Congressional Voters
17.251
Fall 2004
Turnout
Presidential election
House election in presidential year
House election in non-presidential year
Turnout as pct. of voting age population
60
50
40
30
1930
1940
1950
1960
1970
Election year
1980
1990
2000
How to Calculate Turnout Pct.
• Turnout Pct. = Turnout / VAP
not
• Turnout / Registered
• New measure: Turnout / Voting eligible
population
Turnout/VEP vs. Turnout/VAP
70
60
VEP
50
40
30
20
20
30
40
50
VAP
60
70
Variation in Turnout
1998 & 2000
50
10
40
30
6
Count
Pct.
8
20
4
10
2
0
0
0
40000
80000 120000 160000 200000 240000 280000 320000 360000 400000
Turnout 2000
Mean = 223,000
0
100000
200000
2002
Mean = 171,469
300000
400000
Correlation in Turnout
1998 to 2000
400000
Turnout 2000
300000
200000
100000
0
0
100000
200000
Turnout 1998
300000
400000
Primary & General Election
Turnout, 2000
15
Pct.
Pct.
15
10
5
10
5
0
0
0
40000 80000 120000 160000 200000
20000 60000 100000 140000 180000 220000
Primary 2000
0
40000
80000 120000 160000 200000 240000 280000 320000 360000 400000
General 2000
Explaining (Non-)Voting
• Expected value of voting =
– Benefit the individual receives as a
consequence of the election outcome
– Minus
– The cost of voting
Explaining (Non-)Voting
State of the World
Net Benefit if Citizen
w/out Citizen’s Vote Abstains
Net Benefit if
Citizen Votes
Condition under which
Citizen Should Vote
D wins by more
than 1 vote
BDCitizen
BDCitizen – c
Never
D wins by exactly 1
vote
BDCitizen
BDCitizen – c
Never
D and R tie
(BDCitizen + BRCitizen )/2
BDCitizen – c
(BDCitizen - BRCitizen )/2 > c
R wins by exactly 1
vote
BRCitizen
(BDCitizen +
BRCitizen )/2 – c
(BDCitizen - BRCitizen )/2 > c
R wins by more
than 1 vote
BRCitizen
BRCitizen – c
Never
Salvaging the Calculus
– Voter attention
– GOTV
• Citizen duty
300000
Turnout
• Candidate differenial
• Costs of voting
• Closeness of election
200000
100000
0
50
60
70
80
Winning pct., 2002
90
100
Who is hurt/helped by turnout
• Naïve view: Dems helped by turnout
Who is hurt/helped by turnout
• District view: the “out party”
Campaign intensity
Deciding whom to support
• Ideology
– Downsian logic directly
• Party ID
– Downsian logic by proxy
Party and Ideology Distance as Explanatory
Factors in 2000 Cong’l Elections
House
Party ID
Dem
Ideology of voter
Lib. Mod. Con. Total
.70
.81
.86 .81
Ind.
Rep
Total
.54
.29
.78
.60
.33
.67
Source: 2000 ANES
Note: R’s not asked ideological
placement of House or Senate
candidates
.47
.16
.32
.50
.18
.51
Senate
Party ID
Dem
Ideology of voter
Lib. Mod. Con. Total
.80
.87
.90 .83
Ind.
Rep
Total
.79
.28
.84
.67
.09
.61
.48
.15
.34
.60
.16
.54
Overall voting effect, 2002 House election
Party
identification
Effect of changing from an
Ind. to a Dem.
0.25
(0.02)
Ideology
Effect of changing from a
mod. To a lib.
0.07
(0.02)
Democratic
incumbent
Effect of changing from a
open seat race to a Dem. Inc.
0.18
(0.02)
Constant
0.54
(0.02)
R2
.49
N
586
A Word about Primaries
• Party not a useful cue
• Not much research, but…..
– Primary voters are different from general
election voters
• Primary voters are more ideologically extreme, but
• Primary voters are more strategically sophisticated
than general election voters
The Ideological Purity/Electability
Tradeoff
Electability
C2
C1
Purity