Deliberate firesetting: hotspot schools as arenas and preventive actors

ABSTRACT

Firesetting is one of the crime acts most representative of youth crime, and schools are among the buildings most often targeted, causing significant social, material and economic damage. This study examines schools as arenas with particular exposure to deliberate firesetting and as actors interpreting and utilizing their organizational scope to prevent school fires. The focus on school organizations is unique and urgently needed in research on juvenile firesetting, given their pivotal but under-researched role in crime prevention. The study is based on an in-depth analysis of data (interviews, documents and official registry data) related to 20 fire-exposed lower secondary schools in two major Swedish cities. These schools mobilized a broad repertoire of social, situational and structural measures. Interviewed school personnel perceived and responded to firesetting in relation to the institutional school setting, group dynamics, individual characteristics and local context. The schools were generally located in socio-economically disadvantaged areas and faced comprehensive educational and social challenges. The firesetting problem – and paradoxically some well-intentioned preventive efforts – risk adding further dimensions to school segregation and inequality, potentially increasing stigmatization and marginalization.

KEYWORDS:

• Juvenile firesetting
• crime prevention
• lower secondary schools
• residential segregation
• inequality

Introduction

The focus of this article is on schools as arenas of youth firesetting and as actors engaged in its prevention. firesetting is one of the crime acts most representative of youth crime (Kolko, 2002; Mackay, Feldberg, Ashley, & Marton, 2012), and schools are among the buildings most often targeted (Arson Prevention Bureau, 1998; Blomqvist & Johansson, 2009). School fires cause significant social, material and economic damage (MSB, 2011; US Fire administration, 2014). The yearly cost for school fires in Sweden has been estimated to exceed 10% of the total cost of fires in the country, or about 500 million Swedish crowns (USD 70 million) (Johansson, Wahlqvist, & Van Hees, 2012). Moreover, school fires pose a risk to health, injury and human life. They contribute to a general sense of insecurity and disrupt education (Jonsson, Lundqvist, Gell, & Andersson, 2017; Wade, Teeman, Golden, Wilson, & Woodley, 2007), and can have major juridical and economic consequences for the lives of the young people setting them, affecting their prospects and life chances.

Approximately, one in every two Swedish school fires is deliberately set (Blomqvist & Johansson, 2009), and lower secondary schools (with pupils aged 13–15 years) located in larger cities have the greatest exposure (Jonsson et al., 2017). In the past few decades, there has been increased social polarization in Swedish urban areas where different ethnic groups, income levels and forms of housing coexist (Grundström & Molina, 2016). Residential segregation manifests in differences in employment rates, life expectancy and health (Abrahamsson, 2015), and inequality between schools in pupil performance, resources and status (Beach & Sernhede, 2011; Swedish National Agency for Education, 2012). Frustration at the perceived lack of opportunities and respect has fuelled social upheavals (Abrahamsson, 2015), and urban segregation, socio-economic stressors and neighbourhood effects have been related to the frequency of intentionally set school fires (Guldåker & Hallin, 2013, 2014).

In Sweden, comprehensive school fire prevention measures are deployed at the national and local levels, and by public and private actors (Uhnoo, Persson, Ekbrand, & Lindgren, 2015). In 2010, the national government initiated an action plan against school fires (MSB, 2011). However, paradoxical as it may be, the schools’ own fire-prevention strategies received relatively limited attention. The US’ situation is similar, with increased security measures related to firearms and weapons in schools, but less attention to firesetting (Porth et al., 2018). The tendency to deemphasize schools as actors in the prevention of school fires is also reflected in research. Despite the known importance of schools in crime prevention (Brå, 2012; Gottfredson, Cook, & Na, 2012; Lab, 2010), few empirical studies, to our knowledge, have examined their efforts to manage and prevent deliberately set school fires.

In this study, we examine schools both as arenas, in terms of social settings with particular exposure to deliberately set fires, and as actors interpreting and utilizing their organizational scope (Berg, 2007; Gottfredson et al., 2012) to manage and prevent school fires. To pursue this twin line of enquiry, an in-depth analysis of data (interviews, documents and official registry data) related to 20 fire-exposed lower secondary schools in two major Swedish cities (Gothenburg and Malmö) was carried out. The main contribution of this study is descriptive, as it adds further knowledge and understanding of (1) the various forms of fire-related events inside and outside school buildings, (2) how the school personnel interviewed, such as head teachers and caretakers, interpret the causes of school fires and (3) how the individual schools as actors, on the basis of this understanding, actively design, select and reflect upon specific preventive measures.

This is the first study on juvenile firesetting focused on school organizations; it is a unique and urgently needed research perspective given the pivotal but under-researched role of schools in crime prevention. Until now, the sparse literature on deliberately set school fires has focused on perceptions of school firesetting among high school students (Lilja, 2019), the motives reported by school firesetters (Boberg, 2006), convicted school arson offenders (Ekbrand & Uhnoo, 2015; Uhnoo et al., 2015) and the quantitative aspects of the associated socio-structural conditions (Guldåker & Hallin, 2013, 2014). The article proceeds as follows. We begin with a brief literature review. Following a description of our methods, materials and analysis, we then outline and discuss findings. We conclude by summarizing the findings and discussing their implications.

Previous research

Like many other social phenomena, school firesetting is complex and multifaceted, and it results from the interaction of a variety of factors and circumstances (e.g. Ekbrand & Uhnoo, 2015; Uhnoo et al., 2015). Below, we review previous research with an individual-centred perspective on firesetting, on types of school fires and on socio-structural conditions. Finally, we review the literature on school fire prevention.

An individual-centred perspective on firesetting

Firesetting research has been dominated by a psychological and psychiatric perspective, investigating the causes of the phenomenon from an individual-centred perspective and mapping its prevalence in society. This has resulted in many typologies focusing mainly on the motives and individual characteristics of firesetters (e.g. Betsinger, 2006; Del Bove & Mackay, 2011; Doley, Fineman, Fritzon, Dolan, & McEwan, 2011; Mackay et al., 2012). Concerning the fires set specifically by young people, a Swedish study has shown that 80% of youths between 13 and 16 years of age have played with fire at some point (Terjestam & Rydén, 1995). However, a comprehensive Swedish self-report study of 15-year-old adolescents showed that relatively few young people (3%) had set fire to something valuable that did not belong to them. As for their general characteristics, the firesetting youths had lower school grades than their peers, less often viewed school as important to them and less often had teachers they liked. Moreover, they were most often boys (two to four times more often than girls), attracted to typically masculine leisure activities; they demonstrated signs of other problem behaviour and had frequently committed other types of crime as well (Lindgren, Björk, Ekbrand, Persson, & Uhnoo, 2013).

Motives and types of school fires

Studies on the self-reported motives of school firesetters (Boberg, 2006), and the perceptions of school firesetting among high school students (Lilja, 2019), reveal a wide range of motives, such as peer pressure, anger at a teacher, boredom and a desire to be entertained. Furthermore, the students defined school firesetting as a group activity typically done by students who ‘wanted attention, skipped school relatively often or students who had been let down by adults’ (Lilja, 2019, p. 6). Moreover, in the sparse literature on the subject, school firesetters are generally assumed to be pupils or former pupils of the school to which they set fire (e.g. Arson Prevention Bureau, 1998). However, school fires can also be set by people with no connection to the targeted school. In a study of court documents, two main categories of school arson were distinguished; on the one hand, firesetting was directed at the school as an institution, or the organization, pedagogical activities and social environment or milieu represented by the school in question (‘vindictive vandalism’ or ‘obstruction of school activities’), and on the other hand, fires were set by those not related to the school as an institution (‘play vandalism’, ‘destruction of evidence of school burglary’, ‘school fire as a side effect’ or a ‘psychiatric problem’) (Ekbrand & Uhnoo, 2015). In the latter cases, the firesetting could not be straightforwardly linked to pedagogical activities or the social milieu of the schools. However, the design, facility planning and use patterns of the school building(s) after school hours, as well as geographic location, may have significance (Andersson, Dahl, Mårtensson, & Andersson, 2009; Jonsson et al., 2017).

Socio-structural conditions related to fires

Moreover, structural context may also influence firesetting. Research indicates a covariation between arson and structural conditions, such as unemployment and ethnic minority residents (Goebel & Harrison, 2012). Spatio-temporal patterns for arsons in residences, industries and motor vehicles have also been identified (Grubb & Nobles, 2016). Moreover, residential segregation and neighbourhood effects on firesetting have been discussed in an urban Swedish context (Gerell, 2017; Malmberg, Andersson, & Östh, 2013). However, the effects of structural context on deliberate school firesetting in particular have received little attention. One exception is a study finding that intentionally set fires occurred in schools predominantly located within or in close proximity to disadvantaged neighbourhoods characterized by a high proportion of children and adolescents, overcrowded living conditions, economically vulnerable households, low education levels, high levels of unemployment and higher percentages of foreign-born residents (Guldåker & Hallin, 2013, 2014).

Juvenile firesetting prevention

Broadly speaking, prevention of juvenile deliberate firesetting can be divided into two main categories: education, primarily provided by the fire and rescue service, and psychosocial interventions, co-ordinated through the mental health system (Kolko et al., 2008; Palmer, Caulfield, & Hollin, 2007). In addition, a third category can be added, namely technical and situational fire protection approaches, such as sprinkler systems and smoke detectors (see e.g. Wade et al., 2007). Common to the literature on these three categories is that schools, as significant actors in fire prevention, tend to be pushed to the background. One exception is a study suggesting that effective school fire prevention should combine individually focused interventions with situational prevention and efforts to build trust between youth, schools and/or rescue service personnel (Uhnoo et al., 2015). Nevertheless, these recommendations were based on the perspectives of municipality officials and fire and rescue service personnel, and did not consider the school staff members’ own understandings of the school firesetting problem and efforts to manage it.

In summary, existing research has seldom focused on the school as an arena for deliberate firesetting and, to our knowledge, hardly at all on school personnel as actors interpreting and utilizing their organizational scope of action to respond to the problem. This article seeks to fill this gap in the literature, providing a unique perspective from within school organizations on the topic of deliberate school firesetting.

Methods and material

Study design and selection of schools

The study sample consisted of 20 Swedish lower secondary schools situated in large-scale urban settings, namely Gothenburg (Sweden’s second-largest city) and Malmö (Sweden’s third-largest city). The major city context was chosen because in Sweden, the schools most often targeted by firesetters are lower secondary schools located in large cities (Jonsson et al., 2017), and of all the larger cities, the problem has been particularly serious in just the two above-named ones (Blomqvist & Johansson, 2009; Guldåker & Hallin, 2013).

To identify and select so-called hotspot schools (recurrently targeted by fire-setting), we drew on official registry data kept by the Swedish Civil Contingencies Agency (MSB) on emergency fire and rescue service dispatches to school buildings between 2004 and 2013. Ten lower secondary schools from each of the two cities with the highest number of dispatches during the chosen 10-year period were selected. Only incidents registered as ‘started with intent’, ‘children playing with fire’ or ‘fireworks’ were included. The data cover emergency dispatches during both daytime and after school hours. The selection was discussed and validated in a reference group consisting of actors with knowledge of incidents in the two cities under study, representing bodies such as the City Premises Administration in Gothenburg, the emergency fire and rescue services in the two cities, an insurance company owned by the municipality of Gothenburg, and the MSB.

Description of schools in the sample

Official registry data and statistical summaries provided us with relevant background information on the schools in the sample.¹ According to statistics from the Swedish National Agency for Education, almost all of the schools (i.e. 19 out of 20) had fewer pupils who had achieved the nationally specified ninth-grade learning targets than the national average. In many (eight), no more than half or fewer ninth-grade pupils achieved these targets compared with the national average of about three out of every four (77%). In most of the schools examined (i.e. 18 out of 20), the parents’ level of education was below the national average. Moreover, according to reports from the country’s Schools Inspectorate, many of the schools had received criticism for their insecure psychosocial environment. In terms of the geographical location, statistics produced by the two municipalities (Gothenburg City and the City of Malmö) showed that a great majority of the schools in the study were located in the city’s socio-economically disadvantaged areas, with lower educational levels, lower average incomes, higher unemployment, higher proportion of residents from foreign backgrounds and a higher proportion of young people than the average in the respective cities. To conclude, the official statistics indicate the complex challenges that characterize the majority of the schools in the sample. These challenges also became evident during our school visits and in interviews with school staff.

Data and analysis

The study draws on a combination of qualitative data: visits to all schools to observe surroundings, atmosphere and milieu, interviews with key informants and documents. These data were gathered by the authors between April 2014 and March 2015. To supplement these data, we examined the official registry data kept by the MSB on emergency fire and rescue service dispatches to school buildings. This information included the time of the fire incidents, location of the fire (room of origin, inside/outside of the school building), direct cause of the fire, object of origin and severity.

Interviews

To obtain access to the schools and to recruit interviewees, the head teachers were contacted via email with information about the study. They were asked if they wanted to participate in an interview and whether they could help us to contact other school staff, e.g. caretakers, with knowledge of how their school had experienced and addressed the problem of deliberate firesetting. While this led to some scheduled interviews, others were spontaneously conducted during our visits to the schools. A total of 28 semi-structured interviews were conducted (22 face-to-face and 6 by telephone), and the length of interviews varied from about 25 min to 2 h. The majority of the interviews were not audio recorded or transcribed. For unrecorded interviews we took detailed notes, using interviewees’ own words as much as possible, as well as notes on our overall impression of the interview, and typed these up as soon as possible afterwards. The participants completed the approved informed consent process and were guaranteed confidentiality prior to commencing each interview.

Interviews were conducted in 14 out of 20 schools, with strategically selected key personnel such as head teachers (8), caretakers (10) and intendants (1). The head teachers had responsibility for the overall management and day-to-day operations of the schools, while caretakers occupied a strategic position managing the damage caused by fires and were often involved in preventive strategies. Key themes covered during the interviews were: (1) experiences and perceptions of the firesetting problem at the school (including types of incidents and what the interviewees thought had caused the deliberate firesetting) and (2) descriptions of the schools’ firesetting prevention measures (e.g. co-operation with external partners, social prevention, situational prevention and structural prevention). In addition, complementary interviews were conducted with security coordinators at both the school and municipal levels (4), with key personnel at the fire and rescue services in Malmö and Gothenburg (2), a police officer (1), a civil servant at the Swedish Data Protection Authority with knowledge of the regulation of camera surveillance in Swedish schools (1), and a private security company employed by the municipality to enhance the security of public schools (1).

The transcriptions and interview notes were qualitatively analysed for key themes: interpretations and definitions of the problem of school firesetting and accounts of the schools’ preventive work. Because the actual firesetters are hard to identify, and the staff members tend to have limited resources to examine in depth the mechanisms behind specific incidents, their accounts from the interviewees can be seen as partly speculative. However, regardless of whether school staff members’ understanding is based on incomplete information, it is important to consider such information because it guides and legitimizes how the problem is handled by the schools and, thus, has real and concrete consequences for the school environment as well as for the pupils. Put differently, by offering a description of the range of interpretations of the causes and mechanisms behind school fires given by school staff, this study contributes to previous research by providing an enhanced understanding of the rationales behind the preventive measures taken by schools.

Documents

The study also relied on large amounts of documents relating to the 20 schools on the situation and measures undertaken to prevent future fires. First, we collected and analysed documents by the schools, such as data and other material on the schools’ websites, diversity and equal treatment policies, quality reports, codes of conduct (rules and regulations) and any applications for camera surveillance permits. Second, we collected and analysed documents by external actors, such as municipal policy documents (e.g. concerning police reports), reports by the Swedish School Inspectorate, municipal pupil assessments concerning the schools’ psychosocial milieu, and fact sheets and overviews by fire and rescue services. Both types of documents were either downloaded from the home pages of the schools, municipalities or fire or rescue services or sent to us by email.

Overall analysis: matrices for each school

The findings presented below are based on a combined qualitative analysis of interviews, documents and official registry data. The different kinds of data were gathered and organized into one matrix (one document) for each of the 20 schools in the sample to systemize the material, obtain an overview of each individual school and compare schools. In these matrices, a qualitative analysis was performed for each school to assess (1) pupil recruitment, geographical location, school design and how the facilities were used; (2) the specific firesetting problem in each case; (3) the relevant school organizational aspects (such as pupil achievements, psychosocial school environment, educational management, pupil participation, pupil support and structural changes) and (4) the measures (social, situational and structural) used to prevent firesetting. For reasons of confidentiality, matrices for individual schools are not presented. To enhance the validity of the results, the analysis was performed by two researchers in parallel, and different types of empirical materials were combined. The results have also been discussed in a reference group. Some of the results of this study have previously been submitted in a report (in Swedish) to the Swedish Fire Research Board, Brandforsk (Persson & Uhnoo, 2015).

Findings

In this section, we present our findings, focusing on the unique perspective of the school organization. First, we describe the different manifestations of the firesetting problem in the schools examined. Thereafter, the understanding of the firesetting problem by school personnel is discussed in relation to the school setting, the interactive situation, the individual firesetters and the school context. Furthermore, we present our findings on preventive measures taken by the schools in relation to each dimension.

The manifestations of the firesetting problem

The school buildings in our study were exposed to both internal and external deliberate firesetting, during as well as after school hours. The register data on emergency fire and rescue service dispatches to school buildings indicated that, in our sample, two schools were primarily targeted with internal fires during school hours, 11 schools with external fires after school and 7 schools suffered from a combination of both types. However, other material, such as documents and interviews, indicated a substantial number of unrecorded fire incidents at 12 out of 20 schools, emanating mainly from the schools’ tendency to underreport minor internal fire incidents during school hours (cf. Jonsson et al., 2017; MSB, 2011). More specifically, we found that at least 9 of the 11 externally exposed schools had several unrecorded internal fire incidents. In one of several similar cases, registry data indicated two deliberately set internal school fires in a school in one particular year, while complementary data described over 20 such fires in the same period.

According to our interviews, documents and registry data, fire-related incidents inside the school buildings during school hours can be divided into five main types: toilet fires, corridor fires (typically in trash cans), locker fires, intentional false-alarm fires and fires set with firecrackers or other types of fireworks thrown, shot or set off inside the school building. This corresponds well with a previous study on firesetting in Swedish schools, which identify incidents related to firecrackers as a recurrent problem, and toilets and common areas (e.g. stairwells, corridors, small assembly areas and entrance/locker rooms) as the spaces most commonly targeted by daytime school fires (Blomqvist & Johansson, 2009). Research has estimated that only one half of all the school fires break out during school hours when there are both pupils and school staff present in the area (MSB, 2011). The other half are, thus, committed after school hours (i.e. after 5.00pm on weekdays or during weekends and holidays) and can take place either inside or outside the school building.

The main types of fire-related incidents after school hours identified in our study were fires started outside in the schoolyard or close to the school building’s facade (e.g. by setting fire to small vehicles or objects left in the schoolyard, such as construction materials, toys, rubbish or other types of combustibles), fires set in connection with a break-in at the school and fires set by throwing burning or explosive objects into the school building (e.g. Molotov cocktails or fireworks). Prior studies have shown that the most severe and costly school fires generally start after school hours (Blomqvist & Johansson, 2009; MSB, 2011), outside the school building and close to the facade (Johansson et al., 2012). Furthermore, research has indicated that while school fires set during the school hours are often committed by pupils of the targeted school, school fires outside school hours are frequently set by a person or persons with no direct connection to the building or the institution targeted (Ekbrand & Uhnoo, 2015).

The school as an institutional setting: boredom aversion and dysfunctional school environments

The school personnel in our study discussed firesetting as manifestations of boredom aversion and dysfunctional school environments, containing elements of resistance. These types of firesetting can arguably be understood in relation to mandatory school attendance (in Sweden for pupils up to 16 years of age). Juveniles, thus, spend their weekdays in school, not primarily because they want to, but as a result of coercion. Institutional aspects, such as mandatory attendance and additional demands on adaption from formal authorities (e.g. teachers), may motivate resistance practices from pupils disidentifying themselves with the educational system (Willis, 1993). Juvenile firesetting is often presumed to be related to boredom aversion (Boberg, 2006; Lilja, 2019; Perrin-Wallqvist & Norlander, 2003), just like other types of juvenile crimes (Cohen, 1955). This was also a recurrent assumption among the school staff interviewed; ‘It was the only thing they could come up with’, as one caretaker put it. The pupils behind the fires were said to ‘crave some drama’, to want to ‘cause a bit of commotion’ or ‘make something happen’ that could break the monotony of their everyday life at school. According to one head teacher, the youngsters starting school fires were pupils who often shirked lessons but would, nevertheless, ´linger around with nothing particular to do, and then you get these ideas that you might just as well set the wastebasket afire, at least to make something happen: the fire department might show up, and that would be kind of cool´.

None of the interviewees in our study believed the purpose of the intentional daytime fires and firecracker incidents in their school was to burn down the school. Rather, they suggested that pupils started school fires to have the school building evacuated and to disrupt school activities (cf. ‘obstruction of school activities’ in Ekbrand & Uhnoo, 2015), and in that way to skip lessons or avoid sitting a particular test. As research on school arson from a Kenyan context puts it, firesetting can, thus, be a way to get a break from a school setting felt by pupils to be a ‘prison’ (Cooper, 2014).

The schools in our study handled intentional fire alarms with very different and contrasting strategies. One strategy was, as far as possible, to avoid evacuations of the school building because this was assumed to entice pupils to trigger even more false alarms and set more fires. One head teacher explained that in her school the students ‘will not get the pleasure’ of evacuating the school ‘15 times per year’. According to this point of view, if pupils intentionally triggered alarms to avoid a test or a lesson, and their actions resulted in evacuations, the pupils may perceive this as the accomplishment of their goals. The opposite strategy was to make sure to always evacuate if the fire alarm sounded, either to mark the potential severity of fires or to put social pressure on the individuals responsible. This would have a preventive effect, argued one head teacher, because regular evacuations were found to be burdensome by the majority of pupils and teachers, and the minority of pupils behind the intentionally set alarms would not become ‘so popular afterwards’.

In addition, some of the interviewees held dysfunctional school environments responsible for prior fire incidents, involving larger groups of pupils who had also engaged in other kinds of vandalism and transgressive behaviour at school. Youths are aggregated in schools in their peak year of delinquency (Gottfredson et al., 2012), and research suggests (Guldåker & Hallin, 2014) that fire setting can constitute a conflict strategy – a protest against societal institutions. For example, a caretaker reported that fire incidents at his school had been accompanied by murder threats against the school management. In another school, there had been more than 20 incidents of school fires and physical abuse of teachers all in the same year. In a third school, according to news articles, a period of protesting and rioting by pupils, including several incidents of toilet and locker fires, led to the temporary closure of the school.

Prevention strategies associated with the school as an institutional setting

The head teachers, who were responsible for the overall management and day-to-day operations of the schools, had a perspective on school fire prevention clearly marked by their overall understanding of the school’s institutional goals: to educate pupils on certain knowledge, skills and values. They largely understood deliberate firesetting to be part of a wider context and problem, a view also supported by research on the subject (Mackay et al., 2012). A great majority of the school organizations faced comprehensive educational and social challenges, such as many low-performing pupils and, sometimes, insecure psychosocial school environments. Interestingly, the head teachers stressed that best results were obtained from general interventions integrated into the schools’ overall work. These were intended to develop the school’s pedagogical work to support pupils’ learning in a better way, to improve the school’s psychosocial environment and to strengthen social relationships. Such measures are also supported by research, in that poor school performance and attachment to school often precede delinquency in general (Gottfredson, Wilson, & Najaka, 2002), and seem to be positively related to firesetting (Lindgren et al., 2013). Examples of social prevention were deliberately decreasing enrolment to make the school smaller, reorganizing the school in a manner that made it easier to foster closer relationships between teachers and pupils, increasing pupils’ participation and influence in school activities as well as increasing the adult presence in the school. We argue that such interventions may be expensive but are uncontroversial because they are based on a fundamentally positive view of the pupils (Ullucci & Howard, 2015).

Based on the limited attention in research and policy on schools as preventive actors, an important conclusion from the study is that the schools mobilized an extensive and broad repertoire of measures. The social prevention was intended to develop pupils’ awareness, knowledge, norms and values and to foster their capabilities (cf. Brantingham & Faust, 1976; Sahlin, 2008). The measures could be general in character and scope (e.g. increasing pupils’ involvement in their education) or specifically target intentionally set school fires (e.g. providing general fire safety training). They could focus on all the school’s pupils (e.g. interventions to promote a sense of responsibility for the physical environment) or on those pupils deemed to be the most at-risk group for becoming firesetters (e.g. extended contact with the fire and rescue service). Prevention could also address individual pupils identified with firesetting behaviour (e.g. contact with student health and police) (See Table 1).

Table 1. General and school fire-specific social interventions to prevent firesetting in the studied hotspot schools.

	General	Specific
Primary social interventions (all pupils)	Individual: Enhancement of pupil performance through homework support, individualized instruction, etc. Enhancement of physical and psychological well-being through free breakfasts at school, school health services, etc. Relationships: Strengthening of relationships through school cafés run by pupils, supervised recesses, teaching of ethics and values, anti-bullying work. Pupil participation and influence: Class council, student council, involvement of pupils in the shaping and maintenance of their learning environment. Constructive recess activities: Stimulating outdoors environment, library access, computer lab, organized recess activities.	Sense of responsibility for physical environment: Class/form rooms, class toilets, pupils’ involvement in the shaping and improvement of the school’s physical environment, pupils’ involvement in the maintenance and upkeep (re-painting, renovation) of school facilities when needed. General fire safety training of pupils by school staff or fire safety officers. Presence of fire and rescue services at the school and youth club. Fire drills for pupils and school staff.
Secondary social interventions (at-risk pupils)	Student health: Co-operation between school management, special teachers, special needs teachers, school nurse, school doctor, school counsellor, study counsellor, school psychologist. Action programme: Individually tailored programmes for pupils. Group interventions for at-risk youth. Temporary (special) classes. Adult presence at school during daytime.	Temporary/short-term measures: Sessions on fire safety, school visits by fire and rescue services and/or fire victims. Extended/long-term measures: More extended presence of fire and rescue services at schools, youth clubs etc.
Tertiary social interventions (pupils with firesetting history)	Student health. Contacts and dialogue with parents and the pupil e.g. by the school, police, social services. Re-assignment of pupils within school or to another school, suspension, special day school. Police report, legal action.	Student health. Contacts and dialogue with pupils by fire and rescue services or with pupils and their parents by social services. Re-assignment of pupils within school or to another school, suspension, special day school. Police report, legal action.

The interviewees expressed more ambivalence towards strategies directly targeted at firesetting, such as inviting the fire and rescue service to instruct pupils about fire risks and safety and consequences of firesetting. Both head teachers and caretakers expressed concerns that such instructions may trigger further firesetting. In one school, a caretaker recalled this happening during a fire education session when one of the pupils left the room only to set a toilet on fire. Targeted fire education and interventions can take different forms and be received differently by pupils, but measures intended to deter youths from further delinquency or ‘scare them straight’ have generally been shown to be counterproductive (Petrosino, Turpin-Petrosino, Hollis-Peel, & Lavenberg, 2013).

Firesetting as a situated activity: fluctuating, contagious, group-related and situational

The problem of school firesetting was also understood to be fluctuating, contagious, group-related and triggered by situational factors. Caretakers, who had often worked for many years at fire-exposed schools, often described school fires as being highly fluctuating or as ‘coming in waves’ (cf. Grubb & Nobles, 2016). In some cases, the registry data and the school personnel described notably intense periods with several fires lit even during the same day or week. Moreover, the interviewees spoke of the connection between school firesetting behaviour and certain group processes, i.e. actions and events leading to a fire being started could suddenly be ‘set in motion’ and just begin to ‘unfurl’. One caretaker also spoke of ‘contagion’ and described how firesetting behaviour, or a firesetting technique, could suddenly become ‘trendy’ and ‘epidemic’. Another caretaker provided a concrete example of pupils teaching other pupils a specific firesetting technique, namely lighting a piece of paper and inserting it into someone’s locker.

Just like juvenile firesetting in general (Uhnoo, 2015), juvenile school fires are often set in groups (Ekbrand & Uhnoo, 2015). In this study, the interviewees argued that unlike toilet fires, firecrackers and other types of fireworks were typically thrown, fired or otherwise set off by groups of boys. Group firesetting was commonly attributed to group dynamics, as a prank perpetrated to have some ‘fun’ together or with a desire to provoke reactions; for example, youngsters who want to ‘be seen and heard’ tried to impress their peers with how ‘daring’, ‘tough’ and ‘cool’ they could be (cf. Boberg, 2006; Lilja, 2019).

Furthermore, the school staff also proposed, what can be termed, ‘situational’ explanations. These centred on the role of so-called foreground factors (cf. Katz, 1988) or situation-specific aspects to explain what triggered firesetting in a specific situation. For example, absence of adult control was identified as one crucial factor, often meaning that fires had been set in a secluded and, thus, little monitored corridor or toilet in the school building or a schoolyard obstructed from the view of neighbours or passers-by. The situation could also be pupils carrying lighters to school, the presence of overfilled waste bins or combustible material, such as construction materials stored in a visible place on school grounds or easy access to firecrackers and fireworks. Firework shooting and firecracker throwing were clearly related to Christmas and New Year celebrations. These start in November every year and calm down only when all the fireworks are consumed, a head teacher explained.

Prevention strategies associated with firesetting as a situated activity

Situational crime prevention aims to reduce criminal opportunities by tackling the situations in which crime occurs, using physical, environmental or technical measures to change them (e.g. Felson & Boba, 2010). Because firesetting, firework shooting, firecracker throwing and intentional false-alarm triggering is inexpensive, easy to implement and the perpetrators take a minimal risk, the aim of situational measures in the schools examined was to make such activities more difficult, riskier and less rewarding.

Since toilets are particularly exposed to firesetting, situational prevention measures to prevent internal firesetting were often adopted for toilets by minimizing the amount of easily lit materials (e.g. overfilled waste bins) or requiring a key/pass to control access. A problem related to the latter strategy mentioned in the interviews was that pupils sometimes forgot to close the door behind them. An open toilet door in a school with strict access control to the toilets may trigger firesetting because it tempts a ‘beat the system’ effect, motivating pupils to outsmart the school board’s efforts to control and surveil them. Another situational intervention aimed at making firesetting riskier was to increase the adult presence. For example, a caretaker told us how he put a stop to a period of intense firesetting by moving his office to a secluded corridor particularly exposed to fires. Another successful measure the interviewees talked about was promoting a greater sense of responsibility among pupils towards their physical school environment, for instance, by giving them more say in how that environment was shaped and maintained and assigning classes their own classrooms or toilets.

Besides target hardening, in the form of locks and alarms, several of the situational interventions to prevent external fires involved action to minimize access to combustible materials and otherwise reduce opportunities for vandalism. One head teacher said:

We try to make sure that there is no debris or combustible materials stored against the building walls, always double checking this before the holidays and so on. […]. There’s no need to tempt fate.

Arguably, some situational interventions – although regarded as effective and well-motivated in certain cases – are more problematic, for example, providing passes to control access to toilets. Interventions to limit access to combustible materials – paintings, curtains, comfortable and functional seating inside the building – could have negative consequences for the school environment, making it less attractive and stimulating for pupils. During our school visits, it became obvious that some of these measures had resulted in austere and uninspiring surroundings and interiors, including heavy-duty vandal-proof benches and tables permanently affixed to the ground or the building walls, and minimal aesthetic input in the form of decoration. In these schools, the areas frequented by the pupils, thus, contrasted starkly with the look of the cosy staff rooms decorated with sofas, curtains and indoor plants.

The individual firesetters

The school staff also discussed the firesetting problem in terms of who had set the fires. It was proposed that when the firesetters, at some point, left the school, the problem would also go away. As one of the head teachers described it, the situation at the time was ‘calm’ at the school, although ‘it would only take one lost soul’ for the problem to resurface. In some cases, the people who set the fires were identified by surveillance cameras and in other cases, the firesetters had confessed to the offence. However, the interviewees frequently stated that they did not know exactly who started the fires in their school but that they often had strong suspicions. These could be based on observations, such as the fact that certain pupils ‘are always right there, nearby, when the fires start’. However, they could also be based on perceptions about which pupils in the school most often engage in transgressive behaviours; these pupils could then be described by interviewees in terms such as ‘misbehaving kids’ who ‘stick out’, ‘wreak havoc’ or ‘constantly brawl’. According to one caretaker, it was always the same set of individuals behind all acts of vandalism in the school, from water damage and spray-painted graffiti to broken windows, false alarms and actual fire incidents. This corresponds well with research indicating that juvenile firesetting behaviour is often linked to other types of delinquent behaviour (Uhnoo et al., 2015).

At the same time, the firesetters could also be pupils that the school staff never expected to demonstrate such behaviour. One repeatedly discussed perpetrator group was the quieter individuals, both boys and girls, who were ‘not doing well’ because of their general life situation, pupils who were ‘lost’, ‘frustrated’ or ‘needing help’. These individuals tended to act alone and typically set fires to toilets. In the literature, juveniles in this category are referred to as ‘cry for help firesetters’ (Slavkin & Fineman, 2000). A third group discussed by the school staff consisted of pupils described as particularly high-risk firesetters, owing to a psychological problem or so-called specific difficulties that made them more prone to firesetting than others. According to one head teacher, when there had been an earlier series of toilet fires in her school, the school staff ‘knew very well which pupil it was; this person was fascinated by fires’. This particular pupil was transferred to another school only to continue his firesetting behaviour.

Prevention strategies related to individual firesetters

Instead of targeting specific sub-groups or certain individuals, the preventive work often targeted the school pupils in general. As one head teacher said:

I think what works best in all sorts of cases is that you make sure you have a good, well-functioning school. And that you pay attention to the pupils who are not doing that well: offer them a possibility to be heard, to be taken seriously, to become part of the joint process of finding solutions to problems. And that you do that before they maybe end up feeling like they have to set things on fire to call attention to themselves.

The overall work to create a good, well-functioning school was, thus, presumed to result in fewer pupils setting fires. General measures were applied when the firesetters were not identified, which often was the case. However, the school also used strategies to identify firesetters in order to intervene by involving student health professionals, initiating contact with parents, or suspending or re-assigning pupils (within the school or to another one) or taking legal action.

Another prevention strategy, used by the majority of schools (17/20), was the installation of indoor and/or outdoor security cameras. According to information received from the authorities responsible for camera surveillance licences, this practice has rapidly become more common in Sweden, mirroring the development in recent years in countries such as the United Kingdom (cf. Taylor, 2010). However, an interesting finding was that while 8 out of 10 schools in Malmö had installed indoor cameras, none of the schools in Gothenburg had. On the contrary, head teachers in Gothenburg expressed strong disapproval of surveilling their pupils during school hours. One explained that he did not want to ‘stoop so low’. However, in Gothenburg, outdoor cameras were installed in 7 out of 10 schools, while only 4 out of 10 schools in Malmö used outdoor video surveillance. The interviewees thought of video surveillance as a potential strategy to curb firesetting during a particularly difficult period by exposing those responsible for it. However, more prolonged or permanent use may be problematic, considering its negative impact on privacy (cf. Taylor, 2010). Only a few schools seemed to have analysed the impact of video surveillance from a before-and-after perspective, and even fewer appeared to have removed the cameras based on any observed drop in firesetting frequency following their installation during a peak period. In this sense, one could, thus, justifiably speak of a degree of normalization of exceptional measures (cf. Flyghed,2000).

The area surrounding the school – socio-structural factors related to firesetting

Schools in poor, urban areas tend to experience more disorder than other schools (Gottfredson et al.,2002). Intentionally set fires in Malmö occur in schools predominantly located within or in close proximity to disadvantaged neighbourhoods (Guldåker & Hallin, 2013, 2014), and in Gothenburg, in areas with lower levels of education among adults and higher proportions of foreign-born people and of failed grades (Lindgren et al., 2013). Similarly, the schools in our study were located in the cities’ socio-economically disadvantaged areas, with lower educational levels, lower average incomes, higher unemployment, higher proportion of residents from foreign backgrounds and a higher proportion of young people than the city average. This was reflected in discussions with school personnel of the relationship between social conditions in the surrounding areas, in most cases the schools’ primary catchment areas, with the fires. For example, one head teacher spoke about the life circumstances of young people in his catchment area, describing major social difficulties, language problems, lack of access to after-school leisure activities and cramped living conditions. According to the head teacher, the latter caused young people to avoid staying at home and linger around the school building, or even inside it, after their school day was over. Such unorganized spare time without adult supervision can be a risk factor for delinquency (Wikström & Treiber, 2016). The school atmosphere was also impacted by the social situation in the surrounding area, such as the presence of street gangs and the extent of their influence. School crimes tend to be closely linked to community crimes (Gottfredson et al., 2012), and one head teacher explained that one after-hours firesetting case in her school had proven to be about an ‘entrance exam’ to a criminal network operating in the area. Another example recounted by a head teacher concerning the influence of the surroundings was when a school staffroom was firebombed through the window one evening at the same time as several other buildings in the area were vandalized (cf. Grubb & Nobles, 2016), apparently all by one group of youths.

Structural prevention

Structural prevention aims at transforming the schools’ internal and external conditions in a more dynamic and profound way (cf. Guldåker & Hallin, 2013, 2014; Sahlin, 2008). Schools have organizational leeway but are also affected by structural conditions in fundamental ways, such as by residential and school segregation (Gottfredson et al., 2002; Ullucci & Howard, 2015). The head teachers interviewed for our study described different forms of structural prevention related to the following factors: collaboration with local actors (e.g. with social service and police to prevent and manage transgressive behaviour and social unrest in the school and neighbourhood); compensatory economic resources from municipalities and the state to develop the individual school; policy reforms (e.g. local policies about firecrackers and the Education Act about disciplinary issues) and major reorganizations and alterations to improve school environments (e.g. change pupil recruitment patterns that contribute to school polarization, limiting the number of pupils in lower secondary school and rebuilding the school).

Concluding discussion

This study set out to provide new insights into schools as arenas of youth firesetting and as actors in managing and preventing the phenomenon. Empirically, this article is based on an in-depth analysis of 20 lower secondary schools frequently targeted with deliberate firesetting in the cities of Gothenburg and Malmö, Sweden. The focus is on school organizations – a unique and urgently needed perspective within the research field. Until now, the sparse literature has focused on the quantitative aspects of socio-structural conditions related to school fires (Guldåker & Hallin, 2013, 2014; Jonsson et al., 2017), perspectives of pupils (Lilja, 2019), on the motives of school firesetters (Boberg, 2006) and those of convicted offenders (Ekbrand & Uhnoo, 2015). In addition, school fire prevention has been explored from the perspective of municipality officials and fire and rescue service personnel (Uhnoo et al., 2015). Hence, despite the known potential and importance of schools in crime prevention (Gottfredson et al., 2012), few empirical studies, to our knowledge, have explored their own efforts to manage and prevent deliberately set school fires.

Limitations

This study comes with a couple of limitations that may inform future research. First, to select the most fire-exposed schools, we draw on the MSB’s register data consisting of dispatches by the emergency fire and rescue service to schools. However, our finding shows that many schools manage minor incidents themselves, resulting in extensive dark figures. This under-reporting affects the reliability of the data, so the selection of schools should be considered with some caution. Furthermore, our conclusions are based on a qualitative study in Sweden. Whether our findings are transferable to other schools, cities or country contexts needs to be investigated further in future research. A third and final limitation is that the school staff members’ interpretations of the mechanisms behind school firesetting must be considered with due caution. However, we argue that by offering these descriptions grounded in the interviewees’ local and often shared experiences of fire incidents, this study contributes to an enhanced understanding of the rationales behind the preventive measures.

Main findings

In spite of their limitations, the findings of this study based on a combined analysis of interviews, registry data and documents contribute to a deeper understanding of fire-related events inside and outside of school buildings, interpretations by school personnel of the causes of school fires and how school organizations, on the basis of this understanding, design, select and reflect on preventive measures.

First, this study identifies concrete manifestations of the school firesetting problem, adding an elaborated understanding of the phenomena to previous research (i.e. Blomqvist & Johansson, 2009; Ekbrand & Uhnoo, 2015). In our study, fire-related incidents inside the school buildings during school hours – often committed by pupils of the targeted school – were of five main types: fires in toilets, corridor fires, locker fires, intentional false-alarm fires, and firework shooting and firecracker throwing. Fire-related incidents after school hours (outside as well as inside the school) involved fires started in the schoolyard or close to the school building facade, fires set in connection with a school break-in and those set by throwing burning or explosive objects into the school building. Research indicates that these fires outside school hours are frequently set by a person or persons with no direct connection to the building targeted (Ekbrand & Uhnoo, 2015).

Second, our findings show the complex and diverse ways that school personnel, in particular head teachers and caretakers, interpreted the mechanisms behind deliberate school firesetting. These included expressions of boredom aversion and dysfunctional school environments (e.g. disruption of school activities and rioting), possibly linked to institutional dimensions within the school setting, such as mandatory attendance and demands or coercion from formal authorities such as teachers (Willis, 1993). Additionally, deliberate firesetting was framed as a situated activity: fluctuating (‘coming in waves’), contagious (e.g. particular firesetting techniques suddenly becoming popular), group-related (typically fireworks shooting and firecracker throwing as a prank or to provoke reaction) and situational (triggered by factors such as an absence of adult control, overfilled waste bins, an open toilet door or easy access to fireworks). Moreover, the school staff framed the firesetting problem in terms of who had set the fires. Notwithstanding that the firesetters behind the incidents were seldom identified, three main perpetrator groups were discussed: pupils engaged in various types of transgressive behaviours (typically engaged in intentional false-alarm fires, firework shooting or firecracker throwing), quiet individuals, who were ‘not doing well’ and ‘needed help’ (typically acting alone and setting fires to toilets) and particularly those at high risk owing to psychological factors (e.g. an extreme interest in fires). In addition, because the great majority of the schools were located in socio-economically disadvantaged areas, the school personnel related cramped living conditions and social unrest to the schools’ exposure to fires.

Third, our most interesting finding, given the essential but under-researched role of schools in crime prevention, was that the fire-exposed schools use their organizational leeway to mobilize a broad repertoire of measures of social, situational and structural character to prevent fire incidents. In line with policy and general research on schools and crime prevention (MSB, 2011, pp. 26–29; Gottfredson et al., 2012), the head teachers interviewed placed special stress on the importance of social prevention, integrating firesetting prevention into regular work on learning and the school’s psychosocial environment. Instead of targeting specific sub-groups or certain individuals, preventive work was often directed towards all pupils, with the ambition of creating an overall well-functioning school. We argue that such an approach is productive because it is intended to support pupils (cf. Ullucci & Howard, 2015). Furthermore, the schools used situational measures, such as security cameras, to make fire incidents more difficult, riskier and less rewarding. While some situational interventions are rather uncontroversial, such as restricting access to firecrackers and increasing adult presence, other interventions can be more problematic, such as providing passes to control toilet access and limited access to combustible materials like curtains, and comfortable and functional seating inside the buildings. A combination of video surveillance and a vandal-proof environment may symbolize and communicate distrust to the pupils. Moreover, structural measures were aimed at transforming the schools’ internal and external context and dynamics in a broader sense. Examples of such measures include collaboration with external local actors (police and social service), reorganization of schools (rebuilding them or shutting parts down), policy changes (local and national) and reallocation of economic resources. Schools can play a part in such processes, but major structural transformations go beyond the scope of the individual school organization (cf. Ullucci & Howard, 2015).

To conclude, school fires cause significant educational, social, material and economic damage. In the past few decades, Sweden has experienced increased residential segregation in urban areas, coupled with growing disparities in pupil performance, resources and status between schools (Beach & Sernhede, 2011; Swedish National Agency for Education, 2012). The majority of the schools examined were located in socio-economically disadvantaged areas in Malmö and Gothenburg. Therefore, it can be assumed that broader mechanisms influence the phenomenon of deliberate school firesetting (Guldåker & Hallin, 2014), and that one of several expressions of school segregation may be recurrent exposure to deliberate firesetting. It is well known that growing up in poor families and socio-economically disadvantaged areas impacts pupils in numerous ways (Ullucci & Howard, 2015) and increases the overall risk of committing or being a victim of crime (cf. Pratt & Cullen, 2005; Wikström & Treiber, 2016). Taken together, our findings suggest that exposure to school firesetting has numerous potential negative implications, not only insecure school environments, disruptions of school activities owing to evacuations, vandalization of interiors and exteriors, financial burdens on school budgets, but also – paradoxically as side effects of well-intentioned efforts – austere and uninspiring school environments as well as control, surveillance and distrust of pupils, potentially contributing to experiences of stigmatization and marginalization (cf. Wacquant, 2007). Further research into the pivotal role of school organizations in confronting such serious challenges in an effective, yet ethical, manner is, thus, crucial.

Acknowledgements

We are very grateful for the extensive and constructive comments from two anonymous reviewers and the editor.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Swedish Fire Research Board through a grant awarded to Sofia Persson and Sara Uhnoo, University of Gothenburg, Sweden.

Notes

1. To provide background information of the lower secondary schools in the sample, we drew on official data and statistical summaries collected by the Swedish National Agency for Education (Skolverket) on pupil recruitment and pupil performance for each school in the year 2014 (from the SIRIS database: https://siris.skolverket.se/siris/f?p=SIRIS:164). Furthermore, we drew on official registry data and statistical summaries (‘Area facts’) produced by the two municipalities (Gothenburg City and the City of Malmö) describing the local context of each of the 20 schools in the year 2012, e.g. income level (average income), education levels (proportion of the population with post-secondary education), proportion of foreign-born residents or those born in Sweden with both parents foreign born, unemployment rate and proportion of young people. Both types of registry data are accessible to the public and have been downloaded or collected from the authorities’ websites. The descriptions of the schools in the study are based on statistical summaries provided by the municipalities and the Swedish National Agency for Education, and we have not conducted any statistical analyses of our own based on original quantitative data (Persson & Uhnoo, 2015).