Now that the 2012 election results are certified, we thought it would be a good idea to take a look at how our polling compared to the actual results.
In the presidential race, we underestimated the Obama vote by 3.63% and overestimated Romney by 3.10%. Our effective error range was 3.2%, so not bad over all. This minor skew isn't terribly surprising, considering the significant lack of polling results from the "under 30" crowd. I expect this bias to matter less for the relatively minor elections over the next 3 years, as the turnout of that group will also be significantly depressed.
The governor's race turned out slightly differently. The projection for Rob McKenna was more or less spot on. Give all the undecided and "Other" votes to Jay Inslee, and you have the election. Obviously that isn't relevant for the poll. Unlike the president, where we underestimated the spread by 6%, we only underestimated the governor race by 3.5%. All the results here are within our margin of error. If this election wasn't held in conjunction with the presidential election, I suspect the outcome might have been different.
As we move down the ballot, we notice the number of under or bad votes is rapidly increasing. By attorney general, 7.8% of the ballots didn't have one of the two candidates marked. On Attorney General We projected a 5.34% spread, the actual number was 6.96%. This suggests it is largely that younger, liberal crowd that fails to complete their ballot to the end.
|Poll||Actual||Delta||Attorney General Candidate|
|46.74%||53.48%||-6.74%||Bob Ferguson, Prefers Democratic Party|
|42.00%||46.52%||-4.52%||Reagan Dunn, Prefers Republican Party|
I wouldn't want to spoil everything right away, so I'll save the Referendum and Initiative results for later!