This is nonsense. As Nate explained, polls can have errors +/- 3 points, and those errors tend to be correlated across polsters, time, and space. If something h

This is nonsense. As Nate explained, polls can have errors +/- 3 points, and those errors tend to be correlated across polsters, time, and space. If something happens 1/3 or 1/4 the time when sampling from those polls, it cannot be considered shocking when it happens in real life.