What Works to Reduce Crime?: A Summary of the Evidence

PART THREE: REDUCING THE OPPORTUNITIES FOR OFFENDING

Introduction

Our third broad strategy for reducing crime is, in a sense, a 'fall-back 'or last resort. It is to reduce the opportunities for crime so that those, for whom the first two strategies have been ineffective, simply find it more difficult to offend. A key approach is to seek to reduce opportunities for crime through some modification or manipulation of the physical environment. It is an approach that has been heavily influenced by the rational choice perspective articulated by Clarke^[131]. This perspective assumes that the offender makes reasoned decisions about his/her actions by weighing up his/her needs for such things as money, status, sex, and excitement, with the opportunities presented by a particular focus or target. However, this section also considers broader approaches that extend beyond the 'situation' and includes restricting the movement/ activities of those at risk of offending; imposing wider restrictions on society as a whole (such as controls on access to weapons); and occupying potential offenders with recreational diversions at times when they are most at risk of offending.

Situational crime prevention

Situational crime prevention strategies are based on the 'routine activities' theory of crime - that is, that crime occurs where there is a clustering of a motivated offender, the opportunity to offend and the absence of a guardian^[132]. It is assumed that rationality applies not only to acquisitive crime but also to violent offending. While there are some (such as Trasler^[133]) who argue that opportunity-reducing prevention is effective only with 'instrumental' offences (such as property or robbery) but not with 'expressive' offences that are not subject to rational calculus (such as homicide), the empirical literature on the subject (albeit limited) suggests that situational crime prevention can be effective in reducing some violent crimes^[134]. It has also been argued that situational prevention is most effective with casual, uncommitted offenders and least effective with persistent or chronic offenders^[135].

There are a range of situational crime prevention strategies that can be used to reduce the opportunities for crime. These include:

• Increasing formal surveillance using electronic alarms, CCTV, private security patrols or neighbourhood watch;
• Increasing natural surveillance by removing obstacles to line of sight or improving street lighting;
• Increasing access or exit controls;
• Concealing or removing targets (such as lightweight, high value goods that tend to be easy to steal: jewellery; cash; IPods; mobile phones);
• Erecting or strengthening physical barriers such as window locks, double-pane glass, which is both more difficult and noisier to break, and deadbolt locks on doors^[136];
• Restricting access to the tools of crime (such as spray paints);
• Reducing provocation.

Most crime prevention initiatives combine and evaluate more than one strategy at a time. For this reason, there is little evidence about the separate impact of each of these strategies. This section therefore begins with the evidence from initiatives that combine a range of approaches before moving on to the more limited evidence on the impact of individual approaches.

Kirkholt burglary-reduction project

An early initiative that has been evaluated is the Kirkholt burglary-reduction project that ran between 1987 and 1990^[137]. There were two phases to this project. Phase One ran for the first year and involved: removal of prepayment fuel meters (prior to the intervention, 49% of burglaries in Kirkholt involved theft of meter cash) and introduction of exit payment (as a means of preventing swift exit); improved household security; a community support team (the primary role of which was to visit the victims of burglary on the estate, offer support, and put them in touch with appropriate agencies) and a 'cocoon' neighbourhood watch (where close groupings of dwellings share information and support each other). Phase Two introduced three new elements in the second and third years: probation group work programme (which involved offenders engaging in community work such as cleaning up public spaces); a credit union for people living on the estate; and a project for schools to provide recreational activities as an means of preventing vandalism or petty crime. Kirkholt estate was chosen due to its reputation for high crime (the rate of recorded residential burglary was more than double the national rate) and because it had well-defined boundaries.

Two studies looked at the effect of this project on repeat-burglary. The main evaluation^[138] design was a pre and post comparison repeated for each year of the project. The results of the first phase showed that burglary in the area fell from 316 in 1986 to 147 in 1987. That meant they fell to 40% of the pre-initiative level within five months of the start of the programme. As the second phase was introduced, residential burglaries continued to decrease. The annual number of residential burglaries fell from the pre-year to the first post-year by 38%, by 67% in the second year and by 72% in the final year. Given that the incidence of burglary had been reduced, the researchers concluded that the Kirkholt initiative resulted in a substantial reduction in burglaries over the three years of the project.

Secure by Design (SBD)

Another crime prevention initiative that combines a number of strategies and which has been well evaluated in Secure by Design. Secure by Design (SBD) forms part of a commitment to designing out crime set out in the UK government's crime prevention strategy: Cutting Crime: A New Partnership 2008-2011 (Home Office, 2007). SBD is essentially about promoting good practice in urban planning to ensure that designers, architects, planners and builders consider crime prevention measures during the design stage of any proposals. The key principles that are built into the planning to help reduce crime, disorder and fear of crime are:

• minimum standards of physical security;
• a minimum number of access points;
• maximum natural surveillance (such as windows overlooking gardens and clear sightlines with no high walls);
• good management and maintenance policies;
• and informal social control through a mix of dwellings.

There have been a number of evaluations of the SBD schemes that show they have been effective in reducing crime.

In a major evaluation of 25 SBD and 25 matched non-SBD estates, Armitage^[139] found that for developments refurbished to the SBD standard, total crime fell by 55% relative to the pre-SBD period. The evaluation also involved a comparison of police-recorded crime rates in SBD initiative and non- SBD initiative estates after the initiative had been implemented. Recorded crimes were analysed for the period between each estate's completion and March 2000. The prevalence rate of total crime (the proportion of dwellings which were offended against at least once) was less in the SBD sample - 44% of dwellings were offended against once or more on non-SBD estates, compared with 37% on SBD estates. Furthermore, there were almost twice the number of burglary offences within the non-SBD sample. The additional cost of building a house to SBD standards at the construction stage has been estimated by Armitage at £440. On the basis that the average cost of burglary to the victim (at 1999 prices) was £2,300^[140], Armitage argues "the extra expenditure required to build or refurbish housing to SBD standards would appear to be a worthwhile investment"^[141]. Furthermore, a survey of residents suggests that fear of crime was lower amongst those living on SBD estates.

In 2007, Armitage and Monchuk carried out a re-evaluation of the SBD areas evaluated in 2000. Their re-evaluation shows that SBD continued to reduce crime and the fear of crime - SBD developments sustained their crime reduction benefits and continued to experience less crime, fewer visual signs of disorder and less fear of crime amongst residents than their non SBD comparisons^[142].

In Scotland, Teedon and Reid^[143] conducted an evaluation of SBD in Glasgow. The results revealed that total housebreaking crime reduced by 61% following the introduction of SBD. This is compared to a reduction of just 17% in the comparison area. In this case, the total housebreaking figure was made up of the three categories: attempted housebreaking; housebreaking with intent; and theft by housebreaking. Even when 'attempts' were removed from the analysis, the reduction was still extremely positive (a 55% reduction).

Household level measures

While the initiatives reviewed above were implemented at an area level, there is also evidence that measures introduced at individual household level can also be effective in preventing burglary. Budd's multivariate analysis of British Crime Survey data strongly suggests the effectiveness of security measures such as deadlocks and window locks or grilles^[144] . Knight and Pascoe reviewed both British and international studies of the available evidence and reached the same conclusion^[145]. More recently, findings from a study of burglary in the UK, USA and the Netherlands found security measures in the home to be one of four variables affecting victimisation rates across all three countries^[146]. The other three variables identified were age of occupant, lone parent household status, and urbanization.

Vandalism

The crime category of 'vandalism etc.' covers a number of more detailed crimes such as reckless conduct (with and without firearms) and accounts for almost half (48%) of the fall in total recorded crime in Scotland between 2006-07 and 2012-13. A guide produced by the Home Office suggest that projects to engage young people in diversionary activities are effective in reducing vandalism/criminal damage combined with 'target hardening' measures (i.e. via improving natural surveillance opportunities) as well as work with young people in schools to raise awareness of the impact and consequences of vandalism. It is also suggested that attempts to reduce this type of offence benefits particularly from a partnership approach, in terms of engaging the various agencies who are affected including local people and businesses^[147]. Misuse of alcohol was also highlighted as an exacerbating factor in increasing the risk of vandalism and consequently tackling this as a root cause is also considered key^[148].

Vehicle crime

Internationally, vehicle crime has fallen substantially over the past twenty years and this has been a major factor in the overall decline in crime rates. Tseloni et al. (2010) examined aggregate crime rates across 26 countries from 1988 to 2004 and found that over this period there was a mean 77.1% reduction in theft from cars and a 16.8% drop in theft of cars, and that these crime types fell by roughly the same rate across countries. In Scotland, since 1992 theft from/of a vehicle has fallen from 22% to 4% of all crime. Between 2011-12 and 2012-13 alone there was 20% drop in vehicle crime^[149].

Advances in technology and vehicle security, such as engine immobilisers and central locking, have made cars much more difficult to steal, and to steal from. In 1998 the EU made application of the device in all new cars mandatory. Engine immobiliser technology works by blocking a vehicle's electrical circuits when the key is not in the ignition, thus preventing hot-wiring (a technique commonly used by car thieves prior to the introduction of the device). In a Netherlands based study van Ours and Vollaard found the introduction of the engine immobiliser to have reduced the overall rate of car theft by half between 1995-2008 (taking into account the displacement of theft to older cars without the device)^[150]. It is also argued that because vehicle crime has served as a common 'debut' crime, particularly for young men, that making vehicle crime more difficult has had a knock-on effect on crime rates overall as fewer young people will be drawn into a criminal career^[151].

Business crime

Two studies have examined the effectiveness of situational crime prevention measures in reducing crime against small businesses. The first initiative aimed at reducing small business crime in Liverpool and sought to reduce repeat burglaries against small businesses in deprived neighbourhoods^[152]. The measures used included burglar alarms, CCTV, roller shutters, window locks and detection devices as well as advice on managing affairs such as keeping limited amounts of money, not disclosing information to strangers and being careful with keys. Comparisons were made between the intervention areas and other non-residential areas using police recorded crime data. The evaluation found that, with the exception of robbery, the number of businesses experiencing crime reduced in both areas during the evaluation period. The reductions were largest for burglary, attempted burglary, shoplifting, fraud and forgery. However, the reduction in the prevalence of burglary was greater in the intervention areas - 59% reduction compared with a only 7% reduction in the non-intervention areas.

The second study was of a three year project in two areas of Leicester^[153]. Measures were introduced to businesses that were thought to be at high risk of victimisation. Measures included temporary silent alarms with direct lines to the police, cover CCTV, forensic traps designed to obtain footwear marks and a hidden movement detector that triggered an audible alarm. The findings based on the analysis of two victimisation surveys showed that both the incidence and prevalence of most crime categories investigated fell from the first pre-implementation survey and a second survey two years later. The incidence of burglary fell by 41% and the prevalence by 36%. The findings of the surveys also show that repeat-burglary victimisation fell over the same period by 6%. The findings from the police-recorded crime data analysis for the areas covered by the scheme also show an overall reduction of non-domestic burglary. Although data for the county and the rest of England and Wales showed a general decline in non-domestic burglary, this decline was not as great as that experienced in the scheme areas. The authors conclude that crime affecting businesses decreased substantially in the target areas. However, the authors caution that the evidence is not clear that the work of the initiative played a significant part in this reduction.

There is also evidence that measures to tackle credit card and other frauds can be extremely effective. The exact measures and checks applied will vary depending on whether the purchase is being made face to face, by telephone or via the internet. Three evaluations have examined attempts to prevent credit card fraud at the point of sales, and suggest that reductions in magnitude of 82 - 90% have been achieved^[154].

Increasing formal surveillance

A review of the evidence on the effectiveness of public space CCTV in reducing crime was conducted and published by the Scottish Government in 2009^[155]. This review found that very little evaluative research into the effectiveness of CCTV had been conducted since 2000. A Campbell review of the effectiveness of CCTV^[156] found that the available evidence shows that CCTV has a "modest but significant" impact on crime, and that it is most effective at reducing crime in car parks. The review included 44 studies which, at a minimum, involved pre-and post- measures of crime in experimental and control areas, but as there have been no RCTs conducted on the effectiveness of CCTV the researchers identified a need for more robust studies with longer follow-up periods. The available evidence provided minimal evidence to suggest that CCTV effectively deters crime, and in cases where crime does appear to be deterred, this effect is generally short-lived. Evidence consistently shows that CCTV may be more effective in deterring crime in smaller and less complex areas than large city centres. The opinions of convicted offenders largely suggest that cameras are not perceived as a threat, particularly in situations fuelled with alcohol. There is some evidence of a diffusion of benefits in terms of crime reduction to surrounding areas following CCTV installation but, like deterrence, these effects appear to diminish with time. By comparison, the evidence on whether or not CCTV simply displaces crime to neighbouring areas (the 'displacement' effect) remains inconclusive though there is some evidence to suggest that crime displacement may occur on a small scale, within the local (CCTV targeted) area itself.

There is also evidence that CCTV may be more effective in terms of increased crime detection than it is in terms of deterrence, particularly in the case of violent crime. The use of CCTV is therefore also relevant (perhaps more relevant) to the branch of our reducing crime model that focuses on deterring potential offenders by increasing the risk of prosecution by improving the availability and quality of evidence.

Increasing natural surveillance

One of the most common methods of increasing natural surveillance and one which has been evaluated separately is improved street lighting. Although improved street lighting may be implemented for a range of reasons, other than crime prevention, it can be used to reduce the opportunity and increase the perceived risk for offenders. This is supported by theories that emphasise natural, informal surveillance as key to crime prevention. Jacobs^[157], for example, highlighted the role of good visibility, combined with natural surveillance (for example, a window overlooking a path or road) as a deterrent to crime.

A separate meta-analysis by Welsh and Farrington^[158] of 13 improved street lighting evaluations found that they were effective in reducing crime. They found that crime was 21% lower in the experimental areas than in the control areas. Welsh and Farrington concluded, overall, that the results of their systematic review indicate that improved street lighting significantly reduces crime^[159] . However, there was some evidence that it is more effective in reducing property crimes (burglary, vehicle crimes and theft) than violent crimes (including robbery and assault).

Among the studies reviewed by Welsh and Farrington were five that were carried out in the UK. These were carried out in a variety of settings, including a parking garage and a market, as well as residential neighbourhoods. Three evaluations specified the degree of improvement to lighting: by five times in Stoke on Trent and by two times in Bristol and Dudley. The outcome measure was based on police recorded crime for three studies and on victim surveys in the other two cases (Stoke on Trent and Dudley). Uniquely, as well as using self-reports of victimisation of young people and measures of fear of crime, the Dudley project also evaluated the impact of the improved street lighting using self-reported delinquency surveys of young people^[160]. The results of this survey revealed that young people thought that crime had decreased in the area with improved street lighting and also reported that their own fear of crime had decreased. Results of the meta-analysis of the five British studies confirm that improved street lighting had caused a decrease in crime in the experimental areas and that there had been a diffusion of benefits. According to Welsh and Farrington, there were 38% fewer crimes in the experimental areas compared to the control areas. Although the evidence suggests that the extent of impact is dependent on the characteristics of situation and on other concurrent situational interventions, the review did not identify what were those optimal circumstances^[161]. However, the authors did suggest that improved street lighting may be more effective in stable, homogenous communities than in unstable, heterogeneous communities.

Welsh and Farrington also suggest that the introduction of street lighting may reflect a new or recovered sense of community pride^[162]. It has been suggested, elsewhere, that investment in improving the management and maintenance of a neighbourhood can be an important means of strengthening community confidence, cohesion and, in turn, informal social control^[163]. In the Dudley evaluation referred to above, there was some evidence of an increased sense of collective efficacy following improvements in street lighting. The local tenants had complained to the local authorities about the poor lighting. The local authority improved the lighting and the tenants reported that they thought the quality of their life had improved. This prompted the tenants association to obtain a £10million grant for a programme of neighbourhood improvements from the Department of the Environment. A similar chain of events was noted in the evaluation of street lighting in Stoke on Trent. Although Welsh and Farrington say that they cannot be certain that the causal ordering occurred in all other street lighting evaluations, they argue that, at least in some studies, improved street lighting increased community pride and collective efficacy and this increased collective efficacy had a role in addressing and reducing crime. This interpretation seems consistent with the discussion in Section Two about the link between collective efficacy and reduced crime.

Finally, on the subject of natural surveillance, a number of studies (e.g. Newman^[165] and Poyner and Webb^[166]) have found that busier streets with some pedestrian movement have experienced reduced levels of recorded crime.

Access and exit controls

Access control focuses on reducing opportunities for crime by denying access to potential targets and creating a heightened perception of risk in offenders. Access control can include:

• informal/natural (e.g. limiting the number of access and exit points);
• formal/organised (e.g. security personnel); and
• mechanical (e.g. locks and bolts).

A number of studies, notably by Newman^[167] and Poyner and Webb^[168] suggest an association between design features and levels of crime particularly features that allowed unrestricted pedestrian movement through residential complexes. Research suggest that areas with unregulated access have more crime than areas with street layouts with more restricted access.

Welsh and Farrington examined five evaluations of defensible space that included street barriers or street closures - four were in the USA and one was in the UK. These initiatives were focused on general crime reduction in the USA (for example to prevent offenders from driving away from a robbery) and in the UK to prevent kerb-crawling^[169]. They concluded, based on the evaluations, that there is fairly strong and consistent evidence that street barriers or street closures are effective in preventing crime in inner-city neighbourhoods. However, there are differing opinions on how this mechanism works. Cornish and Clarke argue that the physical barrier acts to deflect offenders away from crime targets, whilst others suggest that the improvement is due to increased surveillance due to people feeling safer and being out and about.

In London, two attempts to reduce street-level prostitution used road closures, rerouting and an increased police presence. After road closures (and an increase in policing prior to the road closures) in Finsbury Park, soliciting and kerb-crawling virtually disappeared with little recorded displacement and reported crime fell by 50 %^[170]. In Streatham a similar project reported a decline in traffic flows along major thoroughfares, a reduction in arrests of kerb-crawlers (although there may be several explanations of this) and residents reported a decline in prostitution at street level^[171].

Ekblom^[172] reports on an alley-gating project in Birmingham where 80% of burglaries were committed using access from rear alleys. After erecting 62 alley-gates, steel palisade fencing, the distribution of 400 ultraviolet property-marking kits and stickers and a local newsletter, a 53% decline in burglaries was reported. However, given the number of interventions applied in this example, it is impossible to identify the separate impact of restricting access through alley-gating.

Another form of access control is the use of security guards and place managers (e.g. bus conductors, receptionists). Welsh and Farrington found five evaluations of security guards - two from the USA and one from Canada, the Netherlands and the UK. The UK study was a multi-measure project with security guards at a car park in Basingstoke supplemented by other interventions: erecting a fence around most of the site; pruning trees to ensure natural surveillance; and building a public footpath on one side. The project was evaluated by the Home Office as being highly effective. The US studies involved the 'Guardian Angels' - a non-profit, volunteer organization of unarmed citizen crime patrollers with chapters in a number of cities. The studies assessed the 'Guardian Angels' as having uncertain or mixed results^[173].

Welsh and Farrington also looked at place managers. They found only two evaluations, both in the UK, that met the criteria for inclusion in their systematic review. They describe a study from 1988 of a concierge scheme in South Kilburn. The concierge performed three functions: acting as a receptionist; providing general assistance to residents; and maintaining block security by controlling access through the main entrance. The evaluation found that, compared to neighbouring residential high-rise housing blocks on the estate, the experimental site showed a number of benefits over a one year follow up period including fewer repairs to communal areas and lifts due to a reduction in vandalism. The second evaluation was of a multi-level parking garage which had severe problems of thefts of and from parked vehicles. The intervention involved the presence of parking attendants as place managers. Two years after the project police-reported vehicle crimes were down by half in both the experimental and control area. There was therefore no discernible impact. Poyner^[174] argues that parking lot strategies may curb the theft of vehicles but not theft from vehicles. Furthermore, Welsh and Farrington^[175] note that there is limited information on the deployment of parking attendants.

In conclusion, Welsh and Farrington comment that the low number of high quality evaluations make it difficult to say whether or not security guards and place managers are effective. Furthermore, the multi-measure approaches applied making it difficult to isolate the impact of specific crime prevention techniques^[176].

Overall, the evidence on access control is mixed. Street closures and barriers do appear to be effective in preventing crime. Similarly, automatic gates and locked cash boxes that staff cannot open or access have shown significant reductions in fare-dodging and attempted robberies. However, the evidence of other forms of access control using security personnel, door staff or caretakers is unclear.

With regard to exit controls, new automatic gates have been shown to reduce fare evasion on public transport and reportedly increase ticket sales by 10% in comparison with control stations without gates^[177]. New gates at transit stations in New York have led to a reduction in arrests^[178].

Alley-gating is an approach to situational crime prevention which involves installing lockable gates at the end of alleyways running behind houses to prevent access for potential offenders. In Liverpool, for example, over 3000 gates were installed over a period of six years. Bowers et al. evaluated the impact of the scheme from implementation in 1999 until 2002^[179], comparing crime figures for the gated area with a suitable comparison area for periods pre- during and post- implementation of the gates. The researchers also looked at crime rates in buffer zones adjacent to the gated area in order to check for any displacement effects. The results demonstrated a 37 per cent reduction in crime in the gated area, once general changes in crime rates in the surrounding area had been taken into account. There was also a diffusion of benefits to surrounding areas. The evaluation concluded that for every £1 spent on alley-gating £1.86 had been saved.

There is evidence to suggest that the benefit of situational crime prevention measures may not be sustained over time, unless initiatives are maintained and managed closely. The success of alley-gating, for example, depends on residents using their keys and keeping gates locked, as well as the gates being maintained and not vandalised. Armitage and Smithson conducted a survey with residents in the gated area and a comparison area with no gating, in order to assess the continuing impact of the alley gates (up to 2006), and to explore residents' perceptions of crime and safety in their local area^[180]. Residents living in gated areas consistently reported experiencing less crime and ASB, and feeling safer in and around their homes than those in non-gated areas. When asked about their perceptions of the alley gates, 92% of residents felt they had made their street safer, and 83% felt that the gates had helped improve the image of the street. A physical survey of residents' properties found that gates were locked in 99% of cases and there was only evidence of damage to the gates in 3% of cases. Overall, reductions in crime and ASB appear to have been maintained over time, and the findings also indicate that there were further benefits in terms of improving and maintaining the tidiness and appearance of the streets.

Concealing or removing targets

Efforts to conceal or remove targets include off-street parking, gender neutral phone directories, pre-paid phone-cards (which mean money does not have to be stored in telephone booths) and removable car stereos. The best documented examples of successful target removal involve various cash reduction measures to reduce the potential threat of robbery. Two evaluations of attacks on bus drivers provide evidence that these measures are effective. A number of US cities removed accessible cash with the introduction of exact fares and prohibited bus drivers from giving change. Chaiken, Lawless and Stevenson^[181] reported a 90% reduction in bus driver robberies in New York following these changes. The Stanford Research Institute^[182] reported similar findings in its review of the effect of exact fare systems in 18 other cities. Time lock cash boxes at betting shops in Australia reduced robberies considerably compared with the control group^[183] . Clarke and Goldstein^[184] found that theft of household appliances installed by builders in new houses was significantly reduced when these items were not fitted until new owners were in residence. There was also found to be no evidence of displacement.

Erecting or strengthening physical barriers

As a real and symbolic barrier to crime, protective screens around drivers on buses have been shown to significantly reduce assaults on drivers^[185] . Ekblom found that the after the installation of bullet-proof barriers to protect post office clerks, robberies were 65% lower in those post-offices than in sub-post offices not fitted with screens^[186].

However, commentators have also note that excessive use of target hardening tactics can create a "fortress mentality" whereby residents withdraw behind physical barriers and the self-policing capacity of the built environment is damaged, effectively working against SBD strategies that rely on surveillance, territoriality and image^[187].

Restricting access to the tools of crime

Ekblom and Tilley^[188] suggest that removing resources for crime can be an important mechanism in crime prevention. They present a list of seven resources that offenders may utilise to commit crime ranging from personality traits and skills to tools or 'crime facilitators'. They note that some offences cannot be committed without tools (for example, graffiti without spray paint) and some 'tools' make offences easier to commit (for example, a jacket with a hood for street robbery, baggy clothes or clothes with lots of pockets for shop lifting, ladders for housebreaking). However, robust evaluations of initiatives to restrict access to tools of crime are not available.

Reducing provocation

Crowding has been shown to be related to levels of urban crime. Studies of bars and pubs show how their design and management plays an important role in generating violence or preventing it. For example, violence increases with size of venue, and where the clientele is dominated by young men, who do not know each other and cannot manoeuvre without jostling. Macintyre and Homel found that nightclub violence was reduced by floor plans that regulated traffic flow and minimized unnecessary jostling^[189].

Research also consistently points to a relationship between the presence of bars and prevalence of crime in the surrounding area. Homel and Clarke^[190] suggest that the behaviour of bartenders and bouncers may contribute to violence in these places. Evidence from the USA and Australia, summarised by Eck^[191], suggest that changing the management of drinking places (from staff training and changes in legal liability of bartenders) is a promising method for prevention of drinking-related offences.

In other contexts, segregating rivals supporters at sports events and avoiding congestion on the streets after events, helps to manage and reduce the risk of violence^[192]. There is also evidence (from the UK) that simply widening aisles at an open-air market can significant reduce in robberies^[193].

Factors associated with the effectiveness of situational crime prevention

Before moving on from the evidence on situational crime prevention, it is worth briefly summarising which factors the literature suggests are associated with success.

Poyner^[194] carried out a large evaluation of 122 crime prevention projects. Studies were selected if they were available in English, accessible at the time of review and reported on outcome effectiveness. Each study was given a numerical score depending on whether there was good evidence of a positive effect, some evidence of an effect, no evidence of an effect or whether the evidence showed that crime actually increased (a negative effect). There were a number of studies that showed that situational crime prevention was effective in reducing crime and a number that were inconclusive. Evaluations of access control, place managers (e.g. bus conductors, receptionists), target removal and physical barriers were generally positive. However, evaluations of steering-column locks, security campaigns, security surveys, preventative police patrol, CCTV, street lighting and property marking were generally negative or inconclusive.

An evaluation^[195] of 21 projects included in the Home Office Reducing Burglary Initiative sought to determine whether there was an association between the intensity of the programme and changes in burglary rates before and after the schemes were implemented. The interventions across the 21 projects ranged from location specific crime prevention (use of locks), publicity campaigns, youth diversion schemes and property marking. The study was based on pre- and post measures of burglary rates that were compared for different levels of intensity of the schemes. The outcomes for the scheme areas were also compared with the same outcomes measures for the police force area as a whole. The researchers made a distinction between input and output intensity. Inputs are described as including purchasing equipment (e.g. window locks) , project staffing and training. Outputs included, for example, the number of locks installed or hours worked by offenders under supervision. It was found that output intensity (the amount of crime prevention implemented, for example, by installing locks) was correlated with burglary reduction across the 21 projects. However, there was no correlation between input intensity (the amount of money spent on the schemes) and outcome in terms of burglary reduction. The same evaluation identified the most successful SBDs tended to be those that implemented location specific situational crime prevention measures, such as target hardening (e.g. fitting locks or alley-gates) to individual properties. In addition, the successful SBDs included stakeholder interventions (e.g. residents' associations, accredited tenants' schemes) that often involved some form of stand alone publicity campaign; and, those SBDs that publicised the interventions via newspaper articles and/or radio interviews.^[196] This suggests that publicity of crime prevention scheme is an important factor in the success of an intervention that includes a location-specific preventative measures.

Laycock^[197] summarises the results of an evaluation of a property marking scheme in South Wales conducted in 1984 and concludes not only that this form of target-hardening is effective in preventing burglary, but also that publicity can be a key factor in its success. The programme used three methods to achieve a reduction in burglary: publicity at the launch; door to door visits by police or special constables; and free property marking equipment and door or window stickers. In addition, the chief constable sent a letter to all residents informing them of the launch. In the 12 months before the launch 128 burglaries were reported to the police. In the 12 months after the launch 74 burglaries were reported (a 40% drop). Laycock also noted that the number of burglaries in the second 12 month period, after the launch, reduced further to 66 reported burglaries. The reductions in both the first and second year were higher than expected and this was attributed (at least in part) to the initial and continuing publicity surrounding the programme.

Research evaluating the use of police decoy vehicles in Stockton also suggested that the drop in vehicle crime was due to the publicity surrounding the initiative^[198]. In a large scale study of 21 domestic burglary prevention projects, approximately half of the schemes evaluated set up local stand-alone publicity campaigns and these were found to be the most successful in terms of burglary reductions. Furthermore, the study found evidence of a significant reduction in burglary in the three months that immediately preceded implementation as a consequence of pre-implementation publicity^[199] A number of studies^[200] note that news of a crime prevention initiative was found (based on interviews with offenders) to change the perceptions offenders had of the risk and effort required and that they adjust their behaviour accordingly.

Displacement and Diffusion of Benefits

Criticisms of situational crime preventative interventions argue that they "displace" crime to other times and places. Spatial displacement is the form of displacement most commonly recognised, however another five possible forms of displacement have been identified. The six forms of displacement are:

• Spatial (offenders switch from targets in one location to targets in another location)
• Temporal (offenders change the time at which they commit crimes)
• Target (offenders change from one type of target to another)
• Tactical (offenders switch from one form of crime to another)
• Offender (new offenders replace old offenders who have been removed from crime or have desisted from crime)
• Offence (offenders switch from one crime type to another).

Research evaluations show that crime displacement is never total or 1:1 (in other words, it is not the case that for every crime prevented in the intervention area, another one crime was displaced to another area). In an early review on displacement by Hesseling^[201] 55 studies of situational prevention programmes were examined. In 22 of the studies there was no evidence of displacement (and in six of these there was evidence of a diffusion of benefits), while in 33 there was some displacement but the evidence suggested that it was usually very limited. Rather, there is a diffusion of benefits as reductions of crime (or other improvements) are achieved in areas that are close to the crime prevention intervention, even though those areas were not actually targeted by the intervention itself. A review of 102 evaluations of situational crime prevention projects found that displacement was observed in 26% of the projects reviewed. However, where spatial displacement did occur it tended to be less than the treatment effect, suggesting that the intervention was still beneficial^[202].

APPROACHES THAT EXTEND BEYOND THE 'SITUATION'

Restrictions on access to weapons

Evidence suggests that weapons (particularly knives) are used in a high proportion of violent crimes in Scotland. The Scottish Crime and Justice Survey 2012/13^[203] estimates that weapons were used in 23% of all violent crimes. The most common type of weapon used was a knife - used in 41% of violent crimes. This was followed by a bottle (in 24% of violent crimes), sticks/clubs/hitting implement (in 24%), bricks/stones and screwdrivers/other stabbing implements (in 18%) and pistol/rifle or a shot gun (in 1% of violent crimes). Recent research evidence has therefore tended to focus on the issues of knife carrying and knife crime^[204].

The evidence also highlights the extent of knife carrying among young people in Scotland. Self-report data collected from the Edinburgh Study of Youth Transitions and Crime in the late 1990's and early 2000s revealed that 38% of 12 to 17 year olds (both male and female) report having carried a weapon at some point.^[205] Amongst those who reported carrying a knife or some other kind of weapon between the age of 12 and 17, only a small proportion were persistent offenders^[206]. The majority of young people who claim to have carried a weapon report that they do so as a means of self-defence^[207]. The use of non-bladed weapons, such as bottles, bricks and bats, is as common as the use of knives^[208].

A range of different approaches have been adopted to reduce knife carrying, notably: police 'stop and search' initiatives; knife amnesties; education and awareness-raising; and mandatory minimum prison sentences.

Stop and Search

Research with knife carriers suggests that the risk of being the subject of a police 'stop and search' was reported to be a strong factor in the decision not to carry a weapon among some young people, while other interviewees reported the limited impact or that they simply switched the type of weapon that they carried^[209]. The researchers note that interviewees who were almost 16 or older, were more aware, compared to younger interviewees, of the risks of being caught with a knife (police stop and searches and of the risk of a prison sentence) and as a result deterred from carrying a knife^[210].

An intensive police 'stop and search' campaign was a feature of the Strathclyde Police initiative 'Operation Blade'. A review of this initiative examined the effect on accident and emergency attendances and found a reduction in the number of serious stabbings for ten months during and after the intervention^[211]. However, the increased use of 'stop and search' was part of a wider programme of measures that included a knife amnesty (see below), increased police presence, training for bar stewards, and education/ awareness sessions in schools^[212]. It is therefore difficult to distinguish the effectiveness of the separate elements of this initiative.

Amnesties

Knife amnesties are defined periods during which individuals are encouraged to surrender knives to the police without being prosecuted for handling an offensive weapon. Although national and local knife amnesties have been implemented in England and Scotland, the evidence of impact is complex. As discussed above, the review of "Operation Blade" found that a knife amnesty was among a range of strategies that together were found to be associated with a reduction in serious stabbings^[213]. However, Eades and his colleagues, in their review of knife crime policy in England, specifically make the point that even if amnesties succeed in removing some knives, the volume of knives available in households, shops and other sources means that a replacement is easily acquired^[214]. They compare knife to gun amnesties that remove weapons that are significantly more controlled (e.g. with the requirement to obtain a licence and register guns).

Education campaigns

Education and awareness-raising campaigns targeting young people and children have been carried out both in England and Scotland. The 'No Knives Better Lives' initiative, was introduced in 2009 and has been rolled out to Local Authorities across Scotland on an opt-in basis^[215]. As with English campaigns, there is no robust evidence yet on the effectiveness of these campaigns in reducing knife carrying and knife crime^[216]. However, there is on-going evaluation of No Knives Better Lives. In recognition that delivery of No Knives Better Lives (NKBL) is set up differently to meet the needs of the local area, the NKBL delivery team in Youth Link Scotland have conducted small-scale evaluations of specific aspects of the programme including peer education to raise awareness of the risks of knife carrying. In addition, the SCCJR recently conducted a review of the evidence on the effectiveness of knife crime interventions which highlighted that education based interventions offer most promise in terms of tackling knife crime^[217].

Mandatory minimum sentences

Concerns over perceived levels of knife carrying and knife crime have been central to calls for mandatory minimum sentences for knife carrying and in England, the maximum sentence for carrying a knife in public without lawful reason was increased from two to four years under the Violent Reduction Act 2006. Similarly, in Scotland the introduction of the Criminal Justice (Scotland) Bill 2013, an amendment to the Criminal Justice (Scotland) Act 1995, intends to increase the maximum sentence for unlawful possession of a knife from four to five years. While the use of mandatory minimum sentences is more relevant to the section of the model that deals with deterrence, given that it has been used as part of a strategy for reducing knife carrying, it also deserves a mention here.

The evidence on the likely impact of mandatory minimum sentences is mixed. In England and Wales, the introduction in 2004 of the mandatory minimum sentence for possession of a firearm was not associated with anticipated reductions in firearm offences rates^[218]. In the U.S., mandatory sentences have proven ineffective in reducing serious crime or petty theft rates^[219]. There have been no large-scale evaluations of mandatory sentencing for knife carrying in the UK. However, findings from the monitoring of the Tackling Knives Action Programme (TKAP) activities introduced in England and Wales in June 2008, which included tougher sentences for carrying a knife, show some reductions in certain crimes. This included a reduction in recorded homicides, and "all violence" and robbery offences in the target age group (19 years and under)^[220]. However, as with the assessment of knife amnesties above, because the introduction of tougher sentences was part of a wider package of measures it is not possible to assess the extent to which mandatory sentences have been responsible for these reported reductions. The lack of statistically robust comparison areas also means these results should be interpreted with caution. It is however encouraging that the risk of attracting a prison sentence, if caught carrying or using a knife, was cited as a key reason for not carrying a knife among young people interviewed in Bannister et al's study^[221]. (Those who were interviewed were in or associated with a youth 'gang' and either carried, or had carried, a knife.)

Commentators cite a number of reasons why mandatory minimum sentences for knife carrying are may have limited impact on the weapon use. As has already been argued in the section on deterring crime and as the Halliday report, Making Punishments Work: A Review of the Sentencing Framework for England & Wales, noted 'It is the prospect of getting caught that has deterrence value, rather than alterations to the 'going rate' for severity of sentences'^[222]. Given its concealed nature, knife carrying is difficult to detect without concerted police effort (like 'stop and search). Tougher sentences and mandatory minima are therefore unlikely to produce deterrent effects unless accompanied by a perception that the risk of detection is high. Increased sentences, and specifically mandatory minima also rely on the assumption that offenders are rational and calculating individuals who weigh the associated costs and benefits before deciding to offend. However, Bannister et al's research has found that many young people, who either carried knives or had previously done so, were either unclear or incorrect as to the precise legal consequences of knife carrying/using. Finally, researchers have also noted that even if mandatory minima can have a deterrent effect on some individuals, there is a high likelihood of displacement to alternative types of weapons that are not covered by mandatory sentencing^[224].

Restrictions on access to alcohol

The link between alcohol and drug use and crime has already been established in the section of the paper that explored underlying causes of offending. This section considers the evidence on the impact of changes in the physical and economic availability of alcohol on crime rates and finds that there is very strong evidence that regulating the availability of alcohol can reduce the harm caused.

Evidence suggests that there is a strong link between the affordability of alcohol and consumption. The Policy Memorandum that accompanied the Alcohol Etc. (Scotland) Bill in November 2009^[225] and Business and Regulatory Impact Assessment (BRIA) for the Alcohol (Minimum Pricing) (Scotland) Bill^[226] states that there is strong evidence from numerous studies conducted in European countries, the United States of America, Canada, New Zealand and elsewhere, that levels of alcohol consumption in the population are closely linked to the retail price of alcohol. As alcohol becomes more affordable, consumption increases and so does alcohol-related harm. As the relative price increases, consumption goes down. Comprehensive research by the University of Sheffield that has been carried out for both the UK Government and the Scottish Government^[227] also suggests the introduction of Minimum Unit Pricing (MUP) for alcohol will reduce health harm, lead to a fall in crime and lead to considerable financial savings in the criminal justice system.

A European Union report^[228] that explores the evidence of the effectiveness of a range of policy interventions to reduce alcohol harm also concluded that raising and implementing a minimum age of purchase for alcohol, and reducing the availability of alcohol through restrictions on the number and density of outlets and the days and hours of sale all reduce alcohol related harm. More specifically, some of the findings were that:

• There is evidence that if opening hours for the sale of alcohol are extended then more violent harm results;
• A review of 132 studies published between 1960 and 1999 found very strong evidence that changes in minimum drinking age laws can have substantial effects on youth drinking and alcohol-related harm (particularly road traffic accidents), often for well after young people reached the legal drinking age; and
• Alcohol outlet density is, in general, positively associated with alcohol consumption and alcohol-related problems: the higher the density, the higher consumption and problems will likely be.

A World Health Organisation synthesis of the evidence on the effectiveness of interventions in alcohol control reaches the same conclusions^[229] - "There is substantial evidence showing that an increase in alcohol prices reduces consumption and the level of alcohol-related problems…In addition, stricter controls on the availability of alcohol, especially via a minimum legal purchasing age, government monopoly of retail sales, restrictions on sales times and regulations of the number of distribution outlets are effective interventions. Given the broad reach of all these measures, and the relatively low expense of implementing them, they all are highly cost-effective." It goes on to compare the effectiveness of these approaches to tackling alcohol misuse with approaches that focus on education and advertising and concludes "Various educational approaches have been developed to reduce alcohol consumption. Although they are growing in popularity, there is little evidence of their effectiveness. Similarly, current research findings only show limited effects both on advertising and advertising bans".

An evaluation of the 2005 Licensing Act (which came into full force in 2009) was undertaken as part of the MESAS programme^[230]. The evaluation identified examples of good practice among the license trade but also areas where enforcement measures could be improved. The report set out a series of recommendations for local and national partners.

Diversionary recreational opportunities

Another approach to restricting the ability of people to offend is to occupy their time constructively by providing recreational opportunities. Through programmes such as Cashback for Communities and Positive Futures, government has been investing in a range of positive recreational opportunities, focused principally on young people and taking a number of forms: spanning sports, physical activities, arts and culture. Recreational opportunities are thought to positively impact on offending levels by reducing boredom in young people and decreasing the amount of time available for unstructured and unsupervised leisure time. Evidence has shown that young people who spend more time 'hanging out' or who have 'nothing to do' are more likely to be involved in offending behaviour^[231]. Other studies^[232] have shown activities with low structure or little adult supervision to be a predictor of deviant behaviour.

While there is some evidence to support the positive outcomes of recreational activities, the evidence is not entirely straightforward. In an evaluation of youth diversionary projects funded by Glasgow Housing Association (GHA), several benefits to communities were identified by residents and stakeholders, including: reductions in reports of crime and antisocial behaviour; reductions in fire-setting; less gang activity; and increased accessibility of parks and open spaces for residents to use. In addition, participants (young people) reported a healthier lifestyle and reductions in drinking alcohol, though findings on their involvement in crime and antisocial behaviour were mixed^[233]. However, the research did not look at whether these benefits were sustained over time, or whether any identified reductions in crime and ASB were greater than in other areas where diversionary schemes were not operating. Smith and Waddington^[234] and Crabbe^[235] argue that it is unrealistic to assume that providing recreational opportunities will prevent young people from simultaneously committing crime. However, although there is a lack of systematic evidence of a causal relationship between participation in recreational activities and a reduction in offending, many large scale diversionary projects have demonstrated some success in reducing offending through the reduction of crime figures (i.e. Splash, Splash Extra, Positive Activities for Young People, Midnight Basketball). Overall though, it was not possible to account for the impact that other social processes and interventions may have had on crime rates. Indeed, several studies^[236] have concluded that there is a lack of credible empirical evidence on the impact of recreational activities on offending due to the short term funding associated to projects, the complexity of defining and measuring outcomes and a lack of understanding about the varied causes of criminality. As demonstrated by Crabbe et al^[237] in their evaluation of Positive Futures, recreational activities alone do not reduce offending. Rather, it is the personal and social outcomes that support such change that should be the focus of projects. It is suggested that recreational activities may be beneficial as they can successfully be used as a 'hook' upon which to layer a developmental approach which combines social support, positive role models and other protective factors.

CONCLUSIONS

There is sufficient evidence to support situational crime prevention approaches. The evidence suggests that making even simple changes to the environment can be effective in reducing crime (for example, improving street lighting, introducing exact fare requirements on public transport, reducing overcrowding in public venues and using time-locked cash boxes). Not only can we expect these approaches to impact on acquisitive crime but efforts to reduce overcrowding in public venues and surrounding areas suggest that manipulating the characteristics of the environment can also impact on violent crime. The evidence also suggests that concerns about the potential displacement effect are largely unfounded and that there is even evidence of a diffusion of benefits in some cases. Furthermore, the crime preventative potential (it seems) can be further enhanced where the intervention is accompanied by widespread publicity. Where this publicity can be disseminated to potential offenders it can impact on crime rates even before an intervention is implemented.

The potential that structural improvements might make to increasing community pride and collective efficacy should not be overlooked. This has been discussed elsewhere in this paper but it is useful to be reminded that situational crime prevention might affect changes in crime through a less direct route and that responding to community requests for intervention can have benefits that extend beyond those generated by, for example, the installation of a CCTV camera or an additional street light. The effect that situation crime prevention is likely to have on community cohesion should be a factor in the decision to invest in this type of intervention but we should also be mindful of concerns expressed in the literature about the potential negative impact of erecting physical barriers that segment communities rather than draw them together.

There is already an emphasis on 'designing out' crime through urban planning in the Secure by Design scheme and this has been shown to be effective. The findings from the evaluations of Secure by Design, street lighting and access and exit controls suggests that the evidence presented above has important implications for architects and planners (both urban and transport). The findings are therefore relevant to National Outcome 10 from the Scottish Government's National Performance Framework: We live in well-designed and sustainable places where we are able to access the amenities and services we need. Our definition of 'well-designed' should extend to seeking to minimise the opportunities for crime and there is clearly scope for joint working between justice and colleagues working to improve the built environment. There may also be a role for government to work with businesses to ensure that they are aware of measures that have been proven to help reduce crime and to help them to share good practice across organisations.

Other relevant approaches mentioned in this section of the model include restricting access to weapons and to drugs and alcohol. Given that by far the most common weapon used in Scotland is a knife, the relevant evidence has tended to focus on restricting access to these. Although it is difficult to differentiate between the impact of various strategies, broad strategies that focus on restricting access to knives and apply a range of approaches have been found to be associated with reductions in knife injuries. There is very strong evidence that regulating the availability of alcohol can reduce the harm caused. Policies that focus on increasing the price of alcohol, raising and implementing a minimum age of purchase, restricting the number and density of outlets and the restricting days and hours of sale have all been shown to be effective.

Finally, despite reviewing a broad range of material, it was not possible to find any documented robust evidence of a causal relationship between participation in diversionary recreational activities and a reduction in crime. This is largely because of the difficulty in controlling out the impact of other social processes and interventions to which participants are exposed. However, it has been suggested that although recreational activities alone may not reduce offending, they are beneficial in helping engage young people in positive activities which may lead to the provision of greater social support, positive role models and other protective factors. Overall however, the evidence strongly suggests that the key to having a sustained impact on crime is to address the underlying problems which drive young people to offend.