USC/Dornsife tracking poll megathread

[Adam Griffin]

Just found this.  Back in 2016, this economist constantly reweighted the 2016 LAT/USC poll based on the American Community Survey and other demographic data.  Not only did his "reweighted" version come within 0.2% of the final margin on Election Day 2018 (his final margin was Hillary +1.9, the actual was Hillary +2.1), but his estimations basically worked perfectly as a leading indicator of the RCP + other poll aggregates that was ahead of them by a week or a little less.

https://twitter.com/ernietedeschi/status/796053756915908608

https://sites.google.com/site/latuscrw/

I went ahead and tweeted the gist to him. His response:

https://twitter.com/ernietedeschi/status/1300392242142683137

[Adam Griffin]

Probabilistic
Biden 52.8 (+0.1)
Trump 40.6 (-0.5)

Traditional
Biden 54.3 (-0.5)
Trump 40.0 (+0.2)

Can someone explain in layman's terms the difference between the two methodologies?

Best I can tell, "probabilistic" only features the two major candidates, asks how likely they are to vote in November & asks how people will vote in November, while "traditional" includes Libertarian and Green candidates & only asks how they would vote if the election were held today. Probabilistic seems more relevant to me based on balance.

Probabilistic

What is the percent chance that you will vote in the presidential election? (0-100%)

If you do vote in the election, what is the percent chance you will vote for Joe Biden (Democrat), Donald Trump (Republican), Someone else? (Answers add to 100%)

Traditional

If the election were being held today, would you vote for: Joe Biden (Democrat), Donald Trump (Republican), Jo Jorgensen (Libertarian), Howie Hawkins (Green), Undecided, Would not vote in the election. Undecided are asked: As of now, do you lean more toward voting for Joe Biden (Democrat), Donald Trump (Republican), Someone else, Do not lean toward any candidate.

https://uasdata.usc.edu/index.php?r=eNpLtDKyqi62MrFSKkhMT1WyLrYyNAeyS5NyMpP1UhJLEvUSU1Ly80ASQDWJKZkpIKaxlZKJgZGSdS1cMGzwEts