Which burglary security devices work for whom and in what context? Briefing Note 1 September 2013 Background This is the first briefing note from the ESRC funded Secondary Data Analysis Initiative project Which burglary security devices work for whom and in what context?. The team consists of: Professor Andromachi Tseloni, Principal Investigator, Nottingham Trent University Dr Louise Grove, Co- Investigator, Loughborough University Professor Nick Tilley, Co- Investigator, University College London Professor Graham Farrell, Co- Investigator, Simon Fraser University Rebecca Thompson, Research Fellow, Nottingham Trent University The team is being advised by a panel of experts drawn from across academia, public, and third sectors. Project Aims Domestic burglary is a high volume crime affecting many households. As well as substantial financial loss and damage to property, it causes high levels of anxiety about the possibility of being burgled. Surveys documenting public priorities about crime place burglary at the top. Burglar alarms and other security devices in principle deter potential burglars. Insurance premiums are discounted when a fully operating burglar alarm exists in the home due to claims about the effectiveness of burglar alarms and other security devices in the marketing literature, but no systematic research studies have been undertaken to assess their effectiveness in different areas, accommodation types and occupants characteristics. The present research project is precisely concerned with such an assessment. Phase I Project Activity The analysis has used data from the British Crime Survey (BCS)/ Crime Survey in England and Wales Survey (CSEW). All sweeps of the survey 1992-2011/12 have been included. The innovative methodology employed here, developed as part of the team s past research is the Security Impact Assessment Tool (SIAT) (odds ratios of targeted households compared to the population) which produces Security Protection Factor (SPF) via comparing the odds ratios of individual or combinations of security devices against no security. The analysis is limited to those security devices examined in the BCS/CSEW: external lights on sensor; internal lights on timer; windows locks; security chains; burglar alarm; double locks or deadlocks; window bars; dummy box; CCTV. Window bars and dummy boxes are examined 1
individually but they do not enter the analysis of security configurations due to their low and diminishing use over time. One caveat of this work is that the security of the most vulnerable population is unknown due to the limited number of Victim Forms per respondent and crime seriousness classification in the BCS/CSEW. Phase I Key Findings Increase in the use and number of and improvements in the availability of security devices coincided with the unprecedented burglary falls. Individual and combined security devices prevent burglaries with entry more than attempts. Individual security devices reduce burglaries up to a third. BUT the existence of burglar alarm on its own is counterproductive (see Figures 1 and 2) Combinations of security devices in general afford up to 50 times more protection than no security. Figure 1: Security Protection Factors across individual security devices against burglary with entry over time (significant at 5% level unless solid white bar) BA = burglar alarm DL = double locks SC = security chains WL = window locks IL = internal lights on timer EL = external lights on sensor CC = CCTV 2
Figure 2: : Security Protection Factors across individual security devices against attempted burglary over time (significant at 5% level unless solid white bar) Next steps The next phase of the research will focus on commonly used combinations of security devices in the BCS/CSEW, and begin to draw out the burglary risk and security availability over particular population subgroups. In order to develop this phase, the Advisory Committee is asked: (a) to suggest security devices configurations that may be of particular policy interest, such as packages offered by the building industry or minimum security requirements by the insurance industry and (b) to consider which population subgroups may be of particular interest from a policy perspective. Other areas of analysis will examine burglary modus operandi, and the effect of Neighbourhood Watch. 3
Which burglary security devices work for whom and in what context? Briefing Note 2 Background March 2014 This is the second briefing note from the ESRC funded Secondary Data Analysis Initiative project Which burglary security devices work for whom and in what context? The team consists of: Professor Andromachi Tseloni, Principal Investigator, Loughborough University Dr Louise Grove, Co- Investigator, Loughborough University Professor Nick Tilley, Co- Investigator, University College London Professor Graham Farrell, Co- Investigator, Simon Fraser University Dr Rebecca Thompson, Research Assistant, Loughborough University The team is being advised by a panel of experts drawn from across academia, public, and third sectors. Project Aims Domestic burglary is a high volume crime affecting many households. As well as substantial financial loss and damage to property, it causes high levels of anxiety about the possibility of being burgled. Surveys documenting public priorities about crime place burglary at the top. Burglar alarms and other security devices in principle deter potential burglars. Insurance premiums are discounted when a fully operating burglar alarm exists in the home due to claims about the effectiveness of burglar alarms and other security devices in the marketing literature, but no systematic research studies have been undertaken to assess their effectiveness in different areas, accommodation types and occupants characteristics. The present research project is precisely concerned with such an assessment, using data from the Crime Survey for England and Wales (formerly British Crime Survey) from 1992 onward. Phase II Project News In Autumn 2013, Machi and Becky made the move from Nottingham Trent University to Loughborough University. Their new contact details are: Dr Thompson: R.Thompson@lboro.ac.uk Professor Tseloni: A.Tseloni@lboro.ac.uk The project email is now: burglary.security@lboro.ac.uk Activities The project team has been actively engaging with activities to promote the project and initial findings. Particular highlights have included: The team attended the European Society of Criminology s annual conference in Budapest in September 2013. You can see details at our project blog: http://bit.ly/1geueka
During December, the project was represented at Nesta s Alliance for Useful Evidence Christmas Networking Reception; The Crime Surveys User Conference, organised by the UK Data Service; Social Research Association s annual conference; and initial project findings were presented at the workshop: Social and economic propagation in complex systems (EHESS, Paris). On Tuesday 28 January, 2014, Professor Andromachi Tseloni gave evidence to the Justice Select Committee on Crime Reduction Policies, alongside Richard Garside, Ben Page, and Professor Mike Hough. You can view the video of the meeting at http://www.parliamentlive.tv/main/player.aspx?meetingid=14742 The team have been attending meetings with various groups and organisations throughout Phase II to raise awareness of the project, including Charnwood Community Safety Partnership, Society of Evidence Based Policing, and several police forces. Nottingham Crime and Drugs Partnership plan to launch a pilot of target hardening against burglary based on the Phase I and II project s findings. Details will be presented by James Rhodes, Policy and Performance Manager of the Nottingham Crime and Drugs Partnership at the next Advisory Committee workshop, 26 th March. On Monday, 17 th March 2014, Rebecca Thompson successfully passed her viva. Both the External and Internal Examiners were very impressed with Becky s work and we are all very proud of her! Phase II Project Research The project is now at the half way stage. Work has focused on common and effective combinations of security devices in the BCS/CSEW, as shown in Phase I, and has started to draw out the burglary risk and security availability over particular population subgroups. Phase II Key Findings Increase in the availability and number of security devices coincided with the unprecedented burglary falls. Combinations of security devices on average afford 30 times more protection than no security The most effective combination of security devices is to have both indoor lights on a timer and external lighting on a sensor, with double door locks (or deadlocks) and locks on windows (hereafter abbreviated to EIWD ) This seems to be the combination that confers the best value for number of devices protection consistently since 2001 Groups highlighted as being at increased risk of burglary included households with low income, those consisting of 3 or more adults and households in inner cities and urban areas. Renters were also at heightened risk, particularly if they lived in social housing. Some households are less likely to own effective combinations of security devices: those on lower incomes, with household composition of 3 or more adults, private and social renters, and households who don t own a car. The following figures focus on results relating to the most effective combination of devices, External lights on sensors, Internal lights on a timer, Window locks and Door locks or deadlocks (EIWD).
Figure 1: Social renters compared to homeowners: Percentage difference in the odds of EIWD availability and of experiencing a burglary by BCS sweep group. Figure 2: Private renters compared to homeowners: Percentage difference in the odds of EIWD availability and of experiencing a burglary by BCS sweep group (bars shaded in white are not statistically significant at the 10 per cent level).
Figure 3: Households with children compared to those without: Percentage difference in the odds of EIWD availability and of experiencing a burglary by BCS sweep group (bars shaded in white are not statistically significant at the 10 per cent level). Figure 4: Single adult households compared to two- adult households: Percentage difference in the odds of EIWD availability and of experiencing a burglary by BCS sweep group (bars shaded in white are not statistically significant at the 10 per cent level).
Figure 5: Three or more adult households compared to two- adult households: Percentage difference in the odds of EIWD availability and of experiencing a burglary by BCS sweep group (bars shaded in white are not statistically significant at the 10 per cent level). Next steps The next phase of the research will look in much more detail at area and population characteristics (using data from the Census). Modus Operandi analysis will ascertain whether methods of entry have changed over time and if this has been influenced by the increased availability of security (see Appendix 1 for some initial findings about MO). We would like to ask the Advisory Committee to consider: In your experience, which types of areas are particularly prone to burglary victimisation? Which population features, if any, do you regularly see in high burglary risk areas? How to overcome the lack of area identification prior to 2008/09 CSEW for studying area differences in burglary risk and security availability and effectiveness?
Appendix 1: Excerpts taken from Tilley, N. and Farrell, G. (in print) Security Quality and its Role in the Crime Drop. This chapter provided an argument plus preliminary evidence that the quality of security was critical to the drop in residential burglary of the 1990s. The car theft studies are a particular instance of the security hypothesis which suggests that increases in the quantity and quality of security led to dramatic declines in crime (Farrell et al. 2008). Yet the emphasis of the car theft studies to date has been mainly on the quantity of security. This largely reflects the availability of data that counts the spread of the number of vehicles with different types of security devices. To date there have not been separate indicators of the quality of security. This was not too important for the car crime studies because changes in quantity were much the same as changes in quality. The issue of quality is more important however when it comes to other crime types. We suggest it is fundamental. Figure 1 offers our main finding. It shows that the decline in burglaries was mainly a decline in burglaries where some type of security was overcome. That is, overcoming security was increasingly less likely to be the means of entry, from which, in light of the other evidence, we think it reasonable to infer that this was because of improvements in the quality of the security. Figure 1: Means of entry as Indicator of Role of Security (Burglary with Entry) We view the evidence presented here as supporting the security quality hypothesis while recognising that further research and further evidence from other sources, would be beneficial. Farrell, G, Tilley, N, Tseloni, A, and Mailley, J. (2008). The crime drop and the security hypothesis. British Society of Criminology Newsletter, 62, 17 21.