Bloomberg released a poll today (full pdf below) that is markedly different from their traditional polling methodology.   The poll is a representative sample of 802 random Ohio ‘likely voters’ with no additional weighting, smoothing or adjustments by the pollster, it’s just raw voter data for the sample.  The result:

What makes this specific media poll interesting is that media polls rarely (can’t remember any) present raw unfiltered likely voter data without applying their own weights, and modifying the results to fit the pollsters’ assumptions of who will show up on election day.
This poll doesn’t weight anything or assume/project a specific turnout ratio.

From the poll […] “Party breakdown for the poll was 33 percent Republican, 29 percent Democrats, and 34 percent independents. Exit polling shows that Ohio’s electorate in the 2012 presidential election was 38 percent Democratic, 31 percent Republican, and 31 percent independent, while in 2004 it was 40 percent Republican, 35 percent Democratic, and 25 percent independent.” (link)

All other media polls have taken the raw response data and reconfigured it, weighted it, to mirror the 2012 turnout: 38(D), 31(R) and 31(I); which, as we have discussed at great length, is the critical flaw in virtually all of this year’s media polling.
Common sense will tell you, based on primary results and other metrics, the 2016 electorate turning out to support Donald Trump (or oppose Clinton) is not the same 2012 electorate who turned out to support Romney (or oppose Obama).
This Bloomberg Poll refreshingly doesn’t try to make 2016 equivocal to 2012 in their poll presentation.  It is also quite stunning they didn’t.
Unfortunately, they don’t provide the crosstabs so we must rely on the article.  From the article –  The Ohio demographic groups where Trump has the biggest edge over Clinton are white middle-class men (+43 percentage points), white men overall (+27 percentage points), and white middle-class women (+23 percentage points).  [Sounds like the typical Monster Voter sans the race]
If every poll just presented raw likely voter data in this manner, you would be hard pressed to find a single national or state poll showing Hillary Clinton with any lead (or even close).
Perhaps now that Bloomberg has broken the ice, this will trigger other media polling outlets to stop weighting their poll results based on 2012 models that are not comparable.
Yeah, I know… we won’t hold our breath either.
Here’s the article, and Here’s the data:
[scribd id=324026766 key=key-f9E3MIfHHAZpgS5QJjkX mode=scroll]