What happened on Super Tuesday.

10:00 p.m.: Two weeks ago, I pointed out that John Edwards had beaten John Kerry among independents in four of the last eight primaries in which that number had been measured. Worse yet, Edwards had beaten Kerry among crossover Republicans in the last six primaries in which that number had been measured. The implication was that Edwards might be more electable than Kerry.

How did that theory hold up in today's primaries? Let's look at the numbers. To be extra fair to Edwards, I'll focus on the five exit-polled states in which he campaigned. Here's how he did vis-à-vis Kerry among independents and crossover Republicans:

Not exactly a compelling stat line for Edwards. He did beat Kerry among crossover Republicans in three of the four states in which that number was measured. But Kerry beat him convincingly among crossover Republicans in California—the country's most populous state—breaking Edwards' unbeaten streak in that category. More important, Kerry beat Edwards among independents in three of the five contested exit-polled states. (Kerry's win in Maryland was narrow, but so was Edwards' win in Ohio.) That's consistent with Kerry's previous record. He beat Edwards among independents in six of the first 10 exit-polled contests.

I'm not saying these polls perfectly reflected the preferences of most independents. In Maryland, for example, registered independents like me couldn't vote in the Democratic primary. But to substantiate the argument that he was more electable, Edwards needed to beat Kerry among self-described independents in most of the states he targeted today. He didn't.

The numbers only get worse for Edwards when you factor in Connecticut, Massachusetts, and Rhode Island, the three states in which Edwards was on the ballot but didn't campaign. Kerry won independents in all three states. (Republicans couldn't vote.)

Worse yet for Edwards, Kerry won "agrees with you" voters in four of today's five key states. (Edwards won them in Georgia.) These are the voters who say they chose their candidate because he "agrees with you on the issues," not because he "can defeat Bush." They aren't voting on electability. They're voting for the guy they want. Before today, they had voted for Edwards in four of the last eight states. Today, they went for Kerry.

I've heard a lot of explanations for why Edwards, despite his talents, couldn't win more than one primary this year. But the thing that stands out to me is a question at the tail end of every exit poll: "Which one candidate quality mattered most in deciding how you voted today?" In Georgia, among voters who chose "cares about me," "honest and trustworthy," or "positive message," Edwards creamed Kerry. Among those who chose "stands up for what he believes," the candidates tied. Among those who chose "can defeat Bush," Kerry crushed Edwards. But if most of the folks who voted in Georgia today preferred Edwards on the other criteria, why did they think Kerry was more likely to defeat Bush? The answer, I think, lies in the final category: voters who chose their candidate because he had the "right experience." Those voters went for Kerry by a devastating margin of 80-14.

This was the landslide number in every state. In Ohio, Edwards beat Kerry among "cares about me," "honest and trustworthy," and "positive message" voters. Kerry won "stands up for what he believes" voters and demolished Edwards 71-19 among "right experience" voters. In New York, Edwards ran close to Kerry in most categories but lost "right experience" voters 82-7. In Maryland, the Kerry margin in that group was 83-8. In California, it was 81-6.

True, in all of these states, the "right experience" vote was much smaller than the "can defeat Bush" vote. But by process of elimination, it looks to me as though the first judgment was the basis for the second.