election polling & forecasting 2004

21
But first….. Something of topical interest … Psephology The M&M graph…

Upload: dennis-sweitzer

Post on 18-Jun-2015

280 views

Category:

Documents


0 download

DESCRIPTION

A statistics presentation I gave at work using simulation to predict election results from state-by-state polling data.

TRANSCRIPT

Page 1: Election Polling & Forecasting 2004

But first…..

Something of topical interest …

Psephology

The M&M graph…

Page 2: Election Polling & Forecasting 2004

Poll of polls for October 8th, 2004 - Posted October 15th, 2004 Dead Heat

Before the second and third debates, Bush's lead has dropped to less than 0.1%... http://tis.goringe.net/pop/pollofpolls.html

Page 3: Election Polling & Forecasting 2004

Reweighting by Party ID? Shifting Party Affiliation

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

'52 '56 '60 '64 '68 '72 '76 '80 '84 '88 '92 '96 '00

Year

Strong Republican

Weak Republican

IndependentRepublican

IndependentIndependent

IndependentDemocrat

Weak Democrat

Strong Democrat

Page 4: Election Polling & Forecasting 2004

“Dewey Beats Truman” Chicago Tribune, Nov 3, 1948 Sources of bias: •  “distribution of telephones favored wealthy

Dewey voters rather than poor Truman supporters”

•  “the pollsters stopped polling two weeks or more before the election.”

•  Convenience Sampling

Page 5: Election Polling & Forecasting 2004

Golden Age of Polling? •  >95% Households have a telephone

Easy to get random sample •  Until recently: good response •  Now: caller id, answering machines, etc.

30% response rate

Cell Phones: No landline èYounger, Democratic

Page 6: Election Polling & Forecasting 2004

Online Polling?

Page 7: Election Polling & Forecasting 2004

Demographics

•  "As people do better, they start voting like Republicans--unless they have too much education and vote Democratic, which proves there can be too much of a good thing." --Karl Rove

Page 8: Election Polling & Forecasting 2004

Electoral College Map

Page 9: Election Polling & Forecasting 2004

Electoral College Approach

KERRY ELECTORAL VOTE PROJECTIONS from State-Level Poll Data 14-Oct-2004 Bush > 55% 15 110 Bush total: Bush 50-55% 12 146 Kerry 50-55% 15 187 Kerry total: Kerry > 55% beginning ending Estimated Kerry Vote Share estimated probability of Kerry win if election held now* state EV stdev AL 9 44.3% 0.7% 1.0% AK 3 37.0% 1.2% 1.0% AZ 10 46.1% 3.2% 1.0% AR 6 53.1% 3.1% 97.8% CA 55 54.6% 2.8% 99.0% number of states Total Electoral Votes 256 9 95 282 51 538 polls used in estimate: 10/7/04 10/12/04

number of states

Total Electoral

Votes

Bush > 55% 15 110 Bush total:Bush 50-55% 12 146 256Kerry 50-55% 15 187 Kerry total:Kerry > 55% 9 95 282

51 538

beginning ending10/7/2004 10/12/2004

KERRY ELECTORAL VOTE PROJECTIONSfrom State-Level Poll Data

18-Oct-2004

polls used in estimate:

Page 10: Election Polling & Forecasting 2004

Basis for Simulation

State EV K.share SDAL 9 44.3% 0.7%AK 3 37.0% 1.2%AZ 10 46.1% 3.2%AR 6 53.1% 3.1%CA 55 54.6% 2.8%CO 9 46.8% 2.2%CT 7 55.1% 2.7%DE 3 54.5% 2.4%DC 3 89.5% 0.7%

Page 11: Election Polling & Forecasting 2004

Simulate Elections 1 Random Electoral Votes

Rand# SD Mean Normal Kerry BushAL 0.396 0.73% 44.35% -5.84% 9AK 0.281 1.20% 37.03% -13.66% 3AZ 0.314 3.15% 46.08% -5.44% 10AR 0.244 3.12% 53.15% 0.98% 6CA 0.105 2.78% 54.63% 1.14% 55CO 0.441 2.19% 46.76% -3.56% 9CT 0.113 2.70% 55.15% 1.87% 7DE 0.44 2.40% 54.50% 4.13% 3DC 0.474 0.73% 89.49% 39.44% 3

22-Jul 4-Aug 30-Aug 2-Sep 7-SepElect Wins 76.0% 73.0% 56.4% 37.0% 39.2%

Loses 23.6% 26.1% 42.2% 61.2% 59.7%Ties 0.4% 0.9% 1.4% 1.8% 1.1%Odds 3.22to1 2.80to1 1.34to1 0.60to1 0.66to1

Page 12: Election Polling & Forecasting 2004

Estimating States (1)

•  Historical data: Compare state outcomes vs national

•  Current data: Adjust national poll by historical difference

•  Combine with recent state polls

Example: http://www.mydd.com/story/2004/5/28/115448/006

Page 13: Election Polling & Forecasting 2004

Estimating States (2) Election data:

1980-2000 (6 elections) state vote counts

For pair of states (2550 pairs): Regress each states vote on all other states

For each available poll:

Predict votes in all other states

For each state: Use the median of predicted votes

www.pollkatz.homestead.com/files/kerryEVproj.htm

Page 14: Election Polling & Forecasting 2004

Forecasting Models Predict Vote from: Economic conditions,

social climate, incumbency, “party fatigue”, etc., etc.

Limited by: Too little data Result: perfect predictions of past elections Explanatory, not predictive

Page 15: Election Polling & Forecasting 2004

Bread & Peace Equation 14

1 20

lnjt t j t

jVote R CUM KIAα β λ β−

=

= + Δ +∑

40

45

50

55

60

65

-1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Weighted-average growth of real disposable personal income per capita during the presidential term

Tw

o-pa

rty

vote

sha

re

Real Income Growth and the Two-Party Vote Share of the Incumbent Party's Presidential Candidate

%

%

1980

1972

1996

1952

1968

19641956

1984

1988

1992

1976

1960

Vietnam

Korea

Votes as a function of:

• Income growth

• War Casualties

2004: Predicts ~53% Bush

Explanatory, not predictive!

(Who’s income?)

Page 16: Election Polling & Forecasting 2004

Incumbent Advantage or Disadvantage?

•  Forecasting models: Incumbent advantage

•  Final Polls vs. Final Votes: Incumbent disadvantage

Apparently, undecideds lean toward challenger (at about 2:1)

Page 17: Election Polling & Forecasting 2004

My Favorite Forecasting Model Electability = 4P - V - S + R + 9G

+ 95DCI + 95GEN + 95NUC •  Elections since 1932 •  Predicts all elections since 1932 •  Developed using stochastic trials

(i.e., guessing until something worked)

•  Source: Annuals of Improbable Results

http://members.bellatlantic.net/~vze3fs8i/air/pres2004.html

2004: Bush, 70; Kerry -20

Page 18: Election Polling & Forecasting 2004

Another Favorite Washington Redskins, Last home game prior to election

Redskins Win è Incumbent wins Redskins Lose è Incumbent loses

•  True for entire history of Washington Redskins (15 elections)

(1932 & earlier: Boston Braves, no predictive power) •  October 31, 2004: vs Green Bay

Not anymore 28 to 14 defeat, favor Kerry

Page 19: Election Polling & Forecasting 2004

University of Iowa's Electronic Market

http://128.255.244.60/graphs/graph_Pres04_WTA.cfm

Page 20: Election Polling & Forecasting 2004

Obligatory Bayesian Methods A Bayesian Truth Serum for Subjective Data. Prelec,

Drazen. Science, Vol 306, Issue 5695, 462-466 , 15 October 2004

•  Reward based system •  Respondents “compete” •  Counterbalances tendency to agree with

perceived majority •  Best: large samples, rational participants

Page 21: Election Polling & Forecasting 2004

Bayesian Truth Serum Example: Q1: Do you prefer painting A or B? Q2: Which would others prefer? •  Compute Information score+prediction score

using sums of logarithms for each respondant and etc etc •  Truth Telling is Bayesian Nash equilibrium (I.e.,

reduced payoffs for anything else) •  Does it work? When? For who? For only Bayesians?

What about frequentists? Under what conditions? Cost? Etc...

•  A work in progress