Saturday, February 9, 2008

Super Tuesday: Statistical Breakdown

Coming out of Super Tuesday, with a number of surprising results in hand, we sat down to try and make sense of the outcome. We determined to track some of the major assumptions about the voting bloc for both Hillary Clinton and Barack Obama. We did so by examining the exit polls found here.

We limited our inquiry to 6 states that we considered either "battleground" states, or where the results seemed to defy projections. The inquiry was limited to states with surprising results or "battleground" states because we believed those states to have the most relevance moving forward in the election process. These states were: California, Florida, Massachusetts, Missouri, New Mexico, and Nevada. The numbers are somewhat skewed in Florida and Nevada because those states held their primaries prior to John Edwards leaving the race. However, what we will call the "Edwards Effect" has been accounted for in the conclusion.

We divided the polling numbers into 6 categories: Age, Race, Gender, Education, Income, and Political leanings. Each category is divided into sub categories. These categories along with raw numbers about where the votes fell can be seen in this spreadsheet (you have to have a Google account, and then log into Google Spreadsheets). The available statistics varied in some states. When this happened, we either disregarded the unavailable category, or considered a different dividing line. Neither tactic seemed to skew the numbers in any significant way.

The assumptions we considered were: Age--Barack Obama carries the youth vote; Race--Black voters side heavily with Obama, while Clinton carries the Hispanic vote; Gender--Women lean heavily towards Clinton; Education--Voters with a higher level of education are drawn to Obama; Income--Wealthier voters tend to side with Obama; Political leanings--Voters who identify as "liberal" side with Obama, while "moderates" and "conservatives" are drawn towards Clinton.


1. Gender: The assumption regarding gender is correct. The male vote is split essentially evenly between Clinton and Obama, with Obama having a strong lead in only one state under consideration, while the female vote is overwhelmingly in favor of Clinton. In the four Super Tuesday states we considered, Clinton had a majority of votes in three, and in two states she had a commanding lead. She also had a majority in the states where Edwards was a factor. If anything, Edwards dropping out only strengthened her hold on the female vote.

2. Political Leanings: The assumptions about political leanings turned out mixed. Amongst "Liberal voters," Clinton and Obama are essentially even, with Clinton perhaps being slightly ahead. In the two states where Edwards was a consideration, the "Liberal vote" was heavily skewed to Clinton. Obama gained ground on Super Tuesday, winning the two smaller states in a landslide, but Clinton was still comfortably ahead in the two larger states we considered. The moderates came out heavily in Clinton's favor, both pre- and post-Edwards. The conservatives also went to Clinton. While Edwards was still in the race, Clinton and Obama drew in one state, with Clinton comfortably ahead in another. In the Super Tuesday states, Clinton and Obama drew once, with one landslide for Obama, one for Clinton, and one state going to Clinton by a comfortable margin. Clinton also again carried the larger states.

3. Income: The assumption about income was correct. Clinton easily carried the low income vote, with the exception of an essentially even split in one Super Tuesday state. Post-Edwards, her hold on the lower income vote actually strengthened. The middle income vote was closer, but still trended towards Clinton. Clinton held a lead in both states where Edwards was a consideration, with Obama gaining ground on Super Tuesday. Obama and Clinton each took two Super Tuesday states, with Clinton again taking the large states. The wealthy vote went to Obama. In the Edwards states, Clinton was well ahead of Obama, but lost ground on Super Tuesday. In the four Super Tuesday states we considered, Obama had one landslide, two comfortable wins, and Clinton claimed one comfortable victory. While the number of states won seems even, the wealthy vote is clearly trending towards Obama.

4. Education: The thesis about education is improperly formed. The prevailing assumption is that people with a higher level of education trended towards Obama. While Obama seems slightly ahead amongst well-educated voters, Clinton continued her trend of claiming the two large Super Tuesday states, and did not lose measurable ground post-Edwards. The High School educated and un-educated votes were both heavily in Clinton's favor. A more accurate statement, then, is that the two candidates split the well-educated vote, while the less educated are drawn towards Clinton.

5. Race: The assumption about race is absolutely correct. Obama won the Black vote in a landslide across the board. Clinton claimed a similar victory amongst Hispanics. The division amongst White voters is somewhat closer, but still ultimately in Clinton's favor. In the two states where Edwards was a factor, Clinton took the White vote comfortably. On Super Tuesday, Obama and Clinton each took a landslide in one of the small states we considered, while Clinton won a landslide in one of the large states, and a close victory in the other. We did not consider Asian or other races in our numbers, because only California had those statistics.

6. Age: The assumption about age is again mixed. While most assume Obama takes the youth vote comfortable, the numbers simply do not show that. In the Edwards states, Obama took one landslide, with one draw. Clinton actually gained ground on Super Tuesday, drawing both large states, with Obama claiming a landslide in one small state. The youth vote is certainly in Obama's favor, but this notion that the youth vote turns out for Obama in droves is, quite simply, flawed. Amongst the middle aged, Clinton and Obama are essentially drawn. Obama gained ground on Super Tuesday by splitting two states to two. But, as we have seen consistently, Clinton took the large states. Amongst the elderly voters, Clinton has a comfortable lead across the board--and actually gained ground with Edwards' departure.


While a good number of the commonly held assumptions seem to ring true, the largest surprise came from the age distribution. As discussed immediately above, the youth vote turns in Obama's favor, but he certainly claims no commanding control over that segment. Another small surprise came from the education levels. While the commonly held assumption is that educated voters are drawn to Obama, that assumption seems off the mark. A better way to phrase this is that the less educated are drawn towards Clinton, while the well-educated are essentially split.

The most telling element of these numbers comes from Edwards' presence. The numbers clearly show that once Edwards departed from the race his backers sided with Obama. This is unsurprising because Edwards and Obama were more closely linked, idealistically, than Edwards and Clinton. However, there was always the possibility that Clinton could have gained a big enough lead in the first few states that Edwards' backers went to her side, simply because she was the front runner. Because this scenario did not play out, we really have a compelling race ahead.

Pause for a moment, though, to think about how much different the race would look today had Edwards stayed in until Super Tuesday. After the first several states, the polls clearly showed that the nomination would come down to Obama-Clinton, but Obama trailed because the core of the party split amongst he and Edwards. This scenario played out very similarly with the Republican party. After the first several states, the Republican nomination seemed to clearly be between Mitt Romney and John McCain, with Romney slightly trailing McCain because the core of the Republican party--religious conservatives--split between he and Mike Huckabee. While Huckabee had little to no chance of winning the nomination, he stayed in the race anyway and claimed more delegates on Super Tuesday than Romney, because he carried the southern states and their fundamentalist Christian voters. Huckabee's showing on Super Tuesday so weakened Romney's campaign that he had to withdraw only a few days later, making the remaining primary season little more than a coronation party for McCain. Thankfully for Democrats, Edwards bowed out nobly before Super Tuesday. This allowed Obama and Clinton to wage an even handed campaign for the core of the party. Because Edwards did not needlessly leech votes from either Clinton or Obama, we now have anything but a drawn out Clinton coronation. We have a true race, with candidates on seemingly equal footing. It should be exciting.

No comments: