More Weird Survey USA polls--NC & OH Senate races

[WI_DEM]

I'm not sure I trust Survey USA this cycle. They are worse than Rassmussen and some of their findings are just plain weird such as their poll in Washington Senate--where Rossi was up by 7 and tied with Murray in Seattle--an area she got 59% in the open primary (the respected Elway poll of WA came out yesterday and gave Murray a 9-point lead including 61% in Seattle) and then they have consistently shown younger voters as favoring the GOP--they may not vote, but I doubt they have suddenly turned into republicans. Anyways here are a couple of their new polls--NC is particularly brutal:

North Carolina

2010 Senate
58% Burr (R), 34% Marshall (D), 6% Beitler (L) (chart)

In looking at the demos more weirdness they have Burr getting 26%! of the black vote--I don't think so and tied with Marshall in the Raleigh/Durham area?

SurveyUSA
9/10-13/10; 641 likely voters, 3.9% margin of error
Mode: Automated phone
(SurveyUSA release

2010 Senate
49% Portman (R), 40% Fisher (D) (chart)

Weird stuff here too with Portman getting 23% of the black vote, running nearly even in Cleveland and with double digit leads in Dayton and Toledo??

www.pollster.com

[Alcibiades]

1. I'm dubious about these NC results

There's simply no way Burr is doing that well among black voters, or here in the Triangle.