Political scientists have developed all sorts of questionnaires to measure political knowledge. Some questions focus on the basic structure of the government -- how many Supreme Court justices are there, who runs the House of Representatives. Others ask voters to name their own Congressman and Senators, or even state legislators (which I can't do). Other focus on current political issues, questions like what TARP was or whether budget deficit is going up or down. These surveys repeatedly show that the voters who know the most about the government are often wrong about current hot-button issues:
In 2006, the political scientists Christopher Achens and Larry Bartels presented a paper titled "It Feels Like We're Thinking: The Rationalizing Voter and Electoral Democracy." In it, Achens and Bartels make a point that is so obvious we often forget its implications: "Very few politically consequential facts are subject to direct, personal verification."This effect becomes strongest when you look at conspiracy theorists; people who think JFK was assassinated by Martians know many more facts about the assassination than people who think Oswald probably shot him.
In other words, an informed voter rarely knows anything firsthand, the way we know the sky is blue and the sun rose this morning. Everything she knows is taken on trust; an informed voter is only as good as her information sources. And because we all get to choose which information sources to believe, voters with more information are not always more informed. Sometimes, they're just more completely and profoundly misled.
Looking at the 1996 election, for instance, Achens and Bartels studied whether voters knew the budget deficit had dropped during President Clinton's first term (it had, and sharply). What they found will shake anyone who believes more information leads to a smarter electorate: how much voters knew about politics mattered less than which party they supported. Republicans in the 80th percentile of political knowledge were less likely to answer the question correctly than Democrats in the 20th percentile of political knowledge. . . .
Similar experiments have shown similar self-deception among Democrats when the questions favor Republican ideas or politicians. Achens and Bartels's conclusion is grim: much of what looks like learning in American politics is actually, they argue, an elaborate performance of justifying the beliefs we already hold. "Most of the time, the voters are merely reaffirming their partisan and group identities at the polls. They do not reason very much or very often. What they do is rationalize."
And this is why Democrats (like Hillary Clinton this year) have been pushing so hard to get more poor and minority people to vote, and Republicans keep trying to stop them. Because how people vote is very largely determined by who they are, not what they know.