|
|
|
Regarding the Huffington Post chart above, there’s THIS:
Given the torrent of incoming data, the model now reports nearly complete certainty about Obama’s narrow leads in the most crucial tipping-point states of Iowa and Ohio, but that statistical confidence assumes that the final polls are collectively accurate and unbiased. When we factor in the historical potential for polling error, the probability of an Obama win falls to roughly 90 percent in Ohio and Iowa. An Obama win in those states is thus still very likely, but a 1-in-10 possibility still exists, given the typical historical pattern, that the polls ...
Read more
