In the weeks leading up to the national election, we paid very close attention to the statistical analyses and polling data assessments updated daily by Nate Silver and FiveThirtyEight blog. Silver became a star with his predictions for the 2008 election cycle; he hit every mark.
FiveThirtyEight was extraordinarily accurate in tracking the probabilities leading to the Obama victory. When the Miami Herald put a top of the fold, front page story of a late Mason Dixon poll showing Romney with a 6 point advantage, I quickly turned to FiveThirtyEight where that poll was an extraordinary anomaly.
It was another dismal moment for the Herald, that otherwise endorsed Obama. Nate Silver sets a very high bar for data analysis in American politics.
FiveThirtyEight was extraordinarily accurate in tracking the probabilities leading to the Obama victory. When the Miami Herald put a top of the fold, front page story of a late Mason Dixon poll showing Romney with a 6 point advantage, I quickly turned to FiveThirtyEight where that poll was an extraordinary anomaly.
It was another dismal moment for the Herald, that otherwise endorsed Obama. Nate Silver sets a very high bar for data analysis in American politics.
3 comments:
Was the Miami Herald's choice of hiring Mason-Dixon polling a mistake or a strategy? Was the decision to put up a front page story on the (skewed) results predicting a Romney landslide a mistake or a strategy? Was the fact the story didn't even mention the existence of more than a dozen contradictory polls a mistake or a strategy? Is the fact that even today the Miami Heralds editorial lamenting the miserable voting process doesn't hold local republican officials (Gimenez) and state officials (Scott and Republican leaders) accountable and call for an investigation still just a mistake or a strategy? Just asking.
All very reasonable questions.
Pleeeeease stop referring to what Mr. Silver does as "predictions".
They are NOT predictions. They are probabilities. Mr. Silver is not "right" because Obama won. He would not have been "wrong" had Romney won.
The probability of a truly random die rolling a 3 is 1/6 or about 17%. However, that probability is not "wrong" when the die rolls a 3.
Like Mr. gimeleteye, I checked out the 538 blog and ignored everything else. :)
Post a Comment