Obtaining guilty knowledge in human intelligence interrogations: Comparing accusatorial and information-gathering approaches with a novel experimental paradigm

Abstract

Substantial research has assessed interrogations seeking to obtain a criminal confession, and consequently much has been learned regarding the potential problems with confession evidence. However, an increasing focus on counter-terrorism, and therefore intelligence interrogations, reveals an obvious gap in the literature. Intelligence interrogations are primarily focused on collecting information from individuals as opposed to a confession linked to an alleged event, and little of the extant psychological literature can speak directly to such a scenario. The current research developed an experimental paradigm to test interrogation approaches in an intelligence-gathering context, providing a method for gathering empirical data on human intelligence collection. In the first implementation of this paradigm, accusatorial and information-gathering interrogation strategies were tested using a procedure high in psychological realism. Results indicate that an information-gathering approach yields more relevant information than an accusatorial approach and leads to more diagnostic impressions by third party observers.

Full text

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights

Author's personal copy
Journal
of
Applied
Research
in
Memory
and
Cognition
2
(2013)
83–88
Contents
lists
available
at
SciVerse
ScienceDirect
Journal
of
Applied
Research
in
Memory
and
Cognition
jo
ur
nal
homepage:
www.elsevier.com/locate/jarmac
Obtaining
guilty
knowledge
in
human
intelligence
interrogations:
Comparing
accusatorial
and
information-gathering
approaches
with
a
novel
experimental
paradigm夽
Jacqueline
R.
Evansa,∗,
Christian
A.
Meissnerb,
Amy
B.
Rossb,
Kate
A.
Houstonb,
Melissa
B.
Russanoc,
Allyson
J.
Horgand
aPsychology
and
Counseling,
University
of
Texas
at
Tyler,
United
States
bPsychology,
University
of
Texas
at
El
Paso,
United
States
cSchool
of
Justice
Studies,
Roger
Williams
University,
United
States
dMRAC
LLC,
United
States
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
26
September
2012
Received
in
revised
form
14
March
2013
Accepted
14
March
2013
Keywords:
Interrogation
Interviewing
National
security
a
b
s
t
r
a
c
t
Substantial
research
has
assessed
interrogations
seeking
to
obtain
a
criminal
confession,
and
conse-
quently
much
has
been
learned
regarding
the
potential
problems
with
confession
evidence.
However,
an
increasing
focus
on
counter-terrorism,
and
therefore
intelligence
interrogations,
reveals
an
obvious
gap
in
the
literature.
Intelligence
interrogations
are
primarily
focused
on
collecting
information
from
individ-
uals
as
opposed
to
a
confession
linked
to
an
alleged
event,
and
little
of
the
extant
psychological
literature
can
speak
directly
to
such
a
scenario.
The
current
research
developed
an
experimental
paradigm
to
test
interrogation
approaches
in
an
intelligence-gathering
context,
providing
a
method
for
gathering
empir-
ical
data
on
human
intelligence
collection.
In
the
first
implementation
of
this
paradigm,
accusatorial
and
information-gathering
interrogation
strategies
were
tested
using
a
procedure
high
in
psychological
real-
ism.
Results
indicate
that
an
information-gathering
approach
yields
more
relevant
information
than
an
accusatorial
approach
and
leads
to
more
diagnostic
impressions
by
third
party
observers.
©
2013
Society
for
Applied
Research
in
Memory
and
Cognition.
Published
by
Elsevier
Inc.
All
rights
reserved.
1.
Obtaining
guilty
knowledge
in
human
intelligence
interrogations:
comparing
accusatorial
and
information-gathering
approaches
with
a
novel
experimental
paradigm
The
extant
research
literature
on
interrogations
in
the
law
enforcement
context
(henceforth,
forensic
interrogations),
primar-
ily
driven
by
the
many
confirmed
cases
of
false
confessions
over
the
last
20
years
(see
www.innocenceproject.org),
has
developed
substantially
since
the
early
1990s.
In
addition
to
case
studies
and
observational
studies
(e.g.,
Drizin
&
Leo,
2004;
Leo,
2008),
the
research
investigating
forensic
interrogations
includes
experimen-
tal
studies
evaluating
interrogation
techniques
that
may
produce
夽This
research
was
funded
by
a
Department
of
Defense
grant
awarded
to
Christian
Meissner
at
the
University
of
Texas
at
El
Paso
and
Melissa
Russano
at
Roger
Williams
University.
Statements
of
fact,
opinion,
and
analysis
in
the
paper
are
those
of
the
authors
and
do
not
reflect
the
official
policy
or
position
of
the
Department
of
Defense
or
the
U.S.
Government.
∗Corresponding
author
at:
University
of
Texas
at
Tyler,
Department
of
Psychology
&
Counseling,
3900
University
Boulevard,
Tyler,
TX
75799,
United
States.
Tel.:
+1
903
566
7140.
E-mail
addresses:
jrevans@uttyler.com,
jacki.evans@gmail.com
(J.R.
Evans).
true
and
false
confessions
(e.g.,
Kassin
&
Kiechel,
1996;
Russano,
Meissner,
Narchet,
&
Kassin,
2005)
and
individual
difference
factors
that
make
certain
people
more
susceptible
to
interrogative
pres-
sure
(e.g.,
Redlich
&
Goodman,
2003;
for
reviews
of
this
research,
see
Kassin
et
al.,
2010;
Kassin
&
Gudjonsson,
2004;
Lassiter
&
Meissner,
2010).
In
contrast
to
our
developing
knowledge
on
forensic
interroga-
tions,
there
is
a
distinct
lack
of
available
research
on
information
gathering
in
a
human
intelligence
(HUMINT)
context
(Evans,
Meissner,
Brandon,
Russano,
&
Kleinman,
2010;
Justice,
Bhatt,
Brandon,
&
Kleinman,
2008).
Although
one
might
expect
much
of
the
available
research
on
forensic
interrogations
to
be
applicable
to
the
HUMINT
setting,
there
are
several
distinctions
between
the
two
contexts
that
may
limit
its
applicability.
One
key
distinction
relates
to
outcome
metrics
-
while
forensic
interrogations
are
engi-
neered
to
secure
a
confession
(Leo,
2008),
HUMINT
interrogations
seek
actionable
information
relating
to
an
event
or
organization
(Evans
et
al.,
2010).
Furthermore,
while
forensic
interrogations
generally
target
individuals
suspected
of
having
committed
illegal
acts,
the
targets
of
HUMINT
interrogations
range
from
high-value
individuals
responsible
for
(potential)
terrorist
incidents
to
indi-
viduals
not
directly
involved
in
any
wrongdoing
but
possibly
possessing
knowledge
of
critical
information.
Given
such
important
2211-3681/$
–
see
front
matter
©
2013
Society
for
Applied
Research
in
Memory
and
Cognition.
Published
by
Elsevier
Inc.
All
rights
reserved.
http://dx.doi.org/10.1016/j.jarmac.2013.03.002

Author's personal copy
84
J.R.
Evans
et
al.
/
Journal
of
Applied
Research
in
Memory
and
Cognition
2
(2013)
83–88
differences,
the
extent
to
which
research
in
forensic
interroga-
tion
contexts
translates
to
HUMINT
contexts
requires
empirical
investigation.
Furthermore
although
headlines
about
controversial
interrogation
methods
in
HUMINT
settings
have
stirred
debate
in
the
United
States,
there
is
currently
no
systematic
experimental
research
on
interrogations
in
this
context
to
support
the
use
of
specific
ethical,
evidence-based
approaches.
Experimental
tests
of
forensic
interrogation
techniques
have
focused
on
those
believed
to
lead
to
false
confessions,
such
as
the
presentation
of
false
evidence
(Kassin
&
Kiechel,
1996;
Perillo
&
Kassin,
2011),
minimization
and
maximization
approaches
(Klaver,
Lee,
&
Rose,
2008;
Russano
et
al.,
2005),
and
the
role
of
perceived
consequences
associated
with
confession
(Horgan,
Russano,
Meissner,
&
Evans,
2012;
Horselenberg,
Merckelbach,
&
Josephs,
2003).
The
majority
of
experimental
research
in
the
area
of
interrogations
has
employed
one
of
two
laboratory
paradigms
(or
slight
variations
thereof).
In
Kassin
and
Kiechel’s
(1996)
proce-
dure,
participants
are
falsely
accused
of
hitting
a
“forbidden”
key
during
a
typing
task.
As
this
procedure
allowed
for
the
examination
of
false
(but
not
true)
confessions,
Russano
et
al.
(2005)
developed
a
paradigm
that
incorporates
a
guilt-innocence
manipulation.
Par-
ticipants
are
set
up
by
a
confederate
to
be
innocent
or
guilty
of
sharing
information
on
a
task
and
are
later
accused
of
academic
dishonesty
by
the
experimenter.
Thus,
the
Russano
et
al.
paradigm
permits
the
examination
of
both
true
and
false
confessions
within
a
context
high
in
psychological
realism.
However,
in
both
of
these
paradigms,
the
primary
outcome
measure
is
provision
of
a
con-
fession
statement,
limiting
the
applicability
of
research
with
this
paradigm
to
HUMINT
contexts.
1.1.
Accusatorial
versus
information-gathering
interrogation
approaches
Recent
research
regarding
forensic
interrogations
has
assessed
the
relative
effectiveness
of
accusatorial
methods
(used
in
the
United
States)
and
information-gathering
methods
(developed
in
the
United
Kingdom)
in
eliciting
true
and
false
confessions
(Meissner,
Hartwig,
&
Russano,
2010a;
Meissner,
Russano,
&
Narchet,
2010b;
Meissner,
Redlich,
Bhatt,
&
Brandon,
2012).
Accusatorial
methods
are
often
employed
when
the
interrogator
believes
the
suspect
to
be
guilty.
Thus,
they
are
confession-oriented
and
designed
to
manipulate
suspects’
perceptions
regarding
the
consequences
of
confession
and
develop
themes
that
encourage
confession.
In
contrast,
over
the
last
several
decades
law
enforce-
ment
in
the
UK
have
moved
to
a
less
confrontational
(Kassin
et
al.,
2010)
information-gathering
approach
(Shawyer,
Milne,
&
Bull,
2009)
in
which
investigators
take
a
more
neutral
role
by
probing
suspects
using
open-ended
questions
and
developing
conversa-
tional
rapport
with
the
goal
of
truth-seeking
(Milne
&
Bull,
1999).
Meissner
et
al.
(2012)
highlight
several
key
differences
between
the
two
approaches.
Information-gathering
methods
seek
to
establish
rapport,
use
direct
positive
confrontation,
use
open,
exploratory
questions,
and
aim
to
obtain
information.
In
contrast,
accusatorial
methods
seek
to
establish
control,
use
psychological
manipulation,
use
closed/confirmatory
questions,
and
aim
to
obtain
a
confession.
Meta-analytic
comparisons
of
these
two
techniques
in
field
studies
(Meissner
et
al.,
2012)
indicated
that
both
accusatorial
and
information-gathering
approaches
lead
to
a
greater
fre-
quency
of
confessions
(versus
control).
However,
the
veracity
of
the
confessions
cannot
be
determined
in
such
studies.
In
con-
trast,
meta-analytic
comparisons
in
laboratory
studies
(Meissner
et
al.,
2012)
suggested
that
(a)
accusatorial
interrogations
(versus
control)
lead
to
more
true
and
false
confessions,
(b)
information-
gathering
interrogations
(versus
control)
lead
to
more
true
but
not
false
confessions,
and
most
importantly,
(c)
information-
gathering
interrogations
(versus
accusatorial)
lead
to
more
true
confessions
and
fewer
false
confessions.
However,
Meissner
et
al.
highlighted
the
need
for
additional
research
comparing
these
tech-
niques,
–
only
five
field
studies
and
12
experimental
studies
were
found
that
met
the
selection
criteria
to
be
included
in
the
meta-
analysis,
and
comparisons
with
an
information
gathering
condition
included
only
three
studies.
The
current
study
seeks
to
add
to
this
research
literature
by
experimentally
comparing
accusatorial
and
information-gathering
approaches,
with
a
focus
on
measures
of
information
gain
in
a
HUMINT-relevant
paradigm
(as
opposed
to
confession
rates,
as
examined
in
previous
studies).
1.2.
The
current
study
A
novel
experimental
paradigm
(described
in
detail
below)
designed
to
model
the
contextual
and
psychological
factors
rel-
evant
to
the
HUMINT
interrogation
context
was
used
to
test
the
relative
efficacy
of
accusatorial
and
information-gathering
interrogative
approaches.
A
critical
element
of
this
paradigm
is
the
relatively
elaborate
transgression
(as
compared
to
previous
research).
This
allowed
for
information
gain,
not
simply
con-
fession
provision,
to
be
used
as
an
outcome
metric.
Consistent
with
the
Meissner
et
al.
(2012)
meta-analysis,
we
predicted
that
information-gathering
methods
would
yield
more
relevant
infor-
mation
and
admissions
relative
to
accusatorial
methods.
In
addition,
the
current
study
evaluated
third
party
observers’
perceptions
of
interrogator
and
interviewee
behavior.
Previous
research
explored
whether
or
not
interrogator
beliefs
could
influ-
ence
the
way
a
suspect’s
behavior
is
perceived.
For
example,
Kassin,
Goldstein,
and
Savitsky
(2003)
compared
participants’
perceptions
of
mock
suspects
as
a
function
of
the
interrogator’s
expectation
of
guilt
or
innocence.
Observers
rated
mock
suspects
in
the
guilty
expectation
condition
as
more
defensive
than
those
in
the
truth
expectation
condition.
Given
that
interrogators’
beliefs
can
lead
interviewees
to
behave
in
a
way
consistent
with
their
beliefs
(see
also
Meissner
&
Kassin,
2002;
Meissner
&
Kassin,
2004),
we
hypoth-
esized
that
the
approaches
employed
by
interrogators
may
also
influence
how
interviewees
behave.
For
example,
someone
who
is
confronted
in
an
accusatorial,
guilt-presumptive
manner
may
react
to
that
treatment
in
a
way
that
is
interpreted
as
deceptive
or
guilty
(e.g.,
become
anxious,
less
cooperative),
regardless
of
actual
guilt.
Thus,
we
predicted
that
behavioral
judgments
for
interroga-
tions
conducted
in
an
information-gathering
style
would
produce
judgments
more
diagnostic
of
“guilt”
when
compared
with
inter-
rogations
conducted
in
an
accusatorial
style.
2.
Method
2.1.
Participants
2.1.1.
Interrogation
phase
Psychology
undergraduates
(n
=
103)
were
recruited
to
partic-
ipate
in
a
study
described
as
a
“Trivia
Challenge”
in
exchange
for
research
credit
in
a
psychology
course
and
the
opportunity
to
earn
up
to
$10.
The
majority
of
the
participants
were
female
(58%)
and
Hispanic
(88%),
with
a
mean
(and
median)
age
of
19.
2.1.2.
Observation
phase
Psychology
undergraduates
(n
=
126)
were
recruited
and
par-
ticipated
in
exchange
for
research
credit.
The
sample
was
mostly
female
(55%)
and
Hispanic
(79%),
with
a
mean
age
of
21(median
=
19).
2.2.
Design
and
procedure
A
2
(interrogation
approach:
information-gathering
vs.
accusatorial)
×
2
(culpability:
guilty
vs.
innocent)
between-
participants
factorial
design
was
employed,
with
participants

Author's personal copy
J.R.
Evans
et
al.
/
Journal
of
Applied
Research
in
Memory
and
Cognition
2
(2013)
83–88
85
randomly
assigned
to
condition.
The
procedure
was
adapted
from
Russano
et
al.
(2005),
aiming
to
capture
critical
elements
of
a
HUMINT
interrogation
with
a
detained
individual
believed
to
have
knowledge
of
events
of
interest,
but
with
unknown
personal
involvement.
This
procedure
allows
the
manipulation
of
whether
or
not
an
interviewee
possesses
relevant
knowledge,
and
it
assesses
outcome
metrics
related
to
information
gain.
Further,
the
paradigm
maintains
a
high
degree
of
psychological
realism,
with
participants’
believing
they
are
being
questioned
about
their
potential
involvement
in
a
deliberate
transgression.
A
female
confederate
and
the
participant
(hereafter
referred
to
as
“the
participants”)
arrived
at
the
lab
and
were
instructed
to
leave
all
belongings,
including
cell
phones,
in
the
main
lab
area.
The
participants
were
then
seated
in
a
small
windowless
room
discretely
wired
for
audio
and
video
recording.
The
participants
were
informed
that:
the
study
was
a
national
assessment
of
college
students’
knowledge;
they
would
be
asked
to
answer
20
general
knowledge
questions
as
a
pair;
they
would
each
receive
50
cents
for
every
correctly
answered
question;
and
they
should
complete
the
questions
together
using
no
other
resources.
Note
that
the
use
of
deception
regarding
the
study’s
true
purpose
was
critical
to
achieve
an
appropriate
level
of
psychological
realism.
The
study’s
true
nature
was
fully
revealed
during
the
debriefing
stage.
After
signing
a
consent
form
and
providing
demographic
information,
the
participants
received
the
20-item
test
(e.g.,
What
is
the
capital
of
Hungary?),
and
they
were
left
alone
in
the
room
to
complete
the
task.
2.3.
Culpability
manipulation
2.3.1.
Guilty
participants
The
confederate’s
script
began
with
her
explaining
that
her
roommate,
Maria,
had
already
completed
the
study.
She
then
retrieved
a
paper
from
her
pocket
with
20
handwritten
questions
and
answers,
which
Maria
had
allegedly
provided.
She
used
this
“cheat
sheet”
to
answer
10
of
the
assigned
questions
and
noted
that
the
remaining
10
questions
did
not
correspond
with
the
copied
questions.
Upon
realizing
the
cheat
sheet
did
not
answer
all
of
the
questions,
the
confederate
called
“Maria”
using
the
cell
phone
in
her
pocket
and
received
answers
to
two
questions
after
Maria
“googled”
them.
Finally,
the
confederate
copied
down
the
questions
missing
from
the
cheat
sheet
for
her
friend,
Alex,
who
she
claimed
would
be
completing
the
study
the
following
week.
The
partici-
pants
ultimately
completed
the
questions
and
submitted
them
to
the
experimenter.
2.3.2.
Innocent
participants
The
confederate
followed
a
similar
script
in
which
she
indicated
Maria
had
said
that
the
study
was
“easy”.
When
the
questions
were
more
challenging
than
expected,
the
confederate
retrieved
her
cell
phone
and
called
Maria
to
complain
that
the
questions
were
unex-
pectedly
difficult.
She
neither
requested,
nor
received,
any
answers
to
the
questions.
Thus,
she
did
not
cheat,
but
she
did
break
an
experimental
rule
by
having,
and
using,
her
cell
phone.
After
the
phone
call,
the
confederate
mentioned
that
her
friend
Alex
would
be
unhappy
the
study
was
so
difficult.
2.3.3.
Interrogation
phase
After
completing
the
questions,
the
participants
were
separated,
purportedly
so
that
an
experimenter
could
review
their
answers
with
them
individually.
The
confederate
was
removed,
and
the
participant
was
told
to
wait
until
someone
returned.
After
five
minutes,
the
interrogator
entered
the
room
and
began
a
scripted
interrogation
of
the
participant.
All
interrogation
sessions
were
covertly
audio
and
video-recorded.
The
interrogator
explained
that
because
so
many
questions
were
answered
correctly,
s/he
became
suspicious
that
cheating
had
occurred.
The
interrogator
explained
that
the
professor
in
charge
of
the
study
instructed
him/her
to
speak
with
both
participants
to
determine
what
happened,
and
indicated
that
if
any
cheating
had
taken
place
the
professor
might
consider
it
a
case
of
academic
dishonesty.
S/he
also
explained
that
after
talking
to
the
participant,
s/he
would
talk
to
the
other
participant.
At
this
point
the
scripts
differed
depending
upon
the
assigned
condition.
All
interrogators
were
kept
blind
to
the
participant’s
culpability
condition
and
to
the
hypothesized
influence
of
the
interrogation
methods.
2.3.4.
Accusatorial
script
The
interrogator
used
a
combination
of
maximizing
and
fear-
based
approaches
(e.g.,
“I
want
you
to
know
that
it’s
not
looking
good
for
you
guys.
This
has
the
potential
to
be
a
really
big
deal
since
this
is
a
national
project.
.
.”)
and
comforting,
emotionally
manipu-
lative,
minimizing
approaches
(e.g.,
“.
.
.if
you
are
honest,
that
will
probably
go
a
long
way,
and
show
my
professor
that
you
feel
bad
about
what
happened.
That’s
what
I
would
do.”).
This
approach
was
designed
to
play
upon
the
participants’
anxiety
and
manipulate
their
perceptions
of
the
consequences
of
admitting
versus
denying
that
rule-breaking
took
place.
2.3.5.
Information-gathering
script
The
interrogator
took
a
truth-seeking,
information-gathering
approach
(e.g.,
“Please
take
a
minute
to
try
and
remember
exactly
what
happened.
.
.
I
really
need
to
get
to
the
bottom
of
this.
.
.”)
and
employed
elements
of
the
cognitive
interview
(Fisher
&
Geiselman,
1992).
For
example,
the
rules
of
the
questioning
were
laid
out
(e.g.,
“I’m
going
to
try
not
to
interrupt
you,
since
you’re
the
one
with
the
information,
not
me.”).
After
soliciting
information
from
the
par-
ticipants
via
free
recall,
interrogators
asked
participants
to
explain
what
happened
from
the
confederate’s
perspective.
In
contrast
to
the
accusatorial
script,
this
script
was
designed
to
be
cognitively
challenging,
making
it
more
difficult
to
lie
and
also
increasing
the
number
of
details
likely
to
be
recalled.
2.3.6.
Closing
the
interrogation
At
the
end
of
the
interrogators’
scripted
monologues,
all
partic-
ipants
were
asked
“So,
did
you
guys
cheat?”
This
was
followed
by
four
opened-ended
questions
designed
to
obtain
as
many
details
as
possible
(e.g.,
“Why
don’t
you
tell
me
the
whole
story,”
“Is
there
anything
else
you
can
tell
me?”).
The
last
of
these
was
“Okay,
I’m
going
to
go
and
speak
with
the
other
person
now.
Is
there
anything
else
you’d
like
to
add
to
what
you’ve
said
so
far?”
The
interrogator
then
thanked
the
participant
and
left
the
room.
2.4.
Debriefing
Following
the
interrogation,
the
experimenter
entered
the
room
immediately
to
debrief
the
participant.
For
all
participants,
this
involved
(a)
probing
for
suspicion,
(b)
explaining
the
deception
involved
in
the
study
and
that
the
session
had
been
recorded,
(c)
asking
the
participant
to
complete
a
debriefing
questionnaire
regarding
his/her
experiences
in
the
study,
(d)
offering
the
par-
ticipant
the
option
to
sign
a
new
consent
form
permitting
the
researchers
to
use
their
data
and
recorded
interview,
(e)
discussing
the
importance
of
keeping
the
procedures
and
deception
employed
by
the
study
confidential,
and
(f)
paying
participants
$10
for
com-
pleting
the
study.
The
experimenter’s
goal
during
debriefing
was
to
ensure
that
the
participants
understood
that
they
were
not
in
any
trouble,
were
comfortable
with
what
happened,
and
understood
why
the
use
of
deception
was
necessary.

Author's personal copy
86
J.R.
Evans
et
al.
/
Journal
of
Applied
Research
in
Memory
and
Cognition
2
(2013)
83–88
Table
1
Information
gain
and
admissions
as
a
function
of
culpability
and
interrogation
approach.
Condition
Number
Interrogation
Proportion
admit
Proportion
of
Length
Confederate
Admissions
details
(0–15)
in
seconds
had
phone
(guilty
only)
Approach
Culpability
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Info-
gath.
Innocent
3.08
2.78
212.32
112.68
0.64
0.49
Guilty
5.70
1.95
202.57
173.71
0.95
0.22
0.64*
0.30
Total
4.22*
2.76
208.53*
137.29
0.78*
0.42
Accusatorial Innocent
2.25
2.49
112.63
74.47
0.50
0.51
Guilty
4.41
2.51
107.50
34.32
0.58
0.50
0.45*
0.32
Total
3.43*
2.71
109.87*
55.95
0.55*
0.50
Total Innocent
2.67ˆ 2.65 160.30 106.28
0.57
0.50
Guilty
4.94ˆ
2.37
139.19
111.40
0.73
0.45
0.53
0.32
Total
3.81
2.75
150.23
108.62
0.65
0.48
Note:
The
interrogation
approach
main
effect
was
significant
(ps
≤
.05)
for
all
four
measures,
as
indicated
by
‘*’.
The
main
effect
of
culpability
was
significant
for
number
of
details
only
(p
<
.001),
as
indicated
by
‘ˆ
’.
No
interactions
were
significant.
2.5.
Observation
phase
A
2
(interrogation
approach:
information-gathering
vs.
accusatorial)
×
2
(culpability:
guilty
vs.
innocent)
mixed
fac-
torial
design
was
employed.
Participants
were
randomly
assigned
to
an
interrogation
approach
and
watched
videos
of
both
inter-
viewees
with
and
without
guilty
knowledge.
Participants
were
instructed
that
some,
but
not
all,
interviewees
had
guilty
knowl-
edge,
but
they
were
not
given
a
specific
ratio
of
interviewees
with
or
without
guilty
knowledge.
The
amount
of
information
the
interviewees
disclosed
to
the
interrogator
within
the
videos
was
controlled
for
so
that
the
volume
of
information
provided
was
similar
across
the
two
interrogation
conditions.
In
other
words,
we
ensured
that
it
was
not
the
case
that
the
videos
chosen
for
one
interrogation
approach
yielded
more
guilty
knowledge/admissions
than
videos
selected
for
the
second
approach.
The
videos
averaged
five
minutes
in
length,
and
participants
viewed
a
series
of
eight
videos.
The
presentation
order
was
randomized
and
participants
answered
several
questions
about
each
video
following
its
pre-
sentation.
These
questions
addressed
six
interviewee
behaviors
(nervousness,
cooperativeness,
believability,
confidence,
amount
of
information
provided,
and
effort
exerted),
and
the
amount
of
pressure
placed
on
the
interviewee
by
the
interrogator.
All
responses
were
provided
on
eight-point
Likert
scales
(1
=
“Not
at
all”/“None”
to
8
=
“Extremely”/“A
lot”).
3.
Results
All
means
and
standard
deviations
are
reported
in
Table
1.
3.1.
Information
gain
Two
measures
of
information
gain
were
examined:
quantity
of
relevant
details
and
participant
talking
time.
Both
were
analyzed
via
a
2
(guilt
vs.
innocence)
×
2
(accusatorial
vs.
information-
gathering
approach)
between-participants
ANOVA.
Quantity
of
details.
Two
coders
scored
each
interrogation
video
for
the
presence
of
15
relevant
details
per
participant
(lists
var-
ied
depending
on
culpability
condition).
All
inter-rater
reliability
indices
were
acceptable
(for
number
of
details,
r
=
.96,
p
<
.001;
for
categorical
scoring,
Cohen’s

≥
.90).
Therefore,
the
primary
coder’s
scorings
were
used
for
all
analyses.
Guilty
participants
reported
more
details
than
innocent
par-
ticipants,
F(1,94)
=
22.42,
p
<
.001,
d
=
.90,
and
participants
in
the
information-gathering
condition
reported
more
details
than
those
in
the
accusatorial
condition;
F(1,94)
=
4.39,
p
=
.039,
d
=
.29.
The
interaction
was
not
significant,
F(1,94)
=
0.20,
p
=
.653,
2
p<
.01.
Participant
talking
time
(in
seconds).
There
was
no
culpabil-
ity
effect,
F(1,84)
=
0.12,
p
=
.733,
d
=
.19;
however,
participants
in
the
information-gathering
condition
spent
more
time
talking
than
those
in
the
accusatorial
condition,
F(1,84)
=
20.05,
p
<
.001,
d
=
.94.
Again,
the
interaction
was
not
significant,
F(1,
84)
=
0.01,
p
=
.916,
2
p<
.01.
3.2.
Admissions.
Presence
of
phone
(all
participants).
We
assessed
via
logis-
tic
regression
whether
participants
admitted
to
the
presence
of
the
phone
(a
key
admission),
as
a
function
of
their
culpability,
the
interrogation
approach,
and
the
interaction
term.
The
over-
all
model
was
significant;
X2(3)
=
13.31,
p
=
.004.
An
interrogation
approach
main
effect
demonstrated
that
the
information-gathering
approach
led
to
more
admissions
than
the
accusatorial
approach,
B
=
2.60,
wald
=
5.64,
p
=
.018.
There
was
no
significant
culpability
effect
and
no
interaction,
B
=
0.35,
wald
=
0.39,
p
=
.531,
and
B
=
2.02,
wald
=
2.66,
p
=
.103,
respectively.
Cheating
(guilty
participants
only).
Finally,
coders
scored
whether
participants
admitted
to
“cheating”
elements
of
the
incident.
Specifically,
they
coded
whether
the
participant
mentioned
(a)
the
cheat
sheet,
(b)
the
other
participant
receiving
answers
over
the
phone,
and
(c)
the
other
participant
copying
the
questions.
Overall,
78%
admitted
to
the
cheat
sheet,
59%
admitted
to
getting
answers
on
the
phone
(this
is
different
from
admitting
a
call
was
made),
and
8%
admitted
that
the
other
participant
copied
the
questions
for
someone
(16%
admitted
to
nothing
at
all).
Predictably,
as
they
were
never
directly
accused
of
any
of
these
three
cheating
acts,
no
innocent
participants
made
any
false
admissions.
The
proportion
of
admissions
made
was
calculated
for
each
guilty
participant,
and
analysis
indicated
that
guilty
participants
made
more
admissions
in
the
information-gathering
condition
than
in
the
accusatorial
con-
dition,
t(50)
=
2.13,
p
=
.038,
d
=
.61.
3.3.
Observation
phase
An
average
score
was
calculated
for
each
of
the
seven
intervie-
wee
behavior
ratings,
as
well
as
the
interrogator
pressure
rating.
A
2
×
2
multivariate
analysis
of
variance
(MANOVA)
was
con-
ducted
to
determine
the
effect
of
culpability
(guilt
vs.
innocence)
and
interrogation
approach
(information-gathering
vs.
accusato-
rial)
on
observer
ratings.
The
analysis
yielded
significant
main
effects
of
both
culpability,
Wilks’

=
.42,
F(7,
118)
=
23.78,
p
<
.001,
2
p=
.60,
and
interrogation
approach,
Wilks’

=
.84,
F(7,
118)
=
3.31,
p
=
.003,
2
p=
.16.
The
interaction
between
culpability
and
interro-
gation
approach
was
also
significant,
Wilks’

=
.89,
F(7,
118)
=
2.15,

Author's personal copy
J.R.
Evans
et
al.
/
Journal
of
Applied
Research
in
Memory
and
Cognition
2
(2013)
83–88
87
Table
2
Behavioral
ratings
as
a
function
of
culpability
and
interrogation
approach.
Condition
Observer
ratings
of
behavior
Confidence
Nervousness
Int.
pressure
Approach
Culpability
Mean
SD
Mean
SD
Mean
SD
Info-
gath.
Innocent
5.77
1.18
3.54
1.12
4.06
1.38
Guilty
5.35
1.11
5.04
1.04
4.33
1.29
Total
5.56
0.96
4.29
0.89
4.20*
1.27
Accusatorial Innocent
5.74 1.11 4.05 1.40
5.20
1.50
Guilty
5.26
1.16
5.06
1.27
4.94
1.39
Total
5.50
0.96
4.55
1.21
5.07*
1.33
Total Innocent
5.76ˆ
1.14
3.79ˆ
1.29
4.63
1.55
Guilty
5.31ˆ
1.13
5.05ˆ
1.16
4.64
1.37
Total
5.53
0.96
4.42
1.07
4.64
1.37
Note:
The
interrogation
approach
main
effect
was
significant
for
interrogator
pressure
only
(p
<
.001),
as
indicated
by
‘*’.
The
main
effect
of
culpability
was
significant
for
confidence
and
nervousness
only
(p
<
.001),
as
indicated
by
‘ˆ
’.
The
interaction
effects
were
also
significant
(p
<
.05)
for
nervousness
and
interrogator
pressure.
p
=
.044,
2
p=
.11.
Univariate
ANOVAs
for
each
dependent
variable
were
conducted
as
follow-up
tests
to
the
MANOVA.
Significant
effects
were
observed
in
ratings
of
perceived
confidence,
ner-
vousness,
and
pressure
to
provide
information.
These
effects
are
discussed
below,
with
means
and
standard
deviations
presented
in
Table
2.
A
2
(culpability)
×
2
(interrogation
approach)
repeated
meas-
ures
ANOVA
revealed
that
third
party
observers
rated
innocent
interviewees
as
more
confident
than
guilty
interviewees,
F(1,
124)
=
16.68,
p
<
.001,
d
=
.40.
The
interrogation
approach
main
effect
was
not
significant;
F(1,
124)
=
0.13,
p
=
.719,
d
=
.05,
nor
was
the
interaction;
F(1,
124)
=
0.09,
p
=
.760,
2
p=
.001.
Univariate
analyses
also
revealed
that
guilty
interviewees
were
rated
as
more
nervous
than
innocent
interviewees;
F(1,
124)
=
142.20,
p
<
.001,
d
=
1.03.
The
interrogation
approach
effect
was
not
significant;
F(1,
124)
=
1.98,
p
=
.162,
d
=
.24,
but
it
did
inter-
act
with
culpability;
F(1,
124)
=
5.44,
p
=
.021,
2
p=
.04.
Post
hoc
t-tests
(critical
alpha
=
.025)
showed
that
the
difference
between
nervousness
ratings
of
the
innocent
interviewees
across
the
two
interrogation
approaches
was
significant,
with
the
innocent
inter-
viewees
being
rated
as
more
nervous
in
the
accusatorial
condition
than
the
information-gathering
condition,
t(1,
62)
=
2.27,
p
=
.025,
d
=
.41.
There
were
no
significant
differences
in
nervousness
rat-
ings
of
guilty
interviewees
across
the
two
interrogation
method
conditions,
t(1,
62)
<
1,
p
=
.924,
d
=
.02.
Finally,
univariate
analyses
also
indicated
that
greater
pres-
sure
on
the
interviewee
was
observed
during
the
accusatorial
approach
compared
to
the
information-gathering
approach
F(1,
124)
=
14.22,
p
<
.001,
d
=
.67.
The
culpability
effect
was
not
signif-
icant
F(1,
124)
=
.005,
p
=
.946,
d
<
.01,
but
there
was
a
significant
interaction,
F(1,
124)
=
8.99,
p
=
.003,
2
p=
.07.
Post
hoc
t-tests
(crit-
ical
alpha
=
.025)
showed
the
interaction
was
similar
to
that
for
nervousness
ratings.
Observers
rated
the
pressure
on
innocent
interviewees
in
the
accusatorial
condition
as
higher
than
those
in
the
information-gathering
condition;
t(1,
62)
=
4.42,
p
<
.001,
d
=
.79.
The
effect
for
guilty
participants
was
similar,
but
the
effect
was
smaller;
t(1,
62)
=
2.56,
p
=
.012,
d
=
.46.
4.
Discussion
The
present
research
assessed
interrogation
approach
effective-
ness
within
a
HUMINT
context
using
a
new
experimental
paradigm.
An
information-gathering
approach
yielded
more
critical
details
and
resulted
in
a
more
talkative
interviewee
than
an
accusatorial
interrogation
strategy.
Furthermore,
a
greater
frequency
of
admis-
sions
was
secured
during
information-gathering
interrogations
relative
to
accusatorial
interrogations.
Finally,
results
indicate
the
information-gathering
approach
significantly
reduces
perceived
nervousness
on
the
part
of
innocent
participants
when
compared
with
the
accusatorial
approach.
In
addition,
innocent
participants
were
perceived
to
have
endured
significantly
less
pressure
from
the
interrogator
under
the
information-gathering,
as
opposed
to
the
accusatorial,
approach.
4.1.
Practical
applications
The
current
study
replicated
the
previously
found
superior-
ity
of
information-gathering
approaches
(Meissner
et
al.,
2012)
using
a
different
metric:
information
gain
as
opposed
to
con-
fession
diagnosticity.
This
suggests
that
information-gathering
approaches
may
be
effective
within
HUMINT
contexts
as
well
as
law
enforcement
settings.
While
law
enforcement
agencies
in
the
U.K.
currently
use
this
approach
(i.e.,
the
PEACE
model;
Shawyer
et
al.,
2009),
U.S.
law
enforcement
agencies
have
gener-
ally
continued
to
rely
upon
accusatorial
approaches
(Gudjonsson,
2003).
Additional
research
comparing
these
two
approaches
is
needed,
and
we
are
hopeful
that
others
will
seek
to
test
the
effec-
tiveness
of
the
information-gathering
approach
in
both
forensic
and
HUMINT
contexts,
and
in
both
field
and
laboratory
set-
tings.
Similar
to
findings
regarding
interrogator
expectations
(Kassin
et
al.,
2003),
our
manipulation
of
interrogation
approach
influ-
enced
how
nervous
the
innocent
interviewees
were
perceived
to
be,
but
not
how
nervous
the
guilty
interviewees
were
perceived
to
be.
The
same
pattern
was
found
for
impressions
of
interrogator
pressure.
Specifically,
innocent
interviewees
in
the
information-
gathering
condition
were
perceived
as
less
nervous,
and
their
interrogators
were
perceived
as
applying
less
pressure,
compared
to
the
accusatorial
method.
In
contrast,
guilty
interviewees
were
perceived
as
equally
nervous
under
information-gathering
condi-
tions
as
accusatorial
conditions,
and
the
difference
in
interrogator
pressure
across
interrogation
approaches
was
smaller
for
guilty
interviewees.
These
findings
may
help
to
explain
previous
findings
that
innocent
individuals
are
less
likely
to
provide
false
admis-
sions
under
information-gathering
interrogation
approaches
than
accusatorial
approaches
(Meissner
et
al.,
2012).
This
pattern
of
findings
suggests
there
may
be
different
mech-
anisms
at
work
for
guilty
versus
innocent
individuals,
such
that
anxiety
for
guilty
participants
may
result
from
their
transgres-
sion,
whereas
anxiety
for
the
innocent
participants
are
more
likely
to
result
from
the
interrogation
itself
(see
Narchet,
Meissner,
&
Russano,
2011).
As
the
observers
were
blind
to
the
culpability
condition
of
the
interviewees
and
unaware
of
the
interroga-
tion
approach
manipulation,
these
findings
suggest
that
the
information-gathering
approach,
compared
to
the
accusatorial
approach,
may
create
greater
observable
differences
between
the

Author's personal copy
88
J.R.
Evans
et
al.
/
Journal
of
Applied
Research
in
Memory
and
Cognition
2
(2013)
83–88
behavior
of
guilty
and
innocent
interviewees,
as
well
as
the
behav-
ior
of
the
interrogator.
4.2.
Future
directions
An
important
avenue
for
future
research
is
to
modify
the
paradigm
to
investigate
the
provision
of
false
information.
The
current
procedure
was
specifically
developed
to
assess
informa-
tion
gain
under
optimal
conditions
for
accurate
memory
reporting
(no
leading
questions,
no
cued
recall,
no
ambiguous
events).
For
this
reason
false
information
was
not
provided
by
our
participants.
However,
the
paradigm
can
be
easily
modified
to
increase
the
likeli-
hood
of
obtaining
false
information
for
those
researchers
interested
in
information
quality.
As
with
any
study
taking
a
first
step
into
a
new
field,
our
study
has
caveats.
Chief
among
these
is
that
no
laboratory-based
experiment
can
perfectly
simulate
a
real
HUMINT
interrogation.
In
addition
to
the
relatively
low
stakes
used
here,
our
participants
–
mostly
Hispanic,
female,
students
attending
a
Univeristy
in
the
United
States
–
are
not
representative
of
the
larger
population
of
people
who
are
the
subjects
of
intelligence
interrogations
world-
wide.
However,
at
this
time,
we
have
little
reason
to
believe
that
the
underlying
processes
of
interrogation
(and
resultant
effects)
would
differ
systematically
for
our
sample
versus
that
of
a
more
representative
sample.
Nonetheless,
this
is
an
empirical
question
ripe
for
future
research,
and
we
hope
that
future
research
may
improve
upon
our
work
by
creating
more
applicable
scenarios
(e.g.,
considering
the
role
of
language,
culture,
ethnicity)
with
more
rep-
resentative
samples.
This
paradigm
and
the
above
areas
of
future
research
are
in
line
with
our
goal
of
advising
relevant
parties
regarding
inter-
rogation
techniques
that
will
be
effective,
bringing
together
our
knowledge
of
the
psychology
of
interrogations
with
practitioners’
knowledge
of
interrogation
technique
implementation
in
the
field.
It
is
more
useful
to
approach
the
issue
of
interrogations
in
this
manner,
as
opposed
to
simply
informing
practitioners’
regarding
which
techniques
should
be
avoided
(see
Meissner,
Hartwig,
et
al.,
2010a;
Meissner,
Russano,
et
al.,
2010b).
We
have
taken
a
first
step
in
researching
a
HUMINT
context
within
a
laboratory
set-
ting.
Given
current
military
and
intelligence
operations
aimed
at
combating
terrorism,
we
believe
it
is
imperative
that
researchers
identify
effective,
evidence-based
methods
for
collecting
human
intelligence
with
non-cooperative
individuals.
For
this
reason,
we
hope
to
see
researchers
tackle
this
problem
with
as
much
enthu-
siasm
and
creativity
as
they
have
applied
to
the
problem
of
false
confessions.
References
Drizin,
S.
A.,
&
Leo,
R.
A.
(2004).
The
problem
of
false
confessions
in
the
post-DNA
world.
North
Carolina
Law
Review,
82,
894–1007.
Evans,
J.
R.,
Meissner,
C.
A.,
Brandon,
S.
E.,
Russano,
M.
B.,
&
Kleinman,
S.
M.
(2010).
Criminal
versus
HUMINT
interrogations:
The
importance
of
psychological
sci-
ence
to
improving
interrogative
practice.
Journal
of
Psychiatry
&
Law,
38(1–2),
215–249.
Fisher,
R.
P.,
&
Geiselman,
R.
E.
(1992).
Memory-enhancing
techniques
for
investigative
interviewing:
The
cognitive
interview.
Springfield,
IL:
Charles
C.
Thomas.
Gudjonsson,
G.
H.
(2003).
The
psychology
of
interrogations
and
confessions:
A
handbook.
Chichester,
England:
John
Wiley
&
Sons.
http://dx.doi.org/10.
1002/9780470713297
Horgan,
A.
J.,
Russano,
M.
B.,
Meissner,
C.
A.,
&
Evans,
J.
R.
(2012).
Minimi-
zation
and
maximization
techniques:
Assessing
the
perceived
consequences
of
confessing
and
confession
diagnosticity.
Psychology,
Crime
&
Law,
18,
65–78.
http://dx.doi.org/10.1080/1068316X.2011.561801
Horselenberg,
R.,
Merckelbach,
H.,
&
Josephs,
S.
(2003).
Individual
differences
and
false
confessions:
A
conceptual
replication
of
Kassin
and
Kiechel
(1996).
Psy-
chology
Crime
&
Law,
9,
1–8.
http://dx.doi.org/10.1080/10683160308141
Justice,
B.
P.,
Bhatt,
S.,
Brandon,
S.
E.,
&
Kleinman,
S.
M.
(2008).
An
analysis
of
scientific
support
for
Army
Field
Manual
2-22.3
interrogation
methods.
Washington,
DC:
DCHC/DoD.
Kassin,
S.
M.,
Drizin,
S.
A.,
Grisso,
T.,
Gudjonsson,
G.
H.,
Leo,
R.
A.,
&
Redlich,
A.
D.
(2010).
Police-induced
confessions:
Risk
factors
and
recommendations.
Law
&
Human
Behavior,
34,
3–38.
http://dx.doi.org/10.1007/s10979-009-9188-6
Kassin,
S.
M.,
Goldstein,
C.
C.,
&
Savitsky,
K.
(2003).
Behavioral
confirmation
in
the
interrogation
room:
On
the
dangers
of
presuming
guilt.
Law
and
Human
Behavior,
27,
187–203.
http://dx.doi.org/10.1023/A:1022599230598
Kassin,
S.
M.,
&
Gudjonsson,
G.
H.
(2004).
The
psychology
of
confessions:
A
review
of
the
literature
and
issues.
Psychological
Science
in
the
Public
Interest,
5,
33–67.
http://dx.doi.org/10.1111/j.1529-1006.2004.00016.x
Kassin,
S.
M.,
&
Kiechel,
K.
L.
(1996).
The
social
psychology
of
false
confessions:
Com-
pliance,
internalization,
and
confabulation.
Psychological
Science,
7,
125–128.
Klaver,
J.
R.,
Lee,
Z.,
&
Rose,
V.
(2008).
Effects
of
personality,
interrogation
techniques
and
plausibility
in
an
experimental
false
confession
paradigm.
Legal
and
Crimi-
nological
Psychology,
13,
71–88.
http://dx.doi.org/10.1348/135532507X193051
Lassiter,
G.
D.,
&
Meissner,
C.
A.
(2010).
Police
interrogations
and
false
confessions:
Cur-
rent
research,
practice
and
policy
recommendations.
Washington,
DC:
American
Psychological
Association.
http://dx.doi.org/10.1037/12085-90
Leo,
R.
A.
(2008).
Police
interrogation
and
American
justice.
Cambridge,
MA:
Harvard
University
Press.
Meissner,
C.
A.,
Hartwig,
M.,
&
Russano,
M.
B.
(2010).
The
need
for
a
positive
psychological
approach
and
collaborative
effort
for
improving
practice
in
the
interrogation
room.
Law
&
Human
Behavior,
34,
43–45.
http://dx.doi.org/10.
1007/s10979-009-9205-09
Meissner,
C.
A.,
&
Kassin,
S.
M.
(2002).
He’s
guilty!:
Investigator
bias
in
judgments
of
truth
and
deception.
Law
&
Human
Behavior,
26,
469–480.
http://dx.doi.org/10.1023/A:1020278620751
Meissner,
C.
A.,
&
Kassin,
S.
M.
(2004).
You’re
guilty
so
just
confess!
Cognitive
and
behavioral
confirmation
biases
in
the
interrogation
room.
In
G.
D.
Lassiter
(Ed.),
Interrogations,
confessions,
and
entrapment:
Perspectives
in
law
&
psychology
(pp.
85–106).
New
York:
Klumer
Academic/Plenum
Publishers.
Meissner,
C.
A.,
Redlich,
A.
D.,
Bhatt,
S.,
&
Brandon,
S.
E.
(2012).
Interview
and
interrogation
methods
and
their
effects
on
true
and
false
confessions.
Campbell
Systematic
Reviews,
13,
1–53.
http://dx.doi.org/10.4073/csr
2012.13
Meissner,
C.
A.,
Russano,
M.
B.,
&
Narchet,
F.
M.
(2010).
The
importance
of
a
labora-
tory
science
for
improving
the
diagnostic
value
of
confession
evidence.
In
G.
D.
Lassiter,
&
C.
A.
Meissner
(Eds.),
Police
interrogations
and
false
confessions:
Cur-
rent
research,
practice
and
policy
recommendations.
Washington,
DC:
American
Psychological
Association.
Milne,
R.,
&
Bull,
R.
(1999).
Investigative
interviewing:
Psychology
and
practice.
Chich-
ester,
England:
Wiley.
Narchet,
F.
M.,
Meissner,
C.
A.,
&
Russano,
M.
B.
(2011).
Modeling
the
influence
of
investigator
bias
on
the
elicitation
of
true
and
false
confessions.
Law
and
Human
Behavior,
35,
452–465.
http://dx.doi.org/10.1007/s10979-010-9257-60
Perillo,
J.
T.,
&
Kassin,
S.
M.
(2011).
Inside
interrogation:
The
lie,
the
bluff,
and
false
confessions.
Law
and
Human
Behavior,
35,
327–337.
http://dx.doi.org/10.1007/s10979-010-9244-52
Redlich,
A.
D.,
&
Goodman,
G.
S.
(2003).
Taking
responsibility
for
an
act
not
commit-
ted:
The
influence
of
age
and
suggestibility.
Law
&
Human
Behavior,
27,
141–156.
http://dx.doi.org/10.1023/A:1022543012851
Russano,
M.
B.,
Meissner,
C.
M.,
Narchet,
F.
M.,
&
Kassin,
S.
K.
(2005).
Investigating
true
and
false
confessions
within
a
novel
experimental
paradigm.
Psychological
Science,
16,
481–486.
http://dx.doi.org/10.1111/j.0956-7976.2005.01560.x
Shawyer,
A.,
Milne,
B.,
&
Bull,
R.
(2009).
Investigative
interviewing
in
the
UK.
In
T.
Williamson,
B.
Milne,
&
S.
Savage
(Eds.),
International
developments
in
investiga-
tive
interviewing.
Devon,
UK:
Willan
Publishing.