YouGov MI/CO/NH/WI/NC/FL/PA/NV/IA/OH/GA: Clinton up, tied in GA

[Ozymandias]

This isn't a "real" or "pure" 50-state poll-- IIUC, then like Morning Consult from a few weeks ago, it uses a technique called MRP to model characteristics of voters across the country, and then applies that model to its entire national sample to make predictions for each state.

That's why there's no sample size per state-- because every polled voter has a small effect on every state (which of course varies from state to state, depending on the characteristics of the voter).

[Ozymandias]

(Personal note): I like this daily tracker, think it has a legit methodology and should be included in the 538 model

Actually, the results in most states are so similar to the margins that 538 predicts that I suspect YouGov uses the 538 predictions to help calibrate their model, because the fit is generally just as good in lightly-polled states as in heavily-polled states, which you wouldn't expect from an entirely independent model.

So I think 538 using these results-- which again, aren't actually a poll-- would be redundant...

[Ozymandias]

https://today.yougov.com/us-election/

Clinton's lead back up to 4.6 in YouGov's national survey (which is an actual poll, unlike modeled state numbers):

Clinton 48.2 (+0.8 )
Trump 43.7 (-0.8 )
Johnson 4.1 (+0.0)
Stein 1.9 (-0.1)

"Our election model includes 78512 interviews and has been updated with 7695 new interviews in the past 24 hours. It represents YouGov's best assessment of the current state of the race."

Full methodology: https://today.yougov.com/news/2016/10/03/election-model-methodology/

[Ozymandias]

EV prediction and trend: