1
A
Peek
Inside:
Using
Provisional
Ballots
to
Explore
Photo
Identification
Laws
and
Voter
Turnout
at
the
Polls
Provisional
ballots
can
be
a
valuable
tool
in
assessing
how
electoral
rules,
such
as
Indiana’s
new
photo
identification
law,
impact
voter
turnout
within
the
polls
themselves.
They
provide
a
unique
measurement
of
how
many
people
encounter
administrative
difficulties
within
the
polls,
forcing
them
to
vote
via
alternative
means.
They
also
leave
a
record
of
how
many
of
those
provisional
votes
are
subsequently
verified
after
election
day.
This
paper
has
3
major
aims:
1)
make
the
case
for
how
provisional
ballots
can
move
forward
the
investigation
of
photo
identification
requirements,
2)
generate
hypothesis
about
what
results
might
look
like,
and
3)
examine
data
from
the
2006
election
cycle.
Crawford
v.
Marion
County
Election
Board,
the
recent
Supreme
Court
case
upholding
Indiana’s
controversial
photo
identification
requirement,
rested
upon
a
state
legislature’s
ability
to
regulate
a
perceived
threat
to
the
electoral
system
via
any
means
that
would
not
be
facially
discriminatory.
The
court
relied
upon
the
lack
of
evidence
that
photo
identification
requirements
necessarily
suppressed
turnout
among
disadvantaged
groups
in
its
ruling,
while
at
the
same
time
deciding
that
the
law
did
not
require
the
demonstration
of
actual
voter
fraud
at
the
polls.
In
the
presence
of
suspected
voter
fraud,
the
ruling
declared
that
the
legislature
had
the
right
to
impose
rigid
standards
to
protect
against
perceived
threats
so
long
as
they
were
not
clearly
discriminatory.
The
burden
of
proof
rests
almost
purely
upon
the
opponents
of
photo
identification
to
demonstrate
the
impact
of
the
law
is
in
fact
disproportionately
felt.
Photo
identification
requirements
have
been
studied
only
for
the
past
4
years,
as
they
have
come
into
existence,
but
other
administrative
procedures
have
been
studied
for
far
longer,
providing
guidance
to
contemporary
researchers
on
how
to
investigate
the
effect
of
new
identification
standards.
As
early
as
1909,
identification
requirements
at
the
poll
(in
this
case
simply
matching
a
declared
name
with
a
pre‐registered
address)
were
demonstrated
to
reduce
turnout
in
New
York
City
(Lapp
1909).
Large
declines
in
aggregate
turnout
were
noted
in
this
study,
the
assumption
being
that
this
was
primarily
due
to
reductions
in
poll
fraud.
The
simpler
electoral
procedures
of
the
time
permitted
a
straightforward
interpretation
of
how
turnout
altered
after
voters
were
required
to
register
ahead
of
time
and
identify
themselves
on
election
day.
Research
in
the
succeeding
decades
has
grown
more
statistically
sophisticated,
as
have
poll
regulations,
and
have
expanded
to
examine
a
variety
of
administrative
regulations,
but
have
often
relied
upon
similar
techniques
to
assess
Convention