Hillary Clinton lost the US election despite some polls putting her chances of winning at 99%.

In the run up to the vote pollsters spent huge sums of money speaking to thousands of Americans. They were careful to collect the best possible data from representative samples, and they applied their finest statistical minds to analysing the numbers. Yet almost no-one predicted that Donald Trump would win. So – our question this week – why did the polls get it so wrong? Our expert witnesses explain why polling is getting harder, and why many pollsters weren’t – despite that – very far wide of the mark.

(Photo: Hillary Clinton and Donald Trump listen during the town hall debate at Washington University in St Louis, Missouri, October 2016. Credit: Getty Images)