write a paragraph
tareq1144
TheEffectivenessofNeighborhoodWatch.pdf
Campbell Systematic Reviews 2008:18 First published: 10 December, 2008 Last updated: 10 December, 2008
The Effectiveness of Neighborhood Watch Trevor Bennett, Katy Holloway, David Farrington
Colophon
Title The effectiveness of neighborhood watch
Institution The Campbell Collaboration
Authors Bennett, Trevor Holloway, Katy Farrington, David
DOI 10.4073/csr.2008.18
No. of pages 48
Last updated 10 December, 2008
Citation Bennett T, Holloway K, Farrington D. The effectiveness of neighborhood watch. Campbell Systematic Reviews 2008:18 DOI: 10.4073/csr.2008.18
Copyright © Bennett et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Keywords
Contributions None stated.
Support/Funding The Campbell Collaboration University of Glamorgan, United Kingdom
Potential Conflicts of Interest
There is no conflict of interest. However, one of the studies included in the review was conducted by Trevor Bennett.
Corresponding author
Trevor Bennett Centre for Criminology University of Glamorgan Pontypridd, CF37 1DL United Kingdom Telephone: + 44 1443 482 236 E-mail: [email protected]
Campbell Systematic Reviews
Editors-in-Chief Mark W. Lipsey, Vanderbilt University, USA
Arild Bjørndal, Norwegian Knowledge Centre for the Health Services & University of Oslo, Norway
Editors
Crime and Justice David B. Wilson, George Mason University, USA
Education Chad Nye, University of Central Florida, USA Ralf Schlosser, Northeastern University, USA
Social Welfare Julia Littell, Bryn Mawr College, USA Geraldine Macdonald, Queen’s University, UK & Cochrane Developmental, Psychosocial and Learning Problems Group
Managing Editor Karianne Thune Hammerstrøm, The Campbell Collaboration
Editorial Board
Crime and Justice David Weisburd, Hebrew University, Israel & George Mason University, USA Peter Grabosky, Australian National University, Australia
Education Carole Torgerson, University of York, UK
Social Welfare Aron Shlonsky, University of Toronto, Canada
Methods Therese Pigott, Loyola University, USA Peter Tugwell, University of Ottawa, Canada
The Campbell Collaboration (C2) was founded on the principle that systematic reviews on the effects of interventions will inform and help improve policy and services. C2 offers editorial and methodological support to review authors throughout the process of producing a systematic review. A number of C2's editors, librarians, methodologists and external peer- reviewers contribute.
The Campbell Collaboration P.O. Box 7004 St. Olavs plass 0130 Oslo, Norway www.campbellcollaboration.org
1
THE EFFECTIVENESS OF NEIGHBORHOOD WATCH
CAMPBELL REVIEW
Reviewers
Professor Trevor Bennett Director Centre for Criminology University of Glamorgan Pontypridd, CF37 1DL United Kingdom
Telephone: International + 44 (1443 482236) Fax: International + 44 (1443 484507) Email: [email protected]
Dr Katy Holloway Research Fellow Centre for Criminology University of Glamorgan Pontypridd, CF37 1DL United Kingdom
Telephone: International + 44 (1443 483419) Fax: International + 44 (1443 484507) Email: [email protected]
Professor David Farrington Professor of Psychological Criminology Institute of Criminology University of Cambridge Sidgwick Avenue CB3 9DT United Kingdom
Telephone: International + 44 (1223 335360) Fax: International + 44 (1223 335356) Email: [email protected]
2
ABSTRACT
Background Neighborhood watch (also known as block watch, apartment watch, home watch and community watch) grew out of a movement in the US during the late 1960s that promoted greater involvement of citizens in the prevention of crime. Since then, interest in neighborhood watch has grown considerably and recent estimates suggest that over a quarter of the UK population and over forty per cent of the US population live in areas covered by neighborhood watch schemes.
Objectives The primary aim of this review is to assess the effectiveness of neighborhood watch in reducing crime.
Criteria for inclusion of studies Neighborhood watch sometimes comprises a stand-alone scheme and sometimes includes additional program elements. The most common combination of program elements is the ‘big three’ (neighborhood watch, property marking and security surveys). Studies were selected for inclusion in the review if they were based on a watch scheme either alone or in combination with any of the other ‘big three’ elements. The main quality control was that the studies should be based on random allocation or a pre-post test design with a comparison area.
Search Strategy Studies were identified by searching 11 electronic databases. In addition, studies were sought using on-line library catalogues, literature reviews, lists of references, and published bibliographies. Leading researchers in the field were also contacted when there was a particular need to do so.
Data collection and analysis The narrative review was based on 19 studies (covering 43 evaluations) and the meta-analysis was based on 12 studies (covering 18 evaluations). The data used included police-recorded crimes and self-reported victimizations.
Main results The main finding of the narrative review was that the majority of the schemes evaluated indicated that neighborhood watch was effective in reducing crime.
The main finding of the meta-analysis was that the weighted mean odds ratio for all studies combined was 1.19 using the fixed effects method and 1.36 using the random effects method. The results of both methods show that neighborhood watch was associated with a reduction in crime of between 16 per cent and 26 per cent.
Reviewers’ conclusions This review concludes that neighborhood watch is associated with a reduction in crime.
This review should be cited as: Bennett Trevor, Holloway Katy, and Farrington David. “The Effectiveness of Neighborhood Watch”. In: The Campbell Collaboration Reviews of Intervention and Policy Evaluations (C2-RIPE).
3
BACKGROUND FOR THE REVIEW
Introduction
Neighborhood watch (also known as block watch, apartment watch, home watch and community watch) grew out of a movement in the US that promoted greater involvement of citizens in the prevention of crime (Titus, 1984). One of the first recorded neighborhood watch programs in the US was the Seattle Community Crime Prevention Project launched in 1973 (Cirel et al., 1977). One of the first recorded neighborhood watch schemes in the UK was the Home Watch program implemented in 1982 in Cheshire (Anderton, 1985).
Since the 1980s, the number of neighborhood watch schemes in the UK has expanded considerably. The report of the 2000 British Crime Survey estimated that over a quarter (27%) of all households (approximately six million households) in England and Wales were members of a neighborhood watch scheme (Sims, 2001). This amounted to over 155,000 active schemes. A similar expansion has occurred in the US. The report of The 2000 National Crime Prevention Survey (National Crime Prevention Council, 2001) estimated that 41 per cent of the American population lived in communities covered by neighborhood watch. The report concluded, ‘This makes Neighborhood Watch the largest single organized crime prevention activity in the nation’ (p.39). Considering such large investments in terms of resources and community involvement, it is reasonable for researchers to ask whether neighborhood watch is effective in reducing crime.
The theory of neighborhood watch
The most frequently recorded mechanism by which neighborhood watch is supposed to reduce crime is by residents looking out for suspicious activities and reporting these to the police. The link between reporting and crime reduction is not usually elaborated in the literature. However, it has been argued that visible surveillance might reduce crime as a result of its effect on the perceptions and decision making of potential offenders. Hence, watching and reporting might deter offenders if they are aware of the propensity of the local residents to report suspicious behavior and if they perceive this as increasing the risks of being caught.
Neighborhood watch might also lead to a reduction in crime through the reduction in opportunities for crime. One method discussed in the literature is through the creation of signs of occupancy. Some of the methods by which members of neighborhood watch schemes might create signs of occupancy were discussed in the report of the Seattle scheme (Cirel et al., 1977). These include removing newspapers and milk from outside neighbors’ homes when they are away, mowing the lawn, and filling up trash cans. The way in which signs of occupancy might reduce crime might be through the effect that this has on the perceptions of potential offenders in terms of their likelihood of getting caught.
Neighborhood watch might also lead to a reduction in crime through the various mechanisms of social control. Informal social control is not one of the mechanisms for reducing crime stated in the publicity material of these schemes. Nevertheless, they might indirectly serve to enhance community cohesion and increase the ability of communities to control crime (Greenberg, Rohe, and Williams, 1985). Informal social control can affect community crime
4
through the generation of acceptable norms of behavior and by direct intervention by residents.
It is also possible that neighborhood watch schemes might reduce crime through enhancing police detection. Neighborhood watch might serve to increase in the flow of useful information from the public to the police. An increase in information concerning crimes in progress and suspicious persons and events might lead to a greater number of arrests and convictions and result (when a custodial sentence is passed) in a reduction in crime through the incapacitation of local offenders (Bennett, 1990).
It is also feasible that neighborhood watch might reduce crime through the other components of the program package. It has been argued that property marking might lead to a reduction in crime as a result of making the disposal of marked property more difficult (Laycock, 1985). This might reduce offending rates if potential offenders viewed marked property as increasing the risk of detection. Home security surveys might lead to a reduction in crime as a result of making it physically more difficult for an offender to enter the property (Bennett and Wright, 1984).
Program elements
Neighborhood watch is often implemented as part of a comprehensive package. The typical package is sometimes referred to as the ‘big three’ and includes neighborhood watch, property-marking and home security surveys (Titus, 1984). Some programs include a third or fourth element such as a recruitment drive for special constables, increased regular foot patrols, citizen patrols, educational programs for young people, auxiliary police units, and victim support services.
Neighborhood watch schemes vary in terms of the size of the area covered. Some of the earlier schemes in the US and the UK were based on areas covering just a few households. More recent schemes sometimes cover many thousand households. One of the smallest schemes included in the review was the ‘cocoon’ neighborhood watch program in Rochdale in England covering just one dwelling and its immediate neighbors (Forrester, Frenz, O’Connell, and Pease, 1990). One of the largest was the Manhattan Beach neighborhood watch scheme in Los Angeles covering a population of over 30,000 residents (Knowles, 1983).
Neighborhood watch schemes can be both public and police initiated. Schemes launched in the UK during the early period of a program tended to be police initiated (e.g. the early neighborhood watch schemes in the Metropolitan Police District). More recently, neighborhood watch schemes have been launched mainly at the request of the public. Some police departments continue initiating their own schemes, even when the program is fully developed. A program implemented in Detroit, for example, maintained a section of police- initiated schemes in order to promote neighborhood watch in areas that were unlikely to generate public-initiated requests (Turner and Barker, 1983).
In the US, block watches are usually run by a block captain who is responsible to a block co- ordinator or block organizer. The block co-ordinator acts as the liaison person to the local police department. Neighborhood watch schemes in the UK often include street co-ordinators (equivalent to block captains) and area co-ordinators (equivalent to the block organizer).
5
There is little information in the literature on the number and type of neighborhood watch meetings. The evidence that does exist suggests that some schemes have public meetings that involve all of the residents participating in the scheme, while others have meetings that involve only the organizers of the scheme (Bennett, 1990).
The funding of neighborhood watch schemes is nearly always a joint venture between the local police department and the scheme members through their fund-raising activities. The relative contribution of the two sources varies considerably. Some schemes in the United States are provided with no more than an information package from the local police. Others are provided with police facilities for the production of newsletters and the use of police premises for meetings (Turner and Barker, 1983). Apart from police funding, the majority of schemes are encouraged to raise some funds from other sources such as voluntary contributions, local businesses, and the proceeds of fêtes, and raffles.
Crimes targeted
There is a consensus in the literature that the main aim of neighborhood watch is crime prevention. There are small variations among the programs in terms of which crimes are targeted. The vast majority of programs identify residential burglary as the sole or most important target crime of neighborhood watch. Some programs list other offences that it is hoped neighborhood watch will reduce. The list of other offences is sometimes specific (e.g. 'street robberies, auto thefts and vandalism') and sometimes general (e.g. 'street crime' and 'property crime').
Previous reviews
There are several reviews of evaluations of neighborhood watch programs. One of the earliest conducted in the US was by Titus (1984) who summarized the results of nearly forty community crime prevention programs. Most of these included elements of neighborhood watch. The majority of studies was conducted by police departments or included data from police departments. Nearly all of the studies found that neighborhood watch areas were associated with lower levels of crime. However, most of the evaluations were described as ‘weak’ on the grounds that they offered no comparison group.
Another review of the literature looked mainly at community watch programs in the UK (Husain, 1990). The study reviewed the results of nine existing evaluations and conducted an original analysis of community watch in six additional locations using police-recorded crime data. The review of existing evaluations concluded that there was little evidence that NW prevented crime.
One of the most recent reviews of the literature on the effectiveness of community watch programs selected only evaluations with the strongest research designs. The authors included only studies that used random assignment or studies that monitored both watch areas and similar comparison areas without community watch. The review found just four evaluations that matched these criteria. The results of these evaluations were largely negative. The authors concluded, ‘The oldest and best-known community policing program, Neighborhood
6
Watch, is ineffective at preventing crime’ (Sherman et al, 1997, p. 353; see also Sherman and Eck, 2002).
Objectives of the review
The primary aim of this review is to assess the effects of neighborhood watch on crime.
The primary objectives of the review are:
1) To operationalize the inputs (e.g. neighborhood watch) and the outcomes (e.g. crime) for the purpose of conducting the review. 2) To identify studies that have evaluated the effect of neighborhood watch on crime. 3) To identify a list of studies that meet the minimum criteria of scientific rigor. 4) To obtain a comparable measure of effect size for the most rigorous studies. 5) To arrive at a conclusion about the effectiveness of neighborhood watch.
METHODS
Criteria for inclusion and exclusion of studies in the review
Types of intervention
Neighborhood watch is often implemented alongside other programs. In practice, this is done in two main ways:
a) Neighborhood watch schemes often include elements of other programs within the project. Watch schemes are sometimes described as comprising ‘the big three’ (neighborhood watch, property marking, and security surveys). The additional elements (property marking and security surveys) are viewed as part of neighborhood watch when implemented as part of a package.
b) Neighborhood watch schemes (either single watch schemes or ‘the big three’) are sometimes implemented alongside other related schemes (such as environmental improvements and neighborhood organizing programs) as part of a comprehensive (multi- project) program.
The following types of intervention will be included in the review:
a) stand-alone neighborhood watch schemes (comprising solely a watch component). b) neighborhood watch schemes that include ‘the big three’ (neighborhood watch, property marking and security surveys) as long as there is a watch component. c) neighborhood watch schemes that include two components of ‘the big three’ as long as there is a watch component.
7
In other words, for the purpose of the review we are defining neighborhood watch to mean stand-alone neighborhood watch schemes and neighborhood schemes with additional related elements.
Types of participants
Watch programs can be based on a diversity of populations, including boat owners, farmers, and business employees, and a diversity of locations, including car parks, yacht marinas, and the countryside. The current review is based on schemes involving residents living in neighborhoods.
Types of mediating processes
One of the most important defining elements of neighborhood watch is the mechanism by which the project aims to reduce crime. The main mechanisms of the ‘watch’ part of neighborhood watch schemes are:
a) residents operating as the ‘eyes and ears’ of the police (i.e. surveillance) b) residents reporting suspicious behavior to the police or neighborhood co-ordinator c) residents interacting and working together to solve problems (which might strengthen
social cohesion, collective efficacy, community activism, and other mechanisms of informal social control.)
The mechanisms described above rule out neighborhood wardens and similar citizen patrols. Citizen patrols are based: (a) on the appointment of residents to a particular role, and (b) on agreement to conduct particular duties such as patrolling the streets. Watch schemes are based solely on residents operating in their capacity as residents.
Types of outcome
The review focuses mainly on the impact of neighborhood watch schemes on crime. The types of crimes covered in the review are those that neighborhood watch might be able to reduce. These include the following:
a) crimes against residents b) crimes against dwellings c) other (street) crimes occurring in the watch area
When crime measures are based on police recorded crimes, the main outcome measure is the total number of crimes recorded in the areas studied. When crime measures are based on victimization surveys, the main outcome measure is the prevalence of victimization.
Types of evaluation design
The criteria for selecting rigorous evaluations are based on the Maryland Scientific Methods Scale (SMS) (Sherman and Eck, 2002). This is a five-point scale ranging from level 1 (the weakest design) to level 5 (the strongest design) in terms of overall internal validity. Sherman and Eck (2002) argue that evaluations should be at least level 3 in order to conclude, with a reasonable level of certainty that the program worked. The current review of evaluations also uses this level as the minimum acceptable for inclusion in the review. This level requires that
8
the evaluation must comprise at least a comparison of one or more experimental units and one or more comparable control units over time. Hence, the minimum requirement for inclusion of evaluations in the review of neighborhood watch is that they are based on both before and after surveys and experimental and comparison areas.
Search strategy for identification of studies
Criteria for selecting studies The review included published and unpublished literature. It was based on documented evaluations. There was no restriction on country of origin. The evaluations had to be available in English. There was no restriction on source sector (e.g. academic, government, policy, voluntary, etc.). There was no restriction in terms of year (e.g. year of implementation, study, or publication). There was no restriction of the time period covered by the evaluation (e.g. short-term or long- term effects).
Sources used for selecting studies
The following search strategies were used:
1) search on-line databases (especially for reports and articles) 2) search on-line library catalogues (especially for books) 3) search reviews of the literature on the effectiveness of neighborhood watch in preventing
crime 4) search bibliographies of publications on neighborhood watch 5) contact leading researchers
The following datasets were searched:
1) IBSS (International Bibliography of the Social Sciences) 2) Web of Science 3) Criminal Justice Abstracts 4) National Criminal Justice Reference Service Abstracts 5) Sociological Abstracts 6) Psychological Abstracts (PsycINFO) 7) Social Science Abstracts 8) Government Publications (Home Office) 9) Dissertation Abstracts (ASSIA) 10) ProQuest 11) C2-SPECTR
Search terms
The following search terms were used in the database searches:
neighborhood watch, neighbourhood watch, street watch, block watch, apartment watch, home watch, community watch, home alert, block association, crime alert, block clubs, crime watch, ‘big three’.
9
Description of methods used in primary research
The main types of research design used for evaluating neighborhood watch schemes have been discussed in previous sections. The most common is some kind of quasi-experimental design. The review includes only the strongest of these designs. In practice, quasi- experimental designs were selected only if they included before and after measures in experimental and comparable control areas.
Criteria for determination of independent findings
Evaluations sometimes produce multiple outcome measures. These can occur when:
(1) there are multiple methods of measuring the same outcome, and (2) when the same outcome is measured at multiple points in time.
When multiple outcome measures are provided (e.g. multiple outcome measures of crime) we listed the results for each measure. However, the analysis is based on only one measure. The measure chosen is based on a system for prioritizing the results (i.e. burglary first, followed by all property crimes and then all crimes). When the same outcome is measured at multiple points in time, we have selected the year before and the year after the implementation of the scheme as the first choice. Failing this, we chose other periods in accordance with the above priority system (i.e. periods nearest to the point of implementation were chosen first).
Details of coding categories
The information extracted included: author, publication date, study date, location, physical context of the intervention, type of intervention, duration of the intervention, duration of the evaluation, sample size, other interventions employed at the time, outcome measures, data source, research design, results, author(s) conclusion.
Statistical procedure and conventions
Meta-analyses were carried out to determine an overall effect size. Odds ratios were calculated for each evaluation and a weighted mean odds ratio was calculated for all studies combined based on the guidelines summarized in Lipsey and Wilson (2001). See meta- analysis section for details of the statistical methods used.
Attrition of publications at various stages
Table 1 presents the results of the literature searches. The first section of the table displays the number of publications identified (i.e. the number of ‘hits’) from the literature searches described above. A total of 1,595 publications were identified from the searches. The second section of the table shows the number of publications that was provisionally selected for inclusion. Overall, 335 publications were selected as potentially relevant evaluations. Criteria
10
for selecting publications at this stage were based on a review of titles and abstracts. Publications that were clearly NOT evaluations of neighborhood watch were excluded. The 335 potentially relevant publications included 110 publications that had been identified previously. Hence, 225 unique publications were selected for potential inclusion in the review. The third section of the table displays the number of selected (non-duplicated) publications that were obtained and not obtained. Of the 225 selected publications, 137 (61%) were obtained. The main reasons for not obtaining publications were that they could not be located following various attempts to obtain them by inter-library loan, through the Internet, or by contacting the authors. The fourth section of the table shows publications eligible for inclusion. Thirty publications were eligible and 107 were ineligible. The main reason for ineligibility (n=60) was that the publication did not include an evaluation of neighborhood watch (see Table A1 for details of all ineligible publications). Eleven of the eligible publications presented results that were included in another eligible publication. In each of these cases, the most detailed publication was selected for inclusion in the review. This resulted in 19 publications that presented findings from 19 unique studies. Some studies included evaluations of more than neighborhood watch program. In total, these 19 studies covered evaluations of 43 separate neighborhood watch schemes. The last section of the table shows the number of studies that were included in the meta-analysis. Of the 19 studies, 12 were suitable for inclusion in the meta-analysis on the grounds that they provided sufficient data to conduct the analyses required. In total, these 12 studies covered evaluations of 18 separate neighborhood watch schemes.
11
Table 1: Attrition of publications
Stage Total
IDENTIFIED PUBLICATIONS 1,595
WHETHER PUBLICATION SELECTED
Selected 335 Non-duplicates 225 Duplicates 110
Not selected 1,260 Total 1,595
WHETHER NON-DUPLICATE PUBLICATIONS OBTAINED
Obtained 137 Not obtained 88 Total 225
WHETHER OBTAINED PUBLICATIONS ELIGIBLE FOR NARRATIVE REVIEW
Eligible publications 30 Studies 19 Evaluations 43
Not eligible publications 107 Total 137
WHETHER STUDIES ELIGIBLE FOR META ANALYSIS
Eligible studies 12 Evaluations 18
Not eligible studies 7 Total 19
Notes: Publication = a published document. Study = a research project. Evaluation = an evaluation of a single neighborhood watch scheme. The results of a research project (study) might be reported in more than one publication. A study might present the results of more than one evaluation.
Description of studies meeting the eligibility criteria
The searches described above resulted in 19 studies eligible for inclusion in the review covering 43 evaluations of neighborhood watch schemes. Table 2 provides a description of these studies.
The first column provides details about the author and year of publication. The majority of studies (n=12) were published in the 1980s, when interest in evaluating neighborhood watch was at its height. The second column shows that nine studies reported findings about neighborhood watch schemes operating in the UK and eight reported findings about schemes
12
operating in the US. The two remaining studies reported findings from Canada and Australia. The third column shows that findings from 18 of the 43 separate evaluations were included in the meta-analysis. The fourth column indicates that the majority of evaluations were based on neighborhood watch schemes combined with at least one other element (n=30). Eight of these included the ‘big three’ (i.e. neighborhood watch, property marking and security surveys). The fifth column presents information about the research design. All of the evaluations used a pre-test post-test experimental-control design. The sixth column shows the size of the scheme area (i.e. the number of residents, dwellings, roads or census tracts). The seventh column presents information about the characteristics of the area in which the neighborhood watch scheme was operating (i.e. the experimental area). The last column in the table describes the comparison (or control) area in which neighborhood watch schemes were NOT operating.
13
T a b le
2 : D
e s c r ip
tio n
o f s
tu d ie
s m
e e tin
g th
e e
lig ib
ility c
r ite
r ia
f o r in
c lu
s io
n in
th e r
e v ie
w
A uthor/ publication
date of m ain report
used in the review
C ountry/C
ity/area U
sed in m
eta- anal- ysis
Schem e
elem ents
R esearch design and evaluation
period [1] Size of schem
e area
Schem e area
C om
parison areas
A nderton (1985)
U K
, C heshire,
N orthw
ich D
ivision
Y es
N W
, Im
proved security, Property m
arking
Pre-Post, E xp-C
on, 1 year before and 1 year after
629 schem es
covering 14,300 dw
ellings
A police division of the C
heshire C
onstabulary w ith the largest num
ber of w
atch schem es (629 schem
es covering 14,300 dw
ellings).
The county as a w hole.
B ennett(1990)
U K
, L ondon,
W im
bledon * Y
es N
W ,
Property M
arking, Security Surveys
Pre-Post, E xp-C
on, 12 m onths
before and after 13 roads
A residential area of m
ainly ow ner-
occupied hom es
M atched com
parison som e distance from
the experim
ental area
U K
, L ondon,
A cton
Y es
N W
Pre-Post, E xp-C
on, 12 m onths
before and after 9 roads
A residential area of m
ainly ow ner-
occupied hom es
M atched com
parison som e distance from
the experim
ental area
B ennett and Lavrakas
(1989) U
SA , B
altim ore
N W
, O
ther related activities
Pre-Post, E xp-C
on, 1 year before and 1 year after
not stated Inner-city com
m unities w
ith relatively low
socio-econom ic levels usually racially
m ixed w
ith high crim e problem
s
A ‘ring’ around the target area tw
o census tracts w
ide
U SA
, B oston
ditto ditto
ditto ditto
ditto U
SA , B
ronx ditto
ditto ditto
ditto ditto
U SA
, B rooklyn
ditto ditto
ditto ditto
ditto U
SA , C
leveland ditto
ditto ditto
ditto ditto
U SA
, M iam
i ditto
ditto ditto
ditto ditto
U SA
, M inneapolis
ditto ditto
ditto ditto
ditto U
SA , N
ew ark
ditto ditto
ditto ditto
ditto U
SA , Philadelphia
ditto ditto
ditto ditto
ditto U
SA , W
ashington ditto
ditto ditto
ditto ditto
C irel et al. (1977)
U SA
, W ashington,
Seattle Y
es N
W ,
Property m
arking, Security surveys
Pre-Post, E xp-C
on, 1 year before and 1 year after
not stated Five census tracts w
ithin the city. Tw
o federal tracts adjacent to the selected schem
e census tracts.
Forrester, C hatterton and
Pease (1988) U
K , R
ochdale, K
irkholt Y
es C
N W
, Im
proved security, V
ictim s support
Pre-Post, E xp-C
on, 1 year before and 1 year after
2,280 dw ellings
A residential area of public housing tw
o m
iles outside the tow n centre covering just
over 2,000 households.
The rem ainder of the subdivision
14
H enig (1984)
U SA
, W ashington
D C
, T he First
Police D istrict
Y es
N W
Pre-Post, E xp-C
on, 1 year before and 3 years after
not stated A
n area of the city w ith a population over
50 per cent black and under one-third ow
ner-occupied and highest crim e rate of
the seven police divisions.
The police division as a w hole.
H ulin (1979)
U SA
, C alifornia,
Fontana N
W ,
Target hardening, Property m
arking
Pre-Post, E xp-C
on, 1 year before and 1 year after
not stated A
sm all, high crim
e area of m ainly
apartm ents in a sem
i-rural location. Four dem
ographically sim ilar control area
w ith sim
ilar pre-test crim e rates.
H usain (1990)
U K
, B irm
ingham ,
H andsw
orth W ood
N W
Pre-Post, E xp-C
on, 2 years before and after
7,000 hh H
igh-status residential area. M
atched com parison adjacent to the
experim ental area
U K
, B righton,
Saltdean N
W Pre-Post, E
xp-C on, 2 years before
and after 1,500 hh
A large high-status residential area on the
Sussex coast M
atched area som e distance from
the experim
ental area U
K , B
urnley, H
argher C lough
N W
Pre-Post, E xp-C
on, 2 years before and after
1,000 hh D
ensely populated sm all tow
n com prising
m ainly m
anual w orkers
The non-schem e part of the tow
n
U K
, M anchester,
T rafford, Stretford
N W
Pre-Post, E xp-C
on, 2 years before and after
11,000 hh A
large racially m ixed area
com prising a
variety of dw elling types
N on-m
atched area near, but not adjacent to, the experim
ental area U
K , Preston,
G range E
state and M
oor N ook
N W
Pre-Post, E xp-C
on, 2 years before and after
1,600 hh T
w o large local authority housing areas
N on-schem
e dw ellings close to the
experim ental area
U K
, Sutton C
oldfield N
W Pre-Post, E
xp-C on, 2 years before
and after 3,600 hh
Several fairly affluent areas on the periphery of a large conurbation
Tw o police beats rem
ote from the
experim ental areas, but in the sam
e police sub-division
Jenkins and Latim er
(1986) U
K , M
erseyside, Y
ork A venue
Y es
N W
Pre-Post, E xp-C
on, 1 year before and 1 year after
57 dw ellings
A n area of 57 privately ow
ned but poor quality houses.
The sub-division as a w hole.
U K
, M erseyside,
B urford C
rescent Y
es N
W Pre-Post, E
xp-C on, 1 year before
and 1 year after 67 dw
ellings A
n area of 67 privately ow ned older
houses The sub-division as a w
hole.
U K
, M erseyside,
V illiers C
rescent Y
es N
W Pre-Post, E
xp-C on, 1 year before
and 1 year after 43 dw
ellings A
n area of 43 privately ow ned affluent
suburban houses The sub-division as a w
hole.
U K
, M erseyside,
R ed A
cre Y
es N
W Pre-Post, E
xp-C on, 1 year before
and 1 year after 97 dw
ellings A
n area of 97 dw ellings ow
ned by the local council w
ith poor am enities and high
unem ploym
ent
The sub-division as a w hole.
K now
les, Lesser and M
cK ew
en (1983) U
SA , L
os A ngeles,
M anhattan B
each N
W ,
Property M
arking, Security im
provem ent
Pre-Post, E xp-C
on, 1 year before and 1 year after
31,000 residents
A m
iddle-class beach com m
unity of approxim
ately 31,000 residents covering an area of six square m
iles.
Eight neighborhood jurisdictions
Latessa and T ravis
(1987) U
SA , C
incinnati, C
ollege H ill
B W
, Target hardening
Pre-Post, E xp-C
on, 1 year before and 2 years after
17,327 population
A com
m unity of over 17,000 population
w ithin the city com
prising one-third black and one quarter aged under 18.
The city as a w hole.
15
Lew is, G
rant and R
osenbaum (1988)
U SA
, C hicago,
N orthw
est B
W ,
C om
m unity
organizing, C
om m
unity action
Pre-Post, E xp-C
on, 1 year before and 1 year after
7 census tracts T
he area covered 7 census tracts w ith a
m ean percent w
hite of 91% .
M atched com
parison areas of three equivalent census tracts.
U SA
, C hicago,
N ortheast
B W
, C
om m
unity organizing C
B patrol
Pre-Post, E xp-C
on, 1 year before and 1 year after
2 census tracts T
he area covered 2 census tracts w ith a
m ean percent w
hite of 60% M
atched com parison areas of three
equivalent census tracts.
U SA
, C hicago,
B ack of the Y
ards B
W ,
C om
m unity
organizing
Pre-Post, E xp-C
on, 1 year before and 1 year after
1 census tract T
he area covered 1 census tracts w ith a
m ean percent w
hite of 61% .
M atched com
parison areas of three equivalent census tracts.
U SA
, C hicago,
A uburn-G
resham B
W ,
C om
m unity
organizing
Pre-Post, E xp-C
on, 1 year before and 1 year after
2 census tracts T
he area covered 2 census tracts w ith a
m ean percent w
hite of 0% .
M atched com
parison areas of three equivalent census tracts.
U SA
, C hicago,
E dgew
ater B
W ,
C om
m unity
developm ent,
W histleStop
Pre-Post, E xp-C
on, 1 year before and 1 year after
3 census tracts T
he area covered 3 census tracts w ith a
m ean percent w
hite of 62% .
M atched com
parison areas of three equivalent census tracts.
Low m
an (1983) C
anada, V ancouver
Y es
N W
, Property m
arking Pre-Post, E
xp-C on, 2 years before
and after not stated
A residential district in V
ancouver (W est
E nd South).
O ne of the areas close to the project area
w hich had no substantial border contact
w ith the schem
e location.
M atthew
s and T rickey
(1994a) U
K , Leicester,
N ew
Parks Y
es N
W ,
School W atch,
Target hardening, N
ew sletters,
Sum m
er play schem
e
Pre-Post, E xp-C
on, 2 years before and 17 m
onths after 4,100 dw
ellings A
large residential area on the outskirts of a C
ity com prising m
ainly sem i-detached
housing. T he area w
as dom inantly w
hite w
ith unem ploym
ent rates of tw ice the
national average.
Seven surrounding estates.
M atthew
s and T rickey
(1994b) U
K , Leicester,
E yres M
onsell Y
es N
W ,
School W atch,
Target hardening, N
ew sletters,
Sum m
er play schem
e
Pre-Post, E xp-C
on, 1 year before and 1 year after
3,500 dw ellings
A residential area of m
ainly w hite
population w ith alm
ost half unem ployed
on the outskirts of the city com prising
3,500 dw ellings over an area of 0.7 square
m iles.
Four surrounding estates.
M ukherjee and W
ilson (1988)
A ustralia, V
ictoria N
W Pre-Post, E
xp-C on, 2 years before
and after 617 w
atch schem
es-600- 900 hh per schem
e
A state com
prising 23 police districts. Sixteen of these police districts
had w atch
program s at the tim
e of the evaluation in June 1986. T
he proportion of the population covered in these districts by w
atch schem es varied (6%
to 44% ).
N on-N
W areas w
ithin V ictoria
16
R esearch and Forecasts
Inc. (1983) U
SA , M
ichigan, C
rary-St. M ary’s
Y es
N W
M eetings,
Security surveys
Pre-Post, E xp-C
on, 1 year before and 2 years after
12,880 population
A 155-block m
ainly residential area w ith a
population of about 12,880. The area w as
about 65% w
hite. M ost of the hom
es w ere
houses w ith som
e apartm ents and som
e com
m ercial buildings.
A n area w
ith sim ilar dem
ographic and crim
e characteristics four m iles aw
ay from
the experim ental area.
Tilley and W ebb (1994)
U K
, B irm
ingham ,
Prim rose estate
Y es
N W
, Target hardening,
Pre-Post, E xp-C
on, 1 year before and 2 years after
885 dw ellings
A predom
inantly w hite residential
neighborhood on the edge of a city w ith
three-quarters of household on state benefits and one-third single-parent fam
ilies
The rem ainder of the sub-division
U K
, R ochdale,
B elfield
Y es
H W
, Target hardening, Property m
arking
Pre-Post, E xp-C
on, 1 year before and 2 years after
668 dw ellings
A public housing area tw
o m iles from
the city centre w
ith a large A sian com
m unity
and high unem ploym
ent rate.
The city as a w hole
U K
, R ochdale,
B ack O
’ T h’ M
oss Y
es H
W ,
Target hardening, Property m
arking
Pre-Post, E xp-C
on, 1 year before and 2 years after
A bout 700
dw ellings
A public housing area w
ith som e
affordable private hom es on the outskirts
of the city w ith a m
ainly w hite population
and high unem ploym
ent rate.
The city as a w hole
V eater (1984)
U K
, B ristol,
K ingsdow
n Y
es N
W ,
Security surveys, Property m
arking
Pre-Post, E xp-C
on, 1 year before and 1 year after
5,000 population
A residential area of m
ainly older properties som
e of w hich
w ere run dow
n and som
e renovated including both private and public housing.
A control area of sim
ilar residential com
position in an adjacent neighborhood
17
RESULTS
Two methods can be used to summarize the results of the selected studies. The first is a narrative review, which presents details of the studies and the results obtained. The findings are presented in the form of the relative percentage change in crime in the experimental and control areas. The review also includes the author(s) conclusion and other textual comments found in the research publication. The second method is a meta-analysis, which involves recalculating the published findings to produce a common effect size across studies. The main advantage of a narrative review is that it is possible to include more studies in the review. The main disadvantage is that it is difficult to obtain an overall finding for all studies combined. The main advantage of a meta-analysis is that a single weighted mean effect size can be calculated for groups of studies or all studies combined. The main disadvantage is that it can only be used when there is sufficient information provided in the original report to conduct the analysis. In the following section we present the findings of both methods.
Narrative review
The results of the narrative review are presented in Table 3 below. This is followed by short paragraphs describing the methods and findings of each study in turn.
In order to determine the overall effectiveness of neighborhood watch for the narrative review, it was necessary to determine whether or not the program was effective in reducing crime. In the current review, effectiveness was determined by calculating the relative percentage change of the experimental and comparison area over time. Studies were excluded from the analysis if they did not provide the data that would enable the percentage change to be calculated (e.g. if the results were presented in graphical form only). In total, 24 of the 43 evaluations presented the necessary data. If the experimental area outperformed the comparison area (i.e. crime decreased by more or increased by less), the program was deemed to have a positive effect on crime. If the comparison area outperformed the experimental area, the program was deemed to have a negative effect.
In those cases where there was more than one eligible finding presented within an evaluation, one finding was selected using the system of prioritization outlined above (i.e. findings relating to residential burglary were selected as first priority, followed by findings for property crime and findings for all crimes).
The results of the narrative review show that 19 of the 24 evaluations included in the analysis found that neighborhood watch was associated with a reduction in crime based on relative percentage change (as described above). Conversely, five evaluations found that neighborhood watch was associated with an increase in crime. Technically, it is possible that neighborhood watch could have caused the increase in crime. This could be because watch programs attract offenders as they might suggest that there is something worth stealing in the area. However, it is also possible that watch programs increase recorded crime as a result of increases in reporting rates among residents. Overall, the majority of studies included in the narrative review show that neighborhood watch is associated with a reduction in crime.
18
T a b le
3 : O
u tc
o m
e e
f f e c tiv
e n
e s s o
f n
e ig
h b o r h
o o d w
a tc
h
A uthor/ publication date
of m ain report used in
the review
C ountry/C
ity/area D
ata source
O utcom
e m easure
E xperim
ental area (pre to post) crim
e changes
C om
parison area (pre to post) crim
e changes
R esult
% crim
e difference R
elative % change
O utcom
e (based on relative %
change)
A nderton (1985)
U K
, C heshire,
N orthw
ich D ivision
PD B
urglary 7436-6720
41149-42332 Exp -10% C
on +3% -13%
Positive
B ennett (1990)
U K
, L ondon,
W im
bledon * SR
B urglary
5.1% -4.0%
2.9% -2.1%
Exp -21.6% C
on -27.6%
+6% N
egative
U K
, L ondon, A
cton SR
B urglary
5.9% -8.1%
2.9% -2.1%
Exp +37.3% C
on -27.6% +65%
N egative
B ennett and Lavrakas (1989)
U SA
, B altim
ore SR
V ictim
ization no data
no data D
iff. change = neg. N
on-num erical result
- U
SA , B
oston ditto
ditto ditto
ditto D
iff. change = none N
on-num erical result
- U
SA , B
ronx ditto
ditto ditto
ditto D
iff. change = none N
on-num erical result
- U
SA , B
rooklyn ditto
ditto ditto
ditto D
iff. change = none N
on-num erical result
- U
SA , C
leveland ditto
ditto ditto
ditto D
iff. change = pos. N
on-num erical result
- U
SA , M
iam i
ditto ditto
ditto ditto
D iff. change = none
N on-num
erical result -
U SA
, M inneapolis
ditto ditto
ditto ditto
D iff. change = none
N on-num
erical result -
U SA
, N ew
ark ditto
ditto ditto
ditto D
iff. change = neg. N
on-num erical result
- U
SA , Philadelphia
ditto ditto
ditto ditto
D iff. change = none
N on-num
erical result -
U SA
, W ashington
ditto ditto
ditto ditto
D iff. change = none
N on-num
erical result -
C irel et al. (1977)
U SA
, W ashington,
Seattle SR
B urglary
6.2% -2.4%
10.4% -10.0%
Exp -61.3% C
on -4.0% -57%
Positive
Forrester, C hatterton and
Pease (1988) U
K , R
ochdale, K irkholt
PD B
urglary 512-317
2843-2880 Exp -38% C
on +1% -39%
Positive
H enig (1984)
U SA
, W ashington D
C ,
T he First Police D
istrict PD
B urglary
4-0 2745-1778
Exp -100% C
on -35% -65%
Positive
H ulin (1979)
U SA
, C alifornia,
Fontana PD
B urglary
no data no data
Exp -25.7% C
on +10%
-36% Positive
H usain (1990)
U K
, B irm
ingham ,
H andsw
orth W ood
PD B
urglary graphical data only
graphical data only
N o change in expected
and observed crim e totals
before and after the schem
e launch
N on-num
erical result -
U K
, B righton, Saltdean
PD B
urglary graphical data only
graphical data only
B urglary rates w
ere below
expected levels for the first year of the schem
e and then raised above expected values
N on-num
erical result -
U K
, B urnley, H
argher C
lough PD
B urglary
graphical data only graphical data only
There w as no clear pattern
in the differences betw een
N on-num
erical result -
19
the schem e and
com parison areas over
tim e
U K
, M anchester,
T rafford, Stretford
PD B
urglary graphical data only
graphical data only
In the first year of the schem
e the num ber of
burglaries w ere higher
than expected and in the second year they w
ere low
er
N on-num
erical result -
U K
, Preston, G range
E state and M
oor N ook
PD B
urglary graphical data only
graphical data only
In the first year of the schem
e burglary rates w
ere as expected and in the second year low
er than expected
N on-num
erical result -
U K
, Sutton C oldfield
PD B
urglary graphical data only
graphical data only
B urglary levels w
ere higher than expected during the first 18 m
onths of the schem
e and low er
than expected in the next 6 m
onths
N on-num
erical result -
Jenkins and Latim er (1986)
U K
, M erseyside, Y
ork A
venue PD
B urglary
8-6 378-384
Exp -25% C
on +2% -27%
Positive
U K
, M erseyside,
B urford C
rescent PD
B urglary
1-12 393-472
Exp +1100% C
on +20% +1,080%
N egative
U K
, M erseyside, V
illiers C
rescent PD
B urglary
4-1 880-628
Exp -75% C
on -29% -46%
Positive
U K
, M erseyside, R
ed A
cre PD
B urglary
17-5 1006-754
Exp -71% C
on -25% -46%
Positive
K now
les, Lesser and M
cK ew
en (1983) U
SA , L
os A ngeles,
M anhattan B
each PD
B urglary
no data no data
Exp -27.7% C
on +12.9% -41%
Positive
Latessa and T ravis (1987)
U SA
,C incinnati,
C ollege H
ill PD
B urglary
no data no data
Exp -11% C
on -2% -9%
Positive
Lew is, G
rant and R
osenbaum (1988)
U SA
, C hicago,
N orthw
est SR
V ictim
ization m
ean 0.92-0.73 m
ean 0.86-0.77 Exp -21% C
on -11% -10%
Positive
U SA
, C hicago,
N ortheast
SR V
ictim ization
m ean 1.0-1.23
m ean 1.23-0.9
Exp +23% C
on -27% +50%
N egative
U SA
, C hicago, B
ack of the Y
ards SR
V ictim
ization m
ean 1.3-1.43 m
ean 1.32-1.14 Exp +10% C
on -18% +28%
N egative
U SA
, C hicago, A
uburn- G
resham SR
V ictim
ization no sig. change
not stated Exp no sig. change
N on-num
erical result -
U SA
, C hicago,
E dgew
ater SR
V ictim
ization no sig. change
not stated Exp no sig. change
N on-num
erical result -
Low m
an (1983) C
anada, V ancouver
PD B
urglary 145-97
14-14 Exp -33% C
on 0% -33%
Positive
20
M atthew
s and T rickey
(1994a) U
K , Leicester, N
ew
Parks PD
B urglary
226-181 174-145
Exp -20% C
on -17% -3%
Positive
M atthew
s and T rickey
(1994b) U
K , Leicester, E
yres M
onsell PD
B urglary
231-286 183-266
Exp +24% C
on +12% -21%
Positive
M ukherjee and W
ilson (1988)
A ustralia, V
ictoria PD
B urglary
graphical data only graphical data only
Police divisions identified as having high levels of N
W show
ed greater reductions in residential burglary than those identified as having low levels or no N
W .
N on-num
erical result -
R esearch and Forecasts Inc.
(1983) U
SA , M
ichigan, C rary-
St. M ary’s
PD B
urglary 253-131
206-197 Exp -48% C
on -4% -44%
Positive
Tilley and W ebb (1994)
U K
, B irm
ingham ,
Prim rose estate
PD B
urglary 37-22
104-93 Exp -41% C
on -11% -30%
Positive
U K
, R ochdale, B
elfield PD
B urglary
40-40 1205-1352
Exp 0% C
on +12% -12%
Positive
U K
, R ochdale, B
ack O ’
T h’ M
oss PD
B urglary
40-35 1205-1352
Exp -13% C
on +12% -25%
Positive
V eater (1984)
U K
, B ristol, K
ingsdow n
PD A
ll crim es
652-486 467-610
Exp -25% C
on +31% -56%
Positive
N otes: SR
=Self report. PD =Police data. Positive = a reduction in crim
e. N egative = an increase in crim
e.
21
Short summaries of the 19 studies (covering 43 evaluations) included in the narrative review are presented below.
Anderton (1985) Anderton (1985) conducted an evaluation of a ‘Home Watch’ scheme in Northwich in Cheshire. This was one of the first evaluations of neighbourhood watch in the UK. The study was based on a comparison of police-recorded crimes measured 18 months before and 30 months after the launch of the scheme. The crime rates for Northwich were compared with the crime rates for Cheshire as a whole. The results showed that the number of burglaries in Northwich decreased by 10 per cent, compared with an increase of three per cent across the county as a whole. Anderton (1985) concluded that, ‘It appears from the experience in Cheshire so far that Home Watch is one of the most effective, efficient and successful crime prevention initiatives ever undertaken’ (p.53).
Bennett (1990) Bennett (1990) evaluated the effectiveness of neighbourhood watch schemes in two areas of London (Wimbledon and Acton). The evaluations were based on crime and public attitude surveys in the two areas before the schemes were implemented and again one year after their implementation. Similar surveys were conducted in matched comparison areas some distance from the experimental areas. In Wimbledon, crime decreased by a greater amount in the control area than in the experimental area (28 per cent compared with 22 per cent). In Acton, crime increased by 37 per cent in the experimental area and decreased by 28 per cent in the control area. The author concluded that the findings were ‘not encouraging’ (p.110). Overall, the results suggested that residents in the neighbourhood watch areas experienced either no better or worse rates of victimization than in the comparison.
Bennett and Lavrakas (1989) Bennett and Lavrakas (1989) investigated the effectiveness of neighbourhood watch schemes in 10 US cities (Baltimore, Boston, Bronx, Brooklyn, Cleveland, Miami, Minneapolis, Newark, Philadelphia and Washington). The research was based on a pretest - posttest design with a non-equivalent control groups. The comparison areas were selected by drawing a ‘ring’ around the experimental area approximately two census tracts wide. Monthly crime statistics revealed no differences between the experimental and control areas in seven of the ten evaluations and a negative differential change (where crime decreased less in the experimental area than in the comparison area) in two of the cities. Only one area showed a positive differential change (where the experimental area experienced a larger decrease in crime than the control). The authors concluded that the programs ‘did not seem to achieve the ‘ultimate’ goal of crime reduction’ (p.361).
Cirel et al. (1977) Cirel et al. (1977) conducted one of the first evaluations of the effectiveness of neighbourhood watch in the United States. The evaluation, based in Seattle, Washington, included a telephone and door-to-door surveys of residents one year before the launch of the scheme and one year after. Two census tracts adjacent to the neighbourhood watch area was used as a comparison. The results showed that the rate of burglary decreased by a substantially greater amount in the experimental areas than in the control areas (61 per cent compared with 4 per cent). The authors concluded that participating in community crime prevention, ‘significantly reduces the risk of residential burglary victimization’ (p.79).
22
Forrester, Chatterton and Pease (1988) Forrester, Chatterton and Pease (1988) evaluated a burglary prevention project in Kirkholt, an area of public housing near Rochdale (a town 10 miles north of Manchester) in the UK. A package of measures was introduced as part of the project, including ‘cocoon’ neighbourhood watch. The evaluation was based on the analysis of pre- and post-test police-recorded crime rates in the experimental area (Kirkholt) which were compared with crime rates in the remainder of the police sub-division. The results showed that domestic burglaries decreased by 38 per cent in the experimental area compared with one per cent in the remainder of the sub-division. The authors concluded that there had been a, ‘large absolute and proportionate reduction in domestic burglary during the initiative’ (p.19).
Henig (1984) Henig (1984) conducted an evaluation of neighbourhood watch in a police district in Washington, DC. The impact of block watch on crime was assessed by examining the levels of police-recorded crime in the sample of participating blocks in the year before and after the scheme had been launched. This was compared with crime rates for the police district as a whole and for the city as a whole. The results showed that over the evaluation period the level of burglary decreased by 100 per cent (from 4 to 0 burglaries) in the sample area and by 35 per cent (from 2745 to 1778 burglaries) in the police district as a whole. The author concluded that neighbourhood watch was associated with a reduction in burglary among participating blocks.
Hulin (1979) Hulin (1979) evaluated the effectiveness of a neighbourhood watch scheme in a high crime area of Fontana, California. Using police-recorded crime data for the year before and the year after the scheme, the author compared changes in residential burglary rates in Fontana with changes in burglary in four demographically similar control areas with similar pre-test crime rates. The results showed a decrease in residential burglary of more than 25 per cent in the experimental area compared with increases ranging from 10 to 25 per cent in each of the control areas. Hulin (1979) concluded that the results were ‘positive’ and indicated that neighbourhood watch was ‘an effective crime prevention instrument’ (p.30).
Husain (1990) Husain (1990) evaluated the effectiveness of neighborhood watch schemes in six UK cities (Birmingham, Brighton, Burnley, Manchester, Preston and Sutton Coldfield). In each evaluation, police-recorded crime data were used to compare crime rates before and after implementation of neighborhood watch in the experimental areas. These changes were then compared with changes in crimes rates in six control areas. The findings were presented mainly in graphical form showing relative percentage change in crime in the program areas. In three of the six areas, the implementation of neighborhood watch was accompanied by an improvement in crime. In the other three areas, there were no improvements. The author concluded ‘...in three of the six areas the introduction of NW has been accompanied by some improvement in the crime situation (p.66)’, while, ‘...the results from three other areas are less convincing (p.67).
Jenkins and Latimer (1986) Jenkins and Latimer (1986) conducted evaluations of neighbourhood watch schemes in four areas of Merseyside in the UK. Each of the four evaluations examined the number of crimes recorded by the police in the year before and the year after the scheme had been implemented. In three of the four areas, the experimental area experienced larger decreases in
23
the number of burglaries than in the sub-division as a whole. In the fourth area (Burford Avenue), burglary increased by more than 1,000 per cent (from 1 to 12). The authors concluded that there is, ‘an indication that Homewatch is having an effect, certainly initially, in reducing the instances of burglary within an area and to a lesser extent the total crime’ (p.12). However, they warned that results of the Burford Avenue scheme ‘should not be ignored and indicate that Homewatch is not a panacea for reducing crime’ (p.12).
Knowles, Lesser and McKewen (1983) Knowles, Lesser and McKewen (1983) evaluated the effectiveness of a neighbourhood watch programme in a residential suburb on the western boundary of Los Angeles County in the USA. The evaluation examined changes in the rate of police-recorded residential burglaries in the 12 months before and the 12 months after the programme had been implemented. These were compared with burglary rates in comparison areas (comprising eight neighbouring jurisdictions). The results showed a decrease in burglary of 28 per cent in the experimental area, compared with an increase of 13 per cent in the comparison area. The authors explained that the atmosphere of co-operation fostered by the programme, ‘provided for the achievement of a common goal – crime control’ (p 38).
Latessa and Travis (1987) Latessa and Travis (1987) conducted an evaluation of a block watch programme implemented in the College Hill area of Cincinnati in the USA. College Hill is described by the authors as the fifth largest community in the city with a population of over 17,000 residents. Using police-recorded crime data, burglary rates in College Hill in the year before and after the scheme were compared with burglary rates in the city of Cincinnati as a whole. The figures show that burglary in the experimental area decreased by 11 per cent, while burglary in Cincinnati as whole decreased by two per cent. The authors concluded that College Hill experienced a decrease in the amount of recorded crime during the course of the programme.
Lewis, Grant and Rosenbaum (1988) In another US study, Lewis, Grant and Rosenbaum (1988) evaluated the effectiveness of five block watch schemes in Chicago, Illinois. Crime and public attitude surveys were conducted in the experimental and matched control areas before the launch of the schemes and again one year after the launch. Only one of the five experimental areas experienced a reduction in victimizations. Two of the experimental areas, however, showed a statistically significant increase in victimizations per respondent. The authors concluded in their original report that the results, ‘force us to seriously address the possibility of both theory failure and program failure in this field’ (Rosenbaum, Lewis and Grant 1985, p.170).
Lowman (1983) Lowman (1983) investigated the effectiveness of neighbourhood watch in a residential district of Vancouver, Canada. The evaluation was based on a comparison of crime rates in an experimental area (the neighbourhood watch pilot project area) and three control areas in which neighbourhood watch had not been implemented. The results showed that the number of burglaries decreased by 33 per cent in the experimental area with no change in the comparison areas. The author concluded that the reduction in the experimental area ‘may be indicative of a deterrent effect of the program’ (p.295).
Matthews and Trickey (1994) Matthews and Trickey (1994) conducted an evaluation of a neighbourhood watch scheme in the New Parks area of Leicester in the UK. Police-recorded crime data were used to
24
determine changes in crime rates in the experimental area in the 12 months before and after the launch of the scheme. Comparable data were obtained for seven nearby control areas. The results showed that the number of burglaries decreased in the experimental area and increased in the control area. However, in the following year the rate of burglary increased. The authors explained that this reduction in burglary was ‘welcome’ but somewhat ‘short-lived’ (p 67).
Matthews and Trickey (1994) In a second evaluation of neighbourhood watch in Leicester, Matthews and Trickey (1994) evaluated the effectiveness of a neighbourhood watch scheme on the Eyres Monsell housing estate. Police data were used to examine changes in the number of burglaries in the year before the launch of the scheme and in the year following implementation. Data were also collected for four other housing estates in the area close to the Eyres Monsell estate. Over the study period, the number of burglaries on the Eyres Monsell estate increased by 24 per cent. The number of burglaries on the Saffron Lane estate (the estate with the most similar pre-test burglary rate) also increased over the study period, but the increase was approximately half that of the experimental area (12 per cent). The authors concluded that the outcome of the project as a whole was positive, although ‘not particularly remarkable’ (p.50). However, the rapid increase in the number of burglaries in 1994 was ‘a cause of considerable concern’ (p.50).
Mukherjee and Wilson (1988) Mukherjee and Wilson (1988) evaluated the effectiveness of neighborhood watch in Australia focusing on the state of Victoria. Using police data, the authors compared changes in crime rates in areas with high levels of neighborhood watch with areas that had medium, low levels, or no neighborhood watch. The findings were presented in terms of whether neighborhood watch had a ‘good’, ‘average’ or ‘poor’ effect on reducing crime over the two-year period evaluation period. The results showed that police divisions with high levels of neighborhood watch showed greater reductions in residential burglary than those with low levels or no neighborhood watch. The authors concluded that their findings ‘lend very reasonable support to the objective of neighborhood watch in suppressing residential burglary’ (p.5).
Research and Forecasts Incorporated (1983) In a US study, Research and Forecasts Incorporated (1983) conducted an evaluation of neighbourhood watch in a residential area of Detroit, Michigan. The study used police data to compare changes in crime rates in 155-block experimental area (Crary-St Mary’s) with changes in a matched control area four miles away. In both neighbourhoods, crime rates for the 12 month period before and after implementation of neighbourhood watch were examined. The results showed that burglary rates decreased by a substantially greater amount in the experimental area than in the control area (48 per cent compared with 4 per cent). The authors explained that reported crime statistics showed, ‘a substantial reduction in Crary-St Mary’s that is not matched by the statistics for the control neighbourhood’ (p.34).
Tilley and Webb (1994) Tilley and Webb (1994) present findings from 11 evaluations of individual burglary reduction schemes implemented as part of the Safer Cities Program in the UK. Three of the 11 evaluations (in Birmingham-Primrose estate, Rochdale-Belfield estate and Rochdale-Back O’Th’Moss estate) met the eligibility criteria for inclusion in this review. Each evaluation employed a pretest-posttest research design and compared crime rates in the experimental area with crime rates in either the remainder of the sub-division or in the city as a whole. In all three evaluations, the experimental area outperformed the control area. In the two
25
Rochdale evaluations, the control area experienced an increase in crime, while the experimental area experienced a decrease or remained stable. In Birmingham, both the experimental and control areas experienced a decrease in crime, but the decrease was greater in the experimental area (41 per cent compared with 11 per cent). The authors described the schemes as a, ‘great success’ in terms of ‘reducing crime and as an example of crime prevention work’ (p.4).
Veater (1984) In an early UK study, Veater (1984) evaluated a neighbourhood watch scheme in Kingsdown, Bristol. The evaluation was based on pre-test and post-test victim and public attitude surveys conducted in the scheme area. A comparison was also made of crime rates in an adjacent area to the neighbourhood watch scheme using police-recorded crimes. The findings showed that crime decreased by 25 per cent in the experimental area, but increased by 31 per cent in the comparison area. The author noted that the increase might be a result of crime displacement. He concluded that, ‘the neighbourhood watch concept has potential if adequate resources are made available ...’ (p.5).
Meta-analysis
In order to carry out a meta-analysis of the effects of neighborhood watch, a comparable effect size measure is needed for each evaluation, together with its variance (see Lipsey and Wilson, 2001). All evaluations included in the review employed the same research design (pre-test and post-test measures for experimental and control groups). The majority (n=15) of evaluations used police-recorded data to provide an outcome measure of crime. The remainder (n=3) used self-report victimization surveys. The two types of data require different methods to obtain an odds ratio (OR). These methods are described below.
The outcome measure in each study was the number of crimes (i.e. burglaries, property crimes, or all crimes, in that order) recorded by the police (police-recorded crime data) or the number of people victimized (survey data). There were no evaluations included in the review that provided sufficient information (i.e. standard deviations) to allow ORs to be calculated from mean offending rates. Hence, the meta-analysis is based on ORs derived solely from frequencies or proportions.
Police-recorded crime data
The best measure of effect size for findings based on crimes and victimization is the OR. In practice, the differences in the levels of crime in experimental and control areas recorded in police crime data are not strictly speaking ORs as they are based on events rather than proportions of people experiencing the event. However, for simplicity and clarity the term has been used throughout to describe differences found in both police data and survey data.
26
The OR is calculated as shown in the following table.
Before intervention After intervention Experimental a b Control c d
where a, b, c, d are numbers of crimes OR = a*d/b*c
The null, or no effect, value of the OR is 1.0. To the extent that the OR exceeds 1.0, it might be concluded that the intervention (i.e. neighborhood watch) was possibly beneficial. To the extent that the OR falls below 1.0, it might be concluded that the intervention was possibly harmful. It is technically possible that some schemes might serve to increase the number of recorded crimes (e.g. it has sometimes been argued that increased surveillance will lead to an increase in the number of crimes reported to the police).
The variance of the OR is calculated from its natural logarithm (LOR).
VAR (LOR) = 1/a + 1/b + 1/c +1/d
In order to produce a summary effect size in a meta-analysis, each effect size (here, LOR) is weighted by the inverse of its variance (1/V). This estimate of the variance is based on the assumption that total numbers of crimes (a, b, c, d) have a Poisson distribution. If the number of crimes has a Poisson distribution, its variance should be the same as its mean. However, the large number of changing extraneous factors may cause over-dispersion; that is, where the variance of the number of crimes VAR may exceed the number of crimes N. The analysis was therefore adjusted to deal with the problem of possible ‘over-dispersion’ (i.e. greater than expected variance). Hence, the standard formula for V(LOR) was multiplied by an over- dispersion factor D, where
D = VAR/N.
Farrington, Gill, Waples, and Argomaniz (2007) estimated VAR from monthly numbers of crimes and found the following equation:
D = .0008*N + 1.2
D increased linearly with N and was correlated (.77) with N. The median number of crimes in their study was 760, suggesting that the median value of D was about 2. However, Farrington et al. (2007) argued that this is an overestimate because the monthly variance is inflated by seasonal variations, which do not apply to N and VAR. Nevertheless, in order to obtain a conservative estimate, V(LOR), calculated from the usual formula above, was doubled in all cases involving police-recorded crime data. This adjustment corrects for over- dispersion within studies, not for heterogeneity between studies. In order to test the effects of assuming different over-dispersion factors, the value of V(LOR) was also trebled (rather than doubled) and the results were recalculated. The results showed that there was no change in OR for the fixed effects method and only a very small change in OR for the random effects method. There was also either no change or a very small change in the confidence intervals across the methods.
27
Survey data
For studies based on self-report surveys of residents before and after the intervention, the OR was calculated from the log of OR (LOR) using the formula below:
LOR = Ln (a2*d2/b2*c2) - Ln (a1*d1/b1*c1)
where a2, b2, c2, d2 are ‘after’ numbers and a1, b1, c1, d1 are ‘before’ numbers.
Before intervention After intervention Crime No crime Crime No crime
Experimental a1 b1 a2 b2 Control c1 d1 c2 d2
The variance of LOR is calculated using the following formula:
VAR (LOR) = 1/a1 + 1/b1 + 1/c1 + 1/d1 + 1/a2 + 1/b2 + 1/c2 + 1/d2
This method is based on comparing before and after ORs. This was considered preferable to comparing after ORs only as this would not control for pre-existing differences between the experimental and control areas.
Individual effect sizes
Table 4 summarizes the 18 evaluations included in the meta-analysis. The table shows that 15 evaluations had an OR greater than one and three had an OR less than one. Hence, in the majority of evaluations, neighborhood watch was associated with a reduction in crime. Four of the 15 evaluations with an OR of greater than one were statistically significant (Research and Forecasts Inc. 1983, Anderton 1985, Veater 1984 and Forrester et al. 1988). This can be seen graphically in the forest plot shown in Figure 1. The graph shows a clear pattern of small positive effects.
Mean effect sizes
An important aim of a meta-analysis is to calculate a weighted mean effect size (here, the OR).
There are two ways of calculating the weighted mean effect size. In the case of the fixed effects (FE) method, each effect size is weighted by the inverse of its variance (1/VAR), so that studies based on larger samples are given greater weight than those based on smaller samples. The FE method is based on the assumption that the effect sizes are homogeneous in the sense that they are all drawn from a random distribution of effect sizes about some mean. However, the effect sizes might violate this assumption and be significantly heterogeneous. One method of addressing the problem of heterogeneity is to use the ‘random effects’ model. The random effects (RE) method minimizes heterogeneity by adding a constant to the variance of each effect size (for the formula, see Lipsey and Wilson, 2001, p.119).
28
Fixed effects model
Table 4 shows that the weighted mean OR for the 18 evaluations combined was 1.19 using the FE model. This finding was statistically significant (z=7.25, p<.0001). An OR of 1.19 can be interpreted to mean that crime increased by 19 per cent in the control area compared with the experimental area or decreased by 16 per cent (1/OR) in the experimental area compared with the control area.
Random effects model
The 18 studies were significantly heterogeneous according to the Q statistic (Q=35.72, 17 d.f., p<.0001). Therefore, the RE model was used. The weighted mean OR for the 18 evaluations combined was 1.36 using the RE model. An odds ratio of 1.36 means that crime increased by 36% in the control area compared with the experimental area or decreased by 26% in the experimental area compared with the control area.
We should emphasize that our estimate of the variance of ORs, while the best available at present, are not exact figures and may be slightly inaccurate. Therefore, there may be some inaccuracy in the weightings used in our meta-analyses. The main consequence of this is that the confidence intervals around the weighted mean effect sizes may be slightly inaccurate.
29
Table 4. Individual and mean effect sizes
Author Date Outcome measure
OR CI z p of z
Cirel et al. (1977) 1977 Burglary 2.38 0.87-6.53 1.69 ns Lowman (1983) 1983 Burglary 1.49 0.49-4.53 0.71 ns Research and Forecasts Inc. (1983) 1983 Burglary 1.85 1.23-2.77 2.96 <0.004 Henig (1984) 1984 Burglary 2.59 0.12-57.52 0.60 ns Anderton (1985) 1985 Burglary 1.14 1.08-1.20 5.03 <0.0001 Veater (1984) 1985 All crimes 1.75 1.38-2.22 4.61 <0.0001 Jenkins and Latimer (1986) [1] 1986 Burglary 1.35 0.30-6.13 0.39 ns Jenkins and Latimer (1986) [2] 1986 Burglary 0.10 0.01-1.80 -1.56 ns Jenkins and Latimer (1986) [3] 1986 Burglary 2.85 0.13-63.52 0.66 ns Jenkins and Latimer (1986) [4] 1986 Burglary 2.55 0.62-10.51 1.29 ns Forrester, Chatterton and Pease (1988) 1988 Burglary 1.64 1.32-2.02 4.57 <0.0001 Bennett (1990) [1] 1990 Burglary 0.92 0.27-3.11 -0.13 ns Bennett (1990) [2] 1990 Burglary 0.51 0.16-1.65 -1.12 ns Matthews and Trickey (1994a) 1994 Burglary 1.04 0.69-1.58 0.19 ns Tilley and Webb (1994) [1] 1994 Burglary 1.50 0.65-3.50 0.95 ns Tilley and Webb (1994) [2] 1994 Burglary 1.12 0.60-2.11 0.36 ns Tilley and Webb (1994) [3] 1994 Burglary 1.28 0.67-2.46 0.75 ns Matthews and Trickey (1994b) 1994 Burglary 1.17 0.82-1.69 0.87 ns
Total n=18
Fixed Effects 1.19 1.13-1.24 7.25 <0.0001 Random Effects 1.36 1.15-1.61 3.63 <0.0004
Notes: An odds ratio of 1.19 means that crime increased by 19% in the control area compared with the experimental area or decreased by 16% in the experimental area compared with the control area (1/OR). An odds ratio of 1.36 means that crime increased by 36% in the control area compared with the experimental area or decreased by 26% in the experimental area compared with the control area.
30
Figure 1: The effectiveness of neighborhood watch
Jenkins and Latimer (1986) [3] Henig (1984) Jenkins and Latimer (1986) [4] Cirel et al. (1977) Research and Forecasts Inc. (1983) Veater (1984) Forrester et al. (1988) Tilley and Webb (1994) [1] Lowman (1983) Jenkins and Latimer (1986) [1] Tilley and Webb (1994) [3] Matthews and Trickey (1994b) Anderton (1985) Tilley and Webb (1994) [2] Matthews and Trickey (1994a) Bennett (1990) [1] Bennett (1990) [2] Jenkins and Latimer (1986) [2] Fixed Effects Random Effects
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
0.00 0.01 0.10 1.00 10.00 100.00
Odds Ratio Weighted mean FEOR=1.19; p=.0001; Q=35.72; p<.005 Weighted mean REOR=1.36; p=.0003; Q=13.35; p=ns
Moderator analyses
Overall, the meta-analysis has shown, using both the FE and RE models, that neighborhood watch was associated with a significant reduction in crime. However, it is possible that the results will vary by specific characteristics of the program being implemented or by the research design of the evaluation. The results of the moderator analyses which investigate this are presented in Table 5 below.
Type of comparison area
It is possible that there are differences in results depending on whether studies used non- equivalent or equivalent comparison areas. It could be argued that research based on non- matched areas are more likely to produce a positive result due to regression to the mean in the experimental area (which might have been selected at a time when crime was unusually high and likely to fall) but not in the comparison area (which might have been selected at a low point in its crime cycle and likely to rise). In order to test for this, the studies included in the meta-analysis were split into two groups based on the nature of the comparison area (i.e.
31
whether it was ‘matched’ or ‘not matched’). The meta-analysis was then repeated. The results showed that the difference between these ORs was statistically significant with studies based on matched areas showing larger effect sizes than those based on unmatched areas. This finding is counter to the effect hypothesized above. One reason for this result is that the ‘not matched’ comparison area comprised the broader police area and sometimes included the experimental area. Under these circumstances it is conceivable that any movements in the broader area will be reflected in the experimental area, which would result in no apparent neighborhood watch effect. The differences could also be explained by other factors or unmeasured differences between the groups.
Type of data
It is also possible that the effectiveness of evaluations varies in terms of the type of data collected. It was argued earlier that the method of calculating ORs was slightly different using the police and survey data and that this might result in different findings. The data are also different in that the latter includes non-reported offences. In order to test for this, the 15 evaluations that collected police data were compared with the three evaluations that collected data from self-report surveys. The results showed that the difference between the two ORs was not statistically significant. Hence, the effectiveness of neighborhood watch programs did not vary by the type of data collected. This provides a justification, therefore, for combining police and survey data in the overall analysis.
Type of scheme
It might be expected that NW schemes based on limited versions of the program might be less likely to show an effect than schemes based on more comprehensive versions. In order to test for this, the studies were split into two groups based on program type (i.e. whether it was NW alone or NW with one or more additional elements of the ‘big three’). The results showed that the mean difference between ORs was not significant. Hence, the type of program did not independently affect outcome.
Size of scheme area
It could be argued that larger schemes might be more effective than smaller schemes on the grounds that a greater number of neighbors are looking out for suspicious behavior. It could also be argued that smaller schemes might be more effective than larger schemes as the interaction between neighbors who know each other well might be more concentrated. Overall, there was no statistical difference in the ORs of larger and smaller schemes.
Year of publication
It might be the case that early schemes might be more effective than later schemes on the grounds that the motivation and interest in the program was highest at its inception. It is also possible that the reverse might be the case with motivation and expertise increasing over time. The results show that there was no significant difference in the outcomes of earlier compared with later schemes.
32
Publications status
Another possible variation in results might relate to publication status. It has been hypothesized that publishers are more likely to publish evidence of success than evidence of no effect or failure. This is sometimes referred to as ‘publication bias’. In order to test for this, evaluations were identified as published or unpublished. Research was defined as published if it was reported in a book, journal or official government report, as these were likely to have been externally reviewed before distribution. Evaluations were defined as unpublished if they were police reports or reports from survey research companies, as these were less likely to have been externally reviewed before distribution. The mean OR was then calculated for each group. The results showed that the difference between the mean ORs was statistically significant. In other words, the results support the publication bias thesis by showing that published evaluations tended to provide evidence of a stronger neighborhood watch effect than unpublished evaluations.
Country
Finally, it is possible that schemes operating in different countries have different effects as a result of a variety of factors including the environmental context, the nature of the program implemented or the methods of evaluation. The mean OR for studies conducted in the USA and Canada was 1.87 (n=4) compared with 1.18 for the UK (n=14). The difference between the ORs was statistically significant (p<0.05). In other words, evaluations of neighborhood watch conducted in the US and Canada were significantly more likely to show a reduction in crime than studies conducted in the UK. It is difficult to explain such variations because of the large number of factors that could potentially affect outcomes. The main measurable difference between the comparison countries was that there were proportionately more matched studies in the US and Canada (3 of 4) than in the UK (5 of 14). It was shown earlier that matched studies more frequently showed a favorable outcome than non-matched studies. However, there are many other plausible explanations for the difference.
33
Table 5. Variations in mean effect sizes by features of the methods and the program using
the Fixed Effects method
No. of studies
OR CI z p of z Q p of Q Sig. of diff in
OR
Type of comparison Matched 8 1.48 1.26-1.73 4.84 <0.0001 11.96 ns <0.01 Not matched 10 1.16 1.11-1.22 6.09 <0.0001 15.74 ns
Type of data Police data 15 1.19 1.13-1.24 7.24 <0.0001 31.72 <0.005 ns Survey data 3 1.14 0.60-2.18 0.40 ns 3.98 ns
Type of scheme NW only 8 1.30 0.90-1.87 1.41 ns 4.65 ns ns NW plus 10 1.19 1.13-1.24 7.13 <0.0001 30.82 <0.001
Size of scheme area Small 11 1.27 0.93-1.72 1.52 ns 8.63 ns ns Large 7 1.19 1.13-1.24 7.10 <0.0001 26.9 <0.01
Year 1977-1988 11 1.19 1.14-1.25 7.20 <0.0001 32.77 <.0.001 ns 1989-1994 7 1.12 0.90-1.39 1.01 ns 2.64 ns
Published Published 8 1.51 1.26-1.80 4.56 <0.0001 6.35 ns <0.01 Not published 10 1.17 1.11-1.22 6.28 <0.0001 21.81 <0.01
Country UK 14 1.18 1.12-1.23 6.86 <0.0001 28.82 <0.01 <0.05 USA/Canada 4 1.87 1.31-2.67 3.47 0.0005 0.43 ns
All studies 18 1.19 1.13-1.24 7.25 <0.0001 35.72 <0.01 ns=not significant
34
CONCLUSION
The results of previous systematic reviews of neighborhood watch presented in the introduction were divided in terms of the conclusions drawn. Titus (1984) concluded that neighborhood watch was effective, but noted that the research methods used to investigate this were weak. Husain (1990) concluded that there was little evidence that neighborhood watch worked. Sherman and Eck (2002) concluded that neighborhood watch was ineffective in reducing crime.
The main findings of our narrative review were that just over half of the schemes evaluated (19) showed that neighborhood watch was effective in reducing crime, while only six yielded negative effects. The main finding of the meta analysis was that neighborhood watch was associated with a relative reduction in crime of between 16 and 26 per cent. The generally positive findings of the narrative review are consistent with the favorable effect found in the meta analysis. Hence, the dominant finding of our review, using both methods, is that neighborhood watch is effective in reducing crime.
However, the limitations of both the methods used in the original studies and in the narrative review and meta analysis need to be taken into account. A particular problem with the original studies is that the experimental and comparison areas were rarely wholly equivalent and sometimes not equivalent at all. The narrative review is limited in that it is based on simple counts of changes in reported crime that provide only a simplified measure of effectiveness. The main problem with the meta analysis is that it is restricted to a not necessarily representative subset of all studies. Also, the estimates of the variance of ORs might be slightly inaccurate.
One notable problem is that the studies used in the narrative review and the meta analysis were different. In order to examine the effect of this on outcome, we split the studies included in the narrative review into two groups. Studies included in the meta analysis were more likely to show a positive effect (78% positive) than those excluded (28% positive). If all studies in the narrative review had been used in the meta analysis then the result would have been less positive. As the meta analysis provides the stronger overall finding, it would have been more difficult to conclude that neighborhood watch was effective.
It is also not immediately clear from the research evidence why neighborhood watch is associated with a reduction in crime. According to theory (see above) neighborhood watch might be effective in increasing surveillance, reducing opportunities and enhancing informal social control. Unfortunately, very few studies provide information on the mechanisms by which neighborhood watch might reduce crime. It is therefore difficult to determine from current research how neighborhood watch works.
Research implications
There are a number of implications that can be drawn from the review for future research on the effectiveness of neighborhood watch.
First, the review has drawn attention to the common problem of a relatively small number of good-quality studies in terms of research design. Among the 27 studies that were excluded on
35
grounds of methodological quality, 19 had no comparison group and 8 presented only post- test data on crime.
Second, coupled with this, it is unclear why evaluations of neighborhood watch stopped abruptly in the mid 1990s. It is possible that researchers felt that the effectiveness or ineffectiveness of neighborhood watch was already established and that there was no need for further investigation. As a result, the effectiveness of neighborhood watch in more recent times is largely unknown. It would have been helpful if more recent evaluations of neighborhood watch had been conducted in order to determine current effectiveness.
Third, none of the studies was based on random allocation of areas to treatment or control conditions. Instead, all studies were based on some version of a quasi-experimental design. This is almost certainly a result of the difficulties involved in implementing community- based programs in areas where communities have not requested them. It is difficult to conduct a randomized experiment with areas as the unit of assignment. However, quasi- experimental designs are not ideal and some writers have argued that they can over-estimate the positive effects of schemes as a result of selection effects whereby the subjects or schemes most likely to change are included in the experimental group (for a discussion see Wilson, Mitchell, and MacKenzie, in press).
Fourth, a particularly important problem for the current review was that less than half of the eligible studies reported data that were suitable for a meta-analysis. This was either because studies presented the results using an unusual statistical notation or left out the data entirely (e.g. when the results were presented in graphical form only). It would be helpful if published evaluations included, at a minimum, raw data, cell sizes and other relevant information in order to facilitate future meta-analyses.
Finally, very few evaluations disaggregated the findings in a way that would show differential effects for subgroups and provide detailed information on the features of the program. As there might be variations in outcome according to the type of program implemented or the type of area in which it is implemented it is important that this information should be included in a research report.
Implications for policy
Neighborhood watch has often been described as one of the most widespread methods of reducing crime. It is supported by UK and US governments and is popular among the public and the police (Sims, 2001). The current review provides some support for this level of implementation. However, little is known about the factors that influence whether or not it is effective. The results of this review have shown that there is some variation across schemes in terms of the outcomes achieved. Governments and those responsible for crime prevention policy should investigate differences between more effective and less effective schemes in order to guide good practice.
36
PLANS FOR UPDATING THE REVIEW
We will update this review every two years.
ACKNOWLEDGEMENTS
We would like to acknowledge the financial support of the Campbell Collaboration and the University of Glamorgan for supporting the preparation of the protocol and review. In particular, we are indebted to David Wilson for his methodological assistance and comments. Helpful comments were also provided by the anonymous reviewers.
STATEMENT CONCERNING CONFLICT OF INTEREST
There is no conflict of interest. However, one of the studies included in the review was conducted by Trevor Bennett.
35
Studies included in the narrative review
R eport
num ber
Study num
ber R
eference
1 1
A nderton, K
.J. 1985 T he E
ffectiveness of H om
e W atch Schem
es in C heshire, C
hester: C heshire C
onstabulary.
2 2
B ennett, T
. 1987.A n E
valuation of T w
o N eighborhood W
atch Schem es in L
ondon: Final R eport to the H
om e O
ffice R esearch and Planning U
nit. C
am bridge: Institute of C
rim inology.
3 B
ennett, T . 1988. A
n assessm ent of the design, im
plem entation and effectiveness of neighborhood w
atch in L ondon. H
ow ard Journal of C
rim inal
Justice, 27, (4), pp. 241-255.
4 B
ennett, T . 1990. E
valuating N eighborhood W
atch. C am
bridge Studies in C rim
inology L X
I. U K
: G ow
er.
5 3
B ennett, S.F. and L
avrakas, P.J. 1989. C om
m unity-based crim
e prevention: an assessm ent of the E
isenhow er Foundation’s N
eighborhood Program
. C rim
e and D elinquency, 35, 3, pp.345-364
6 4
C irel, P., E
vans, P., M cG
illis, D . and W
hitcom b, D
. 1977. C om
m unity C
rim e Prevention Program
, Seattle: A n E
xem plary Project. W
ashington D
.C .: G
overnm ent Printing O
ffice.
7 L
indsay, B .D
. 1979. T he com
m unity crim
e prevention program . In H
.J. Schneider (ed.). V ictim
in International Perspective. Papers and E ssays
G iven at the T
hird International Sym posium
on V ictim
ology 1979 in M unster. W
est G erm
any: W alter de G
ruyter & C
o.
8 L
indsay, B . and M
cG illis D
(1986) C ityw
ide com m
unity crim e prevention: an assessm
ent of the Seattle program . In R
osenbaum (1986)
C om
m unity C
rim e Prevention. D
oes it W ork? L
ondon: Sage.
9 5
Forrester, D ., C
hatterton, M . and Pease, K
. 1988. T he K
irkholt B urglary Prevention Project, R
ochdale.C rim
e Prevention U nit: Paper 13. L
ondon: H
om e O
ffice.
10 Forrester, D
., Frenz, S., O ’C
onnell, M ., et al. 1990. T
he K irkholt B
urglary Prevention Project: Phase II. C rim
e Prevention U nit: Paper 23.L
ondon: H
om e O
ffice.
11 Pease, K
. 1991. T he K
irkholt Project: preventing burglary on a B ritish public housing estate. Security Journal, 2, (2),pp. 73-77.
12 6
H enig, J.R
. 1984. C itizens A
gainst C rim
e: A n A
ssessm ent of the N
eighborhood W atch Program
in W ashington, D
.C . O
ccasional Paper, C enter
for W ashington A
rea Studies. W ashington: G
eorge W ashington U
niversity.
36
R eport
num ber
Study num
ber R
eference
13 7
H ulin, J.O
. 1979. C om
m unity based crim
e prevention project. C rim
e Prevention R eview
, C alifornia State D
epartm ent of Justice, 6, (3),pp.26-34.
14 8
H usain, S. 1989. The Im
pact of N eighborhood W
atch on R ecorded C
rim e. U
rban Policy R esearch U
nit: U niversity of Southam
pton.
15 H
usain, S. 1990. N eighborhood W
atch and C rim
e: A n A
ssessm ent of Im
pact.L ondon: Police Foundation.
16 9
Jenkins, A .D
. and L atim
er, I. 1986. E valuation of M
erseyside H om
ew atch Schem
e. M anagem
ent D evelopm
ent and Force Planning U nit:
M erseyside Police.
17 10
K now
les, L ., Lesser, C
. and M cK
ew en, F. 1983. B
urglary prevention: a citizen initiated and operated neighborhood w atch program
. Police C hief,
50, pp.36-38.
18 11
L atessa, E
.J. and T ravis, L
.F. 1986. E valuation of the C
ollege H ill C
rim e Prevention Program
. U nited States: U
niversity of C incinnati.
19 12
L ew
is, D .A
., G rant, J.A
., and R osenbaum
, D .P. 1988. Social C
onstruction of R eform
: C rim
e Prevention and C om
m unity O
rganizations.N ew
B
runsw ick: T
ransaction B ooks.
20 R
osenbaum , D
.P., L ew
is, D .A
. and G rant, J.A
. 1986. N eighborhood-based crim
e prevention. A ssessing the efficacy of com
m unity organizing in
C hicago. In D
.P. R osenbaum
(ed.), C om
m unity C
rim e Prevention: D
oes It W ork? L
ondon: Sage.
21 R
osenbaum , D
.P. 1987. T he theory and research behind neighborhood w
atch: is it a sound fear and crim e reduction strategy? C
rim e and
D elinquency, 33(Jan), (1),pp.103-134.
22 R
osenbaum , D
.P., L ew
is, D .A
., and G rant, J.A
. 1985. T he Im
pact of C om
m unity C
rim e Prevention Program
s in C hicago: C
an N eighborhood
O rganizations M
ake a D ifference? Final R
eport: V olum
e O ne. Illinois: C
enter for U rban A
ffairs and Policy R esearch, N
orthw estern U
niversity.
23 13
L ow
m an, J. 1983. T
arget hardening burglary prevention and the problem of displacem
ent phenom ena. In T
. Flem ing (ed.) D
eviant D esignations:
C rim
e, L aw
and D eviance in C
anada. T oronto: B
utterw orths: pp. 277-304.
24 14
M atthew
s, R . and T
rickey, J. 1994. E yres M
onsell C rim
e R eduction Project.Leicester: U
niversity of L eicester, C
entre for the Study of Public O
rder.
25 15
M atthew
s, R ., and T
rickey, J. 1994. T he N
ew Parks C
rim e R
eduction Project. L eicester: C
entre for the Study of Public O rder, U
niversity of L
eicester.
26 16
M ukherjee, S. and W
ilson, P. 1988. N eighborhood W
atch: issues and policy im plications. T
rends and Issues N o. 8.A
ustralia: A ustralian Institute
37
R eport
num ber
Study num
ber R
eference
of C rim
inology.
27 17
R esearch and Forecasts Inc. 1983. T
he Figgie R eport, Part IV
: R educing C
rim e in A
m erica –
Successful C om
m unity E
fforts.O hio: Figgie
International.
28 18
T illey, N
. and W ebb, J. 1994. B
urglary R eduction: Findings From
Safer C ities Schem
e, C rim
e Prevention U nit Paper 51. L
ondon: H om
e O ffice.
29 E
kblom , P., Law
, H . and Sutton, M
. 1996. Safer C ities A
nd D om
estic B urglary, H
om e O
ffice R esearch Study 164. L
ondon: H om
e O ffice.
30 19
V eater, P. 1984. E
valuation of K ingsdow
n N eighborhood W
atch Project, B ristol. B
ristol: A von and Som
erset C onstabulary.
38
Studies excluded from the narrative review
E xcluded
study num
ber
R eference
R eason for exclusion
1 A
ldridge,R . 1990. D
on’t botch the w atch. Police R
eview , 98, pp.1046-1047.
N ot an evaluation
2 A
ldridge, R .C
. 1989. N eighborhood W
atch: T he E
xperience of Four Schem es. H
om e O
ffice C rim
e Prevention U
nit. L ondon: H
om e O
ffice. N
ot an evaluation
3 A
lpert, G .P., and Piquero, A
. 1998. C om
m unity Policing: C
ontem porary R
eadings. U nited States: W
aveland Press.
N ot about neighborhood w
atch
4 A
very, J. 1986. N eighborhood w
atch in N ew
South W ales: one year on. A
ustralian P olice Journal, 40, (2),
pp.47-50. N
ot an evaluation
5 B
aker, T .E
., B aker, J.P. and Z
ezza, R . 1999. N
eighborhood w atch: a leadership challenge. FB
I L aw
E
nforcem ent B
ulletin, 68, (2),pp.12-18. N
ot an evaluation
6 B
arton, A . 2000. A
tale of tw o projects: the grow
th and developm ent of tw
o neighborhood w atch schem
es in South W
ales. C rim
e Prevention and C om
m unity Safety A
n International Journal, 2, (3),pp.7-16. N
o control area
7 B
ellair, P.E . 2000. Inform
al surveillance and street crim e: a com
plex relationship. C rim
inology, 38, (1), pp.137-169.
N ot about neighborhood w
atch
8 B
ennett, T . 1988. Im
pact of neighborhood w atch on residents and crim
inals. Journal of Security A
dm inistration, 11, (2),pp.52-59.
N ot an evaluation
9 B
ennett, T . 1989. Factors related to participation in neighborhood w
atch schem es. B
ritish Journal of C
rim inology, 29, (3),pp. 207-218.
N ot an evaluation
10 B
ennett, T . 1992. T
hem es and variations in neighborhood w
atch. In D .J. E
vans, N .R
. Fyfe and D .T
. H erbert,
(eds.), C rim
e, Policing and Place – E
ssays in E nvironm
ental C rim
inology, pp. 272-285. L ondon: R
outledge. N
ot an evaluation
11 B
ennett-Sandler, G . 1979. C
itizen participation in policing: issues in the social control of a social control agency. In, R
.G . Iacovetta, and D
.H . C
hang, (eds.), C ritical Issues in C
rim inal Justice. U
nited States: C
arolina A cadem
ic Press.
N ot about neighborhood w
atch
12 B
ennion, C ., D
aw e, A
., H esse, B
.H ., Joshua, L., M
cG loin, P., M
unn, G . and Tester, S. 1985. N
eighborhood W
atch: T he E
yes and E ars of U
rban Policing? U niversity of Surrey, Sociology D
epartm ent.
N o control area
13 B
eyer, L .R
. 1993. C om
m unity Policing: L
essons from V
ictoria. A ustralia: A
ustralian Institute of C
rim inology.
Post-test only w ithout random
allocation
14 B
rantingham , P.L. and B
rantingham , P.J. 1990. Situational crim
e in practice. C anadian Journal of
C rim
inology, 32, pp.17-40. N
ot an evaluation
15 B
rodeur, J.P. 1995. C om
parisons in Policing: A n International Perspective. U
nited States: A vebury
Publishing C om
pany. N
ot an evaluation
39
16 B
row n, L
. 1992. N eighborhood W
atch: A L
iterature R eview
. E dinburgh: T
he Scottish O ffice C
entral R
esearch U nit.
N ot an evaluation
17 B
row n, L
.P. and W ycoff, W
.A . 1987. Policing H
ouston – reducing fear and im
proving service. C rim
e and D
elinquency, 33, (1), pp.71-89. N
ot about neighborhood w atch
18 B
uerger, M .E
. 1994. T ale of tw
o targets: lim itations of com
m unity anticrim
e actions. C rim
e and D
elinquency, 40, (3),pp.411-436. N
ot an evaluation
19 B
urden, O .P. 1979. C
itizens’ initiative crim e prevention. In, R
.G . Iacovetta, and D
.H . C
hang, (eds.), C ritical
Issues in C rim
inal Justice. U nited States: C
arolina A cadem
ic Press. N
otan evaluation
20 B
urgess, L. 1990. A block president’s perspective on citizen participation and research. A
m erican Journal of
C om
m unity Psychology, 18, (1), pp.159-161.
N ot an evaluation
21 C
arnall, P. 1994. N eighborhood W
atch: Fulfilling the Potential. L ondon: H
om e O
ffice Police D ept.
N ot an evaluation
22 C
lad, J. 1986. N eighborhood w
atch. Far E astern E
conom ic R
eview , 131, pp.15-17.
N ot an evaluation
23 C
larke, R . (ed.). 1992. Situational C
rim e Prevention: Successful C
ase Studies. N ew
Y ork: H
arrow and
H eston. 2
ndE d. 1997.
N ot an evaluation
24 C
onnelly, N .M
., H arries, K
.D ., and H
erbert, D .T
. 1987. D ysfunction in the urban system
; an evaluation of neighborhood crim
e prevention in T ulsa, O
klahom a. Journal of E
nvironm ental System
s, 17, (2), pp. 115-130. Post-test only w
ithout random allocation
25 D
arian-Sm ith, E
. 1993. N eighborhood w
atch – w
ho w atches w
hom ? R
einterpreting the concept of neighborhood: A
ustralia and U nited States. H
um an O
rganization, 52, pp. 83-88. N
ot an evaluation
26 D
arw ick, N
. and H igginbotham
, G .E
. 1984. International policing – special focus. Police C
hief, 51, (6),pp. 31-34.
N ot about neighborhood w
atch
27 de Savornin L
ohm an, P.M
., and V an-D
ijk, A .G
. 1988. N eighborhood W
atch in the N etherlands: T
he L essons
of Six Projects: A Sum
m ary.T
he H ague: N
ational C rim
e Prevention B ureau.
N o control area
28 D
elaney, M . 1983. T
he K ingsdow
n beat experim ent: the w
ay forw ard. A
von and Som erset C
onstabulary. N
ot an evaluation 29
D oleschal, E
. 1984. Prevention of C rim
e and D elinquency. U
nited States: N ational C
ouncil on C rim
e and D
elinquency. N
ot about neighborhood w atch
30 D
onnerm eyer, J.F. and M
ullen, R .E
. 1987. U se of neighbors for crim
e prevention: evidence from a state-
w ide rural victim
study. Journal of the C om
m unity D
evelopm ent Society, 18, (1),pp.15-29.
N o control area
31 D
onnison, H . and Scola, J. 1986. N
eighborhood W atch: Policing the People. L
ondon: L ibertarian R
esearch and E
ducation T rust.
N o control area
32 D
uncan, S. 2001. N ew
successes w ith old problem
s. L aw
and O rder, 49, (10), pp. 101-104.
N ot about neighborhood w
atch 33
E vans, D
.J. and H erbert, D
.T . 1989 G
eography of C rim
e. U nited K
ingdom : R
outledge. N
ot about neighborhood w atch
34 Flem
ing, J. (2004) W orking Together: N
eighborhood W atch, R
eassurance Policing and the Potential of Partnerships. R
egN et A
ustralian N ational U
niversity. N
ot an evaluation
35 G
arofalo, J. and M cleod, M
. 1989. T he structure and operations of neighborhood w
atch program s in the
U nited States. C
rim e and D
elinquency, 35(3), pp. 326-344. N
ot an evaluation
36 G
illham , J.R
. 1992. Preventing R esidential B
urglary: T ow
ard M ore E
ffective C om
m unity Program
s. N ew
Y
ork: Springer-V erlag.
Post-test only w ithout random
allocation
37 G
ore, L .K
. 1986. T he Future of N
eighborhood W atch Program
s: Y ear 2000.C
alifornia: C om
m ission on
N ot an evaluation
40
Peace O fficer Standards and T
raining. 38
G rant, S.E
. 2002. A n A
nalysis of C om
m unity B
lock C lubs as a V
ehicle of Social C apital: U
sing E -911
statistics (N ew
Y ork). N
ew Y
ork: State U niversity of N
ew Y
ork at B uffalo.
N o control area
39 G
reenberg, M .A
. 1991. V olunteer police: the people’s choice for safer com
m unities: advantages over
neighborhood w atch and the guardian angels. Police C
hief, 58,pp. 42-44. N
ot an evaluation
40 G
reenberg, S.W ., R
ohe, W .M
. and W illiam
s, J.R . 1985. Inform
al C itizen A
ction and C rim
e Prevention at the N
eighborhood level. W ashington D
C : U
S D ept of Justice, N
ational Institute of Justice. N
ot an evaluation
41 H
ale, D .C
. and Leonik, R .G
. 1982. Planning com m
unity-initiated crim e prevention. Journal of Police Science
and A dm
inistration, 10, (1), pp. 76-82. N
ot an evaluation
42 H
art, G .T
. 1974. H om
e alert – crim
e prevention through police-citizen cooperation. C rim
e P revention
R eview
, 1, (4), pp.18-24. N
ot an evaluation
43 H
iller, A . 1986. Som
e recent developm ents in com
m unity policing -part 1. A
ustralian Police Journal, 40 (2), pp. 54-62.
N ot an evaluation
44 H
irschfield, A . and B
ow ers, K
.J. 1997. T he effect of social cohesion on levels of recorded crim
e in disadvantaged areas. U
rban Studies, 34, (8) pp.1275-1295.
N o control area
45 H
irst, J. 1988. T he pow
er of the w atch. N
ew Statesm
an and Society, 1, pp. 26-27. N
ot an evaluation 46
H oness, T
., M aguire, M
. and C harm
an, E . 1993. V
ehicle W atch and C
ar T heft: A
n E valuation. C
rim e
Prevention U nit Series N
o. 50. L ondon: H
om e O
ffice Police D epartm
ent. N
ot about neighborhood w atch
47 H
ope, T . 1988. Support for neighborhood w
atch a B ritish C
rim e Survey analysis. In T
. H ope and M
. Shaw
(eds.). C om
m unities and C
rim e R
eduction. L ondon: H
M SO
N ot an evaluation
48 H
ourihan, K . 1987. L
ocal com m
unity involvem ent and participation in N
eighborhood w atch: a case-study in
C ork, Ireland. U
rban Studies, 24,pp.129-136. N
ot an evaluation
49 H
ow ard, M
. 1996. C rim
e w atch. N
ew Scientist, 149, p. 47.
N ot about neighborhood w
atch 50
H usain, S. 1988. N
eighborhood W atch in E
ngland and W ales: A
L ocational A
nalysis. L ondon: C
rim e
Prevention U nit, H
om e O
ffice. N
o control area
51 H
usain, S. and B right, J. (eds.) 1990. N
eighborhood W atch and the Police. Sw
indon: C rim
e C oncern.
N ot an evaluation
52 H
usain, S. and W illiam
s, H . 1993. N
eighborhood W atch: A
G uide For Successful Schem
es.Sw indon: C
rim e
C oncern.
N ot an evaluation
53 Jespersen, A
. 1987. Projecting cuts in crim e. Police R
eview , 95, pp.2206-2207.
N ot an evaluation
54 K
enney, D .J. 1989. Police and Policing: C
ontem porary Issues. U
nited States: Praeger. N
ot an evaluation 55
K essler, D
.A . and D
uncan, S. 1996. Im pact of com
m unity policing in four H
ouston neighborhoods. E
valuation R eview
, 20, (6), pp.627-669. N
o control area
56 K
lein, L ., L
uxenburg, J. and K ing,M
. 1989. Perceived N eighborhood C
rim e and the Im
pact of Private Security. Society for the Study of Social Problem
s (C onference Paper.)
N ot an evaluation
57 L
ab, S.P. 1988. C rim
e Prevention: A pproaches, Practices and E
valuations. C incinnati: A
nderson Publishing C
o. N
ot an evaluation
58 L
atessa, E .J. and T
ravis, L .F. 1987. C
itizen crim e prevention: problem
s and perspectives in reducing crim e.
Journal of Security A dm
inistration, 10, (1),38-51. N
o control area
41
59 L
aV igne, N
.G . and W
artell, J. (eds.). 2000. C rim
e M apping C
ase Studies: Successes in the field (V ol. 2).
W ashington D
C : Police E
xecutive R esearch Forum
. N
ot about neighborhood w atch
60 L
aycock, G ., and T
illey, N . 1995. Policing and N
eighborhood W atch: Strategic Issues. C
rim e D
etection &
Prevention Series Paper, N o. 60.L
ondon: U K
H om
e O ffice Police R
esearch G roup.
N ot an evaluation
61 L
eblanc, A .G
. and W illiam
s, D . 1978. C
rim e prevention –
dollars and sense. Police C hief, 45, (3), pp.24-26.
N ot an evaluation
62 L
ee, L .J. 1995. A
n E valuation of a N
eighborhood W atch Program
in T aipei M
etropolitan Police D epartm
ent in T
aiw an, T
he R epublic of C
hina. D issertation A
bstracts International, A : T
he H um
anities and Social Sciences, 56, (2, A
ug).
Post-test only w ithout random
allocation
63 L
ee, L .J., C
heurprakobkit, S. and D eng, F. 1999. N
eighborhood w atch program
s in T aiw
an: police attitudes, crim
e rate and com m
unity support. Journal of Police Science and M anagem
ent, 2, (1), pp. 57-77. Post-test only w
ithout random allocation
64 L
eland, L .S., R
ushton, W .E
. and W atson, J.E
. 1984. T he effectiveness of prom
oting neighborhood w atch
door-to-door. P olice C
hief, 51, (6),pp. 42-44. N
ot an evaluation
65 L
ow enberg, R
.E . and B
andurraga, R .A
. 1982. N eighborhood w
atch: a com m
unity resource. Journal of C
alifornia L aw
E nforcem
ent, 16, (3), pp. 159-164. N
ot an evaluation
66 L
yddy, F. 1998. C rim
e-w atch confusions. Psychologist, 11, (6), p. 297.
N ot about neighborhood w
atch 67
M cK
eow n, K
. & B
rosnan, M . 1998. Police and C
om m
unity: A n E
valuation of N eighborhood W
atch and C
om m
unity A lert in Ireland. Ireland: D
epartm ent of Justice, E
quality and L aw
R eform
. Post-test only w
ithout random allocation
68 M
cL ean, H
.E . 1991. W
atching out for H onolulu. L
aw and O
rder, 39, (6),pp. 58-62. N
o control area 69
M cPherson and T
hom son 1985. A
n E xam
ination of the Im plem
entation, O peration and E
ffectiveness of the Four Pilot N
eighborhood W atch Schem
es introduced in the C ity of E
dinburgh on 14 thM
ay, 1984. L othian
and B orders Police.
Post-test only w ithout random
allocation
70 M
eese III, E . 1988. N
eighborhood W atch: a crim
e success story. T he Police C
hief, 55, p.10. N
ot an evaluation 71
M eredith, C
. and Paquette, C . 1992. C
rim e Prevention in high-rise rental apartm
ents: findings of a dem
onstration project. Security Journal, 3, (3),pp. 161-168. N
o control area
72 M
urphy, R .B
. and H orton, S. 1975. Focus on burglary –
a m anagem
ent approach to prevention of crim es.
Police C hief, 62, (11),pp. 30-34.
N ot an evaluation
73 N
ew ton, A
.M . 1978. Prevention of crim
e and delinquency. C rim
inal Justice A bstracts, 10, (2),pp. 245-266.
N ot an evaluation
74 O
lson, R .L
., L avrakas, P.J., R
osenbaum , D
.P., K am
inski, F. L ow
enberg, R .E
., B andurraga, R
.A ., K
now les,
L ., L
esser, C ., M
cK ew
en, F. and M cC
am pbell, M
.S. 1983. R educing residential burglary. Police C
hief, 50, (2), pp.26-41. (C
heck authors)
N ot an evaluation
75 Pate, A
.M . 1989. C
om m
unity-oriented policing in B altim
ore.In D .J. K
enney (ed.), Police and Policing: C
ontem porary Issues. U
nited States: Praeger. N
ot about neighborhood w atch
76 Payne, D
. 1989. N eighborhood W
atch: A ttitudes of R
esidents to Schem es in E
dinburgh and G lenrothes.
E dinburgh: T
he Scottish O
ffice C entral R
esearch U nit.
N o control area
77 Pengelly, R
. 1996. V ehicle w
atch schem es: a com
parative study. Focus on Police R
esearch and D
evelopm ent, 7, pp.42-43
N ot an evaluation
78 R
einer, R . 1995. C
om m
unity Policing in E ngland and W
ales. In, J.P. B rodeur. C
om parisons in Policing: A
n International Perspective. A
ldershot: A vebury.
N ot an evaluation
42
79 R
obinson, M .B
. 2000. From research to policy: preventing residential burglary
through a system s approach.
A m
erican Journal of C rim
inal Justice, 24 (2), pp. 169-179. N
ot about neighborhood w atch
80 R
osenbaum , D
.P. 1988. A critical eye on neighborhood w
atch: does it reduce crim e and fear? In T
. H ope and
M . Shaw
(eds.) C om
m unities and C
rim e R
eduction. L ondon: H
M SO
. N
ot an evaluation
81 R
osenbaum , D
.P. 1988. C om
m unity crim
e prevention: a review of w
hat is know n. In, D
.J. K enney (ed.),
1988, Police and Policing: C ontem
porary Issues. U nited States: Praeger.
N ot an evaluation
82 R
udoff, G .S. 1995. C
rim e w
atchers use cellular phones: program in D
ade county succeeds in reducing crim e.
L aw
and O rder, 42, (12),pp. 32-33.
N o control area
83 R
udoff, G .S. 1996. C
ell Phones R educe M
ajor U rban C
rim es. A
m erican C
ity and C ounty, 111, (4), pp.19-20.
N o control area
84 Salcido, R
.M ., O
rnelas, V . and G
arcia, J.A . 2002. A
neighborhood w atch program
for inner-city school children. C
hildren and Schools, 24, (3),pp.175-187. N
ot about neighborhood w atch
85 Sansfacon, D
. and W elsh, B
. 1999. C rim
e Prevention D igest II: C
om parative A
nalysis of Successful C
om m
unity Safety. C anada: International C
entre for the Prevention of C rim
e. N
otan evaluation
86 Schneider, H
.J. (ed.). V ictim
in International Perspective. Papers and E ssays G
iven at the T hird International
Sym posium
on V ictim
ology 1979 in M unster. W
est G erm
any: W alter de G
ruyter & C
o. N
ot an evaluation
87 Security 21 (2004) N
eighborhood W atch: A
R eport on Status of N
eighborhood W atch Schem
es N
ot an evaluation 88
Sherm an, L
.W ., Farrington, D
.P., W elsh, B
.C .and M
acK enzie, D
.L. (eds.) 2002. E vidence B
ased C rim
e Prevention. N
ew Y
ork: R outledge.
N ot an evaluation
89 Silw
a, l. 1983. Suburban crim e and citizen action. In, D
.J. K enney (ed.), 1989, Police and Policing:
C ontem
porary Issues. U nited States: Praeger.
N ot an evaluation
90 Skogan, W
. 1995. C om
m unity Policing in the U
nited States. In, J.P. B rodeur (ed.), C
om parisons in Policing:
A n International Perspective. A
ldershot: A vebury.
N ot about neighborhood w
atch
91 Skogan, W
.G . 1988. C
om m
unity organizations and crim e. In, M
. T onry and N
. M orris (eds.), C
rim e and
Justice: A R
eview of R
esearch, V olum
e 10.C hicago: U
niversity of C hicago Press.
N ot about neighborhood w
atch
92 Skogan, W
.G . 1989. C
om m
unities, crim e and neighborhood organizations. C
rim e &
D elinquency, 35 (3), pp.
437-458. N
ot about neighborhood w atch
93 Sm
ith, B .W
., N ovak, K
.J. and H urley, D
.C . 1997. N
eighborhood crim e prevention: the influences of
com m
unity-based organizations and neighborhood w atch. Journal of C
rim e and Justice, 20, (2),pp. 69-86.
Post-test only w ithout random
allocation
94 T
he L ibertarian R
esearch and E ducation Trust 1985. N
eighborhood W atch in O
ne Police D ivision. W
orking Paper N
um ber Five.
N o control area
95 T
illey, N . 1993. A
fter K irkholt –
T heory, M
ethod and R esults of R
eplication E valuations. C
rim e Prevention
U nit Series Paper N
o. 47. L ondon: H
om e O
ffice. N
ot an evaluation
96 T
itus, R .M
. 1984. R esidential burglary and the com
m unity response In R
. C larke and T
. H ope, (eds.) C
oping w
ith B urglary: R
esearch Perspectives on Policing. B oston: K
luw er-N
ijhoff, pp. 97-130. N
ot an evaluation
97 T
onry, M . and M
orris, N . C
rim e and Justice: A
R eview
of R esearch, V
olum e 10.C
hicago: U niversity of
C hicago Press.
N ot an evaluation
98 T
opp, N . 1987. D
etroit’s proactive approach to law enforcem
ent. FB I L
aw E
nforcem ent B
ulletin, 56, (8), pp. 22-25.
N ot an evaluation
43
99 U
nited States N ational C
rim e Prevention C
ouncil. 1989. Success of com m
unity crim e prevention. C
anadian Journal of C
rim inology, 31, (4),pp. 487-506.
N ot an evaluation
100 W
adm an, R
.C ., and B
ailey, S.E . 1993. C
om m
unity Policing and C rim
e Prevention in A
m erica and E
ngland. U
nited States: O ffice of International C
rim inal Justice.
N ot about neighborhood w
atch
101 W
aller, I. 1979. W hat reduces residential burglary: action and research in Seattle and T
oronto. In H .J.
Schneider (ed.). V ictim
in International Perspective. Papers and E ssays G
iven at the T hird International
Sym posium
on V ictim
ology 1979 in M unster. W
est G erm
any: W alter de G
ruyter & C
o.
N o control area
102 W
ashnis, G .J. 1976. C
itizen Involvem ent in C
rim e Prevention.Lexington: L
exington B ooks.
N ot an evaluation
103 W
hitehead, K . 1989. W
onder schem e or w
hite elephant? Police R eview
, 97, pp.1822-1824. N
ot an evaluation 104
W illiam
s, H . and Pate, A
.M . 1987. R
eturning to first principles – reducing the fear of crim
e in N ew
ark. C
rim e and D
elinquency, 33, (1),pp.53-70. N
ot about neighborhood w atch
105 Y
anay, U . 1994. T
he ‘big brother’ function of block w atch. International Journalof S
ociology and Social Policy, 14, (9),pp. 44-58.
N o control area
106 Y
arw ood, R
. and E dw
ards, B . 1995. V
oluntary action in rural areas: the case of N eighborhood W
atch. Journal of R
ural Studies, 11, (4),pp.447-459. N
o control area
107 Y
arw ood, R
. and G ardner, G
. 2000. Fear of crim e, cultural threat and the countryside. A
rea, 32, (4),pp. 403- 411.
N ot about neighborhood w
atch
44
REFERENCES
References (Included studies)
Anderton, K.J. 1985 The Effectiveness of Home Watch Schemes in Cheshire, Chester: Cheshire Constabulary.
Bennett, T. 1990. Evaluating Neighborhood Watch. Cambridge Studies in Criminology LXI. UK: Gower.
Bennett, S.F. and Lavrakas, P.J. 1989. Community-based crime prevention: an assessment of the Eisenhower Foundation’s Neighborhood Program. Crime and Delinquency, 35, 3, pp.345- 364.
Cirel, P., Evans, P., McGillis, D. and Whitcomb, D. 1977. Community Crime Prevention Program, Seattle: An Exemplary Project. Washington D.C.: Government Printing Office.
Forrester, D., Frenz, S., O’Connell, M., et al. 1990. The Kirkholt Burglary Prevention Project: Phase II. Crime Prevention Unit: Paper 23. London: Home Office.
Henig, J.R. 1984. Citizens Against Crime: An Assessment of the Neighborhood Watch Program in Washington, D.C. Occasional Paper, Center for Washington Area Studies. Washington: George Washington University.
Hulin, J.O. 1979. Community based crime prevention project. Crime Prevention Review, California State Department of Justice, 6, (3), pp.26-34.
Husain, S. 1990. Neighborhood Watch and Crime: An Assessment of Impact. London: Police Foundation.
Jenkins, A.D. and Latimer, I. 1986. Evaluation of Merseyside Homewatch Scheme. Management Development and Force Planning Unit: Merseyside Police.
Knowles, L., Lesser, C. and McKewen, F. 1983. Burglary prevention: a citizen initiated and operated neighborhood watch program. Police Chief, 50, pp.36-38.
Latessa, E.J. and Travis, L.F. 1986. Evaluation of the College Hill Crime Prevention Program. United States: University of Cincinnati.
Lewis, D.A., Grant, J.A., and Rosenbaum, D.P. 1988. Social Construction of Reform: Crime Prevention and Community Organizations. New Brunswick: Transaction Books.
Lowman, J. 1983. Target hardening burglary prevention and the problem of displacement phenomena. In T. Fleming (ed.) Deviant Designations: Crime, Law and Deviance in Canada. Toronto: Butterworths: pp. 277-304.
45
Matthews, R. and Trickey, J. 1994. Eyres Monsell Crime Reduction Project. Leicester: University of Leicester, Centre for the Study of Public Order.
Matthews, R., and Trickey, J. 1994. The New Parks Crime Reduction Project. Leicester: Centre for the Study of Public Order, University of Leicester.
Mukherjee, S. and Wilson, P. 1988. Neighborhood Watch: issues and policy implications. Trends and Issues No. 8. Australia: Australian Institute of Criminology.
Research and Forecasts Inc. 1983. The Figgie Report, Part IV: Reducing Crime in America – Successful Community Efforts. Ohio: Figgie International.
Tilley, N. and Webb, J. 1994. Burglary Reduction: Findings From Safer Cities Scheme, Crime Prevention Unit Paper 51. London: Home Office.
Veater, P. 1984. Evaluation of Kingsdown Neighborhood Watch Project, Bristol. Bristol: Avon and Somerset Constabulary.
References (Additional studies)
Cook, T.D. and Campbell, D.T. (1979) Quasi-Experimentation: Design and Analysis Issues for Field Settings. Chicago: Rand McNally.
Farrington, D.P. (2003) ‘Methodological quality standards for evaluation research’, Annals of the American Academy of Political and Social Science, Vol. 587, pp.49-68.
Farrington, D.P., Gill, M., Waples, S.J. and Argomaniz, J. (2007) ‘Studying the effects of CCTV on crime: meta-analyses of a national evaluation’, Journal of Experimental Criminology, Vol. 3, pp. 21-38.
Jones, H.E. (2005) Measuring effect size in area-based crime prevention research. Unpublished M.Phil. thesis, Statistical Laboratory, Cambridge: University of Cambridge.
Lipsey, M.W. and Wilson, D.B. (2001) Practical Meta-analysis. Thousand Oaks, California: Sage.
National Crime Prevention Council (2001) The 2000 National Crime Prevention Survey. Washington: DC. National Crime Prevention Council
Sherman, L.W. and Eck, J. (2002) Policing for crime prevention. In L.W. Sherman, D.P. Farrington, B.C. Welsh, and D.L. MacKenzie (Eds.) Evidence-based Crime Prevention. London: Routledge (pp. 295-329).
Sherman, L.W., Gottfredson, D.C., MacKenzie, D.L., Eck, J., Reuter, P. and Bushway, S. (1997) Preventing Crime: What Works, What Doesn’t, What’s Promising. Washington, D.C.: US Office of Justice Programs.
46
Sims, L. (2001) Neighborhood Watch: Findings from the 2000 British Crime Survey. Research Findings 150. London: Home Office.
Titus, R., (1984) 'Residential Burglary and the Community Response'. In: Clarke, R.V.G. and Hope, T. (Eds.). Coping with Burglary. Boston: Kluwer-Nijhoff.
Welsh, B.C. and Farrington, D.P. (2002) Crime Prevention Effects of Closed Circuit Television: A Systematic Review. Home Office Research Study 252. London: Home Office.
