Journal
of
Applied Psychology Copyright 1985
by the
American
Psychological
Association,
Inc.
1985,
Vol.
70, No. 3,
556-564
0021-90IO/85A00.75
Improving
Eyewitness Identifications From Lineups:
Simultaneous
Versus
Sequential Lineup Presentation
R. C. L.
Lindsay Gary
L.
Wells
Queen's
University
University
of
Alberta
Kingston,
Ontario, Canada Edmonton, Alberta, Canada
Staged crime research
has
demonstrated
the
utility
of
controlling
the
conduct
of
lineups
as a
means
of
reducing
false
identifications with little
or no
apparent
decline
in the
rate
of
correct identifications
by
eyewitnesses (e.g., Lindsay
&
Wells,
1980;
Malpass
&
Devine,
1981a; Wells, 1984).
A
recent variation
in
lineup
procedure shows that
a
blank
lineup, which includes
no
suspects,
can
reduce
the
rate
of
false
identifications
if it
precedes
the
actual lineup.
However,
there
are
several
practical problems that make
it
unlikely
that police
will
accept this
procedure. Sequential lineup presentation
is
proposed
as a
means
of
accomplishing
the
same goals
of
reducing
false
identifications with little
or no
loss
in
accurate
identifications.
A
crime
was
staged
for 240
unsuspecting eyewitnesses either
individually
or in
pairs.
One
quarter
of the
eyewitnesses attempted identifications
in
each
of
four
lineup conditions:
Six
pictures
were
presented either simultaneously,
as
used
in
traditional
procedures,
or
sequentially,
in
which
yes/no judgments
were
made
for
each picture; each procedure either contained
the
photograph
of
the
criminal-confederate
or a
picture
of a
similar
looking replacement. Sequential
lineup presentation
significantly
reduced
false
identifications
but did not
signifi-
cantly
influence
correct identifications
when
compared
with
the
simultaneous
procedure.
This resulted
in an
overall increase
in
diagnosticity
ratio
(Wells
&
Lindsay,
1980) using
the
sequential procedure.
The
data
are
interpreted
as
supporting
the
conclusion that sequential presentation
of
lineups
can
reduce
false
identifications
of
innocent suspects
by
reducing eyewitnesses' reliance
on
relative-
judgment
processes.
Wells
(1978)
argued
that
two
types
of
(Loftus,
1979);
however,
relatively
little
staged-
variables
could
be
distinguished
in
eyewitness
crime,
system-variable
research
has
dealt
with
research:
(a)
estimator
variables,
such
as
race
issues
of
eyewitness
identification.
or
sex,
which
may
affect
eyewitness
accuracy
Investigation
of
eyewitness
identification
but are not
controllable
in
actual
cases
and
accuracy
has
demonstrated
considerable
vari-
(b)
system
variables,
such
as
lineup
structure,
ance
in the
rate
of
correct
and
false
identifi-
which
not
only
affect
accuracy
but
also
can
cations.
In
particular,
false
identifications
of
be
controlled.
The
potential
value
of re-
innocent
people
have
been
obtained
from
as
searching
system
variables
is as a
means
of few as 12%
(Leippe,
Wells,
&
Ostrom,
1978)
improving
on
current
procedures.
To
date,
or as
many
as 78%
(Malpass
&
Devine,
system-variable
research
has
been
conducted
198
la)
of the
subjects
in
some
conditions
of
on
questioning
procedures,
and the
resultant
staged
crime
experiments.
If
this
variance
in
accuracy
of
eyewitness
memory
has
been
rate
of
false
identification
is
produced
by
reflected
in
descriptions
of
people
and
events
differences
in
system
variables,
substantial
_
reductions
in
the
risk
of
real
world
identifi-
_
This
research
was
supported
by
a
grant
from
the
â„¢
lion
errors
™&«
be
obtained
by
following
Social
Sciences
and
Humanities
Research
Council
of
appropriate
procedures.
Canada
to the first
author.
Lineup
and
photo-spread
identification
Requests
for
reprints
may be
sent
to R. C. L.
Lindsay,
procedures
typically
involve
the
presentation
Department
of
Psychology,
Queen's
University,
Kingston,
t
.,
evew
i
t
ne<isr^
of one
<iusnect
embedded
Ontario,
Canada
K7L
3N6
or
Gary
L.
Wells,
Department
to
tne
eyewitnesses)
or one
Suspect
emoeaaea
of
Psychology,
University
of
Alberta,
Edmonton,
Alberta,
among
toils
or
distractors.
A
foil
Or
Canada
T6G
2E9.
distractor
is a
stand-in
who is not a
suspect
556
SIMULTANEOUS
VERSUS
SEQUENTIAL
LINEUP
PRESENTATION
557
but
instead
is
known
to be
innocent (e.g.,
police
officers
or
persons drawn
from
jail
cells). Although some lineups
may be
com-
posed
entirely
of
suspects,
the
opinions
of
legal
experts (e.g.,
Sobel,
1973)
and
psychol-
ogists
(e.g.,
Ellison
&
Buckhout,
1981;
Wells
&
Lindsay,
1980)
clearly argue against such
practices.
The
advantages
of
having known-
innocent
foils
in
lineups
are
numerous,
and
single-suspect lineups
are
therefore
the
state-
of-the-art
model
on
which
the
current study
is
based.
Lineups
may be
live
(the
so-called
corporeal
lineup)
or
color photographs.
In-
creasingly, police departments
are
using
the
photographic method
for
several
reasons,
in-
cluding
the
greater ease with which acceptable
distractors
can be
found
and the
fact
that
right
to
counsel does
not
accrue
to the
suspect
in the
case
of
photographic identification
procedures.
Although
the
latter reason seems
especially questionable (see Grano,
1984),
it
should
be
noted that there
is no
evidence
that photo identifications
are
less
reliable
than
live
identifications,
even
when
the
wit-
nessed event
was a
live
event (Shepherd, Ellis,
&
Davies,
1982).
Lindsay
and
Wells
(1980) outlined
the
possible outcomes that
can
result
from
a
lineup
identification procedure using
the
sin-
gle-suspect model.
If the
lineup contains
the
perpetrator,
the
eyewitness
can
make
an ac-
curate identification,
a
foil
identification,
or
an
incorrect
nonidentification.
If the
lineup
does
not
contain
the
perpetrator,
the
eyewit-
ness
can
make
a
false
identification,
a
foil
identification,
or a
correct nonidentification.
In
the
real world case,
only
foil
identifications
are
known
at the
time
to be
eyewitness errors.
(Indeed, this
is one of the
reasons
why it is
important
in
actual cases
to
have
foils
rather
than
all
suspects—so
that
at
least
one
type
of
error
can be
detected with certainty.) Iden-
tifications
of the
suspect
and
nonidentifica-
tions
may be
either accurate
or
inaccurate
depending
on
whether
the
lineup contains
the
actual perpetrator.
It
is
important
to
keep
in
mind that
the
identification
of a
foil
is a
known error
and
does
not
function
as a
false
identification
in
the
true
sense—for
example, charges will
not
be
brought against
Officer
Jones
if he is
identified.
Thus, when using
the
single-suspect
lineup
model,
false
identifications
can
occur
only
in
perpetrator-absent lineups. This
im-
portant observation
has
paved
the way for
Bayesian
analyses
of
lineups
(Wells
&
Lindsay,
1980).
Using this Bayesian approach, Wells
and
Lindsay outlined
the two
main statistical
properties
of
lineup identifications, noniden-
tifications,
and
foil
identifications; namely
diagnosticity
and
informativeness.
Diagnos-
ticity
is a
directional likelihood ratio, whereas
informativeness
is a
combination
of the
like-
lihood ratio
with
prior probabilities.
The
simpler
of the
two, diagnosticity, captures
the
important elements
of the
lineup procedure
because higher levels
of
diagnosticity produce
greater
informativeness
for all
prior
proba-
bilities
greater
than
zero
(0) and
less
than
one
(1.0).
The
diagnosticity
of an
identifica-
tion
procedure
is
defined
as the
ratio
of
accurate identifications
to
false
identifications.
As
indicated earlier, accurate identifications
occur only
in
perpetrator-present lineups,
and
false
identifications occur only
in
perpetrator-
absent lineups. Similarly,
the
diagnosticity
of
foil
identifications
is the
ratio
of
foil
identi-
fications in
perpetrator-absent
lineups
to
foil
identifications
in
perpetrator-present lineups.
The
diagnosticity
of
nonidentifications
is the
ratio
of
nonidentifications
in
perpetrator-ab-
sent lineups
to
nonidentifications
in
perpetra-
tor-present lineups. Obviously,
the
higher
the
diagnosticity ratio
the
more
informative
or
diagnostic
the
identification decision
from
the
lineup
will
be
with regard
to the
guilt
or
innocence
of the
suspect.
Higher diagnosticity
ratios
for
identifications
can
result
from
in-
creases
in
accurate identifications, decreases
in
false
identifications,
or
both. These
may
result
from
better witnesses
or
better testing
procedures.
The
current study randomly
as-
signed
witnesses
to
testing procedures
so
that
diagnosticity
differences
are
attributable
to
testing procedures.
There
are
three previous studies that
have
systematically
varied
eyewitness-identification
testing
procedures
in
attempting
to find im-
provements
in
diagnosticity.
1
Malpass
and
'
Another
study (Malpass
&
Devine,
198lb)
attempted
to
vary
testing
procedures
with lineups,
but its
status
is
not yet
clear
because
of the
lack
of a
perpetrator-absent
lineup
control.
558
R.
C. L.
LINDSAY
AND
GARY
L.
WELLS
Devine
(198la)
staged
a
vandalism
and had
eyewitnesses attempt
an
identification under
conditions
in
which
the
perpetrator
was
pres-
ent in or
absent
from
the
lineup
and
eyewit-
nesses were
led to
think
the
vandal
was
present
or
"may
or may
not"
be
present.
Their results were
a
clear demonstration
of
the
importance
of
telling eyewitnesses
that
the
true
perpetrator
may not be in the
lineup.
Their results showed
a
robust decline
in the
likelihood
of
false
identifications
(from
78%
down
to
33%)
and no
loss
in
accurate iden-
tifications
(75%
to
83%)
for the "is
present"
versus
"may
or may not be
present" condi-
tions,
respectively. Their results translate
into
diagnosticity
ratios
of .96 and
2.52, respec-
tively,
for the two
sets
of
instructions. This
is
a
profound
demonstration
of the
potential
for
meaningful
improvements resulting
from
system
variable research
with
lineups.
Lindsay
and
Wells
(1980)
demonstrated
that
further
improvement
can be
obtained
even
under conditions
in
which eyewitnesses
are
told
that
the
perpetrator
may or may not
be
present.
Specifically,
they varied
the
phys-
ical
similarity
of
lineup members
and
found
that lineups containing consistently poor
foils
(i.e.,
foils
that
fail
to
match
the
general
description
of the
perpetrator) produce lower
levels
of
diagnosticity than
do
lineups con-
taining
distractors
that tend
to
match
the
general description
of the
perpetrator.
It is
important
to
note
for
later purposes that
the
similarity variable
had no
effect
on
willingness
of
eyewitnesses
to
attempt
an
identification.
Instead,
high similarity resulted
in a
spread
of
choices
from
the
innocent suspect
to
known-innocent
foils
(in
perpetrator-absent
lineups)
without
comparable spread
from
the
guilty
suspect
to
foils
(in the
perpetrator-
present lineup).
Recently,
Wells
(1984)
proposed that
the
improvements
in
diagnosticity resulting
from
instructions
(Malpass
&
Devine,
198la)
and
similarity
(Lindsay
&
Wells,
1980) could
be
understood
to
result
from
a
tendency
for
eyewitnesses
to
make
"relative
judgments."
That
is,
eyewitnesses tend
to
choose
the
lineup
member
who
most looks like
the
perpetrator
relative
to the
other lineup members (cf.
a
comparison
of
each lineup member
to
one's
recollection
on a
more
"absolute"
basis
of
comparison). Clearly,
the
relative-judgment
strategy
is
adequate
when
the
perpetrator
is
present
in the
lineup. However,
when
the
perpetrator
is
absent,
the
relative-judgment
strategy necessarily elicits
errors.
Telling eye-
witnesses that
the
perpetrator
may not be in
the
lineup helps reduce
false
identifications
because
it
makes
apparent
the
fallacious
use
of
the
relative-judgment strategy under per-
petrator-absent conditions.
The way in
which
the
similarity variable
affects
diagnosticity
may
also
be
understood
in
terms
of the
relative-judgment conceptu-
alization.
The
pattern
of
data
in the
Lindsay
and
Wells
(1980)
study shows that increases
in
similarity between
foils
and
suspect have
little
or no
effect
on
witnesses' ability
to
choose
the
suspect when
the
suspect
is
guilty
(i.e.,
he
remains
the
best
choice relative
to
other lineup members)
but
serves
to
spread
choices
to
foils
in the
perpetrator-absent line-
ups
(i.e.,
where
the
suspect
is
innocent).
Based
on the
relative-judgment conceptu-
alization,
Wells
(1984) proposed that those
eyewitnesses
most prone
to
making relative
judgments
could
be
"screened"
or
"lured"
by
a
blank lineup.
A
blank lineup
is one
con-
taining
no
suspect
(i.e.,
composed entirely
of
foils
who are
known
to be
innocent).
He
exposed
unsuspecting students
to a
staged
crime
and
subsequently asked them
to at-
tempt
to
identify
the
culprit. Half
of the
192
eyewitnesses participated
in the
standard
identification
procedures, with
48
viewing
a
lineup
containing
a
photo
of the
criminal
and
48 a
lineup
in
which
the
guilty party
was
replaced
by a
similar other.
The
remain-
ing
96
participants
first
viewed
a
"blank"
lineup
containing
no
suspect
and
then were
shown
one of the
lineups seen
by
other
witnesses.
Eyewitnesses
who
made
a
selection
from
the
blank lineup (those
that
Wells's
procedure would screen out) were
significantly
less
likely
to
accurately
identify
the
guilty
party,
significantly
less
likely
to
make
an
accurate
no-identification
decision,
and
sig-
nificantly
more
likely
to
identify
another
in-
nocent person
from
the
second lineup than
eyewitnesses
who
made
a
no-identification
decision
from
the
blank lineup. Compared
to
the
nonscreened control group, eyewitnesses
who
successfully
passed
the
screening test
SIMULTANEOUS
VERSUS
SEQUENTIAL
LINEUP
PRESENTATION
559
(i.e.,
those
not
making
an
identification
from
the
blank lineup) made
significantly
fewer
false
identifications
and
significantly
more
correct no-identification
decisions
in the ab-
sence
of the
criminal
as
well
as
significantly
fewer
incorrect identifications (choices
of
foils)
in
the
presence
of the
criminal.
Clearly
the
screening
of
eyewitnesses that
tend
to
employ
a
relative-judgment strategy
successfully
increased
the
accuracy
of
resul-
tant
identification
decisions.
However, some
practical limitations might reduce police
will-
ingness
to
adopt this technique.
Among
their
objections
would
be the
requirement
of
con-
ducting
two
lineups, thus recruiting twice
as
many
foils.
All
eyewitnesses making
a
selec-
tion
from
the first
lineup would
be
dis-
carded—not
a
problem
for a
researcher
who
simply
runs another student through
a
staged
crime
but a
potentially serious loss
in a
real-
world
case
involving
a
single eyewitness. Fur-
thermore,
the
technique
may
cease
to be of
value
if the
public became aware
of the
fact
that
two
lineups
were
commonly used
and
that
the
"real"
suspect appears
in the
second
(a
technique recently portrayed
in
Hill
Street
Blues).
What
is
needed
is a
procedure that reduces
the
tendency
to
employ
a
relative-judgment
process without
(a)
eliminating eyewitnesses
from
the
"real"
identification test,
(b) in-
creasing
the
cost
and
workload
for the
police
(i.e.,
by
increasing
the
number
of
people
required
to
conduct
a
lineup),
or (c)
allowing
potential eyewitnesses
to
predict when
or
where
the
true suspect
is
likely
to
appear
in
the
procedure.
The
absolute number
of
people
required could
be
reduced
by
limiting
the
"arrays"
to
less than
six
people each. Simi-
larly,
the
public could
be
prevented
from
anticipating
the
appearance
of the
suspect
by
randomly assigning
the
suspect
to an
array
while
informing
the
witness that more than
one
array
will
be
employed.
The
knowledge
that more than
one
array
or
lineup
will
be
presented ought
to
reduce eyewitnesses' reli-
ance
on
relative judgments because
the
wit-
ness
has no
means
of
determining
in
which
array
the
suspect
will
appear.
If
this
reasoning
is
correct,
the
only remaining issues
to be
determined
are the
optimal number
of
targets
per
array
and
total
number
of
arrays.
Because
an
ideal procedure would avoid
police objections,
the
total
number
of
people
involved
in the
identification
procedure
could
be
limited
to the
number typically employed
in
a
single lineup. Assuming
a
six-person
lineup,
for
example,
the
proposed technique
could employ
two
sets
of
three, three sets
of
two,
or six
individuals presented
sequentially.
Because
the
idea
is to
prevent
as
much
as
possible
the
opportunity
for
eyewitnesses
to
make
relative judgments,
we
opted
for the
sequential
technique,
in
which
the
eyewitness
sees
one
lineup member
at a
time, makes
a
yes/no decision
for
each lineup member
on
first
viewing,
and is not
informed
of the
total
number
to be
viewed.
If the
relative-judgment
strategy
is a
major source
of
false
identifica-
tions,
the
sequential procedure ought
to be a
powerful
means
of
preventing relative judg-
ments;
it
should
force
eyewitnesses
to
compare
each lineup member
to
their recollection
of
the
perpetrator
using some
absolute
standard
of
recognition rather than considering
who
"most"
looks
like
the
perpetrator.
Method
Participants
Introductory psychology students
were
recruited
by
telephone
to
participate
in a
study
of
"complex
infor-
mation processing."
A
total
of 243
Queen's
University
students
volunteered
to
participate. From
1 to 4
students
took
part
in
each session.
Procedure
The
experimenter
met the
participants
at the
laboratory
door,
led
them into
a
large room
(9
m
X 10
m),
and
then
led
them into
a
small
cubicle
(3 m X 3 m)
containing
a
table
and two
chairs
(if
more than
two
students
were
available
for a
session,
a
second similar cubicle
off the
same
large
room
was
employed).
The
cubicle(s)
had a
counter
at
waist
level
on two
walls
with
drawers
and
cupboards below
the
counter
and
cupboards above
the
counter.
After
seating
the
participants,
the
experimenter
explained
that
she had to
leave
briefly
to
"get some
forms"
and
closed
the
door
to the
cubicle
on her way
out.
Approximately
30 s
after
the
experimenter
left,
a
21-year-old
male Caucasian entered
the
cubicle. Looking
directly
at the
subjects
he
stated that
he
"didn't
know
anyone
was in
this
room"
and
would
be
"out
of the way
in
just
a
minute."
He
then searched through several
drawers
and
cupboards along both walls
finally
finding a
calculator
in a
leather
case.
With
the
calculator
in
hand,
he
crossed
to the
door
while
looking
in the
direction
of
the
(now) eyewitnesses
and
left,
closing
the
door behind
him.
The
entire
event
lasted
approximately
20 s.
560
R. C. L.
LINDSAY
AND
GARY
L.
WELLS
The
experimenter
returned
shortly
after
the man had
left
and
explained
that
the
thief
was a
confederate
and
that
the
students
had
witnessed
a
staged crime.
All
participants received
an
explanation
of the
true purpose
of
the
experiment
(to
investigate
eyewitness
identification
accuracy)
and
were asked
to
sign
a
consent
form.
2
Witnesses
were
then
separated
(if
more than
one
witness
was
in the
cubicle)
and
completed
an
open-ended
descrip-
tion
of the
criminal, then
a
detailed questionnaire about
his
appearance.
Approximately
5
min
after
viewing
the
crime,
the
eyewitnesses were asked
to
attempt
to
identify
the
criminal
from
photographs.
Identification
Procedures
Half
of the
eyewitnesses
(simultaneous
presentation
conditions)
were
shown
a
board
on
which
was
mounted
a set of six
photographs
of
male Caucasians
in
their
early
20s.
Each
picture
was a 7 cm by 7 cm,
head
and
shoulder
view
from
the
front.
For 60 of
these witnesses
the
array
contained
a
photograph
of the
criminal confederate,
whereas
for the
remaining
60, his
picture
was
replaced
with
a
photograph
of a man who resembled
him.
The
replacement photograph
was
selected
on the
basis
of
pilot
testing
to be the
most
similar
in
appearance
to the
confederate
of
18
available
photos
of men fitting the
general
description. Although some studies
and
real-
world
cases
have
demonstrated that innocent suspects
may
be
identified
even
when
they
are not
particularly
similar
in
appearance
to the
criminal,
a
stronger test
of
the
value
of
sequential lineup presentation
was
provided
by
having
a
similar-looking innocent suspect.
The
pho-
tographs were numbered
from
1 to 6 and
were
presented
in
two
rows
of
three photographs. Each witness
was
also
provided
with
an
identification
form. Across
the
middle
of the
form
were
six
boxes containing
the
numbers
1
through
6.
Below
there
was a
seventh
box
labelled
"none
of
them."
The
instructions asked
the
witness
to put an
"X"
through
the
appropriate box.
At the
bottom
of the
form
the
witnesses
were
asked
to
rate
on a
7-point
scale
how
certain they were
that
their identification decision
was
correct.
The
experimenter placed
the
picture
array,
covered
by the
identification
form,
on the
table
in
front
of the
eyewitness.
Before
the
eyewitness looked
at the
photographs,
the
experimenter pointed
to and
read aloud
the
following
statement
from
the top of the
identification
form:
"Remember,
as in a
real identification situation,
the
guilty
party
may or may not be
present."
After
reading this statement
the
experimenter immediately
left
the
room
and
returned
3 min to 5 min
later.
All
witnesses
completed
the
identification
form
in the
alloted
time.
Witnesses
in the
sequential
lineup
presentation con-
ditions were provided
with
a
different
identification
form.
The
identically worded warning
that
the
criminal
may or
may
not be in the
lineup
was at the top of the
form.
The
next
line
read
"Is #1 the
person
you
saw?
No,
Yes"
and
was
followed
by a
7-point certainty
scale.
Eleven more
lines
followed,
identical
in
every
regard except
that
the
number
varied
from
2
through
12. The
experimenter
explained that
the
eyewitness
would
be
shown
a
sequence
of
individual
photographs
and
must
decide
for
each
whether
or not it was a
picture
of the
criminal.
It was
emphasized that
the
witness could take
as
long
as he or
she
wished
to
decide
but
that each photograph
would
only
be
seen once.
The
experimenter held
in her
hand
a
"deck"
of 12
photographs,
face
down. Thus
the
experi-
menter
first
read
the
warning, then asked
"Is
number
one
the
person
you
saw?,"
then turned
a
photograph over
in
front
of the
witness
and
waited
for the
witness
to
circle
either
no or yes and a
second number indicating
how
certain
the
witness
was of
this decision.
After
both
responses,were
circled,
the
photograph
was
placed
on the
bottom
of the
deck
and the
procedure repeated. Following
the
sixth
photograph
the
experimenter explained that
there would
be no
further
pictures. Witnesses were
intentionally
led to
believe that there might
be
more
than
six
photographs
to
reduce
any
tendency
to
increase
the
probability
of
making
a
"yes" response
as the end of
the
pile approached.
Half
of the
sequential
lineup
eyewitnesses
were
shown
six
photos
including
a
picture
of the
confederate-thief.
The
remaining
half
saw a
photo
of the
innocent suspect
rather
than
the
guilty
party.
The
same
five
photographs
of
foils
were
used
in all
conditions.
Twenty
eyewitnesses
from
each
of the
four
experimental conditions (simulta-
neous
vs.
sequential presentation
by
perpetrator present
vs.
absent)
saw the
suspect's photograph
in the first,
third,
or fifth
position.
The
order
of
foils
in
Positions
2,
4 and 6 was
constant,
and the
position
of all
foils
was
constant
for a
given position
of the
suspect,
regardless
of
mode
of
presentation.
For the
sequential presentation
conditions,
the
confederate arranged
the
order
of the
deck
in
advance,
and
thus
the
experimenter
was
blind
to
the
position
of the
suspect's
picture
at
least
until
it was
turned
over.
She was
instructed
to
ignore
the
pictures
and
to
restrict
her
gaze
to the
witnesses' identification
forms
to the
best
of her
ability.
The
presence
of the
experimenter
in the
room during
the
identification procedure
is
confounded
with
mode
of
presentation. However, absence
of the
experimenter
for
the
simultaneous procedure
is
consistent
with
attempts
to
minimize experimenter
or
police
officer
effects
(as
recommended
by
Brooks, 1983). Keeping
the
experi-
menter blind
to
position
and
asking
her to
look only
at
the
response sheet rather than
the
photographs
was
intended
to
reduce
or
prevent
such
effects
in the
sequential
presentation conditions.
After
the
identification
procedure
was
completed,
all
witnesses
who had
identified
the
suspect
(guilty
or
inno-
cent)
were asked
if
they
would
be
willing
to
testify
in a
mock
trial
later
in the
year.
The
purpose
of
this measure
was
not
only
to
recruit eyewitnesses
for a
subsequent
2
Three subjects
refused
to
sign
the
consent
form.
Upon
questioning
they
reported
that
the
signing
of a
"legal
document"
for
such
a
mundane procedure seemed
like
"overkill"
and
they suspected that
the
form
was
deceptive
(it was
not). These subjects
did not
participate
further.
In
spite
of
this response
from
these three people,
recent research suggests
that
debriefing
subject-witnesses
at
this point
has no
appreciable
effect
on the
results
(Murray
&
Wells,
1982).
In
other words,
it
seems
that
incidental
encoding
is the
primary prerequisite
for
creating
an
analog
to the
actual eyewitness situation
and
that
continuing
the
ruse beyond
this
point
is
unnecessary.