After the contests on March 4th, I wondered if conservative talk show host Rush Limbaugh's urging Republicans to "cross over" and vote for Sen. Clinton had any affect on the outcome of the race. Initially, the media had said there is no real way to determine if his efforts had any effect. It seemed to me there was, so I did some analysis a week or so ago of the exit poll results (see below). I guess I wasn't the only one curious. There is a story in the Boston Globe that indicates there may have been some affect.

Here are the exit polling results for Republicans voting in Democratic contests for some states before March 4th. The percentage after the state is the percentage of the Democratic primary vote cast by self-identified Republicans.

Obama Clinton
Louisiana (5%) 53% 17%
Wisconsin (9%) 72% 28%
Virginia (8%) 72% 23%

Here are the results since March 4th.

Obama Clinton
Ohio (9%) 49% 49%
Texas Primary (9%) 53% 46%
Mississippi (12%) 25% 75%

There are two things that are particularly striking. First, the percentage of the Republican vote that Sen. Clinton received in the March 4th contests and beyond is much higher than she received before March 4th. Second, the turnout and percentage of the Republican vote Sen. Clinton received after Sen. McCain secured the Republican nomination is even higher still (i.e., when Republicans had little incentive to participate in their own primary, they came out even stronger for Sen. Clinton).

Of course, every experiment needs its control. Here are the exit poll results for self-identified independents who voted in Democratic contests.

Obama Clinton
Louisiana (12%) 53% 36%
Wisconsin (28%) 64% 33%
Virginia (22%) 69% 30%

Here are the results since March 4th.

Obama Clinton
Ohio (22%) 50% 48%
Texas Primary (25%) 49% 48%
Mississippi (17%) 53% 43%

Now this is not the perfect control for a few reasons. Foremost is that each of these states are indeed different and so you would expect some variability from state to state (although the numbers before and after March 4th are pretty stable when considered independently). Another consideration is that news events just before the March 4th contests probably boosted Sen. Clinton's standing. Nonetheless, the numbers for independents do not change as much as the Republican numbers. This smaller change is likely due to two factors. First, in previous contests Sen. Obama did not have as large an advantage over Sen. Clinton among independents as he did among Republicans so there was not as much ground to gain. Second, Mr. Limbaugh's plea had a larger affect on Republicans than on independents. Why would Mr. Limbaugh's plea have any affect on independents? Perhaps conservative-leaning independents who favor Sen. McCain over either of the two potential Democratic nominees and viewed Sen. Clinton as a more beatable opponent did indeed cross over. But it certainly seems reasonable that fewer independents would act in such a way than partisans. How much each of these factors contributed to the ultimate result is hard to say. Therefore it is also hard to say if the Limbaugh voters changed the ultimate winner of the Texas primary. What can be said is that if Sen. Obama had enjoyed the same advantages over Sen. Clinton among the Republican and independent voters in Texas as he did in previous contests, he would have won. However, with the proportional awarding of delegates the Democratic party employs, the net effect of such a win would have only been a swing of about 5 delegates.

The Texas primary result is also interesting in contrast to the Texas caucus result where Sen. Obama is winning 56% to 44% (the count has been halted but the Texas Democratic party has stated they expect Sen. Obama's margin to hold). The caucus result is complicated by two factors. First, the superior organization of the Obama campaign gives it an advantage in caucuses. Second, it is unlikely that Limbaugh primary voters would spend the time and effort to caucus for Sen. Clinton. Again, how much each of these factors contributed to Sen. Obama's win in the caucus, and ultimate win in Texas, is hard to say. Do the Texas caucus results show the organizational advantage of the Obama campaign or does it reflect the true choice of Texas Democrats?