Every day from now until Canada’s federal election, a new poll will be released somewhere. Some will be right, some wrong, but you won’t know which until October. Some will be biased, but maybe not intentionally so. Some will be good polls that get no coverage, and some will be bad ones that spawn giant headlines. If you ignore all of them, you won’t be an informed voter. If you pay attention to every new set of numbers, you’ll go crazy.
So how can you tell a good poll from a bad one? When is it a party or leader’s jump a sign of a real shift and not an outlier? How can you tell which events are actually driving public opinion, and which ones will fade without a lasting impact on the race? Why do we like to talk about the polls being wrong when they’re right far more often than not? What do the reporters and analysts writing and talking about polls get wrong when new data lands? And how can you catch them in the act?
GUEST: David Coletto, CEO of Abacus Data