The counting is now close enough to complete to determine which public pollsters got it right and who was wrong regarding the Rossi-Murray Senate race.
Picking the winner doesn't make a poll accurate. What matters is whether or not the actual results fall within your poll’s margin of error (MOE). The MOE differs, depending upon how the poll was done.
Sen. Patty Murray won by 4.6 percent. These pollsters were within the MOE with their final poll, and, therefore, got the race "right."
UW's Washington Poll: predicted Murray +6, difference between predicted and actual, 1.4 percent; MOE 4.3 percent.
FOX/Rasmussen: predicted Murray +2; difference 2.6 percent; MOE 3 percent.
CNN: predicted Murray +8; difference 3.4 percent; MOE 3.5 percent.
Marist: predicted Murray +1; difference 3.6 percent, MOE 4 percent.
And the final polls from these three firms were outside the MOE and, therefore, "wrong."
SurveyUSA: final had the race even; difference 4.6 percent; MOE 3.8 percent.
Rasmussen: predicted Rossi +1; difference 5.6 percent; MOE 4 percent.
PPP: predicted Rossi +2; difference 6.6 percent; MOE 2.2 percent.
And finally, the Elway poll was released three weeks before the election, too far out to be considered a final poll. Still, that poll predicted Murray +15, for a difference from the actual results of 10.4 percet, with an MOE of 4.6 percent.
A tip of the hat to Matt Baretto and his team at the UW. At the time, I questioned his method for determining likely voters, but he came the closest to predicting the Democratic spike in turnout that cancelled out the Republicans and the "enthusiasm gap" the GOP had opened up.
