Thursday, February 7, 2008

What the race will look like on Feb. 13th, 2008

So right now there's a dearth of polling data for the upcoming contests, and previous data this political season has been spotty at the very best. Since there's not much data now, and the data we have has sucked (especially for somebody whom I'll never trust again: John Zogby), I might as well just make some up. Given this, I'm making up my own projections for the coming up race.

So below find my projections for the Feb 9th to Feb 12th races. Note these aren't based upon much data... all we have recently is the Survey USA poll out of Washington from Feb 4th 2008 showing Obama with a 13% lead over Clinton (53% to 40%, with 7% undecided). Since Survey USA has actually been one of the more reliable polls this time around, I'm going to trust them, and assume a 55% to 43% final margin in Washington. For the rest of the data I simply am using some intuition, and general knowledge about the electorate and knowledge of how the past contests have gone.

The chart below shows projected vote percentages in each state. Again, this isn't based upon much data yet, though I do plan to revisit it when more data is available. Instead, it's based upon mainly understanding the demographics of the areas.



Rationales for each state:
  • Washington: Latest poll by Survey USA showed Obama with a 13% lead over Clinton.
  • Louisiana: According to Wikipedia, the Black population Louisiana is somewhere around 34%. So this is even more than Georgia, and Georgia went for Obama 67% to 31% for Clinton. I was conservative in giving Obama 60% of the vote. In actuality, this may turn out to be a huge win, perhaps with Obama winning 65-70%.
  • Nebraska: I don't have a good sense of this state, but I think we can assume it will be like the other caucus states and go for obama heavily. I am just assuming it will be like nearby caucus states such as Kansas and North Dakota, where Obama won 74% and 61% of the vote respectively.
  • Maine: I have no idea here, but I suspect it will go to Obama. But just to be conservative, I gave Clinton the benefit of the doubt here with a 49% to 48% victory
  • Virginia: I would guess VA would go for Obama also by a thin margin given the number of college towns and the large Black population. However, the black population is not nearly as large (about 20% of the electorate) as states like Maryland, Georgia, Louisiana (all around or above 30%), so I'm just giving the benefit of the doubt to Clinton again with a 52% to 47% victory. Note: Updated to a 59% to 40% victory for Obama based upon newer polling data.
  • Maryland: Should be pretty solid for Obama given that it has very similar demographics to Georgia, and perhaps a more progressive electorate in general than some of the southern states.
  • DC: Very solid for Obama. I'm giving him an 85% to 15% victory in DC. This is because 60% of the population of DC is Black, and Obama will probably win 85% of these votes. I might have overstated the margins here, but I doubt by very much.

The below chart is a *very* naive approach to calculating the number of delegates coming out of these states. I just took the number of available pledged delegates and multiplied it by the proportion of the vote that the candidate received. Very simplistic, but I don't have the interest in analyzing each voting district.



This shows the number of delegates won overall when all is said and done in these states.



This is a stacked graph showing delegates already pledged, projected from these upcoming primaries/caucuses, and superdelegates. A note about superdelegates though: I think including these is very suspect by the TV networks, as these delegates can change their minds whenever they want. Also, as far as I know (correct me if I'm wrong) superdelegates votes will not even count if the race is decided before the convention. So even though I think it's misleading, I'm including it for the sake of no information hiding.



Below, I included the actual raw data I used.

More to come later.

Thanks a lot to Poblano from daily kos (http://poblano.dailykos.com) for the inspiring me to work on this.

-Marc

3 comments:

Unknown said...

This is a creative and detailed analysis. The question becomes- how many votes will it take to determine a winner at the convention?

Marc said...

Added rationales for each state.

Marc said...

Updated based upon recent polling data that gives Obama a 15-20% advantage in Virginia.

Blog Archive