Hats off to Hillary Clinton for remaining in the game. She did what she needed to do for her campaign which was to stay alive, though that really is not what she needed to do to win the nomination, which now seems beyond her grasp. She still is up against almost impossible odds to win the nomination, as she is over 150 pledged delegates behind even after her wins last night. In this diary, I'll use past election data from this year, and old exit polls from Mississippi (2000, and 2004) to predict the vote and margins in Mississippi and Wyoming.
Wyoming
This will be brief, because it's a caucus state, and it's in the West, and we all know the results of those, but just in case anybody forgot, here are the results of the previous Midwest, and Mountain West caucus states:
Nebraska - Obama: 68%, Clinton: 32%
Colorado - Obama: 67%, Clinton: 32%
Idaho - Obama: 79%, Clinton: 17%
Kansas - Obama: 74%, Clinton: 26%
North Dakota - Obama: 61%, Clinton: 37%
--------------------------------------------------------
Average - Obama: 69.8%, Clinton: 28.8%
Looking at the demography of Wyoming, it's very similar to states the states I mentioned above such as Idaho. I think for this race, we will have to conservatively assume that in Wyoming, Obama will performs at least as well as his least good performance in those states, so North Dakota. If Obama does slightly better than North Dakota (since it's next to Idaho), we can conservatively assume that Obama will take about 65% of the vote there, which is still below the average for these caucus states, and according to Slate's delegate estimator, he would get 8 delegates to Clinton's 4.
Net WY prediction: Obama +4
Mississippi
My intuition tells me that this is going to be a huge win for Barack Obama, but I want to back this up with some data. First of all, the Black population of Mississippi is over 37%, according to the 2005 census. Of course, since a majority of Black voters are registered democrats, they represent a much larger portion of the Democratic vote than 37%. Luckily the Roper center at Uconn has a really nice collection of primary exit polls going back to 1976. When I looked at the numbers, I was very surprised at just how large of a proportion the Black vote is in previous Mississippi primaries. For the purposes of this, I think it makes sense to look at 2000 and 2004 exit polling data, since the demographics might have changed since then.
- N=508, 286 Black (56.2%), 183 White (36.0%)
- N=830, 530 Black (63.8%), 255 White (30.7%)
-----------------------------------------------------
Average: Black: 60%, White: 33.4%, Other: 6.6%
I think I'm being conservative here, but given that a Black candidate is running, I would assume that Black voters will turn out in at least the numbers they came out in 2004. I would suspect at higher turnout, but just to be very conservative let's get three cases of turnout Low (2000), Medium (average of 2000 and 2004), and High (2004). I'll just assume a 50-50 split amongst the (100 - (Black voters + White voters)), to come up to 100% here. Assuming these, I think we can use previous voting behavior in the deep south as a indication of how Mississippi will vote, which is largely along racial lines, with many more White voters voting for Barack Obama, than Black voters voting for Hillary Clinton. So we can look at two nearby states that recently voted: Alabama and Louisiana.
From the exit polls in Louisiana, we find that White voter split was Clinton: 58% Obama: 30%, and the Black voter split was Obama: 86%, Clinton 13%. From the exit polls in Alabama, we find that the White voter split was Clinton: 72%, Obama 25%, and the Black voter split was Obama: 84%, Clinton: 15%. So we can take the average of these, and assume that Clinton would capture: 65% of the white vote and 14% of the Black vote, and Obama would get 27.5% of the White vote, and 85% of the black vote. Using the three turnout models I explained above, the final results would be:
Low Black voter turnout: Obama: 61.5%, Clinton: 35.1%
Medium Black voter turnout: Obama: 63.4%, Clinton: 33.4%
High Black voter turnout: Obama: 65.4%, Clinton: 32.5%
If Obama won between 61% and 64%, the Slate delegate estimate is Obama: 21, Clinton: 12. If Obama won greater than 65% but less than 68%, the slate delegate estimate is Obama: 22, Clinton 11 In this case, I'm going to assume somewhere between medium to high Black voter turnout for this scenario, because that's what I think we will see there.
Net MS prediction: Obama +11
So overall, I predict Obama picking up a net of 15 delegates in the next two contests. What do you think?