Happy new year

Yesterday we had the final poll of the year – the fortnightly Opinium poll for the Observer, which showed topline figures of CON 29%(nc), LAB 39%(nc), LDEM 8%(nc), UKIP 15%(+1). I’m normally somewhat wary about polls conducted over holiday periods due to the potential of getting skewy samples. In this case the poll was conducted between the 21st and 27th of December, so right over the Christmas period itself, but the results don’t seem to be out of line, in fact they are almost unchanged from a fortnight ago.

With that out of the way I was intending to write a nice end of year round up, but time has gotten away from me, so instead I’ll put together some looks forward to the year ahead over the next week or two. Until then, have a happy new year.

The last monthly poll of the year that was still outstanding, ICM for the Guardian, turned up on Christmas Day of all times. Topline figures were CON 32%(nc), LAB 40%(nc), LDEM 13%(nc) – the figures are all typical of ICM’s polling of late (the comparatively high Liberal Democrat level of support is methodological, and normally due to the reallocation of a proportion of don’t knows to the party they voted for last time, which usually produces a higher Lib Dem score and a lower Labour lead).

Depending on what TNS BMRB and Opinium are doing with their regular polls over the Christmas break, this may well be our last poll of the year.


YouGov’s final Sunday Times poll of the year is online here. Topline figures are CON 33%, LAB 43%, LD 10%, UKIP 8%.

YouGov repeated their semi-regular trust tracker they have asked since 2003. While people in the survey said the Mitchell affair had made them trust the police less, the percentage saying they trusted local police officers was actually almost unchanged – 67% trust local police officers a great deal or a fair amount, typical of all the times YouGov have asked the question in the last few years. There was a slight drop in the proportion of people who trusted senior police officers with 47% of people saying they trusted them. While this is the lowest YouGov have recorded, it is not a significant change from the 49% who said they trusted senior police officers in November.

In contrast the Savile affair does seem to have significantly damaged trust in the BBC. Today’s figures actually show a slight increase in trust in the BBC since YouGov last asked the question at the very height of the Savile affair – now 51% of people say they trust the BBC, compared to 44% a month ago – however this is still a significant fall from the BBC’s 60% trust rating at the start of the year.

Looking at the more specific questions on plebgate, Andrew Mitchell is now narrowly more believed than the police – 31% think Mitchell is telling the truth, 28% the police, 41% don’t know. 43% of people still think that Mitchell probably called the police officer a pleb, but this is a drastic change from September when 69% of people thought he did.

Despite the turnaround in opinion people still think Mitchell was right to resign by 49% to 26%. 29% of people think he should be offered another government job, 40% think he should not. Even if Mitchell’s own version of events is proven to be true, many people think that should be a resigning matter anyway – asked about Mitchell’s admitted version of events, where he said to the officer “I thought you lot were supposed to f—— help us”, 38% of people still think this would be a resigning matter, 44% do not.

When a party goes up or down in the polls there is inevitably speculation as to why. Sadly it’s not normally very good speculation… there is always a temptation for people to follow the logic of I think issue X is very important therefore issue X is the cause of the recent shift in the polls. Inferences from polls are not always much better than that – people who are supporting party Y are more likely to think X, therefore X has caused the increased support for that party. It sounds okay, but what about issues A, B and C which weren’t asked in the poll?

Daily polling does at least give us an idea of when movements in public opinion have happened, and therefore make inferences about what events may have caused them. The graph below shows a five day rolling average of UKIP’s support in YouGov’s daily poll since the end of 2011.

You can see there are two big increases – the first was the Budget in 2012, nothing to do with immigration or Europe or any of those issues we associate with UKIP, the thing that co-incided with an increase in UKIP support more than anything else was the budget. My guess, given the demographic make up of UKIP’s vote, that the granny tax and the messages it sent out were the most important factor there. UKIP’s support then faded away a bit, had a couple of lumps and bumps during the autumn and then shot up again during November when there was an almost perfect storm for them – the run up to the EU budget summit, a decent performance in the police elections, the Rotherham fostering row, the speculation over a Con-UKIP pact and finally the solid by-election performance at the end of the month, all combining to produce far more news coverage than the party could normally dream of. It is possible that the gay marriage issue since then has helped keep their support up.

All of this is still a far cry from proving what causes the ups and downs in UKIP support, after all, correlation does not prove causality. There could have been other events at the same time that got less attention, but it is normally a fairly good pointer.

Note also the biggest drop in UKIP support, back at the end of 2011 at the time of David Cameron’s veto in Europe. As I wrote the other day, Europe isn’t actually the main driver of UKIP support, so if the Conservatives suddenly became more anti-European UKIP would not vanish like magic… but it is an issue that plays to the sort of values that drive UKIP voters, so neither is it irrelevant.

This morning’s YouGov poll for the Sun has topline figures of CON 30%, LAB 43%, LDEM 11%, UKIP 10%. It isn’t the biggest Labour lead YouGov have shown this month, but (leaving aside a rather outlandish 29% back in April) it matches the lowest Conservative score this Parliament. UKIP, meanwhile, remain in double figures on 10%.

Yesterday afternoon there was also the weekly poll by TNS BMRB, which has topline figures of CON 30% (+4), LAB 40% (-1), LD 7% (-1), UKIP 12% (-4), OTHER 10% (+1). The increase for the Conservatives and the drop in UKIP support is probably no more than a reversion to the mean after an unusually high score for UKIP in their poll last week.