Saturday, April 30, 2005

Pollwatching

There's no poll evidence to confirm my view about the relative insignificance of this week's upheavals over Iraq, although I think we can expect that on Sunday and Monday.

The YouGov/Telegraph poll on Friday threw up some interesting results. Firstly, the Labour vote seems to be firming up at around 36% and we've seen the expected Liberal Democrat campaign bounce push them up to 24% from 21% at the start of April. The real bad news is for the Tories, who have seen their vote share slide from 35% to 32% over the same period, suggesting a Tory/LibDem swing. This doesn't surprise me that much, as I suspect that Labour/LibDem swing voters will have shifted over the Iraq issue and are already on-board with the Liberal Democrat vote. Overall, the vote is settling, as the undecided voters make their minds up - 27% down to 23% since the start of April. Of course, the fieldwork for this poll was done over the period of the Iraq revelations, so we won't really see any significant impact yet. The 'real alternative' approach for the LibDems is also having some impact, as the wavering voters are generally shifting towards them, but if Labour can keep pushing the line that the fight is between them and the Tories, half of the the undecideds would plump for Labour, against 36% for the Conservatives. Despite all of the grief, Blair has also firmed up his lead as the best PM, up slightly to 36% - 12 points clear of Howard and 18 points ahead of Kennedy. ICM ran a poll for the Guardian this week, which produced some different results. Here, the Labour vote topped 40%, 7 points clear of the Tories, with the Liberal Democrats down on 21%. While people liked and trusted Charles Kennedy, they didn't find him charismatic and 44% still viewed Tony as the best PM on offer - double the number who backed Howard or Kennedy.

On the issues, both the YouGov/Telegraph poll and one that YouGov did for Sky comment on issues indicate the unimportance of Iraq in this election - 11% of the electorate consider it one of the three deciding factors. If anyone doubts the effectiveness of a campaign based on prejudice, the Tories have shifted immigration up into the second spot, with 42% of the electorate putting it into their top three. You can see why, as the Tories have a commanding lead on this issue over Labour of 18 points. The traditional Tory issue of crime has also climbed up and they have held a 7 point lead over Labour. Unfortunately for them, the other major issues - health, education and the economy all hold double figure Labour leads.

ICM questions on the issues only ask about which will be the most important, rather than looking at a basket of key issues. This is handy for assessing top priorities, but I suspect that most undecided voters will choose on the basis of a small range of issues. However, Iraq is only a key issue to around 3% and immigration to 8%. Health, the economy, law and order, education and tax are all strengths for the Labour party.

A Labour victory seems assured (despite dire threats about a LibDem protest vote benefitting the Tories, comprehensively exploded by the Independent today) and this is backed by the British Electoral Survey results from Essex University, which shows a widening Labour lead over the Tories and a Tory/LibDem swing. To confirm this theory - the bookmakers Paddy Power are already paying out on a Labour win.

I've said for a while that this election will be decided on turnout and there's an interesting article by Anthony Wells that discusses how different pollsters weight answers on certainty to vote. MORI's method of only accounting for those certain to vote seems to have an inbuilt Tory bias - they tend to be the most reliable voters. ICM weight each respondent by their likelihood to vote on a scale of 1-10, so allowing more for individual habits. Along with their accuracy in 1997 and 2001, this is one of the reasons I trust the ICM polls more than others.

That said, I think that there will be a number of unpredictable results - Labour will hold a number of awkward marginals (including Yardley, I suspect) while the LibDems will make some surprising gains against the Tories. I'm standing by an estimate of a majority of 70, but there's still a lot to play for. Now's the time to start looking at the weather forecast as well - turnout depends on good weather!

No comments: