British Journal of Criminology Advance Access published April 18, 2014
doi:10.1093/bjc/azu020
BRIT. J. CRIMINOL
BRIDGING STRUCTURE AND PERCEPTION
On the Neighbourhood Ecology of Beliefs and Worries About Violent Crime
Ian Brunton-Smith*, Jonathan Jackson and Alex Sutherland
Keywords: violence, fear of crime, collective efficacy, neighbourhood effects, disorder,
concentrated disadvantage
Scholarly interest in the spatial patterning of crime has been widespread ever since the
work of the Chicago school (Park and Burgess 1924; Thrasher 1927). Understanding
how crime is geographically distributed is vital to criminological knowledge and policing policy and practice, with a number of spatial theories of crime emerging over the
years, including social disorganization theory (Shaw and McKay 1942), routine activity
theory (Cohen and Felson 1979), crime pattern theory (Brantingham and Brantingham
1993) and more recently situational action theory (Wikström et al. 2012).
A particularly influential explanation centres upon collective efficacy as an emergent property of a neighbourhood. On this account, it is the shared values and shared
propensities for collective action that inform the strength of informal control mechanisms at the local neighbourhood level (for an excellent review, see Sampson, 2012).
Originating from a series of analyses in Chicago that linked the spatial patterning of
concentrated disadvantage, residential instability and population heterogeneity to differences in levels of violence through variations in the levels of collective efficacy in these
areas (Sampson et al. 1997, 1999, 2002; Morenoff et al. 2001), studies in cities across
the world have tested the importance of collective efficacy for inhibiting violent crime
(Zhang et al. 2007; Oberwittler and Wikström 2009; Odgers et al. 2009; Armstrong
et al. 2010; Mazerolle et al. 2010). There is a good deal of support for the importance of
collective efficacy in the United States, with structural characteristics of the neighbourhood seeming to inhibit the development and sustenance of low-level informal social
control processes, removing certain barriers to violent crime.
*Ian Brunton-Smith, Department of Sociology, University of Surrey, Surrey GU2 7XH, UK; [email protected];
Jonathan Jackson, Department of Methodology and Mannheim Centre for Criminology, LSE, UK; [email protected]; Alex
Sutherland, Institute of Criminology, University of Cambridge, UK; [email protected]
Page 1 of 24
© The Author 2014. Published by Oxford University Press on behalf of the Centre for Crime and Justice Studies
(ISTD). All rights reserved. For permissions, please e-mail: [email protected]
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Applying Robert Sampson’s (2012) work on interdependent spatial patterns in a new setting, we
link structural characteristics of the neighbourhood to public beliefs and worries about neighbourhood violence via two intermediate mechanisms: (1) collective efficacy and (2) neighbourhood disorder. Analysing data from face-to-face interviews of 61,436 individuals living in 4,761 London
neighbourhoods, we find that the strength of informal social control mechanisms and the extent of
low-level breaches of common standards of behaviour communicate information about the prevalence and threat of violent crime in one’s neighbourhood. Moreover, collective efficacy partially
mediates many of the statistical effects of structural characteristics of the neighbourhood on beliefs
and worries about violent crime. Theoretical implications of the findings are discussed.
BRUNTON-SMITH ET AL.
The social ecology of beliefs and worries about violence
In our analysis of the spatial patterning of public perception, we distinguish between
beliefs about violence (perceptions of the prevalence of gun, knife and gang crime in
one’s local neighbourhood) and worries about violent victimization (rumination about the
Page 2 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
But outside of the United States, the picture is less clear. In one recent study,
Sutherland et al. (2013) did not find a strong link between collective efficacy and violent
crime in London. While neighbourhood characteristics such as concentrated disadvantage and population mobility predicted both collective efficacy and violent crime, collective efficacy did not seem to be the mechanism linking neighbourhood conditions in
London to violence. Similarly, Bruinsma et al. (2013) found no link between collective
efficacy and crime in the Hague (but see Sampson and Wikström (2007) who found
some evidence for the role of collective efficacy in explaining variation in violence in
Stockholm). Also of note is Kaylen and Pridemore’s (2013) study, which found only
mixed evidence in rural communities across England and Wales that social disorganization (which in part reflects the absence of collective efficacy) mediated the impact
of neighbourhood characteristics (socioeconomic status and ethnic heterogeneity) on
property and violent crime.
In this paper, we consider the idea that structural characteristics of the neighbourhood shape not actual violent crime but rather public beliefs and worries about violent crime.
We examine whether collective efficacy and neighbourhood disorder mediate the relationship between neighbourhood conditions (such as concentrated disadvantage) and
public perceptions and anxieties regarding violent crime. On the one hand, collective
efficacy refers to shared values and shared propensities for collective action. On the
other hand, neighbourhood disorder refers to aspects of the social and physical environment that indicates to the observer (1) a lack of control and concern and (2) the values
and intentions of others that share the space (Wilson and Kelling 1982; Skogan 1990;
LaGrange et al. 1992; Innes 2004; Jackson 2004; Brunton-Smith 2011). Using a dataset
based on 61,436 individuals living in 4,761 neighbourhoods, we examine whether collective efficacy and disorder bridge neighbourhood structure and individual perception.
We explore whether (1) certain structural characteristics of the neighbourhood help to
inhibit the development and sustenance of low-level informal social control processes
and the ability of local people and institutions to regulate public behaviour; (2) shared
values and propensities to act communicate to observers that violence is a problem and/
or threat; and (3) visible symbols of the failure of local people and authorities to regulate behaviour and secure social control (i.e. breaches of normative standards of public
behaviour) also send signals of potential violent conduct in one’s neighbourhood.
Our paper proceeds as follows. We begin with an overview of our conceptual model,
specifying a number of potential pathways through which structural and social neighbourhood characteristics predict beliefs and worries about violent crime. We then
describe the Metropolitan Police Public Attitudes Survey data source used in this analysis. This is followed by the results from a multilevel structural equation model (SEM)
where we link the structural characteristics of local neighbourhoods to beliefs about
the prevalence of violence and worry about becoming a victim of violent crime through
collective efficacy and neighbourhood disorder. To close, we discuss our results in light
of existing evidence.
BRIDGING STRUCTURE AND PERCEPTION
Page 3 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
personal risk of being mugged and robbed or physically attacked in the local streets
by a stranger). Modelling each separately is important for two reasons. First, people
could think violence is a local problem but not necessarily worry about their own risk
of victimization; they may believe that violence occurs in their neighbourhood but they
do not feel personally vulnerable or susceptible (see Killias 1990; Ferraro 1995; Jackson
2013). Second, if more serious crimes tend to cluster more strongly in certain geographical ‘hot-spots’ (Sherman et al. 1989; see also Braga 2005; Johnson 2010) then beliefs
about gun, knife and gang crime may be more strongly shaped by neighbourhood context and neighbourhood conditions (assuming of course that people’s perceptions of
local crime problems partly ‘track’ the reality of crime in their area).
Our study builds upon prior work not only into collective efficacy but also into the
links between neighbourhood disorder and fear of crime (in the United States: Perkins
and Taylor 1996; Rountree and Land 1996; Sampson 2009; in the United Kingdom:
Farrall et al. 2009; Brunton-Smith and Sturgis 2011). This research helps us understand
the relationship between the micro-environment and fear of crime in two ways (for
reviews of the fear of crime literature, see Hale 1996; Farrall et al. 2009). First, direct
signs of crime can indicate real risk of crime. Second, fear of crime may be provoked
by the features of the urban environment (i.e. cues) that are symbolic of the capacity
of the community to exercise informal social control. The importance of a neighbourhood’s capacity to regulate its residents’ behaviour has been frequently cited (Bursik
1988; Bursik and Grasmick 1993; Sampson and Raudenbush 1999) with Jacobs (1961:
31–2) clearly articulating this: ‘The public peace ... is not kept primarily by the police,
necessary as the police are. It is kept primarily by the intricate, almost unconscious,
network of voluntary controls and standards among the people themselves and [is]
enforced by the people themselves.’
There is thus reason to believe that beliefs and worries about crime are driven by
levels of informal social control and public signs of the breach of normative standards
of behaviour. In turn, collective efficacy and neighbourhood disorder may be driven
by structural characteristics of the neighbourhood, such as concentrated disadvantage
and residential stability (Sampson et al. 1997). To our knowledge, however, no study
has used a multilevel SEM framework to capture individuals within neighbourhoods;
measured structural characteristics of neighbourhoods; modelled both collective efficacy and disorder as neighbourhood properties that mediate neighbourhood structure on public perception; and assessed both people’s beliefs about the prevalence of
local violence and worries about violent victimization. Our contribution is to bring
together hitherto disparate strands of research and theory, exploiting a powerful new
dataset of individuals nested in London neighbourhoods (see also Jackson et al. 2012b;
Sutherland et al. 2013).
Figure 1 gives an overview of the analytical framework structuring the current study.
Our neighbourhood structural measures reflect the range of characteristics of local
neighbourhoods that have been linked to beliefs and worries about violent crime in
prior research (Covington and Taylor 1991; Brunton-Smith and Sturgis 2011). Research
drawing on the social disorganization tradition has pointed to higher levels of worry
in more disadvantaged and ethnically diverse local areas as well as those experiencing greater levels of population mobility (Sampson et al. 1997; Kershaw and Tseloni
2005). Studies have also pointed to greater worry in more urban areas, areas with a
younger population and a reduced capacity for effective adult supervision, and areas
BRUNTON-SMITH ET AL.
Concentrated
disadvantage
Population
mobility
c
Urbanicity
Adult: Child
ratio
a
Beliefs about
violence
h
Worries
about
violence
f
g
b
Neighbourhood
disorder
Housing
structure
Ethnic
diversity
e
d
Prior levels of
crime
Fig. 1 A schematic overview of the pathways linking neighbourhood structural characteristics to
beliefs and worries about violence.
containing more vacant properties and terraced accommodation (Sampson and Groves
1989). Other work has emphasized the importance of the neighbourhood crime rate
(Brunton-Smith and Sturgis 2011).
If collective efficacy and neighbourhood disorder are the bridges through which
structural characteristics of local neighbourhoods shape worry and beliefs about violence, we would expect that their inclusion would explain some or all of the differences between these neighbourhoods that are initially observed. If, e.g., the reason that
residents of more socioeconomically disadvantaged neighbourhoods are more likely to
believe that violent crime is a problem is because these neighbourhoods possess weaker
informal control mechanisms—as well as physical and social cues signalling that informal control is weaker (e.g. unrepaired windows or signs of graffiti)—then incorporating collective efficacy and disorder in a model should weaken the direct pathways from
neighbourhood characteristics to beliefs and worry. If, on the other hand, disorder and
Page 4 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Age structure
Collective
efficacy
BRIDGING STRUCTURE AND PERCEPTION
1. Neighbourhood characteristics influence the extent to which people in a given
neighbourhood collectively come to share values and are willing to act upon those
values; and
2. The strength of local and informal social control mechanisms—separate to social and
physical cues of disorder and the failure of formal and informal social control mechanisms—generates information about the prevalence and personal risk of crime.
Consider next neighbourhood disorder. According to Ferraro (1995: 15), incivilities are
‘…low-level breaches of community standards that signal an erosion of conventionally
accepted norms and values’. Signs of neighbourhood breakdown and the deterioration
of social controls are said to influence public perceptions of crime risk (see also Lewis
and Maxfield 1980; Skogan 1990). If control is seen to have been exerted, a sense of
security is engendered. But if a situation arises where residents think control should be
exerted but it is not (e.g. ‘why isn’t anyone doing something about this?’) then (logically)
a sense of insecurity is more likely to be engendered. Consistent with this, prior UK work
has linked worry about crime to certain neighbourhood features. Brunton-Smith and
Sturgis (2011) matched data from personal interviews with more than 100,000 Crime
Survey of England and Wales respondents to census data about the structural composition of local neighbourhoods. Of particular interest is their finding that recorded
crime and neighbourhood structural characteristics all had direct and independent
statistical effects on residents’ expressed fear of crime, and that interviewer-rated disorder mediated some of these effects. In Figure 1, the set of pathways that we test from
neighbourhood structural characteristics to disorder and from disorder to beliefs and
worries about violence—adjusting for any potential impact via collective efficacy—refer
to two more connected propositions:
3. Neighbourhood characteristics influence the extent to which social and physical
disorder develops and is sustained; and
4. Social and physical cues of disorder that signal failure of formal and informal social
control mechanisms—separate to the strength of local and informal social control
mechanisms—generates information about the prevalence and personal risk of
crime.
Page 5 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
collective efficacy are not the central mechanisms, the direct statistical pathways from
neighbourhood structure to beliefs and worries would remain.
Moving to the middle layer of Figure 1, consider collective efficacy. Prior research
suggests that high levels of perceived community efficacy and tight social structure
(with low levels of anonymity and distrust) may inhibit fear of crime. Ross and Jang
(2000) found that people’s perceptions of informal social ties ‘buffered’ the statistical effect of disorder on fear and mistrust. Jackson (2004) showed that perceptions
of social cohesion and informal social control predicted perceived risk in the same
way as judgements of disorder. Sampson et al. (1997) demonstrated that the statistical
effects of concentrated disadvantage, population heterogeneity and residential stability on beliefs about violence were partly mediated by variations in levels of collective
efficacy in Chicago. Accordingly, the set of pathways that we test from neighbourhood
structural characteristics to collective efficacy and from collective efficacy to beliefs
and worries about violence (adjusting for any potential impact via disorder) refer to two
connected propositions:
BRUNTON-SMITH ET AL.
Data and analytical strategy
Our contribution, in sum, is to link neighbourhood structural composition, signs of disorder and collective efficacy to beliefs and worries about violent crime. It is important to
assess perceptions and anxieties about crime and criminal victimization, in part, because
fear of crime can shape well-being, social distrust, precautionary behaviour and people’s
decisions to move in and out of certain neighbourhoods, having knock-on effects on the
zoning and trajectory of cities (Coy and Pöhler 2002; Lemanski 2006) and the public
health of citizens (Stafford et al. 2007; Jackson and Stafford 2009; Lorenc et al. 2012).
We draw on data from the Metropolitan Police Public Attitudes Survey (see Jackson
et al. 2012b), which is representative of residents of London aged 15 and over, with data
covering the period between 2007 and 2010. The survey adopts a multistage clustered
design, with 267 residents sampled from each of London’s 32 boroughs per quarter (full
details available from CELLO mruk Research, 2009). The survey has an annual response
rate of 60 per cent, with 160 interviews conducted in each borough, and a total analytic
sample of 61,436.1 This ensures adequate coverage across the full distribution of local
areas in London to produce a robust assessment of the contribution of area differences.
In addition to detailed survey data for each respondent, we also know their residential
address, allowing us to attach administrative data about the local area to each record.
A resident’s local area is represented here by Lower layer Super Output Areas
(LSOA). This is a census geography, with each area comprised of an average of 600
households grouped together based on spatial proximity, natural boundaries and
social homogeneity (ONS, 2011). Approximately 150 LSOA are grouped within each
London Borough. This is a reasonable representation of local areas, and certainly
1
Comparisons with the 2001 census reveal that the sample is broadly representative of the population of London, with a slight
under-representation of young (aged 15–34) white residents—see CELLO mruk Research (2009).
Page 6 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Before we turn to the study, we highlight the importance of including disorder and collective efficacy simultaneously in the model. We examine whether social and physical cues
of disorder communicate information above and beyond the strength of the informal social
control processes wrapped up in the mechanism of collective efficacy. Conversely, a statistical effect of collective efficacy on people’s beliefs and worry about violent crime (adjusting
for neighbourhood disorder) is consistent with the idea that people infer the prevalence
and threat of violent crime from the strength of informal social control processes that
operate at the neighbourhood level, and that these are separate to disorderly cues. People
may believe that violence is inhibited when people share values and are willing to act upon
those values for the collective good. If collective efficacy remains a predictor of people’s
beliefs and worries (conditioning on neighbourhood disorder), then collective efficacy
may convey information above and beyond more concrete social and physical cues of disorder, like graffiti, litter and young people hanging around in public space.
Finally, we treat worries and beliefs about violent crime as related but conceptually
distinct reactions to the threat of potential victimization. Allowing worries and beliefs
to correlate (rather than regress one on the other) reflects the fact that we are agnostic
about the direction of causal influence between the two constructs (and interested only
in assessing the conditional correlation).
BRIDGING STRUCTURE AND PERCEPTION
Multilevel Structural Equation Modelling
We estimate a multilevel structural equation model, combining individual survey data
with administrative data about local areas from the 2001 census and police recorded
crime. This approach allows us to directly test the structural pathways linking neighbourhood context to worries and beliefs via collective efficacy and neighbourhood disorder outlined in Figure 1. Like regression-based multilevel models, multilevel SEM
enables identification of the variability in beliefs and worries that occurs between
neighbourhoods, in addition to the variability that occurs within each neighbourhood.
Distinct from more traditional multilevel approaches, however, multilevel SEM also
provides a suitable latent variable modelling framework to measure collective efficacy,
neighbourhood disorder, worry about crime and beliefs about violence, as well as allowing us to test for the existence of indirect (mediation) effects (Muthén and Muthén,
2012).
As in Sampson et al.’s (1997) study, we examine whether structural characteristics of
the neighbourhood predict people’s perceptions of violence, with neighbourhood collective efficacy and signs of disorder working as statistical mediators. We assume that
neighbourhood qualities (such as deprivation) inhibit social organization and people
then infer from the lack of ‘shared expectations about order and control, activated ties,
and acts of informal control’ (Sampson, 2012: 20) the extent of the violent crime problem. We also consider whether concrete social and physical cues of disorder predict
people’s beliefs and worries about local violent crime.
Collective efficacy, neighbourhood disorder, worry about crime and beliefs about
violence are all conceptualized as latent variables measured (with error) by a series
of observed indicators collected for each respondent. At the individual level, the factor structure represents individual assessments of collective efficacy and perceived
disorder, as well as personal worries and beliefs about violence. Since our substantive interest is in the processes operating at the neighbourhood level, we do not
impose additional constraints on the pathways between the four latent variables in
the within model, allowing them all to be correlated with one another. At the neighbourhood level, the intercepts of each indicator are treated as random variables
with the same factor structure as the individual model. The four resultant latent
Page 7 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
a more useful area geography than the sorts of area boundaries more traditionally
relied on such as electoral wards, enabling us to study processes that may be operating at the very local level (Oberwittler and Wikström 2009). Haynes et al. (2007) have
demonstrated a good degree of overlap between these ‘artificial’ neighbourhoods,
neighbourhoods as subjectively defined by ‘experts’ (such as town planners) and residents’ own perceptions of their ‘neighbourhood’. This is particularly so because LSOA
are constructed by grouping together households that are similar in housing type
(amongst other things), with housing type identified as an important aspect of how
residents ‘sort’ their neighbourhood from others. However, we cannot ignore the fact
that these are administrative boundaries constructed for the dissemination of census
data, not to represent local neighbourhoods. As such, it is important to assess our
findings at different spatial scales—something we return to in the discussion of our
results.
BRUNTON-SMITH ET AL.
Collective efficacy
Following Sampson et al. (1997), collective efficacy was measured using a total of six
items covering social cohesion and informal social control. All questions were measured on a 5-point scale from strongly disagree (1) to strongly agree (5):
Cohesion/trust
r People in this neighbourhood can be trusted (trust).
r People act with courtesy to each other in public spaces in this area (courtesy).
r You can see from the public space here in the area that people take pride in their
environment (pride).
Informal social control/willingness to intervene
r If I sensed trouble whilst in this area, I could get help from people who live here
(help).
r The people who live here can be relied upon to call the police if someone is acting
suspiciously (call police).
r If any of the children or young people around here are causing trouble, local people
will tell them off (intervene).
These six items were combined to produce both a single ‘efficacy’ score for each individual and a neighbourhood-level latent variable that represented collective efficacy.
ICCs indicate that 5–7 per cent of the variability in these items occurs between neighbourhoods, suggesting that whilst there was clearly a neighbourhood component, residents also differed markedly in their levels of efficacy—a point we return to in the
discussion. Following initial exploratory analyses, additional residual error correlations
between ‘help’ and ‘call police’, and ‘trust’ and ‘courtesy’ were also included at the
individual level.
Page 8 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
variables at the neighbourhood level therefore represent the neighbourhood average levels of collective efficacy, perceived disorder, worry about crime and beliefs
about violence (Muthén and Muthén 2012). Initial exploratory models suggested
good discriminant validity across the latent variables in both the individual and
neighbourhood models, with no strong cross-loadings identified by the modification indices. The full factor structure for the individual and neighbourhood models
is included in Appendix Table A1, along with estimates of the neighbourhood intraclass correlation (ICC) for each item.
The model is estimated using Full Information Maximum Likelihood (FIML)
using Mplus version 7 (Muthén and Muthén 2012). This approach deals efficiently
with missing data, which is treated as Missing At Random conditional on the
range of individual and neighbourhood measures included in the model (Rubin
1987).
BRIDGING STRUCTURE AND PERCEPTION
Signs of neighbourhood disorder
The same strategy was adopted to measure signs of neighbourhood disorder. Here we
relied on individual assessments of the extent that they view signs of physical and social
disorder as a significant problem in the local area (defined as within 15 min’ walk).
These were treated as imperfect estimates of the overall prevalence of signs of disorder within the area. A total of five items were combined, with all items measured on a
4-point scale from ‘not a problem at all’ (1) to ‘a very big problem’ (4).
Physical signs of disorder
Social disorder
r How much of a problem are noisy neighbours or loud parties in this area (noisy
neighbours)?
r How much of a problem are teenagers hanging around on the streets (teenagers)?
r How much of a problem are people being drunk or rowdy in public places
(drunkenness)?
Exploratory analysis confirmed that a single latent factor efficiently represented the variability across these items at both the individual and neighbourhood levels, indicating
that individuals did not make a conceptual distinction between the sorts of disorderly
actions traditionally identified as social and those classed as physical. Between 7 and 16
per cent of the variability in these items occurred across neighbourhoods.
Worry and beliefs about violent crime
It is common in fear of crime research to ask intensity questions and differentiate
between crimes. For example, Ferraro and LaGrange (1987: 715) recommend (1) using
phrases like ‘how afraid’ (which is a ‘helpful way to examine an emotional state of fear’),
(2) making reference to particular crimes (which focuses the mind of the respondent
and allows different crimes to be viewed in different ways) and (3) asking about ‘your
everyday life’ rather than using a hypothetical format (which brings a touch of reality
to the questions). Given the UK focus of the current study, we follow in the British tradition of talking about ‘worry’ rather than ‘fear’ (for discussion about the meaning and
measurement of fear of crime in the UK context, see Gray et al. 2011; Farrall et al. 2009).
Worry about becoming a victim of violent crime was measured using two items covering
worry about being mugged or robbed and worry about being physically attacked. The
response alternatives were ‘not at all worried’, ‘not very worried’, ‘fairly worried’ and
‘very worried’.
Worry about violent crime
r How worried are you about being mugged or robbed in this area (mugged/robbed)?
Page 9 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
r How much of a problem are rubbish or litter lying around (rubbish)?
r How much of a problem are vandalism, graffiti and other deliberate damage to property or vehicles (vandalism)?
BRUNTON-SMITH ET AL.
r How worried are you about being physically attacked by a stranger in the street in this
area (attack)?
Individual perceptions of violence were captured using three survey items covering perceptions of knife crime, gun crime and gang crime in the local area. Respondents were
asked to indicate whether they thought these crimes were a ‘major problem’, ‘a minor
problem’ or ‘not a problem at all’.
Beliefs about violent crime
Exploratory analyses confirmed that these should be treated as distinct but related
latent variables, with a correlation of 0.35 at the neighbourhood level. This indicates
that in areas where worries about violent crime are higher than average, beliefs about
crime will generally be higher than average, but allows for the possibility of a different
pattern of structural pathways between neighbourhood characteristics and worries and
beliefs. Assessment of item ICCs indicated that approximately 10 per cent of the variation in worry about crime is partitioned between neighbourhoods, with between 8 and
19 per cent of the variability in beliefs occurs between neighbourhoods, confirming
the utility of exploring the mechanisms that link neighbourhood structure to worries
and beliefs. Examination of modification indices led to the inclusion of a correlated
error between beliefs about gun and knife crime at the individual level, as well as a link
between beliefs about teenagers hanging around (a component of perceived disorder)
and beliefs about gang crime. No correlated errors were suggested at the neighbourhood level.
Neighbourhood level measures
We used principal components analysis with 2001 census data to construct the
majority of our neighbourhood measures. 2 ‘Concentrated disadvantage’ was a principal components score with high component loadings for the proportions of individuals in receipt of income support; lone parent families; individuals/families in
social housing; unemployed; households without a car; professional or managerial
class; owner-occupied tenancy. ‘Population mobility’ combined in- and outmigration with occupancy rates and the amount of non-domestic land use. ‘Urbanicity’
comprised the extent of domestic land use, green space, population density and
agricultural land in a given LSOA. ‘Age structure’ combined the proportions of
the resident population (1) under 18 and (2) over 65. ‘Housing structure’ was the
proportion of vacant property, terraced property and flats in a neighbourhood.
2
The component loadings for these measures from Sutherland et al. (2013) are given in Appendix Table A2. For full details
of how these were constructed, see Brunton-Smith and Sturgis (2011).
Page 10 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
r To what extent do you think knife crime is a problem in this area (knife).
r To what extent do you think gun crime is a problem in this area (gun).
r To what extent do you think gang crime is a problem in this area (gang).
BRIDGING STRUCTURE AND PERCEPTION
These measures were standardized before entry into the model. We also included
the adult-to-child ratio and a measure of ethnic diversity (the herfindahl concentration formula)—ranging from 0 (representing an area with a single homogenous
ethnic group resident in the area) to 1 (an area with an even mix of different ethnic
groups). Finally, we added information on the neighbourhood crime level, derived
from the Index of Multiple Deprivation. This was a compound measure of all crime
types to capture historical variation in crime between neighbourhoods. Data were
available from a total of 4,760 LSOAs in London, with an average of 13 respondents
per area.
Individual measures
Results
To test whether collective efficacy and neighbourhood disorder operate as intermediate mechanisms linking structural characteristics of local neighbourhoods to beliefs
and worries about violence, we estimated a multilevel SEM (root mean square error
of approximation (RMSEA), 0.023; confirmatory fit index (CFI)/Tucker–Lewis Index
values, 0.939/0.920; see Hu and Bentler 1999).
Table 1 includes details of the spatial variations in collective efficacy and neighbourhood disorder across local areas (pathways a and b in Figure 1). Consistent with
Sutherland et al. (2013), we found that collective efficacy was weaker in more disadvantaged neighbourhoods. Collective efficacy is also weaker in more urban areas and
in neighbourhoods with a housing structure characterized by more terraced accommodation and vacant properties. Conversely, collective efficacy is stronger in more
diverse neighbourhoods, areas with a younger age structure and areas with a higher
adult-to-child ratio. Neighbourhood disorder is generally higher in more disadvantaged neighbourhoods, more urban areas, areas with a higher adult-to-child ratio
and higher levels of crime. We also identify a significant negative correlation between
collective efficacy and neighbourhood disorder, with neighbourhoods that have a
reduced propensity to mobilize to control deviant behaviour experiencing higher
levels of disorder than neighbourhoods where collective efficacy is stronger. Taken
together, these results point to several potential pathways through which collective
efficacy and disorder may mediate the relationship between neighbourhood structure and beliefs about crime, whilst also confirming that the two processes are closely
wrapped up with one another.
Page 11 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
To account for individual differences in perceptions of crime and worry about crime,
our multilevel SEM included a number of individual controls covering gender, age, marital status, ethnicity, length of residence, tenure, employment status and social class.
Including these variables adjusted estimates of neighbourhood differences for uneven
sample composition within each area (so-called compositional effects) and improved
the estimation of the neighbourhood latent measures of collective efficacy and disorder.
This ensured that apparent differences between areas reflect real differences in average
levels of worry and perceptions of violence, rather than chance differences in the sample
characteristics within those areas. Finally, we also controlled for the survey quarter.
BRUNTON-SMITH ET AL.
Table 1
Neighbourhood structural model predicting collective efficacy and neighbourhood disorder
(pathways a and b)
SE
−0.44***
0.05
−0.12***
0.07**
0.12***
−0.37***
0.14***
0.00
0.04
0.03
0.02
0.03
0.03
0.03
0.03
0.02
0.18***
0.03
0.09***
0.01
0.11***
0.01
0.01
0.04*
0.03
0.03
0.02
0.02
0.03
0.03
0.02
0.02
−0.27***
0.023/0.939/0.920
17026.6/529
58,725, 4,759
0.03
Unstandardized measure.
Derived from the Index of Multiple Deprivation.
*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
a
b
Having demonstrated the close link between the structural characteristics of local
neighbourhoods and both the strength of informal social control mechanisms and the
extent of low-level breaches of common standards of behaviour, we now turn our attention to whether collective efficacy and neighbourhood disorder act as intermediate
mechanisms between neighbourhood structure and beliefs and worries about violence.
Table 2 includes estimates of the direct effects of neighbourhood characteristics on
worries and beliefs (pathways c and d from Figure 1) as well as the effects of collective
efficacy and neighbourhood disorder (pathways e, f, g and h). This also includes estimates of the indirect links from neighbourhood characteristics to worries and beliefs
through collective efficacy and neighbourhood disorder. Finally, this includes details of
the total effects, which represent the combined impact of neighbourhood characteristics on worries and beliefs about violent crime from the direct and indirect pathways.
Comparison of the total and direct effects reflect changes to the magnitude of the relationships between neighbourhood characteristics and worries and beliefs about violent
crime once the effects of collective efficacy and neighbourhood disorder have been
adjusted for. The indirect effect estimates allow us to determine whether these mediation effects are felt through collective efficacy and/or neighbourhood disorder.
Considering first the total effects of neighbourhood structural characteristics on
worries and beliefs about violence, we find clear evidence of systematic differences
Page 12 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Neighbourhood Level
Collective efficacy
Neighbourhood disadvantage
Population mobility
Urbanization
Youth population
Ratio Adult:Childa
Housing structure
Ethnic diversitya
I Prior crimeb
Neighbourhood disorder
Neighbourhood disadvantage
Population mobility
Urbanization
Youth population
Ratio Adult:Childa
Housing structure
Ethnic diversitya
I Prior crimeb
Correlations
Collective efficacy—disorder
Model fit (RMSEA/CFI/TLI)
χ2/df
Sample size (individual, area)
Standard estimate
Table 2
Total, direct and indirect effects of neighbourhood structural characteristics through collective efficacy and disorder
Total effect
Indirect effects via
wdisorder (path
b*g and b*h)
Direct effects
(paths c and d)
Standard
estimate
SE
Standard
estimate
SE
Standard
estimate
SE
Standard
estimate
SE
−0.02
−0.06*
0.02
−0.07***
−0.12***
0.13***
0.32***
0.05*
−0.32***
0.25***
0.03
0.03
0.02
0.02
0.03
0.03
0.02
0.02
0.04
0.02
0.14***
−0.02
0.04***
−0.02*
−0.04***
0.12***
−0.05***
0.00
0.02
0.01
0.01
0.01
0.01
0.02
0.01
0.01
0.05***
0.01
0.02***
0.00
0.03***
0.00
0.00
0.01*
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
−0.20***
−0.05*
−0.05*
−0.05*
−0.11***
0.01
0.37***
0.04*
0.03
0.03
0.02
0.02
0.03
0.03
0.02
0.02
0.18***
0.11***
0.14***
0.03
−0.18***
0.17***
0.08***
0.08***
−0.42***
0.10***
0.03
0.03
0.02
0.02
0.03
0.03
0.02
0.02
0.04
0.03
0.19***
−0.02
0.05***
−0.03**
−0.05***
0.16***
−0.06***
0.00
0.02
0.01
0.01
0.01
0.02
0.02
0.01
0.01
0.02**
0.00
0.01**
0.00
0.01**
0.00
0.00
0.00
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
−0.02
0.13***
0.08***
0.06**
−0.14***
0.01
0.14***
0.08***
0.03
0.03
0.02
0.02
0.03
0.03
0.02
0.02
0.35***
58,725, 4,759
0.03
Unstandardized measure.
Derived from the Index of Multiple Deprivation.
*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
a
b
BRIDGING STRUCTURE AND PERCEPTION
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Page 13 of 24
Worry about violent crime
Neighbourhood disadvantage
Population mobility
Urbanization
Youth population
Ratio Adult:Childa
Housing structure
Ethnic diversitya
I Prior crimeb
Collective efficacy (path f)
Neighbourhood disorder (path h)
Beliefs about violent crime
Neighbourhood disadvantage
Population mobility
Urbanization
Youth population
Ratio Adult:Childa
Housing structure
Ethnic diversitya
I Prior crimeb
Collective efficacy (path e)
Neighbourhood disorder (path g)
Correlations
Worry—belief
Sample size (individual, area)
Indirect effects via
collective efficacy
(path a*e and a*f)
BRUNTON-SMITH ET AL.
Page 14 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
in worries and beliefs about crime across London. In common with existing research
(Covington and Taylor 1991; Kershaw and Tseloni 2005; Brunton-Smith and Sturgis
2011), worries and beliefs are generally higher in more ethnically diverse local areas,
areas with a higher crime rate and areas with a higher proportion of terraced accommodation and vacant properties. In contrast, both are generally lower in areas with a
higher adult-to-child ratio. We also find some differences between beliefs and worries.
Residents of more disadvantaged neighbourhoods, urban areas and areas experiencing higher levels of population mobility generally believe violent crime to be more of a
problem. Yet, neighbourhood disadvantage and the level of urbanization are unrelated
to worry about crime, and worry is actually lower in areas experiencing greater population mobility and with a younger age structure. The lower levels of worry in areas
experiencing greater population mobility runs contrary to the ideas of social disorganization theory (Shaw and McKay, 1942) and points to the possibility that population turnover has a qualitatively different meaning for some local residents—signifying
rapid urban development and gentrification—which may counteract the fear enhancing impact of the presence of unfamiliar neighbours. Similarly, whilst we might expect
lower levels of worry and beliefs in areas with a higher adult-to-child ratio (reflecting
the presence of ‘capable guardians’), the lower worry in areas with a larger youth population is unexpected.
We also find strong support for the theorized role that collective efficacy plays in
shaping beliefs and worries. Residents of neighbourhoods characterized by a stronger
degree of collective efficacy have correspondingly lower levels of worry and beliefs
about violent crime (on average) than residents of areas where collective efficacy is
weaker. Neighbourhood signs of low-level disorder are also influential, with higher levels of worry and beliefs about violence evident in areas that had higher levels of neighbourhood disorder.
Crucially, there is also strong evidence that collective efficacy is a plausible mechanism linking neighbourhood structure to worries and beliefs about violent crime.
Across both beliefs and worries, a similar pattern of indirect effects is evident, with
large indirect effects when considering neighbourhood disadvantage and the neighbourhood housing structure and more moderate effects detected when considering
ethnic diversity, urbanization, the age structure and the age ratio. In contrast, neighbourhood disorder does not have a strong mediation effect, with considerably weaker
(albeit significant) indirect effects apparent when considering neighbourhood disadvantage, urbanization, the age ratio and the level of crime (this effect is restricted to
worry about crime).
Collective efficacy completely mediates the relationship between neighbourhood disadvantage and beliefs about violence. The increased tendency for residents of disadvantaged neighbourhoods to believe violence is problematic occurs indirectly via the
lower levels of collective efficacy in these areas. The higher levels of disorder in areas
characterized by more disadvantage are also important although this effect is substantially weaker. The picture is consistent, albeit less straightforward to interpret, when we
consider worry about victimization. Here, lower levels of collective efficacy generally
found in more disadvantaged neighbourhoods act to attenuate the positive direct effect
of disadvantage on worry. A weaker suppression effect of neighbourhood disorder is
also evident. The net result of these competing relationships is a broadly neutral total
effect of disadvantage on worry. Without accounting for the role of collective efficacy
BRIDGING STRUCTURE AND PERCEPTION
Page 15 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
(and to a lesser extent neighbourhood disorder), the lower levels of worry amongst
residents of more disadvantaged neighbourhoods are masked. Reduced levels of worry
in more disadvantaged communities is counter-intuitive, standing at odds with most
existing evidence, but perhaps reflects important differences in the levels of adaptation
to risk throughout London communities. Residents of more disadvantaged neighbourhoods believe that violent crime is more of a problem because of the reduced capacity
for informal control in these neighbourhoods. But these same residents do not worry
more about crime. In fact, they worry less than residents of more prosperous neighbourhoods (with the same levels of collective efficacy and neighbourhood disorder),
suggesting that the threshold at which risk is translated into worry is higher in more
disadvantaged neighbourhoods. A fuller examination of the potentially protective role
that concentrated disadvantage may have on worry within the context of London (when
both the fear inducing effects of reduced collective efficacy and greater prevalence of
disorder have been accounted for) is beyond the scope of this research but may be a
fruitful area of future enquiry.
The link between the neighbourhood housing structure and beliefs and worries
about violence is also completely mediated by collective efficacy. The higher levels of worry
and beliefs about violence in areas with a more terraced accommodation and vacant
properties are completely explained by the lower levels of collective efficacy in these
areas (see Table 1), indicating that the statistical effect is felt through collective efficacy.
Weaker mediation effects of collective efficacy are evident when considering the
links between the neighbourhood age structure and worries and beliefs about crime.
The reduced levels of worry and beliefs about violence in areas with a higher ratio of
adults to children are felt partly through the protective effects of higher levels of collective efficacy in these areas. Similarly, the higher than average levels of worry and beliefs
about violent crime in more urban areas are partly the result of the weaker levels of
collective efficacy in these areas. We also find weak evidence that the higher than average levels of worry in areas with an increased crime rate can be partly explained by the
higher levels of neighbourhood disorder in these areas.
Collective efficacy also alters the magnitude of the relationship between ethnic diversity and beliefs/worry about crime. Here we see that accounting for indirect effects via
collective efficacy strengthens the association between diversity and worries and beliefs
about violent crime. This can be explained with reference to the higher levels of collective efficacy identified in more diverse neighbourhoods (Table 1). Part of the positive relationship between diversity and perceptions and worries is initially suppressed
because collective efficacy is also higher in these neighbourhoods. Since collective efficacy has a negative relationship with worries and beliefs about crime, when differences
in collective efficacy have been accounted for, the fear enhancing effect of diversity
becomes more visible.
In contrast, neighbourhood disorder is much less clearly implicated as a suitable
mechanism linking neighbourhood structure to beliefs and worries about crime.
Here we find only moderate to weak indirect effects of neighbourhood disadvantage,
urbanization, adult-to-child ratio and crime via disorder. Despite clear patterning of
disorder across neighbourhoods and a consistent link to worries and beliefs about violence, neighbourhood disorder does not appear to be a good candidate mechanism for
explaining the spatial patterning of perceptions of crime that result from differences
in neighbourhood structure.
BRUNTON-SMITH ET AL.
Discussion
Page 16 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
The results of our analysis first suggest that neighbourhood collective efficacy exerts a
downward pressure on individual perceptions of crime and worry about crime. London residents living in areas characterized by a greater degree of interpersonal trust and social
cohesion amongst neighbours, with a concomitant increased capacity for informally
controlling disorderly behaviour, are less likely to worry about becoming the victim of
crime and less likely to believe that violent crime is a problem in their neighbourhood.
Importantly, this is evident even when taking into account signs of disorder in a neighbourhood. Our modelling and data structure allow us to disentangle the fear enhancing effects that environmental cues of a lack of social control and regulation have from
the more direct sources of these limited social control mechanisms represented by collective efficacy.
In common with much existing research (Wilson and Kelling 1982; Skogan 1990;
Sampson and Raudenbush 2004; Farrall et al. 2009), neighbourhood disorder is also
implicated as a key driver of residents’ beliefs about the prevalence of violence within
the neighbourhood, as well as their levels of worry about being victimized. Residents
of areas with more signs of physical and social disorder such as graffiti, vandalism and
the presence of teenagers hanging around are more likely to believe that violent crime
is a problem and will also tend to worry more about their own risk of being victimized.
That this effect remains even having controlled for informal social control mechanisms
open to neighbourhoods indicates that visible signs of the failure of formal and informal
social control mechanisms also appear important to residents when forming judgements about the extent of violence and their own likely risk of victimization.
The consistent links with collective efficacy and disorder have important implications
for understanding the dynamics of individual fear of crime and perceptions of crime.
Residential areas possessing more effective networks of social control exert a reassuring
pressure on residents, whose outlook appears to be shaped by these collective sentiments. This appears to operate alongside more frequently reported fear-inducing effect
of signs of disorder within the neighbourhood—typically viewed as indirect signifiers of
an absence of effective social control and regulation mechanisms within the area.
Collective efficacy is also a suitable candidate to further understand the link between
neighbourhood structural characteristics and perceptions demonstrated in previous
research. Sampson et al. (1997) found that differences in levels of collective efficacy
almost completely explained the differences in perceived neighbourhood violence that
were evident between areas based on their levels of crime, concentrated disadvantage,
immigrant populations and ethnic diversity. Our conclusions within the context of the
city of London are more modest, with collective efficacy partially mediating the differences in worry and beliefs about violent crime initially observed between local areas.
Most notably, observed differences in beliefs about the prevalence of violent crime
between areas based on concentrated disadvantage are completely mediated by collective
efficacy. It seems that it is not disadvantage per se that promotes increased perceptions
of violence, but the perceived reduced capacities for control in these neighbourhoods.
Variations in beliefs about violence and worry based on neighbourhood housing structure are also completely mediated by collective efficacy.
However, we find no mediating effect of collective efficacy when considering links
between neighbourhood crime rates and perceptions. This is consistent with the
BRIDGING STRUCTURE AND PERCEPTION
Limitations of the study
The consistent links between neighbourhood structure, collective efficacy and perceptions of crime and worry about crime appears, prima-facie, convincing. But inherent in
our approach are a number of limitations that must be addressed before we can conclude
that collective efficacy and disorder play a role in shaping public perceptions. Collective
efficacy is an abstract concept, measured indirectly with a series of imperfect indicators
asked of survey respondents. Data from the same respondents were also used to measure individual perceptions of crime and worry about crime. Therefore, it is plausible that
the strong associations reflect these shared measurement properties, rather than representing a distinct directional pathway from collective efficacy to perceptions. To mitigate
this problem, we repeated all analyses with a random half of respondents from each area
used to construct a neighbourhood collective efficacy measure (by aggregating individual
responses on the six individual items to the neighbourhood level), and the other random
half used to estimate a traditional multilevel mediation model using this aggregate measure (results available on request). Comparing these results to our multilevel SEM based on
all observations shows a more moderate effect when using the split sample design, but does
not change any of our substantive conclusions. The same strategy was also used to assess
neighbourhood disorder, with the results again robust to the alternative measurement
strategy. We therefore discount shared measurement as an explanation for our findings.
Studies of neighbourhood effects are also regularly challenged on their choice of spatial scale—with the possibility that results are an artefact of the choice of scale (Manley
et al. 2006; Hipp 2007; Weisburd et al. 2009). To check on the sensitivity of our results to
spatial scale, we repeated all analyses at MSOA level (consisting of approximately 5,000
households, with areas again respecting physical boundaries like roads and waterways).
Here a similar picture emerges, with collective efficacy and neighbourhood disorder
again exerting consistent downward pressure on perceptions of all forms of criminal
activity (results available on request). It seems unlikely, then, that the role of collective
efficacy and disorder are artefacts of our choice of spatial scale. Rather, they seem to be
robustly associated with perceptions of crime at different spatial scales.
Page 17 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
findings of Sutherland et al. (2013), who failed to identify a consistent link between
collective efficacy and neighbourhood crime rates. This suggests that the reach of collective efficacy is qualitatively different in London when compared to other cities like
Chicago. When taken alongside the differential effect of ethnic diversity on collective
efficacy in London (see also Sturgis et al. 2014), this reinforces the need to consider the
role of collective efficacy as more context specific—something that is intimately tied
to the broader compositional structure of the area under study. London is historically
diverse and has a unique geographical structure that does not neatly map onto the
grid structure of cities like Chicago. This may mean that whilst collective sentiments
of cohesion and trust, married to an increased capacity to regulate the behaviour of
residents, may promote feelings of increased security amongst residents, such feelings
are of insufficient strength to actually promote increased mobilization against criminal
activity. A direct test of this hypothesis is beyond the capacity of this paper (and the
available data), but it highlights a need for further detailed examination of the various
ways that collective efficacy may operate in different contexts, and whether a sufficient
threshold is required before collective sentiments transform into collective action.
BRUNTON-SMITH ET AL.
Final words
Acknowledgement
Professor Betsy Stanko (Mayor’s Office for Policing and Crime) for access to MetPAS
data.
References
Armstrong, T. A., Katz, C. M. and Schnelby, S. M. (2010), ‘The Relationship Between
Citizen Perceptions of Collective Efficacy and Neighbourhood Violent Crime’, Crime and
Delinquency, online first: 1–22.
Page 18 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
We conclude with a possible direction for future research. In our analysis, we restricted
our focus to neighbourhood processes using neighbourhood summaries of disorder
and collective efficacy, yet there are clearly individual differences among neighbourhood residents, in that not everyone sees their neighbourhood in the same way. Put
simply, two individuals living in the same environment can come to quite different conclusions about the same environmental cues. A good deal of work suggests that people’s
perceptions of their neighbourhood are dependent not just on the social and physical
cues in the local environment, but also on the respondent’s relationship to that environment and others who inhabit it (Sampson and Raudenbush 2004; Farrall et al. 2009).
Some individuals might judge certain stimuli as ‘disorderly’ while other individuals in
the same environment might not (e.g. graffiti), with people ascribing meaning to disorder they see in front of them, with an individual’s existing attitudes, beliefs or prejudices providing a filter through which they experience and interpret their environment
(see also Jackson, 2004). Future research should attempt to disentangle individual perceptions from collective properties. It should also examine whether trust in the police
(cf. Hohl et al. 2010; Bradford 2011) and perceptions of police legitimacy (cf. Jackson
et al., 2012a; Dirikx and van den Bulck, 2014) are linked to both individual perceptions
and neighbourhood levels of collective efficacy and disorder.
In summary, we have provided evidence that collective efficacy and neighbourhood disorder are two bridges linking structural characteristics of the neighbourhood
to people’s beliefs and worries about local violent crime. We have built on a recent
London-based study, which found that collective efficacy was not a particularly important mechanism through which neighbourhood characteristics shaped actual violence
(Sutherland et al. 2013), contradicting well-established work from the United States,
particularly Chicago (Sampson et al. 1997, 1999, 2002). Reanalysing Sutherland et al.’s
(2013) data, we found that collective efficacy and neighbourhood disorder do seem to
act as sources of information for local people. Building on much prior work (LaGrange
et al. 1992; Ferraro 1995; Perkins and Taylor 1996), we have found that people infer the
prevalence and threat of violence from the strength of informal social control mechanisms and the presence of social and physical cues of disorder. Collective efficacy seems
to be a particularly important mediator of structural neighbourhood characteristics on
people’s beliefs and worries—a bridge by which factors like concentrated disadvantage
shape public perceptions of violence and risk.
BRIDGING STRUCTURE AND PERCEPTION
Page 19 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Bradford, B. (2011), ‘Convergence not Divergence? Trends and Trajectories in Public
Contact and Confidence in the Police’, British Journal of Criminology, 51: 179–200.
Braga, A. A. (2005), ‘Hot Spots Policing and Crime Prevention: A Systematic Review of
Randomized Controlled Trials’, Journal of Experimental Criminology, 1: 317–42.
Brantingham, P. and Brantingham, P. (1993), ‘Nodes, Paths and Edges: Considerations
on the Complexity of Crime and the Physical Environment’, Journal of Environmental
Psychology, 13: 3–28.
Bruinsma, G. J. N., Pauwels, L. J. R., Weerman, F. M. and Bernasco, W., (2013), ‘Social
Disorganization, Social Capital, Collective Efficacy and the Spatial Distribution of Crime
and Offenders: An Empirical Test of Six Neighbourhood Models for a Dutch city’, British
Journal of Criminology, 53: 942–63.
Brunton-Smith, I. (2011), ‘Untangling the Relationship between Fear of Crime and
Perceived Disorder: Evidence from a Longitudinal Study of Young People in England
and Wales’, British Journal of Criminology, 51: 885–99.
Brunton-Smith, I. and Sturgis, P. (2011), ‘Do Neighborhoods Generate Fear of Crime?:
An Empirical Test using the British Crime Survey’, Criminology, 49: 331–69.
Bursik, R. J. (1988), ‘Social Disorganisation and Theories of Crime and Delinquency:
Problems and Prospects’, Criminology, 26: 519–51.
Bursik, R. J. and Grasmick, H. G. (1993) Neighbourhoods and Crime: The Dimensions of
Effective Community Control. Lexington Books.
CELLO mruk research (2009), Public Attitude Survey Technical Report (Q1 2009/10). CELLO
mruk research.
Cohen, L. E. and Felson, M. (1979), ‘Social Change and Crime Rate Trends: A Routine
Activity Approach’, American Sociological Review, 44: 588–608.
Covington, J. and Taylor, R. B. (1991), ‘Fear of Crime in Urban Residential
Neighbourhoods: Implications of Between- and Within- Neighbourhood Sources for
Current Models’, The Sociological Quarterly, 32: 231–49.
Coy, M. and Pohler, M. (2002), ‘Gated Communities in Latin American Megacities: Case
Studies in Brazil and Argentina’, Environment and Planning B: Planning and Design, 29:
355–70.
Dirikx, A. and Van den Bulck, J. (2014), ‘Media Use and the Process-based Model for
Police Cooperation: An Integrative Approach toward Explaining Adolescents’ Intentions
to Cooperate with the Police’, British Journal of Criminology, 54: 344–65.
Farrall, S., Jackson, J. and Gray, E. (2009), Social Order and the Fear of Crime in Contemporary
Times. Oxford University Press.
Ferraro, K. F. (1995), Fear of Crime: Interpreting Victimization Risk. State University of New
York Press.
Ferraro, K. F. and LaGrange, R. L. (1987), ‘The Measurement of Fear of Crime’, Sociological
Inquiry, 57: 70–101.
Gray, E., Jackson, J. and Farrall, S. (2011), ‘Feelings and Functions in the Fear of Crime:
Applying a New Approach to Victimisation Insecurity’, British Journal of Criminology, 51:
75–94.
Hale, C. (1996), ‘Fear of Crime: A Review of the Literature’, International Review of
Victimology, 4: 79–150.
Haynes, R., Daras, K., Reading, R. and Jones, A. (2007), ‘Modifiable Neighbourhood
Units, Zone Design and Residents’ Perceptions’, Health and Place, 13: 812–25.
BRUNTON-SMITH ET AL.
Page 20 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Hipp, J. R. (2007), ‘Block, Tract, and Levels of Aggregation: Neighborhood Structure and
Crime and Disorder as a Case in Point’, American Sociological Review, 72: 659–80.
Hohl, K., Bradford, B. and Stanko, E. (2010), ‘Influencing Trust and Confidence in the
Metropolitan Police: Results from an Experiment Testing the Effect of Leaflet-drops on
Public Opinion’, British Journal of Criminology, 50: 491–513.
Hu, L.-T. & Bentler, P. M. (1999), ‘Cutoff Criteria for Fit Indices in Covariance Structure
Analysis: Conventional Criteria Versus New Alternatives’, Structural Equation Modeling, 6:
1–55.
Innes, M. (2004), ‘Signal Crimes and Signal Disorders: Notes on Deviance as Communicative
Action’, British Journal of Sociology, 55: 335–55.
Jackson, J. (2004), ‘Experience and Expression: Social and Cultural Significance in the
Fear of Crime’, British Journal of Criminology, 44: 946–66.
—— (2013), ‘Cognitive Closure and Risk Sensitivity in the Fear of Crime’, Legal and
Criminological Psychology. Online early access. 5th October 2013.
Jackson, J., Bradford, B., Hough, M., Myhill, A., Quinton, P. and Tyler, T. R. (2012a),
‘Why do People Comply with the Law? Legitimacy and the Influence of Legal Institutions’,
British Journal of Criminology, 52: 1051–71.
Jackson, J., Bradford, B., Stanko, E. and Hohl, K. (2012b), Just Authority? Trust in the Police
in England and Wales. Oxon: Routledge.
Jackson, J. and Stafford, M. (2009), ‘Public Health and Fear of Crime: A Prospective
Cohort Study’, British Journal of Criminology, 49: 832–47.
Jacobs, J. (1961), The Death and Life of Great American Cities. Vintage Books.
Johnson, S. D. (2010), ‘A Brief History of Crime Concentration’, European Journal of Applied
Mathematics, 21: 349–70.
Kaylen, M. T. and Pridemore, W. A. (2013), ‘Social Disorganisation and Crime in Rural
Communities: The First Direct Test of the Systemic Model’, British Journal of Criminology,
53: 905–23.
Kershaw, C. and Tseloni, A. (2005), Predicting Crime Rates, Fear and Disorder based on Area
Information: Evidence from the 2000 British Crime Survey. Home Office.
Killias, M. (1990), ‘Vulnerability: Towards a Better Understanding of a Key Variable in the
Genesis of Fear of Crime’, Violence and Victims, 5: 97–108.
LaGrange, R. L., Ferraro, K. and Supancic, M. (1992), ‘Perceived Risk and Fear of Crime:
Role of Social and Physical Incivilities’, Journal of Research in Crime and Delinquency, 29:
311–34.
Lemanski, C. (2006), ‘Residential Responses to Fear (of Crime Plus) in Two Cape Town
Suburbs: Implications for the Post-Apartheid City’, Journal of International Development, 18:
787–802.
Lewis, D. A. and Maxfield, M. G. (1980), ‘Fear in the Neighbourhoods: An Investigation
of the Impact of Crime’, Journal of Research in Crime and Delinquency, 17: 160–89.
Lorenc, T., Clayton, S., Neary, D., Whitehead, M., Petticrew, M., Thomson, H.,
Cummins, S., Sowden, A. and Renton, A. (2012), ‘Crime, Fear of Crime, Environment,
and Mental Health and Wellbeing: Mapping Review of Theories and Causal Pathways’,
Health and Place, 18: 757–65.
Manley, D., Flowerdew, R. and Steel, D. (2006), ‘Scales, Levels and Processes: Studying
Spatial Patterns of British Census variables’, Computers, Environment and Urban Systems, 30:
143–60.
BRIDGING STRUCTURE AND PERCEPTION
Page 21 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Mazerolle, L., Wickes, R. and McBroom, J. (2010), ‘Community Variations in Violence:
The Role of Social Ties and Collective Efficacy in Comparative Context’, Journal of Research
in Crime and Delinquency, 47: 3–30.
Morenoff, J. D., Sampson, R. J. and Raudenbush, S. W. (2001), ‘Neighborhood Inequality,
Collective Efficacy, and the Spatial Dynamics of Urban Violence’, Criminology, 39: 517–59.
Muthén, B. and Muthén, L. (2012), Mplus User’s Guide, 7th edn. Muthén & Muthén
Oberwittler, D. and Wikström, P.-O. H. (2009), ‘Why Small is Better: Advancing the
Study of the Role of Neighbourhood Contexts in Crime Causation’, in D. Weisburd, W.
Bernasco and G. Bruinsma, eds., Putting Crime in its Place: Units of Analysis in Spatial Crime
Research, 35–59. Springer.
Odgers, C. L., Moffitt, T. E., Tach, L. M., Sampson, R. J., Taylor, A. and Matthews,
C. L. (2009), ‘The Protective Effects of Neighbourhood Collective Efficacy on British
Children Growing up in Deprivation: A Developmental Analysis’, Developmental Psychology,
45: 942–57.
ONS (2011), ‘More About Areas—Further Areas’, available online at http://www.neighbourhood.statistics.gov.uk.
Park, R. E. and Burgess, E. W. (1924), Introduction to the Science of Society. University of
Chicago Press.
Perkins, D. D. and Taylor, R. B. (1996), ‘Ecological Assessments of Community Disorder:
Their Relationship to Fear of Crime and Theoretical Implications’, American Journal of
Community Psychology, 24: 63–107.
Ross, C. E. and Jang, S. J. (2000), ‘Neighborhood Disorder, Fear, and Mistrust: the Buffering
Role of Social Ties with Neighbors’, American Journal of Community Psychology, 28: 401–20.
Rountree, P. and Land, K. C. (1996), ‘Perceived Risk versus Fear of Crime: Empirical
Evidence of Conceptually Distinct Reactions in Survey Data’, Social Forces, 74: 1353–76.
Rubin, D. B. (1987), Multiple Imputation for Nonresponse in Surveys. Wiley.
Sampson, R. J. (2009), ‘Disparity and Diversity in the Contemporary City: Social (Dis)order
Revisited’, British Journal of Sociology, 60: 1–31.
—— (2012), Great American City: Chicago and the Enduring Neighborhood Effect. University of
Chicago Press.
Sampson, R. J. and Groves, W. B. (1989), ‘Community Structure and Crime: Testing Social
Disorganization Theory’, American Journal of Sociology, 94: 774–802.
Sampson R. J., Morenoff, J. and Earls, F. (1999), ‘Beyond Social Capital: Spatial Dynamics
of Collective Efficacy for Children’, American Sociological Review, 64: 633–60.
Sampson, R. J., Morenoff, J. D. and Gannon-Rowley, T. (2002), ‘Assessing Neighbourhood
Effects: Social Processes and New Directions in Research’, Annual Review of Sociology, 28:
443–78.
Sampson, R. J. and Raudenbush, S. W. (1999), ‘Systematic Social Observation of Public
Spaces: A New Look at Disorder in Urban Neighbourhoods’, American Journal of Sociology,
105: 603–51.
—— (2004), ‘Seeing Disorder: Neighborhood Stigma and the Social Construction of
Broken Windows’, Social Psychology Quarterly, 67: 319–42.
Sampson, R. J., Raudenbush, S. and Earls, F. (1997), ‘Neighborhoods and Violent Crime:
A Multilevel Study of Collective efficacy’, Science, 277: 918–24.
Sampson, R. J. and Wikström, P. O. (2007), ‘The Social Order of Violence in Chicago
and Stockholm Neighborhoods: A Comparative Inquiry’, in I. Shapiro, S. Kalyvas and T.
Masoud, eds., Order, Conflict, and Violence, 97–119. Cambridge University Press.
BRUNTON-SMITH ET AL.
Page 22 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Shaw, C. R. and McKay, H. D. (1942), Juvenile Delinquency in Urban Areas. University of
Chicago Press.
Sherman, L. W., Gartin, P. R. and Buerger, M. E. (1989), ‘Hot Spots of Predatory Crime:
Routine Activities and the Criminology of Place’, Criminology, 27: 27–55.
Skogan, W. (1990), Disorder and Decline: Crime and the Spiral of Decay in American Neighbourhoods.
University of California Press.
Stafford, M., Chandola, T. and Marmot, M. (2007), ‘Association between Fear of Crime
and Mental Health and Physical Functioning’, American Journal of Public Health, 97:
2076–81.
Sturgis, P., Brunton-Smith, I., Jackson, J. and Kuha, J. (2014), ‘Ethnic Diversity and the
Social Cohesion of Neighbourhoods in London’, Ethnic and Racial Studies, 37.
Sutherland, A., Brunton-Smith, I. and Jackson, J. (2013), ‘Collective Efficacy and
Violence in London’, British Journal of Criminology, 53: 1050–74.
Thrasher, F. M. (1927), The Gang: A Study of 1,313 Gangs in Chicago. University of Chicago
Press.
Weisburd, D., Bruinsma, G. and Bernasco, W. (2009), ‘Units of Analysis in Geographic
Criminology: Historical Development, Critical Issues, and Open Questions’, in D.
Weisburd, W. Bernasco and G. Bruinsma, eds., Putting Crime in its Place: Units of Analysis
in Spatial Crime Research. Springer.
Wikström, P. O., Oberwittler, D., Treiber, K. and Hardie, B. (2012), Breaking Rules: The
Social and Situational Dynamics of Young People’s Urban Crime. OUP.
Wilson, J. Q. and Kelling, G. L. (1982), ‘Broken Windows’, Atlantic Monthly, 29–38.
Zhang, L., Messner, S. F. and Liu, J. (2007), ‘A Multilevel Analysis of the Risk of Household
Burglary in the City of Tianjin, China’, British Journal of Criminology, 47: 918–37.
BRIDGING STRUCTURE AND PERCEPTION
Appendix
Table A1
Factor loading
SE
0.85
0.77
0.00
0.01
0.71
0.64
0.79
0.01
0.01
0.01
0.69
0.70
0.71
0.66
0.65
0.55
0.00
0.00
0.00
0.00
0.01
0.01
0.64
0.76
0.49
0.70
0.66
0.01
0.00
0.01
0.00
0.00
0.38
0.28
0.15
0.27
0.01
0.01
0.01
0.01
1.00
0.82
0.00
0.02
0.09
0.10
0.96
0.91
0.92
0.01
0.01
0.01
0.19
0.15
0.08
0.89
0.90
0.87
0.93
0.66
0.75
0.01
0.02
0.02
0.01
0.03
0.02
0.07
0.05
0.06
0.06
0.07
0.07
0.87
0.98
0.86
0.84
0.92
58,725, 4,759
0.01
0.01
0.01
0.01
0.01
0.15
0.10
0.15
0.07
0.16
Page 23 of 24
ICC
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
Individual level
Worry about violent crime
Mugged/robbed
Attack
Beliefs about violent crime
Knife
Gun
Gang
Perceptions of efficacy
Trust
Courtesy
Pride
Help
Call police
Intervene
Perceived disorder
Rubbish
Vandalism
Noisy neighbours
Teenagers
Drunkenness
Correlated errors
Gun—knife
Help—call police
Trust—courtesy
Teenagers—gang
Neighbourhood level
Worry about violent crime
Mugged/robbed
Attack
Beliefs about violent crime
Knife
Gun
Gang
Collective efficacy
Trust
Courtesy
Pride
Help
Call police
Intervene
Neighbourhood disorder
Rubbish
Vandalism
Noisy neighbours
Teenagers
Drunkenness
Sample size (individual, area)
Latent variable measurement model
BRUNTON-SMITH ET AL.
Table A2
Rotated factor loadings for neighbourhood measures based upon 2001 census data for all LSOA
across England (n = 32,482)
Rotated component matrixa
Population
mobility
Urbanicity
Age
structure
Housing
structure
0.890
0.836
0.852
0.828
0.817
−0.780
−0.620
0.044
−0.143
0.237
−0.109
−0.098
−0.045
0.314
0.292
0.454
0.405
0.015
0.227
0.274
0.485
0.202
0.024
−0.008
0.200
0.371
0.155
−0.374
0.110
−0.130
0.242
−0.052
0.868
0.856
0.778
0.570
0.552
−0.398
−0.251
0.459
0.095
0.529
0.155
0.149
0.019
0.212
0.338
−0.024
−0.163
0.919
−0.905
0.807
−0.628
0.048
0.098
0.244
0.262
0.376
0.039
−0.148
−0.138
0.259
0.214
0.125
0.293
0.078
0.119
−0.089
0.174
−0.529
0.044
0.030
0.162
−0.095
0.082
0.126
0.074
−0.083
0.144
0.702
−0.902
−0.145
0.242
−0.105
0.085
0.142
−0.153
0.130
0.069
−0.343
0.040
0.174
−0.048
−0.094
0.037
0.123
0.203
−0.057
−0.066
−0.287
0.136
−0.043
0.541
0.690
−0.502
a
Rotation converged in six iterations. Extraction method: Principal component analysis. Rotation method:
Varimax with Kaiser Normalization. Values in bold indicate strong component loadings.
Page 24 of 24
Downloaded from http://bjc.oxfordjournals.org/ by guest on October 9, 2014
% income support
% lone parent
% local authority housing
% unemployed
% no car
% professional or managerial
% owner occupied
% domestic land use
% green space
Population density
% agricultural land
% in migration
% out migration
% single parent non-pensioners
% non-domestic land use
Occupancy rating
% under 18
% over 65
% vacant property
% terraced property
% flats
Concentrated
disadvantage