Polling is an inexact science at best, and it’s hard to know who to trust. That said, Quinnipiac and Public Policy Polling are two of the least trustworthy, as their liberal biases skew their samples and are generally used to create the result rather than measure it.
Rasmussen is generally the most reliable, but in the end, it all comes down to who has the most accurate likely voter model, and no one knows that until after the election.
I could not disagree more about Rassmussen Reports. Scott Rassmussen has left the organization, and not on particularly good terms. Nate Silver, who like it or not has been dead on about polling issues, cited Rassmussen and PPP as polling firms to be ignored.
Rasmussen was blown out of the water in 2012. At one time, they had credibility, but seem to be badly skewed in their sampling methodologies.
The good news is only 62% of Sarvis voters are definetly voting for him with 18% likely to change their mind.