Is American political polling a dying enterprise? Polls in 2016 were excellent at forecasting the final vote totals nationwide, but not nearly as good at forecasting outcomes in individual states. In 2020, polls under-counted support for Donald Trump even more than they did in 2016.
Why is that? And if polling is in trouble, why are polls for other political races much more consistent and effective? What accounts for the differences? Our guests discuss it: