And now for something entirely different . . .

In a book entitled British Security Coordination, published in 1998, which was an internal history of the British intelligence campaign during the Second World War against German influence in the Americas, the following statement appears: ‘As WS learned, there was little doubt that Gallup deliberately adjusted his figures in Dewey’s favour in the hope of stampeding the electorate thereby . . .’

The WS was William Stephenson, a Canadian, and head of the BSC (British Security Coordination), and himself the subject of a later book, A Man Called Intrepid.  The Gallup mentioned was George Gallup, founder and director of the famous political polling organisation.  The Dewey was Thomas E. Dewey, the Republican candidate in the 1944 election against Roosevelt.

Whether true or not, the above statement was of no great surprise to me, who had always believed in the dubiousness of opinion polling.

There are several ways in which logically polls could be skewed.

One is through deliberate interference from the top, as is suggested in the case of Gallup, though nowadays, one imagines, it would be much more difficult to achieve.

Another can occur at the interview level, especially nowadays when so many young people—and the ground-level pollsters tend in general to be young—are ideologically committed.  In the two most recent Irish referendums, one could make a pretty accurate guess as to how people were likely to vote simply by looking at them, and choosing one’s interviewees accordingly.

A third way lies in the framing of the questions and the supplied range of readymade answers.

As well, of course, we all tend to be somewhat vague in our opinions, and, when effectively door-stepped by a questioner, tend to reach for the most convenient answer in order to hide our ignorance.

But the really interesting part of the quotation above is the belief that a poll could have the effect of ‘stampeding the electorate’ to vote for a particular candidate or party.  Although if such was the case with regard to Dewey, it was a failure, because he lost quite substantially to Roosevelt.

Yet the suggestion is that a poll can be more than simply a reflection of opinion at a certain point in time; it can, like a self-fulfilling prophecy, actually help bring about the outcome it is predicting.  Is this true?  I have always thought so, even long before the matter of Gallup ever arose.  Yet why should this be the case?

The best answer I can come up with is reflected in the Grand National.  Imagine someone who has 50p each way on the winner—and there are usually plenty of them.  Their winnings tend to be paltry, yet the fact is that they are still generally over the moon.  And this is because they have picked the winner, especially when they can see loads of their friends and acquaintances who haven’t.  In other words, they receive a psychological boost from it, some sense of increased status vis a vis the rest of the world.

And I really do think that this is something inherent in the human heart, this desire to be associated with winners, even if only forecast winners, in whatever field it may be.  And especially so in the context of referendums and elections, where unless you have firm convictions, which many people tend not to have, you are wide open to suggestion and to being swayed.

Which is why for a very long time I have been opposed to the idea of the publication of polls in the weeks or months before voting, especially where political parties etc. tend to commission a selection of polls, publishing only those that give them the most positive profile.

Indeed, I tend to be suspicious of polling in general.

The psychological mechanism that makes people vulnerable to manipulation can be viewed in different ways.  Certainly ‘authority’ seems to be a factor in it.  It is generally recognised by behavioural scientists now that there in an inbuilt tendency in humans to obey authority.  This is something individually and socially variable in its effect, reflected no doubt in the fact that Germans always seem to come top of the various surveys.

The question is what constitutes ‘authority’; and who possesses it?