2004 Delegate Race
Notes: the intent of this exercise is to determine the delegate count at the end of the large group of primaries on February 3. Totals at this point should go a long way toward determining which candidates might have the momentum and money to survive until Super Tuesday. Where possible, I've used the latest polling data in determining how the delegates are distributed. Democratic party rules dictate that candidates are distributed proportionately based on popular vote. However, this doesn't mean there's a general 1-1 mapping of polling percentage to delegate percentage. Historical votes tend to be a little more lopsided than pre-election polls would indicate. Also, states that use a caucus process (like Iowa) usually have a "floor" of 15%. This doesn't prevent candidates polling less than this number from acquiring delegates, but does mean that those candidates have to have locally heavy representation to get on the board. In states where polling is old, I've swallowed my assumptions and put the numbers up pretty much as the old polls would dictate. However, we have new polls from several states this week, so the number should reflect the current thinking.
In some few states, I have no polling information. In these cases, I've tried to extrapolate from past trends or neighboring states. For example, I made North Dakota over in the image of Iowa (sorry Dakotans).
Iowa, 45 delegates+
Gephardt 25
Dean 19
Kerry 01
New Hampshire, 22 delegates
Dean 13
Kerry 05
Clark 02
Edwards 02
Arizona, 55 Delegates
Dean 21
Clark 16
Kerry 09
Lieberman 09
Delaware, 15 delegates#
Dean 10
Kerry 05
Missouri, 74 delegates#
Gephardt 33
Dean 20
Kerry 12
Lieberman 09
New Mexico, 26 delegates+
Dean 15
Clark 11
North Dakota, 14 delegates#+
Gephardt 08
Dean 06
Oklahoma, 40 delegates#
Gephardt 18
Dean 15
Kerry 07
South Carolina, 45 delegates
Edwards 18
Lieberman 10
Dean 10
Clark 04
Gephardt 02
Sharpton 01
Predicted Total after 3 February
Dean 129 (129)
Gephardt 086 (098)
Kerry 039 (043)
Clark 033 (040)
Lieberman 028 (022)
Edwards 020 (008)
Sharpton 001 (004)
+caucus state
# based on very old polls or a lot of guesswork
#+ both of the above
Dean loses a little here, gains a little there, and holds steady. Gephardt's numbers edge up in Iowa, but go down overall mostly because of a significant decline in South Carolina. Kerry slips a bit. Clark also loses a little of that initial luster. Lieberman edges up. Edwards shows significant improvement in SC and also makes the board in New Hampshire. Sharpton drops to a single delegate, mostly because of changes in the formula for delegate allocation. For the same reason, Braun dropped from the chart entirely.
The changes this week come from two factors: changes in the polls and changes in the formula. There was additional polling in Iowa, NH, and SC that affected the numbers. However, the numbers changed everywhere due to a new formula that should better reflect how numbers in the polls translate into delegates. This formula was calculated by looking at polling information before the 1992 and 2000 races then generating a "poll weighting" factor for each state. Granted, this is not perfect as the further you go down the line the more the vote skews to reflect the percieved frontrunner, but you have to base the numbers on something. The new formula tends to somewhat flatten the exaggeration I had previously given to leaders in a state, while chopping off those hanging at the bottom of the polls. If you can imagine the old "bell-shaped curve," my previous distribution formula had tended to be heavily weighted at the front with a tear-drop fade at the back. The new curve is more gentle at the front and steeper at the rear. When running the 1992 numbers, it seems considerably more predictive.