Push Me, Pull Me (April 14 2010)
Election polling has evolved since the 1970s and pushes parties towards the centre ground
Election campaigns have always been changing and evolving. The idea of an election campaign as a coherent story, unfolding over time, does not really apply to some contests such as 1950 and 1955, when there was an almost aimless meandering across the policy landscape punctuated by the big speeches by the party leaders and other main figures. The election campaign took more shape in 1959 – essentially the first television election – and through the 1960s and 1970s. There have been polls since 1945, but the close election of 1964 was arguably the first in which the ‘horse race’ aspect of an election started to dominate in the media with polls providing the evidence for the state of the race every few days. In 1970, the Conservative campaign deployed the first effective use of modern negative campaigning. Nine years later, Saatchi & Saatchi famously revolutionised election advertising. But in between the fall of Edward Heath and the rise of Margaret Thatcher, there was another important development.
Perhaps the first poll-driven election campaign was Harold Wilson’s Labour campaign of February 1974, which used MORI to pilot the party to a narrow and somewhat improbable victory. Wilson was brilliantly reactive, responding to campaign events such as bad economic figures and government missteps and changing the terms of debate from the Tories’ ground of ‘Who Governs?’ to Labour’s – the government’s record on prices and industrial relations. Wilson astutely dodged around the contents of Labour’s manifesto, which was a left-wing programme of nationalisation he had no intention of implementing. The purpose of the polls in this election was not to look at headline figures, but at the underlying attitudes and opinions on the issues that were driving political choice. The polls have been used to craft the parties’ narratives during the election ever since. Although focus groups were used in the 1970s, their real heyday has been since 1997 when they have tested party messages and the electorate’s perceptions exhaustively. The 1990s also saw the rise of the ‘grid’, in which party messages, issues and leading figures are deployed in a rigorously planned fashion.
On one level, a modern election campaign is a fierce contest for control of the narrative. Each party is attempting to tell their story about the state of Britain and what needs to be done, and calculates that if they are able to do so unobstructed then their narrative will convince people to vote for them. However, it is never possible to get the message across unobstructed. The parties compete with each other, with the media (now in all its electronic wildness) and with events (both ‘known unknowns’ like announcements of economic statistics, and ‘unknown unknowns’ that happen unpredictably) for control over the agenda. Crudely, a party ‘wins’ a day in the campaign if the real events that are prominent in the media conform to what was planned in its grid. As well as all this chaotic competition, there is also feedback – that the profusion of public and private polls is giving everyone nearly instant knowledge of what is going well or badly, and who is ahead and behind. Polls can change a campaign instantly, with a good poll creating confidence or complacency, and bad ones causing lurches into despair. It is not without reason that they are sometimes compared to mood-altering drugs.
The 2010 campaign has more published polls than ever before, with the almost-daily YouGov series for the Sun being a notable innovation. To continue the drug analogy, the political classes are getting habituated through heavy use, and this is perhaps not a bad thing. Every now and then there is a ‘rogue’ poll outside the normal range of sampling error, and these tend to attract the most interest in the media for the standard ‘man bites dog’ reason that the unusual is news. In past elections, rogue polls have sometimes had major influence, famously so in 1987 when one precipitated ‘Wobbly Thursday’ in the Tory campaign. But thanks to the sheer weight of polling, rogues are now likely to be swamped by polls that are closer to the mark. It will take a proper trend, not an outlier, for voting intention polls to change the tone of an election.
It is at a deep level that the parties’ strategies are influenced by the findings from their private polling about what the public wants to hear. One cannot blame the parties for using the best available techniques for crafting and putting out their messages. Nor can one blame the electors for thinking that the parties sound the same, because they are all talking at the same swing voters in marginal seats in the same sort of language. It is a consequence of the electoral system. And a question, perhaps, for another day.