Small differences in question wording...

Over the last week Civitas released a North Carolina poll and Quinnipiac released an Ohio poll on topics similar to ones we had polled on recently, and found favorability numbers for various politicians that were very different from the ones we got:

Politician	PPP Favorability	Civitas (NC) or Quinnipiac (OH) favorability
Rob Portman	28/23	21/6
Jennifer Brunner	34/36	34/10
Tim Ryan	26/27	12/3
Lee Fisher	40/32	33/10
Richard Burr	33/30	37/10
Heath Shuler	21/21	17/2
Roy Cooper	44/23	32/2

In every case we found a lot more respondents having an opinion of the elected officials one way or the other than the other companies did. The reason? When we asked about favorable/unfavorable we provided party labels and the other companies didn't. So with these folks for the most part little known, our ratings in many cases are as much a reflection of party identification as they are of any real attitudes toward the politicians themselves.

I don't think asking the question one way or the other is either right or wrong, but when you see big differences between numbers polling companies put out on the same topics this is just another example of how nuances in question wording can yield disparate results.