Analysis: Even in a close election, random chance means polls should be showing a broader range of results. That raises the question of whether we’re in for another polling surprise.
Not entirely. A few months ago AOC was discussing how her own internal polling of her own district ended up under-estimating her support by around 10 percentage points. It was in that hour long talk she gave explaining why she was still supporting Biden as the candidate, before he dropped out.
Polling has always been tricky, but I think in the past decade its gotten nigh-impossible. These institutions now seem to be more focused on not losing their jobs than actually trying to gauge support for a politician.
Makes me wonder if issue polling instead of politician polling is better. I imagine it probably is a little bit, but I’m not sure.
Go down to the district level and it’s a different story.
Not entirely. A few months ago AOC was discussing how her own internal polling of her own district ended up under-estimating her support by around 10 percentage points. It was in that hour long talk she gave explaining why she was still supporting Biden as the candidate, before he dropped out.
Polling has always been tricky, but I think in the past decade its gotten nigh-impossible. These institutions now seem to be more focused on not losing their jobs than actually trying to gauge support for a politician.
Makes me wonder if issue polling instead of politician polling is better. I imagine it probably is a little bit, but I’m not sure.
District polls of PA-07, -08, and -10 suggesting either a 2020 redux or better.
State polls show a tie.
Which is right? Both?
This article has to do with pollseters herding to results. They are making the same assumptions. District level polls is one proof.
Why not none of them?