Sunday, December 20, 2009

Wishful Thinking & "Reliable" Sources

From Chris Whiteside's blog

Earlier this evening the Labour PPC for Skipton and Ripon, Claire Hazelgrove, put this comment on Twitter:

"Have now heard rumour from a number of reliable sources of new MORI poll placing the Tories on 37, and Labour on 34. Let's wait and see!"

The actual results of the IPSOS MORI poll concerned, taken some time ago but only published this evening, came out a couple of hours later and were rather different from that rumour: they were

CONSERVATIVE 43% (Up from 37% in the last MORI poll)
LABOUR 26% (Down from 31%)
LIB/DEM 20% (Up from 17%)
OTHERS 11% (Down from 16%)

Which only goes to show that if there is one thing sillier than reading too much into one opinion poll it is paying attention to rumours about an opinion poll.

I'm not at all excited by this poll. How can you take seriously a polling organisation which shows such huge monthly fluctuations? The last three months have shown the Tory lead go from 17% down to 6% and back up to 17%. Ridiculous.


Tapestry said...

quite. Bob Worcester, writing today on claims that the PBR was enough to bounce this months result back up, but faĆ­led to explain how MORI managed to grant Labour a large majority in London in their 6% effort in November, based on the views of only 50 people.

The extraordinary thing is that Labour are still propagating the Hung Parliament narrative. See Labour Do Hung Parliament Narrative To Death.

Hythlodaeus said...

Given the variance in a number of polls (YouGov, Mori, and Populous to name but three), I suspect what we are seeing is the public in full dithering mode. On one hand, Brown isn't exactly at his peak just now, but on the other Cameron is coming across as nearly schizophrenic in his policy aims, saying different things to different groups.

There is, however, far too much importance being given to individual polls, when people should be seeking an average across polls conducted over a period of time to help counter error.

Personally, I just wish Gallup still did election polling in the UK.

The King of Wrong said...

@tapestry: Labour need to talk up a hung parliament, as it's their only hope.

Basically, they have to stop people voting as people are not going to vote for them - so a low turn-out is good. As is disenfranchising as many non-Labour supporters as possible - it's obvious that postal vote fraud, missing ballot registrations and general bureaucratic incompetence can have a significant effect on an election, so that's something to watch carefully.

But how best to encourage a low turn out? Talk of a Labour victory would be implausible, and would drive Tory activism, tactical voting for the Conservatives in marginal seats and complacency from Labour supporters; talk of a certain Tory victory would lead to complacency and "well it doesn't matter if I don't vote" attitudes from the right, but would also discourage Labour donors - which they can't afford to lose; talk of a hung parliament means that everyone will lose interest and it emboldens the Lib Dems and other minority parties to campaign and split the non-Labour ballot in the feeble hope that they'll get to have Influence in a coalition government.

A corollary to that is that any story likely to cause disaffection with mainstream politics will be pushed: cross-party expenses scandals, political stalemates, the failure of COP15, EU bureaucracy, immigration, etc. etc.

Anonymous said...

Iain, I'm pretty sure the 6% one was the real rogue one amongst the pigeons, it seems the methodology was wrong on that one, was it intentional to create a shock poll to sell it, i wonder?

Although i'm almost sure the real lead is somewhere around 15 putting us in landslide territory, i can't see anyone wanting to vote labour.

Moriarty said...

The point is not whether you get excited or not. What's more interesting is whether the commentariat who made so much of the "narrowing" of the Tory lead will adjust their narrative in light of this latest poll. My guess is not. To be fair to Adam Boulton (not something I enjoy) he's at least maintained a consistent scpeticism re the "Tory lead narrowing" nonsense.

Unknown said...

I'm no statto but there does seem to be an explanation for the polls.
The Tory vote - apart from the MORI 6% one - seems fairly steady at 40 or above. It is the Labour vote that is variable.
That could be because of methodological reasons, or it could be because the Labour core vote of around 30% is getting shaky.
It would also explain Brown's adoption of a core vote strategy.
If that is true the 'hung parliament' narrative could actually help the Conservatives.
Apart from the 25% of the electorate who would vote for a monkey if it was wearing a red rosette, Brown is cordially loathed.
The possibility of a hung parliament and with it the possibility of Brown hanging on might be enough to drive undecideds and supporters of other parties to the Conservatives.
If I was with the Conservative high command I would keep banging on about the electoral mountain they have to climb to win. It might shove a few undecided Brown haters into their camp.

Anonymous said...

Nobody ever reads the small print on these polls. The margin for error is typically +-3% , and that's +-3% on each party's vote.

For example, if the true Tory vote is 40%, and the true Labour vote 30%, the poll will show anywhere between Tory 37% / Labour 33% (a 6%Tory lead), and Tory 43% / Labour 27% (a 16% Tory lead).

What's more, the confidence level is only 95% - in other words, 1 in every 20 polls will produce an even bigger error.