A rumour seems to be doing the rounds that the most recent YouGov poll indicates that, under a Conservative government, 50% of people would vote for an independent Scotland, with only 41% in favour of retaining the United Kingdom.

And in yesterday’s Herald, Robbie Dinwoodie even claimed of the poll that “it is the first time Alex Salmond's strategy of using a Tory victory to fuel support for independence in a referendum has been tested in this way”.

But neither the 50% figure nor Dinwoodie’s comment are correct. There have been at least two other fairly recent polls asking such a “what if...?” type question.

The first was an August 2007 YouGov poll for the Sunday Times which found 23% support for independence and that a further 9% might “consider” supporting separation in the event of a Conservative win, although it should be noted that a great many of the potential switchers were drawn from the 39% of SNP voters who didn’t actually favour independence!

The second was a March 2008 MRUK poll, blogged about by Douglas Fraser here, which again found 23% support for independence. But with MRUK’s wording a much higher 18% suggested that a Tory win might encourage them to “consider” the possibility.

And so it is with this week’s YouGov poll. As a single-option survey, the apparent 34% backing for independence isn’t comparable with the previous two polls, but like the other two the question about a Conservative government was worded not in terms of certainties, but only about whether it might make people “more likely or less likely” to back independence. About 8% indicated each of “much more likely” and “slightly more likely”.

Hence the 50% total, which is incorrect as it assumes that all those who said they might be tempted to vote for separation would actually switch. It also ignores the “less likely” countertendency.

But in any case, that these “what if...?” questions are unreliable is clearly demonstrated by the fact that multi-option polls, in which support for independence is gauged against that of the other main constitutional options, show virtually static support, with averages of 22% both for the six such polls conducted in 2007 and the group of three undertaken so far this year.

Yet the August 2007 YouGov poll found that 10% of people said they would be more likely to support independence if Alex Salmond really impressed as First Minister, and the March 2008 MRUK survey found a 22% figure under similar circumstances.

With the latest YouGov poll finding 41% backing for Alex Salmond as the best candidate to be First Minister and none of the other party leaders – or, in Labour’s case, leadership candidates – scoring even beyond single figures, it seems clear that, in relative terms at least, Salmond has really impressed very many people. It is with some regret that I confess that despite my many and fundamental disagreements with the First Minister, even I would probably count myself in that 41%.

So why then, with Salmond’s personal rating so high, is support for independence not growing? In the single-option context, it even appears to be on the wane: from 41% in April to 34% now. Why, when a Conservative government now seems almost inevitable, is that not actually filtering through into people’s constitutional preferences?

The basic answer is to be found in this clever analysis on the UK Polling Report blog, which succeeds in utterly discrediting questions in the “would X make you more or less likely to vote Y” style.

The simple truth is that while many people of a generally leftist disposition might loathe the idea of a Conservative election win, the notion that something as ordinary and non-threatening as a moderate, centre-right party winning a term in office might cause vast numbers of people to want to dismantle their nation, is little less absurd than reactionary domestic pledges such as a threat of divorce if one’s husband or wife goes directly from their work to the pub, or a threat of spousal murder if one is obliged to spend an unreasonable length of time traipsing around shops!