FL - Bloomberg - Trump +2 (4 way)

[TheDeadFlagBlues]

Selzer's assumptions might work in Iowa but they might not be transferable to other parts of the country. For those who aren't aware, Selzer's methods are entirely based on voting file information and, to my knowledge, her screens are very strict. This is a great strategy for polling people for caucuses but it's risky before a presidential election because it has the effect of being unable to capture surges in first-time voters or low-voting propensity groups actually voting. This is probably sensible in Iowa. When was the last time there was a voter registration surge and turnout surge among old, lower income whites? However, it seems pretty clear that Latinos are turning out this year in Florida...

"Trump has only a -5 favorability rating in this poll (Clinton is at -9)." If you believe this, I feel sorry for you tbh.

[TheDeadFlagBlues]

their model works just fine...they just assume a low latino turnout cause those normally don't vote.

and that's a stupid assumption!