The Huffington Post has an interesting piece by Mark Blumenthal about how the major polling firms did this year: Mark Blumenthal HuffPo article on polling The National Council on Public Polls released their report this week. Here is their basic conclusion:
"Generally speaking, the national and state polls this year did okay," NCPP president Evans Witt told The Huffington Post, though "they didn't come quite as close to matching the election results as they had, for example, in 2008."
The average error for final week estimates of support for each candidate as calculated by NCPP was slightly higher in national polls this year (1.4 percentage points) than in 2008 (0.9), 2004 (0.9) or 2000 (1.1), but was lower than in 1996 (2.1), 1984 (2.4) or 1980 (3.1).
The article goes on discuss the details, but my takeaway below the squiggle.
1. The polls were pretty accurate again. As witnessed by Nate Silver and all the sober minded analysts, Obama was never truly behind in the polls if analyzed state by state.
2. The National polls reflected more volatility largely because of our polarized country (witness the 'fiscal cliff' negotiations). In the non-swing states, huge majorities on one side or the other could show great movement after key election moments (the debates, the 47% remark etc.), but not alter the basic result.
3. As one report after the Florida debacle in 2000 made clear, you can never achieve perfection in elections. We have too many voters and localities to expect everything to go perfectly. The same goes with polling. There are simply too many variables to expect absolute accuracy.
4. Still, the polling firms who read the electorate most accurately were those who didn't try to over-interpret who was a "likely voter". PPP and Pew in particular were terrific in letting their interviews do the talking. Gallup and Rasmussen (not to mention the internals in the Romney campaign) seemed bent on figuring out the composition of the electorate FIRST and then laying that screen over their polling. For Gallup, Ras etc. to consistently show a Romney lead when 90% of the other polls didn't was an absurd display of hubris - "We don't care who the public says they are going to vote for - we know better!" seemed to be their attitude.
5. In the end, the polls in 2012 weren't bad. If a few notable polling firms had just taken the voters at their word that they were going to show up on election day, the polling average would have been close to predicting a 3% Obama victory (it stands at 3.7% right now). And, if the Media (particularly the analysts on the right) had read the polls fairly, there wouldn't be this public perception that the polls did worse than they really did considering the defiant outliers.