Topic: 2008 election
Rush Limbaugh, being the enormous taint that he is, encouraged Republicans to vote for Hillary Clinton last night in Indiana in order to "create chaos". Did it work? Here are the numbers, according to MSNBC's exit polls.
In Indiana, 10 percent of voters in the Democratic primary were Republicans. Of those, 54 percent went for Clinton and 46 percent went for Obama. But it would be assuming too much to conclude that the "Limbaugh effect" is therefore eight percent; we need to compare these numbers to numbers from similar states.
So, first we need to determine what counts as a "similar state." I'll use the following criteria:
- The state must have an open primary (both Republicans and Democrats are allowed to vote);
- The state must have voted AFTER it became a two-person race (i.e., February 5 and after);
- No home states (New York, Illinois, Arkansas, and Hawaii are thrown out);
- Data must be available.
Using these criteria, we narrow it down to seven states: Alabama, Missouri, Virginia, Wisconsin, Ohio, Texas, and Mississippi (data was not available for many states). At first glance, this sample may appear to be more pro-Obama than the nation as a whole; he wins these seven states by an average of 54.1 to 44.4, whereas his lead nationally (not counting Michigan and Florida) is 49.6 to 47.4. However, six of these seven states have substantial black populations, which Obama won by about 90-10 in each case. This distorts the picture a little, because the Republican vote in these states was almost surely overwhelmingly white (only 6.2 percent of black voters are registered as Republicans). So, if we look only at the non-black vote in these seven states, we get a number that is much more favorable to Clinton: 59.3 for Clinton to 38.4 for Obama.
The point of the previoius paragraph is simply this: this sample is fairly representative of the nation, and if anything it skews pro-Clinton. It is, I think, a decent proxy for Indiana. And here's how the Republican vote broke down in these seven states:
Alabama: 5% Republican, Clinton 72%, Obama 25%
Mississippi: 12% Republican, Clinton 75%, Obama 75%
Missouri: 6% Republican, Clinton 21%, Obama 75%
Ohio: 10% Republican, Clinton 49%, Obama 49%
Texas: 9% Republican, Clinton 46%, Obama 53%
Virginia: 7% Republican, Clinton 23%, Obama 72%
Wisconsin: 9% Republican, Clinton 28%, Obama 72%
Average: 8% Republican, Clinton 42%, Obama 56%
So, Clinton went from averaging a 14 point loss among Republicans to a sudden 8 point win. Could the Limbaugh effect be responsible for this 22 point swing among Republicans? Actually, that sounds about right when you consider this: the Republican turnout was 2 percent higher in Indiana than in our sample. 2 percent of the total vote in Indiana is about 25,300 votes. And 22 percent of the Republican vote is an almost identical number: 27,830. So, by comparing the numbers in Indiana to the numbers in the national sample, I would estimate that somewhere around 26,000 of Clinton's votes in Indiana last night were Limbaugh supporters trying to prolong the Democratic primary.
Amazingly, those votes change the outcome. If we subtract 26,000 votes from Clinton's total, Indiana switches from a narrow win for Clinton to a narrow win for Obama (about 623,000 votes for Obama and about 616,000 for Clinton - 50.3 to 49.7).
I hate to give Rush Limbaugh credit for anything, but he may have actually affected the outcome in Indiana. Not that he'll affect the outcome of the primary as a whole; Obama has been pretty much a statistical lock for more than a month now. And if he wants the Republicans to keep the White House in November, he'll have to come up with a scheme to elect a man whom supports an unpopular war, untenable fiscal policies, and an irresponsible energy policy. It will take a lot more than 26,000 votes to do that.