The opinion polls were memorably wrong about the 1992 general election. They were so wrong that some of them picked the wrong winner. And they were all wrong in the same direction - overestimating Labour and underestimating the Conservatives. An industry post-mortem established that one of the principal causes was a tendency among some Conservative voters not to want to admit how they were going to vote. They were dubbed the "shy Tories."
This tendency was first described by the German social scientist Elisabeth Noelle-Neumann, founder of the Allensbach Institute, who defined it as a "spiral of silence." Noelle-Neumann argued that public opinion is a tangible force that has a direct effect on people's decisions. The spiral of silence theory holds that people have a kind of sixth sense which allows them to know the prevailing public opinion, even without access to polls, in much the same way that they sense what is fashionable in other realms. At the same time, most people have an in-built resistance to being socially isolated. As a result, the wider the perceived distance between majority views and a person's own views, the less likely people are to express what they think.
The spiral of silence among Conservative voters during the 1990s was evident in significant numbers of Tory supporters defining themselves as "don't knows," or even as Labour voters, when asked by pollsters (and, more tellingly, friends and family) how they were going to vote. When the "don't knows" in polls of the period were analysed by how they had voted in the past, they were overwhelmingly former Tory voters, and as a result, at successive elections the Tories performed better than predicted.
The response of some pollsters was to put in place an adjustment process which re-allocated a proportion of the "don't knows" to the party a voter had previously voted for. The company that pioneered this approach in Britain, ICM, was by far the most accurate at the subsequent general elections.
For nearly a decade this methodological response to the spiral of silence had the effect of adjusting upwards the raw poll support for the Tory party - so that those using the approach consistently showed lower Labour leads than those who did not. Many pundits, as well as rival pollsters, came to regard it as an innately pro-Tory adjustment and to view it with corresponding suspicion.
But the spiral of silence adjustment does not axiomatically increase Tory poll ratings: it increases the chances of a poll accurately measuring support for any party that is out of fashion. The duration of the Tory party's estrangement from the centre of gravity of public opinion can be measured by the fact that its support was afflicted with the spiral of silence for so long - certainly for more than a decade from the late 1980s.
For two or three years after the Conservatives were expelled from office in 1997 there remained "shy Tories," and there continued to be a marked disagreement about their level of poll support between those pollsters that adjusted for the spiral of silence and those that did not.
Then, in early 2000, the adjustment stopped making any difference: the spiral of silence was gone; the number of Tory supporters who were reluctant to admit their voting intentions was no greater than for any other party.
With hindsight we can infer from the trajectory of the polls that the disappearance of the spiral of silence was only partly because the stigma of admitting to voting Tory had diminished. It was also a function of rapidly rising discontent with Labour. The tangible political mood stopped being to give Labour the benefit of the doubt, and started to be about disappointment, even bitterness, at their failure to deliver what they promised.
About a year ago, the spiral of silence returned - this time not among Tories, but among Labour voters. The spiral of silence adjustment made by some research companies to their raw data started to increase Labour's poll support. The trend has persisted now for a year and is slowly growing, now improving Labour's poll position by as much as 4 per cent relative to the other parties.
For the first time in 20 years there are more shy Labour voters than shy Tories. The spiral of silence has reversed itself - indicating a significant subterranean shift in the momentum of the political tide, albeit one whose full effect may not be seen for several years. Labour is likely to remain in government for some time yet. If the new spiral of silence plays out as the last one did, it will be many years before the perception of unfashionability that causes it drives Labour's shy supporters to switch sides. But analysis of the spiral of silence may well allow us to look back and point to the moment the Labour government's end began.
This new spiral of silence brings with it a risk that the Conservative party will as a result believe itself to be much closer to power than it really is, halting any impulse to respond further to the new centre of gravity of public opinion, and misaligning its message - and tone - as a result.
This risk is all the greater for the fact that the spiral of silence among Labour voters is highly unlikely to unwind on 10th June - when there is a glut of mid-term elections, but one that most voters see as neither important nor interesting enough to engage with. Labour supporters in the spiral of silence are quite likely not to vote at all in the European elections, the local elections or the London mayoral elections. The Tories will do deceptively well in June. It is when it matters - at the next general election - that history suggests the spiral of silence will count.
