The Daily Kos pollster conducted a poll of Virginia voters last month and found Barack Obama leading Governor Palin by a decent margin in a state that he won by around six points in 2008. What most people found questionable about the poll was the 50% approval rating that it showed for Obama in the state. Considering that Obama’s average national approval rating is nowhere near 50%, it’s extremely unlikely that his approval rating would be at 50% in Virginia as the state is not more liberal than the nation as a whole.
In any event, Roanoke College released a poll today that casts doubt on the Virginia poll that the Daily Kos pollster released last month. It found Obama’s approval rating in Virginia at 36% rather than the 50% approval that the Daily Kos pollster found.
What would account for such a huge disparity between pollsters? The two pollsters found two dramatically different ideological samples. The Daily Kos pollster used a sample comprised of 44% moderates and 37% conservatives. The Roanoke College poll used a sample comprised of 45% conservatives and 35% moderates. Keep in mind too that the Roanoke College poll was of Virginia residents, not registered or likely voters. A poll of residents, as opposed to actual registered or likely voters, typically will not produce a sample that is too favorable to the Republican.
If we plug in the numbers that Daily Kos pollster found in the Obama/Palin matchup among liberals/moderates/conservatives (84/13, 72/21, 10/78, respectively) but use the liberal/moderate/conservative ideological sample (17, 35, 45, respectively) found by Roanoke College, we effectively have a tie between the two. There’s reason to believe that the ideological composition that Roanoke College found for the Virginia electorate is correct considering the conservative shift we have seen in the electorate over the past two years.