Title: Does Early Voting Really Improve Voter Turnout The View from the Academy
1Does Early Voting Really Improve Voter
Turnout?The View from the Academy
- Paul Gronke, Director
- Eva Galanes-Rosenbaum, Research Director
- Early Voting Information Center
- July 16, 2007
- Prepared for the National Association of
Secretary of States Summer Meeting, Portland OR
2Who am I?
- Professor of Political Science at Reed College
and Director of the Early Voting Information
Center - Primary fields of research are voting behavior,
public opinion, survey administration and
statistics - Recent projects
- Carter/Baker Commission report on Voting by Mail
- 2006 Carnegie Corporation project on 2004
Election Day Survey - 2007 EAC Contract on 2006 Election Administration
Survey - Where you can find out more about me
- Email earlyvoting_at_reed.edu
- Web http//earlyvoting.net, http//www.reed.edu/
gronke - Blog http//electionupdates.caltech.edu
3Agenda for today
- Theories of voting and the early voter
- Voting as A Cost/Benefit decision Because EV
decreases cost, turnout should increase - Voting as Consumption Because voting is a
marginal decision, EV should have small effects
(positive and negative) on turnout - Voting and the Social Context If states make it
easier for parties, candidates, and other
organizations to identify early voters, turnout
should increase - Empirical studies of turnout and early voting
- How scholars and policymakers can work together
4State of the World Early Voting is coming and
you cant stop it
6 states cast ballots before Iowa Caucus
12 states cast ballots before New Hampshire
Major share of delegates at stake in early voting
Thanks to
Source Fitzgerald 2005
5Voting as a Cost-Benefit Decision
- Why should anyone vote at all?
- Scholars often conceptualize voting as a
cost-benefit calculation - Vote if C lt pB
- The probability that a citizens vote will make a
difference (p) multiplied by the perceived
benefit from voting (pB) must be greater than the
cost of voting (C) in order for a citizen to
vote. - Prediction rational citizens should not vote and
in equilibrium, turnout should be zero
6How to increase turnout if you believe the
Cost-Benefit Model
- The p (probability that one vote will make a
difference) is virtually nil thus - Efforts to increase turnout work on B or C
- B most GOTV efforts for candidates and policy
campaigns work here, trying to make potential
voters feel that the benefit gained from voting
is large (or that the cost of not turning out
will be severe) - C a few campaigns, and most civil servants
operate here, trying to lower the costs of voting
by making polls easier to access, making
information accessible, etc. - Early Voting may increase turnout by lowering the
costs of voting.
7Voting as consumption
- Many political scientists modify the classic
cost-benefit model of turnout - Vote if C lt pB D
- D has been variously described as civic duty or
value of democracy. - Consumption because you enjoy the benefit
whether or not your candidate wins. You
consume voting. - Prediction turnout should equilibrate at some
level greater than zero, and vary according to
the consumptive value of the election.
8How to increase turnout if you believe the
consumption model
- Manipulate D
- Encourage a sense of civic duty
- Describe voting as a way to express your voice
(Rock the Vote), not to win or lose for policy
purposes - Structure voting so as to encourage community
activity and civic rituals - Early voting
- Should have no impact, because small changes on
the cost side wont translate into big changes on
the turnout side - Should have a negative impact because it destroys
old civic rituals - Should have a positive impact because it creates
a new civic ritual (voting at the kitchen
table). - Regardless, any effects are likely to be small.
9Voting and the Social Context
- Criticism of homo economicus. What about homo
societushumans embedded in a social context? - Institutions play a role by getting people out to
vote - Institutions play a role by teaching skills and
providing resources needed to help navigate
political waters - Evidence
- Mobilization can overcome barriers to voting
- Citizens with greater social connections vote
more frequently - Church attendance
- Members of civic organizations
- Volunteerism
10How to increase turnout if you believe in
mobilization
- Encourage the health of American civic life
- Fund political parties and not just candidate
organizations - Encourage volunteerism and civic engagement no
more bowling alone - Early voting
- May have a positive impact if states and
jurisdictions make permanent absentee lists and
early voting returnseasily and cheaply available
(statewide?) - Allows civic organizations to target mobilization
efforts - Promotes cost efficiency among candidates and
party organizations - May level the campaign playing field and lower
campaign spending
11VBM and Turnout Empirical Evidence
- Academic verdict on VBMs effect
- Small increase in turnout
- EVIC estimates are that voting by mail has
increased turnout in Presidential elections
approximately 5 - One California estimated a decrease of 2.6
- VBM does not expand the electorate
- Tends to draw in regular voters, does not attract
new voters - What we dont know yet
- Changing landscape of voting by mail
- Concerns about novelty effect and other
special circumstances - Effects likely vary in higher and lower profile
contests
12Other Modes and Turnout
- Liberalized absentee voting
- EVIC compared primary types of non-precinct place
voting since 1960, in Federal elections, and
found no impact (other than VBM) - Other studies have found consistently positive
but small (less than 3) and often statistically
insignificant - In Person Early Voting and Voting Centers
- No measurable impact on turnout
- Early voting allows parties to harvest core
voters - What we dont know yet
- Based primarily on studies in Texas and Colorado
- Widespread party mobilization efforts directed at
early voters
13Who votes early? The Early Electorate
- Resource Rich voters (better educated, higher
income) - Committed voters (voters with strong ideological
leanings and partisan ties, older voters?) - Voters who are inconvenienced by traditional
precinct place voting (e.g. voters facing longer
commutes, older voters?) - Early voting does not expand the electorate and
reinforces existing biases in American politics - Early voting methods increase voter turnout on
average (across many different types of
elections), primarily by making it easier for
current voters to continue to participate, rather
than by mobilizing nonvoters into the electorate.
- Very consistent with the theoretical
expectations
14Looking to the future
- Electoral reforms that make voting more
convenient but may exacerbate pre-existing biases - Retention is easy to fix. Likely voters1 who
dont participate already are those who respond
best to lower physical coststheyre already
engaged, wealthy, educated, or habitual voters - Expanding the electorate is hard.
- Turning unlikely voters into likely voters
involves much more than administrative changes - 1. Likely voters are those who share
characteristics with habitual voters older,
wealthier, politically and socially engaged,
educated, etc. The expectation that a member of
these groups will turn out to vote is higher than
for others.
15How EV research can help you
- Study EV as a way to improve election
administration - Give poll workers extra time to learn in a
hands-on environment - Allow many problems to be caught and dealt with
before the huge crowds of Election Day - Give voters the option of returning later if
something isnt working or if access is a
problem - Study EV as a way to increase turnout and expand
the electorate - Experimental studies of different methods of
voter outreach and voter assistance - Different methods for different citizens
- Study EV across jurisdictions to establish best
practices
16How you can help research
- Tracking the number of early votes in an election
is a source of considerable frustration - States such as Texas, Oregon, Florida make data
readily available, in real-time and permanently - Other states dont make these numbers available
easily - Often cant produce the data when directly
solicited - Many dont track EV at the state level (must go
to each county) - Inconsistent definitions
- Quick collection and analysis of national EV
rates is currently impossible - Take full advantage of statewide voter
registration systems and link to voter history
files