This morning’s YouGov poll for the Sun has topline figures of CON 34%, LAB 44%, LDEM 9%, UKIP 7%. The ten point Labour lead is right back to normal after what looks like an obvious outlier yesterday.

Meanwhile the weekly poll from TNS BMRB also has a ten point lead and shows no significant move from a week ago. Topline voting intentions are CON 31% (nc), LAB 41% (-1), LDEM 9% (nc), UKIP 9% (+1).

Today’s YouGov poll for the Sun has topline figures of CON 35%, LAB 39%, LDEM 10%, UKIP 8%. The four point lead looks very much like an outlier, so I would treat it will some scepticism unless later polls this week show a similar pattern.

YouGov also repeated their semi-regular tracker about trust in various professions following the BBC’s recent troubles. The proportion of people saying they trusted BBC News journalists to tell the trust was down from 57% last month to 44% now, and for the first time marginally more people said they didn’t trust BBC journalists than said they did.

To put this in context, BBC News journalists are still more trusted than journalists on other channels or newspapers, but there has been a sharp decline in recent years. The impact of the Newsnight affair is just a further blow to an already declining reputation. When YouGov first asked the question back in 2003, prior to Andrew Gilligan and the Hutton Report, 81% of people said they trusted BBC News journalists to tell the truth. The drop from 81% to 44% is close to a halving of public trust in BBC journalism over the last decade.

The long term trends show a number of interesting patterns over that decade. There has been a decline in trust towards most groups, the exception being those that were not particularly trusted to start with. So doctors, teachers, local police officers and judges remain the most trusted professions, and the only ones trusted by more than half the population. Police chiefs are trusted by 49% of people, down from 72% back in 2003. The recent revelations over Hillsborough do not, incidentally, appear to have do any particular damage, the drop in trust towards police chiefs came back between 2003 and 2006.

All journalists have seen a drop in their trust ratings, though this has effected tabloid journalists the least (because very few people trusted them to begin with). Trust in mid-market newspapers like the Mail and the Express has halved over the decade, from 36% in 3003 to 18% now. Trust in the broadsheets has fallen from 65% to 38%. Surprisingly there appears to be very little lasting effect from the phone hacking affair. If you look at the figures from July 2011 – conducted when the phone hacking scandal was at its height – there is an obvious drop in trust towards newspaper journalists, trust in broadsheets fell by 6 points, in mid-market papers by 5 points, in tabloids by 4 points. However, if you look at the figures from January 2012 and since then trust in the newspapers appears to have recovered to the sort of figures there were showing prior to phone hacking.


Nadine in the jungle

Having written last week about the way that political trivia gets far more political coverage than hard politics it feels almost remiss of me not to have mentioned this YouGov polling from the middle of last week (indeed, if it wasn’t for the BBC management’s rapid implosion it would probably be the big political news still – sigh).

Anyway, for the record 77% of people thought it was wrong for Nadine Dorries to appear on “I’m a Celebrity” compared to 11% who thought it was the right decision. 73% thought the Conservatives were right to suspend her from the party until she explains her decision.

(For those appalled by such triviality, fear not, I am sure there will be polling on the BBC along before too long, and on Thursday we have some proper elections, albeit, not ones we’ve had much real polling on. More on those in the coming days)

This week’s YouGov results for the Sunday Times are online here. Topline voting intention is CON 32%, LAB 44%, LDEM 8%, UKIP 8% (so towards the higher end of the normal variation around a ten point lead). Approval ratings for the party leaders are minus 16 for David Cameron, minus 18 for Ed Miliband and minus 55 for Nick Clegg.

On the regular economic trackers the increase in the “feel good factor” (the proportion of people who think they will be better off in the next 12 months minus those who think they will be worse off) that we saw after the good GDP figures three weeks ago has now unwound and we are back to the levels of pessimism we saw pre-October. 10% expect their position to get better in the next 12 months, 50% expect it to get worse – we will obviously require more sustained good economic news in order to see a real turnaround in economic optimism.

On law and order, the Conservatives retain a small lead as the most trusted party, with 24% to Labour’s 19%. 41% of people think policing has got worse in the last couple of years, asked about crime levels 20% think they have gone up in their local area, 12% gone down and 53% stayed the same.

Looking specifically at the Police and Crime Commissioner questions, 28% of people in areas with elections say they are 10/10 certain to vote in this week’s election. Turnout is notoriously difficult to predict in opinion polls – people invariably overestimate their own likelihood to vote – but at general elections the proportion of people who say they are 10/10 certain to vote has not been a bad guide to actual turnout. 28% however still sounds quite high considering some of the predictions we’ve seen, just lower than the sort of level more energetically contested local elections usually get.

20% of people support the introduction of elected commissioners, 34% do not, 46% say they don’t know, underlining the lack of interest in or awareness of the policy. People think it will make the police more accountable by 24% to 8%, but very few seem to think it will make any difference to standards of policing or levels of crime.

36% of people see Rowan Williams has having done a good job as Archbishop of Canterbury, 25% a bad job (39% don’t know, perhaps a sign of the role’s diminishing prominence). Opinions from people who identified themselves as belonging to the Church of England were a little more positive, 49% though Rowan Williams had done a good job, 25% a bad job. It’s important to note that the Anglican figures are for people who self-identified as being Church of England – many of them will be Christian in a purely notional, cultural sense. For example, 43% of people who said they were Church of England actually attend a church only once a year or less and only 49% say they believe in God. Practicing Anglicans who regularly attend church may have different views, but there are not enough of them in a national sample to get representative figures.

One of the most important things in understanding public opinion on politics is quite how little attention most people pay to it. I constantly see comments in here asking what effect a news event will have on voting intention. Most of the time the answer is none. This is should be evident from looking at the polls, which have had a steady Labour lead of 10 points or so for months, despite lots of “things” happening. The reasons are that firstly people aren’t watching anyway – most people don’t read broadsheet newspapers or pay much attention to the news, secondly, those that do normally interpret events and stories through their pre-existing political preferences, so they are more likely to re-inforce their existing views than change them.

If you are reading this website in the first place, you are probably a bit of a political anorak. At the very least you are interested in politics. Most people are not, and no matter how little attention you think people pay to political events, you are probably still *vastly* overestimating it. To illustrate it, think of something you care absolutely nothing for – celebrity magazines perhaps, soap operas you don’t watch, US baseball, whatever doesn’t float your boat. Do you know what the big story was in that field last week, how has it changed your opinion of the big players in that field, do you even know who the big players are? That’s most normal people’s attitude to what happened at PMQs this week.

The lack of public interest and awareness of what is going on in politics first really struck me looking at some Populus polling Lord Ashcroft did for his book “Smell the Coffee” back in 2005. It is easy to ask people if they have heard about a story, but it is not the best way – they may not want to look stupid, they may misremember having heard about the story*, the act of prompting people about the story may make people remember it when they’ve forgotten it and so on.

What Lord Ashcroft and Populus did was each day, between January 2005 and election day, ask 250 people if they recalled anything the Conservative party had done or said that week. There was no prompting, it was just what people recalled. Most of them noticed noting at all – sometimes up to 90% of people had noticed nothing (recall, this was just before an election was interest was at its height). The biggest score of anything was Michael Howard pledging to cut immigration, which peaked with 30% of people noticing it straight after Howard’s announcement. The key Conservative election pledges on things like cleaner hospitals, cutting taxes, more police were normally recalled by well under 5% of people. Never forget how little of politics gets through.

Anyway, seven years later and we have another bit of polling from Lord Ashcroft on a similar vein. He’s asked people to list what political news stories they have heard over the last few weeks, again unprompted. The most recalled, by far, is Andrew Mitchell and plebgate, which was recalled by 33% of people, followed by George Osborne not paying for a first class ticket which 13% of people recalled and 8% who recalled stories about the Scottish independence referendum. 7% recalled cuts to child benefit, 6% recalled the story about MPs “swapping flats” to claim more expenses. 5% recalled the increase in GDP figures and the row about prisoners voting, everything else was below 5%.

The second half of Ashcroft’s poll gave people a prompted list of stories and asked if they had heard of them, and also how important they were. The proportions of people claiming to have heard of a story were higher, but “plebgate” still came top (many of the other differences were timing related – in the unprompted question people cleared tended to give stories from the previous couple of days, when the prompted question included things from weeks or months back). People did tend to rate the solid policy stories as more important than the “political soap opera” stories, but the “soap opera” stories were more widely recalled.

What it shows, especially the unprompted question, is that people are more likely to pay attention to and remember the rather trivial but human stories that they think are unimportant than stories about policies and proposals. People may say they think it is comparatively unimportant that Mitchell called a policeman a pleb… but a third of people recalled it unprompted. Try getting a third of people to recall a party’s tax or economic policy unprompted. One might be seen as petty and one might be seen as important, but if people are only aware of the petty one is it going to inform their view of the party.

That’s different, however, from saying they necessarily make an impact. As I said at the start of this post, the reason most events don’t have any impact on voting intention or other trackers isn’t just that they aren’t noticed, it is also that people view them through the prism of their existing political views. So if a Labour MP does something awful, Conservative supporters will probably think it is disgusting and corrupt and must taint the whole of the Labour party… but they weren’t supporting Labour anyway. Labour supporters will probably tend to take a more charitable view, it was an understandable mistake, just one rogue MP and there are bad apples in all parties, the leadership acted strongly to punish them, etc (and of course, it works the other way round if a Conservative MP does something awful).

Even if they do have an effect, it is probably so subtle it is impossible to measure. No one is, in three years time, going to think “Well, the Conservatives have done well in government, I think David Cameron is the better leader, but one of their MP was a bit rude to a policeman three years ago so I’m voting Labour”. However, they might well think “the Conservatives are out of touch with ordinary people and look down on those less wealthy, they aren’t the party for me, I’m voting Labour”. Some of that view could have been contributed to, or reinforced by, a Conservative MP allegedly calling a policeman a pleb. Could we ever prove or disprove this through an opinion poll, not really, no. If we had polls tracking whether people thought the Conservatives were in touch or not from before and after the event we could infer it – but it is tricky to isolate an event, and most changes cannot be distinguished from margin of error variation.

In short, as I’ve said here before, there are three ways of understanding public opinion and its impact on people’s views. The first is crude support or opposition – do people approve or disapprove, like or dislike something. Very easy and straightforward to measure – they don’t like politicians swearing at policemen. The second is salience – is is important to them compared to other issues? Are they even away of it? This is trickier to measure, but in this case we know a significant proportion of people were aware of the Andrew Mitchell story, but also that most didn’t think it was that important. Thirdly, what impact does it have on their wider perception of the party – does it make them think the Conservatives are more out of touch, just reinforce existing views, or neither? We really can’t tell, and its not something that polling can easily tell us.

(*on people misremembering stories that they haven’t actually heard about, the Ashcroft poll today included two fake stories. 14% of people said they had heard at least something about Labour MP Audrey Cockburn using union funds to decorate her flat. Given neither she nor her flat exist, those 14% of people are wrong)