Seven Lives from Mass Observation

Recent events have famously—notoriously—called into question the validity of opinion polling. The British general election of 2015 and the referendum of 2016, and the apparently irresistible rise of Donald Trump: all these have proven problematic for the pollsters. True enough, it has always been possible to find some polls that, with hindsight, cleaved closer to the end result than others. In fairness, it must also be acknowledged that the final percentages which yielded Tory victory, Brexit, and President Trump were closer to most of the polls than many critics have acknowledged. But the fact that even psephologists seem baffled—apparently unable to find a representative sample of people or ask them the right questions or interpret their answers appropriately—is disquieting nonetheless. If this is true of professional, supposedly scientific polling, then it should lead us all to be very cautious about using other, less structured sorts of survey. When local newspapers, for instance, claim that attitudes to the European Union have suddenly, radically changed, with ‘a massive swing from Brexit to Remain’ reported by the Newcastle Chronicle in December 2016, it is worth remembering that the self-selected, self-appointed participants in small-scale polls are necessarily unrepresentative. Simply participating makes them bound to have an axe to grind

Brexit · Economics · European union · Hindsight bias · Newspaper · Opinion poll · Political science · Politics · Polling · Presidential system · Public opinion · Referendum · Sociology · Victory · Data Analysis and Archiving · Psychology · Law · Social Psychology