Testing public opinion

I'm sure you can bear to wait a week or so for a comment on the second and final debate, so-called, between the two presidential candidates.

What I, or for that matter any other individual, thought about it is of no consequence. What matters is what the people who saw it thought about it. In other words, the millions who tuned in to one or other of the four national networks on Thursday evening or rather the 2000 or so people tapped as a statistical sample of the whole. I'm told there are still people, otherwise intelligent, who say, 'How can they possibly tell what a whole nation feels about anything by interviewing 1100 or 2000 individuals?'

We went into that in a talk after the death of George Gallup, who was surely not the father of statistical method, but he did, more than anybody before him, refine it and apply it successfully as a gauge of what used to to be called public opinion.

Polling, so-called, started before he came on the scene. It was not, in the beginning, remotely scientific and the naiveté of it was exposed and exploded, once for all, in 1936 when an American magazine, the Literary Digest, polled the voters of the United States about the forthcoming presidential election between the incumbent, President Franklin Roosevelt, and his Republican opponent, the governor of Kansas, Alfred Landon.

The Digest announced on the eve of the election that Governor Landon was going to be a clear, a very comfortable, winner. In the result – this was before Hawaii and Alaska had joined the Union – in the result, Roosevelt took 46 states and Landon took two, the rural New England Republican states of Maine and Vermont. Up to that time, even the most astute politicians who had not travelled around the country used to guide their hunches with old political maxims and mysterious sayings.

For many years, the most northern, north-eastern of the New England states, Maine – anyone remember the book or the movie, 'Way Down East?' – Maine was watched, almost revered, as a barometer of national politics and it did happen that Maine always seemed to vote for the winner in presidential elections. Hence, the saying, one of the first bits of Americana I picked up, 'As Maine goes, so goes the nation!'.

After the Roosevelt massacre of Governor Landon in '36, Roosevelt's campaign manager amended the old folk saying, 'As Maine goes, so goes Vermont'. We never heard the original again.

The Literary Digest, it came out, had simply telephoned a lot of people at random and asked them who they were going to vote for. To begin with, the actual gesture of telephoning cut out of the sample all those people who hadn't got telephones which, 52 years ago, was a considerable number of Americans. Secondly, the question was put bluntly and undoubtedly isolated only people with their minds made up. No account was given of the undecided, in fact, the notion of a body of undecided opinion had not occurred to the pollsters.

Well, I won't poke around any longer in the wound suffered by the poor old Literary Digest. It – the magazine – didn't merely go out of the polling business, it went out of business and its dramatic collapse was what actually inspired the young George Gallup to ponder and work on the prospect of applying statistical method to public opinion.

I think I once mentioned the time I talked to Dr Alfred Kinsey, the American zoologist who, in 1948, released a bombshell on the world. He'd turned from his study of bats and mice to the human being, specifically to the sexual habits of the human male. He was the first person, I believe, to generalise about human sexual conduct not from the cases, the people, he'd known, but from several years of interviewing men, over 18,000 in all, using scrupulous interviewing techniques and meeting criteria acceptable to professional statisticians.

When he published this first book of findings, 'Sexual Behaviour in the Human Male', it appeared – it was proved – that men did not, in their behaviour, fall into a few simple categories accepted for generations by doctors, parsons, publicists, novelists and other students of human behaviour. Dr Kinsey shocked everybody, in public anyway, by describing very wide variations in male sexual behaviour.

The shock was severe in proportion to the depth of your feeling about how people ought to behave. This, swore Dr Kinsey, is how they did, and inevitably the professionals, the doctors, the psychologists, the novelists, who'd always assumed they had a monopoly on knowing about human behaviour, were the first to protest and say the Kinsey sampling was irregular and the manner of personal interviewing must have been prejudiced, and so on.

But Dr Kinsey's work was a landmark and, after him, more and more people, the public opinion pollsters especially, went to work to refine their techniques and see how small a sample of the public could be dependable.

Well, after a lecture at the New York Academy of Medicine, I got hold of Dr Kinsey and, in time, asked him just how far he thought he or any other statistician might go in using a sample as an accurate reflection, microcosm, of the whole population. I was thinking then of people's opinions, war, peace, political issues, so on. He said, quite slowly and gravely, 'The time is coming when we can go down to Time Square and choose a sailor, a housewife, a businessman, a waitress and a truck driver, perhaps no more than a dozen individuals, and tell you pretty accurately that they represent the gamut, the variation of American public opinion'.

I goggled, as I was expected to do. I never took it up with him after that but I can say for certain that we are not there yet. However, the techniques are advanced enough to make a careful nationwide sample of something over a thousand adults a reliable guide to the ebb and flow of public opinion. The New York Times and the Columbia Broadcasting System together published such a poll on Thursday morning, on the eve, or the dawn, of the Bush-Dukakis debate.

This was not a survey, as often sloppy news reports say, of 'the American people' – the American people, whoever they are, except 245 million individuals – the American people don't vote, not even all the registered voters vote. Only about 52, 53 per cent usually vote – the smallest participation of the voters in any participatory democracy.

However, for the first time that I've noticed anyway, this latest Times-CBS poll selected 1518 people qualified to vote and asked them all first how sure they were that they would vote at all. This brought the surveyed sample down from 1518 to 1009 and before the interviews continued, it was decided that 509 were almost certain to stay away from the polls. The main questions were more interesting than most because they addressed an anxiety about Mr Bush's choice of vice president that has, if anything, deepened since the two vice presidential candidates, Senator Benson and Senator Quayle, debated a week ago.

Soon after that debate, the three or four most responsible pollsters reported much the same conclusion, that while the young Senator Quayle started out firm and confident, reciting practised answers to questions he'd expected, he wobbled badly when he got one or two he hadn't anticipated, especially the question, what would he do if he suddenly became president?

He was thrown. 'Err...'. He would pray. He then desperately bypassed the question by saying it was hypothetical and what the questioner really meant was, what were his qualifications? He then went on about his experience in the Senate, but that's not what the questioner wanted to know. Three times they put the same question. Nothing.

Well, the Times-CBS poll asked, 'If you had to vote only for vice president?' The result 55 per cent for Benson, 30 per cent for Quayle. Two voters in three, invading party lines, do not think Mr Quayle is qualified to be president. It's the one drag on the Bush campaign.

The Dukakis camp is naturally greatly cheered by this thought, but not so cheered by a Bush advantage that has been there from the start. It's called the Solid South and, these days, it's an ironical phrase because it was invented to describe, through the first half of this century, the certainty that the whole South would go wholly for the Democrats.

Eisenhower broke the bloc and Nixon took most of it away and in 1980 and 1984, the whole South went for Reagan. The Solid South was now solidly Republican.

And in the South, today, Bush leads by 51 per cent to 36. Now since it's not the total number of votes but the electoral votes that count, Mr Bush could win in each of nine states by only one vote and take the whole electoral pot. Not counting Texas, it looks to him and to the Dukakis camp, as if he would take 97 electoral votes down there – more than a third of the 270 electoral votes needed to win.

It's a giant lead to have to overtake.

This transcript was typed from a recording of the original BBC broadcast (© BBC) and not copied from an original script. Because of the risk of mishearing, the BBC cannot vouch for its complete accuracy.

Letter from America audio recordings of broadcasts ©BBC

Letter from America scripts © Cooke Americas, RLLP. All rights reserved.