Back to school: an analysis of repeat school victimization in Malmö

ABSTRACT

This study examines repeat school victimization, and is an extension of previous research. The distributions of two property crimes (burglary and theft) and three crimes against the person (assault, threats and harassment) in compulsory schools are compared between municipal and independent schools in Malmö, as well as between school areas with varying economic status. The time course between incidents is compared between crime types and between areas with low and high economic status, respectively. The study is based on reported crime data, and is geographically limited to the municipality of Malmö. The data relate to the period 2015–2017. The study’s results show a skewed distribution for victimization in general, with municipal schools and schools in areas with low economic status being most affected. The majority of repeat school victimization takes place within a month of the previous incident, and this time frame applies to all four crime types. Burglaries and assaults are repeated more quickly in the most deprived areas, while for theft and threats/harassment, there is very little difference in the time course between different areas. Victimization differences and preventive implications are discussed based on the theories of routine activities, social disorganization and macro-level strain.

KEYWORDS:

• Repeat victimization
• school
• Malmö
• routine activities
• social disorganization
• macro-level strain

Introduction

Repeat victimization is defined as a target being repeatedly exposed to crime. A target can be a person or household, but it can also be an object (Grove & Farrell, 2010), such as a school. In the 1980s and 1990s, researchers noted the importance of studying repeatedly victimized schools, a research topic that was first introduced in the UK with a focus on repeat school burglaries (Bowers et al., 1998; Burquest et al., 1992; Hope, 1986), which was later extended in Sweden (Lindström, 1997). Since the 1990s, Swedish schools have continued to change, whereas the development of this specific field of research has come to a complete halt. Instead, a great deal of focus has been directed at residential burglaries, despite the fact that several studies have established that schools constitute a more vulnerable target (Bowers et al., 1998; Brå¹, 2001; Guidi et al., 1997; Townsley et al., 2000).

This is not to say that the research on repeat victimization in connection with schools is limited. Plenty of research has studied repeat victimization at the individual level among students, with a special focus on repeated exposure to violence and bullying (e.g. Perry et al., 1988; Tillyer et al., 2018). However, crimes committed against or at a school do not only affect specific individuals. A school should be viewed as a contextual unit consisting of many interdependent characteristics: the building itself, which should keep what is on the inside safe; students and educational staff, with all their related backgrounds and varying capabilities, who are forced to interact with each other on a daily basis; norms and rules (including the school’s policy with regard to reporting crime to the police), which each individual school determines on the basis of its own risks and needs; and the neighbourhood in which the school is located, with its varying conditions. And when a crime occurs, this whole entity is affected. In the event of a burglary, the building may be physically damaged, but the financial loss and potential subsequent feelings of unsafety affect both students and staff, and also the local community. When a person is assaulted or threatened, the individual may constitute the direct victim, but the conflict itself naturally also affects the others who are present (Nofziger, 2009), affecting the school as both a workplace and learning environment. The aim of this study is therefore to build on previous research and to examine and compare repeat victimization between schools, and not between individuals. Hence, reported criminal incidents that have taken place against or at a school will be referred to as school victimization.

To date, only two studies have specifically examined repeat victimization among schools. The first, by Burquest et al. (1992), presented a distribution and a general time course for repeat school property victimization (mainly burglary). The second, by Lindström (1997), made an extension by also including violent crime and by examining the relationship between victimization rates among schools and the social status of school areas. However, both studies – like all studies – have their shortcomings, and several issues that were raised could not be answered due to the limited amount of reported crime data. The current study will address some of the shortcomings associated with these prior studies, and will also answer some of the questions that they were unable to examine. The study will describe the victimization distribution between comprehensive schools for both property crimes (burglary and theft) and crimes against the person (assault, threats and harassment). It will also examine the victimization rate in relation to the economic status of different areas, and compare the time course between incidents across crime types and areas. Finally, the study will also compare the victimization rate between municipal schools and independent schools (i.e. schools that are not owned by the municipality), which appeared in Sweden during the 1990s and which have become increasingly common over time (Östh et al., 2013; Sandahl, 2020; Yang Hansen & Gustafsson, 2016). This issue was not examined in the studies described above.

The aim of this study is thus to analyse repeat school victimization in Malmö, focusing on school type, school location and time course. The term repeat school victimization refers to whether a school has reported the same type of crime twice or more, as defined in the previous research (e.g. Bowers et al., 1998; Brå, 2001; Burquest et al., 1992; Lindström, 1997). The ambition is to bring new life to a research area that has been on pause for over 20 years, but also to provide insights into the type of schools in which preventive measures should be implemented, in which areas, and how quickly.

The importance of studying repeat school victimization in a Swedish context

During the 1990s, a major school reform took place in Sweden, which meant that the responsibility for schools was transferred from the state to the municipalities, and that schools became voucher-financed (i.e. paid for each attending student). It also meant that independent schools run by private actors – often profitmaking companies (Bunar & Ambrose, 2016) – became more numerous, and that children and their parents were given the freedom to choose which schools to attend (ibid; Sandahl, 2020). Although the majority still choose the nearest school within their catchment area (Bunar & Ambrose, 2016), an increasing number of children attend schools located further from home (ibid; Östh et al., 2013) because they are perceived as being better. According to parents, a ‘better’ school is a safe school, comprised primarily of Swedish-born students of a higher-class background (Bunar & Ambrose, 2016). The parents who make these active choices are the more socio-economically advantaged (Sandahl, 2020; Yang Hansen & Gustafsson, 2016), and those who are more engaged in their children’s education (Bunar & Ambrose, 2016). In many cases, parents may want the best for their children, but are hindered from making an informed choice by limitations with regard to time, money and social networks (ibid.). The consequence of this has been segregated schools with varying student compositions, educational achievement (Östh et al., 2013; Yang Hansen & Gustafsson, 2016) and dominant basic values (Bunar & Ambrose, 2016; Yang Hansen & Gustafsson, 2016), with the most socially, economically and academically advantaged students appearing to cluster in the independent schools (Bunar, 2010).

Meanwhile, the law states that Swedish schools must strive for equity, for the children’s best interests and must offer all students – regardless of which school they choose – a safe environment (Education Act (SFS 2010:800) Ch. 1, paras. 9–10, Ch. 5, Para. 3). Whether the goals and guidelines of this legislation are being met should also be examined from a victimological perspective.

Children and adolescents spend the majority of their daytime in school, and it is thus important that the school environment is safe. However, in Sweden, youngsters report that school is the most common location for crime victimization, and those exposed to crime develop anxiety and fear (Brå, 2020). Feelings of unsafety have been shown to affect students’ mental health (Nijs et al., 2014), school avoidance (Hughes et al., 2015) and academic achievement (Lacoe, 2020). Thus, a school that repeatedly reports crime to the police should consequently be a school with unsafe students who are more likely to feel bad mentally and perform poorly. In addition, educational staff must also feel safe at their workplace. Fear and unsafety affect teachers physically and emotionally, and also affect their teaching abilities (Moon & McCluskey, 2020; Vettenburg, 2002; Wilson et al., 2011). It is therefore likely that repeatedly victimized schools will have lower educational standards. This would then result in not all schools being able to provide an equally good education and an equally safe environment in line with the goals of the Swedish legislation.

A school’s situation is not limited to the school itself but must be observed in a community context. In the disadvantaged areas of Malmö, the SEI values² of comprehensive schools are the most negative, and the proportions of those qualifying for upper secondary school are the lowest (statistics from Skolverket³). The conditions that can be found in deprived areas are thus also found within the local schools (Bunar, 2010), which is not surprising since the majority of students live in the school neighbourhood (Bunar & Ambrose, 2016). Behavioural problems and crime will also be reflected in schools, as ‘unsafe schools are lodged within unsafe neighborhoods, and (…) serious violators of school rules and the student perpetrators of criminal acts are the products of the homes and communities that breed them’. (Menacker et al., 1990, p. 76). However, a school can also be a tempting target for a motivated outsider, such as a violent intruder (Hellman & Beaton, 1986) or a burglar who lives nearby (Vandeviver & Bernasco, 2020). On the other hand, research indicates that schools also have an impact on the school neighbourhood. Residential areas adjacent to schools have a higher crime rate (Willits et al., 2013), and when schools with the lowest performing and the least law-abiding students are closed, neighbourhood crime decreases (Steinberg et al., 2019). It is therefore important to prevent schools from being repeatedly victimized, not only for the sake of both students and educational staff, but also for the sake of the local community.

Routine activities, social disorganization and macro-level strain

Routine activities theory (RAT), as developed by Cohen and Felson (1979) assumes that the following three factors must converge in time and space for a crime (and thus victimization) to occur: (1) a motivated offender, i.e. someone with ‘criminal inclinations and the ability to carry out those inclinations’ (Cohen & Felson, 1979, p. 590), (2) a suitable target, such as a school, an item or a person, and (3) the absence of capable guardians, e.g. burglar bars, anti-theft alarms or a teaching assistant being present at break time. If any of these three factors are eliminated from the equation, crime and victimization should not take place. Depending on an individual’s daily activities and routines, the risk of being exposed to a crime will vary (Cohen & Felson, 1979). However, routine activities are also crucial from the motivated offender’s perspective, as he often chooses his targets in areas and places that he knows well and which are within the perpetrator’s daily activity patterns (Brantingham & Brantingham, 1993). People at a school, such as students and staff, are involved in joint routine activities 5 days a week, and the school thus becomes a place where the interaction between a motivated offender, a suitable target and an absence of capable guardians is most likely to occur at some point during the day (Popp & Peguero, 2011). Crimes committed by outsiders will also most likely be committed by offenders who live in the school area (e.g. Bernasco, 2010; Vandeviver & Bernasco, 2020) and who pass the school during their daily activities (Kleemans, 2001).

Motivated offenders are more numerous in certain areas (Piscitelli & Doherty, 2018) and in certain schools (Sandahl, 2020). Shaw and McKay (1942) discovered that crime rates were higher in deprived areas and explained the connection using the concept of social disorganization. In socially disorganized areas – which are characterized by concentrated disadvantage – it becomes more difficult to establish anti-criminal values, while the material benefits of committing crime are more obvious. In such areas, the number of motivated offenders increases, and the targets that they choose will be made based on RAT (Piscitelli & Doherty, 2018). However, integrating RAT with the theory of social disorganization can become problematic. On the one hand, motivated offenders may be more clustered in low-income areas, but in high-income areas suitable targets may be more numerous and more attractive (ibid.; Hipp, 2007).

The proportion of motivated offenders may also be greater in certain schools, regardless of the area in which the schools are geographically located. Sandahl (2020) employed macro-level strain theory (MST) to better understand school variation in violence and general offending and concluded that school-contextual strain may have an impact on students’ motivation to commit crimes. According to Agnew (1999), strain will be greater in deprived units (such as in deprived schools) since the people in these units will perceive difficulties in achieving financial success, high status and respect in legitimate ways. This type of contextual strain will manifest itself in the form of an aggregated anger and frustration, and the motivation to commit crime will become stronger. In such schools, not only will the proportion of motivated offenders grow, but the risk will also increase that angry and frustrated individuals will interact and end up in conflict with each other during the course of a school day (ibid.). The risk of being victimized – both as direct, but also indirect victims who are only witnesses to the crime (Nofziger, 2009) – will therefore increase, which will in turn further intensify the aggregated anger and frustration. As Agnew (1999, p. 128) has stated: criminal victimization is in fact ‘one of the most serious types of strain to which individuals are subject’. Schools will thus vary in terms of their level of strain, their proportion of motivated offenders and their victimization risk.

Previous research on repeat school victimizations

One of the first studies to find a skewed distribution of burglaries among British schools was published in the 1980s, and when the schools were compared, a difference in opportunities to break in without being detected was noted (Hope, 1986). While the less victimized schools ‘were situated in areas of greater population density’, the schools that were most exposed were in ‘quiet, suburban neighborhoods away from main roads’, meaning that ‘once inside, burglars could operate without fear of being seen’ (ibid., p. 76).

This first British study was followed by two others, both of which were based on property crimes (mainly burglaries) reported to the police in Merseyside, Liverpool over the course of a year. Burquest et al. (1992) found that of 33 schools, only 21% had been victimized once, while the remainder had been victimized twice or more. Bowers et al. (1998) noted that while repeat home burglaries were well researched, not much had been published on repeat victimization of non-residential properties. Having then studied and compared both residential burglaries and non-residential burglaries, they found that educational properties were the most likely property type to be repeatedly victimized (ibid.).

The first Swedish study to analyse repeat school victimization, was a replication of the study conducted by Burquest et al. (1992) Lindström, (1997). The sample consisted of 96 schools located in Stockholm, and the study included both property crimes and school violence. The period examined was once again one year. Similar to the results found in Merseyside, the Stockholm study noted a similarly skewed distribution for property crimes (ibid.). The number of schools reporting school violence was few, but among them, 22% accounted for more than half (52%) of the reported crimes, making them ‘multiple violence schools’ (ibid., p. 125). A second Swedish study focused on the municipality of Örebro (using the municipality of Sundsvall as a control area) and on crimes reported to the police between 1997 and 1999 (Brå, 2001). Several different crime types were included, with repeat school burglaries being the most frequent type of repeat victimization events found for property crime.

Although these studies have their advantages, they are also subject to a number of limitations. The most important shortcoming is that the focus was mainly directed at repeat burglaries. Although the study by Burquest et al. (1992) was extended to include other types of property crime, it is unclear exactly which types of crime were included in the data. The only study that has included crimes against the person (Lindström, 1997) is also vague on this point, since the concepts of school violence and violent crime are broad and not defined in the article. The current study has tried to avoid this lack of clarity.

Repeat school victimization in relation to school location and time course

Repeat victimization is more common in deprived areas (Johnson et al., 1997), and this seems also to be true for repeat school victimization. The numbers of repeatedly burgled educational properties were much higher in the most disadvantaged areas of Merseyside (Bowers et al., 1998). In Lindström’s (1997) Stockholm study, the average victimization rate for property crime in schools was much higher in low social status areas than in school areas with high social status. Although statistically significant differences were not established for violent crime, it was observed that reported school violence was also primarily a problem in schools in low-status areas (ibid.). In the Örebro study, the schools with the highest rates of burglary were also mainly located in deprived areas (Brå, 2001).

Most of the repeat school victimization events occurred within the first month (Bowers et al., 1998; Brå, 2001; Burquest et al., 1992; Lindström, 1997). Research has shown that a short time course between incidents is a result of same-offender involvement, i.e. that the same motivated offender commits crimes against the same victim or object within a short period of time (e.g. Ashton et al., 1998; Bernasco, 2008; Everson, 2003; Johnson et al., 2009; Lantz & Ruback, 2017). In the UK, over 50% of all repeat burglaries in educational properties took place within one month of the previous incident (Bowers et al., 1998). Similar patterns have been found in Sweden; just over 40% of all repeat school burglaries in Örebro occurred within the first month (Brå, 2001). In the studies conducted by Burquest et al. (1992) and Lindström (1997), the temporal concentration was even more pronounced; repeat property incidents took place within a month in almost 80% of cases (74% in Stockholm and 79% in Merseyside). In addition, Lindström (1997) compared the time course between crime types and between areas with high and low social status. For both property and violent crime, the proportions of repeat incidents that occurred within a month were higher in low status areas. As was noted by Lindström (1997), however, since school violence was very rare in the Stockholm data, particularly in high-status areas, it was not possible to draw any firm conclusions. Among the types of crime examined, it was mainly repeat thefts that took place within one month of the previous incident, with repeat burglaries in second place and repeat vandalism in third. Repeat school violence was also primarily reported within the first month, but to a lesser extent than the other three crime types examined (ibid.).

The current study

This study is inspired by previous research conducted by Burquest et al. (1992) and Lindström (1997), while avoiding the above-described limitations. The conclusions that can be drawn from the study are whether preventive measures should be directed at a certain type of school, in certain areas, within a certain period of time, and whether such a preventive approach can be applied to both property and personal crime. A difference in victimization may mean that Swedish schools are unequal in terms of being able to offer students and educational staff a safe daytime environment, which can in turn lead to differences in physical and mental health, and also in educational achievement. This is contrary to the goals set out in the Swedish Education Act. Victimization differences and preventive implications will be discussed based on the theories of routine activities, social disorganization and macro-level strain. The following questions will be addressed:

How is victimization distributed among compulsory schools in Malmö?

Does victimization differ between municipal and independent schools?

Does victimization differ between school areas with varying economic status?

How quickly do repeats occur following a previous incident, and are there any differences between crime types and between school areas with low and high economic status, respectively?

Data and methods

Reported crime data

The reported crime data were obtained following approval from the Swedish Ethical Review Authority (Dnr. 2019–00206). A description of the project with contact information was given on Malmö University’s website. No complainant has contacted me regarding a desire to opt out of the study. The reported crimes have been registered by the Swedish Police Authority and are geographically limited to the municipality of Malmö. The offences included in the study had occurred and been reported between 2015 and 2017. The information obtained for each criminal incident is as follows: offence code, date of offence, registration date, complainant (a person, a school, or a school organizer/owner), and the address at which the crime occurred. The offence types that have been included are (1) school burglary, (2) theft in school, (3) assault in school and (4) threats and harassment in school (these two offences have been combined into a single category). The essential factor for selecting an offence for inclusion in the study was that the crime had occurred against a school or on school grounds.

Other types of data

All between-area comparisons are focused on Malmö’s 10 districts,⁴ which are easy to find on Malmö City’s website. The economic status of the districts was specified on the basis of the official statistics from Statistics Sweden that are available on this same website. The variables used for this purpose were (1) the average proportion of unemployed residents, (2) the average proportion of residents receiving welfare benefits, and (3) the residents’ average incomes from employment and capital for the years 2015–2017, for each district. Based on these variables, the districts were ranked and sorted into four economic-status categories: low, medium low, medium high and high.

A complete list of all the municipal and independent schools in Malmö for the years 2015–2017, and the districts in which these were located, was obtained from the Schools Administration in Malmö. Almost all of Malmö’s compulsory schools are included in the material; of 106 compulsory schools, a reported offence of one or more of the types included in the study had been reported by 100 of these schools. An offence had been reported by 75 of the 80 municipal schools included in the sampling frame, and by 25 of 26 independent schools. Smaller schools for children with special needs, such as children with autism, learning disabilities or children with problem behaviours such as aggression, are also included.

The selection process

The material obtained from the police was limited to crimes that had both taken place (based on the crime date) and been reported to the police (based on registration date) between 1 January 2015 and 31 December 2017. Crimes reported during these 3 years but which had taken place prior to January 2015 were thus excluded. Previous Swedish research has shown that burglary and theft are the most common types of repeat victimization in schools (Brå, 2001; Lindström, 1997), and these were therefore included as the study’s two property crimes. The crimes against the person that young Swedish people state that they are most commonly exposed to are assault, threats and sexual offences (Brå, 2020). Since very few reported sexual offences had taken place in schools, this offence type was replaced by harassment, which had been reported to a greater extent. It should be noted that sexual offences have most likely taken place at the schools, but have not been reported due to the sensitive nature of sex crime. The crimes against the person that were included are thus assault, and threats/harassment (the latter two were combined into a single category in order to obtain a sufficiently large number of offence reports).

In the next step, all reported incidents were examined manually. All crimes against the person that had occurred in a place other than a school were excluded. In a few cases, a school was listed as the site of the crime but the address was incorrect; such reports were removed. The requirement was that these crimes had taken place on school grounds, and thus reports that stated the offence had occurred outside a school, or next to a school, were excluded from the sample. All reports that only stated the address of the crime but contained no further information about the actual location were also removed. This means that many crimes against the person that have in fact taken place at a school, but that the police have failed to register as crimes at a school, have not been included in the sample.

The examination of the reported property crimes was simpler, since school burglary and thefts in school have their own offence codes. All the reports that did not name a specific school, but included an exact address, were carefully processed. Many addresses could be linked to schools, and such reports were included in the sample. In cases where the address could not be linked to a school, or where the address was incomplete in any way, the reported incident was excluded. In cases where the same address covered both a school and another type of property, the incident was coded as unspecified. These reports were sent back to the police, who investigated whether a location was indicated in the written offence description contained in the register. Following this process, only two unspecified incidents remained, one burglary and one theft. These two offences were excluded from the sample.

Limitations and problems with reported crime data

Studying repeat school victimization on the basis of incidents reported to the police by no means constitutes a new approach (Bowers et al., 1998; Brå, 2001; Burquest et al., 1992; Lindström, 1997). However, the results in this study will only show a distribution for reported victimization incidents and not for those that have actually occurred. During the selection process, shortcomings were also noted regarding the register itself. Although police officers are asked to state the location of the crime when registering reports, many fail to do so. For this reason, many incidents had to be excluded. However, the use of a three-year observation period, instead of the 1-year period employed in other studies, should compensate for this. Farrell et al. (2002) found that a 3-year window captured 57% more repeats than a one-year window. This speaks in favour of the current study.

Despite an extended observation period, however, the problem remains that crimes that have been reported before 2015 and after 2017 are not included. A single incident in the study may well be a repeat in reality.

Data analyses

To investigate whether the victimization distribution differs between municipal and independent schools, a comparison of the average victimization rate between all municipal schools (n = 80), and all independent schools (n = 26) was conducted with the help of an independent two-sample t test.

Another research question examined in the study is that of whether the victimization distribution differs depending on the economic status of the area in which the school is located. A comparison between the four types of crime was conducted using one-way ANOVA with a post-hoc Tukey Test. The average victimization rate was calculated for all of the 106 schools. The distribution of the schools across different types of area is as follows: 17 schools in low-status areas; 31 schools in medium low-status areas; 20 schools in medium high-status areas and 38 schools in high-status areas.

Results

General distribution of school victimizations

A total of 367 school burglaries, 418 thefts in school, 535 assaults in school and 327 incidents of threats/harassment in school were reported. While 50 schools had reported all four crime types, only 6 had not reported any of these crime types during the 3-year period examined. Regardless of the type of crime, approximately three quarters of the schools had reported repeats. The most affected schools had reported a maximum of 16 burglaries, 23 thefts, 36 assaults and 23 cases of threats/harassment.

Upon closer examination, the distribution of property crime shows that 11% of the schools that had reported a property crime account for 32% of all property crimes, while 20% of the schools that had reported a property crime account for 49% of all such offences. A similar distribution can be found for crimes against the person. 7% of the schools that had reported a crime against the person account for 27% of all crimes against the person, and 20% of the schools that had a crime of this kind account for 55% of all such offences.

Repeat school victimizations and school type

As shown in Table 1, there is a statistically significant difference in the mean victimization rate for burglary, assault and threats/harassment between school types, with these crime types being reported to a greater extent by municipal schools than by independent schools. The difference between types of school is most obvious for assault. Among the independent schools, a maximum of 7 assaults has been reported, compared with the municipal schools, where the maximum number of reported assaults was 36 during the 3-year period examined. The average victimization rate for assaults is 5.01 higher among the municipal schools than among independent schools.

Table 1. School victimization by school type and type of crime

School type	Number of victimization incidents	Number of victimization incidents (range)	Average victimization rate	Std. deviation
Burglary
Municipal	320	0–16	4.00	4.20
Independent	47	0–5	1.81	1.65
t = – 3.846, p = .000
Theft
Municipal	348	0–17	4.35	4.35
Independent	70	0–23	2.69	4.59
t = – 1.666, p = .099
Assault
Municipal	502	0–36	6.28	7.41
Independent	33	0–7	1.27	1.85
t = – 5.539, p = .000
Threats/harassment
Municipal	294	0–23	3.68	4.58
Independent	33	0–4	1.27	1.37
t = – 4.158, p = .000

A statistically significant difference between school types was not found for theft. However, it should be noted that the average victimization rate is somewhat higher for municipal schools compared to independent schools.

Repeat school victimizations and school location

A statistically significant difference in the average victimization rate between school areas based on economic status was established for burglary, assault and threats/harassment, but not for theft (see Table 2). When comparing the average victimization rate between Malmö’s ten districts, it can be seen that theft offences mainly seem to take place in schools in the districts with the highest economic status. During the years 2015–2017, all schools within these districts had reported at least two thefts.

Table 2. School victimization incidents by school area status and type of crime

School area status	Number of victimization incidents	Average victimization rate		Std.Deviation	Tukey’s HSD Comparisons
					Low	Md. Low	Md. high	High
Burglary
Low	110	6.47		5.26	–	.063	.000	.005
Md. low	116	3.74		3.71		–	.150	.764
Md. high	31	1.55		1.73			–	.529
High	110	2.89		3.29				–
F = 6.26 (df = 3), p = .001
Theft
Low	97	5.71	5.27		–	–	–	–
Md. low	94	3.03	3.52		–	–	–	–
Md. high	67	3.35	2.82		–	–	–	–
High	160	4.21	5.25		–	–	–	–
F = 1.51 (df = 3), p = .216
Assault
Low	165	9.71		9.00	–	.109	.192	.002
Md. low	162	5.23		7.68		–	1	.357
Md. high	108	5.40		5.55			–	.417
High	100	2.63		4.24				–
F = 4.68 (df = 3), p = .004
Threats/harassment
Low	118	6.94		4.94	–	.001	.010	.000
Md. low	76	2.45		4.53		–	.977	.956
Md. high	58	2.90		3.04			–	.819
High	75	1.97		2.95				–
F = 7.07 (df = 3), p = .000

Since no significant difference between school areas was found for theft, no post-hoc test was conducted.

Although there is a significant difference between areas for three of the crime types, a clear pattern emerges; the type of area that primarily differs significantly from the others is comprised of the areas with low economic status. This is especially true for crimes against the person, for which a clear difference in average victimization rates can be distinguished between the school areas with the lowest and highest economic status, respectively.

Burglary differs in that the lowest average victimization rate is found in school areas with medium high economic status. This can probably be explained by the fact that the city centre district is included in this type of area, and is also the district in which the schools are least exposed to burglary. Almost 40% of the compulsory schools located in the city centre had not reported a single burglary during the three-year period of the study, and the average victimization rate among the compulsory schools within this district was only 1.56 burglaries per school.

Repeat school victimizations and time course

As can be observed in Figure 1, the majority of repeat incidents occur within the first month following the previous incident. Assault is the type of crime that is most often repeated within the first month in Malmö’s schools; 41% of the repeat school violence occurred within a month of the preceding incident, of which 35% of the repeat incidents had taken place within a week. 37% of the repeat school burglaries occurred within the first month, and as many as 43% occurred within a week of the preceding incident. Threats and harassment are the crime category with the greatest dispersion over time.

Figure 1. Time course for repeat school victimization by crime type.

Figure 2 shows the proportion of repeat incidents occurring within the first month in areas with the lowest and the highest economic status. A difference between areas can be noted for burglary and assault, but not for theft or threats/harassment. In the areas with the lowest economic status, 45% of the repeat school burglaries occurred within the first month, of which 42% occurred within the first week. In the districts with the highest economic status, burglaries are not repeated as quickly; here 24% of the repeats had taken place within a month, and 35% of these within a week.

Figure 2. Proportion of repeat school victimization incidents occurring within one month, by crime type and type of area.

Results from χ² test for differences in proportions: burglary (p = .005); theft (p = .970); assault (p = .029); threats/harassment (p = .947).

Almost half of the repeat assaults occurred within the first month in low-status areas, as compared with only one-third in high-status areas. Of those that occurred within a month, however, more were repeated within the first week in high-status than in low-status areas (35% compared to 25%).

Repeat thefts and threat/harassment incidents occurred in almost identical proportions within the first month in both types of area. However, these incidents occurred within the first week to a somewhat greater extent in low-status areas.

Conclusions

The results show that victimization is skewed among compulsory schools in Malmö, both in terms of property and personal crimes. The most vulnerable schools are municipal schools, and schools in areas with low economic status. The majority of the repeat incidents take place within a month – and often within a week – of the previous incident. Such a time course was noted for all four crime types. Repeat burglaries and assaults occur more quickly in areas with low economic status, compared to areas with high economic status. For repeat thefts and threats/harassment, the difference in the time course is negligible.

Discussion

Lindström (1997) found a statistically significant difference in the average victimization rate between areas with varying social status for property crime, but not for violent crime. In this study, there is a significant difference for all types of crime with the exception of theft. During the 3 years covered by the study, many theft offences have taken place in schools in high-economic-status districts. This demonstrates the problem raised by Hipp (2007) and by Piscitelli and Doherty (2018) in linking social disorganization with RAT, in that motivated offenders may be more numerous in areas with concentrated disadvantage, whereas crime targets may be more suitable and attractive in high-income areas. Schools here might then contain more theft-prone items that trigger a propensity for theft among the school’s own students, or that attract outsiders. For other types of crime, schools in deprived and thus the most socially disorganized areas are the most affected, with the reservation that the lowest mean victimization rate for burglary was found in districts with a medium–high economic status. This group of districts includes the Malmö city centre, which has the lowest level of reported school burglaries in the entire city. The explanation for this might possibly be that presented by Hope (1986), who found that there were more burglaries in areas where burglars were able to work undisturbed; it may seem logical to choose not to break into buildings that are located in areas where people are constantly passing by and thus assuming the role of capable guardians.

Previous research has noted a clear time course for repeat school victimization (Bowers et al., 1998; Brå, 2001; Burquest et al., 1992; Lindström, 1997), and a similar time course was found in this study among the schools in Malmö. The relatively brief time noted between victimization incidents may once again indicate same-offender involvement. Burglary is a crime type that has been the primary focus of previous studies. In Malmö, 37% of repeat school burglaries occurred within the first month, and this is comparable with the pattern reported for repeat school burglaries in Örebro (Brå, 2001). Assault constitutes the crime type with the highest level of repeat crime within the first month among Malmö’s schools, which is the opposite of what Lindström (1997) found for the schools in Stockholm.

A statistically significant difference in the average victimization rate has also been established between municipal schools and independents schools, for all crime types with the exception of theft. The most pronounced difference between these two types of schools was found for assault. Previous studies have shown that socio-economically advantaged parents are primarily those who make use of the opportunity to choose their children’s school (Sandahl, 2020; Yang Hansen & Gustafsson, 2016), and many socially, economically and academically advantaged children end up in independent schools (Bunar, 2010). Given that the average SEI value is more positive among Malmö’s independent schools than among the municipal schools, this type of student allocation is also found here. Students at more strained schools are ‘painfully aware and daily reminded of their schools’ low status, bad reputation, and low expectations held for them by society’ (Bunar, 2010, p. 10). The macro-level strain that according to Agnew (1999) creates aggregated anger and frustration may thus be greater in Malmö’s municipal schools, hence creating offender motivation and angry individuals who are forced to interact with each other within a restricted space. The victimization itself also causes strain, which will lead to even more anger and conflicts (ibid.), and thus a second, and third victimization becomes inevitable. The fact that there are more thefts than other types of crime in independent schools may possibly also indicate a difficulty in combining RAT with MST. Motivated offenders may be clustered at municipal schools, but there may be more attractive targets to steal in independent schools.

Preventive strategies

Based on RAT (Cohen & Felson, 1979), victimization/repeat victimization can be prevented by performing any of the following three things: making targets less attractive, making guardians more capable or reducing the motivation of potential offenders. The first two are easier to implement in the form of immediate measures to prevent the repeat incidents that are likely to occur within the first month or even within the first week, and these could be directed at motivated offenders both within the school and among outsiders. Acute interventions to avert property crime may involve improving surveillance in the form of quickly installing new alarms and having security guard patrol the site at night and on weekends. Over time, it is possible to implement measures such as installing window bars, securing items that are prone to theft, such as computers, and allowing others to use the school premises outside school hours. Many of these efforts could be intensified during the first few months, and then eased or removed when the risk of a repeat incident has diminished. With regard to crimes against the person, staff should immediately have an investigative conversation with the offender and the victim, and conduct a survey of the places and times when there is a risk of repeat violence and threats. If the risk of a new conflict is greatest during the transition between two classrooms, or during break-times, a school might have a staff member supervise on these occasions, or require the offender to spend his breaks indoors during the coming week. The schools that should primarily focus on the early introduction of prevention measures are those that are located in deprived areas. However, repeats also occur quickly in areas with high economic status, and for crime types such as theft and threats/harassment, there is very little difference in the time course between incidents.

In the longer term, schools should work with more general preventive measures to counteract both property and violent crime. Many conflicts between students occur during breaks, which can be prevented with the help of staff supervision at these times and staff-led break activities. In research, the presence of adults has been emphasized as an effective strategy for creating safety and preventing crime in schools (Odenbring & Johansson, 2019). Students who present a particular problem should be investigated and be given help by the student health team, and increased supervision from student assistants or student coordinators. Teaching focused on norms, values and morals, as well as social skills training for those who need it, should be scheduled and performed by competent staff. Mentors in Violence Prevention (MVP) is an example of a prevention programme that has been implemented both internationally and in Sweden, and which has shown positive effects on both attitude and behaviour among students (Bruno et al., 2020). In this way, a motivation to commit crime if not being eliminated could at least be reduced.

Limitations and future research

It is time to resume research interest in schools as a target for repeat victimization. Swedish schools are already segregated in terms of student composition and educational achievement (Östh et al., 2013; Yang Hansen & Gustafsson, 2016), despite the equity goals specified in the Swedish Education Act, and according to this study, schools are also segregated in terms of victimization and safety. Schools in deprived areas, and schools with strained students, do not need the added problems entailed by frequent victimization, such as poorer physical and mental health among staff and students, and in the end poorer school standards. This study represents a good start to reawaken interest in this issue, but it has shortcomings, and should therefore be viewed as a point of departure for further studies that may improve upon it. Studying school victimization on the basis of crimes reported to the police constitutes a major problem. The dark figure for offences such as assault and threats/harassment is substantial (Brå, 2018), and the fact that many reported incidents had to be excluded due to a lack of information on the crime location has naturally affected the study’s results. A better picture might possibly be obtained using the reports that schools are required to write in connection with incidents of crime and abuse at school. Schools’ policies for reporting crime to the police appear to be arbitrary, with no clear guidelines for schools to follow. The explanation for differences in victimization may thus be a simple question of policy differences. There is ongoing competition between schools, and parents themselves have stated that order and safety are factors that are taken into account when choosing a school for their children (Bunar & Ambrose, 2016). The significance of the way in which schools choose to deal with crime becomes even clearer when the small number of special needs schools included in the material are examined more closely. In these schools, where one might expect more reports of assault and threats/harassment – since these include schools for children with aggression and crime problems – very few offences, and in some cases none, had been reported during the 3 years examined in the study. It appears likely that many crimes against the person have in fact taken place at these schools, but that these have not been reported. The students at these schools are probably known to the social services, and the incidents may therefore be handled directly by them. Further, violence may perhaps be more normalized at such schools, and something that may be expected. A closer look at schools’ varying policies regarding the reporting of crime, and the ways in which incidents that are not reported are dealt with, would be of interest in the future research. Lunneblad et al. (2019) and Vainik (2017) have already initiated such research.

In addition, the conclusions presented in this study are theoretical. The fact that there is a difference in the number of victimization incidents between areas with different economic status does not mean that it is concentrated disadvantage that acts as a risk factor; it may be some other characteristic of these areas that has this effect. The studies performed by Burquest et al. (1992), Lindström (1997) and myself have primarily been descriptive, and to move forward, research needs to examine differences in victimization at a deeper level.

Further, although the results reported in the study are primarily applicable to Swedish conditions, there are Nordic countries whose school system is similar to Sweden’s, such as that in Finland (Östh et al., 2013), who can benefit from the results and the implications for future research.

In summary, this study suffers from a number of limitations. However, the picture that emerges is that most schools are victimized at some point, and that for a majority, this happens repeatedly. Repeat school victimization means insecurity among students, teachers and local communities, and as a result a poorer school experience for children. This conclusion has not changed since the 1990s; schools remain a place that should be prioritized by actors working with crime prevention and victim support.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. The Swedish National Council for Crime Prevention

2. Socioeconomic Index value = An estimated risk that a student will not qualify for upper secondary school, based on his or her school results and socio-economic background. The estimated probabilities at the student level are summed to the school level, and then summarized in an index score that provides guidance as to how much support is needed in each school.

3. The Swedish National Agency for Education

4. Centrum, Kirseberg, Västra Innerstaden, Södra Innerstaden, Rosengård, Husie, Limhamn-Bunkeflo, Hyllie, Fosie and Oxie