| Let's start our consideration of the UNH poll results with their partisan breakdowns, since that's been the place their sample is clearly inconsistent. July registrations, statewide, broke down as 30.5% Democratic, 31.1% Republican, and 38.3% undeclared. The 5.5% of the UNH first district sample and 7% of the second that's unregistered makes it difficult to assess compare directly, but in any case, what I'm most interested in is the sample from poll to poll.
Here's how that goes:
| Dem | Rep | Ind |
| July Poll 1st District | 25.5 | 27.2 | 41.3 | | April - 1st District | 26.3 | 25.9 | 41.8 | | January - 1st District | 37.1 | 26.6 | 28.1 |
| Dem | Rep | Ind |
| July Poll 2nd District | 25.2 | 28.2 | 38.2 |
| April - 2nd District | 27.5 | 21.5 | 42.9 | | January - 2nd District | 29.3 | 24.7 | 38.7 |
But the key point I've been trying to make is that the fluctuations in the sample of this poll from one to another make it problematic to directly compare them.
So from January to April, Carol Shea-Porter's net favorability dropped from +26 to +11 at the same time as Andy Smith's sample of Democrats dropped from 37% (which was way too high) to 26%. From April to July, Andy Smith's first district sample remained relatively stable, and Shea-Porter's favorability dropped further as her opponents worked hard to negatively define her, but her head-to-head match-ups with Jeb Bradley and John Stephen changed very little.
Meanwhile, in the second district, the sample had both Democrats and Republicans losing ground to undeclareds in April, and Republicans jumping considerably in July with both Democrats and undeclareds dropping. And lo and behold! Hodes dropped in his head to head match-ups with Bob Clegg and Jennifer Horn. Who ever would have thought that when you raise the Republicans in your sample by nearly 7 points, the Democratic incumbent's head-to-heads will suffer a bit?
So, with that said, take it for what it's worth: That is, somewhat more than ARG, but not a whole lot at least up until the pre-election tracking poll starts.
If you still want the numbers, here they are. I'm including April numbers in parentheses, even though the comparison is, as I've detailed, problematic:
First district MoE 6.4%
Shea-Porter 40 (39)
Bradley 46 (45)
Shea-Porter 42 (43)
Stephen 36 (35)
Second district MoE 6.3%
Hodes 44 (51)
Clegg 25 (24)
Hodes 43 (52)
Horn 23 (25)
(Edited to add MoE.)
Update: In case anyone's wondering, here are the statewide sample numbers.
| Dem | Rep | Ind |
| January | 32.9 | 25.6 | 33.6 |
| April | 26.9 | 23.7 | 42.4 |
| July | 25.3 | 27.6 | 39.9 |
I should also note, some of these fluctuations aren't at all significant. I'm just laying them all out because some are and I think it's important to have as much information as possible available. |
