As a former market researcher, I’ve been watching the apparent decline in the UKIP vote, as shown by the polls, with some scepticism.
The important voters, as we all know, are the ones who voted UKIP for the first time in the European election. UKIP’s success or failure in the General Election will depend on the extent to which these voters repeat their UKIP vote this time. There are a lot of people in this position, millions, probably more than 10% of those likely to vote. They are well enough aware of the importance of their voting decision. They know that UKIP is a new and dramatic political phenomenon. They have thought about it a lot, often discussed it within their families and sometimes with a few friends.
But even in the family, and very often when talking to friends, they’ve found that the very mention of UKIP may arouse strong feelings, often very negative feelings. So they have become a little reserved and careful what they say. In many, perhaps most cases, they have not even made up their own minds how they will vote in May, or even whether or not they will vote at all.
When they happen to interviewed on a political survey, they’re probably somewhat reluctant to declare their voting intentions. In answer to the straight question, they prevaricate. The easiest thing for them to do is to fall back on telling the interviewer or the computer how they usually vote, or how they have voted in the past, or to say they haven’t decided. My experience of interviewers is that they are always eager to get on to the next question, and they will usually accept any answer provided it fits into the box. If the respondent says “I have always voted Labour” it will go down as a Labour vote. If they say “I’m afraid I often don’t vote at all” it will go down as a non-voter.
This suggestion is not an entirely new one. Something called the “Shy Tory Effect” was identified after the opinion polls had failed to predict the Conservative victory in the 1992 election. The idea has gained wide currency and is quoted in US analysis of poll discrepancies there too. It is suggested that voters for some parties tend to be more reluctant to disclose their voting intentions than voters for others. Pollsters try to discount this effect by weighting methods.
The “Shy UKIP Effect’
No doubt the “Shy Tory Effect” is still affecting polls and it’s certainly still being allowed for. But what about the “Shy UKIP effect”? There is no weighting for that. Yet any embarrassment which a voter may currently feel about disclosing an intention to vote Conservative is surely negligible compared with the embarrassment which is currently affecting many of those who intend to vote UKIP.
We have had recent evidence of the shy UKIP effect in Heywood and Middleton. An Ashcroft poll before the by-election there gave Labour a 19-point lead. In fact Labour won by only two points. By contrast, the Ashcroft poll in Clacton, where voting took place on the same day, proved to be quite accurate. Why the difference? Well, the difference was that in Clacton, a vote for UKIP was much more socially acceptable because it was, after, all a vote for the candidate who, a few weeks before, had been the sitting MP. Much of the UKIP vote in Clacton was a Carswell vote. By contrast, in Heywood and Middleton, the UKIP candidate was a new name to most voters.
If what happened in Heywood and Middleton is anything to go by, UKIP will do a great deal better in the coming election than the polls currently suggest. And those in the media and the old parties who are writing UKIP off are going to get the shock of their lives.