I do prophesy th’ election lights …

… “on Fortinbras”, says Hamlet. A victory for the strong-arm man. I happen to be teaching Hamlet this week, a play very interested in succession and election. Also, though, a play in which people lie a lot. One thing that strikes me about the current election is the huge discrepancy between nearly all the opinion polls and the result. Exactly as in the 2015 Conservative victory in the UK, 2016 Brexit and 2016 Trump. It’s not my field of expertise, but this further debacle for the polling industry does suggest to me that there is a fundamental problem with the social science model of human behaviour, in which statistics, surveys and mathematical projections are used to make predictions. It was striking that the only pollster to go against the grain, in 2020 as well as 2016, was Robert Cahaly of The Trafalgar Group, who said that the premise of his methodology is that “people lie all the time”. Which they – or at least Claudius, Polonius, Rosencrantz and Guildenstern – do in Hamlet.

Was it not self-evident that Trumpists, who are predominantly non-college-educated and almost entirely suspicious-of-“mainstream”-media, would either lie or refuse to answer when confronted with an orthodox poll that (a) used the social science model that comes out of college education, and (b) would be reported in, and therefore associated with, the mainstream media? Shouldn’t it be obvious by now that people act more on the basis of emotion than reason and that a “rational”, quasi-scientific methodology is a hopelessly inadequate way of predicting human behaviour? As so often, the humanities – the lessons of history, the thought experiments of Shakespeare – offer a far more useful predictive tool than the social sciences.

What we need next time around is a new opinion polling model based on Shakespearean as opposed to statistical principles!