re: #113 Obdicut

There is no way to know why they are getting an inaccurate picture of the electorate (by party registration) without knowing their randomization algorithms and precise weighting models. What we do know is that their output appears to be out of line with the current reality. This is supported by the fact that their poll is way outside the consensus Margin of Error.

In the end, we cannot know for sure. One can argue with a lot of evidence that turnout is highly unlikely to reflect a 7-point Democratic advantage.