by Metsfanmax on Thu Nov 08, 2012 10:36 am
Night Strike wrote:Metsfanmax wrote:The answer is yes. I saw absurd claims by conservative pundits saying Romney would get as many as 320 electoral votes. These people fail at basic statistics and should stop making predictions.
And if Romney had done ~2 percent better in Florida, Virginia, Ohio, and 1 or 2 other states, he would have been over 300 electoral votes. It wasn't like Obama won these states by as large of margins as in 2008; he squeaked them out and the electoral college made his victory look more decisive than the reality of it.
OK, but he didn't, and the people who understood the statistics made clear that doing 2 percent better in
enough states to win was unlikely. Doing 2 percent better in Florida? If that was all Romney needed to win, they would have been more justified. But doing better than polls predicted in 4-8 swing states? Statistically, that's an outlier. I don't blame conservatives for saying Romney had a
chance of winning; I blame conservatives like George Will, who actually predicted
on Monday that Romney would get 321 electoral votes. There's no reasoning in that. Just punditry.
Furthermore, most of the conservative pundits believed the 2008 demographics were outliers, especially since those demographics returned to past levels in the 2010 election, so they had based their predictions on averaging the past several elections together, accounting for the big change in 2008 but not giving it heavy weight. It turned out that 2008 wasn't an outlier, which is why Obama narrowly won.
The problem with the pundits is that time and again, their gut feelings on how things will turn out end up being proven for what they are, which is just a hunch with no scientific basis behind it. They don't get it right more than half the time.