Preventing Repeat Residential Burglary:

A meta-evaluation of two Australian demonstration projects

Contents | Acknowledgements | Executive Summary | Chapter 1: About repeat victimisation | Chapter 2: The demonstration projects | Chapter 3: Meta-Evaluation | Chapter 4: Programme guide and resource tools | Appendix 1: Summary of research study outcomes | Appendix 2: Process evatuation framework | References

CHAPTER 2

The demonstration projects

This chapter provides an overview of the features and reported findings of the demonstration projects in Queensland and South Australia. More detailed information about each project is available in the individual evaluation reports prepared by each jurisdiction. The following chapter discusses the outcomes and identified issues in the two demonstration projects. The term 'break and enter' is used consistently throughout this chapter, even when referring to the outcomes of one jurisdictional evaluation report that used the word 'burglary'.

Background

Residential break and enter was identified as a priority area for crime prevention in 1997 through a collaborative process between the National Anti-Crime Strategy Lead Ministers from all States and Territories and the Federal Attorney-General.

Two demonstration projects began in 1998, supported by federal funding. One was conducted in Queensland, designed and administered by the Criminal Justice Commission and managed in partnership with the Queensland Police Service. The programme aimed to provide an enhanced police response to residential break and enter, with a specific focus on repeat victimisation. The other was established in South Australia, designed to prevent repeat victimisation through a range of local community resources and services that are alternative and additional to current practice in preventing and reducing residential break and enter. It was developed and managed by the Crime Prevention Unit of the South Australian Attorney-General's Department.

The Commonwealth funded a meta-evaluation to address the overall aims of the pilot projects, which were to:

• trial two alternative methods of applying a problem-solving approach - one police based, the other community based - to reduce residential break and enter offences
  
• develop a better understanding of the characteristics of the phenomenon of break and enter, including repeat victimisation, in an Australian setting
  
• identify specific intervention strategies effective against repeat residential break and enters
  
• produce material that could facilitate programme implementation in other locations⁴⁴.

The two demonstration projects began in late 1998 and continued for at least 12 months in the two jurisdictions. Each programme was evaluated by the jurisdiction operating the demonstration project. Queensland and South Australia completed individual evaluation reports in 2000 and 2001. The key outcomes from those reports are summarised in the following sections. More detailed information is available in the individual jurisdiction reports⁴⁵.

Objectives and operation of the demonstration projects

Project objectives

Queensland demonstration project

The specific objectives of the Beenleigh Break and Enter Reduction Project are set out in the Queensland evaluation report⁴⁶ as:

• evaluate the effectiveness of strategies to reduce overall break and enter victimisation rates and levels of reported repeat victimisation in areas with high residential break and enter rates
  
• assess the effects of this approach on victims' perceptions of the quality of policing services received in response to the break and enter incident
  
• identify the information requirements and the administrative and organisational conditions needed to enhance the ongoing police response to residential break and enter offences
  
• develop a programme guide to support the implementation of effective break and enter reduction strategies in other locations
  
• increase understanding of the phenomenon of repeat victimisation.

South Australian demonstration project

The primary aim⁴⁷ of the South Australian Residential Break and Enter Prevention Pilot Project was "to improve the quality of life in our community by exploring whether it is possible to reduce the incidence of break-ins to people's homes particularly by preventing that portion of residential break and enter offences that victimise the same dwelling repeatedly" (referred to as repeat victimisation).

Secondary aims48 were:

• "to trial a community-based response to reduce residential break and enter offences, including the use of community volunteers
  
• to produce materials that can facilitate project implementation in other locations
  
• to develop a better understanding of the characteristics of the phenomenon of break and enter, including repeat victimisation [in the selected sites]
  
• identify specific intervention strategies effective in reducing residential break and enter and preventing repeat victimisation for residential break and enter".

Intervention design

Queensland demonstration project

The approach taken by the Beenleigh Break and Enter Reduction Project was a police-based crime prevention response. It involved three different levels of intervention provided by police. The first two levels focused on the individual residence. All victim addresses received the first level standard response and only repeat victims received the more intensive second level of intervention. The third level was an area-level response targeting hot spots. Hot spots refer to small areas with high rates of break and enter.

The Queensland evaluation report describes the operation of the programme in detail. The following overview summarises the information provided in that report.

The first level of intervention was described as the Stopbreak response. It required the police officers responding to the reported incident to:

• complete a crime report, conduct a preliminary investigation of the scene, and arrange for other officers to attend as required for more intensive investigation
  
• provide a Stopbreak folder of crime prevention material which contained a burglary prevention booklet, crime prevention brochures, and a property-marking kit
  
• conduct a Home Security Quick Assessment of the victim's residence to identify the steps victims need to take to reduce the risk of further incidents
  
• contact near neighbours and encourage them to take immediate steps to improve their own home security.

The second level of intervention was termed the Hot Dot response, directed at preventing multiple re-victimisation. It was triggered when there had been a previously reported break and enter or related property offence (such as stealing, wilful property damage, or theft of a motor vehicle) at the address within the past 12 months. This second level of intervention targeting repeat victims involved a visit by the police project officer to identify and respond to any factors, such as poor security, that may have contributed to the repeat victimisation. The project officer's tasks were to:

• conduct a repeat victimisation survey with the resident and, where necessary, make suggestions on how to minimise the risk of re-victimisation
  
• where relevant, provide the resident with a portable security alarm, locks, property engravers, timer devices, or similar security equipment according to the assessed security vulnerabilities of that dwelling or identify the residence as requiring extra police patrols
  
• distribute a package containing burglary prevention advice to near neighbours.

The third level of intervention was the Hot Spot response, designed to prevent or reduce residential break and enter by applying an area-wide approach to both victims and other residents in an area of about 200 metres radius. This response included:

• police and volunteers conducting door-knocks and letter-box drops at each residence in the hot spot area to increase community awareness of the break and enter problem in the area
  
• increasing police patrols, particularly around victimised addresses in the hot spot
  
• offering free home security assessments and providing assistance with property marking to residents in the area
  
• encouraging the establishment of Neighbourhood Watch groups
  
• assisting residents to conduct a safety audit
  
• providing specialised burglary prevention training to residents.

South Australian demonstration project

The approach taken by the South Australian Residential Break and Enter Prevention Project was a community-based crime prevention response with trained volunteers delivering services to victims in their local community.

The South Australian programme evaluation report⁴⁹ describes the operation of the programme in detail. The following overview summarises the information provided in that report.

The intervention comprised five elements:

• security advice tailored to the individual dwelling and victim, where volunteers undertook a security audit and recommended measures to improve security
  
• direct linkage to engravers for property marking, where volunteers made an immediate referral to a free property engraving service
  
• neighbourhood contact, either directly by the volunteer or by the victim, to provide an information kit on prevention measures
  
• informal support for victims, where volunteers provided an opportunity for victims to talk about their experiences and any related concerns
  
• referral to other agencies, where volunteers referred victims to various services for counselling and other forms of support if there was an apparent need.

In addition, in one of the two intervention sites, victimised residents had access to a service providing and installing locks to the value of $200 at no cost to the resident. These monies were made available through project funds as part of the South Australian intervention programme.

The same package of intervention elements was available to the victimised address regardless of prior history, that is, whether the incident triggering the intervention response was the first one reported during the programme period or whether there had been previous reported break and enters at that address before the programme began.

Evaluation design

Queensland demonstration project

The Queensland outcome evaluation compared the total number of break and enter offences and incidence and prevalence of repeat victimisation during the 12 months of the operation of the intervention programme (November 1998 to October 1999) against figures for the immediately preceding 12-month period. Percentage change between the two time periods in the intervention location was compared to change over the same two time periods in both a geographically contiguous control site and a non-contiguous control site.

South Australian demonstration project

The South Australian outcome evaluation compared the total number of break and enter offences between the 20-month period from the start of the project in November 1998 to the immediately preceding 20 months in the two intervention areas. It used a contiguous control site for each intervention area, and a common non-contiguous control site using survival analysis, a statistical technique that models the risk of 'failure' (in this case, of a repeat burglary occurring) over any period up to the maximum time of risk, and which allows for control over the length of the follow-up in which the address is 'at risk'. Repeat victimisation was evaluated using survival analysis over three and six-month time periods, dated from the point of time that the particular household was victimised rather than the programme start date.

Return to the top

Reported results from the demonstration projects

The results of each demonstration project are presented in three sections below. The first section on outcomes provides an overview of the key findings reported by each project. The second section on specific results presents key statistics and analyses from each report relevant to the outcomes. The third section outlines the main implementation issues reported by each project. Strategies to address these issues are presented in the programme guide in chapter 4.

Outcomes

Queensland demonstration project

The key results reported in the Queensland outcome evaluation50 were:

• a modest decline in the absolute number of repeat victimisations and a substantial reduction in the probability of becoming a repeat victim
  
• a reduction in the number of offences in the targeted hot spots during the intervention period without any apparent displacement to neighbouring areas
  
• no evidence that the project had an effect on the total number of residential break and enters in the intervention area
  
• an increase in the likelihood of victims acting on police information and/or advice about ways to improve home security
  
• no major differences between the trial area and control areas in the level of expressed satisfaction with the way that police responded to the break and enter.

South Australian demonstration project

The key findings reported in the South Australian outcome evaluation51 were:

• overall, repeat victimisation remained stable in the intervention area but rose in the control area
  
• intervention did not reduce the total number of break and enters in the intervention area during the intervention period
  
• break and enter offences showed promising downturns in the six-month follow-up period after the intervention was completed in the intervention but not the control area
  
• intervention was more successful in one of the two intervention sites.

Common conclusions

Both jurisdictional evaluation reports concluded the intervention was successful in addressing repeat victimisation, but was not effective in reducing overall break and enter rates.

Specific results

Queensland demonstration project

The number of repeat incidents and the number of victims experiencing at least one repeat both declined during the 12-month period that the programme was operating in the intervention area, compared to the immediately preceding 12-month period. There was a 16 per cent drop in the number of repeat victims and a 15 per cent decline in the number of repeat incidents.

TABLE 5
Number of repeats between pre-project and the project periods in the trial area (Queensland) 

	Victims		Incidents
	Pre-project	Project period	Pre-project	Project period
Total incidents	563	686	617	732
Total repeats	45	38	54	46
% change in repeats	-15.6%		-14.8%
Source: Criminal Justice Commission, 2001:15

Analysis of changes in the number of repeat victims in the two control sites showed an increase of 86 per cent in the adjacent area and a decline of 18 per cent in the non-contiguous control area.

TABLE 6
Number of repeats between pre-project and the project periods in the control areas (Queensland) 

	Control 1	Control 2
Pre-project period	36	95
Project period	67	78
% change in repeats	+86.1	-17.9
Source: Criminal Justice Commission, 2001:15

The Queensland report also provided an analysis of differential risk, that is, the probability of being a victim compared to the probability of being a repeat victim. The probability of victimisation was calculated by dividing the total number of victims by the total 'at risk' population (i.e. the number of dwellings).

The probability of experiencing a further incident after an initial break and enter was calculated by dividing the number of repeat victims by the total number of victims. The ratio of increased risk is the ratio of repeat victimisation probability to initial victimisation probability for each time period and area.

TABLE 7
Probability of reported repeat victimisation in trial and control areas (Queensland) 

	Pre-project	Project period
Trial area
Probability of being a victim	0.034	0.041
Probability of being a repeat victim	0.087	0.059
Ratio of increased risk	2.56	1.44
Control 1
Probability of being a victim	.044	.038
Probability of being a repeat victim	.043	.093
Ratio of increased risk	1.02	2.45
Control 2
Probability of being a victim	.043	.035
Probability of being a repeat victim	.085	.087
Ratio of increased risk	1.98	2.49
Source: Criminal Justice Commission, 2001:16

The probability of becoming a repeat victim fell in the intervention area while the overall risk of victimisation increased over the two time periods. Conversely, the probability of becoming a repeat victim increased and the risk of overall victimisation declined in both control areas. The ratio of increased risk dropped from 2.6 to 1.4 in the intervention area. The ratio more than doubled in the adjacent control site and rose from 2 to 2.5 in the non-contiguous site.

The effect of the level 3 Hot Spot intervention was assessed by comparing the average monthly number of break and enter incidents occurring during the hot spot intervention period with the months immediately before the intervention and those immediately after the intervention ended. The immediately surrounding geographic area was used as a control site in each case.

TABLE 8
Monthly average number of break and enter incidents in hot spots and surrounds (Queensland) 

	Hot Spot 1		Hot Spot 2
	hot spot	surrounds	hot spot	surrounds
Pre-intervention period	2.80	1.80	1.67	0.83
Intervention period	0.67	1.67	0.00	0.00
Post-intervention period	2.75	1.25	0.50	0.00
% change pre and during intervention period	-76.1	-7.2	100.0	100.0
Source: Criminal Justice Commission, 2001:17

There was a 76 per cent reduction during the intervention period in the first hot spot and no change in the surrounding area. There were no incidents during the intervention period in both the second hot spot and its surrounding area. After the intervention ceased, the number of break and enters returned to pre-intervention levels in the first hot spot and rose to one-third of previous levels in the second hot spot. The Queensland evaluation report concluded that the hot spot intervention resulted in reduced offending but only on a temporary basis.

The total number of break and enters increased during the intervention period in the trial area compared to the immediately preceding 12 months, but declined in both control sites. Total incidents fell in all three areas in the eight-month follow-up period after the programme concluded.

FIGURE 1
Monthly average number of offences (Queensland) 

Source: Criminal Justice Commission 2001:18
 

There was a 19 per cent increase in reported residential break and enters in the intervention area between the intervention period and the immediately preceding 12 months. In contrast, there was a 13 per cent decline in the adjacent control site and a 12 per cent decrease in the non-contiguous control area. An adjusted analysis, that excluded the effects of a two-month period with atypical trends attributed to the effect of a single prolific offender, showed a four per cent increase in the intervention area compared to 17 per cent and seven per cent reductions in the two control sites. This adjusted analysis reduced the magnitude of difference between outcomes for the intervention and control sites. However, the intervention site still showed increased levels and the control areas declined.

Victims in the intervention area were more likely than those in both control sites to improve their home security in response to the security advice provided through the programme.

TABLE 9
Percentage of victims improving home security (Queensland) 

	Trial area	Control 1	Control 2
Took any action to improve home security	77%	65%	66%
Average number of actions taken per victim	3.7	1.9	1.4
Source: Adapted from Criminal Justice Commission, 2001:22

Two-thirds (67 per cent) of surveyed victims in the intervention area stated they were satisfied with the way that police had handled the matter, compared to 63 per cent of victims in the adjacent control site and 71 per cent in the non-contiguous control area. The percentage expressing dissatisfaction in the intervention area (13 per cent) was equal to or lower than percentages in the two control sites (16 per cent and 13 per cent respectively).

South Australian demonstration project
A survival analysis comparison of the 20-month period before and after the programme began showed no difference in cumulative failure rates between these two time periods for the combined intervention areas. The intervention areas did not show any change in repeats despite a 31 per cent increase in total break and enters. There was an 11 per cent increase between the equivalent time periods for the non-contiguous control area. The adjacent control area showed a seven per cent decline in repeats that parallels the six per cent decrease in total break and enters.

TABLE 10
Change in repeat break and enter cumulative failure estimates at 532 days' exposure (South Australia) 

	20 months before the programme		20 months during and after the programme		% increase in cumulative failure
	Estimate	Std. error	Estimate	Std. error
Combined intervention areas	0.11	0.006	0.11	0.011	0.0
Non-contiguous control area	0.15	0.004	0.17	0.006	11.3
Adjacent control areas	0.14	0.008	0.13	0.011	-7.1
Source: Morgan, 2001:29

Return to the top

Patterns of repeat incidents were also analysed over periods of three and six months from the time that the individual address received the intervention. The latter analysis included only those residences receiving the intervention early enough during the programme's operation to allow for a full six-month period to be assessed (about 80 per cent of cases). The time periods for the control areas relate to the equivalent time before and after the reported incident for that household.

TABLE 11
Percentage of residences experiencing a repeat within the relevant time period (South Australia) 

	Three-month follow-up		Six-month follow-up
	Pre-period	Post-period	Pre-period	Post-period
Combined intervention areas	4.25	4.15	6.40	6.52
Combined control areas	4.54	5.67	7.44	8.55
Source: Adapted from Morgan, 2001:30 and calculations from additional data provided to the meta-evaluation.

When repeats are calculated over a three-month period, the combined intervention area results show a small decline, from 4.25 per cent of residences experiencing a repeat incident before the intervention to 4.15 per cent within the three months after the intervention - a decrease of two per cent. The combined control areas show an increase from 4.54 per cent to 5.67 per cent between the two comparable time periods, a rise of 25 per cent.

In the six-month analysis, the combined intervention areas showed a small increase between the first and second time period, from 6.40 per cent to 6.52 per cent. This represents a rise of less than two per cent in the percentage of residences experiencing a repeat within six months after the intervention. The control areas rose by 15 per cent, from 7.44 to 8.55 per cent of residences experiencing a repeat.

The total number of break and enter incidents reported to police increased in the combined intervention areas by 31 per cent between the 20-month period before the programme began and the 20-month period after its introduction. The combined control areas showed a lower level of increase of 17 per cent between the two time periods.

TABLE 12
Change in break and enter incidents before and after the intervention (South Australia) 

	20 months before the programme	20 months during and after the programme	% increase
Intervention area 1	2240	3036	35.5
Intervention area 2	1122	1379	22.9
Combined intervention areas	3362	4415	31.3
Non-contiguous control area	7262	8875	22.2
Adjacent control areas	1762	1658	-5.9
Combined control areas	9024	10533	16.7
Source: Adapted from Morgan, 2001:27

Intervention and control areas show different trends over the pre- and post-programme 20-month periods. Annualised figures (calculated to take into account the different numbers of months in each of the four time periods of interest) show a decline in the six-month follow-up period for the intervention areas and an increase in the control sites. Total break and enters fell by 13 per cent between the final six months of the programme period and the six-month follow-up period in the two intervention areas combined. In comparison, the combined control areas showed a six per cent increase.

FIGURE 2
Annualised number of offences (South Australia)

Source: Criminal Justice Commission 2001:18
 

Analysing the two intervention areas separately provides a different pattern of outcomes for both repeat and total break and enter figures. One intervention area shows a rise in repeats in the 20-month survival analysis while the second shows a reduction. The same pattern of results emerges when comparing the percentage of residences experiencing a repeat within three months and within six months. The first intervention area also shows a greater increase in the total number of break and enters.

TABLE 13
Differences in outcomes between individual intervention areas 

	Intervention area 1	Intervention area 2
Percentage increase in cumulative failure using 20-month survival analysis	27.2	-18.2
Change in % of residences experiencing a repeat break and enter within 3 months	8.7	-21.8
Change in % of residences experiencing a repeat break and enter within 6 months	8.3	-11.0
Change in total break and enter incidents	35.5	22.9
Source: Adapted from data reported in Morgan, 2001 and calculations from additional data provided to the meta-evaluation.

The control areas also show different patterns of results. The adjacent control area results show a reduction in repeat incidents in the 20-month survival analysis and the three-month analysis, and a small rise in the six-month analysis. The non-contiguous control area shows increases in all three. Total break and enters also declined in the adjacent control area and rose in the non-contiguous control site.

Implementation issues

The major implementation issues identified in the Queensland and South Australian evaluations are summarised below. They are described in more detail in the individual jurisdiction reports. These and other issues, together with strategies to address them, are discussed in the programme guide section of this report. The major issues common to both jurisdictions are also considered under key success factors in the following section.

Queensland demonstration project

The major implementation issues reported in the Queensland evaluation related to:

• a need to revise programme objectives in the light of early programme experience, specifically, (i) the types of strategies that could be applied were limited in various ways, so that the programme was more appropriately considered a test of differentiated policing responses than of problem-solving policing as originally intended, and (ii) the most appropriate solution often required a combination of multiple intervention responses and the number of cases where only a single intervention strategy was applied was too small to allow meaningful statistical comparison; therefore it was not possible to assess the relative effectiveness of specific intervention responses as originally intended
  
• one element of the intended Stopbreak response was not fully implemented, with only an estimated one-third of near neighbours contacted (reasons given by police for not distributing the advice cards were that they failed to see the benefit of leaving an advice card or did not have the time to distribute them)
  
• difficulties in mobilising residents in hot spot areas to participate in the project, for example, only 16 residents attended the first and eight attended the second burglary prevention workshops designed as part of the Hot Spot intervention, despite a door-to-door campaign inviting residents to attend
  
• problems in proper completion of basic procedures by some operational police in the first two months of the project
  
• initial difficulties in using volunteers to assist in delivering programme elements that were time-critical, given volunteers were not always able to commit to the times and dates on which visits to victimised residences were arranged
  
• difficulties in identifying repeat victim addresses from existing police records, through both system issues (the electronic database for crime reporting did not have a built-in analytic capacity and did not interface well with other software applications) and data quality issues, for example, misspelling of addresses or wrong codes assigned to incidents.

In addition, the report identified several other issues affecting evaluation outcomes, such as:

• low overall incidence of break and enter victimisation in the intervention site during the programme period, making it difficult to separate programme effects from random or seasonal effects, and introducing potential for atypical events (such as the single prolific offender) to dramatically affect overall comparisons
  
• low incidence of repeat victimisation, making it difficult to demonstrate statistically the effectiveness of initiatives
  
• limited timeframe, which restricted the operation of the strategies and tactics to those that could be implemented within a relatively short period
  
• operational constraints, such as competing demands on police personnel that limited training of operational police in programme requirements, as well as considerable staff movements during the programme
  
• extraneous influences affecting evaluation measures, such as the introduction of new crime prevention initiatives in control areas during the evaluation period.

South Australia demonstration project

The key issues reported in the South Australia programme evaluation related to:

• volunteer related aspects of the project, such as recruitment, selection and monitoring, for example, difficulties in recruiting sufficient numbers of volunteers in one site and unsuitability of a small number of volunteers identified only after the selection process, whose involvement in the programme was subsequently terminated
  
• capacity of volunteers to deliver interventions and meet the requirements of their role, particularly achieving competency across all five strands of the intervention by all volunteers
  
• management and resource issues in relation to using volunteers, including meeting timeframes and procedural requirements
  
• inconsistency between professional role expectations and the traditional culture of volunteering, for example, in introducing quality monitoring demands and requirements to meet strict time commitments or notify of changes in availability
  
• capacity of police to participate in a collaborative approach, including complying with notification procedures that identified victimised addresses for involvement in the project
  
• feasibility of community groups taking responsibility for the programme.

In addition, the South Australian evaluation reports identified other issues relevant to the operation and evaluation of the programme's outcomes, including:

• a low rate of involvement of eligible victims in the project, with under one-third (32 per cent) of residential burglaries reported to police in the intervention areas during the 14 months of the programme's operation referred to the programme by police
  
• the potential for the programme to encourage victims to report break and enters to police that may otherwise have gone unreported
  
• the possibility of micro-level displacement, where a determined offender prevented from committing a successful break and enter at a residence where there has been programme intervention, continues to make attempts in the local neighbourhood until successful.

Common implementation issues are discussed in more detail in the next chapter, and strategies for addressing them are set out in the programme guide.

Conclusions reported in the evaluations

The Queensland evaluation report52 concluded that the project succeeded in getting operational police to give greater priority to prevention and victim issues and that it was possible to improve the police response to residential break and enter without imposing additional burdens on operational police. The evaluation report concluded that in changing the way in which police respond to break and enter, it was possible to reduce the risk of repeat victimisation and decrease offending in hot spots.

The South Australia outcome and programme evaluation reports53 concluded that the intervention was successful in preventing repeat break and enter of households in the intervention period and that volunteers were able to offer an effective service to victims of break and enter in their community.

The evaluation reports from both jurisdictions identified various implementation issues that would need to be considered in establishing similar programmes in other locations.

Chapter summary

Queensland and South Australia conducted separate crime prevention demonstration projects in 1998-99. Both aimed to reduce the rate of overall break and enter in the selected intervention areas relative to identified control sites by focusing on repeat victimisation.

The approach taken by the Queensland project was a police-based crime prevention response, involving three different levels of intervention provided by police. All victimised addresses received a Stopbreak response. Repeat addresses received a Hot Dot response. An area-level Hot Spot response was provided to residents in selected small areas with very high rates of break and enter.

The approach taken by the South Australian project was a community-based crime prevention response with trained volunteers providing prevention services. There were five elements: informal victim support, security advice, property marking, neighbourhood contact, and referral to other agencies for support.

The evaluation reports by both jurisdictions report a successful outcome in relation to repeat victimisation, but not in addressing overall break and enter rates. The jurisdictional reports identify various implementation issues. These issues will be considered in more detail in the next chapter.

Return to the top