It's Rush Wot Won It
In the days running up to these last primaries, Rush Limbaugh told his national audience of conservatives to vote in the Democratic race.
I want Hillary to stay in this, Laura. This is too good a soap opera. We need Barack Obama bloodied up politically, and it's obvious that the Republicans are not going to do it and don't have the stomach for it, as you probably know. We're getting all kinds of memos from the RNC, saying we're not going to be critical there. Mark McKinnon of McCain's campaign says he'll quit if they get critical over Obama. This is the presidency of the United States we're talking about. I want our party to win. I want the Democrats to lose. They're in the midst of tearing themselves apart right now. It is fascinating to watch, and it's all going to stop if Hillary loses. So, yeah, I'm asking people to cross over and, if they can stomach it—I know it's a difficult thing to do, to vote for a Clinton, but it will sustain this soap opera, and it's something I think we need. It would be fun, too.
Go and check the exit polls. In Wisconsin, Republicans made up 9 percent of the Democratic primary vote. Obama won them 72-28 over Clinton. Just as tellingly, 14 percent of primary voters said they were "conservative," and Obama won them 59-40, a bigger margin than he won with liberals or moderates. Tactical voters who said Obama stood a better chance of winning in November? They went for him 87-13.
Now, look at Ohio. Once again 9 percent of voters were Republicans, but Obama and Clinton split them evenly, 49-49. Once again, 14 percent of voters were "conservatives," and Obama and Clinton split them 48-48. (Obama did better with them than he did with liberals and moderates.) Those tactical voters who thought Obama could win gave him a 80-18 victory, a margin twelve points smaller than the margin in Wisconsin.
It's a similar story in Texas, where Limbaugh has the most listeners of any of these states. Obama won the Republican vote 52-47, but conservatives (22 percent of all voters, up from 15 percent in the Kerry-Edwards primary) went against Obama. For the first time since Super Tuesday, they were Clinton's best ideological group: She won them 53-43. And Clinton won 13 percent of the people who said Obama was the most electable candidate.
Ohio didn't wind up being very close, but Clinton won the Texas primary by about 98,000 votes out of 2.8 million cast. If the exits are right, about 252,000 of those voters were Republicans, and about 618,000 were conservatives. Clinton truly might have won the Texas primary on the backs of Rush Limbaugh listeners.
What's this mean? Psychologically it's hilarious: Every joke that's ever been told about how the right needs the Clintons to survive is true. Hillary Hatred is the gas, the ethanol, and the rocket fuel of the staggering GOP. Logistically, it might mean the end of GOP crossover voting if the Democrats get their game together and pass new primary/caucus reforms when this Ragnarok draws to a close. (In the short term I can't decide if it's better for Hillary or Obama, but it's a probably a relief to both campaigns that Pennsylvania will be Democrats-only.)
UPDATE: The Texas theory is being challenged in the comments, so I'll add one data point. In 2004, Al Sharpton got 3.7 percent of the vote in Texas. Among the Republicans who crossed over to vote in the race, he got 10 percent. He got 2 percent of liberal voters, 4 percent of moderates, and 7 percent of conservatives. There's no way to explain this unless you assume some eventual Bush voters were making mischief for the Democrats.
UPDATE II: An e-mailer to the Corner has the most persuasive evidence yet. An astounding number of voters took Democratic ballots, voted for president, then left the rest of their ballots blank.
The undercount in the D primary was almost 700,000 ballots out of 2.86 million. By contrast, the undercount in the R primary was about 164,000 ballots out of 1.38 million. In the 2004 general election, the dropoff from president to railroad commissioner (the next race on the state ballot) was less than 400,000 out of 7.4 million.
It's reasonable to assume many of these voters were "screw the Dems" Limbaugh listeners.