FL - Bloomberg - Trump +2 (4 way)

[JerryArkansas]

Another point from Steve Schale:

You must be logged in to read this quote.

Bloomberg's likely voter model sucks.

They're (well, Seltzer is) generally one of the most accurate pollsters.  It's tough to argue with past results, even if such results don't guarantee future performance.  Take a look at their Iowa Senate polls in 2014 that everyone else thought were outliers.

When the poll had Trump getting 44% of Hispanics, the poll is just bad.

[JerryArkansas]

When the poll had Trump getting 44% of Hispanics, the poll is just bad.

Trump gets 36% vs 51% in two-way according to this poll. With MOE of 7%

Read wrong number, but the principle still stands. 

[JerryArkansas]

When the poll had Trump getting 44% of Hispanics, the poll is just bad.

Yes, let's unskew the polls.  That always turns out really well.

I'll offer an alternative explanation -- Floridian Cuban Hispanics have typically voted Republican in greater percentages than Hispanics from other regions.  That hasn't been the case this cycle because, quite frankly, Trump is toxic to ethnic minorities.  However, it's possible that this group of voters is coming "home", either as a result of Rubio's candidacy or otherwise.

Or something more obvious the poll is an outlier.

[JerryArkansas]

When the poll had Trump getting 44% of Hispanics, the poll is just bad.

Yes, let's unskew the polls.  That always turns out really well.

I'll offer an alternative explanation -- Floridian Cuban Hispanics have typically voted Republican in greater percentages than Hispanics from other regions.  That hasn't been the case this cycle because, quite frankly, Trump is toxic to ethnic minorities.  However, it's possible that this group of voters is coming "home", either as a result of Rubio's candidacy or otherwise.

Or something more obvious the poll is an outlier.

Also possible, for sure.  I'll just say, though, that the Atlas population tends to describe all polls highly unfavorable to their preferred candidate as "outliers".  That isn't always the case -- again, look at Seltzer's final Ernst +7 poll in 2014 when the other pollsters were herding around +3ish.  In fact, Ernst won by more than 8.

When the average has her up by three, and this comes out 5 points to the right, it seems to look like an outlier to me.  Something that far off should be treated with salt by all.
And Selzier does poll white people very well, when it comes to other groups, they seem to throw sh**t at the wall.