[usafa92]

Huffington Post tries to discredit the USC/LA Times poll and in doing so let's us in on an important tidbit. The poll is weighted to the 2012 electorate. Many have pointed out that Gallup's voter ID affiliation is showing an even split amongst Democrats and Independents. Actually if you look at 2012 and 2008, the Gallup partisan id the week of the election almost exactly mirrored the final split on exit polls. So, an argument can be made that Trump is actually up 5 or so points in this poll of the current affiliation ratings are used.

Lastly, Kellyanne Conway had a tweet Sunday that referenced the USC poll and she mentioned that this was the state of the race today. My takeaway from that is that Trump's internal polling is pretty much in line with the USC poll. Maybe better if they have more insights into the partisan if split.

[usafa92]

FYI

[usafa92]

That should read even split between Democrats and Republicans.

[Albion Wilde]

Ping

[Robert DeLong]

but their actual value of Trump up by 0.9 percentage points could be way off

It most definitely is. His lead is much larger.

[mac_truck]

I’d say just participating in the tracking poll exponentially increases the chances these people will actually vote their preference in November.

Not sure you can say the same thing about folks wrestling with their choices at home.

[God luvs America]

Florida, Ohio as well as Pennsylvania/Michigan /Virginia /Wisconsin all that matters.

[NYRepublican72]

Of course the left-leaning Huffington Post is trying this tact. The USC/LA Times Poll is based on the RAND Continuous 2012 Presidential Election Poll, which accurately predicted the 2012 margin within .7% of the final margin between Romney and Obama in that period.

That was one of the closest spreads of the cycle.

The left doesn’t like a result that matches the carefully weighed results like Monmouth and Marist. Those polls assume a certain race/gender/age turnout, which may or may not actually be where the electorate is come November.

Monmouth had such a terrible Ohio sample on Monday that they had to adjust race/gender/age to such an extent that a R+4 turnout from raw data turned into a D+4 sample.