Government by opinion poll

Opinion polls have become a mainstay of representative democracy. One thing we can be sure of over the next few weeks is the daily updates on the waxing and waning of voters’ views. Political parties test their campaign broadcasts with private polls and focus groups, choosing those that best sway target sections of the public. Few organisations – from local councils to campaign groups, from newspapers to government departments – get by nowadays without the services of a pollster.

The rise and rise of opinion polls parallels a growing disengagement from public life. Today’s model citizen is less a member of a political party than a survey respondent. It is by answering yeses, noes and maybes in response to pollsters’ prompts that the public is represented in the political sphere. This is a passive electorate, which speaks only when spoken to and fits its opinions into pre-manufactured forms. While the political subject forces his or her views upon leaders, the survey respondent is an object to be prodded and tested.

Opinion polls moved to the centre of UK political life in the 1980s and 90s, a time when attachment to political parties weakened, membership fell, and the numbers of inactive members grew. Polls were brought in as a way of getting in touch with the growing ranks of floating voters.

Parties had used polls before, but they did so erratically and mainly for the purpose of presentation rather than policymaking. Polls were first introduced in 1937, after the success of the Gallup Poll in predicting the 1936 US election. By the 1960s the main parties had established relations with pollsters: the Opinion Research Centre was set up in 1965 at the behest of the Tory Party, and Research Services Ltd had links with the Labour Party (1). But at this stage policymaking was mainly driven by a combination of political necessity and pressure from party bases. It was only when the politics of left and right began to lose its fire, and party members disengaged, that polls were adopted wholesale.

According to Matt Henn, author of the book Opinion Polls and Volatile Electorates, the number of national campaign polls in general elections rose from one in 1945, to 25 in 1970, to 47 in 1997. The Labour Party’s spending on polls was £25,000 in 1979, rocketing to £100,000 in 1983 and £200,000 in 1992. The Tory Party’s grew correspondingly, reaching £60,000 in 1979, £96,000 in 1983, and £250,000 in 1992.

The index to watch in the Labour Party is the position of Philip Gould, pollster and key member of New Labour. In the 1983 election he was out in the cold, as Michael Foot fought on the ticket of nuclear disarmament, high taxation, and an expansion of public ownership. By 1987 Gould was playing a key part in the presentation of Neil Kinnock’s campaign. By 1992 he was leading on policymaking, telling Labour’s leaders that they should change their policy on tax because his polls found that this was the main reason people baulked at voting Labour. According to Gould, his Shadow Communications Agency exuded ‘an aura of quiet, mysterious power’, growing in strength as the parliamentary party floundered (2).

With the victory of New Labour in 1997, polls moved to the heart of government. During the summer of 1998 Labour commissioned MORI to put together a ‘People’s Panel’, a group of 5,000 people to be interviewed over the next four years about the progress of public service delivery. Although they were interviewed separately, they formed in effect a pseudo-plebiscite in which government tested the framing and results of its policies.

Dr Lawrence Pratchett, director of the Local Governance Research Unit at De Montfort University in Leicester, notes that opinion polling is now an everyday part of local authorities’ service delivery. ‘Since the 1999 Local Government Act there has been a large-scale move towards opinion polling. Before that there was some polling, but it was mainly through local initiatives. Today the majority of local authorities use citizens’ panels, say 1,000 people, to whom they send a questionnaire every few months.’ MORI claims that over the past decade it has carried out more than 1,000 projects for 250 local authorities. Meanwhile, a scan down ICM’s client list finds all the main government departments, including the Cabinet Office, Foreign Office, Home Office and the Department for Culture, Media and Sport.

Opinion polling was first used by corporations as a way to get inside the heads of isolated consumers, the better to push their buttons and get them to buy products. There is no debate between consumers and corporations about what would be best to produce. Instead corporations release a product on to the market and ask: how do you like it? In their endless consultations, politicians are effectively asking the same thing: do you like it when we do this; how about if we do that? Rather than, this is what we believe and we want to convince you it’s right.

The USA’s less established party system has meant that opinion polls moved faster and further into politics, with ‘approval ratings’ reported every few days on the main political figures. The website Pollingreport.com keeps a running tally: we learn that at the start of October 2004 Bush had a 53 per cent favourable response, declining to 51 per cent on 9-10 October, and rising to 55 per cent on 14-16 October. ‘Managing reputation’ has become the key preoccupation of prime ministers and presidents, as well as CEOs.

But the difference is that parties and campaign groups use poll results in political arguments, whereas market research is of interest only to the corporation concerned. When organisations lack an engaged mass membership, poll results provide moral authority, a phantom demos in whose name they can speak.

In the recent debate about foxhunting both sides marshalled polls in their defence. John Barter, head of the British Polling Council, which was set up partly in response to political scraps over poll results, says that the general polling regulator, the Market Research Society, was inundated with people querying each other’s results. ‘The two sides of an issue would try to cast doubt on the other side’s polling, and make complaints about the professionalism of the companies involved.’ Organisations allied to the British Polling Council publish their research methodologies in an effort to dampen disputes – but it’s possible that this will just provide the different camps with more ammunition.

The politicisation of polls has a deadening effect on public debate: rather than arguing their corner and trying to influence popular opinion, organisations pick through each other’s results, taking up this method or that calculation. This is a strategy for non-intervention, assuming that the public is a passive lump that has already made up its mind.

What we are seeing is representative democracy going through the motions. Opinion polling is a way of giving the appearance of a connection between political bodies and the public. In contrast to the messy maelstrom of public debate, we have the sanitised world of the People’s Panel, where individuals (kept safety apart) respond to a pre-determined set of questions. The fact that polling requires special bodies with special scientific techniques for gauging what people really think indicates its artificiality. Polling bodies fight ferociously about methodology, each claiming to have a hotwire to that elusive beast, public opinion.

Rather than argue what they believe, leaders try to follow the shifts and moods of the polls. We have now reached the logical consequence of that process: Your Party, a polling organisation posing as a political party (3). Its MPs vow to bow to the results of a constituents’ poll on any political issue, and argue that case in parliament even if it is contrary to their own beliefs. ‘MPs will often have their own point of view, and where appropriate they should try to persuade their constituents’, says the party’s candidate for Holborn and St Pancras David Woods, not very convincingly.

All this isn’t the fault of polls, of course. ‘There is nothing intrinsic to polling that means that political debate is deadened’, says Peter Kellner, chairman of YouGov. ‘What would it be like if you had no polls at all – would you get back to an Athenian idyll?’ Polls aren’t evil American techniques corrupting British political life; instead they are pulled in to plug the hole in representative democracy. Pollsters are just businesses providing a service that people want to buy.

In his excellent history of the transformation of the public sphere, Jürgen Habermas makes a distinction between public opinion and public opinion polls (4). Public opinion isn’t just the sum of individual views, but instead is forged in the white heat of public debate. Only by debating one another in rational argument do citizens come to firm points of view. While opinion polls take a snapshot of private individuals, public opinion is the mindset of society as a whole. This is perhaps why opinion polls fluctuate from hour to hour: the opinions we come to in the privacy of our living rooms are often a lot more tentative than those reached through rational argument.

Opinion polls are notoriously unreliable and variable, failing to predict the Tory victory in 1992 and falling wide of the mark in 1997. One MORI poll last week put the Tories on 39 per cent and Labour on 34 per cent; but another MORI poll for the Observer last Sunday put Labour on 40 per cent and the Tories on 33 per cent.

As 5 May approaches, we could well ask: what is the difference between an election and an opinion poll? If the parties bicker rather than engage in political debate, and if most people switch off for the duration of the campaigns, there is perhaps little to choose between the conclusions of MORI and the results of the Electoral Commission. Perhaps 5 May will just be one big opinion poll.

(1) Opinion Polls and Volatile Electorates, Matt Henn, 1998

(2) The Unfinished Revolution, Philip Gould, Abacus, 1999, p116

(3) See Your Party

(4) Structural Transformation of the Public Sphere, Jürgen Habermas, MIT, 1991