With mounting panic and terror (sorry about the cliches, but that’s how dreams go), I contemplate the awful reckoning at hand: the exams are only a week away and, barring miracles, I am certain to fail. Then I wake up with an immense feeling of joy, relief and liberation, not just because it was only a dream but because I realise that there is not the smallest chance that I could ever find myself in such a situation again.

I finished with exams 30 years ago and, even if I were to take one again as a result of some Third Age study course, the results couldn’t possibly matter in the way that they did in my youth.

Even in these days of extended higher education, the same goes for the majority of the voting population. And that is why I have always been sceptical when people say that education is going to be a big “issue” in an election and that it could significantly affect the vote.

Unlike health, taxation, pensions, crime and jobs, education does not reach the most visceral parts of the voters. Some people dream about heart attacks or car crashes and frantic searches for hospitals with spare beds; others will dream about being burgled or assaulted or raped. Their relief on waking is surely tempered by the knowledge - or, more precisely, the belief - that such drama is all too likely. The same must apply to those who dream about having to live on the pitiable state pension or about losing their jobs or being forced to retire early.

Education is about aspiration but, as any politician knows, fear is a much stronger emotion. I accept that people care about their own children - and fear for their future - but parenthood is a relatively brief interval in the average voting lifetime. Parents understand that it takes several years for political change to have any real effect in the classroom. And most parents, if they are honest, will acknowledge that their greatest fears are for their children’s health. It was Jennifer’s ear, not Johnny’s dyslexia, that dominated the 1992 election campaign.

All three main parties put education at the centre of their manifesto launches last week. But at the beginning of almost every election campaign that I can remember, politicians and press have proclaimed schools and colleges as a central issue and then proceeded to ignore them for the next three weeks.

Education policies are usually thrust forward for our admiring attention at the moment when politicians are assuring us that they intend to concentrate on their “positive messages”. Almost invariably, as polling day approaches, the negative takes over. Nightmares, usually conjured from the past, are the currency of politics. Fear of a return to the 1930s dole queues put Labour in office in 1945; fear of another winter of discontent has kept it out since 1979. Education simply lacks nightmares of comparable power and resonance.

Will it be any different this time? Should I abandon my scepticism about education’s capacity to swing an election? The early signs are not encouraging.

A couple of weeks ago, Gillian Shephard was due to announce new proposals on testing. She was swept aside to allow Michael Heseltine to launch some preposterous scare about Labour policy on trade union recognition.

But I rang Bob Worcester, director of the polling organisation MORI, and asked his opinion. “Can education influence the outcome of elections?” “You bet it can!” he replied. And he sent me some figures that seem to prove his point dramatically.

For more than a decade, MORI has been asking samples of the electorate to name “the important issues facing Britain today”. At the beginning of 1987, only 11 per cent mentioned education, against 75 per cent mentioning unemployment. At the end of last month, education got 43 per cent, unemployment 39 per cent. Education, indeed, was judged second only to the NHS in importance. This is a remarkable change, and quite a recent one: in December 1994, education was still getting only 18 per cent (which put it behind law and order and the economy as an issue) against 55 per cent for unemployment.

But I am not going to abandon my scepticism so easily. Poll respondents are apt to give the answers expected of them. People have been told they can stop worrying about the economy: unemployment is down, inflation, under control, growth steady. Even Labour doesn’t have much interest in spreading alarm because the experience of 1992 suggests that, when people are worried about the economy, they prefer the devils they know.

What people should worry about apparently - and all three parties, to say nothing of the entire national press, agree on this - is education. Indeed, to listen to the politicians, you would think that the only thing now holding Britain back from Singapore-sized growth rates was the incompetence of the schools. This is exactly what they want the public to think, because if anything does go badly wrong with the economy, then teachers, not politicians, can be blamed.

In these circumstances, it would be surprising if voters persisted in regarding education as unimportant. But I’m not convinced that makes the subject into a campaign issue, still less into one that swings votes. This is mainly because it now requires the skills of an old-style Kremlinologist to interpret the differences between Labour and the Conservatives.

One party, I think, favours foundation schools, the other locally-maintained schools, but I wonder if even Messrs Major and Blair can remember which is which, or what they are supposed to be. Either way, it’s not the stuff of nightmares.