Discussion about this post

User's avatar
Shane's avatar

Certainly feels like the model is being way too aggressive hedging against the convention bounce, and is effectively giving Trump a bounce instead.

Expand full comment
John Endres's avatar

The Pennsylvania polling shift has the appearance of Simpson's Paradox*. The model is mostly weighting pollsters who have Trump ahead, but those pollsters are actually showing better results for Harris than they were the last time they polled. Curious what others think of this.

*If you take each pollster on its own, Harris is improving, but the composition of pollsters (and their weights) is changing in Trump's favor.

Expand full comment
375 more comments...

No posts