The Times - of all outlets - interviews Simon Rosenberg, the author of the Hopium Chronicles substack, about his prediction that Biden will win in November.
One thing I think he says that might be worth emphasizing is that we have a sort of "polling industry mindset" that really isn't predictive of much in politics, especially this far out from the election. Nate Silver, Nate Cohn and others have all made us feel like polling experts when we ask to look at crosstabs, sample sizes and screens. The reality is that polling is an approximation of any given moment. Polls are what they are, but they are not elections. Arguments that Trump shocked the world in 2016 so therefore he is going to shock the world in 2024 are fallacious reasoning. It's equally as likely that Biden wins Texas and Florida; polls can be wrong in the other direction. Witness 2022.
We have started to see some real movement towards Biden as it becomes clear that it will be Biden v Trump. However, it's important to remember that traditionally few professionals used to pay any attention to polling before Labor Day. It was always considered largely noise.
No comments:
Post a Comment