The Distribution of Crime Victimisation in the Population
| Author | Alan Trickett,Tim Hope |
| DOI | 10.1177/026975800801500103 |
| Published date | 01 January 2008 |
| Date | 01 January 2008 |
International
Review
ofVictimology. 2008,
Vol.
15,
pp.
37-58
il:l
A B
Academic
Publishers
-
Printed
in
Great
Britain
THE DISTRIBUTION OF CRIME VICTIMISATION IN
THE POPULATION
TIM HOPE and ALAN TRICKETT*
Keele
University,
UK
and
University
of
Manchester,
UK
ABSTRACT
This
paper
addresses
the
question
of
how
best
to
determine
the
appropriate
theoretical
model
for
explaining
the
frequency
distribution
typically
observed
in
self-report
crime
victimisation
surveys
of
general
adult
household
populations.
The
contemporary,
prevailing
approach
is
characterised
as
a
'double-hurdle
model'
of
exposure
to
victimisation
risk
that
focuses,
separately,
upon
the
transition
initially
from
a
non-victim
to
a
victim
state
(the
'lifestyle-exposure'
hypothesis),
and
thence
upon
the
transition
to
a
subsequent,
specific
level
of
risk
(the
'repeat
victimisation
hypothesis').
An
alternative
model
-
the
'immunity
hypothesis'
-
is
proposed
with
the
aim
of
addressing
some
of
the
theoretical
and
empirical
difficulties
identified
in
the
current
approach.
This
model
takes
the
form
of a
compound-Poisson
generalisation
of
the
Negative
Binomial
statistical
distribution.
Its
chief
difference
from
the
current
approach
is
its
assumption
of a
general
tendency
in
the
population
towards
'immunity
from'
rather
than
'exposure
to'
crime
victimisation
risk.
An
outflow
table
of
data
on
household
property
crime
victimisation
from
a
longitudinal
panel
survey
is
analysed.
The
results
provide
support
for
the
hypotheses
derived
from
the
immunity
model.
Keywords:
Household
Crime
Victimisation.
Repeat
Victimisation.
Double
Hurdle
Model.
Immunity
Model.
Negative
Binomial
Distribution.
Crime
victimisation
surveys
of
general
populations typically ask their
respondents
to
report
the
numbers
of incidents of a
set
of
offences
that
they
or
their
households
have
experienced
over
a particular
recall
period (usually
no
more
than
the
previous
12
months).
Naturally, tabulations of
responses
to
such
questions
yield
frequency
distributions
for
the
number
of
crime
victimisations
experienced
by
the
sampled
respondents.
Typically
also,
those
experiencing
one
or
more
incidents
over
the
recall
period
are
usually
considered
to
be
'victims',
•
Tim
Hope
is
Professor
of
Criminology
at
Keele
University.
Corresponding
address:
School
of
Sociology
and
Criminology,
Keele
University, Staffordshire
ST5
580,
UK
(t.j.hope@crim.keele.ac.uk).
Alan
Trickett
was
formerly
at
the
School
of
Economic
Studies, University
of
Manchester,
UK.
This
article
was
previously published
in
French
as
Hope,
T.
and
Trickett, A.
(2004).
La Distribution
de
Ia
Victimisation
dans
Ia
population.
Deviance
et
Societe,
28 (3), 385-404.
The
authors gratefully acknowledge
the
permission
of
the
Editors
of
Deviance
et
Societe
to
reprint
the
English
language
manuscript
version,
and
the
invitation
and
editorial contribution
of
Philippe Robert
and
Renee
Zauberman,
the
editors
of
the
Special Edition
(Un
autre
regard
sur
Ia
delinquance)
in
which
it first
appeared.
We
are
especially grateful
to
Renee
Zauberman
for
the
French
translation, particularly
of
'le
modele
du
double
saut'.
38
and
thus
distinguished
from
non-victims,
while
persons
reporting
more
than
one
incident
have
come
to
be
termed
'multiple
victims'
(Hope
et
a/.,
2001)
or,
following
Farrell
and
Pease
(1993},
'repeat
victims'.
Because
the
surveys
are
samples
of
the
general
population,
so
these
categories
have
also
come
to
be
thought
of
as
representative
of
the
crime
victimisation
experiences
ofthe
general
population
1•
Since
the
surveys
also
ask
a
range
of
other
questions
about
respondents'
backgrounds,
experiences
and
so
on,
these
questions
have
been
correlated
with
crime
victimisation
frequencies
and,
in
turn,
have
been
considered
as
causal
antecedents,
or
at
least a priori 'risk
factors'
2,
to
victimisation
events;
and
used
to
support,
substantiate
or
test
various
explanatory
theories
of
crime
victimisation.
The
principal
purpose
has
been
to
investigate
the
reasons
for
the
probability
of
respondents
falling
into
one
of
the
victim-categories
and,
especially,
to
discover
what
it
is
that
distinguishes
'victims'
from
'non-victims'.
Yet
since
the
earliest
empirical,
statistical
investigations
of
crime
victimisation
survey
data,
two
issues
have
been
apparent:
first,
that
crime
victimisation,
at
whatever
level
of
frequency,
is
apparently
a
relatively
rare
event
-
there
are
many
more
non-victims
than
victims,
and
respondents
reporting
higher
victimisation
frequencies
are
even
rarer.
Second,
despite
its
rarity,
there
are
more
higher-frequency
crime
victims
than
would
be
expected
by
chance,
at
least
under
the
assumptions
of
the
simple
Poisson
probability
model
(Nelson,
1980)
3•
The
argument
of
this
paper
is
that
these
two
observations,
and
the
challenges
of
investigation
they
present,
have
been
paid
far
less
attention
than they
should
by
researchers
(including
ourselves)
and,
consequently,
a
picture
of
the
distribution
of
crime
victimisation
has
emerged
that
is
at
once
puzzling
and
misleading.
With
the
aid
of
some
survey
data
analysed
below,
we
hope
to
identify
some
of
the
key
reasons
why
the
puzzles
have
arisen
and
to
suggest
a
new
strategy
for
modelling
crime
victimisation
risk.
Our
general
solution
is
to
posit
two
polar
risk
groups
in
the
general
population
- a
group
relatively
immune
to
victimisation,
and
a
group
suffering
multiple
victimisation
over
time,
whom
we
shall
term
chronic
victims.
As
a
framework
for
research,
this
model
may
yield
some
more
productive
answers
than
hitherto
to
the
problems
of
understanding
the
frequency
of
crime
victimisation
in
the
population.
THE
'DOUBLE-HURDLE'
MODEL
OF
CRIME
VICTIMISATION
Theories
of
crime
victimisation
have
developed
closely
alongside
the
analysis
of
large-scale
victimisation
surveys
(Miethe
and
Meier,
1994t.
Characterising
the
way
in
which
analysis
has
followed
conceptual
concerns,
there
can
be
thought
to
have
been
two
phases
of
development,
each
with
a
predominant
emphasis
on
one
aspect
of
the
frequency-distribution.
Taken
together,
these
might
be
thought
of
in
terms
of
a
double-hurdle
conceptual
framework
for
the
To continue reading
Request your trial