We've long had some skepticism about any perfect matchup in Oregon (the point is inapplicable in Washington and Idaho) between a decision to register as a member of a political party, and the decision about who will get that person's vote.
A useful post on Oregon Catalyst got after some of this, with some specific results from recent exist polling.
Because not all states register voters by party like Oregon does, exit pollsters ask a party identification question “Do you consider yourself a Democrat, Republican or Independent,” rather than a party registration question (“Are you registered to vote as a . . . ?”). This party identification question is probably more accurate than a party registration question when it comes to gauging actual voter sentiment. In 2010 Oregon exit polling, 36% of those who turned out said they considered themselves Democrats, 27% Republicans and 37% Independents. By contrast, the Oregon Secretary of State party registration data among those who voted was 44% Democrat, 36% Republican and 20% other.
The makeup of the 2010 voter turnout in Oregon was identical to that of 2008. In both years, 36% of those who voted were Democrats, 27% were Republicans and 37% were Independents. By contrast, the party identification breakdown nationally in 2010 was 35% Democrat, 35% Republican and 29% Independent. Further, in Oregon those who voted in 2010 said they had voted for Barack Obama in 2008 by 16% edge (52-36%), almost identical to the 17% edge Obama had over John McCain in the 2008 election in Oregon. Nationally, the recalled vote for President in 2008 among those who voted in 2010 was 45% Obama/45% McCain.
The post also notes that a lot of the result in Oregon from last month's election had to do with the nature of voter turnout. The numbers certainly seem to support that.