ELECTION MODEL: SENSITIVITY ANALYSIS

[TruthIsAll]

Sensitivity Analysis (“What-if”)

A powerful feature in the Election Model is sensitivity analysis, viewing the effects of different undecided voter allocation assumptions on Kerry’s projected popular vote, electoral vote, and win probability. The base case assumption is that Kerry will get 60% of the undecided vote. But what if he does better than that? Or worse? The model analyzes Kerry’s prospects for the five cases in which he gets 50%, 55%, 60%, 67% and 75% of the undecided/other vote. It does this for both the 18-poll National model and the State Electoral Vote model.

Based on Oct. 16 data, let’s analyze the five cases.

State Model: (Win probability of at least 270 EV)

Allocation Popular Electoral Win probability

50% 50.80% 311 94.7%
55% 51.12 319 97.4

60% 51.44 327 99.0

67% 51.89 336 99.7
75% 52.40 345 100.0



National Model:(Win probability of at 50% of the popular vote)

Allocation Popular Win probability

50% 50.31% 79.4%
55% 50.65 95.9

60% 50.99 99.6

67% 51.47 100.0
75% 52.01 100.0



So, what does this tell us?

Kerry’s chances of winning are excellent even if he splits the undecided vote with Bush.

If Kerry gets 2/3, he has close to a 100% chance of winning with 336 Electoral votes.

If Kerry gets 3/4, he has virtually a 100% chance of winning with 345 Electoral votes.

http://www.geocities.com/electionmodel/