The Limbaugh Effect in Indiana and North Carolina was real, but fairly small; however, it was enough to reduce Obama's margin of victory in North Carolina, and it easily provided enough votes to give Clinton her 23,000-vote margin of victory in Indiana.
According to exit polls, in Indiana, among the 11% who described themselves as voting Republican in the past, Clinton won 53-45%, not much different from Clinton's support among Democrats. However, due to fears about possible legal action against Republicans trying to subvert the Democratic primary, it's possible that Limbaugh listeners either lied about their party identification or who they voted for.
There are other ways to try to measure this. Overall, 16% of Indiana voters called themselves conservatives, and 65% supported Clinton (among the 4% who were very conservative, Clinton won approximately 70%).
Also, in an Obama-McCain match-up, 19% of the Indiana voters would vote for McCain, and of these 87% had voted for Clinton. By contrast, in a projected Clinton-McCain match-up, 17% would vote for McCain, and only 58% of them had voted for Obama. What does this mean? Basically, about 7% of the overall voters said they voted for Clinton but support McCain no matter what. About 2% of the voters voted for Obama but support McCain against either case. Taking these numbers into account, I would guess that approximately 5% of the Indiana electorate in the Democratic primary consisted of Republicans seeking to vote against Obama to create chaos in the Democratic Party.
Openly racist voters were a bigger factor. Fully 10% of the Indiana voters in the exit polls were whites who said that race influenced their voting, and 79% of these voters supported Clinton. So approximately 8% of the voters in Indiana were whites voting against Obama because of race (of course, many of these racists may be Limbaugh listeners).
As in previous primaries, gender helped Clinton in Indiana while race hurt Obama. Among the voters who said that gender was not a factor in their vote, Clinton and Obama split the vote equally. It was women voting for Clinton because of her gender that created her likely margin of victory in Indiana.
In North Carolina exit polls, the Limbaugh effect was equally powerful, and perhaps even more so. There were more self-described conservatives (22%), and they voted 51-42% for Clinton, but among the 8% who were "very conservative," Obama actually won by a small margin. About 6% of the voters were self-described Republicans (and they supported Clinton 56-35%). However, it's odd that self-described Independents split evenly between Obama and Clinton, while Obama won Democrats easily, which is the reverse of the case in Indiana and many other states. This suggests that some of the dittoheads in North Carolina were calling themselves independent.
The Limbaugh effect becomes clear in this stat: 19% of the voters said they would support McCain in a match-up against Obama, and 82% had voted for Clinton; that means they were disgruntled voters who say they may not support the other Democrat. But of the 15% who would support McCain in a match-up against Clinton, only 45% had voted for Obama. This means that a large proportion of the anti-Clinton voters voted for her, and that's the Limbaugh effect. I would estimate that around 5% of the voters in North Carolina were Republicans who voted for Clinton despite hating her.
In North Carolina, racism was less of a factor than in other states, while sexism was a bigger factor. Nine percent of the voters were whites who said race impacted their vote, and 59% supported Clinton. That's a net of only 5-6% racist voters, less than in Indiana. By contrast, 21% of voters said gender was a factor and Obama was supported by them 52-43 (however, Obama's margin was even larger among those who said gender was not a factor). Interestingly, as a total number, more women said that gender was a factor and voted against Clinton than men who did so.
Hillary Clinton wrote in an email to her supporters, "This victory is your victory, this campaign is your campaign, and your support has been the difference between winning and losing." Let's hope she sent that message to Rush Limbaugh, who did more than anyone to give her a victory in Indiana.
Overall, the North Carolina and Indiana exit polls show that the Limbaugh effect is real and does affect the final margins. But the exit polls also show that racism and sexism are alive and well in American voting, even within the Democratic Party. Obama will be the Democratic nominee, and we will need to find a way to defeat racism along with John McCain in the November elections.