“If I were a rich man, daidle deedle daidle…” Even as someone who runs a website on opinion polls, I doubt if I had £750,000 burning a hole in my pocket I would blow the lot on commissioning opinion polls. Still, much to my delight Lord Ashcroft has done just that and the mass of polling data he commissioned from YouGov and Populus is now up for our perusal on Populus’s website. If you want a full copy of Michael Ashcroft’s actual report into why the Conservatives didn’t win the 2005 General Election you’ll have to buy a copy of the report from Politicos.
I’m not going to attempt a precis of Ashcroft’s work, it’s an interesting study, albeit low on any actual conclusions aside from an attack on the Conservative strategy of targeting 164 seats which, while probably entirely correct, doesn’t actually stem from the polling data the study is based on. If you really want to read it you’ll have to cough up ten quid. That said, you can’t commission 12 large tranches of opinion polling without finding something interesting out. I’ve had a peer through the tables, though the sheer volume of them means there is probably much more to look at, but there a few points that leap out.
Firstly we might as well look at issue the polls were intended to address – the Conservative election campaign. As part of his research Ashcroft commissioned a daily tracker poll from Populus, one of far greater duration than that in the Times. As most readers will know, Populus’s tracker poll for the Times ended in tears, when a rogue sample over the bank holiday weekend threw their results out in the final few days of the campaign. While that illustrates one of the potential pitfalls of tracker polls, it doesn’t mean they aren’t a useful tool, and there are some interesting findings in Ashcroft’s tracker
The most interesting part of the tracker was an open ended question, asked every day, asking people if they had noticed anything the Conservative party had done or said in the last few days. By using an unprompted open ended question Populus were able to track the Conservative campaigns that people actually noticed, rather than those which sailed straight over the public’s heads. This is particularly telling for the Conservatives – in the run up to the election all the speeches and campaigns about more police, cleaner schools and so on failed to connect with the public. Only three policy announcements were really noticed by the public: the promise to reduce immigration, which over 30% of people told Populus that they had noticed when it was first launched, and which peaked again after further announcements about testing immigrants for HIV and so on; tax cuts, which peaked after the launch of the James Report and the announcement on council tax; and the announcement of a crackdown on illegal Gypsy sites. During the campaign itself people’s recall of the Conservative message was totally dominated by the issue of immigration, which was recalled by up to 17% of people. The next most prominent policy was cleaner hospitals, recalled by a relatively paltry 5% of people. Despite the fact that other issues were being spoken about by Michael Howard, either the media or the public had latched on to immigration as almost the sole theme of the campaign, the result of which seemed to be to make the Conservative campaign seem nasty and negative.
After the election YouGov asked people which party they had thought had the better campaign – most people thought Labour and the Lib Dems had both had good campaigns. Most people thought the Conservatives had a poor or very bad campaign. When asked to choose words to describe the campaigns, people thought the Lib Dems were “positive” and “moderate”; they thought Labour’s campaign was “confused” and “depressing” – but equal numbers of people thought it was “positive” and “negative”; the Conservative campaign was seen as “negative”, “aggressive” and “confused”. When asked about the leaders’ own campaigns, Charlie Kennedy was seen as “friendly” and “moderate”; Tony Blair was seen as “dishonest” and “nervous”; Michael Howard earned the unenviable descriptions of “aggressive”, “creepy” and “petty”.
Does this matter though, after all, if a campaign works who cares if people think its negative? The polls suggest that the campaign did indeed have a negative effect. In the Populus follow-up poll 44% of those who didn’t vote Conservative agreed with the statement that “Some people who didn’t vote Conservative have said that they were seriously considering doing so but were put off by how the Conservatives came across during the election campaign. Is this true of you or not?” I’m dubious about those figures – it stretches credulity to think that two-thirds of the electorate seriously considered voting Conservative at some point suring the campaign much though. Leaving the precise numbers aside though, there’s a clear message that the perception of running a negative and aggressive campaign hurt the Conservative party’s chances. Now, it’s possible that this was cancelled out by the damage it did to Labour, though it seems unlikely, Blair was, after all, already seen as dishonest without the Conservative party having to say it.
Looking towards the future, a Populus poll in January, covering 10,000 people, asked those people who were not currently considering voting Conservative whether they could consider voting Conservative at some point in the future. 48% said they could consider it at some point, 52% they they could never vote Tory (the same question was asked immediately after the election in May, when 49% of those who didn’t vote Tory said they might consider it in future). Populus then asked this group of people – people who weren’t voting Tory, but could possibily consider it in the future – about a number of criticisms that had been made about the Conservative party. The three that people thought were most true were that the party was opportunist, the party was too male dominated and the party had weak leadership. The statement that people thought was least true was that the party was in some way bigoted or narrow-minded. This statement was also rejected by younger people and AB social groups, the groups most likely to hold liberals opinions on such issues, suggesting that this really isn’t the Conservative party’s problem.
Populus then asked the same people about a series of possible reforms that might make them more likely to vote Tory. Easily the most popular would be for the Conservatives to concentrate more on health and public services, followed by putting forward policies to reduce government interference in people’s lives, and making the party younger and more female. The least popular reform would be to take a more sceptical line on Europe – although even that would make 56% of target voters more likely to vote Tory.
The first of these two sets of statements was put to people again immediately after the election. There was little change. People still thought the Conservatives were opportunitistic and they still thought they had weak leadership. Strangely people no longer thought they were too male – there hadn’t been any strong female leads during the campaign, in fact Michael Howard has often seemed to be a lone figure, so it is hard to explain why. Justine Greening had a lot of coverage the day after the election, but still!
Moving on there were some interesting bits of data about tactical voting, particularly in the first Populus poll, back in November 2004, which was conducted in 160 Conservative target seats and included some questions on tactical voting behaviour. In Labour held Conservative targets, Populus asked people who said they would vote Lib Dem or “Other” how they vote if it seemed their first choice didn’t have a chance of winning locally. 52% said they would still vote for their first choice. The remainer split almost exactly down the middle between people who would vote tactically against Labour, and people who would vote tactically against the Conservatives. Later in the campaign Populus asked Lib Dem supporters in 12 Lab/Con marginal seats how would would vote, bearing in mind that their seat was likely to be a close race between Labour and the Conservatives; in nearly every seat Lib Dem supporters were as likely to vote tactically against Labour as they were against the Conservatives.
Such decisions do, of course, depend upon the electorate being able to accurately gauge whether not their seat is a marginal, and which parties are actually in the running. Back in Populus’s November poll, when asked if people thought they were in a marginal or a safe seat, a quarter of people had no idea. The rest were split pretty evenly between thinking their seat was safe or marginal. Bearing in mind that this survey was only carried out in marginal seats, this means only around a third of people were able to accurately identify their own seat as a marginal. As political anoraks, it is very easy to forget that most people don’t give a monkey’s about such things as how marginal their local seat is.
People were similarly unaware of which parties were actually in the running in their seats – a YouGov poll at around the same time and in the same seats asked people who they thought would win the local seat – 9% of people in Lab/Con marginals thought the Lib Dems would win, 11% in Con/LD marginals thought Labour would win. This didn’t get much better as the campaign progressed – by April 7% of people in Lab/Con marginals still thought the lib Dems were most likely to win their seat.
Even if they did, it doesn’t appear to have quite the effect as you might think. 28% of people said living in a Conservative/Labour marginal makes them more likely to vote, but this included 24% of Lib Dem voters. Only 3% of Lib Dem voters and 5% of “Other” voters said living in a Con/Lab marginal made them less likely to vote – surprisingly this was pretty much in line with Conservative and Lib Dem voters. 22% of people said they would base their voting intention on whether they wanted Labour or the Conservatives to win their local seat. Here there was a substantial difference when it came to party alliegence. Liberal Democrats in Con/Lab seats were, unsurprisingly, most likely to say they were basing their vote on the national campaign, while Labour voters were most likely to say they were basing their decision on the local race, suggesting their were significant numbers of Labour voters who didn’t much care for the Labour government, but didn’t want to let the Tories in locally.
Moving on to the Conservative/Lib Dem marginals, marginality made much more difference to Lib Dem voters. 45% of Lib Dem voters said the fact they were in a Con/LD marginal made them more likely to vote, and 43% of them said their decision was based more on the local result than the big picture. In contrast to Lib Dem voters, when Labour voters in such seats were asked about tactical voting were still far more likely to vote Lib Dem to keep out the Tories than vice-versa.
When Populus phoned back some of those people it had surveyed after the election. 18% of people said that, in the event, they had voted tactically. Tactical voting continued to work against the Conservatives, but not overwhelmingly so – 49% voted tactically against the Conservatives, but 34% voted tacitically against Labour. If you look at the cross-tabulation with how people actually voted though, considerably more people told Populus they had voted Tory to stop Labour winning than voted Labour to stop the Conservatives winning (however some people were obviously confused by the question – 32 said they voted tactically against Labour, but voted Labour).
That said, a second follow-up poll by YouGov, this time only in Lab/Con marginals, found contrasting results; marginally more people voted Labour for tactical reasons than voted Conservative. Whichever is true, Labour certainly no longer seem to be gaining any significant advantage through tactical voting. (Again, the YouGov survey seems to show some strange results on tactical voting questions. In the 130 Labour/Con marginals – i.e. seats where the Lib Dems are third – 14% said they voted Lib Dem tactically because their first choice had no chance of winning).
Finally, while most polling tends to concentrate on the Conservative party’s target seats, a separate Populus poll covered the Conservative Lib Dem battle, polling 9 Conservative held Lib Dem target seats (Orpington, Folkestone & Hythe, Taunton, Westmoreland & Lonsdale, Maidenhead, Haltemprice & Howden, Surrey SW, West Dorset and Eastbourne).
One of the reasons behind Michael Howard’s strong performance in his own seat (and presumably the reason why all party leaders do well at elections) is immediately obvious: 80% of people correctly named Howard as their local MP. Recognition for the other MPs ranged between 31% for John Horam and 58% for David Davis. Being correctly identified by 56% of respondents in his constituency didn’t help Tim Collins, but then his opponent Tim Farron was the Lib Dem candidate with the second highest profile (the highest was Chris Maines, but then this was the fourth election in a row for him in Orpington. It was only Tim Farron’s second go.)
In Lib Dem target seats Labour voters were far more likely to vote tactically against the Conservatives than against the Liberals by a margin of about 4-1. The Lib Dem “decapitation strategy” seemed to have little effect – the proportion of people who said it made them more likely to vote Lib Dem was cancelled out by the people who said it made them more likely to vote Conservative. In Folkestone & Hythe it was perhaps even counter-productive, more people rallied to Howard’s support than were encouraged to try and oust him.
Populus went on to ask about how people viewed their local MPs and, broadly speaking, people had very good opinions of them: even Labour and Lib Dem voters thought their local MPs would work hard for them if they went with a problem, were working on local issues and were closely involved with the constituency. This is perhaps an explanation for why the Lib Dem decapitation strategy largely didn’t work – the Liberal Democrats often campaign on the strength of their candidate being a hard working local campaigner, pushing forward the actual candidate. When asked if they might vote for the best candidate, regardless of party, about two-thirds of Lib Dem voters said yes, while only half of Conservative voters said yes. In short, Lib Dem voters are more likely to vote for the candidate, rather than the party, and far from being seen as remote figures, the shadow cabinet figures they targetted seemed to be largely well-respected in their constituencies. Sadly Populus didn’t break down these figures by seat, so we cannot tell if the two defeated Conservative MPs, Tim Collins and Adrian Flook, were particularly badly thought of by constituents.