LA Times / USC Daybreak poll

[/dev/null]

http://graphics.latimes.com/usc-presidential-poll-dashboard/

 

Interesting polling method. Instead of taking the traditional random sample of ~1000 for each poll they recruited a random panel of ~3000 and poll a random sample of ~400 of their panel every day asking the same 3 questions:

 

(1) percentage chance you will vote in the presidential election

(2) percentage chance you will vote for Clinton, Trump, or someone else

(3) percentage chance Clinton, Trump or someone else will win

 

They're coming up with much different results than traditional polling methods

[TH3]

http://graphics.latimes.com/usc-presidential-poll-dashboard/

 

Interesting polling method. Instead of taking the traditional random sample of ~1000 for each poll they recruited a random panel of ~3000 and poll a random sample of ~400 of their panel every day asking the same 3 questions:

 

(1) percentage chance you will vote in the presidential election

(2) percentage chance you will vote for Clinton, Trump, or someone else

(3) percentage chance Clinton, Trump or someone else will win

 

They're coming up with much different results than traditional polling methods

By traditional you mean the ones that are traditionally correct? Edited August 21, 2016 by baskin

[drinkTHEkoolaid]

That is interesting

[KD in CA]

By traditional you mean the ones that are traditionally correct?

 

Do those traditions include 2000, 2004, the 2016 Republican primary, the Brexit vote, etc., or just landslides?

Edited August 22, 2016 by KD in CA

[unbillievable]

Just pick out one random guy and ask him the three questions 1000 times to be more efficient.

[/dev/null]

By traditional you mean the ones that are traditionally correct?

Spoken like a French officer manning the Maginot Line

[Dante]

By traditional you mean the ones that are traditionally correct?

Maybe, just maybe, people hate that evil communist troll more than they hate Donald?