Here's another thing for me to file under "I wanted to do something like this, but now I don't have to because someone else did it."
Washington Post reporter Philip Bump has a short but sweet post on The Fix that includes this neat little graphic:
A week before Election Day, I posited on the accuracy of the polls in key Senate races and compared candidates' poll positions in Senate contests in the same states four years ago with the 2010 election results. I found that, in every U.S. Senate race except the one in Colorado, the candidate who was ahead in the RealClearPolitics rolling average of polls won, but in many states, the polls greatly exaggerated or undertold the winner's eventual margin of victory. By comparison, Bump's analysis shows that, this year, the polls were more skewed in one direction, although - with the exception of North Carolina - they were correct about who would win.
The most important line from Bump's piece was its last sentence: "In this case, the polls largely predicted the correct winners, but -- with select exceptions such as the Des Moines Register poll last weekend that showed Joni Ernst (R) ahead by as wide a margin as occurred -- the scale of the victories was way off." This is what I suggested could be the case in my October 28 post.