and fractures within a group of societies for which the political parties and their programmes are mere ciphers, at best.

Take the latest findings on the monarchy. You couldn't in honesty call them a surprise, but you could question the responses of the party machines. Thus, as we reported, "Only a third of Scots want a republic." That's a clear and obvious minority. But it is also a substantial minority that finds no representation whatever among mainstream parties chasing the will o' the wisp opinions of the majority.

That chase never ends. The parties conclude, rationally no doubt, there is no choice in the matter. It means, though, that often enough the effort to educate and inform is minimal. Are immigrants causing crime? There are 43 per cent who think so, despite the fact that, as yesterday's editorial observed, crime rates are falling. The tide of ignorance sweeps all before it while parties drift and grub for Ukip votes. It's how quotidian politics is done. Some take pride in the fact.

The Better Together parties congregating on Monday at the old, half-finished Edinburgh monument known as Scotland's Disgrace - some jokes are too easy - were examples of the phenomenon. All those last legislative words on devolution, all those lines in the sand, are long gone. The absolute refusal to offer a devo max question on the ballot paper has been consigned to the big file called "inconvenient history". The people want more powers? Since when did these noble parties fail the people?

As guarantees floated like chaff in the Edinburgh air, the question of why it took obduracy so long to melt into obliging eagerness didn't detain those who find No Thanks an inspirational slogan. Perhaps there's a referendum coming. But there was a deeper question for the half-a-loaf faction, one obvious to anyone who takes more than a first slice from the latest TNS Scotland survey: an awful lot of people preparing to vote on things as they might be don't know much about things as they are.

So 60 per cent hadn't heard that Unionism already means to provide "additional powers" through the Scotland Act 2012. Questioned, just 59 per cent of those who did claim to be aware of the legislation could identify at least one area of competence at stake. Of the minority who had a clue about the last attempt to expand devolution, only 24 per cent and 22 per cent, respectively, knew about the headline-grabbers, the power to set a Scottish income tax rate and to grant Holyrood a borrowing capacity.

On what basis, then, does anyone parade on Calton Hill with yet more new-minted promises of "powers"? How, if it crosses your mind, do you explain how pledges are to be reconciled should Scotland oblige by voting No? The independence claim, in contrast, is simple: all power to a sovereign parliament unless that parliament decides, case by case - but mind how you go - otherwise. When ignorance tops a poll, the promises of multiple complications result in a guddle. Some call that a strategy.

When a campaign grows fast and furious, context matters. Pollsters never fail to explain that their surveys are snapshots, fading from the moment they are taken, but activists are troubled only rarely by that detail. The difference between settled, underlying attitudes and the opinions of the moment doesn't often get a look in. When a polling company notes "margin of error" the first word gets all the attention.

Pitch the NatCen/BSA work into an independence campaign, therefore, and the result is a feeding frenzy at the pick 'n' mix counter. A majority of Scots (just outside the margin of error) would keep Trident boats on the Clyde while a bigger group (10 per cent above the England and Wales tally) would dispense with nuclear stupid-machines altogether? Which bit of survey work has real implications for SNP policy?

Ask instead whether people in England and Wales would object to sharing a currency with an independent Scotland. BSA says that all the Unionists who claim "English voters wouldn't wear it" are either wrong or deceitful. In this, there is a figure of 69 per cent running against Better Together and George Osborne, the Chancellor.

A representative of No Better Thanks (or whatever the latest formulation is) could well object. They could say - though oddly, they do not - that a historic survey of more than 3000 is no snap poll. The bulk of NatCen's work was completed last September, long before Mr Osborne intervened with his Sermon on the Pound. Perhaps the Chancellor moved hearts and minds.

Less ambitious polls says otherwise, with a certain clarity. Last September, surveys found the vote for independence running at between 27 per cent and 32 per cent; this month the raw Yes vote is between 36 per cent and 39 per cent. No has fallen from a high of 59/52 per cent last autumn to 46/43 per cent this summer. That's the proper context for any news of public attitudes.

Perhaps still more Better Together promises of still more powers are in order, then. A small margin of support for Trident has not coloured those recent snapshots, after all. Nor has gloomy Mr Osborne. Instead, the effort to inform seems to be paying more dividends, of a non-Union sort, than the effort to bamboozle in the name of public opinion.

One of the oldest jokes in politics is the one about the mushroom strategy. How do you treat voters? Keep them in the dark and pour manure over them. The jokers never talk about the crop that results.