Watching cricket at Lord’s on one of those days when the appearance of a saucer-sized cloud overhead induces batsmen to peer down the wicket as if a pea-souper has descended and to feel their way about the crease as though their willow in their hands was a white stick, I and scores of others groaned as the umpires conferred and reached for their light meters.

When, inevitably, a few minutes later, the players were offered the light and dawdled back to the pavilion, the man sitting next to me, who looked like Lord Emsworth up from the country for the day, tapped me on the knee and confided, “The trouble with those confounded meters is that it makes the thing look scientific.”

And that, dear reader, is exactly the trouble with opinion polls. It makes the thing look scientific. To their satisfaction and enrichment, the pollsters are the soothsayers of the age. Whereas Caesar had the augurers who, plucking the entrails of an offering could not find a heart within the beast and thought that was a pretty rum omen, Bill Clinton has the heirs of Dr Gallup who pluck feathers from the minds of the electorate. Unlike Caesar, however, Bill acts on the vapourings of the soothsayers. In 1995, for example, he went “on a mountain vacation with hi-tech gear” because a poll showed that swing voters liked camping, hiking and technology. Later, he confessed that he had not enjoyed the experience.

It is unlikely that Slick Willy would have crept under canvas at high altitude had his advice come from the lights of a disembowelled tomcat. He went because he believed the thing was scientific.

It is perhaps a tribute to the greater wisdom of politicians in our more mature democracy that they are beginning to smell rotting innards in the polls themselves. Tony Blair and his advisers are said to have reached the conclusion that people lie to pollsters, particularly about their willingness to pay more tax. This, says one report, is not an uncommon phenomenon, even among people who have been assured that their answers will be treated in the strictest confidence. Sex surveys, for instance, are notorious for eliciting dishonest answers, as are inquiries into drinking habits. People seem to take the healthy view that what they get up to between the sheets and what they pour down their throats at other times, is no business of prodnoses with clipboards. However, a British disinclination to give offence, either verbal or physical, prompts those surveyed to smile politely and lie through their teeth.

There is another reason for lying, according to Stewart Steven of The Mail on Sunday. When MORI recently asked people who they had voted for at the last general election, it was able to report a solid Labour lead of 9.5 per cent. “People aren’t prepared to admit to holding unfashionable views,” concludes Steven.

Perhaps stung by a growing suspicion that the thing is not scientific after all, Gallup last week announced a change in its methodology. Instead of face-to-face interviews conducted in the street and in people’s homes, the organisation will henceforth base its political polling on telephone interviews obtained by “random digit dialling”.

Explaining the change, Gallup says trends since 1992 have put in doubt the traditional method’s reliability: “In particular, Gallup’s interviewers appear to have been reaching too many staunchly working class voters and not enough of the better-off.”

This probably reflects well on the working class, whose natural bonhomie and good nature makes them more approachable, and therefore amenable to answering a series of tomfool questions, than the better-off who tend to be more stuck-up and standoffish.

Telephoning people at random, however, will not overcome the problem of separating the mendacious from the truthful. Indeed, it could be argued that lying over the phone is easier than doing so face-to-face since the liar is spared the questing eye of the interviewer.

Isn’t it time to admit that polling is no more scientific than geomancy or clairvoyancy, and no more accurate than phrenology or divination, but every bit as much fun? The media certainly think so, which is why so many radio stations and newspapers invite people to take part in telephone polls on a variety of topics and issues, from the size of Fergie’s bottom to the desirability of gun controls, two subjects, incidentally, that might be more closely related than at first appears. The surveys are profitable since the calls are charged at a high rate, and they are entertaining. But not everyone likes them. Some maintain they are unscientific and easily rigged. Professor Robert Worcester, chairman of MORI, says: “These are voodoo polls, they are a farce, completely self-serving and very dangerous because they mislead the public.” Far better to stick to the traditional arrangement whereby the public misleads the pollsters.

Professor David Butler, Oxford University psephologist and pioneer of the swingometer, describes telephone polls as monstrous and intolerable. “It’s a kind of deception. They should have enormous health warnings.”

But Professor A Spokesman of the BBC says: “We realise the polls are unscientific. But they are not really riggable.”

Professor Dickie Bird takes much the same view of his light meter.