Given that everything depends on the turnout model, the polls are also useless when it comes to giving us a useful prediction for a UK general election. So we are looking at the next election as a real unknown. That's useful in one regard in that it should stop anyone getting a huge shock on the day, but only because no one really has expectations to be dashed!
However fortunately amongst all the polling we've had a crop of independence referendum polls which do tell us something quite interesting. Firstly the turnout model for the independence referendum should not suffer from the issues we've seen in the General Election. That's due to the high level of turnout that we experienced during the referendum and can expect again. Also, as I've shown before, polls on independence tend to favour the Yes movement and overestimate support for the SNP.
However I'm beginning to wonder if that effect is waning, and that might be what we are seeing in the polls now. We'll know soon enough when we compare the polling for the SNP versus the actual votes they achieve on Thursday.
An SNP win whilst losing the battle?
I expect a resounding "win" for the SNP on Thursday as the split unionist vote to enables them to save quite a few high profile SNP names (such as Angus Robertson) I still think they will lose some ground in votes and seats. The big issue for the SNP however is even with a General Election "win" in Scotland (it will still be a loss at UK level because that is mathematically inevitable for the SNP) their flagship policy continues to decay in support.
I usually leave this chart to the end but I think it's probably the best place to start. When you just look at the Yes vote alone, i.e. exclude the No votes and Don't Knows, something remarkable has been happening of late.
In the past I liked this chart as it helped to give us a better understanding of the true Yes vote. That was based on the evidence that "shy no voters" were "dont know" in the pre independence referendum polls.
The trouble with the chart however is different polling methods would render different Dont Know levels, PanelBase for instance would tend to have few don't knows. Therefore the only way to look at this chart was through a poly line of best fit, this would help to smooth out the different methods and let us see the underlying strength of the Yes vote.
You don't need an explanation of what's going on to see that the Yes vote is in retreat. Lots of flag waving Yes supporters marching through Glasgow continues to demonstrate the lessons the Yes side (and perhaps UK Labour's big crowds) should have learned from the referendum. The enthusiasm of the minority does not make a majority.
This point is made obvious in the standard charts with the linear line of best fit (splitting don't knows equally amongst the Yes and No votes) showing that No continues to rise with time.
The 5 factor poly trend also shows and even more dramatic trend of late with the case for independence in serious decline. Some of this may be due to No voters becoming more confident in their position (that would certainly line up with the rise in the Tory vote and the reported strong support for Labour amongst young voters in recent polls).
So it may be a reduction of the shy no effect (we should be able to establish that on Friday) but is almost certainly also a waning of the case for independence. No wonder that it's number 10 on the SNP manifesto.
ELECTION DAY UPDATE
-------------------------------------------------------------------------------------------
We've had a new Survation phone poll on election day so I thought it would be worth updating the trends. The poll is potentially an outlier or is just a further crystallisation of the effects I was discussing above.
We'll be able to tell much more when the election results come in, but right now this looks really bad news for the SNP.
No comments:
Post a Comment