What opinion pollsters do after elections

"Most people desperately want the election to be over and done with. I am the opposite. I hope it never comes."

The speaker is tall, handsome, silvering Robert Q Sample, boss of the mighty Mururoa Poll empire. As he sits atop Mururoa Towers, the tall, handsome, silvering office block in London from which he directs the opinion- tasting of the nation, he stares out over the population of Britain who are, figuratively speaking, his bread and butter.

"Did you know that 67 per cent of the British people get the words `figuratively' and `literally' mixed up?" asks Sample, uncannily reading my thoughts. "In a survey we did in 1987, two-thirds of the nation could not see anything wrong with the statement `I literally had kittens'. It should, of course, be `I figuratively had kittens', because `figurative' is the word to use when talking of the opposite of `real'."

Which is odd, I point out, when you think that in the world of painting it is quite the opposite.

"Before election-time," orates Sample, "we in the opinion-testing industry are in the news, in demand and on fire. This is our time! This is when the voice of the poll is heard in the land! This is when the good and the bad, the ugly and the wealthy, beat a path to our door to find out what people are thinking. Better still, they come to ask us what people want to hear. If we tell them that the voters are sick of sleaze, they go away and strike sleaze off the agenda. Then they come back to us to find out if it has made any difference."

And after the election?

"After the election we are as much in demand as a man with a broken leg at a disco," says Robert Q Sample, his eyes going suspiciously moist. "Whoever heard of an opinion poll being taken the day after an election? The week after? The month after? Slowly, slowly, after that it revives, when people start asking questions like, `How do you think the new government is doing?' and, `How were its first 100 days?'. But it would be fair to say that nine out of 10 pollsters take a long holiday after any big election."

He stares out of the window ruminatively, down at the streets below, where Londoners like ants scurry to and fro.

"Odd, isn't it?" he says, half to himself."It's mid-afternoon. Most people are at work. Then who are all these people walking about in London? Where are they all going?"

There is a short pause.

I suggest that he must have done a poll on this as well.

"Oh yes, we did. How could I ever forget that poll?"

There is another pause and a slight sniff of emotion.

"We stopped a sample of people walking through London streets in mid- afternoon and simply asked them where they were going. And do you know what? Most of them lied to us. They all looked shifty anyway, but most people had something to hide. They were late back to work after lunch, or back from the bookie, or fresh from an assignation. This was the first time we came to realise, reluctantly, that people do lie in answer to pollsters' questions. For me it was a day of tragic implications."

And have people been lying to you in the 1997 election run-up?

Robert Q Sample turns a gaze on me brimming with tears. "How else can you explain the wild vicissitudes? One moment it is `Labour heads for landslide', then `Labour's lead melts away', and next moment `Labour heads for landslide' again. How can it vary so wildly unless people are lying to us? Do you think we are doing it for fun?"

In the ensuing short pause, I hear what I can only describe as a chuckle. Robert Q Sample is chuckling.

So that rogue poll in The Guardian last week ...?

"I'll tell you something, my boy. We are doing it for fun. When the polls tell the same story, we like to vary it a little. If the polls never change, people start to think that they don't need polls, and that would never do. So we play around a little with the results. But you must never print this. Do you promise?"

I give him my solemn promise and leave, more thoughtful than when I had arrived.