Identifying micro-spatial and temporal patterns of violent crime and disorder in the British city centre

Identifying micro-spatial and temporal patterns of violent crime and disorder in the British city centre

Applied Geography 21 (2001) 249–274 www.elsevier.com/locate/apgeog Identifying micro-spatial and temporal patterns of violent crime and disorder in t...

497KB Sizes 0 Downloads 102 Views

Applied Geography 21 (2001) 249–274 www.elsevier.com/locate/apgeog

Identifying micro-spatial and temporal patterns of violent crime and disorder in the British city centre A.L. Nelson, R.D.F. Bromley*, C.J. Thomas Department of Geography, University of Wales Swansea, Swansea SA2 8PP, UK Received 9 September 1999; received in revised form 9 October 2000; accepted 11 October 2000

Abstract Using variables recorded by the police on when and where violent crime and disorder occur, the paper emphasizes the value of using a combination of different types of micro-spatial information in identifying patterns of violence. Evidence from Cardiff and Worcester shows primary clusters at night in the pub/club leisure zones; and secondary clusters during the shopping day in major retail streets. Disorder data also reveal subsidiary afternoon clusters near licensed premises, and a late-night confluence flashpoint at a node of pedestrian activity. Police geographical information systems should incorporate precise spatial and temporal variables to enhance our understanding of violent crime, to facilitate targeted policing and to assist in creating safer city centres.  2001 Elsevier Science Ltd. All rights reserved. Keywords: Alcohol consumption; City centre; Fear of crime; GIS; Police records; Violent crime

Research on the micro-spatial geography and temporal characteristics of violent crime has been neglected in favour of identifying broader patterns. Most recent research has been undertaken at the macro-spatial scale, often comparing the number of violent crime incidents between different regions or cities of the UK, or simply in relation to longitudinal trends. As a result, the importance of establishing information systems that contain precise spatial and temporal variables has not been fully recognized. Using basic techniques derived from Crime Pattern Analysis, this paper seeks * Corresponding author. Fax.: +44-1792-295-955. E-mail address: [email protected] (R. Bromley). 0143-6228/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 3 - 6 2 2 8 ( 0 1 ) 0 0 0 0 8 - X

250

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

to demonstrate how particular variables relating to where and when violent crime occurs are valuable in identifying the geography of city-centre violent crime. A micro-spatial approach, combined with the interrelated dimension provided by temporal data, contributes to a knowledge of the precise patterning of violent crime and disorder, and this is becoming increasingly important for the police and planners. Violent crime and disorder are known to cluster in urban as opposed to rural areas (Mirrlees-Black, Budd, Partridge, & Mayhew, 1998), with the greatest concentration in city centres. Greater knowledge of the distribution of violent crime and disorder in the city centre offers inferential insights into the causal factors of crime and is an important aspect of risk assessment in responding to crime. The identification of spatial ‘hot spots’ and temporal concentrations enables the targeted deployment of police resources to prevent the generation of crime and disorder, by policing the ‘hot spots’ more effectively. Likewise, the identification and explanation of emerging patterns of crime enables more effective planning of the built and functional environment in order to reduce the susceptibility of certain locations to crime. At a time when a concern for the vitality and viability of town and city centres has been combined with a recognition of the negative implications of the potential for violent disorder, a greater understanding of such security issues is of particular significance. This is reflected in the Crime and Disorder Act 1998, which places responsibility upon the police, the local authority and other agencies to respond to community safety and crime issues through the creation of partnerships. These partnerships generally employ a community safety officer, undertake audits of crime and disorder, and advise on strategies to maintain the attractions of city centres. The paper begins by briefly examining some of the existing research findings and methodologies on violent crime, before using empirical data from Cardiff and Worcester to explore the spatial and temporal patterns of violent crime. The aims are not only to elucidate the micro-spatial and temporal characteristics of violent crime in the city centre, but also to emphasize the critical value of establishing detailed information systems to enhance future analysis.

Violent crime, fear and the city centre The 1997 British Crime Survey estimated that there were 3 381 000 violent crimes committed in 1997, representing approximately 20% of all crimes (Mirrlees-Black et al., 1998). Of these, only the minority classified as ‘domestic’ incidents (25%) are unlikely to be relevant to the current investigation. Such domestic incidents generally occur in the private space of the home, and the current paucity of residential accommodation in the city centre explains their lower level of occurrence in this part of the city. The majority of the remainder of violent crimes in 1997 were perpetrated by either ‘acquaintances’ (43%) or ‘strangers’ (20%). Many of these are likely to have occurred in town and city centres, since the earlier 1995 British Crime Survey indicated that 80% of acquaintance and stranger violence had taken place in a combination of street, public house, club and transport locations, primarily in the evening (50%) (Mirrlees-Black, Mayhew, & Percy, 1996). The remaining 12% of

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

251

violent crimes, classed as ‘mugging’ motivated by theft, are also, in part, likely to involve city-centre locations. Most violence in the city centre thus occurs in the public space of the street or the semi-private spaces of leisure and other commercial facilities. Victimization from violent crime has considerable psychological as well as economic effects. In Britain, a significant minority of around 10% of city-centre shoppers are fearful for their personal safety (Nottingham Safer Cities Project, 1990), while more substantial minorities express anxieties concerning fears of being robbed (c. 20%), or of threats posed by groups of youths and vagrants (c. 40%) (Thomas & Bromley, 1996). At the same time, there has been a steady growth of public houses and night clubs in city centres, increasingly catering for an evening and late-night ‘youth market’ (Oc & Tiesdell, 1997). The growth of what is often stereotyped as a ‘youth culture’, with its tendency towards male domination, along with its association with heavy drinking, drugs and regularly reported late-night incidents of violence (Potter, Sacks, Kresnaw, & Mercy, 1994; Tomsen, 1997), have together served to generate anxiety amongst a substantial proportion of the population and to reduce the attraction of the night-time city centre for many of the middle-aged and elderly (Doeksen, 1997; Bromley, Thomas, & Millie, 2000). A third of respondents in the 1995 British Crime Survey, for example, indicated that they were fearful for their safety when walking alone after dark (Mirrlees-Black et al., 1996). This rose to 47% for women, increasing steadily with age so that 60% of women aged 60 or over expressed anxiety in this situation. Similarly, a study of Swansea, for example, indicates that 62% of the residents rarely visit the city centre in the evening, while 58% considered the city centre to be dangerous in the evening (Bromley et al., 2000). Fear and anxiety associated with actual or potential violence in the contemporary city centre can detract from its vitality and viability (Department of the Environment, 1994) at a time when its commercial success is being threatened by decentralization (Thomas & Bromley, 2000). Avoidance strategies adopted to reduce exposure to risk will have economic impacts upon the vitality and viability of a city centre and further increase social exclusion. Violent crime is, therefore, likely to have both social and economic impacts on the city centre and its users. A survey of Nottingham, for example, suggested that perceptions of danger in the city centre were costing £24 million annually in lost revenue and over 600 jobs (Nottingham Safer Cities Project, 1990). On a nationwide scale, a Home Office Working Party in 1988 indicated that compensation for victims of violent crime amounted to £52 million, while the administration of the Criminal Injuries Compensation Board cost a further £7 million (Taylor, 1994). These costs, however, were only a small proportion of the estimated £5000 million associated with the wider apparatus of the overall ‘criminal justice system’, including policing, social work and the courts. Even so, only a small proportion of violent crime results in criminal convictions. In 1995, for example, of the approximately 5 million ‘reported’ offences, only 301 000 convictions were obtained for violent crimes, with a further 203 000 resulting in ‘cautions’ (Maguire, 1997). The cost of violent crime in the USA is much higher. Estimates of direct physical and psychological costs in 1989 reached $10 billion (Miller, Cohen & Rossman, 1993). In addition, the associated costs of loss of productivity were estimated at $23

252

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

billion, and the economic value of the loss of quality of life, as experienced by the victims, at $145 billion (Miller et al., 1993). The social and economic impacts of violent crime within society are therefore substantial, not least because violent crime is a dimension of crime about which individuals feel considerable fear. Despite its significance, research into the spatial and especially the temporal patterns of violent crime is rather limited. Data sources typically used for studies of violent crime are victimization surveys such as the British Crime Survey, local surveys, information from hospital accident and emergency departments (Shepherd, 1990), and police statistics. Davidson (1989), for example, used crime survey data to link violent incidents to certain environments. These data only enabled him to relate incidents to the type of space rather than to specific places; for example, to a public house, rather than to a particular city-centre or other public house. Police crime data, however, can provide more precise information, but only on those incidents considered as crimes. These police data offer some consistency in that all police forces follow the carefully specified Home Office Counting Rules on what constitutes a violent crime. Nevertheless, flaws in police crime data have been acknowledged, principally due to the under-recording of violence that occurs (Davidson, 1989; Shepherd, 1990, 1994; Levi, 1997; Maguire, 1997). This underreporting may vary across socio-demographic groups, context of crime or environmental setting (Hough & Lewis, 1989), and it is important to recognize the many and complex ways in which crime data are socially constructed. Using data from the 1984 British Crime Survey, Hough and Lewis (1989) estimated that around 60% of woundings were reported to the police but only 31% of common assaults. Under-reporting, however, is not the only problem. Farrington and Dowds (1985), on the basis of their research in Nottingham, suggest that police recording practices influence both the level and the distribution of violent crime. Similarly, Levi (1997), referring to the work of Cohen (1981), indicates the manner in which public awareness of particular crimes can be amplified by campaigning groups and newspapers to create ‘moral panics’. This sometimes initiates legislation against such crimes. More commonly, however, the pressure of public opinion results in the targeting of police effort against crimes such as muggings, child molesting, racial harassment and car theft. This can inflate the incidence of these crimes in the recorded crime statistics. Clearly, there are problems when using police data to measure the extent of different types of violent crime and disorder, but within particular crime groups the data are seen as useful for exploring patterns and characteristics (Hough & Lewis, 1989). The police data, whatever their deficiencies, are undoubtedly the most comprehensive and reliable data source available. Since April 1998, other ‘public disorder’ incidents, such as breach of the peace or drunkenness, have been recorded on police reporting systems, providing a fuller picture of disorder and violence, especially in the city centre. In addition to crime data, this investigation therefore uses information on calls for police assistance, and other disorder incidents not recorded as crimes. Consequently, British official data sources are moving slowly in the direction of providing a more comprehensive coverage of crime in the environment. This is somewhat akin to the experimental study of Sherman, Gartin and Buerger (1989) in Min-

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

253

neapolis, USA, in which all calls for police assistance over a period of a year were recorded by street address or road intersection. Clearly, such data are likely to provide a more comprehensive picture of the ‘crimogenic’ nature of places than is available for most British cities; although even in this case the under- or over-reporting of many types of crime may well occur in an unpredictable fashion. Thus, while the quality of British crime statistics is improving steadily, it is important to stress that due to a combination of the ‘vagaries’ of police reporting practices and the ‘socially constructed’ definition of the nature of most crimes, some bias continues to be an inherent problem. Exploring patterns The analysis of patterns of crime in space and time draws upon the concepts developed in the field of environmental criminology. This approach seeks to explore the emerging patterns of crime by focusing on the nature of the environment in which crimes occur. It examines where and when the incident occurs, rather than concentrating on the offender and associated motivation (Bottoms & Wiles, 1997). The environment, both functional and physical, is viewed as a key factor in the generation of crime and disorder. Within this broad analytical framework, the more specific methodology and data collection techniques of Crime Pattern Analysis have also been developed to explore patterns and trends in crime, in recent years sometimes using Geographical Information Systems (GIS) (Ekblom, 1988; Hirschfield, Brown, & Todd, 1995; Read & Oldfield, 1995; Bowers & Hirschfield, 1999). Incidents of crime and fear of crime tend to concentrate in specific locations at particular time periods (Brantingham & Brantingham, 1993). These ‘hot spots’ (Sherman et al., 1989) of violent crime may not necessarily coincide with ‘hot spots’ of fear of personal victimization. Potential threat of victimization from violent crime shapes people’s perceptions of risk and subsequent behaviour more than any other potential form of victimization. Incidents of violence in the city centre, regardless of place or time, will affect perceptions of insecurity, as the concept of ‘risk’ becomes attached to the locality. Fear is often generated through the presence of visual symbols of criminality, such as social or physical incivilities (Herbert, 1993), and/or a knowledge of the actual crime problem within an area. In the context of the current investigation, the concepts introduced in the ‘routine activities’ approach in the pioneering work of Cohen and Felson (1979) are particularly appropriate. Crime events were characterized as the result of the non-random convergence of motivated offenders and suitable targets and the absence of capable guardians against crime. Consequently, crime incidents display the distinct patterns in space and time that are subsequently characterized as ‘hot spots’. Poyner (1983), for example, investigating a sample of 552 criminal attacks in central Birmingham, noted that pedestrian subways were the location of 13% of such incidents, despite the fact that they accounted for a negligible proportion of total land use. The groundbreaking study of the spatial concentration of crime incidents in Minneapolis by Sherman et al. (1989) investigated the location of 322 976 calls to the police for assistance over a year at the micro-spatial scale of individual addresses or the nearest

254

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

road intersection. Nearly 50% of calls related to only 3% of all the locations, while rapes were confined to a mere 1.2% of locations. In fact, spatial concentration into a relatively small number of ‘hot spots’ was the most common feature of all categories of crime recorded, to such a degree that it was considered possible to anticipate the ‘crimogenic’ nature of places. In a similar vein, work by Brantingham and Brantingham (1993) has analysed the manner in which crime clusters and fear generators are found to focus upon particular ‘nodes, paths and edges’ in the urban environment at specific times, often associated with ‘high-activity areas’. Similar concepts underlie the generation of ‘hot spots’ of fear at the micro-spatial scale. In the context of a university campus, the importance of ‘blocked prospects’ (which reduced visibility), points of ‘concealment’ for potential assailants, and ‘bounded space’ (which reduce the possibility of escape) were all found to be important environmental fear generators (Nasar & Fisher, 1993; Fisher & Nasar, 1995). Of particular relevance to the current investigation, however, is Hope’s (1985) exploratory analysis of 1463 incidents of disorder recorded in the city centre of Newcastle-upon-Tyne on Friday and Saturday nights over a period of 18 months. Like the findings of Sherman et al. (1989), the great majority of incidents were concentrated in a small number of ‘high-activity’ areas. Four streets on which 15% of public houses were located were the scene of 42% of incidents. Additional concentrations were associated with two large dance halls and the Metro underground stations. Temporal concentration was also a marked feature of the pattern, with 48% of the incidents confined to the 30 minutes before and 1 hour after public house closing time. Hope (1985) suggested the explanatory significance of a combination of the emergence of a new youth culture since the 1950s; situational factors associated with the arrangement of space within public houses and nearby public spaces; urban design characteristics that have been inappropriate to the activities of the contemporary night-time city centre; and management and marketing policies instituted by the licensed trade. This analysis provided an early basis for policies designed to implement crime prevention measures in the night-time city centre and continues to suggest avenues for further investigation. In effect, despite the somewhat environmentally deterministic orientation of much environmental criminology and crime pattern analysis, it has increasingly been recognized that the micro-spatial aspects of place and setting can contribute significantly to an understanding of the spatial and temporal patterns of crime and fear in the urban environment. Alcohol consumption and violent crime Crime surveys have highlighted the links between evening leisure behaviour, alcohol and violent crime, in terms of time periods and general locations (Shepherd & Brickley, 1996). Shepherd (1990: 298) found clusters of incidents on Friday and Saturday nights between 22.00 and 03.00. This has been confirmed by the various sweeps of the British Crime Survey, which state that violent crimes are most prominent during weekends, with half the incidents occurring between 18.00 on a Friday and 06.00 on a Monday (Mayhew, Aye Maung, & Mirrlees-Black, 1993: 88;

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

255

Mirrlees-Black et al., 1998). However, the broad pattern of association with evening leisure zones may hide subsidiary or specific hot spots. A majority of incidents of violence are considered to involve drinking and drugs. The 1995 British Crime Survey indicated that in 53% of cases of violent crimes committed by ‘strangers’, and 45% of violent crimes by ‘acquaintances’, the offenders were under the influence of alcohol (Mirrlees-Black et al., 1996). For such incidents involving drugs the respective figures were 11% and 22%, while for cases involving alcohol or drugs, figures of 57% and 53% were recorded. In fact, a growing literature underlines the strong association between violent crime and leisure use of the city centre (Shepherd, 1994; Mirrlees-Black et al., 1998). In addition, the number of violent offences associated with the excess consumption of alcohol has been shown to be increasing (Engs & Hanson, 1994). Shepherd (1994), however, stated that almost all of the evidence supporting the link between alcohol consumption and violent crime was in the form of aggregate data. Spatially, he demonstrated that alcohol-induced violent crime typically occurred in and around bars, adjacent fast food outlets and in taxi ranks. This is consistent with the contention that most cases of street violence occur where young people cluster and become embroiled in disputes when they leave drinking premises (Levi, 1997). The role of alcohol consumption as a factor in violent crime nevertheless continues to be a subject of debate (Tomsen, 1997). Collective drinking and associated behaviour have been identified as creating a volatile setting which is pre-disposed to violence. Tomsen (1997: 94) points to culture, symbolism and emasculating leisure behaviour in the generation of violence. Tuck (1989: 66) suggests that there is a ‘complex amalgam of reasons’ that influence where disorder occurs. The cause of violence may be seen as ‘machismo’, rather than the result of ‘any obvious political, racial or gang motivations’ (Tuck, 1989: 45). Victims of crime Although the emphasis of this paper is on the variables that explain the spatial and temporal patterns of violent crime, these patterns cannot be divorced from the people who are the victims and perpetrators of such crime. The people involved in violent crime clearly reflect the leisure-based late-night context in which the majority of such crimes occur. Violent crime primarily affects the young, particularly young men. Early evidence indicated that the young were the most frequent victims of violent crime, and this high risk was associated with their way of life (Sparks, Genn, & Dodds, 1977). Gottfredson (1984: 7), investigating victimization from data obtained from the 1982 British Crime Survey, found that five times as many people in their teens report victimization than people in their 40s. This suggests a distinct pattern of victimization on the basis of age. Again, Gottfredson (1984) associated the high levels of risk amongst the young with their lifestyles. He based this notion on the fact that the young frequent types of facilities where they are exposed to greatest risk. More recent evidence from successive British Crime Surveys indicates that the pattern persists. Overall victimization rates from violent crimes are nearly twice as high for

256

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

men (6.1%) as for women (3.6%), while the differentials are markedly higher for those aged 16–24 (20.9% for men against 8.8% for women) (Mirrlees-Black et al., 1998). By contrast, the rates fall dramatically for both men (7%) and women (4.6%) aged 25–44. Youthfulness, particularly in combination with gender, are important determinants of risk of violent crime. It is also notable that the pattern of victimization is reflected closely in the characteristics of perpetrators. Offenders involved in violent crime were recorded as predominantly male (82%) in the 1995 British Crime Survey (Mirrlees-Black et al., 1996), a figure that is closely paralleled by those sentenced or cautioned for such crimes (82%) (Maguire, 1997). Of those brought to court, 45% were under 21 years, with a peak at 18 years old. Most of these were poorly educated individuals from lower-status groups or were defined as behaviourally disturbed (Maguire, 1997). However, despite the evidence that the victims of violent crime are mainly young and male, as indicated earlier, women and the elderly tend to be the most fearful. Women and the elderly consider themselves, and are perceived by society, as being at greatest risk from violent crime (Mirrlees-Black et al., 1996; Bromley et al., 2000; Thomas & Bromley, 2000). Thus, those most fearful of potential victimization are generally the groups who are statistically at least risk (Mirrlees-Black et al., 1998). Research on this apparent paradox, however, has yet to provide a definitive explanation (Nasar & Fisher, 1993; Young, 1997). The low levels of victimization of these groups, for example, may well reflect higher levels of conflict-avoidance tactics associated with their physical vulnerability, rather than being paradoxical. It has also been suggested that tolerance of violence and threat is lowest amongst women and the elderly relative to the ‘machismo’ values held by many young men (Young, 1997). Whatever the explanation, however, the resulting levels of fear are associated with significant constraints on the ‘normal’ patterns of behaviour of women and the elderly in contemporary society. Violent crime and associated fear amongst wide sections of the community is an important issue in contemporary British society. This is significant for the constraints that it imposes upon the economic and cultural life of city centres, particularly for evening and night-time functions. The issue has distinctive spatial and temporal dimensions related to hot spots of violence and fear. These appear to be generated by the current dominance of evening and night-time activity patterns by the relatively youthful patrons of public houses and clubs. The previous sections have reviewed an extensive literature drawn from a number of interrelated disciplines, which aim to describe and explain the incidence of violent crime and fear in the city centre. This provides the broad context for the current investigation, which seeks to develop and explore the spatial, temporal and policy issues associated with violent crime in two British city centres.

Researching violent crime and disorder Research on violent crime is directly constrained by the quality and detail of the police records on which it inevitably relies. The weaknesses of police data have

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

257

already been acknowledged, but their value for exploring patterns and characteristics has to be recognized. In addition to police crime data, this paper also uses incident data relating to calls for assistance from the public, providing a dataset on disorder incidents that are not recorded as crimes. Analysis of the crime and disorder incidents datasets was supplemented with more qualitative information from interviews with key police personnel, including the Licensing Sergeant and city-centre Beat Officer. In this paper, the variables selected as key to the research relate to three inter-related categories of information: the typological definition of the location of crime; the spatial definition of the location of crime; and, lastly, the temporal definition of the occurrence of crime. Typological definition Type of location relates to the function of the specific place where the crime is recorded as occurring and includes street, car park, public house, night club or shop. When applied to a small area, the function is often time specific, as in the case of a shopping street that becomes a night-time leisure zone owing to the land-use mix of shops, public houses and clubs and their different hours of operation. GIS often lack such location detail, which would assist in exploring the various factors involved in violent crime. While the spatial definition of the crime data is of paramount importance in the use of crime pattern analysis, the availability of information relating to time-specific functions can be of equal importance. Spatial definition Traditionally, crime analysis is undertaken at police beat or parish level. This provides only large area data on the crime problem, offering no indication of microspatial hot spots. However, police records almost always specify a small area, such as a street or car park, where the crime occurred, and in many cases they provide point information, at the level of the particular address. Point data, rather than area data, are clearly the ideal for identifying crime patterns and crime trends. The extent to which police records diverge from this ideal, however, is explained by briefly describing the information systems currently in place. The use of GIS is becoming more evident and police forces generally claim that they geo-reference all crime data, thus facilitating a spatial analysis of crime patterns. Grid references accurate to 100-m resolution, based upon 1:2500 (c. 25⬙ to a mile) or 1:1250 (c. 50⬙ to a mile) maps, are widely used by the police to locate crimes on a scale roughly equivalent to a street address. This potentially provides the researcher with point data. However, the reality of this scale of spatial definition is more akin to the small area rather than to precise points. Police forces, perhaps using a gazetteer, will commonly use a single 100-m grid reference for an entire street. While the specific address will also be recorded for a burglary or shoplifting offence, violent crime outside the home is likely to be recorded only by the selected grid reference. Alternatively, police forces may use a Post Office grid reference system based on postcodes. In the city centre, each large store is likely to have its own postcode and related

258

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

grid reference, but smaller establishments will be clustered together into a single postcode/grid reference. Other areas, such as car parks or open squares, will have no postcode at all, although such locations can be easily grid referenced. In effect, the potentially precise geo-referencing system used by many police forces actually provides small area spatial definition, rather than the ideal of point data at the city-centre scale. It is clearly possible, given the availability of sufficient clerical resources, to locate crimes at a more precise 1-m scale of spatial resolution using existing Ordnance Survey maps (Denham, 1993). However, such data are currently not routinely available, except in the exceptional situations where a purposebuilt GIS has been established for experimental purposes (e.g. Hirschfield et al., 1995; Bowers & Hirschfield, 1999). Temporal definition The recording of the time of occurrence of violent crime is probably the most reliable and consistent of the three key categories of information. Police records note the date, day of the week and approximate time of each violent incident. Such information is input into the computer information systems, facilitating an exploration of crime by the different hours of the day. Empirical research in Cardiff and Worcester The research focuses on British city centres, which during the day attract large numbers of shoppers. Most city centres have also developed significant concentrations of evening and late-night leisure and entertainment functions, which have increasingly become the preserve of a vibrant youth culture (Bromley et al., 2000). Both day and night-time functions create ‘high-activity’ areas that have the potential to generate violence and disorder. This article uses a range of variables to explore the spatial and temporal aspects of violence in the city centre throughout the 24hour day. For this purpose, 12 months’ data on violent crime were collected for the city centres of Cardiff and Worcester, abstracted respectively from the records of the South Wales and the West Mercia constabularies. Despite occupying different positions in the urban hierarchy, both cities have a similar array of urban functions. Cardiff and Worcester were chosen to represent the major regional city centre and a typical smaller centre of minor regional status. Cardiff, for example, has an immediate urban population of 279 055 (Office of Population Censuses and Surveys, 1994) but also serves as the regional shopping and entertainment centre for almost an additional population of 1 million living throughout South Wales east of Swansea. It is, consequently, the eighth largest shopping centre in Great Britain, reflected in a multiple branch score shopping index of 115 (Hillier Parker Research, 1996). It also has a significant concentration of leisure and entertainment activities on the southern periphery of St Mary Street, Caroline Street and Mill Lane (Fig. 1). A secondary concentration occurs on Queen Street, while a number of additional venues are spread throughout the rest of the city centre. Worcester has a smaller urban population of 81 538 in 1991 (Office of Population Censuses and Surveys, 1994),

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

Fig. 1.

259

Violent crime ‘hot spots’ in Cardiff city centre, 1993.

but also serves the 187 880 residents of the adjacent rural hinterland of the Malvern Hills and Wychavon districts. This is reflected in its position as 49th of the 172 regional shopping centres in Great Britain, with a multiple branch score shopping index of 74 (Hillier Parker Research, 1996). Significant concentrations of leisure and entertainment facilities are found in the Angel Place–Angel Street area immediately to the north of the main shopping concentration and in a secondary concentration to the east of the city centre (Fig. 2). Defining violent crime and disorder In the context of this paper, violent crime covers a variety of offences. These are defined by the Home Office as murder, attempted murder, wounding and assault, rape and indecent assault, and robbery; crimes that are essentially ‘personal’. All the Cardiff data are classified in this way, and include the crime of ‘common assault’, even though this offence was not part of the official Home Office violent crime

260

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

Fig. 2. Functional areas in Worcester city centre.

category until April 1998. The post-April-1998 category of violent crime also includes ‘harassment’. Thus, the Worcester data, for the period since April 1998, relate to the extended violent crime definition and include a small number of cases of harassment. With respect to other violent crimes, the Cardiff and Worcester data are similarly defined. Comparability between the police force areas is ensured as far as possible through the Home Office Counting Rules for crimes. In addition, the more recent Worcester dataset has the advantage of including disorder incidents, many of which are recorded as a result of information supplied

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

261

by telephone from members of the public. These include disorder in the street and public places, drunkenness, breach of the peace and civil disputes. Disorder can be classified as a public disorder crime if it meets the Home Office criteria specified in the Public Order Act for a recordable offence. Obviously the boundaries between the different categories are blurred rather than precise. A street scuffle may be recorded as a disorder incident; more serious fighting would be a public disorder crime; while fighting involving an assault would be a violent crime. The majority of disorder incidents reported, however, will never be recorded as crimes and can be viewed as an indicator of less serious violence. For Cardiff, information on violent crimes covering the whole of 1993 was abstracted directly from 6958 detailed ‘paper’ records for Cardiff Central in that year and then coded into a computerized dataset for analysis. Although it is acknowledged that under-recording occurs, 12 months’ data are likely to reflect the actual spatial and temporal patterns of violence, even if on a reduced scale. In Cardiff city centre during 1993, 409 incidents of violent crime were recorded by the police. This accounted for 6% of the total number of crime incidents for the city centre in 1993. The Worcester city-centre dataset for the whole of 1998 consisted not only of computerized geo-referenced crime data, but also of similarly computerized disorder incident data. The dataset had disadvantages in that the computerized data lacked the detail of micro-spatial characteristics that could be obtained from the handwritten records in Cardiff, but offered advantages in ease of use and the fact that the more recent dataset covers disorder incidents as well as crimes. West Mercia police utilize a Post Office grid reference system based on postcodes, purchased in 1993. Violent crime incidents recorded for a street are usually geo-referenced to a single grid reference, which acts as a cluster site for the whole street, whatever its length. However, the availability of the disorder incident data adds a new and wider dimension to the exploration of violence in the city centre, enabling the micro-spatial and temporal pattern of disorder to be identified. In Worcester, during 1998, there were 231 recorded incidents of violent crime, and over 2000 disorder incidents. Violent crime represented 7.4% of all crime occurring in the city centre during the year. Micro-spatial patterns of violent crime in the city centre The empirical research reported in this paper demonstrates the value of the different typological, spatial and temporal attributes in elucidating the pattern of crime in the city centre. More precise problem identification and exploration is clearly the first step in formulating a strategic response. This section looks specifically at the type of location for violent crime and then at the spatial locations in terms of a small-area definition in order to explore the insights these data offer into the pattern of violence. Type of location The type of location of the incident is considered in order to identify a typology of places, on the basis of function, where risk from violent crime is greatest. How-

262

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

ever, analysis of violent crime by the type of location was only possible through the manual coding of the Cardiff dataset, because the computerized Worcester dataset either omitted the information or recorded only a limited typology of locations. Examination of the Cardiff dataset suggests four aspects of the distribution of violent crime. First, it is clear that the majority (52%) of recorded violent crimes in the city centre take place in the public arena of the street (Table 1). Secondly, it is evident that the semi-private spaces of night clubs, public houses and other licensed premises accounted for a further 18% of incident locations. A third point is less predictable, and reveals that shops are prominent locations for the incidence of violent crimes in Cardiff. During 1993, 9% (37 incidents) of violent crime occurred in shops. Further investigation of such incidents showed that the act of violence generally occurred during the apprehension of a shoplifter, whereby the apprehending security guard became a victim of assault. Investigation of the modus operandi of violent crime and disorder incidents for Worcester indicated a comparable problem of shop-based disorder and violence, usually generated during the arrest of a shoplifter. Such evidence suggests the social dimension of crime reporting, in that incidents involving a security guard are probably more likely to involve police assistance than incidents between other members of the public. A fourth and final point of note from the classification is the unimportance of car parks as locations for violent crime (Bromley & Thomas, 1997). This observation is based on analysis of the Cardiff data, where the handwritten police records almost invariably noted the street, building, car park or other location of crime commission. Cardiff has a cluster of large city-centre multi-storey car parks, and such car parks are known to generate fear, particularly after dark, and particularly by women users (Nelson, 1997; Thomas & Bromley, 2000). Nevertheless, only 9 of the 409 recorded violent crimes in Cardiff city centre occurred in car parks, highlighting the apparent asymmetry between the location of violent crime and the places where fear of personal victimization is greatest. Within Worcester, evidence from the police data was confirmed by the managers and security staff of the three multi-storey car parks, who

Table 1 Location of violent crime, Cardiff city centre, 1993 Location Street Nightclub Public house/licensed premises Shop Car park Cafe/Take-away Bus stop Other Total

No. of incidents 211 37 37 37 10 9 9 59 409

Source: Authors’ analysis of police records, South Wales Constabulary

% of total 51.6 9.0 9.0 9.0 2.4 2.2 2.2 14.4

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

263

were not aware of any incidents of violence occurring in their car parks (Worcester Community Safety Vehicle Crime Strategy Group Meeting: June 2000). Small area location As the ideal of point data was not widely available from the police records, the small area location of incidents of violent crime was used to identify hot spots of violent crime. In this context the small area was usually a street or a segment of a street. Fig. 1 indicates the streets and other specific locations in Cardiff with the highest concentrations of crime. Whether the incidents took place on the street or inside buildings, an examination of the specific location of violent crimes reveals that their overall distribution clearly reflects the location of public houses and night clubs within the city centre. It is not possible from the crime records to determine whether fights in the street were initiated within the pub/club and then spilled out onto the street, or if the incident began on the street. Nevertheless, a clustering of public houses and clubs appears to increase the likelihood of street-based violence, whether these are categorized as general assaults or associated with robbery. This relationship is widely recognized, if rarely quantified, and was demonstrated in the earlier investigation at Newcastle-upon-Tyne (Hope, 1985) and, more recently, for Southport (Bowers & Hirschfield, 1999). Data for Cardiff reveal that the location with the highest number of recorded violent crimes in 1993 was St Mary Street, on the southwestern periphery of the retail core. There are 21 licensed premises on St Mary Street, representing the highest concentration of such establishments in the city centre (Crime Concern, 1992). The high level of recorded violent crime in this location emphasizes the relationship between alcohol consumption and violent crime widely noted in the literature discussed earlier. Moreover, public house and club areas, such as St Mary Street, represent the types of location that attract an evening trade comprising both locals and non-locals. Millward (1988: 5) suggested that as people come into the city from outside, the amount of violence escalates. Such an influx of people will also display temporal characteristics, reflecting the evening leisure use of the city centre. Another minor hot spot of violence in Cardiff city centre is the Caroline Street– Mill Lane area. During 1993, 14 violent crimes were recorded here, representing less than a quarter of those in St Mary Street, even though both streets have a concentration of public houses and clubs. In this particular case, however, supplementary evidence suggests that here the relationship between recorded violent crime and actual violence is somewhat atypical and that the number of recorded crime incidents is unexpectedly low. Interviews with the Crime Prevention Officer (in April/May 1994), backed by evidence of public fear of the streets at night, suggests a generally high level of disorder, which would normally be reflected in a higher level of recorded crime. It is argued that violent crimes are rather lower than they might be in this location for two main reasons: the particular type of violence and the high police presence. First, because most of the violence in Caroline Street and Mill Lane involves fights between young men, an even lower proportion of violent incidents than usual is subsequently reported to the police. Secondly, because there is a parti-

264

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

cularly high police presence in these areas, especially on Friday and Saturday nights, the occurrence of violent crime is depressed. This effect has also been demonstrated by Brumm and Cloninger (1995), who showed how an increase in the number of police on the street reduces the commission of violent crime. However, in the context of Worcester, a month of targeted/high visibility policing was found to have little effect (⬍5%) on the recorded levels of violent crime and reported disorder within the evening leisure zone either during or after the operation. The Cardiff data reveal that Queen Street also emerges as a hot spot, with 63 violent crimes recorded for 1993. Queen Street is the primary retail street within the city, along which four enclosed shopping centres have been developed. It is pedestrianized and houses the main ‘high street’ chain stores. Unlike St Mary Street, it does not have a concentration of public houses and clubs, although a number of entertainment venues and restaurants are located along the street. When the type of location of the crime is linked with the specific location, it is evident that a substantial number of crimes in this area occur inside shops and are associated with the apprehension of shoplifters (Nelson, Bromley, & Thomas, 1996). In fact, much of the inshop violence recorded for Cardiff city centre occurs on Queen Street, which is consistent with the highest concentration of major stores in the city centre. Analysis of the Worcester dataset reveals similar patterns, and offers the additional advantage of providing information on public disorder crimes and incidents. The larger number of incidents enables a more precise pinpointing of minor hot spots of violence. Violent crime and disorder were clustered primarily in the evening leisure zone of the city, in and around Angel Place, Angel Street and The Cross (Table 2 Table 2 Violence and disorder in Worcester city centre, 1998 Streeta

Crimes

Violence Angel Place The Cross Deansway Angel Street Foregate Street St Mark’s Close Lowesmoor High Street Broad Street The Butts The Shambles Corn Market All Saints Parish

39 5 23 12 4 0 12 8 8 11 10 0 231

Disorder incidents

Combined total

170 92 56 67 82 87 62 58 44 32 24 30 1604

229 110 89 89 88 87 76 71 56 46 36 31 1954

Public disorder 20 13 10 10 2 0 2 5 4 3 2 1 119

Source: Authors’ analysis of police records, West Mercia Constabulary a For street locations see Fig. 2. All Saints Parish covers the city centre, including all the streets listed in the table.

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

265

and Fig. 2). In fact, it can be seen for both Cardiff and Worcester that there is a close association between the location of public houses and clubs and incidents of violence. However, the evening leisure zone in Worcester, around the Cornmarket area of the city, did not emerge as a violent crime or disorder hot spot. Although the functional precursors to violence did exist, it appears that other contributory factors were not present. This is most likely associated with social factors such as the slightly older clientele of this leisure zone, a point that highlights the importance of utilizing additional qualitative information derived from interviews. In addition, in both cities a significant number of incidents of violence occurred in the primary retail streets. In Worcester, the shop-related dimension of violence also emerged in the primary shopping areas, focusing on the High Street, The Shambles and Broad Street (Fig. 2). In effect, like Cardiff, the overall spatial pattern of violent crime and disorder was directly associated with the function of the locality. Temporal distribution of violent crime Analysis of temporal variables, particularly when linked with the micro-spatial evidence, further clarifies the pattern of violence and disorder in the city centre. In effect, the type or function of a location is in many cases specific to the day of the week or the time of day: a weekday shopping or business street becomes a leisure zone at weekends and at night. The temporal information facilitates, for example, the separation of shop-related violence from night-time leisure violence. The temporal pattern of the incidence of violent crime in Cardiff and Worcester city centre was considered in terms of the day and time of occurrence. This analysis reveals the specific days and the times of day when levels of violent crimes are at their greatest. Day of incident Violent crime in the city centre is most common at weekends. Saturdays experience 27% of the weekly total for Cardiff, while Fridays and Sundays also experience relatively high proportions (15.4% and 15.2%, respectively; see Table 3). Violent Table 3 Temporal incidence of violent crimes in Cardiff city centre (1993) and Worcester city centre (1998) (%) Day of incident

Cardiffa

Worcesterb

Sunday Monday Tuesday Wednesday Thursday Friday Saturday

15.2 8.1 6.8 13.0 14.2 15.4 26.9

16.5 10.4 11.7 7.8 12.6 17.2 23.8

Source: Authors’ analysis of police records, South Wales and Mercia constabularies. a n = 409. b n = 231.

266

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

crime on Fridays and Saturdays is consistent with the spatial link already established between violence and the frequenting of public houses and clubs. Some of the incidents on Friday nights actually take place during the early hours of Saturday mornings, and the same occurs on Saturday nights, with crimes extending into Sunday mornings. Chi-squared analysis of the Cardiff data confirms the significance of the uneven distribution of violent crime across the week, with the weekend period experiencing significantly more incidents than expected (Chi-squared = 73.87, df = 6, p⬎.001). The same pattern emerges for Worcester, with 24% of violent crime occurring on Saturdays. The evidence from both cities, therefore, confirms the prevalence of violent crime on Friday and Saturday nights (extending into the early hours of the subsequent day), and its association with leisure behaviour (Shepherd, 1990). The daily pattern of occurrence of crime is likely to vary between different locations, reflecting their different functions. A comparison between St Mary Street (an important evening/night leisure zone) and Queen Street (prime retail) in Cardiff shows that over the year Queen Street has far higher numbers of incidents occurring on a Saturday (20) than for any other day (⬍10) (Fig. 3). The high level of Saturday violence recorded for Queen Street is associated with increased levels of both shopping activity and shoplifting on Saturdays (Nelson et al., 1996). The low number of incidents recorded for Sundays suggests that there is little post-midnight violence on a Saturday night/Sunday morning. By contrast, St Mary Street experiences a high number of incidents on Fridays, Saturdays and Sundays. This is linked to behaviour resulting from the patronage of the late-opening public houses and clubs located on St Mary Street. The number of people frequenting St Mary Street on Friday and Saturday nights is greater than on other nights of the week.

Fig. 3.

Daily distribution of violent crimes in Queen Street and St Mary Street, Cardiff, 1993.

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

267

The analysis of the day of incidence of violent crimes demonstrates that there is a distinct weekly pattern, with Fridays and Saturdays emerging as the peak days. This further confirms the link, first, between violence and the frequenting of public houses and clubs, and secondly, between violence and the busiest shopping periods. There is a clear relationship between the incidence of crime and levels of usage of the city centre. For example, violence not only arises from behaviour associated with the frequenting of public houses and clubs, but is also connected with the high level of usage on certain days of the week. These patterns are clarified further by examining the precise temporal distribution of violent crime and disorder incidents. Timings of incidents Analysis of the timings of the incidents of violent crime in the city centre reveals a distinct concentration of incidents at night. For Cardiff, the period 23.00–04.59 experienced 170 incidents of violent crime (42% of the 24-hour total) (Fig. 4). The main daytime period of 09.00–16.59 witnessed only 29% of all recorded violent crime. Thus, just under a third of violent incidents occur during the shopping or business day. Both Cardiff and Worcester show a similar pattern of daytime violent crime, which rises to a peak around 17.00 and then falls off before the evening/night time concentration of crime (Fig. 4). Although the occurrence of violent crime is viewed as reflecting the ‘high activity’ levels associated with concentrations of people (Brantingham & Brantingham 1993), the ratio of crime to activity clearly varies across the 24-hour period. At night, the ratio of crime to activity will be far greater than it is during the day, when the number of crimes will be low in relation to the amount of daytime activity. This highlights the influence of the differing functions and associated behaviours for the different time periods. A more precise comparison of the timing of violence in Cardiff and Worcester city centres reveals interesting differences between them. For Cardiff, the incidence of violent crime extends further into the early hours of the morning. It reaches a peak between 02.00 and 03.00, followed by a dramatic decline after 03.00. This is obviously due to the fact that very few facilities are open beyond this time of night. By contrast, in Worcester, the peak occurs between 24.00 and 01.00, followed by a similarly dramatic decline between 01.00 and 03.00. As a major regional centre, Cardiff has a more extensive hinterland for its larger concentration of public houses and clubs, which tend to have longer opening hours. In Worcester, not only is the leisure zone smaller and closes earlier, but people come from less far afield. The smaller city has a much more confined catchment. This combination of characteristics explains the earlier reduction of violent incidents. Examination of the temporal patterns of violent crime combined with the spatial/functional information also helps to distinguish further between the daytime shop-based violent crime and the night-time leisure/alcohol-related crime. As previously noted, 9% of all Cardiff city-centre violent crime was recorded as occurring in shops. Such shop-related crime is obviously higher in the city’s prime retail area. Thus, a comparison between Queen Street, the main shopping street, and St Mary Street, with its concentration of public houses and clubs, should highlight the dif-

268

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

Fig. 4. Hourly incidence of violent crime in Cardiff city centre, 1993 (n = 398) and Worcester city centre, 1998 (n = 218).

fering violence patterns of these two hot spots. On Queen Street, the majority of the violence (48%) occurs during the shopping day (09.00–16.59). Only 25% of all violence recorded on Queen Street was for the late evening/early morning period (Fig. 5). By contrast, the majority of violent crimes on St Mary Street occurred during the period 23.00–04.59 (62%), reaffirming the association between violence and public houses and night clubs. This dichotomy in the temporal distribution of violent crime between the two streets is confirmed by significance testing. Chi-squared analysis indicates that the

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

Fig. 5.

269

Variations in timings of violent crimes for Queen Street and St Mary Street, Cardiff, 1993.

difference in the temporal distribution of violent crime between the two streets is highly significant (Chi-squared = 23.14, df = 2, p⬎.001). The timings of the incidents occurring on St Mary Street conformed to the general pattern described for city centres, being prevalent during the night-time period. By contrast, incidents of violent crime on Queen Street were primarily a day-time phenomenon. The temporal incidence of violent crime, therefore, emerges as being strongly determined by the function of the area. The identification of this day-time dimension of violent crime and disorder is facilitated by the micro-spatial scale of analysis at street level. The Worcester data for 1998 reaffirm the patterns identified for Cardiff. The data show that during the day, violent crime and disorder incidents are greatest in the primary retail area of the High Street (Fig. 6), and that, as in Cardiff, there is a clear element of daytime shop-related violence. At night, Angel Place emerges as a particular hot spot. Again, like Cardiff, late evening violence and disorder is clustered here, associated with night-time pub/club (alcohol)-related violence. Violence generated in this area is directly linked to people visiting and leaving night clubs while under the influence of alcohol or drugs. However, more interestingly, the wider array of violence and disorder data for Worcester enables a more sensitive exploration of the temporal differences between streets, and suggests two additional features of the pattern of violence and disorder in the city centre. First, the existence of a daytime/early-evening concentration of disorder associated with drunks or vagrants is revealed. Angel Place displayed this secondary daytime peak, considerably earlier than the primary night-time peak. This secondary

270

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

Fig. 6.

The times of disorder incidents by street, Worcester city centre, 1998.

peak may be linked with the high volume of people associated with the outdoor market, and the all-day opening of the public houses located around Angel Place. Fig. 6 shows the rise in incidents between 14.00 and 15.00 after the lunch hour, which then decline until they increase again after 19.00. St Marks Close, a residential street on the edge of the city centre, also emerged as problematic during the early evening. Here the rise in violence occurs after 18.00. Disorder at this location was associated with vagrants and drunks. This cluster is thought to be related to new alcohol by-laws in the city, which have led to the displacement of drunks and vagrants from the retail core. A second pronounced micro-spatial/temporal concentration of violence is also revealed by the Worcester data. A night-time confluence flashpoint associated with pedestrian flows occurs at The Cross around 23.00. This location contains a cluster of fast-food establishments on the main route between a number of public houses and the concentration of night clubs. Disorder at this site is again associated with its functional character, but the particular timing of the violence reflects its role as a route/intersection between various concentrations of evening leisure facilities in the city centre. The confluence flashpoint is, therefore, a particular type of hot spot, related as much to the ways in which pedestrian movement is channelled at night as to the functions of the immediate area. The explanatory value of the ‘routine activities’ approach of Cohen and Felson (1979) and Brantingham and Brantingham’s (1993) concepts of ‘nodes, paths and edges’ is relevant to this phenomenon.

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

271

Conclusions This paper has emphasized the value of using a combination of different types of information in identifying the patterns of violent crime in the city centre by analysing a range of variables recorded by the police, relating to where and when violent crime occurs in the city centre. While it is important to note that the generation of violent crime and disorder in the city centre appears to be the result of a complex interplay of factors, and cannot be explained by any single causal factor, there is no doubt that the micro-spatial and temporal analysis can assist our understanding of the spatial and temporal dimensions of the phenomenon. The evidence from Cardiff and Worcester city centres provides a comprehensive overview of the patterns and enables a number of conclusions to be drawn. Incidents of violence within the city centres of Cardiff and Worcester are associated with particular functions and locations. In both cities, the type or function of locations and the associated behaviour strongly influences the spatial and temporal characteristics of violent crime. Overall, the investigation has shown how two distinct clusters of violent crime and disorder emerge both spatially and temporally in the city centre. During the evening and night, both cities showed a concentration of incidents in and around the pub/club leisure zones. During the day, both cities also reveal a lesser clustering of incidents in the primary shopping streets. A more precise temporal analysis reveals that incidents occurred primarily on Fridays and Saturdays, days notable both for the highest levels of evening leisure activity and for the largest numbers of shoppers in the city centre. Violent crime is at its highest level particularly between 23.00 and 03.00, reaffirming its direct association with the behaviour associated with youthful visitors. Clearly, micro-spatial/temporal exploration offers useful insights into the patterns of crime that are obscured at broader spatial and temporal scales. For example, the detail available in the Worcester dataset revealed that in addition to the late-night peak of violence and disorder, there were secondary afternoon and early evening peaks. The data also revealed the existence of what is termed a confluence flashpoint of violence and disorder, reflecting the movement of pedestrians at critical times in the night-time functioning of the city centre. The availability of point data, not utilized in this research, would permit even more precise identification of the patterns down to a particular street corner or public house (Hirschfield et al., 1995; Bowers and Hirschfield, 1999). The paper has demonstrated the insights that can be derived from an awareness of micro-spatial and temporal patterns of violent crime and disorder, as opposed to a broad area analysis. Although such insights often require clarification on the basis of more qualitative information derived from interviews, the findings emphasize the importance of information systems that permit the exploration of such detailed geographical and temporal analysis. It is very important to establish GIS that facilitate analysis at precise spatial and temporal scales. The different categories of information, whether relating to the type or function of place or temporality, all need to be referenced to specific locations. Without this degree of precision in geo-referencing, a more detailed environmental approach to the understanding of violent crime and

272

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

disorder is impossible. Many current GIS obscure information through the generalized geo- and temporal referencing adopted, and fail to allow the identification of crime-and-disorder hot spots. The cost of the additional clerical effort associated with the provision of such data is likely to offer markedly more effective opportunities for the development of deeper insights from the analysis of patterns of crime for most types of criminal activity. Since the Crime and Disorder Act 1998, the interests of the police, local authorities and academic geographers in identifying and understanding the spatial and temporal patterns of crime and disorder have come closer together. Guidance from the Home Office exhorts the police and the local authorities to examine the causes or ‘generators’ of crime. Precise spatial and temporal information is a critical step in this process. In the context of the contemporary emphasis on the development of multiagency strategies to tackle problems of crime and disorder, it is vital that clear and comprehensive information is available to all participants, rather than simply relying on the expression of the accumulated experience of long-serving police officers. Precise information is essential for the targeting of police resources to prevent the further generation of violence and disorder at hot-spot locations. The provision of the precise kinds of micro-spatial and temporal data suggested in this article is likely to prove invaluable for the targeting of scarce police resources to specific locations at particular times of the day and night. It also has implications for the installation and management of CCTV cameras, which can be concentrated at the key locations and most carefully monitored at the most problematic times. In addition, precise information facilitates prediction in order to assist in the planning of safe city centres. For example, detailed geographical information is a basis for predicting the possible effects of longer opening hours for public houses and clubs and, hence, may inform licensing/planning applications. It helps planners and architects to be aware of the environmental factors that help to generate crime and disorder and, therefore, contributes towards the creation of improved city-centre environments where crime and disorder are at a minimum. Additional detailed analyses of patterns of late-night violence and disorder in a larger number of city centres, for example, might well provide insights into the relative advantages of the alternative planning strategies. In what circumstance and for what facilities, for example, is it better either to concentrate or to disperse the new leisure and cultural activities that continue to be developed in most major and medium-sized city centres? The need to balance the potentially conflicting ‘public order’ or ‘commercial’ perspectives would, of course, also need to be included in the equation. In effect, the value of precise spatial and temporal information in combating violent crime and disorder in the city centre cannot be overstated.

Acknowledgements The authors wish to acknowledge the invaluable cooperation of the South Wales and West Mercia constabularies, and preliminary funding from the Welsh Development Agency. They also thank Nicola Jones and Anna Radcliffe of the Cartographic

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

273

Unit, University of Wales Swansea, for preparing the figures, and to two anonymous referees for comments on an earlier version of this paper. References Bottoms, A. E., & Wiles, P. (1997). Environmental criminology. In M. Maguire, R. Morgan. & R. Reiner, The Oxford handbook of criminology (pp. 305–359). Oxford: Oxford University Press. Bowers, K., & Hirschfield, A. (1999). Exploring links between crime and disadvantage in north-west England: an analysis using geographical information systems. International Journal of Geographical Information Science, 13, 159–184. Brantingham, P. L., & Brantingham, P. J. (1993). Nodes, paths and edges: considerations of the complexity of crime and the physical environment. Journal of Environmental Psychology, 13, 3–28. Bromley, R. D. F., & Thomas, C. J. (1997). Vehicle crime in the city centre: planning for secure car parking. Town Planning Review, 68, 257–278. Bromley, R. D. F., Thomas, C. J., & Millie, A. (2000). Safety concerns in the night-time city centre. Town Planning Review, 71, 71–96. Brumm, H. J., & Cloninger, D. O. (1995). Violent crime and punishment: an application of the LISREL Model. Applied Economics, 27, 719–725. Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: a routine activity approach. American Sociological Review, 44, 588–608. Cohen, S. (1981). Folk devils and moral panics. London: Paladin. Crime Concern (1992). Cardiff city centre: night time safety strategy. Cardiff: Crime Concern. Davidson, R. N. (1989). Micro-environments of violence. In D. J. Evans, & D. T. Herbert, The geography of crime (pp. 59–85). London: Routledge. Denham, C. (1993). Census geography. In A. Dale and C. Marsh, The Census user’s guide (pp. 168– 200). London: HMSO. Doeksen, H. (1997). Reducing crime and fear of crime by reclaiming New Zealand’s suburban street. Landscape and Urban Planning, 39, 243–252. Department of the Environment (1994). Vital and viable town centres: meeting the challenge. Report of the study undertaken by Urban and Economic Development Group (URBED) in association with Comedia, Hillier Parker, Bartlett School of Planning University College London, and Environmental and Transport Planning, Department of the Environment. London: HMSO. Ekblom, P. (1988). Getting the best out of crime analysis. London: HMSO (Home Office Crime Prevention Unit Paper No. 10). Engs, R. C., & Hanson, D. J. (1994). Boozing and brawling on campus: a national survey of violent problems associated with drinking over the last decade. Journal of Criminal Justice, 22, 171–180. Farrington, D. P., & Dowds, E. A. (1985). Disentangling criminal behaviour and police reaction. In D. P. Farrington, & J. Gunn, Reaction to crime: the public, the police, courts and prisons. Chichester: Wiley. Fisher, B., & Nasar, J. L. (1995). Fear spots in relation to microlevel cues: exploring the overlooked. Journal of Research in Crime and Delinquency, 32, 214–239. Gottfredson, M. R. (1984). Victims of crime: the dimensions at risk. London: HMSO (Home Office Research and Planning Unit Report No. 81). Herbert, D. T. (1993). Neighbourhood incivilities and the study of crime in place. Area, 45, 45–54. Hillier Parker Research (1996). Shopping centres of Great Britain: a national survey of retailer representation by trader location. London: Hillier Parker May and Rowden. Hirschfield, A., Brown, P., & Todd, P. (1995). GIS and the analysis of spatially referenced crime data: experiences in Merseyside, UK. International Journal of Geographical Information Systems, 9, 191–210. Hope, T. (1985). Implementing crime prevention measures. London: HMSO (Home Office Research Study No. 86). Hough, M., & Lewis, H. (1989). Counting crime and analysing risks: the British Crime Survey. In D. J. Evans, & D. T. Herbert, The Geography of Crime (pp. 16–37). London: Routledge.

274

A.L. Nelson et al. / Applied Geography 21 (2001) 249–274

Levi, M. (1997). Violent crime. In M. Maguire, R. Morgan, & R. Reiner, The Oxford handbook of criminology (pp. 841–889). Oxford: Oxford University Press. Maguire, M. (1997). Crime statistics, patterns, and trends: changing perceptions and their implications. In M. Maguire, R. Morgan, & R. Reiner, The Oxford handbook of criminology (pp. 135–188). Oxford: Oxford University Press. Mayhew, P., Aye Maung, N., & Mirrlees-Black, C. (1993). The 1992 British Crime Survey. London: HMSO (Home Office Research Study No. 132). Miller, T. R., Cohen, M. A., & Rossman, S. B. (1993). Victim costs of violent crime and resulting injuries. Health Affairs, 12, 186–197. Millward, R. E. (1988). The safe city: municipal strategies for preventing public violence against women. Toronto: City of Toronto, Planning and Development Department. Mirrlees-Black, C., Mayhew, P., & Percy, A. (1996). The 1996 British Crime Survey: England and Wales. London: HMSO (Home Office Statistical Bulletin, Issue 19/96, Research and Statistics Directorate). Mirrlees-Black, C., Budd, T., Partridge, S., & Mayhew, P. (1998). The 1998 British Crime Survey: England and Wales. London: HMSO (Home Office Statistical Bulletin, Issue 21/98, Research Development Statistics). Nasar, J. L., & Fisher, B. (1993). Hot spots of fear and crime: a multi-method investigation. Journal of Environmental Psychology, 13, 187–206. Nelson, A. L. (1997). Fear of parking. Town and Country Planning, 66(3), 3. Nelson, A. L., Bromley, R. D. F., & Thomas, C. J. (1996). The geography of shoplifting in a British city: evidence from Cardiff. Geoforum, 27, 409–423. Nottingham Safer Cities Project (1990). Nottingham Safer Cities Project Steering Group report on safety in the city centre. Nottingham: Nottingham City Council. Oc, T., & Tiesdell, S. (1997). Towards safer city centres. In T. Oc and S. Tiesdell, Safer city centres: reviving the public realm (pp. 222–238). London: Paul Chapman. Office of Population Censuses and Surveys (1994). 1991 Census: key statistics for local authorities. London: HMSO. Potter, L. B., Sacks, J. J., Kresnaw, M.-J., & Mercy, J. (1994). Non-fatal violence, United States, 1994. Public Health Reports, 114, 343–352. Poyner, B. (1983). Design against crime: beyond defensible space. London: Butterworth. Read, T., & Oldfield, D. (1995). Local crime analysis. London: HMSO (Home Office Crime Detection and Prevention Series Paper No. 65). Shepherd, J. (1990). Violent crime in Bristol: an accident and emergency department perspective. British Journal of Criminology, 30, 289–305. Shepherd, J. (1994). Violent crime: the role of alcohol and new approaches to the prevention of injury. Alcohol and Alcoholism, 29, 5–10. Shepherd, J., & Brickley, M. (1996). The relationship between alcohol intoxication, stressors and injury in urban violence. British Journal of Criminology, 36, 546–566. Sherman, L. W., Gartin, P. R., & Buerger, M. E. (1989). Hot spots of predatory crime: routine activities and the criminology of place. Criminology, 27, 27–55. Sparks, R. F., Genn, H. G., & Dodds, D. J. (1977). Surveying victims. Chichester: Wiley. Taylor, I. (1994). The political economy of crime. In M. Maguire, R. Morgan, & R. Reiner, The Oxford handbook of criminology (pp. 469–510). Oxford: Oxford University Press. Tomsen, S. (1997). A top night: social protest, masculinity and the culture of drinking violence. British Journal of Criminology, 37, 90–102. Thomas, C. J., & Bromley, R. D. F. (1996). Safety and shopping: peripherality and shopper anxiety in the city centre. Environment and Planning C: Government and Policy, 14, 469–488. Thomas, C. J., & Bromley, R. D. F. (2000). City centre revitalisation: problems of fragmentation and fear in the evening and night-time city. Urban Studies, 37, 1403–1429. Tuck, M. (1989). Drinking and disorder: a study of non-metropolitan violence. London: HMSO (for Home Office). Young, J. (1997). Left realist criminology? Radical in its analysis, realist in its policy. In M. Maguire, R. Morgan, & R. Reiner, The Oxford handbook of criminology (pp. 471–498). Oxford: Oxford University Press.