Ipsos/Reuters Polls (through DNC week): Clinton +6 (tied 4-way)

[Likely Voter]

It is hard to make apples to apples comparisons with this poll and their release earlier in the week due to overlapping time samples and the fact that they changed their methodology sometime in the last week, not sure if it was halfway through this poll or earlier.  

But with that in mind, here are the somewhat contradictory trends of the July 25-29 sample vs the July 22-26 sample.

LV (Clinton margin +6 since last poll)
Clinton 40 (+1)
Trump 35 (-5)

RV (Clinton +7)
Clinton 40 (+2)
Trump 33 (-5)

4-way:

LV (Trump +3)
Clinton 37 (-3)
Trump 37 (-)
Johnson 5 (-2)
Stein 1 (-2)


RV (Clinton +1)
Clinton 37 (-1)
Trump 34 (-2)
Johnson 8 (+1)
Stein 3 (-)

 

[Likely Voter]

OK, now Reuters says Clinton is actually up by 6 (2-way, LV)

http://www.reuters.com/article/us-usa-election-poll-idUSKCN1092M5

They really need to get their s--t together.

[Likely Voter]

I can only assume that their 2-way and 4-way polls have been two completely separate polls all along. I have changed the OP post to reflect that these are two polls and changed the title to reflect it is 'polls' and not 'poll'. We should just keep this in mind as there are new Reuters polls released in the future. Unlike other polls, you can't really compare the 2-way and 4-way and draw direct conclusions about which candidate is losing votes to the 3rd parties more or less.