Boston Globe assesses current polling

[Pototan]

▪ Ghosts of 2016: Pollsters still feel a visceral dismay over state-level polling misses in 2016, which overestimated Hillary Clinton’s performance in key states. Overcompensating for that mistake could lead pollsters to design samples and weighting schemes in a way that overestimate Republican support this time. For example, some pollsters are asking voters how they voted in 2020 and then weighting based on that. But The New York Times pollster Nate Cohn noted that polls that weighted by past vote have tended to be more favorable to Trump. Pollsters accounting for partisanship in other ways such as weighting the data to partisanship targets could also lean too far to the right.


https://www.bostonglobe.com/2024/10/25/opinion/election-presidential-polls/

[NCDem47]

1. Kamala WILL prevail, but keeping me up at night:

African-American men and young white males (educated and non-educated). Something is brewing there to favor Trump in small ways that may make it a nailbiter. I think motivated females of all political stripes going for Harris will offset that.

[kevink077]

2. 2016 was different

Hillary was (unfairly) unpopular and the public was convinced Trump was not that bad. Many liberals sat out because she was not perfect enough. Not going to happen this time. I expect Harris to win comfortably

[Quixote1818]

3. I've been arguing this for weeks. Nt

[RJ_MacReady]

4. Are they suggesting

Trump is being oversampled?

[Chasing Dreams]

5. No, Trump is overweight(ed)

[Prairie Gates]

6. So, they are experimenting with methods in real time and reporting out those experimental results

Very science-ish.