If one story is more important in the world of polling than all others this year it is the proliferation of surveys using non-traditional methodologies, such as surveys conducted over the Internet and automated polls that use a recorded voice rather than a live interviewer (Interactive Voice Response, or IVR). Of the over 2100 polls we collected in 2006, 34% were conducted by IVR and 19% were conducted over the Internet. Less than half, 47%, were conventional telephone polls with a live interviewer. In this paper we examine the properties of these new methodologies in comparison with conventional polls, using the dynamics of the campaign, the variability of polls (sampling plus non-sampling error), and the eventual outcomes as crucial benchmarks. While the "final poll" performance has been the focus of considerable research on election polling, we use the entire election cycle to also compare how the various methodologies captured campaign dynamics: did they track together or are some flat while others reveal substantial trends? We also estimate the variability of results about the trends (with house effects estimated and accounted for) among the methods. And we consider a final accuracy estimate for each method. The results are important for how we understand these new methods which by shear numbers now account for more than half the polling information released to the public. Are the economies of Internet and IVR polls justified by performance comparable to much more expensive conventional polling? Or are such polls subject to substantial differences which may produce misleading interpretations of elections?