Run your own Excel Interactive Election Model Simulation

[Michael Collins]

This will get around.

[All_The_States]

Can this model be somehow used to check only one specific demographic? There's one demographic I'd like to check in the whole analyzer whatcha-call it and I'm wondering how to do it.

[TruthIsAll]

Input the gender demographic for the 12:22am state exit polls.
These are the real exit polls.
Run the simulation.
Check out the EV and popular vote.

Next...
Input the adjusted final state exit poll gender numbers, or the actual vote percentages (same thing, they forced a match).
These are the not-real exit polls.

Note: The model weights each state poll by its percentage of total voting population, using data from the three elections prior to 2004.

[TruthIsAll]

I have a problem with my home machine and didn't see it until I got to work today.

[TruthIsAll]

Assume the final pre-election polls were correct.
Hint:
Lower Kerry's undecided voter % allocation until the Bush 2-party vote is equal to the final recorded 51.24%.

[TruthIsAll]

A Question Of Letitimacy: Who Really Won In 2004?
Tuesday, 20 September 2005, 1:59 pm
Press Release: Concerned Americans
September 19, 2005
For Immediate Release
From: Concerned Americans

IT'S NOW A QUESTION OF LETITIMACY: WHO REALLY WON IN 2004?

[TruthIsAll]

Just select:
1 for pre-election
2 for exit polls
3 for final

Press F9 (calc) and 200 scenarios will be executed...

The actual state polling and final vote percentages will automatically be entered into the Kerry and Bush columns.

The margin of error will be inserted as well: 4% for state pre-election polls and the MoE as calculated based on the number of exit poll respondents.

If you enter your own polling data, it will override the reference/formula links to the cells which contain the actual data and you will not be able to refresh the cells with the original data, so first save a copy of the original file.

http://www.geocities.com/electionmod...Simulation.xls

[TruthIsAll]

MONTE CARLO ELECTION SIMULATION
by TruthIsAll

Kerry Bush Input Data in Blue cells only for:
Trial Wins 196 4 Kerry % undecided voter:67%
Win Prob 98.0% 2.00% Kerry / Bush poll %
Por Vote% 51.32 48.68
Vote (mm) 62.10 58.90

Analyze: 1
ELECTORAL VOTES 1 Final Pre-election Polls
Mean 323 215 2 Exit Poll (12:22am)
Median 325 213 3 Final Recorded Vote
Maximum 378 289
Minimum 249 160 PRESS F9 TO CALCULATE

Kerry Bush Proj% Prob Poll Exit WinPct
EV Vote% Vote% Kerry Win MOE MOE 98.0%

AL 9 39.00 57.00 41.68 0.0% 4.00% 3.57% 0% 0
AK 3 30.00 57.00 38.71 0.0% 4.00% 3.18% 0% 0
AZ 10 45.00 50.00 48.35 20.9% 4.00% 2.27% 24% 47
AR 6 46.00 48.00 50.02 50.4% 4.00% 2.61% 57% 113
CA 55 49.00 42.00 55.03 99.3% 4.00% 2.22% 100% 199

CO 9 47.00 48.00 50.35 56.8% 4.00% 1.95% 57% 114
CT 7 52.00 42.00 56.02 99.8% 4.00% 3.27% 100% 200
DE 3 45.00 38.00 56.39 99.9% 4.00% 3.48% 100% 200
DC 3 78.00 11.00 85.37 100.0% 4.00% 1.92% 100% 200
FL 27 47.00 47.00 51.02 69.1% 4.00% 1.84% 66% 131

GA 15 42.00 52.00 46.02 2.6% 4.00% 2.48% 3% 6
HI 4 45.00 45.00 51.70 79.8% 4.00% 4.38% 80% 159
ID 4 30.00 59.00 37.37 0.0% 4.00% 3.91% 0% 0
IL 21 54.00 42.00 56.68 99.9% 4.00% 2.60% 100% 200
IN 11 39.00 58.00 41.01 0.0% 4.00% 3.17% 0% 0

IA 7 50.00 44.00 54.02 97.6% 4.00% 1.96% 97% 193
KS 6 37.00 60.00 39.01 0.0% 4.00% 3.65% 0% 0
KY 8 39.00 56.00 42.35 0.0% 4.00% 3.00% 0% 0
LA 9 40.00 48.00 48.04 16.8% 4.00% 2.38% 15% 29
ME 4 50.00 39.00 57.37 100.0% 4.00% 2.20% 100% 200

MD 10 54.00 43.00 56.01 99.8% 4.00% 3.07% 100% 199
MA 12 64.00 27.00 70.03 100.0% 4.00% 3.10% 100% 200
MI 17 52.00 45.00 54.01 97.5% 4.00% 1.98% 95% 189
MN 10 52.00 44.00 54.68 98.9% 4.00% 2.09% 100% 200
MS 6 42.00 51.00 46.69 5.2% 4.00% 3.44% 3% 5

MO 11 44.00 49.00 48.69 26.0% 4.00% 2.11% 31% 61
MT 3 36.00 57.00 40.69 0.0% 4.00% 3.78% 0% 0
NE 5 32.00 61.00 36.69 0.0% 4.00% 3.37% 0% 0
NV 5 49.00 49.00 50.34 56.6% 4.00% 2.13% 56% 112
NH 4 47.00 47.00 51.02 69.1% 4.00% 2.27% 67% 134

NJ 15 50.00 42.00 55.36 99.6% 4.00% 2.49% 99% 198
NM 5 49.00 49.00 50.34 56.6% 4.00% 2.22% 55% 109
NY 31 57.00 39.00 59.68 100.0% 4.00% 2.47% 100% 200
NC 15 46.70 50.30 48.71 26.4% 4.00% 2.10% 23% 45
ND 3 35.00 55.00 41.70 0.0% 4.00% 3.63% 0% 0

OH 20 50.00 47.00 52.01 83.8% 4.00% 2.21% 85% 169
OK 7 28.00 61.00 35.37 0.0% 4.00% 2.38% 0% 0
OR 7 50.00 44.00 54.02 97.6% 4.00% 3.00% 98% 195
PA 21 50.00 45.00 53.35 95.0% 4.00% 2.22% 96% 192
RI 4 56.00 36.00 61.36 100.0% 4.00% 3.30% 100% 200

SC 8 42.00 55.00 44.01 0.2% 4.00% 2.34% 0% 0
SD 3 42.00 52.00 46.02 2.6% 4.00% 2.45% 2% 4
TN 11 47.80 50.30 49.07 32.5% 4.00% 2.29% 35% 70
TX 34 37.00 59.00 39.68 0.0% 4.00% 2.31% 0% 0
UT 5 24.00 69.00 28.69 0.0% 4.00% 3.18% 0% 0

VT 3 53.00 40.00 57.69 100.0% 4.00% 3.56% 100% 200
VA 13 47.00 51.00 48.34 20.8% 4.00% 2.59% 22% 43
WA 11 52.00 44.00 54.68 98.9% 4.00% 2.12% 100% 199
WV 5 45.80 48.60 49.55 41.3% 4.00% 2.35% 45% 90
WI 10 51.00 44.00 54.35 98.3% 4.00% 2.08% 98% 196
WY 3 29.00 65.00 33.02 0.0% 4.00% 3.50% 0% 0

[TruthIsAll]

New: The National Exit Poll sheet.
Revised: The state exit poll sheet. Added probability calculations.

In the National sheet, compare the 12:22am 13047 respondent timeline (which Kerry won) to the Final 13660 which was matched to the vote.

Suggestions:
Change the weightings in the FINAL "How Voted 2000" demographic from the impossible 43 Bush/37 Gore to a possible 39/39.
See the dramatic effect on the Kerry Vote.

[anaxarchos]

They could make up their own "what about this; what about that?" analysis.
Actually, you could make an argument for the 12:22 exit polls' 41/39 split... and the irony is that it would be "false memory"...

The naysays don't read their own research. The most likely appearance of their response error (if at all) is not in Gore voters who switched to Bush but in people who would have voted for Bush but didn't (Rove's famous 3 million evangelicals). Of course, that would mean that the 3 million wouldn't show up in the new voters and the Kerry split there would increase... exactly as your spreadsheet shows.

And Kerry still wins... what a shock.

[TruthIsAll]

You can link to it from the electionmodel site.

[Michael Collins]

Online product development. Where's DZIKA to do the beautiful graphics?

[catluce]

I completely missed it. Been a little bit depressed about everything.

Were you using too many capital letters again?

ETM

[Michael Collins]

with the appropriate explanation and links. Created a handout and everything, color too, until I found outit was $.89 a page...so it's a tasteful b/w with a color poster.

[h2]INTERACTIVE ELECTION MODEL[/h2]
http://www.geocities.com/electionmod...Simulation.xls

This workbook contains three interactive Excel worksheets:

State Polls - for projections and analysis using pre-election and exit polls

National Exit Poll - analyze demographics for the following timelines:

Preliminary (13047 respondents) at 12:22am
Final Exit Poll (13660 respondents) at 1:25pm.
The Preliminary National Exit Poll followed the 4pm (8349 respondents) and 7:33pm (11027 respondents) timelines. Kerry was leading at each of the three timelines. Bush won the Final Exit Poll in which weights and percentages were matched to the vote.

Summary - State and national exit poll demographic statistics.
Data assumptions can be changed to perform “what-if" analysis of state and national exit polls. Alter demographics on the third sheet.

DO YOU OWN ANALYSIS AND TRY TO FIND ONE PLAUSIBLE WAY IN ANY OF THE DATA SHOWING A KERRY LOSS!

http://www.geocities.com/electionmod...Simulation.xls


[h2]TruthIsAll is now posting at www.ProgressiveIndependent.com[/h2]

(any typos, don't want to hear about it; presume much better formatting;)

[TruthIsAll]

http://us.share.geocities.com/electi...Simulation.xls

[TruthIsAll]

You should have Excel 2002 to run the model.

The purpose of this interactive simulation is primarily educational. It's based on the the Monte Carlo simulation model I developed to project the probability of a Kerry electoral vote victory based on final state polling.

Final state pre-election polls included Zogby, ARG, Gallup, Survey USA and others. You can change these percentages.

Press F9 to run the simulation of 200 trials. The probability of a Kerry state win will change as well as his total popular and electoral votes. In the actual Election Model, there were 5000 simulation trials.

The probability of a Kerry electoral win is equal to the percentage of the 200 trials that he wins. The model totals up the electoral votes for the states in which Kerry wins a majority of the popular vote. If the total electoral vote is 270 or more, he wins that trial.

Note that the margin of error (MoE) is assumed to be 4% for each state poll. This is typical for a state poll of 600 respondents.

You can change the percentage of undecided voters who you assume will vote for Kerry. As this percentage increases, so will his final popular vote - and perhap the electoral vote also. Undecided voters have historically voted 60-80% for the challenger. As expected, the exit polls indicated that Kerry won a majority of late undecided voters. And he also won a solid majority of new voters. And an even larger majority of former Nader voters.

That's why Kerry won. And you know that he knows that we know that he knows he did.

Have fun.

http://www.geocities.com/electionmod...Simulation.xls