Domestic burglary drop and the security hypothesis
© The Author(s) 2017
Received: 19 November 2016
Accepted: 31 January 2017
Published: 15 February 2017
This study examines the role of household security devices in producing the domestic burglary falls in England and Wales. It extends the study of the security hypothesis as an explanation for the ‘crime drop’. Crime Survey for England and Wales data are analysed from 1992 to 2011/12 via a series of data signatures indicating the nature of, and change in, the relationship between security devices and burglary. The causal role of improved security is strongly indicated by a set of interlocking data signatures: rapid increases in the prevalence of security, particularly in the availability of combinations of the most effective devices (door and window locks plus security lighting); a steep decline in the proportion of households without security accompanied by disproportionate rises in their burglary risk; and the decline being solely in forced rather than unforced entries to households. The study concludes that there is strong evidence that security caused the decline in burglary in England and Wales in the 1990s. Testing the security hypothesis across a wider range of crime types, countries and forms of security than examined to date, is required both to understand the crime drop and to derive lessons for future crime prevention practice and policy.
KeywordsCrime drop Domestic burglary Security hypothesis Data signatures Crime Survey for England and Wales
Sustained crime falls across a wide array of offences began in the United States in the early 1980s and in many other countries from the early 1990s. These have been referred to as the international crime drop (Tonry 2014; van Dijk et al. 2012) with the suggestion that there may have been a global crime drop (Tseloni et al. 2010). The widespread falls in crime came as a surprise to criminology (Farrell et al. 2008) and have posed a major challenge to those interested in understanding crime trends (Tonry 2014).
Reflecting the earlier crime downturn in the United States, early efforts to explain the crime drop stressed distinctive developments there (Blumstein and Wallman 2000). At least seventeen explanations have been identified in academic studies to date (Farrell et al. 2010; Farrell 2013). Due to their singular focus on the United States and inapplicability elsewhere, many of the early frontrunners have been discounted and categorised, with the benefit of hindsight, as parochial (Farrell et al. 2014; Tonry 2014). Most of the others appear to be contradicted by a range of specific evidence as well as the fact that they lack consistency with broader sets of evidence: They are inconsistent with the fact that crime rose for several decades previously; that some crime types, such as cyber-crimes and theft of some electronic products, have increased; and that there was significant variation in the timing and trajectory of crime declines both between and within countries (Farrell et al. 2014; Tseloni et al. 2010). The surviving hypothesis is the ‘security hypothesis’ (Farrell et al. 2011a): The crime drops are a function of reduced opportunities, which have been largely brought about by increases in the extent and quality of security, an idea first introduced by Clarke and Newman (2006) and by Van Dijk (2006).
The security hypothesis is underpinned by the crime opportunities theoretical framework of rational choice and routine activities. The link between crime and crime opportunities, and the role of security in their reduction, has long been established (Mayhew et al. 1976). The security hypothesis suggests that the infusion of everyday life with increasingly well-designed, unobtrusive and publicly acceptable security has led to the substantial crime drops that have been observed (see Tilley et al. 2015a). In addition there may well be some collective effects beyond the operation of individual devices, especially if potential offenders no longer assume that they can easily commit some types of crime with low risks of apprehension. Indeed increased security is the most common reason for the crime drop according to offenders interviewed in the four Australian states of New South Wales, Queensland, Western and South Australia (Brown 2015a). This fits with the rational choice theoretical perspective, according to which security improvements can be expected to increase the actual or perceived risk and effort of committing crimes or reduce the actual or perceived reward from them (Clarke 2012).
The security explanation of the crime drop is compatible with evidence that (a) different crime types have fallen at different times in different places, reflecting variation in how improved security has spread, and (b) particular types of security measures affect crime patterns differentially, producing distinctive crime-change ‘signatures’ (Farrell et al. 2014; Tseloni et al. 2010). Indeed, one of the advantages of the security hypothesis over other explanations of the crime drop is that it does not assume that all crime has dropped. It would expect crime increases where new crime opportunities emerge. Any new developments may inadvertently create new crime opportunities whose inhibition has not been built in from the start (Pease 1997), especially for CRAVED goods (concealable, removable, available, valuable, enjoyable and disposable) (Clarke 1999; Ekblom and Tilley 2000).1 This appears to have been the case with mobile phone theft (Thompson 2014; Office for National Statistics [ONS] 2013a) and cybercrime (McGuire and Dowling 2013). Thus, interspersed with an overall crime drop for many crimes produced by security increases, the security hypothesis would expect rises in some specific crimes where new opportunities have been created.
The diversity and ubiquity of security improvements create, however, huge challenges for testing the hypothesis as a whole. The task is that of specifying hypotheses that can be tested retrospectively with the available data, following a crime type—and country—specific approach (Farrell et al. 2008, 2010). Efforts to test the security hypothesis so far have focused on car crime. Car crime has fallen dramatically and there is mounting cross-national empirical evidence (based on available data that relate to the devices fitted to the vehicle itself) that central door locking, alarms and electronic immobilisers are especially important contributors to the drops in theft of and from vehicles (Bässman 2011; Brown 2013; Farrell et al. 2011b; Fujita and Maxfield 2012; Kriven and Ziersch 2007; Van Ours and Vollaard 2016). That said, it has to be acknowledged that these do not exhaust the ways in which the security of cars may have been increased.2 To our knowledge, the only previous study that touched upon longer term domestic burglary trends3 and house security comes from the Netherlands: Vollaard and van Ours (2011) conducted a cost-benefit analysis of a government regulation requiring burglary-proof windows and doors in new housing. They established that the built-in security in new homes reduced their burglary risks by 26% per year. It also contributed a net 5% to the overall burglary drop in that country in the decade following the regulation’s introduction.4
The present study contributes to research on the security hypothesis with a particular focus on domestic burglary in England and Wales. The security interventions to prevent domestic burglary have been unsystematic and non-universal: with the adoption of secured by design, some have been building age related, such as in the Netherlands study above, while others have been retrofitted to older properties. By contrast, the security interventions against vehicle crime were universal, implemented at the vehicle production phase and rolled out incrementally with vehicle age.5 Therefore this work moves the discussion on from car crime to explore the security hypothesis in relation to the decline in domestic burglary. It explicitly examines the security hypothesis over a longer period and a wider array of security combinations than previously, as well as with respect to burglars’ modus operandi. Thereby it expands crime signatures analysis to accommodate the challenges due to the different nature of the security interventions in residential properties compared to cars (Farrell et al. 2011b). We argue that the increases in availability of household security devices, coupled with an increased efficacy of security devices, provides a compelling explanation for the decline in burglary in England and Wales since the mid-1990s.
The next section discusses five key research hypotheses which test the effect of security on domestic burglary falls. An overview of the data and analytical strategy of this study follows. The "Findings" section is organised in five subsections each corresponding to a research hypothesis. Thereafter the theoretical and policy implications of the findings are discussed. Recommendations for further research and replications for testing the security hypothesis for the crime drop conclude this study.
Domestic burglary and the security hypothesis
An increase in overall levels of household security.
A decrease in the proportion of households with no security and an increase in their relative burglary risk.6
An increase in the installation of more efficacious security devices and combinations of security devices.
An improvement in the efficacy of the security devices that are fitted.
A much larger drop in burglaries that required that security be overcome than in those where no security had to be overcome.
If these changes are found they provide support for the overall security hypothesis as it relates to domestic burglary.
The analysis presented used data from a national victimisation survey that has been conducted since 1982. The survey is now called the Crime Survey for England and Wales (prior to 2012 it was known as the British Crime Survey), and for consistency will be referred to here as ‘CSEW’.
Overall the figure indicates that both burglary with entry and attempts have dropped dramatically, although the fall in numbers of burglaries with entry began first. Attempts did not start falling until 1997, 4 years after ‘successful’ burglaries. The time lag between the beginning of attempts and burglary with entry falls indicates that burglaries fell due to target characteristics encountered after the target had been selected (such as unanticipated guardianship in the form of security) rather than offenders’ decisions not to target properties.7 Burglars’ ‘hit rate’ (burglary with entry over total number of burglaries) has fluctuated around 59% since 1981 and attained its highest value (63%) within the period of rising crime (1981 and 1991). However it reached its minimum of 53% the year attempts peaked, in 1997. It is worth noting that alongside burglary rates burglars’ ‘hit rate’ has remained fairly stable at 60% since 2004/05. In the remainder of this paper, the focus will be on burglaries with entry on the grounds that attempts may be thwarted by the presence of effective security devices, which the potential offender is unable to overcome, such as locks, or discouraged altogether by, say, internal lighting.8 Before turning to findings, as they relate to these issues, the CSEW data and methodology used in the analysis are described.
The Crime Survey for England and Wales
This study’s evidence is based on analyses of sixteen CSEW sweeps, conducted between 1992 (referring back to events in 1991) and 2011/12, most undertaken in the course of the crime drop. The CSEW is a survey of the adult (16 years or older) population in England and Wales with currently 35,000 respondents per annum and consistently high response rates, between 73 and 83% (Jansson 2007; TNS-BRMB 2012). It is regarded as one of the most rigorous national crime surveys (Hough and Maxfield 2007).9 The CSEW, unlike police data, is not subject to changes in counting rules and offence categories, and therefore provides comparable year on year crime estimates (Van Dijk and Tseloni 2012).
The survey records crime experiences, including domestic burglary. Burglary victims are asked about details of incidents, including the security devices fitted to their properties at the time of the burglary.10 In addition, a randomly selected sub-sample is asked about the security devices fitted to their dwellings at the time of the interview. The security information provided by victims at the time of the burglary is a unique feature of the CSEW.
The number and type of security devices examined in the CSEW has improved slightly over time. Between 1992 and 1996, they included burglar alarms, double locks or/deadlocks, window locks and lights on a timer or sensor switch. Between 1998 and 2007/08 questions about dummy alarms, window bars/grilles and security chains were added, and lights were differentiated between indoor lights, and external lights on a timer or sensor. Between 2008/09 and 2011/12, CCTV cameras were added to the list. In order to obtain adequate sample sizes that enable meaningful statistical analysis of the large number of security combinations generated by the above list, the CSEW sweeps were merged: 1992–1996; 1998–2000; 2001/02 to 2004/05; 2005/06 to 2007/08 and 2008/09 to 2011/12. The Appendix provides details of the data sets, methodology and the security devices examined for both victims and non-victims of burglary over time.
This section places the analytic approach used in the study in a broader methodological context. Strong research designs are most straightforward where a single independent variable can be introduced in controlled conditions to observe its effect, if any, on the dependent variable to test the conjectured causal relationship. In relation to the security hypothesis, a different approach is necessary. The crime drops have occurred and the task is that of specifying hypotheses that can be tested retrospectively with the available data, with statistical confidence in the results. The analysis here does not describe a randomised experiment. Instead it relies on data describing contrasts between ‘treatment’ and ‘control’ samples which occurred in an unstructured manner, both gradually out of landlords’ and home owners’ own initiative and with regards to secured by design as a result of discrete changes in policy.
The CSEW data on security installed both at victims’ homes at the time of burglary and non-victims’11 residences at the time of interview delineate a quasi-natural experiment contrasting burglary risks between households with and without security (Dinardo 2010). This allows testing the security hypothesis via examining any causal effects of security on burglary. Since the cause, security, cannot be manipulated a quasi-natural experiment is not literally an experiment (Shadish et al. 2002). The methodology used to estimate the magnitude and statistical significance of the contrast between households with individual security devices and suites of them and those without security is the security impact assessment tool (SIAT) originally developed to test the effectiveness of car security devices (Farrell et al. 2011b).
The burglary risk for households without security is compared to the risk for households with a particular security device or combination of devices (both with respect to overall risk). The resulting metric, which is termed the security protection factor (SPF), shows how much less (or more) vulnerable a target is with given security devices compared to those with ‘no security’.12
Drawing on the above contrast provided in the CSEW, the present study develops a series of data signatures, whereby crime changes are consistent with expected outcomes from distinct context mechanisms outcome pattern configurations (CMOCs, Pawson and Tilley 1997). Having said this examining context, delineated by area and household type—specific patterns of burglary risks and security uptake trends, is beyond the scope of the current work.13 The methodology, which was originally developed to evaluate policy interventions using a realist evaluation approach, has acquired prominence in studies evaluating situational crime prevention interventions (Pawson and Tilley 1997). Each data signature is a discrete piece of empirical evidence that comprises a component of an overall triangulation approach to evaluation. As Eck and Madensen (2009: 69) highlight ‘[s]ignature changes consistent with expected intervention mechanisms eliminate rival explanations’ whereas those which are ‘inconsistent with the expected intervention mechanism undermine the validity of the conclusion that the intervention produced the crime change.’ The closer an observed outcome follows the expected pattern from the activation of the preventive mechanisms and the fewer alternatives exist, the more confident we can be in attributing causality to it (Pawson and Tilley 1997). The five expected changes outlined in the "Domestic burglary and the security hypothesis" section provide the data signatures pointing to the pivotal role of security in domestic burglary falls.
Was there an increase in overall levels of household security?
a. Was there a decrease in the proportion of households with ‘no security’? and
b. Did their relative burglary risk rise?
Households with ‘no security’ are known to be at very high risk of burglary with entry (Budd 1999) and Fig. 4 confirms this. Households with ‘no security’ were twice as likely as the general population to be burgled in 1992–1996 and nearly eight times more so in 2008/09 to 2011/12. Therefore households with ‘no security’ have experienced a fourfold increase in their relative burglary risk during the crime drop and especially since 2001/02. This agrees with the overarching assumption that security was the main driver for the burglary falls.
Was there an increase in the installation of more efficacious security devices and combinations of security devices?
Some security devices are more effective than others in reducing the risk of burglary with entry based on their SPFs from the 2008/09 to 2011/12 CSEW data (Tseloni et al. 2014). Moreover, a greater number of security devices is generally more effective than fewer although the benefit of more than four devices is negligible. The most effective combination of two devices is window locks and door double or deadlocks (SPF = 13); the most effective combination of three involves adding external lights on a sensor (SPF = 34); and the most effective four adding internal lights on a timer (SPF = 49).
The protection conferred by suites of devices generally exceeds the sum of the impact of the separate devices. The value added from combining individual devices is given by the Net Interaction Effect (NIE) of security configurations. For example, properties fitted with the four ‘WIDE’ devices: window locks (W), indoor lights on a timer (I), door double or deadlocks (D) and external lights on a sensor (E) have an SPF of 49. The sum of the SPFs of the individual devices (W = 7, I = 3, D = 3 and E = 3) is, however, only 16. The WIDE NIE is, therefore, 33 (calculated as 49 − 16). Of the 41 security suites examined, 28 had a positive NIE for burglary with entry. Five had zero NIEs and only eight had negative NIEs, of which more will be said later (Tseloni et al. 2014).16
Was there an improvement in the efficacy of the security devices?
With one exception the trends are clear: The protective power of all commonly installed suites of security devices, increased exponentially over time even more than their presence (see Figs. 3, 5). This is suggestive of increased efficacy (and implied quality improvements) and together with the previous discussion explains why burglary fell sharply while security did not expand to more households after 1998 (see Fig. 4). Notwithstanding the security hypothesis however the greatest ‘efficacy’ gains happened after 2004/05, when burglary rates plateaued. This implies an uneven take up of efficacious security and large variations in burglary trends across areas and population groups (as already seen in the increasing burglary risk of ‘no security’ households and the discussion of Fig. 4) which since 2004/05 have averaged out nationally (see footnote 5).
The addition of an alarm resulted in a consistent increase but a consistent decrease in security in 1992–1996 and 2008/09 to 2011/12, respectively, for the examined combinations.19 The paradoxical findings for EWD and EWBD in 2001/02 to 2011/12 are therefore part of a more general pattern and not exceptional nor confined to England and Wales (Bettaïeb 2015; ICPC 2015). From the various possible explanations for alarms’ ineffectiveness leading conjectures include the changing balance between risks and gains for the prospective burglar that alarms signal and/or alarms normalisation for home/contents insurance coupled with poorer average quality over time (Tilley et al. 2015b). Further research is needed but what the findings so far suggest is that it would be unwise to assume that, although more security devices generally increase security, this is invariably the case.
Was there a much larger drop in burglaries that required that security be overcome than in those where no security had to be overcome?
The purpose of the research reported here was to extend the empirical test of the security hypothesis as an explanation for the widespread crime drop. To do so it examined the relationship between domestic burglary, household security devices and burglars’ modus operandi over time based on data from the CSEW. The findings support the hypothesis while they do not find that all devices were similarly responsible. The use of combinations of security devices appears to have been particularly important, especially door and window locks plus security lighting while, somewhat counterintuitively, alarms were not identified as contributory. The findings reported here align well with the positive findings for the security hypothesis reported earlier in relation to theft of and from vehicles cross-nationally and domestic burglary in the Netherlands.
The findings, however, reflect the limitations of the available data, excellent though the crime survey drawn on is. For example, the proportion of households with storm windows (double glazing) to all external doors and windows more than doubled while the rate of burglary with entry more than halved between 1996 and 2008 (see Tilley et al. 2015a). Moreover, some security measures are targeted at an area rather than individual households, the unit of analysis here. These include secured by design (SBD) housing planning and construction standards; levels of ambient lighting sometimes increased in the interests of crime prevention; alley-gating to restrict access to the backs of properties where covert access may otherwise be readily obtained; and the widespread use of and publicity for DNA-like property marking. The implementation of SBD planning and building standards, first introduced in the UK in 1989, greatly expanded from 1998 (Armitage and Monchuk 2011) which coincides with the sharpest fall in burglaries (between 1997 and 2001/02, Fig. 1). The relationship of all the above changes however with the burglary falls cannot be tested with CSEW data. Consequently, it is quite possible that the overall role of security, when more broadly defined, is understated by the present study.
The data do not speak to mechanisms through which the inhibition of some volume crimes may multiply preventive outcomes. These mechanisms include, for example, the inhibition of the onset of criminal careers by removing the easy opportunity for novice offenders to become criminally involved with their associates (the debut crime hypothesis) or the removal of easy opportunities to raise the money needed for purchasing illicit drugs and becoming habituated to them in ways that foster further criminal behaviour. Testing whether these indirect crime-inhibiting out-workings of security induced falls in common acquisitive crimes lies beyond the scope of this study and provides avenues for future research on the security hypothesis.20 Other suggestions for expanding this work include context (reflected in area or household characteristics)-based analysis of the relationship (and the distributional justice) between security and crime trends; cost-benefit analysis (including carbon footprint costs) of fitting security; and investigating any diminishing marginal return of security in low risk areas (see footnote 5).
The research findings reported here have direct practical implications for maintaining or extending the crime drop. Previous research has shown that security is unevenly distributed across the population (Tilley et al. 2011; Tilley 2012; Tseloni and Thompson 2015). Moreover, the poorest sectors of the population and those who rent property privately or from local authorities (in the USA public housing tenants) are amongst the most at risk of domestic burglary (Hunter and Tseloni 2016; Tseloni and Thompson 2015). Those who have already been victimised are the most vulnerable of all and previous burglary victimisation is the single highest predictive factor of current risk (Farrell and Pease 1993; Osborn and Tseloni 1998). The clear finding that WIDE security devices (Window locks, Indoor lights on a timer, door double or Deadlocks and External lights on a sensor) produce a very substantial (almost 50 times) lower risk of burglary with entry than ‘no security’ suggests that these devices be targeted on the most vulnerable. Targeting could take the form of advice by police, insurance companies or government, subsidies for the installation of devices by charities or crime reduction partnerships, or legislation to decree that these devices be required as minimum standards to be met by property developers and by landlords for their tenants. The findings also suggest rethinking advice regarding burglar alarms. It would be prudent to think twice before recommending an alarm as part of a standard package of security devices to reduce the risk of domestic burglary. Pending further research to understand better why alarms are associated with increased rather than decreased risk of burglary with entry, the requirement by insurance companies that alarms be fitted as a condition for continued coverage or for avoiding increased premiums becomes questionable.
Conclusion: where next?
The empirical evidence reported here lends support to the security hypothesis as an explanation for the crime drop. Research to date has largely established the critical role of vehicle security in reducing car crime cross-nationally. The present study, drawing on triangulation of data signatures from sixteen sweeps of national victimisation surveys covering the period 1991–2011/12 that describe a quasi-natural experiment, shows that increases in the prevalence and effectiveness of house security have been a major driver of the domestic burglary falls in England and Wales. Therefore it expands coverage of the security hypothesis quite significantly.
The security hypothesis is important not only because it relates to the major question currently facing criminology, ‘Why did the long term trend of increasing crime reverse?’, but also because of its clear implications for crime prevention practice and policy. If security has been largely responsible for the massive and unexpected falls in past volume crimes, including burglary and theft of and from vehicles, crime policy should focus on reducing or pre-empting crime opportunities for new and emerging volume crimes, perhaps most notably cyber-crime, that have been facilitated by the internet.21
The results presented here do not prove the security hypothesis for the international crime drop. First, they relate to one crime type in one jurisdiction. Second, they rely on data that do not describe a randomised but a quasi-natural experiment, studying the contrasts recorded by the CSEW, a very strong series of victimisation surveys, between households with and without security devices. Indeed, the available data have been analysed to determine whether the precisely expected patterns that can be elicited are congruent with the security hypothesis. Therefore it has not been possible to control for all changes that have occurred. For example, ideally changes in the quality of security devices, and/or area-based security improvements ought to have been measured and tested in relation to producing the overall drop in burglary. Likewise the research reported here for England and Wales could usefully be replicated in other countries where there has been a drop in burglary (or where, conversely, burglary has risen) to ascertain whether similar household security improvements (or relaxing of security) have been made elsewhere with a similar role in producing drops (or rises) in burglary. Therefore while the present study comprises a step in understanding whether and how security has contributed to the drop in burglary, there is more work to be done.
While there is room for doubt we conclude that for burglary with entry the evidence supporting the primary role of security is now beyond reasonable doubt. This evidence also concurs with that for vehicle crime. We suggest it is now up to others to find evidence that could falsify the security hypothesis as it relates to burglary and vehicle crime or to look to evidence in other jurisdictions and for other crime types to determine whether similar findings suggest likewise that security improvements have led to widespread falls in crime.
Although the security hypothesis has been tested in relation to some specific crimes, what is observed are drops in many types of crime across many countries. Yet wide-ranging security developments, which Clarke (2016) refers to as an avalanche of security, now wash through much of everyday life. There is a need to catalogue those security improvements, establish where and when they were introduced, and to identify and assess the relevant crime pattern change signatures. There is also a need for further research into whether security brought down violence either directly or indirectly. There is, in short, significant potential for further research into the security hypothesis.
Pires and Clarke (2012) have suggested that CRAVED might be updated with CRAAVED suggesting available could be replaced with accessible and abundant.
Many car parks have been made more secure via CCTV cameras, barriers and improved lighting, reducing the risk to cars parked in them (Webb 2005; Smith et al. 2003; Poyner 1992; Tilley 1993). Also, in Australia and England and Wales new car registration arrangements have made it more difficult to change the identity of a stolen car (Brown 2015b; Webb 2005).
A number of studies has examined the short term effects of target hardening and security policy initiatives, including Secured by Design evaluations, on domestic burglary (for example, Armitage and Monchuk 2011; Ekblom 1996).
The study did not examine the effect of other security devices or security fitted retrospectively in existing homes both of which may be responsible for the bulk of Dutch national burglary falls.
With thanks to the Anonymous Reviewer who pointed this out.
The first two hypotheses are distinct: the first refers to an increase in security devices per household and the first part of the second to more households with any sort (single or multiple) of security or less households without any security.
Simultaneous or preceding drop in attempts compared to burglary with entry would have indicated that burglars’ decisions not to target properties (perhaps due to low financial returns from this crime type) might have been a reason behind the burglary fall. However the ‘typical gain’ for an offender does not vary considerably before and after the crime drop. Moreover between 2000 and 2013/14, the average value of goods stolen through burglary increased, alongside a decrease in the number of incidents (Shaw et al. 2015).
Therefore security can exert both a positive and negative effect to attempts requiring security device—specific analyses to address the role of security on attempts falls.
For details on the CSEW methodology and questionnaire, see Hough and Maxfield (2007), Flatley (2014) and the various CSEW Technical Reports, which are accessible via http://www.ons.gov.uk/ons/taxonomy/index.html?nscl=Crime+in+England+and+Wales.
For repeat victims of domestic burglary the security of the first incident experienced is included in this analysis to gauge its preventive effect. An initial burglary increases the demand for and use of security devices (Budd 1999; Philipson and Posner 1996). However victims’ response to a first burglary and whether this alters subsequent burglary risk, whilst acknowledged as of great interest, is outside the scope of the present study.
Non-victims refer to those households which have not encountered any type of burglary. Attempts are not discussed in this paper but in work not reported here they have been analysed similarly to burglary with entry, so they are not removed from the analysis or subsumed with non-victims.
By ‘no security’ we mean respondents who stated they had none of the listed security devices. Therefore this is not directly comparable before and after 1998 as the security devices included in the CSEW change (see Appendix). However we attempt to adjust for this in Fig. 4 and the later Figs. 5 and 6.
Area type- and population group- specific estimates of the relationship between burglary risk and security during the crime drop is indeed beyond the scope of the current work and requires a different methodological approach to avoid omitted variables problems.
These are defined here as security installed in more than 10% of households.
The 1992–1996 set of data is omitted in Fig. 3. The CSEW sweeps have not asked about an identical range of security devices fitted to households before and after 1998 impeding comparisons of exact security combinations with the 1992–1996 period.
Any individual security excludes CCTV cameras which were available to only 0.15% of households as a single device in the period 2008/09 to 2011/12. Such negligible prevalence of CCTV cameras in households in effect does not influence the pre- and post-2008/09 comparisons of ‘no security’ and any suites of security devices. Overall (individually or in combination with other devices) 4.6% of households had CCTV cameras.
These important for the security hypothesis research questions may draw on longitudinal criminal careers data.
We note that street crime did not switch to cyber-crime: the internet arrived too late and cyber-crime is a poor substitute for car theft and burglary (as argued by Farrell et al. 2015).
For example, in the 2008/09 to 2011/12 CSEW data 2.66% of burglaries with entry with available security information were repeats.
context mechanisms outcome pattern configurations
concealable, removable, available, valuable, enjoyable and disposable
concealable, removable, accessible, abundant, valuable, enjoyable and disposable
Crime Survey for England and Wales
door double or deadlocks
external lights on a sensor
external lights on a sensor, window locks, burglar alarm and door double or deadlocks
external lights on a sensor, window locks and door double or deadlocks
indoor lights on a timer
net interaction effect
Office for National Statistics
secured by design
security impact assessment tool
security protection factor
window locks and door double or deadlocks
window locks, indoor lights on a timer, door double or deadlocks and external lights on a sensor
RT and EE analysed the data on security presence and effectiveness and modus operandi trends while working as Research Staff to the project under the supervision of the Principal Investigator (AT) and the Co-Investigators (GF and NT). NT provided the spine of the manuscript and outlined the five key hypotheses. GF provided the methodological approach. AT was responsible for delivering the manuscript, including revisions following Reviewers’ suggestions. All authors worked jointly on the various sections of the paper. All authors read and approved the final manuscript.
This study is part of the research project ‘Which burglary security devices work for whom and in what context?’ undertaken with funding from the Economic and Social Research Council, Secondary Data Analysis Initiative-Phase 1 (ES/K003771/1 & ES/K003771/2). The authors are indebted to the project’s Advisory Committee (http://www4.ntu.ac.uk/app_research/soc/document_uploads/178982.pdf) for their support for the duration of the research project and contribution to this work; Dr. Louise Grove for past collaboration; and the insightful comments and corrections of the Anonymous Reviewers of a previous version. Any errors or omissions are entirely the Authors’ responsibility.
The authors declare that they have no competing interests.
Availability of data and materials
The work is based on publicly available data within the UK Data Service.
Consent for publication
All authors give consent for publication.
The work reported here is part of the project “Which burglary security devices work for whom and in what context?” for which funding was provided from the Economic and Social Research Council, Secondary Data Analysis Initiative-Phase 1 (ES/K003771/1 & ES/K003771/2). The funding body had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Armitage, R., & Monchuk, L. (2011). Sustaining the crime reduction impact of designing out crime: Re-evaluating the secured by design scheme 10 years on. Security Journal, 24, 320–343.View ArticleGoogle Scholar
- Bässman, J. (2011). Vehicle theft reduction in Germany: The long-term effectiveness of electronic immobilisation. European Journal of Criminal Policy and Research, 17(3), 221–246.View ArticleGoogle Scholar
- Bettaïeb, I. (2015). Le Profil des Ménages S’étant Déclarés Victimes d’une Entrée ou d’une Tentative d’entrée dans la Résidence Principale: Exploitation des Huit Enquêtes de Victimation «Cadre de Vie et Sécurité» 2007 à 2014. 11ème Rapport Annuel de l’ONDRP.
- Blumstein, A., & Wallman, J. (2000). The crime drop in America. New York: Cambridge University Press.Google Scholar
- Brown, R. (2013). Reviewing the effectiveness of electronic vehicle immobilisation: Evidence from four countries. Security Journal. doi:10.1057/sj.2012.55.Google Scholar
- Brown, R. (2015a). Explaining the property crime drop: The offender perspective. Trends and issues in crime and criminal justice. Canberra: Australian Institute of Criminology.Google Scholar
- Brown, R. (2015b). Crime prevention design in a vehicle registration system: A case study from Australia. Crime Science, 4, 25. doi:10.1186/s40163-015-0038-1.View ArticleGoogle Scholar
- Budd, T. (1999). Burglary of domestic dwellings: Findings from the British Crime Survey. Home Office Statistical Bulletin 4/99. London: Home Office.Google Scholar
- Clarke, R.V. (1999). Hot products: Understanding, anticipating and reducing demand for stolen goods. Police Research Series, Paper 112. London: Home Office.
- Clarke, R. V. (2012). Opportunity makes the thief. Really? So what? Crime Science, 1(3), 1–9.Google Scholar
- Clarke, R. V. (2016) Criminology and the fundamental attribution error. The Criminologist, 41(3), 1–7.
- Clarke, R. V., & Newman, G. R. (2006). Outsmarting the terrorists. Political science. Westport: Greenwood Publishing Group.Google Scholar
- Dinardo, J. (2010). Natural experiments and quasi-natural experiments. In S. N. Durlauf & L. E. Blume (Eds.), Microeconometrics (pp. 139–153). London: The New Palgrave Economics Collection, Palgrave Macmillan.Google Scholar
- Eck, J. E., & Madensen, T. (2009). Using signatures of opportunity structures to examine mechanisms in crime prevention evaluations. In J. Knutsson & N. Tilley (Eds.), Evaluating crime prevention initiatives (pp. 59–84). Monsey: Criminal Justice Press.Google Scholar
- Ekblom, P. (1996). Safer Cities and residential burglary. European Journal on Criminal Policy and Research, 4(1), 22–52.View ArticleGoogle Scholar
- Ekblom, P., & Tilley, N. (2000). Going equipped: Criminology, situational crime prevention and the resourceful offender. British Journal of Criminology, 40(3), 376–398.View ArticleGoogle Scholar
- Farrell, G. (2013). Five tests for a theory of the crime drop. Crime Science, 2(5), 1–8.Google Scholar
- Farrell, G., Laycock, G., & Tilley, N. (2015). Debuts and legacies: The crime drop and the role of adolescent-limited and persistent offending. Crime Science. doi:10.1186/s40163-015-0028-3.Google Scholar
- Farrell, G. & Pease, K. (1993). Once bitten, twice bitten: Repeat victimisation and its implications for crime prevention. Crime Prevention Unit Paper 46. London: Home Office.
- Farrell, G., Tilley, N., & Tseloni, A. (2014). Why the crime drop? Crime and Justice, 43, 421–490.View ArticleGoogle Scholar
- Farrell, G., Tilley, N., Tseloni, A., & Mailley, J. (2008). The crime drop and the security hypothesis. British Society of Criminology Newsletter, 62, 17–21.Google Scholar
- Farrell, G., Tilley, N., Tseloni, A., & Mailley, J. (2010). Explaining and sustaining the crime drop: Exploring the role of opportunity-related theories. Crime Prevention and Community Safety: An International Journal, 12, 24–41.View ArticleGoogle Scholar
- Farrell, G., Tseloni, A., Mailley, J., & Tilley, N. (2011a). The crime drop and the security hypothesis. Journal of Research in Crime and Delinquency, 48(2), 147–175.View ArticleGoogle Scholar
- Farrell, G., Tseloni, A., & Tilley, N. (2011b). The effectiveness of vehicle security devices and their role in the crime drop. Criminology and Criminal Justice, 11(1), 21–35.View ArticleGoogle Scholar
- Flatley, J. (2014). British Crime Survey. In G. Bruinsma & D. Weisburd (Eds.), Encyclopaedia of criminology and criminal justice (ECCJ) (pp. 194–203). New York: Springer.View ArticleGoogle Scholar
- Fujita, S., & Maxfield, M. (2012). Security and the drop in car theft in the United States. In J. Van Dijk, A. Tseloni, & G. Farrell (Eds.), The international crime drop: New directions in research (pp. 231–249). London: Palgrave Macmillan.View ArticleGoogle Scholar
- Hough, M., & Maxfield, M. (2007). Surveying crime in the 21st century. Crime prevention studies volume 22. Monsey: Criminal Justice Press.Google Scholar
- Hunter, J., & Tseloni, A. (2016). Equity, justice and the crime drop: The case of burglary in England and Wales. Crime Science. doi:10.1186/s40163-016-0051-z.Google Scholar
- Ignatans, D., & Pease, K. (2015). Distributive justice and the crime drop. In M. Andresen & G. Farrell (Eds.), The criminal act: Festschrift for Marcus Felson (pp. 77–87). London: Palgrave Macmillan.View ArticleGoogle Scholar
- Ignatans, D., & Pease, K. (2016). On whom does the burden of crime fall now? Changes over time in counts and concentration. International Review of Victimology, 22(1), 55–63.View ArticleGoogle Scholar
- International Centre for the Prevention of Crime (ICPC) (2015). Prevention of residential burglaries: Lessons drawn from a comparative approach—summary. http://www.crime-prevention-intl.org/fileadmin/user_upload/Publications/Prevention_of_Residential_Burlglaries.pdf. Accessed 27 Jul 2015.
- Jansson, K. (2007). British crime survey—measuring crime for 25 years. London: Home Office.Google Scholar
- Kriven, S., & Ziersch, E. (2007). New car security and shifting vehicle theft patterns in Australia. Security Journal, 20, 111–122.View ArticleGoogle Scholar
- Mayhew, P., Clarke, R. V., Sturman, A., & Hough, M. (1976). Crime as opportunity. Research and planning unit research study, volume 34. London: Home Office.Google Scholar
- McGuire, M. & Dowling, S. (2013). Cyber crime: A review of the evidence. Home Office Research Report 75. London: Home Office.
- Office for National Statistics (ONS). (2013a). Chapter 2: Mobile phone theft. Crime Statistics, Focus on Property Crime, 2011/12 (pp. 1–22). http://www.ons.gov.uk/ons/dcp171776_309652.pdf. Accessed 09 May 2015.
- Office for National Statistics (ONS). (2013b). Crime in England and Wales, year ending December 2013. Statistical Bulletin, 24 April 2014. http://www.ons.gov.uk/ons/dcp171778_360216.pdf. Accessed 19 Jun 2014.
- Osborn, D. R., & Tseloni, A. (1998). The distribution of household property crimes. Journal of Quantitative Criminology, 14, 307–330.View ArticleGoogle Scholar
- Pawson, R., & Tilley, N. (1997). Realistic evaluation. London: Sage.Google Scholar
- Pease, K. (1997). Predicting the future: The roles of routine activity and rational choice. In G. Newman, R. V. Clarke, & S. G. Shoham (Eds.), Rational choice and situational crime prevention (pp. 233–245). Aldershot: Ashgate.Google Scholar
- Philipson, T. J., & Posner, R. A. (1996). The economic epidemiology of crime. Journal of Law and Economics, 39(2), 405–436.View ArticleGoogle Scholar
- Pires, S., & Clarke, R. V. (2012). Are parrots craved? An analysis of parrot poaching in Mexico. Journal of Research in Crime and Delinquency, 49(1), 122–146.View ArticleGoogle Scholar
- Poyner, B. (1992). Situational crime prevention and two parking facilities. In R. V. Clarke (Ed.), Situational crime prevention: Successful case studies (pp. 157–166). New York: Harrow and Heston.Google Scholar
- Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Boston: Houghton Mifflin.Google Scholar
- Shaw, O., Morgan, N., Mineva, I., Hoare, J., & Stevenson, R. (2015). Crime and the value of stolen goods. London: Home Office.Google Scholar
- Smith, D.G., Gregson, M. & Morgan, J. (2003). Between the lines: An evaluation of the secured by design car park award scheme. Home Office Research Study 266. London: Home Office.
- Thompson, R. (2014). How theft and robbery have changed since the 90s. Significance 21 October. http://www.statslife.org.uk/social-sciences/1842-how-theft-and-robbery-have-changed-since-the-90s. Accessed 27 Jul 2015.
- Tilley, N. (1993). Understanding car parks, crime and CCTV. Crime Prevention Unit Series Paper 42. London: Home Office.
- Tilley, N. (2012). Community, security and distributive justice. In V. Ceccato (Ed.), The urban fabric of crime and fear (pp. 267–282). New York: Springer.Google Scholar
- Tilley, N., Farrell, G., & Clarke, R. V. (2015a). Target suitability and the crime drop. In M. Andresen & G. Farrell (Eds.), The criminal act (pp. 59–76). London: Routledge.View ArticleGoogle Scholar
- Tilley, N., Thompson, R., Farrell, G., Grove, L., & Tseloni, A. (2015b). Do burglar alarms increase burglary risk? A counterintuitive finding and possible explanations. Crime Prevention and Community Safety, 17(1), 1–19.View ArticleGoogle Scholar
- Tilley, N., Tseloni, A., & Farrell, G. (2011). Income—disparities of burglary risk and security availability over time. British Journal of Criminology, 51(2), 296–313.View ArticleGoogle Scholar
- TNS-BMRB. (2012). The 2012/13 Crime Survey for England and Wales Technical Report Volume One. http://www.ons.gov.uk/ons/guide-method/method-quality/specific/crime-statistics-methodology/2012-13-crime-survey-for-england-and-wales-technical-report—volume-one.pdf. Accessed 25 Jun 2015.
- Tonry, M. (2014). Why crime rates fall and why they don’t. Chicago: University of Chicago Press.Google Scholar
- Tseloni, A., Mailley, J., Farrell, G., & Tilley, N. (2010). The cross-national crime and repeat victimisation trend for main crime categories: Multilevel modelling of the international crime victims survey. European Journal of Criminology, 7(5), 375–394.View ArticleGoogle Scholar
- Tseloni, A., & Thompson, R. (2015). Securing the premises. Significance, 12(1), 32–35.View ArticleGoogle Scholar
- Tseloni, A., Thompson, R., Grove, L., Tilley, N., & Farrell, G. (2014). The effectiveness of burglary security devices. Security Journal. doi:10.1057/sj.2014.30.Google Scholar
- Van Dijk, J. (2006). What goes up, comes down: Explaining the falling crime rates. Criminology in Europe—Newsletter of the European Society of Criminology, 5(3), 17–18.Google Scholar
- Van Dijk, J., & Tseloni, A. (2012). Global overview: International trends in victimisation and recorded crime. In J. Van Dijk, A. Tseloni, & G. Farrell (Eds.), The international crime drop: New directions in research (pp. 11–36). London: Palgrave Macmillan.View ArticleGoogle Scholar
- Van Dijk, J., Tseloni, A., & Farrell, G. (2012). The International crime drop: New directions in research. London: Palgrave Macmillan.View ArticleGoogle Scholar
- Van Ours, J. C., & Vollaard, B. (2016). The engine immobilizer: A non-starter for car thieves. Economic Journal, 126(593), 1264–1291.
- Vollaard, B., & van Ours, J. C. (2011). Does regulation of built-in security reduce crime? Evidence from a natural experiment. The Economic Journal, 121, 485–504.View ArticleGoogle Scholar
- Webb, B. (2005). Preventing vehicle crime. In N. Tilley (Ed.), Handbook of crime prevention and community safety (pp. 458–485). London: Routledge.Google Scholar