I have more confidence in the data and statistics at
http://fivethirtyeight.com/ .
To me, here's the troubling thing about the partisan conservative polls -- if they use poor data collection and analytic techniques, and people believe them (that the conservative candidate(s) is/are ahead) ... then if the conservatives lose, people can believe the loss was due to the election being rigged (when in fact the conservative message simply failed to get the most votes). That's a potential consequence of sloppy or slanted analyses.
The last thing the Republican party needs is to go into denial and claim that they have a winning message, if/when the party actually needs to change some to capture more of the middle in order to become competitive at the national level again.