Measuring peer victimization and school leadershipA study of definitions, measurement methods and associations with psychosomatic health

Lisa Hellström

Lisa Hellström

| Measuring peer victim

ization and school leadership | 2015:15

Measuring peer victimization and school leadership

The negative consequences of peer victimization on children and adolescents such as worsening academic achievement and mental ill health are major public health concerns which have been subjected to extensive research. However, there are long-standing concerns how to define, measure, and estimate prevalence rates of peer victimization and successful school leadership. The aim of this thesis is to study methods for assessing peer victimization and pedagogical leadership in school. The results show that excluding other forms of peer victimization than bullying have serious implications for the identification of victims and may underestimate the full impact of peer victimization on children. Further, the validation of the Pedagogical and Social Climate (PESOC-PLP) scale is a step towards ensuring valid assessments of pedagogical school leadership. By strengthening the understanding of measurement methods of peer victimization and school leadership the results from the thesis may indirectly contribute to a safe and positive school experience for children and adolescents.

DISSERTATION | Karlstad University Studies | 2015:15 DISSERTATION | Karlstad University Studies | 2015:15

ISSN 1403-8099

Faculty of Health, Science and TechnologyISBN 978-91-7063-626-4

Public Health Science

DISSERTATION | Karlstad University Studies | 2015:15

Measuring peer victimization and school leadershipA study of definitions, measurement methods and associations with psychosomatic health

Lisa Hellström

Print: Universitetstryckeriet, Karlstad 2015

Distribution:Karlstad University Faculty of Arts and Social SciencesCentre for Research on Child and Adolescent Mental HealthSE-651 88 Karlstad, Sweden+46 54 700 10 00

© The author

ISBN 978-91-7063-626-4

ISSN 1403-8099

urn:nbn:se:kau:diva-35192

Karlstad University Studies | 2015:15

DISSERTATION

Lisa Hellström

Measuring peer victimization and school leadership - A study of definitions, measurement methods and associations with psychosomatic health

WWW.KAU.SE

1

ABSTRACT

Background: The negative impact of peer victimization on children and adolescents is a

public health concern and the subject of extensive research. However, there are long-

standing concerns about how to accurately measure and estimate prevalence rates of peer

victimization and whether children and adults define and understand bullying in the same

way. Given that prevalence rates of peer victimization are needed to evaluate mental

health interventions at an individual and school level, methods for estimating the number

of children experiencing peer victimization are crucial. The Swedish Education Act clearly

states that schools should work actively with social, civic, as well as academic objectives

and that the principal is responsible for the preventive and anti-bullying work in school.

PESOC is a Swedish instrument intended for studies on school effectiveness and

pedagogical leadership and has been used for school improvement efforts in Sweden. The

psychometric properties of PESOC have not been examined previously. To get a clearer

picture of the full extent and distribution of the social and mental health consequences of

peer victimization, this thesis explores the methodological ambiguities concerning peer

victimization and school leadership.

Aim: The overall aim of this thesis is to examine methods for assessing peer victimization

and pedagogical leadership in school. The following research questions are asked: 1) To

what degree does a measure of bullying overlap with measures of peer aggression with

respect to the number of students identified as victims and association with self-reported

psychosomatic problems? 2) What are Swedish adolescents’ definition and understanding

of bullying? 3) Can the PESOC-PLP scale be used as a valid measure to assess the

pedagogical leadership of the principal?

Method: This thesis is based on four studies. Study I and II are based on a web-based

questionnaire among compulsory school students. The data were collected in a classroom

setting in the city of Karlstad at two points in time, in 2009 and 2010. Altogether, 2,568

adolescents in school years 7-9 (aged 13-15) participated. Concordance and discordance

between a measure of bullying and a measure of peer aggression were analyzed using

contingency tables and measures of association (phi squared coefficients). Linear

2

regression analyses and analysis of variance (ANOVA) were applied to analyze the

association between psychosomatic problems, bullying and peer aggression. In study III,

data were collected among students in school years 7 and 9 with a web-based

questionnaire and focus group interviews. 128 students completed the questionnaire while

21 students participated in the focus groups (four focus group interviews in total). The

quantitative data were analyzed using contingency table analysis and Chi-square

significance tests. The qualitative data were analyzed using qualitative content analysis. In

study IV, a total of 344 teachers and special-education teachers were asked to rate their

principal’s leadership skills using a web-based questionnaire distributed via e-mail. The

psychometric properties of the PESOC-PLP scale were analyzed using Rasch analysis.

Results: Study I: Among the students reporting being victimized by peers, 13% reported

bullying, 44% reported repeated peer aggression and 43% reported being both bullied and

exposed to peer aggression. Concordance was further elucidated by phi square coefficient

tests revealing that 18% of the variance in either measure was accounted for by the other

measure. The results showed that a measure of bullying and a measure of peer aggression

capture partly different students. Study II: The results show that the proportion of

adolescents with psychosomatic problems is the same regardless of which measure is

used. The two measures are not capturing exactly the same individuals and the proportion

of adolescents with psychosomatic problems related to peer victimization is higher when

measures of bullying as well as measures of peer aggression are included simultaneously in

the analyses. Study III: The adolescents’ understanding and definition of bullying did not

only include the traditional criteria of repetition and power imbalance, but also a criterion

based on the health consequences of bullying. That is, a single but hurtful or harmful

incident could also be considered bullying irrespective of whether the traditional criteria

were fulfilled or not. Further, girls and older students had a more inclusive view of

bullying and reported more types of behavior as bullying compared to boys and younger

students. Study IV: The Rasch analysis indicated two sub dimensions of the PESOC-PLP

scale and analysis of item content revealed two types of leadership: direct pedagogical

leadership and indirect pedagogical leadership. One item showed misfit according to the

Rasch Model and was removed without loss in content validity. The analysis of the sub

dimensions showed satisfying psychometric properties of the instrument, indicating that

3

the PESOC-PLP scale can be useful when measuring the pedagogical leadership of

principals.

Conclusions: Ensuring a safe and positive school experience for children is of great

importance as it is a good predictor of future health. All children have the right to express

their opinions regarding matters that concern them and allowing children’s voices to be

heard are crucial as their understanding may not always conform to the understanding of

adults. This thesis highlights the deficiencies in the way bullying and school leadership are

currently defined and measured. A single measure of bullying, including the traditional

definition, does not fully capture the overall prevalence rate of peer victimization and its

association with psychosomatic problems, nor does it fully reflect adolescents’

understanding and definition of bullying. In order to incorporate adolescents’ perception

into the definition of bullying, the negative health consequences should be included as a

criterion and not only be seen as a consequence of bullying. Further, by showing sound

psychometric properties, the PESOC-PLP scale adds to the research field of pedagogical

leadership. In Sweden and other countries that prioritize comprehensive head teacher

training programs, the PESOC-PLP scale may be used as a tool for evaluating these

programs. By strengthening the understanding of measurement methods of peer

victimization and school leadership the results from the thesis may indirectly contribute to

a safe and positive school experience for children and adolescents.

4

SAMMANFATTNING

Bakgrund: De negativa konsekvenserna av peer victimization1 utgör ett stort hot mot

folkhälsan och är också ett viktigt forskningsområde. Det finns emellertid många

frågetecken om hur man bäst mäter och beräknar förekomst av peer victimization och om

hur barn och vuxna definierar och uppfattar mobbning. Metoder för att mäta förekomst

av peer victimization behövs för att kunna utvärdera interventioner som syftar till en ökad

psykisk hälsa på individ-och skolnivå. Enligt den svenska skollagen ska varje skola aktivt

arbeta med sociala, demokratiska samt akademiska mål och huvudmannen har ansvar för

förebyggande åtgärder mot mobbning i skolan. PESOK är ett svenskt mätinstrument som

har använts i utvecklings- och förbättringsarbetet på svenska skolor med syfte att

identifiera och studera effektiva skolor och pedagogiskt ledarskap. De psykometriska

egenskaperna hos PESOK har dock inte tidigare studerats. För att få en tydligare bild av

den fulla omfattningen och de sociala och psykiska konsekvenserna av mobbning är målet

med denna avhandling att studera peer victimization och ledarskap i skolan med ett fokus

på metod frågor.

Syfte: Det övergripande syftet med denna avhandling är att utforska metoder för att mäta

peer victimization och ledarskap i skolan. Följande forskningsfrågor undersöks: 1) I vilken

utsträckning överlappar ett mått på mobbning och ett mått på aggression med avseende

på antalet elever som identifieras samt sambandet med självrapporterade psykosomatiska

besvär? 2) Hur definierar svenska ungdomar mobbning? 3) Kan PESOC-PLP skalan

användas som ett validerat instrument för att mäta rektorns pedagogiska ledarskap?

Metod: Avhandlingen bygger på fyra delstudier. Studie I och II bygger på en webbaserad

enkät bland grundskoleelever i Karlstads kommun. Enkäterna samlades in på skolor vid

två olika tidpunkter, 2009 och 2010. Sammanlagt deltog 2568 ungdomar i åldrarna 13-15

år (årskurs 7-9). Överensstämmelsen mellan ett mått på mobbning och ett mått på

aggression analyserades med hjälp av korstabeller och sambandsmått (phi squared

coefficients). Linjär regressionsanalys och variansanalys (ANOVA) tillämpades för att

analysera samband mellan psykosomatiska problem, mobbning och aggression. I studie

1 Det finns ingen direkt svensk översättning på begreppet ”peer victimization”, men det engelska begreppet är

övergripande och innefattar såväl mobbning som andra aggressiva handlingar

5

III samlades data in bland elever i årskurs 7 och 9 med dels en webbaserad enkät och dels

med fokusgruppsintervjuer. 128 elever deltog i enkäten medan 21 elever (totalt fyra

fokusgruppsintervjuer) deltog i fokusgrupperna. Den kvantitativa studien analyserades

med korstabeller och signifikans test (Chi2) och den kvalitativa studien analyserades med

hjälp av kvalitativ innehållsanalys. I studie IV deltog totalt 344 lärare och specialpedagoger

i en webbaserad enkät. Enkäten distribuerades via e-post där lärarna ombads värdera sin

rektors ledarskapsförmågor. De psykometriska egenskaperna hos PESOC-PLP skalan

analyserades med hjälp av Rasch analyser.

Resultat: Studie I: Av de elever som rapporterade att de blivit utsatta för peer

victimization rapporterade 13 % att de blivit mobbade, 44 % rapporterade att de blivit

utsatta för upprepad aggression och 43 % rapporterade att de blivit utsatta både för

mobbning och för upprepad aggression. Överensstämmelsen mellan de två måtten

belystes ytterligare med hjälp av phi square coeffiecient tester som visade att 18 % av

variansen i endera måttet utgjordes av det andra måttet. Resultaten visade att ett mått på

mobbning och ett mått på aggression fångar in delvis olika elever. Studie II: Resultaten

visade att andelen ungdomar med psykosomatiska problem var den samma oavsett vilket

mätinstrument som användes. De två måtten fångar in delvis olika individer och andelen

ungdomar som rapporterar psykosomatiska besvär i samband med peer victimization blir

högre då frågor om både mobbning och aggression inkluderas. Studie III: Utöver de

traditionella kriterierna repetition och maktobalans inkluderade ungdomarna även

hälsokonsekvenser av mobbning i sin förståelse för och definition av mobbning. Det vill

säga, en enstaka händelse som sårar eller skadar kan också anses som mobbning oavsett

om de traditionella kriterierna var uppfyllda eller inte. Vidare visade resultaten att flickor

och äldre ungdomar har en vidare syn på vad mobbning är och rapporterade fler typer av

beteenden som mobbning jämfört med pojkar och yngre ungdomar. Studie IV: Rasch

analyser indikerade två underdimensioner av PESOC-PLP skalan och en analys av

innehållet i frågorna i de två dimensionerna visade på två olika typer av ledarskap; direkt

pedagogiskt ledarskap samt indirekt pedagogiskt ledarskap. En fråga visade avvikelser från

Rasch modellen och kunde exkluderas eftersom innehållsvaliditeten inte försämrades.

Rasch analyser av de två underdimensionerna visade tillfredsställande psykometriska

6

egenskaper vilket indikerar att PESOC-PLP skalan kan vara användbar för att mäta

rektorns pedagogiska ledarskap.

Slutsatser: Att säkerställa en trygg och positiv skolupplevelse för barn är av stor betydelse

eftersom det är en viktig förutsättning för deras framtida hälsa. Alla barn har rätt att

uttrycka sig i frågor som rör dem och att lyssna på barns åsikter och uppfattningar är av

yttersta vikt då de inte alltid överensstämmer med vuxnas åsikter och uppfattningar.

Denna avhandling belyser problemen med definitioner och mätmetoder för mobbning

och skolledarskap. Med endast ett traditionellt mått på mobbning fångas den totala

förekomsten av peer victimization och dess samband med psykosomatiska besvär inte in.

Ett sådant mått reflekterar inte heller fullt ut ungdomars förståelse för och definition av

mobbning. För att integrera ungdomars syn på mobbning behöver definitionen på

mobbning inkludera negativa hälsokonsekvenser som ett kriterium och inte bara som en

konsekvens av mobbningsbeteende. Att PESOC-PLP skalan visar goda psykometriska

egenskaper gör att instrumentet kan användas till att utvärdera pedagogiskt ledarskap. I

Sverige och andra länder där man gör stora satsningar på rektorsutbildningar kan

PESOC-PLP skalan användas som ett instrument för att utvärdera dessa typer av

satsningar. Genom att öka förståelsen om olika mätmetoder för peer victimization och

skolledarskap kan resultaten från avhandlingen främja arbetet för en säker och trygg

skolupplevelse för barn och ungdomar.

7

CONTENT

LIST OF PAPERS ................................................................................................................................... 9

AUTHORS’ CONTRIBUTIONS .......................................................................................................... 10

INTRODUCTION ................................................................................................................................. 11

Positive and negative school experiences – a public health perspective ........................................... 13

Health, ill health and mental health ................................................................................................... 14

Stress and psychosomatic problems .............................................................................................. 15

New arenas for peer relations ............................................................................................................ 16

School climate ................................................................................................................................... 18

Effective schools and school leadership ........................................................................................ 18

Pedagogical and Social Climate in School (PESOC) .................................................................... 20

Concepts and definitions in peer victimization research ................................................................... 21

A historical review of the term bullying ........................................................................................ 21

Definitions ..................................................................................................................................... 23

Critique of the current definition of bullying ................................................................................ 27

Measurement methods in peer victimization research ...................................................................... 31

Using different informants ............................................................................................................ 31

Measurement concerns .................................................................................................................. 32

Study context ..................................................................................................................................... 33

Summarized problem formulation ..................................................................................................... 36

AIMS AND OBJECTIVES ................................................................................................................... 36

MATERIAL AND METHODS ............................................................................................................ 37

The ”Preventive School” project ....................................................................................................... 37

Study I ............................................................................................................................................... 38

Study II .............................................................................................................................................. 42

Study III ............................................................................................................................................. 44

Study IV ............................................................................................................................................ 48

Ethical considerations ........................................................................................................................ 51

RESULTS .............................................................................................................................................. 52

Prevalence rates of peer victimization ............................................................................................... 52

Study I ............................................................................................................................................... 53

Study II .............................................................................................................................................. 54

Study III ............................................................................................................................................. 58

Study IV ............................................................................................................................................ 60

8

DISCUSSION ....................................................................................................................................... 63

Discussion of results .......................................................................................................................... 63

Methodological discussion – strengths and weaknesses ................................................................... 72

CONCLUSIONS AND IMPLICATIONS ............................................................................................ 77

FUTURE STUDIES .............................................................................................................................. 79

ACKNOWLEDGEMENTS .................................................................................................................. 80

REFERENCES ...................................................................................................................................... 82

9

LIST OF PAPERS

I. Hellström, L., Beckman, L., & Hagquist, C. (2013). Self-reported peer

victimization: concordance and discordance between measures of bullying and

peer aggression among Swedish adolescents. Journal of School Violence, 12(4):

395-413. DOI:10.1080/15388220.2013.825626

II. Hellström, L., Beckman, L., & Hagquist, C. Does the strength of the

association between peer victimization and psychosomatic health problems

depend on whether bullying or peer aggression is measured? Manuscript

III. Hellström, L., Persson, L., & Hagquist, C. Understanding and defining bullying

– adolescents’ own views. Archives of Public Health 2015, 73:4. DOI:

10.1186/2049-3258-73-4

IV. Hellström, L., Hagquist, C. Psychometric properties of the PESOC-PLP scale,

a Swedish teacher instrument measuring Pedagogical Leadership – a Rasch

analysis. Submitted

10

AUTHORS’ CONTRIBUTIONS

Study I + II

Lisa Hellström participated in the preparation of the questionnaire, the data collection,

the design of the study, conducted the statistical analyses and interpretation of data and

drafted the manuscript. Linda Beckman participated in the preparation of the

questionnaire, the data collection and helped to draft the manuscript. Curt Hagquist

participated in the design of the study, participated in the data analysis and helped to draft

the manuscript. All authors read and approved the final manuscript.

Study III

Lisa Hellström initiated and implemented the study, conducted the data collection,

analyzed the data and drafted the manuscript. Louise Persson participated in the data

analysis and provided critical review of the manuscript. Curt Hagquist participated in the

design of the study, participated in the data analysis and helped to draft the manuscript.

All authors read and approved the final manuscript.

Study IV

Lisa Hellström initiated and implemented the study, conducted the data collection,

analyzed the data and drafted the manuscript. Curt Hagquist participated in the design of

the study, participated in the data analysis and helped to draft the manuscript. All authors

read and approved the final manuscript.

11

INTRODUCTION

Ensuring a safe and secure school experience is important for the protection of children’s

health and is a relevant public health issue as the childhood health status can be seen as a

determinant factor of future health. While strong leadership and social relations can be

seen as a positive determinant for mental health, academic achievement (Gustafsson et al.,

2010) and lower levels of drug use (LaRusso, Romer, & Selman, 2008), negative school

experiences and social relations can be linked to lower student engagement (Mehta,

Cornell, Fan, & Gregory, 2013) and mental ill-health (Gådin, Gillander, & Hammarström,

2003). Peer victimization was once thought of as a character-building rite of passage or

harmless behavior among peers. Today, however, it is well known that peer victimization

is associated with poor school performance (Glew, Fan, Katon, Rivara, & Kernic, 2005;

Lien & Welander-Vatn, 2013), feelings of loneliness and emotional problems (Woods,

Done, & Kalsi, 2009) and mental health complaints (Beckman, Hagquist, & Hellström,

2012; Stapinski, Araya, Heron, Montgomery, & Stallard, 2014). Recent studies have

shown that bullying can have long-lasting effects on both mental and physical health

(Holt et al., 2014) as well as motivation for further education (Liu, Lan, Hsu, Huang, &

Chen, 2014). Peer victimization affect children’s perception of the school as an

unwelcome and unsafe place, and the negative effects associated with peer victimization

are experienced also by students not directly involved (Lacey & Cornell, 2013; Swearer,

Espelage, Vaillancourt, & Hymel, 2010). Because of its’ negative effects, many schools

struggle to find ways to prevent peer victimization to offset the adverse outcomes

(Espelage, 2014). Even though the awareness of the problems associated with peer

victimization is growing, there are long-standing concerns about how to accurately define,

measure and estimate prevalence rates of peer victimization (e.g., Felix, Sharkey, Green,

Furlong, & Tanigawa, 2011; Greif & Furlong, 2006; Monks & Smith, 2006; Smith, Cowie,

Olafsson, & Liefooghe, 2002).

The meaning and manifestations of bullying may differ over time and

between different cultural contexts. The concept of bullying may also mean different

things from person to person, which indicates that there is a great complexity involved in

understanding bullying behavior among children and adolescents (SNAE, 2009b). The

terms peer victimization, bullying, and aggression are sometimes used interchangeably even

12

though they are defined differently. Due to the inconsistent use of measurement methods,

terminology and definitions the prevalence rates of peer victimization can vary greatly

from study to study. The conceptual confusion and measurement ambiguities in

victimization research need further illumination. According to the UN Convention on the

Rights of the Child (UNCRC), every child has the right to be heard in matters that

concerns them (General Assembly, 1989). Still, definitions of bullying are coined by adults

while children’s own understandings and perceptions have largely been neglected. This is

very unsatisfactory since adults and children have somewhat different perceptions of

bullying (e.g., Guerin & Hennessy, 2002; Madsen, 1996). Children’s perception of what

bullying is could be an important missing component in the pursuit of understanding and

addressing bullying in schools (Varjas et al., 2008). Finding effective methods to reduce

peer victimization requires the use of measurement methods that accurately reflect the

challenges facing children today.

Efforts to prevent bullying have been an integral part of the Swedish school

mission since the 1990s. Successful approaches to reduce bullying, mental ill health, and

poor academic achievement are those that attempt to create a supportive school ethos

and school environment by involving all personnel at school (Stewart, 2008). Many

approaches to reduce bullying are based on the assumption that bullying is a systematic

problem reflecting a school context rather than individual characteristics of bullies and

victims (Richard, Schneider, & Mallet, 2012). Measurement methods that assess effective

and successful leadership are important since strong school leaders are important not only

for the anti-bullying efforts in schools, but also for the creation of a positive school

climate and academic success (Dake, Price, Telljohann, & Funk, 2004; Höög, Johansson,

& Olofsson, 2005; Törnsén, 2009; Yukl, 2012). However, there is a scarcity of

psychometrically sound instruments assessing principal performance (Condon & Clifford,

2010; Tejeda, Scandura, & Pillai, 2001). In Sweden, the instrument Pedagogical and Social

Climate in School (PESOC), designed to measure school effectiveness and pedagogical

leadership, has been widely used for school improvement efforts (Grosin, 2002).

However, the psychometric properties of PESOC have not been examined previously,

which calls for an examination of its validity for future use. To compare schools and

determine the impact of leadership management on the individual school more

knowledge and awareness are needed of what is being measured and how the different

13

instruments and techniques are being used (Höög, 2011). Current methodological

ambiguities in research on peer victimization and school leadership, such as lack of valid

measurement methods, inconsistent definitions and operationalizations are of concern to

public health as this lack of knowledge indicates that the understanding of what to do to

reduce every child’s negative experiences in school is incomplete.

Positive and negative school experiences – a public health perspective

School achievement is associated with mental health and poor academic achievement can

lead to truancy and bullying. Such behavioral problems, in turn, increase the risk of

aggravating the peer-teacher relations and produce even weaker school performance.

Since school performance affects mental health and mental health affects school

performance, there is a risk of a vicious circle (Gustafsson et al., 2010). A study by

Juvonen, Wang and Espinoza (2011) reported that peer victimization can account for an

average of a 1.5 letter grade decrease in an academic subject (e.g., math) across three years

of middle school. Further, Glew, Fan, Katon and Rivara (2008) found that for each 1-

point increase in grade point average, the odds of being a victim versus a bystander of

peer victimization decreased by 10%. It is possible that the impact of peer victimization

on academic achievement can be explained by mediating or moderating factors such as

internalizing behaviors, low self-esteem, difficulty concentrating, friendship quality and

social support (See Espelage, 2014). Since young people spend most hours of the day in

school, school is an important arena for health promotion (Gådin, Gillander, &

Hammarström, 2003). The overall goal of the public health work in Sweden is to create

societal conditions for a good health on equal terms for the entire population. It has been

shown that childhood social and educational factors are strongly linked to adult mental

and physical health, and to adult health-related behavior such as smoking, alcohol

consumption, exercise and dietary habits (Wadsworth & Kuh, 1997). Therefore, children

and youth are prioritized target groups in public health promotion. Conditions during

childhood and adolescence constitute one of the main areas in Sweden’s public health

policy and focuses on, among other things, high quality schools and healthy development

(Proposition, 2007/08:110).

14

School exists to meet the needs of children and adolescents and according to

the UN Convention on the Rights of the Child (UNCRC) every child has the right to an

education that develops their full potential and that respects human rights (General

Assembly, 1989). The UNCRC further states that all children have the same rights and

equal worth and no child should be discriminated against. When it comes to children’s

economic, social and cultural rights, the state shall use the utmost of their resources to

satisfy the needs of the children. Ensuring a safe and secure school experience is a

relevant public health issue and can be seen as an implementation of the UNCRC. Since

school attendance is compulsory for Swedish children and adolescents, positive and

negative experiences in school in terms of relationships with peers, teachers and other

leaders are unavoidable. Promoting positive social relations, strong leadership and school

climate as well as preventing peer victimization is important for the health of all children

(Modin, 2012).

Health, ill health and mental health

Health concerns many aspects and areas of society, making it a complex concept that is

not easily defined. There is disagreement concerning the meaning of health since health

can have medical, social, economic, spiritual, and many other components (Larson, 1999).

The most commonly used definition, formulated by the World Health Organization,

states that health is “a state of complete physical, mental, and social well-being and not merely the

absence of disease or infirmity” (WHO, 1948). WHO’s definition highlights both physical and

psychological components of health but has been criticized for demanding states of

complete health (See Larson, 1999). Health can also be seen as a continuum with good

health at one end and ill health at the other end. While it is hard to determine when good

health is attained, it may be easier to recognize when health is not attained, i.e., when

individuals are experiencing ill health (Gadamer, 2003). Ill health can be described in

terms of disease and illness. While disease is defined as disturbed bodily functions, illness

is an individual’s perception of suffering from a disease or experience of mental well-

being or lack thereof (SOU, 2006). There is no generally accepted division between what

15

is considered normal and abnormal conditions in health and this division often varies

across different time periods and cultures (SOU, 2006). This thesis examines the

association between peer victimization and adolescent mental health by investigating self-

reported psychosomatic problems.

Stress and psychosomatic problems

In the last decades, self-reported mental ill health has increased among youth in Sweden

and globally (Hagquist, 2010). Recent reports have shown that adolescent reports of

sadness, anxiousness, sleeping problems and different forms of pain and ache have

increased. In the public debate, these symptoms are often described as signs of stress

(SOU, 2006). For scientific purposes, mental ill health among children is often divided

into two main categories: internalizing and externalizing problems. Externalized problems

often include behaviors such as defiance, aggression and impulsivity, while internalized

problems often include symptoms such as anxiety, depression, eating disorders as well as

self-harm and suicidal behavior. The division of externalizing and internalizing problems

is, however, partly arbitrary and problematic, since both types of problems often occur

simultaneously (SBU, 2010). Stress is a growing problem among children and adolescents.

When debated in media, the concept of stress has a wide meaning. To feel strained, not to

have time to do the things you want to do, to be frustrated, to be worried and having

trouble falling asleep are described as stress, or as signs of stress. In its basic form, stress

is a natural reaction to a situation that is perceived as challenging. Stress becomes harmful

if it is prolonged and does not allow for recovery (SBU, 2010).

From a psychological perspective, stress is when an individual feels that the

strains he or she is exposed to exceed the own ability to cope with the strains. If the

strains become too heavy they can lead to feelings of worry, frustration or tension

(Lazarus & Folkman, 1984). Somatic symptoms which are perceived to be the

consequences of prolonged stress, such as headaches or stomach aches, are often

included in this everyday description of stress (SOU, 2006). These symptoms such as

head, stomach and back pain may precede mental disorders, but they can also be a

16

consequence of mental illness and is therefore often used as an indicator of milder forms

of mental illness symptoms (known as psychosomatic symptoms). Psychosomatic

symptoms are psychologically induced and problems from two or more parts of the body,

such as the stomach, head, leg, chest and back, often co-exist (Petersen, Brulin, &

Bergström, 2006). Psychosomatic symptom is a term often used without definition

although the term can have different meanings. In modern dictionaries, it is defined in the

following two ways: 1) a disorder having physical symptoms but originating from mental

or emotional causes; 2) the influence of the mind on the body, and the body on the mind,

especially with respect to disease.2 In this thesis the term psychosomatic problems is used

according to the first definition, i.e., as physical/somatic symptoms caused by emotional

stress. Signs of mental ill health, such as stress and psychosomatic symptoms, must be

understood in relation to the child’s age, developmental level and the social context.

Whereas worry, distress and trouble sleeping for most children are transient reactions and

part of the normal development, persistent symptoms may inhibit children’s development

and increase the risk of serious problems such as anxiety and depression (SBU, 2010).

New arenas for peer relations

To understand peer relations and bullying among today’s youth one must consider the

social context, including the increasing influence of mass media on the lives of children

and adolescents. Today, most adolescents own a mobile phone and are linked to each

other through facebook, twitter, instagram or other social networking sites (SNS).

Research from the United States shows that among internet users aged 18-29, 89%

reported using SNS (Brenner & Smith, 2013). Further, posting comments to pictures or

profiles is the most popular activity on SNS and 94% of older teens report sharing photos

of themselves on their social media profiles (Livingstone, Haddon, & Ólafsson, 2011;

Madden et al., 2013). Teens also take active steps to shape their reputation on SNS;

research among college students has found that young women remove photos of

themselves with too few “likes” or pictures that are not in accordance with societal ideals

of beauty (Madden et al., 2013; Vitak & McLaughlin, 2011). Given the emphasis placed

2 http://www.thefreedictionary.com/psychosomatic

17

on weight and looks in present day society, bullying based on appearance may have

serious implications for the emotional well-being of young people (Eisenberg, Neumark-

Sztainer, Haines, & Wall, 2006; Thompson, Coovert, Richards, Johnson, & Cattarin,

1995).

Being constantly connected to the internet via smart phones means that

much of adolescents’ social interactions take place on the internet (Cassidy, Faucher, &

Jackson, 2013). Judgment of what is considered acceptable behavior may in some sense

be influenced by the exploitation of hurtful and humiliating behavior portrayed on

television and on the internet. For example, Swedish media recently highlighted cases of

public shaming, i.e., the publishing of offensive pictures and comments, on the picture-

sharing website Instagram which lead to riots and the prosecution of two adolescent girls

(Aftonbladet News, 2013). Another recent case highlighted in Swedish media was the

creation of Instagram accounts where children upload pictures of themselves for others

to rate (Värmlands Folkblad News, 2013). A study among adolescent girls showed that

the more time spent using SNS the higher level of self-objectification, i.e., the

internalization of an observer’s perspective on one’s own body (Fredrickson & Roberts,

1997; Vandenbosch & Eggermont, 2012). However, removing themselves from those

platforms may not be an option as it would mean exclusion from the social community

(Cassidy et al., 2013). The internet, mobile phones and other electronic media provide

young children access to information, culture, communication and entertainment to an

extent impossible to imagine only a few decades ago (UNICEF, 2011). At the same time,

information and communication technologies (ICT) influence many aspects in the lives of

adolescents into which school personnel and parents have little insight (Cassidy, Jackson,

& Brown, 2009; Lenhart, Purcell, Smith, & Zickuhr, 2010). Being brought up in a world

full of technology, younger people may not share older people’s balanced view and a

sense of caution not to trust technology blindly (Fisher, 2013). It has been pointed out

that the changing nature of peer relations may alter the conceptions of what is

unacceptable behavior and how to react to it (Abt & Seesholtz, 1994). To define bullying

among today’s youth, it is imperative to understand the new arenas for peer relations such

as the internet.

18

School climate

School climate is a complicated concept that is frequently used but poorly defined

(Libbey, 2004). Cohen, McCabe, Michelli, and Pickeral (2009) proposed that school

climate “refers to the quality and character of school life…based on patterns of people’s experiences of

school life and reflects norms, goals, values, interpersonal relationships, teaching and learning practices,

and organizational structures” (p. 180). There is extensive research showing that school

climate effect the life of youth and adolescents in many different ways. A positive school

climate, for example, has shown to affect students’ motivation to learn (Goodenow &

Grady, 1993) and is associated with lower levels of drug use (LaRusso et al., 2008) and

decreased student absenteeism (Purkey & Smith, 1983). On the other hand, a negative

school climate has shown associations with negative psychosomatic health (Modin &

Östberg, 2009) and academic failure (McEvoy & Welker, 2000). School climate is also

seen as an important component to include in school improvement efforts and to prevent

bullying (Thapa, Cohen, Guffey, & Higgins-D’Alessandro, 2013). A positive school

climate is of great importance for ensuring a positive school experience, and together with

the variables leadership and quality instruction, school climate is often associated with

what has been defined as effective schools (R. C. Kelley, Thornton, & Daugherty, 2005).

School Effectiveness Research (SER) is a research track focusing on the roles of the

school leaders as well as the organizational structures.

Effective schools and school leadership

The early stages of SER found that academic results and school adaptation among

students could be explained not only by students’ genetic abilities and social factors, but

also by factors relating to the organization of the school and leadership (Brookover,

Beady, Flood, Schweitzer, & Wisenbaker, 1979; Edmonds, 1979; Rutter, Maugham,

Mortimore, Ousten, & Smith, 1979; Sammons, Thomas, & Mortimore, 1997; Weber,

1971). The early stages of SER and the later stages of Educational Effectiveness Research

(EER) focus on the roles of the school leaders as well as the organisational structures and

other school variables in order to investigate what makes good schools (D. Reynolds et

19

al., 2014). A meta-analysis from 2005 covering 35 years of research indicated that school

leadership can have a substantial effect on student achievement (Marzano, Waters, &

McNulty, 2005). It has further been discussed whether different types of leadership styles

affect student achievement differently. Some scholars highlight the principal’s knowledge

of curricular content and instructional materials as most important for successful

leadership (e.g., Robinson, 2007; Stein & Nelson, 2003), while others emphasize the

principal’s ability to empower, support and distribute leadership to others (e.g.,

Leithwood, 2001). A study by Seashore Louis, Dretzke and Wahlstrom (2010) showed

that leadership variables such as focused instruction, shared leadership and teachers’ trust

in the principal are positively related to student learning.

A recent focus in EER is the development of school policy and teaching as

the most important factors for school effectiveness while leadership factors are seen as

less important (Creemers & Kyriakides, 2008; D. Reynolds et al., 2014; Scheerens, 2013).

The ambiguous empirical results of SER are further illustrated by a recent meta-analysis

suggesting that leadership has only a weak effect on student achievement and suggests

that focus should rather be on the content of school policy and the type of activities that

take place in schools (Kyriakides, Creemers, Antoniou, & Demetriou, 2010). The mixed

results concerning the associations between leadership and student outcomes may partly

reflect disparities as regards methodology, e.g., how the direct effects of leadership and

the indirect effects such as school leaders’ influence on teachers are distinguished.

Relatively few studies look at the connection between direct instructional leadership and

student outcomes, while the association between the principal’s indirect influence and

student achievement is generally more accepted (Seashore Louis et al., 2010; Stronge,

Richard, & Catano, 2008). While a great deal of attention has been paid to outcome

measures of student achievement in public education, only limited attention has been paid

to measures of leadership practice that may promote student learning (C. Kelley &

Halverson, 2012; R. C. Kelley et al., 2005). Moreover, the fact that many instruments for

measuring school leader performance are obsolete also warrants updated assessments of

their continued validity (Hallinger, Wang, & Chen, 2013). The scarcity of psychometrically

sound instruments may have contributed to the inconsistent results of different studies on

school effectiveness (Condon & Clifford, 2010; Tejeda et al., 2001).

20

The Swedish Education Act (SFS, 2010:800) includes both social/democratic

and academic goals and highlights the principal as the responsible pedagogical leader. As a

pedagogical leader, the principal is expected to provide opportunities for the education to

develop through discussion of pedagogical matters, leading discussions and meetings,

evaluating teaching activities, supporting the teachers and conducting systematic follow-

ups of students’ academic results (Rapp, 2009; Ärlestig, 2011). There is no agreed upon

definition but an often used definition states that pedagogical leadership is “the influence

a school leader exerts in relation to teachers through various actions, which aim to

influence them to improve their teaching in accordance with the objectives and guidelines

stated in the curriculum and Education Act” (Nestor, 1993, p. 183). Pedagogical

leadership has also been defined to include indirect and direct forms of leadership.

Indirect pedagogical leadership includes providing opportunities for teaching and

learning, for example organising activities by creating a clear structure for teacher- and

student schedules, meeting formats and allocating time for collaboration between

teachers. Direct pedagogical leadership, on the other hand, includes, for example, giving

feedback to teachers on teaching methods and analysing academic results in relation to

teaching methods, giving support, challenging and developing teachers through dialogue

and communication (The Swedish School Inspectorate, 2012; Törnsén, 2009).

Pedagogical and Social Climate in School (PESOC)

Pedagogical and Social Climate in School (PESOC) is a Swedish instrument designed for

studies on school effectiveness and pedagogical leadership. The first version was

developed at the beginning of the 1990s and later versions have been updated based on

the school reforms in 1990 and 2000 (Grosin, 2004). The instrument is based on the

theoretical framework of the early stages of SER and the conceptual understanding of

pedagogical leadership. PESOC includes pedagogical as well as social dimensions of

school climate aiming to capture the dynamic interplay between cultural and structural

factors as well as the interaction between students, teachers and school leaders (Grosin,

2002). The conceptual understanding of PESOC rests on the assumptions that the social

and pedagogical school climate is established by the expectations and norms of the

21

principal, school management and teachers, as well as by their perceptions of the purpose

of school, its potentials and limitations and by the consequent action patterns towards

students, colleagues and parents. In PESOC, the pedagogical leadership of the principal is

included as a sub scale (PLP scale) and is defined as a responsibility of the central

activities in school and aims to capture both the structure of the school (i.e. content and

form of teaching and education as well as the social interaction between the principal,

teacher and students) and the culture of the school (i.e. goal fulfilment, values and

theoretical assumptions of the principal and the teachers) (Grosin, 2002; Höög et al.,

2005). The definition is based on the assumption that student adjustment and actions,

attendance, performance and behavior in school are formed as they become aware of the

adults’ core values, expectations and behavior patterns (Grosin, 2002). In PESOC,

pedagogical leadership could be seen as a wider conceptualisation of instructional

leadership3 (Grosin, 2003). However, since pedagogical leadership is a concept that is

used mainly in a Scandinavian context, it is hard to find a direct analogy to international

research on school leadership (Rapp, 2009). The psychometric properties of PESOC have

not been examined previously and the results of studies using PESOC have not been

published in scientific journals. PESOC has been widely used for school improvement

efforts in Sweden, which warrants an examination of its validity for future use.

Concepts and definitions in peer victimization research

A historical review of the term bullying

Bullying is a historical concept and its meaning has changed along with changes in societal

values (SNAE, 2009b). The acknowledgement of bullying as something problematic is

fairly new, but peer harassment is hardly new in the school setting. Already in the 1700s it

was common for older students to harass new and younger students in order to maintain

the hierarchical structure that existed in the schools. This organized type of peer

harassment, called penalism, manifested itself in very sadistic forms varying from verbal 3 The most frequently used conceptualisation of instructional leadership includes three dimensions; defining the

school’s mission, managing the instructional program, and promoting a positive school-learning climate (Hallinger & Murphy, 1985).

22

harassment to assault (See Bliding, 2002). The phenomenon of bullying was first noted by

ornithologists as a behavior among birds and in the late 1960s, Konrad Lorenz used the

term “mobbing” to explain animal behavior. In ethology, the term mobbing describes a

collective attack by a group of animals on an animal of another species, which is usually

larger and a natural enemy of the group (Olweus, 2001). Lorenz (1968) also characterized

mobbing as actions of a school class or a group of soldiers ganging up against a deviating

individual. The word mobbing4 was first used in the Swedish debate at the beginning of

1970 by the physician Peter-Paul Heinemann (1972), who used it to describe the

phenomenon of aggressive group behavior towards a single child. According to

Heinemann, bullying behaviors occur naturally in humans as a biological species, but are

made possible by certain social factors. It was partially in the role of parent that

Heinemann acknowledged the phenomenon as his adopted son was exposed to degrading

comments about his skin color. Heinemann objected to the acceptance of this type of

behavior and deemed it unacceptable, demanding counteraction (Larsson, 2008).

Heinemann’s argument was that deviation does not cause abuse but rather triggers it. The

notion that the bullied child is different was an obvious assumption. In an interview for a

Swedish newspaper [Dagens Nyheter, November 13, 1969], Heinemann claims that the

prevalence of bullying is higher in urban and overcrowded cities and advises parents with

“deviant adoptive children” to move to small towns, preferably in northern Sweden, and

place the child in a small school (SNAE, 2009b). The deterministic view of bullying as a

naturally occurring behavior has later been criticized and an alternative explanation is that

bullying behavior is learned through interactions with other people (Ahlström, 2009).

In the scientific field the term bullying was first used at the beginning of the

1970s. The psychologist Dan Olweus conducted research on Swedish schoolboys and

found that both personality traits and the home environment were possible factors

associated with bullying involvement. During this time, the perception of bullying was

influenced by the understanding of penalism and was primarily described as a problem

among boys (SNAE, 2009b). In 1978, Dan Olweus’ book Aggression in the schools: Bullies and

Whipping Boys (1978) attracted attention internationally and aroused scientific interest in

bullying as an object of research. Olweus used somewhat different explanations to

describe bullying. Unlike Heinemann, he pointed to the fact that bullying does not have

4 Mobbing was later re-written as Sw. mobbning, which was considered to be more in line with Swedish spelling.

23

to be a collective action by a homogenous group but could also be the actions of

individual members. He also questioned the overemphasis on temporary and ‘spur of the

moment’ reactions and directed attention to bullying as a situation in which an individual

student is exposed to aggression systematically and over long periods of time – either

from another individual, a small group, or a whole class (Olweus, 1978, p. 5; Olweus,

2001). Led by the pioneering work of Olweus, research on bullying has flourished

internationally during the last three decades (Richard et al., 2012). Today, there is hardly a

published study on bullying that does not cite his groundbreaking work.

Definitions

The following section aims to illuminate similarities and differences regarding three

central concepts in peer victimization research: peer victimization, aggression and

bullying.

Peer victimization

Even though the term ‘peer victimization’ is sometimes used synonymously with the term

bullying, peer victimization can be seen as an umbrella term that includes all types of

negative actions between children (Greif & Furlong, 2006). Peer victimization has been

defined as experiences among children of being a target of the aggressive behavior of

other children, who are not siblings and not necessarily age-mates (Hawker & Boulton,

2000). It can be seen as a form of abuse that undermines children’s healthy development

that increases the risk of psychosocial dysfunction (e.g., Kochenderfer-Ladd & Ladd,

2001; Kochenderfer & Ladd, 1996; Olweus, 1991). Peer victimization can take the form

of direct verbal or physical aggression (van der Wal, de Wit, & Hirasing, 2003), indirect

aggression, such as rumor spreading or gossiping (Björkqvist, 1994), or relational

aggression such as social exclusion (Crick & Bigbee, 1998). Peer victimization can also

include victimization through electronic means such as the internet and mobile phones

(Slonje & Smith, 2008). In the present thesis, the term peer victimization is used as a

collective term for bullying and peer aggression. Although much effort has been made to

24

understand the complexity of peer victimization, there are still some conceptual

ambiguities (Greif & Furlong, 2006). Much of the concerns involve distinguishing

bullying from acts of peer aggression.

Table 1. Behavioral and relational features covered by the definitions of bullying and peer aggression

Umbrella term: Peer victimization

Bullying Peer aggression

Intent Yes Yes

Repetition Yes No

Power imbalance Yes No

Aggression

Aggression has been defined as any behavior directed toward an individual with the intent

to cause harm; some researchers include intentionality (e.g., Anderson & Bushman, 2002;

Berkowitz, 1993), while others do not, arguing that the intention behind a behavior is

difficult, perhaps impossible, to measure in children. For instance, Eron (1987) defines

aggression as “an act that injures or irritates another person” (p. 435). The debate

regarding including intent in the definition of aggression also regards the inclusion of

accidental behaviors causing harm. Anderson & Bushman (2002) maintain that accidental

harm should not be viewed as aggressive because it is not purposely aggressive. For the

purpose of the present thesis, aggression is defined as all behaviors intended to hurt or

harm others either physically or psychologically (Berkowitz, 1993).

Human aggression can be viewed as a multidimensional construct including

different forms and functions (Ostrov, Murray-Close, Godleski, & Hart, 2013). In

addition to the different forms of aggression briefly described in the section on peer

victimization above, aggression can also be reactive (e.g., a defensive response to a

perceived threat), or proactive (e.g., unprovoked, aversive behavior intended to harm,

dominate, or coerce another person) (Brown, Atkins, Osborne, & Milnamow, 1996). An

25

affect closely related to aggression is violence, which is described as extreme aggression

with the intent to harm (e.g., kill). All violence is aggression, but many instances of

aggression are not violent. For example, a child pushing another child off a tricycle is an

act of aggression but is not an act of violence (Anderson & Bushman, 2002, p. 29).

Bullying

Although aggression research dates back many decades, bullying as a research subject is

fairly new, beginning in the late 1970s (Griffin & Gross, 2004). In some of the early

definitions of bullying, there was no distinction made between bullying and aggression.

For example, Randall (1991) defines bullying as: “Aggressive behavior arising out of a deliberate

intent to cause physical or psychological distress to others”. Even today the terms aggression and

bullying are often used interchangeably (Leff & Waasdorp, 2013). Rigby (2002) observes

that some researchers may prefer to use the term aggression when talking about bullying

because “it is more familiar to researchers and less tainted, one may think, by everyday

use” (p. 30). In contrast to peer aggression, bullying is more narrowly defined. The

definition of bullying has been, and still is, widely debated and to date there is no generally

accepted definition. The word bullying may also have different meanings and may be

interpreted differently between countries, cultures and children of different ages (Smith et

al., 2002). The varying definitions used in bullying research also make it hard to compare

findings from studies conducted in different countries and cultures (Griffin & Gross,

2004). In addition to the definition proposed by Randall (1991), another early definition

of bullying stated that “bullying is a willful, conscious desire to hurt another and put him/her under

stress” (Tattum & Tattum, 1992, p. 147). This definition was criticized, in particular among

behaviorally oriented scientists, since the definition implied that someone can be a bully

without actively doing anything. Having a conscious desire to hurt someone would be

enough to make someone a bully which would imply that bullying is a state of mind or an

attitude rather than a behavior (Besag, 1989; Rigby, 2006a). Although this view of the

bully as a bad person may be morally appealing, such a view would make it hard to

distinguish bullies from non-bullies. Also, wanting to hurt someone does not mean that

26

you actually will put your thoughts into action (Rigby, 2006a). From a behavioral

psychological view, the definition of bullying focuses on the victim and on actions rather

than attitudes. For example, Dan Olweus (1993) stated: “A student is being bullied or

victimized when he or she is exposed, repeatedly and over time, to negative actions on the part of one or

more students”. He also added that “the student who is exposed to negative actions has difficulty

defending him- or herself”. In the revised version of the Olweus Bully/Victim Questionnaire

(OBVQ) (Olweus, 1996) the definition was somewhat expanded:

We say a student is being bullied when another student, or several students

Say mean and hurtful things or make fun of him or her or call him or her mean and hurtful

names

Completely ignore or exclude him or her from their group of friends or leave him or her

out of things on purpose

Hit, kick, push, shove around, or lock him or her inside a room

Tell lies or spread false rumors about him or her or send mean notes and try to make other

students dislike him or her

And do other things like that

When we talk about bullying, these things happen repeatedly, and it is difficult for the

student being bullied to defend himself or herself. We also call it bullying when a student is

teased repeatedly in a mean and hurtful way. But we don’t call it bullying when the teasing is

done in a friendly and playful way. Also, it is not bullying when two students of about the

same strength or power argue or fight (See Olweus, 2001).

In the 1993 definition, there is no assumption of intention to harm, which was added in

the 1996 definition. Olweus’ definition of bullying is the most widely used formulation of

what constitutes bullying and includes three criteria for bullying behaviors: intent,

repetition and a power imbalance between the perpetrator(s) and the victim. Many

researchers also agree that bullying takes place in the presence of peers and that the victim

does not provoke the bullying behavior (Griffin & Gross, 2004). Following the rapid

developments of information and communication technology (ICT) and widespread use

in society, including increased internet access and use, the forms and platforms for

bullying have changed. Cyberbullying, i.e. bullying using electronic forms of

communication, as a research topic is new and the word cyberbullying did not even exist a

27

decade ago. To date, there is no generally accepted definition of cyberbullying which may

be due to a lack of conceptual clarity (Dooley, Pyżalski, & Cross, 2009; Notar, Padgett, &

Roden, 2013; Tokunaga, 2010; Vandebosch & Van Cleemput, 2008). However, most

researchers define cyberbullying as traditional bullying using an electronic medium.

Specific criteria for cyberbullying have been proposed, such as anonymity and publicity

(Menesini & Nocentini, 2009; Slonje & Smith, 2008; Tokunaga, 2010). While it is claimed

that the easy access to ICT has led to an increase in bullying incidents (e.g., Fisher, 2013),

others argue that cyberbullying is rather a new expression of long-established negative

social relations (Livingstone et al., 2011). This assumption is based on research showing

that victims of cyberbullying often also are victims of traditional bullying and that they

report similar associations with mental health (e.g., Barboza, 2015; Beckman, 2013;

Gradinger, Strohmeier, & Spiel, 2009).

Critique of the current definition of bullying

Even though Olweus’ definition of bullying is the most widely used formulation today of

what constitutes bullying, it is not without flaws. The objections involve the use of the

three criteria of a) intent, b) repetition, and c) power imbalance as ways of separating

bullying from other forms of peer victimization.

Intent

According to the definition, bullying includes intentional and negative actions. This is,

however, also included in the definition of peer aggression. Hence, the purpose and

motivation behind hurtful behavior cannot be used to differentiate between the two

concepts. Intentionality is particularly hard to measure in children’s behavior, which has

been acknowledged in both peer aggression research (Eron, 1987) and bullying research

(Greif & Furlong, 2006). Some researchers argue that it is possible that children can be

“unintentional bullies”, i.e., they may not be aware that their behavior is hurtful. In such a

28

case, there is no intention of harm but there is someone being hurt who perceives the

behavior as bullying (Rigby, 2006b).

Repetition

The definition further states that bullying happens repeatedly, but does not indicate what

is meant by repetition. A study by Solberg and Olweus (2003) showed that victims of

repeated bullying report significantly more psychosocial maladjustment and depression

compared to students reporting occasional incidents of bullying and hence argue that “2-3

times per month” is a reasonable cut point for bullying. Further, Olweus (1993) justifies

the repetition criterion because it excludes, what he calls, occasional violent and “non-

serious” events. However, some researchers claim that occasional incidents of bullying

and aggression can lead to equally traumatic consequences as repeated events due to fear

of renewed attacks and argue that even occasional incidents could be regarded as bullying

(Arora, 1996). By discrediting occasional incidents of bullying there is a risk that children

that are traumatized from single experiences of intense bullying will not be given attention

(Rigby, 2006b). Other problems in assessing peer victimization are the use of different cut

points and time frames for the bullying behavior. The reference period, to which the

respondents are asked to relate their assessment, can sometimes be set to the whole

school year, one school term, or the past 2 or 3 months. Such a wide range of time frames

calls the reliability and comparability of data into doubt (Solberg & Olweus, 2003), but

also gives different meanings to “occasional” bullying. While occasional incidents in the

time frame “a school year” would be less likely to be interpreted as bullying, occasional

incidents in the time frame “2-3 months” could in reality imply repetition.

Power imbalance

Another problem with the definition of bullying is the difficulty to clearly describe what is

meant by “power imbalance”. A study by Green, Felix, Sharkey, Furlong, and Kras (2013)

showed that a single item measure of bullying might not detect the subtle forms of power

29

imbalance that separates it from other forms of peer victimization. While physical

strength, superiority in numbers, a higher position in the hierarchical order or being

verbally skilled may be more obvious sources of power, having unexpected qualities of

assertiveness or support from others may be examples of power of a less obvious kind. In

addition, power imbalance may be situationally bound and can change over time (Rigby,

2006a). Further, the features of power imbalance are not always in alignment. Finkelhor,

Turner & Hamby (2012) illustrates this by asking if a stronger but less popular girl

repeatedly intimidates a weaker but more popular boy, is the power dimension popularity,

gender or physical strength? (p. 2). The indication that power imbalance implies a

difficulty to defend oneself (Olweus, 1996), means that any situation where the “victim” is

able to defend herself or himself is not bullying. However, even if you are able to defend

yourself you might not always choose to do so for various reasons.

The rationale behind the criteria seems to be separating harmful behavior

from less harmful behavior, provided that all three criteria are present (e.g., Olweus &

Breivik, 2014), i.e., all three subcomponents are necessary, but each is insufficient to judge

bullying behaviors on its own (Greif & Furlong, 2006). Despite this, not all studies

claiming to measure bullying include all three criteria in their definition; many bullying

measures ask respondents to indicate the frequency of aggressive actions, and if endorsed

as occurring on more than one occasion (i.e., repetition), will infer that this represents

bullying behavior (Leff & Waasdorp, 2013). It has also been argued that the criteria used

to distinguish different forms of peer victimization need more empirical foundation

(Finkelhor et al., 2012). Organizing a research field around a concept whose definition is

so difficult to specify is a huge drawback (Finkelhor et al., 2012).

Children’s definitions of bullying

The general understanding and definition of bullying are products of an adult perspective

while children’s own definitions and perceptions have largely been neglected. When

students are asked to define bullying, they rarely include the traditional criteria for bullying

(Frisén, Holmqvist, & Oscarsson, 2008), which suggests that their understanding of what

it means to be bullied may differ from adults’ understanding (Beran, 2006). Regarding

30

gender differences, girls tend to omit the traditional criteria and mention the effect on the

target more often compared to boys (Frisén et al., 2008; Naylor, Cowie, Cossin, de

Bettencourt, & Lemme, 2006; Thornberg & Knutsen, 2011). In addition, younger

children tend to report physical aggression as bullying more often, while older children

more often report verbal aggression and social exclusion as examples of bullying (Madsen,

1996; Smith et al., 2002; Swain, 1998). Madsen (1996) found that among pupils aged 6-16

years, 40% reported that bullying must have a negative effect on the victim. In their study

among 10-13 year-old children, Guerin and Hennessy (2002) found that more than half

reported that a physical advantage did not matter in bullying and Varjas et al. (2008)

suggested that bullying is a means to gain power, not a prerequisite for bullying. Further, a

study by Boulton et al. (2002) showed that about 80% of the students agreed that

behaviors such as hitting and pushing, threatening people, or forcing someone to do

something were considered bullying, whereas only half of the students regarded behaviors

such as calling someone names, stealing people’s belongings and telling nasty stories

about other people, as bullying. This could affect the chance of a bystander’s intervention

as research has shown that tendencies to intervene in bullying situations have shown to be

dependent on what types of behavior are considered bullying and what types of behavior

are expected by the peer group (Boulton et al., 2002; Pozzoli & Gini, 2010).

Despite the acknowledgement that children may have a different

understanding of bullying compared to those researching the problem and although every

child has the right to be listened to in matters that concerns them (General Assembly,

1989), children’s understanding of bullying have had little or no impact on the definition.

Rather, suggestions for solving the discrepancies include adjusting children’s definitions

to coincide better with researcher’s definitions (Frisén et al., 2008; Hopkins, Taylor,

Bowen, & Wood, 2013; Maunder, Harrop, & Tattersall, 2010). However, many adults are

unaware of the seriousness and the consequences of bullying in their schools and often

underestimate the number of students being bullied (Bradshaw, Sawyer, & O'Brennan,

2007; Kochenderfer-Ladd & Pelletier, 2008). It has been discussed that the discrepancy

between student and adult understandings of bullying can be an indicator of poor school

climate with regard to awareness of problematic peer behavior and willingness to

intervene (Espelage, Low, & Jimerson, 2014). The differing understandings of bullying

between children and adults could also indicate problems within the anti-bullying

31

education process (Hopkins et al., 2013). Given that accurate prevalence rates are critical

for planning treatment and prevention (Vaillancourt et al., 2008), it is of great importance

to have measurement instruments including definitions that correctly reflect peer-relations

among today’s youth and that capture the entire phenomenon of bullying. Current

methodological uncertainties regarding the definition of bullying and the lack of

understanding of what it means to be bullied need to be further addressed.

Measurement methods in peer victimization research

Using different informants

Different measurement methods such as teacher reports, peer nominations and self-

reports use different informants for assessing peer victimization. Teacher reports typically

ask teachers to observe and rate children on a victimization scale. The usefulness is that

such instruments add an outsider’s perspective on behaviors in school. However, not all

teachers are attentive to or aware of certain behavior, e.g., relational bullying in the form

of rumors or covert teasing or bullying over the internet (Griffin & Gross, 2004). In

addition, since the children and adult view of bullying may differ, teacher reports may be

underreporting prevalence rates of bullying in school (Smith & Sharp, 1994). Peer

nominations often work in a similar way as the teacher reports, asking students to

nominate their classmates in different categories. For example, the students receive a list

with the names of all the children in the class and are asked to make a mark next to any

student’s name that they think fit the described behavior (Perry, Kusel, & Perry, 1988).

Still, the most commonly used instruments in bullying research are self-reports including

questions on how often respondents have been exposed to bullying (Bond, Carlin,

Thomas, Rubin, & Patton, 2001; Crick & Bigbee, 1998; Stevens, De Bourdeaudhuij, &

Van Oost, 2000) or attitudes concerning bullying (Rabiner, Keane, & MacKinnonLewis,

1993). The most widely used self-report questionnaire is the Olweus Bully/Victim

Questionnaire, OBVQ, (Olweus, 1996). OBVQ includes both a definitional method to

measure bullying asking respondents to indicate how often they have been bullied as well

as questions concerning specific types of peer aggression behavior. The correspondence

32

between self-reports and peer nominations has shown to be low. Results from a study by

Scholte, Burk and Overbeek (2013) suggest that self-reported victimization is related to

emotional problems, while peer reported victimization is more indicative of social

problems. Further, a study by Pakaslahti and Keltikangas-Järvinen (2000) showed that the

correlation between peer nominations and teacher-reports are higher than the agreement

between teacher-reports and self-reports. Lowest correlation was shown for peer

nominations and self-reports. Different measurement techniques involve respondents that

may differ in their understanding of what it means to be bullied, which could explain the

low correlation between these measures.

Measurement concerns

In bullying research, assessment has been called the “Achilles’ heel” of prevention efforts

(Cornell, Sheras, & Cole, 2006) and there are concerns whether currently used measures

actually assess the intended subset of peer victimization (Greif & Furlong, 2006). Self-

reported peer victimization is usually measured by using either a single global question

together with a definition of bullying, or questions of specific peer aggression behavior.

The use of a definition is a way to operationalize the concept of bullying while peer

aggression is usually operationalized by including behavior descriptors in sentence form,

asking about exposure to different types of aggression (W. Reynolds, 2003). Prevalence

rates of peer victimization are often higher when assessed with measures of specific

aggressive behaviors compared to measures using a single question about bullying

together with a definition (Sawyer, Bradshaw, & O'Brennan, 2008; Stockdale,

Hangaduambo, Duys, Larson, & Sarvela, 2002). The use of a definition and emotionally

laden labels, like “bullying”, may lead to the underreporting of victimization due to stigma

and shame involved in admitting to being a victim (Bosworth, Espelage, & Simon, 1999;

Greif & Furlong, 2006; Kert, Codding, Tryon, & Shiyko, 2010). Acknowledging yourself

as a victim of bullying could mean recognizing unwanted qualities of a typical victim, such

as anxiety, insecurity, low self-esteem, few friends (Smith et al., 1999), and poor social

skills (Fox & Boulton, 2005). In that sense, it may be easier to admit being subjected to

different types of peer aggression behavior without having to label yourself as a victim of

33

bullying. Assessing peer victimization by asking about exposure to specific aggressive

behavior may avoid the stigma associated with laden labels, but may not discriminate

between bullying and brawl or quarrel among peers as the formal criteria for bullying

usually are not stated explicitly (Furlong, Sharkey, Felix, Tanigawa, & Green, 2010).

Hunter, Boyle and Warden (2007) found that adolescents who are targets of

aggressive behaviors reported a more severe depressive symptomology when they also

reported intent and power imbalance (e.g. the criteria for bullying) compared to when

intent and power imbalance were not reported. However, their study did not use a

definition of bullying, nor the word bullying and thus avoided methodological ambiguities

associated with such a measure, e.g., possible underreporting due to stigma or shame.

Some researchers argue that a clear definition of bullying is needed to minimize room for

subjective interpretations (Solberg & Olweus, 2003), while others argue that children

would rather be likely to use their own understanding of what it means to be bullied

(Smith et al., 2002). While the association between peer victimization and psychosomatic

health problems has been reported in previous literature (e.g., Beckman et al., 2012; Gini

& Pozzoli, 2009; Gådin Gillander & Hammarström, 2003), little is known about whether

this association varies between different measures of peer victimization. Considering the

long-standing concerns about how to accurately measure and estimate prevalence rates of

peer victimization (see Felix et al., 2011), it is high time to analyze the concordance and

discordance and whether the association with psychosomatic health problems differs

between different peer victimization measures.

Study context

The Swedish school system and the role of the principal

The Swedish school system has multiple structures, namely the national level including

the curriculum and other steering documents with objectives for all schools, the

municipality level and the internal structures at each individual school (Ahlström, 2009).

In recent years, there have been major changes in terms of the roles and responsibilities

of the school leaders. In the 1990s, an extensive decentralization of responsibility for

school operations was made from the state to the municipal level. Although school

34

leaders and teachers have been given greater responsibility for managing and improving

school there is a strong state and local control of the school in terms of curriculum and

quality assessments (Ludvigsson, 2009). Since 2010, all newly hired principals are obliged

to undergo a comprehensive head teacher training program (SFS 2009:1521). The

purpose with the program is to improve schools’ goal fulfillment by providing school

leaders with required skills and abilities. Public education in Sweden is regulated by the

Education Act (2010:800), which clearly states that schools should work actively with the

social and civic objectives as well as with the academic objectives (Ahlström, 2009). These

tasks indicate that a school cannot be perceived as fully successful if it fails to achieve

either academically or socially (Ahlström & Höög, 2009; Höög, 2011). The Education Act

further emphasizes that the principal is responsible for student achievement and school

effectiveness as well as for teachers' professional development. The Swedish school

debate since the beginning of 2000 has increasingly focused on result and goal fulfillment

and an important part of efforts to improve today’s schools in a broad sense is to

strengthen the efforts to improve school leadership (Höög & Johansson, 2011).

The schools role in preventing and reducing bullying

During the 1980s, Swedish schools became responsible for the anti-bullying work and

efforts to develop methods to prevent and combat bullying was prioritized. The different

methods represented different views of bullying and have had great influence on the

conceptions and discussions of bullying in Swedish schools. The earlier assumption that

the victim was deviant or different that prevailed in the 1970s slowly subsided and more

attention was directed towards bullying as a manifestation of problems in the democratic

and social climate in school (SNAE, 2009b). Today, schools role in bullying prevention is

emphasized in the curriculum and legal documents. In 2006, the Equality Act came into

force and as of 2009 harassment and victimization are addressed in the Education Act

Chapter 14a (SFS 2010:800) and discrimination in the Discrimination Act (SFS 2008:567).

Both the Education Act and the Discrimination Act state that the school shall take active

steps to prevent and combat abusive behavior and harassment. Furthermore, they indicate

that schools have the obligation to investigate and take action if it comes to their

35

knowledge that a child or a student claims to have been subjected to abusive treatment or

harassment. According to the Education Act, the school shall also ensure that an annual

plan against degrading treatment is established. Similarly, the Discrimination Act states

that an equality treatment plan should be revised annually (Skolinspektionen, 2010).

Instead of using the term “bullying”, the Education Act and the

Discrimination Act use the terms “discrimination” [Diskriminering] and “degrading

treatment” [Kränkande behandling]. According to 14a Chapter 3 § in the Education Act,

degrading treatment is defined as a “conduct that without being discrimination according to the

Discrimination Act violates a child's or student's dignity. This could include harassment based on

overweight, hair color, someone being studious or similar offensive treatment”. Degrading treatment

may also exist without the offender specifying any unique characteristics of the victim.

Pushing, pulling someone’s hair or psychological harassment such as ostracism also

constitute degrading treatment in the legal sense (SFS 2010:800). The terminology used in

the Education Act has been introduced in the school curriculum and the word bullying is

used less seldom (SNAE, 2009a). Skolinspektionen (2010) writes that the definition of

bullying, regarding the criteria of repetition, is too narrow to capture children’s

vulnerability.

36

Summarized problem formulation

Ensuring a safe and secure school experience with optimal conditions for good academic

performance free of peer victimization is important for children’s health and is a relevant

public health issue as childhood health status can be seen as a determinant of future

health. There are, however, long-standing concerns on how to accurately define and

measure peer victimization. Conceptual, operational and measurement difficulties in peer

victimization research affect both prevalence rates and the full understanding of the

negative consequences of bullying behavior. At the same time new forms of peer

victimization over the internet, reflecting the changing social conditions among youth

may have altered adolescents’ and children’s views of what behaviors constitute bullying

which may challenge previous definitions. Little attention has been paid to children’s

understandings when it comes to defining bullying. Given that prevalence rates are critical

in planning treatment and prevention it is of great importance to have measurement

instruments including definitions that correctly reflect peer relations among today’s youth

and that capture the entire phenomenon of bullying. The Swedish Education Act clearly

states that schools should work actively with social, civic, as well as academic objectives

and the principal is responsible for the preventive and anti-bullying work in school. What

constitutes successful leadership, however, is debated and evaluation of instruments

measuring principal leadership has received limited attention. A lack of evaluated

instruments assessing successful leadership could be one explanation for the inconsistent

research findings regarding leadership and student outcomes. To be able to compare

schools and determine the impact of leaders’ work in the individual school, valid

instruments assessing school leadership are needed.

AIMS AND OBJECTIVES

The overall aim of this thesis is to examine methods for assessing peer victimization and

pedagogical leadership in school. The following research questions are asked: 1) To what

degree does a measure of bullying overlap with measures of peer aggression with respect

to the number of students identified as victims and association with self-reported

37

psychosomatic problems? 2) What are Swedish adolescents’ definition and understanding

of bullying? 3) Can the PESOC-PLP scale be used as a valid measure to assess the

pedagogical leadership of the principal?

MATERIAL AND METHODS

The ”Preventive School” project

During 2005-2007, the national Public Health Institute in Sweden (FHI) was

commissioned by the Government to spread knowledge about effective alcohol and drug

prevention programs in schools. The “Preventive School” project (PS) is a wide initiative

in all municipalities and comprehensive schools in Sweden where Skåne and Värmland

were chosen as pilot counties. In the Värmland region Karlstad, Arvika and Sunne were

selected as pilot municipalities. The project included the implementation of different

programs and methods to promote mental health among children and youth (Karlstad

municipality, 2007). The PS project is the empirical basis of the studies in the current

thesis.

During a period of three years (2009-2012), the municipality of Karlstad

[Barn- och Ungdomsförvaltningen, BUF] and Karlstad University [Centre for Research

on Child and Adolescent Mental Health, CFBUPH] conducted a collaborative project,

funded by the national Public Health Institute in Sweden. The overall aim was to promote

a positive school environment, good mental health and increased academic achievement

among youth. Within the project, extensive data collection has been conducted involving

students (school years 4-5 and 7-9), teachers, principals, school nurses, school social

workers and school managers. As a part of the project, CFBUPH has also identified and

described the extensive training initiatives that have been conducted among teachers in

Karlstad municipality since 2006. The specific training initiatives, with the aim to create

greater leadership in the classroom, improving social relations and reaching a higher level

of working atmosphere in the classroom, include programs such as Social Emotional

Learning [SET], Komet, Classroom Management [Ledarskap i klassrummet], Örebro

Prevention Program, Motivational Interviewing [Motiverande samtal], Repulse, among

38

others. Within the project, models for feedback of results to students and teachers have

been developed in close collaboration with selected schools involved in the project.

Continuous feedback has also been given to principals and school managers.

Table 2. Method summary of studies I-IV

Study Aim Study design Study population Analysis I Examine the

concordance and discordance between a measure of bullying and measures of peer aggression

Cross-sectional. Web-based questionnaire. Researchers on site organizing the data collection

1,760 students, School years 7-9 (ages 13-15) from 8 different schools

Contingency tables, measures of

association (Φ2)

II Analyze to what extent the strength of the association between peer victimization and psychosomatic problems among adolescents depends on which of two measures is used, a measure of bullying and a measure of peer aggression

Cross-sectional. Web-based questionnaire. Researchers on site organizing the data collection

2,568 students, School years 7-9 (ages 13-15) from 8 different schools

Contingency tables, linear regression analysis, ANOVA

III Explore adolescents’ definitions of bullying

1) Cross-sectional. Web-based questionnaire + 2) Semi-structured focus group discussions

1) 128 students; 2) 21 students; School years 7 and 9 (ages 13 and 15) from two different schools

1) Descriptive 2) Qualitative content analysis

IV Analyze the psychometric properties of the PESOC-PLP scale, using Rasch analysis

Cross-sectional. Web-based questionnaire distributed by e-mail

344 elementary school teachers, from 30 different schools

Rasch analysis: lack of invariance, targeting, item fit and threshold ordering, differential item functioning (DIF)

Study I

Data collection

The municipality of Karlstad implemented and introduced the study to all high-school

principals and school managers in the city. In a second step, information about the study

was sent out by CFBUPH to all principals and school personnel via email and shortly

39

thereafter all principals were contacted to schedule the data collection. Information and

consent letters to all students and their caregivers were delivered to the schools a couple

of weeks before the anticipated data collection and caregivers for students in school years

7 and 8 were asked to contact the teacher if they did not want their child to participate in

the study. Due to the students’ higher age in school year 9 (15 years), caregiver consent

was not needed. A total of 52 students and/or their caregivers declined participation in

the study. Members of the research team (three doctoral students) collected the data

during regular class-hours, with one doctoral student present at each data collection.

Students were given oral information about the aim of the study, that their participation

was voluntary, that all answers were anonymous and that they could terminate their

participation at any point. Thereafter, the students received the link to the questionnaire

and a unique password randomly generated by the survey tool esMaker (Entergate, 2009).

The questionnaire took approximately 30 minutes to complete. Information regarding

class size (number of girls/boys), participation and absence were collected at each data

collection by the researcher on site. Any questions the students had regarding the

questions in the survey as well as any distractions or disturbances during the completion

of the questionnaires were documented. Researchers revisited every school to collect data

from students that were sick or absent during regular data collection.

Study design and participants

A cross-sectional study design was used in order to collect self-reported data regarding

bullying and peer aggression. The study population included students from eight out of

nine high schools in the ages 13-15 (school years 7-9) in Karlstad. All nine high schools in

the city were invited to participate, of which one school declined to participate in the

study. Data collection was carried out in December 2009. Total response rate was 82.3%

(n=1,760); school year 7 (89%), school year 8 (81%) and school year 9 (82%). The total

response rates were similar for boys (81%) and girls (82%).

40

Instruments

The questionnaire included 109 questions regarding health, bullying, school, alcohol- and

drug habits and physical activity. For the purpose of the study questions on bullying and

peer aggression were used.

The Olweus Bully/Victim Questionnaire (OBVQ)

Questions regarding bullying and aggressive behavior were measured using OBVQ

(Olweus, 1996). The self-report measure of bullying was filled out anonymously and a

definition used in the Swedish version of the questionnaire Health Behavior in School-

aged Children (HBSC) was included (Marklund, 1997):

We say a student is being bullied when another student, or a group of students, says or does nasty

and unpleasant things to him or her. It is also bullying when a student is teased repeatedly in a way

he or she does not like. But it is not bullying when two students of about the same strength quarrel

or fight. It is also not bullying when the teasing is done in a friendly or playful way.

Following the definition was a question regarding bullying victimization: “How often

have you been bullied in school in the last couple of months?” with the response

categories “I haven’t been bullied in the last couple of months”, “Occasionally”, “2-3

times/month”, “About once a week” and “Several times/week”. The time frame “the last

couple of months” is thought to constitute a natural (and not too long) memory unit for

the students. Also, most questions have a clear spatial reference, asking about events and

activities having occurred “at school” (Solberg & Olweus, 2003). The question on bullying

was followed by questions regarding exposure to specific aggressive behaviors. The

questions concerning peer aggression followed after the statement “Has another

student/s in school done anything of the following to you in the last couple of months?”

The behaviors included: “I have been scoffed at, ridiculed, and called nasty names in an

unpleasant and hurtful way”, “Other students have not let me be a part, have on purpose

excluded me from others or totally ignored me (pretended like I did not exist)”, “I have

been hit, kicked, pushed or locked in”, “Other students have been spreading lies or false

rumors about me or have tried to get others to dislike me”, and “I have been threatened

or forced to do things that I did not want to do”. The response categories were “It hasn’t

41

happened to me in the last couple of months”, “Occasionally”, “2-3 times/month”,

“About once a week” and “Several times/week”. OBVQ includes three additional

questions regarding aggressive behaviors. The question “I have been bullied with

unpleasant words, comments or gestures with sexual meaning” was removed from the

analyses since it contained the word bullying. Two questions were excluded due to space

limits, i.e. “I have been robbed of money or other things, or have had things destroyed”,

and “I have been bullied with unpleasant words or comments about my skin color or

immigrant background”.

Data analysis

Classification of victims

Students that reported either traditional bullying or cyberbullying were identified as

bullying victims. Further, as the definition requires repetition, any indication of bullying

victimization (even those reporting being bullied occasionally/once or twice) was

considered bullying. Victims of peer aggression were identified if they reported exposure

to any of the peer aggression behaviors. In addition, a variable including adolescents

reporting exposure to one or more behaviors was constructed to present an overall

assessment of peer aggression. In the descriptive analyses, victims were classified as

occasional peer aggression victims (including students reporting being bullied occasionally)

and frequent peer aggression victims (including students reporting being bullied 2-3

times/month, about once a week, or several times/week). In the statistical analyses, two different

cut points for peer aggression were used and analyzed separately; occasional or more

(including students reporting being bullied occasionally, 2-3 times/month, about once a week, or

several times/week) and 2-3 times/month or more (including students reporting being

bullied 2-3 times/month, about once a week, or several times/week).

42

Statistical methods

Concordance and discordance between a measure of bullying and a measure of peer

aggression were analyzed using contingency tables and measures of association. Phi

squared coefficients (Φ2) are used to examine the association between two binary

variables, i.e., bullying and peer aggression. The phi squared coefficient is a symmetric

index that does not depend on which variable is considered dependent and gives the

proportion of variance in one variable explained by the variance in the other variable (H.

T. Reynolds, 1984). Lastly, the degree to which peer victimization was captured by a

measure of bullying and measures of peer aggression (95% CI), i.e., the unique

contribution of each measure as well as the overlap were examined.

Study II

Data collection

In study II, the data collection was identical to study I (see page 38).

Study design and participants

A cross-sectional study design was used in order to collect self-reported data regarding

bullying, peer aggression and psychosomatic problems. Data collection was carried out in

October 2010 and included students from all eight high schools in the ages 13-15 (school

years 7-9) in Karlstad (n=2568). After the data collection in 2009, one high school in the

city was closed down. The study population included school year 9 students in 2009

(n=586) and school year 7-9 students in 2010 (n=1982). Self-reported data regarding

bullying, peer aggression and psychosomatic problems were collected. The total response

rate in 2010 was 90.3%; school year 7 (93%), school year 8 (90%) and school year 9

(90%). The total response rate was somewhat lower for boys (88%) compared to girls

(92%).

43

Instruments

The questionnaire included questions regarding health, bullying, school, physical activity,

alcohol-, and drug habits. For the purpose of the study, questions on bullying, peer

aggression and psychosomatic problems were used. OBVQ was used to assess bullying

and peer aggression and is described in the section regarding study I.

The Psychosomatic Problems Scale (PSP-scale)

The PSP-scale was developed in Sweden in 1987 and 1988 and includes eight items

intended to capture psychosomatic problems among children and adolescents. The

construction of the items in the PSP scale was partly influenced by existing questionnaires

for self-rated health among adults in general populations (Hagquist, 2008). The items

included in the scale are: ‘‘had difficulty in concentrating’’, ‘‘had difficulty in sleeping’’,

‘‘suffered from headaches’’, ‘‘suffered from stomach aches’’, ‘‘felt tense’’, ‘‘had little

appetite’’, ‘‘felt sad’’ and ‘‘felt giddy’’. Five response categories are given; “never”,

“seldom”, “sometimes”, “often”, and “always” and the time frame concerns the past six

months. A low value on the PSP scale indicates less or no psychosomatic problems and a

high value indicates more psychosomatic problems. The scale has shown good

psychometric qualities such as invariance and high reliability. Psychometric analyses based

on the Rasch Model (Rasch, 1960/1980) justify the summation of raw scores. The Rasch

model offers a method for nonlinear transformation of ordinal raw scores to interval

measures that allows the responses (raw scores) from different items representing

different severity to be summated (Hagquist, Bruce, & Gustavsson, 2009).

44

Data analysis

Classification of victims

Bullying victims were identified as reporting either traditional bullying or cyberbullying.

Further, as the definition requires repetition, any indication of bullying victimization (even

occasionally/once or twice) was considered bullying. Occasional peer aggression includes

adolescents responding “occasional” to at least one type of peer aggression behavior;

frequent peer aggression includes adolescents responding “2-3 times/month”, “Once a

week” or “many times a week” to at least one type of peer aggression behavior. In the

analyses, mutually exclusive groups of victims were created and coded as bullying only,

occasional peer aggression only, frequent peer aggression only, bullying and occasional

peer aggression and bullying and frequent peer aggression.

Statistical methods

The proportion of victimized adolescents reporting psychosomatic problems was

identified using contingency tables. An OLS regression was conducted, using the

psychosomatic problems scale (PSP) as the dependent variable and the mutually exclusive

victimization groups as independent variables using dummy coding. Possible confounding

variables included gender and grade level and were controlled for using dummy coding.

Analysis of variance (ANOVA) was conducted to test for differences in mean values on

the PSP scale for victims of bullying, peer aggression and both.

Study III

Data collection

The two largest high schools in Karlstad were contacted for participation in the study.

The decision to include these particular schools was based on the wide composition of

45

students at both schools. A meeting with the principals of the two schools, one

collaborative partner from the municipality and the researcher in charge (i.e. the author of

this thesis) was arranged and information about the study was given after which the

principals agreed to include their schools in the study.

Questionnaire study

The principals were told to select three classes with equal distributions of boys and girls

(school year 7 students from school A and school year 9 students from school B) from

which the students were asked to participate in a web-based questionnaire. The students

as well as their caregivers were given written information about the study. Students were

informed that their participation in the study was voluntary, that their answers were

anonymous and that they could terminate their participation at any point. The caregivers

for the students in school year 7 were asked to sign a written consent to their child’s

participation in the study. Due to the students’ higher age in school year 9 (15 years),

caregiver consent was not needed. As it turned out to be difficult to get the consent forms

signed by all caregivers of the students in school year 7, the researcher responsible for the

data collection received phone lists for the students and contacted the caregivers to make

sure that they had gotten information about the study and reminded them to sign the

consent forms (either consenting to or declining participation). In cases where the parents

were divorced, both were contacted. The data was collected with a web-based

questionnaire in a classroom setting and the researcher in charge was on site when the

students completed the questionnaire. The students received a password and the link to

the website for the questionnaire which took approximately 10-20 minutes to complete.

Every school was revisited to collect data from students that were sick or absent during

regular data collection.

46

Focus group study

Letters were distributed to all caregivers and students informing them about the study and

asking them to contact their class teacher if they wanted to participate in a focus group

interview about bullying. They were told that the interviews would not be about personal

experiences of being bullied. Each teacher was told to form two focus groups (6-8

students in each group), one with boys and one with girls. If the number of girls and boys

who expressed interest in participation exceeded 6-8 students respectively, they were told

to draw lots to decide who would participate. The students as well as their caregivers were

given written information about the study. The caregivers of the students in school year 7

were asked to sign a written consent to their child’s participation in the study. For

students in school year 9, caregiver consent was not required. The focus group interviews

took place at the students’ schools and lasted about an hour. The moderator, i.e. the

current author, conducted the interviews assisted by a research colleague. All students

were orally informed about the aim of the study and asked to sign a consent form

beforehand stating that they were aware that their participation in the study was

voluntary, that they could choose to refrain from talking about any specific topic during

the interview, that they could terminate their participation at any point and that they

agreed to the recording of the interview. It was emphasized that there are no right or

wrong answers. To make the students feel comfortable sharing information in the group,

they were told not to discuss the shared information with other classmates. The

moderator asked the questions and followed up with questions such as “can you develop

what you just said”, “what do you mean” and “can you give any examples”. Before each

focus group interview ended the students were asked if they had anything to add or if

they thought that something important had been left out of the discussion.

Study design and participants

Data regarding adolescents’ perception of bullying were collected with both

questionnaires and focus group interviews in the spring of 2012. The study population

included students aged 13 and 15 years (school years 7 and 9) from two large schools in

47

Karlstad. The questionnaire included a total of 128 students with a total participation rate

of 82 % (61% girls). Four focus group interviews were conducted including a total of 21

students (8 girls and 13 boys), with separate groups for girls and boys. Each group

consisted of members of the same school year. The number of students in each focus

group was 6, 4, 7 and 4 respectively.

Instrument and interview guide

The web-based questionnaire included questions regarding gender, age and 24 behavior

descriptions of varying conditions based on questions of bullying behaviors used in

Olweus Bully/Victim Questionnaire (Olweus, 1996). The students were asked to answer

whether they considered the different situations to be bullying or not. The students were

also given the opportunity to comment on their responses in an open-ended format. For

the focus group interviews, an interview guide was developed. The main question of

interest in the focus group interviews was “What do you think bullying is?”

Analysis methods

First, different types of behavior that the adolescents considered to be bullying

(questionnaire data) were reported and differences were analyzed among girls and boys as

well as among school year 7 and school year 9 students using Chi square test. In total, 48

significance tests were performed. Therefore, the Bonferroni adjustment (Bland &

Altman, 1995) was applied in order to adjust for the influence of multiple significance

tests. This implies that the significance level for significant differences between girls and

boys as well as between school year 7 and school year 9 students was set to 0.05/48=

0.0010. The risk with Bonferroni adjustment is the increased possibility of making Type

II errors5 (e.g., Perneger, 1998). Second, analysis of the data from the focus group

interviews was conducted using qualitative content analysis, according to the principles

described by Graneheim and Lundman (2004). Each focus group interview was

audiotaped and transcribed verbatim. The transcription of each focus group interview was

5 The probability of accepting the null hypothesis when the alternative is true

48

read through several times to get a sense of the material. Meaning-carrying units which

responded to the aim of the study were extracted and condensed into codes. In order to

identify similarities and differences the codes were compared and then sorted into sub-

categories. As the analysis proceeded, sub-categories were subsequently clarified and

adjusted and one main category emerged. The initial coding of the transcripts was

performed by the first author, and the coded data were examined by the second and third

author for emergent sub-categories. The interpretations were compared and discussed

until consensus was reached. Comparisons were made with the context in each step of the

analysis, to verify the empirical base of the data. The pupils answered in Swedish and the

quotations cited were translated into English after the analysis.

Study IV

Data collection

An e-mail with information about the study was sent out to all teachers and special

education teachers in Karlstad. The list of e-mail addresses was provided by the

municipality and anyone listed who was no longer working as a teacher was excluded

from the study population and encouraged to contact the researcher to be removed from

the list. The questionnaire was sent out by e-mail in which every respondent received the

link to the questionnaire and a unique password randomly generated by the survey tool

esMaker (Entergate, 2009). Each password could only be used once. Each school’s

principal was contacted to make sure that the teachers had received the questionnaire.

Through the survey tool esMaker, a total of four reminders were sent out to respondents

whose questionnaires had not been registered.

Study design and participants

A cross-sectional study design was employed to collect self-report data regarding the

social and pedagogical climate in school. Data collection was carried out in March 2011

and the study population included all teachers and special education teachers in Karlstad.

49

Total response rate was 56.7% (n=344); teachers 55% (n=319) and special education

teachers 96% (n=27).

Instruments

The web-based questionnaire included background questions and the instrument PESOC,

i.e., questions concerning the social and pedagogical climate in school.

Pedagogical and Social climate in school (PESOC)

The original full-extended version of PESOC consists of 95 questions intended to tap

information about 13 domains of the social and pedagogical climate in school (Grosin,

2004). For the purpose of this study, the domain ‘The pedagogical leadership of the

principal’ (PLP) was used. Teachers were asked to rate their principal’s leadership skills on

17 different questions. The number of response categories on the PESOC items is four,

with a cumulative response format: “strongly disagree”, “disagree”, “agree”, and “strongly

agree”. The original instrument consists of 19 questions regarding PLP. After consulting

with principals at two local secondary schools and with the approval of the author of the

instrument (Lennart Grosin), two questions that did not correctly reflect the routines and

organization of the schools were removed. In addition, small changes in wording were

also made regarding two of the original questions.

Data analysis

The Rasch Model

The psychometric properties of the PESOC-PLP scale were examined using Rasch

analysis (Rasch, 1960/1980). A fundamental feature of the Rasch model is the property of

invariance, which means that it is possible to test if the items in an instrument work in the

same way along the latent trait as well as across different sample groups (Andrich, 1988).

Lack of invariance across sample groups like gender is called Differential Item

50

Functioning (DIF) and can be detected by ANOVA of standardized residuals (Hagquist

& Andrich, 2004). The model requires a probabilistic Guttman pattern in the structuring

of responses, meaning that for the same person ability, the probability to endorse an easy

item has to be higher than the probability to endorse a more difficult item (Andrich, 1985;

Hagquist et al., 2009). The correspondence between the data and the Rasch model can be

examined with formal statistics, e.g., chi-square statistics based on comparisons between

observed and expected values. Also graphical representations are valuable tools for

examining the operating characteristics of the items. Item characteristics curves (ICC) are

graphs that show the expected value curve which predict the item scores as a function of

the item- and person-locations on the latent trait (Hagquist, 2001). Further, good targeting

is essential for good measurement. Mistargeting implies lower reliability and may cause

problems to differentiate people along the latent scale (Hagquist et al., 2009). The Rasch

model is also sensitive to the categorization of items which means that possible problems

with the intended successive order of the response categories can be detected, i.e.,

reversed item threshold locations (Andrich, 2005; Hagquist & Andrich, 2004). Further,

signs of multidimensionality may imply violation of the principles of local independency

(Marais & Andrich, 2008; Tennant & Conaghan, 2007). Given that the data fit the Rasch

model, summation of the responses (raw scores) of the different items is justified. The

unidimensional Rasch model offers a method for nonlinear transformation of ordinal raw

scores to interval measures (Andrich, 1988).

In the current study, different aspects of the PESOC-PLP scale were

examined. All analyses were performed using the item analysis program RUMM2030

(Andrich, Sheridan, & Luo, 2010). The Rasch model was used to judge whether the

PESOC-PLP scale meets the requirements of unidimensionality, invariance and proper

item categorization. First, the threshold ordering of all 17 items was examined to detect

possible problems with the intended successive order of the response categories. Second,

targeting, i.e., the distribution of persons relative to the item thresholds on their common

latent trait, was examined. Third, the fit of items to the Rasch model was examined with

graphical illustrations as well as formal test statistics, e.g., chi-square statistics based on

comparisons between observed and expected values. Fourth, unidimensionality was

examined by principal component analysis of the item residuals and independent t-tests of

differences between person locations values from different subsets of items. Fifth, new

51

analyses of two subsets of items were performed and included: item fit analysis, DIF

analysis (ANOVA), and targeting. Sixth, the possible impact of the nested data on

measurement was examined by analysis of response dependence between persons using

the Rasch model.

Ethical considerations

Ethical considerations were taken into account and ethical issues were discussed in the

research team throughout the project. All study participation was voluntary and based on

informal consent. Written information about the aim of the study and assurance of

respondents’ anonymity, confidentiality and the right to withdraw at any point was given.

In Study III, caregivers were contacted to make sure that they had received information

about the study, but were not requested to consent to their child’s participation. Since the

data collections did not include sensitive personal data, the project was not regulated by

the Swedish act concerning Ethical Review of Research Involving Humans. The project

was hence not reviewed by the regional committee of ethics. However, the local Ethical

Review group at Karlstad University reviewed the project and presented no objections

(Study I+II, Dnr. 2009/623, Study III, Dnr. 2012/193, Study IV, Dnr. 2010/707).

52

RESULTS

Prevalence rates of peer victimization

The prevalence rates of peer victimization were similar for the two data collection points

(2009 and 2010); 10% of the students reported being traditionally bullied, 8% reported

being cyberbullied, while 14% reported victimization by either type of bullying at least

occasionally. A somewhat higher proportion of girls, compared to boys, reported being

traditionally bullied (10% vs. 9%) as well as cyberbullied (10% vs. 5%). The prevalence

rates for the different types of peer aggression behavior were generally higher compared

to the prevalence rates for bullying. When combining all five types of peer aggression

behavior into a composite measure, a total of 29% of the adolescents reported being an

occasional victim of at least one type of peer aggression behavior (33% for girls and 24%

for boys), while 13% reported frequent victimization by at least one type of peer

aggression behavior (15% for girls and 12% for boys). Hence, the prevalence rates for the

different types of peer aggression behavior were higher for girls, compared to boys, with

the exception of being hit or kicked (Figure 1).

Figure 1. Prevalence rates of bullying and peer aggression. Cut point “occasionally or more often”

(including response categories occasionally, 2-3 times/month, about once a week and several times/week)

(N=2568).

10 10 15

27

18

9

25

6 9

5

12

22

11 14 16

6 10 8

14

25

15 12

21

6

0

10

20

30

40

50

Traditionalbullying (TB)

Cyberbullying(CB)

EitherTB or CB

Scoffed at Left out Hit, kicked Lies or rumors Threatened

Girls Boys Total

%

53

Study I

First, the degree to which peer aggression was captured by a measure of bullying was

examined. Among students reporting bullying, a total of 33.6% reported being

occasionally victimized by at least one type of peer aggression behavior while 51.6%

reported frequent peer aggression. Being frequently scoffed at, ridiculed or called nasty

names was reported among 34.6% of the bullying victims while 15.6% reported being

frequently threatened or forced to do something. Second, the degree to which bullying

was captured by a measure of peer aggression was examined. Among students reporting

occasional victimization by at least one peer aggression behavior, 16.8% reported being

bullied, whereas 49.6% of students who reported frequent peer aggression also reported

being bullied. Further, among students reporting frequent threats or coercion, 84.1% also

reported bullying victimization, whereas only 50.0% of the adolescents reporting being hit

or kicked as often reported bullying victimization.

The concordance between measures, i.e., the proportion of students

reporting both bullying and peer aggression, was higher when the cut point for peer

aggression was set to “2-3 times/month or more” compared to “occasional”. In addition,

among students who did not report bullying, 8.7% reported being victimized by at least

one frequent type of peer aggression behavior while 27.5% reported being victimized by

at least one occasional type of peer aggression behavior. This indicates that many students

report peer aggression without reporting bullying. Third, the concordance between a

measure of bullying and measures of peer aggression was further examined using phi

squared coefficients, expressing the association between two binary variables. Phi square

coefficients showed that 18% of the variance in either measure was accounted for by the

other measure. Further, a higher proportion among boys compared to girls who reported

peer aggression also reported bullying. This means that even though girls tend to report

that they have been exposed to peer aggression more often compared to boys, boys more

often rate peer aggression in combination with bullying. In fact, phi square coefficients

showed that the variance in measures explained by the other measure was higher for boys

(19.4%) compared to girls (16.6%). For all subgroups examined, chi-square tests revealed

significant relationships (p < .001) between bullying and all specific types of peer

aggression (e.g., bullying and a composite measure of occasional peer aggression, χ2(1, N

54

= 1734) = 207.77, p < .0001 and bullying and a composite measure of frequent peer

aggression, χ2(1, N = 1734) = 312.45, p < .0001). To illustrate the overlap between

measures, a Venn diagram was used to show how many victims were captured by a

measure of bullying, a measure of repeated peer aggression, and how many were captured

by both measures (Figure 2). A total of 302 adolescents were identified as victims of peer

victimization; 13.1% bullying victims, 95% CI [9.3%, 16.9%], 43.6% victims of repeated

peer aggression, 95% CI [38.0%, 49.2%] and 43.3% victims of both, 95% CI [37.4%,

48.6%]. By comparing individual responses on the two measures, the results from paper I

indicate that a measure of bullying and a measure of peer aggression capture somewhat

different pupils.

Figure 2. A Venn diagram illustrating peer victimization captured by a measure of bullying and a measure

of repeated peer aggression (using cut point 2-3 times/month or more for peer aggression).

Study II

The results showed that a measure of bullying captures 52.9% of adolescents with

psychosomatic problems and a composite measure of peer aggression captures 31.5% of

adolescents with psychosomatic problems at the lower frequency cut point and 52.8% of

adolescents with psychosomatic problems at the higher frequency cut point. Higher

proportions of psychosomatic problems are captured when the bullying measure and the

measure of peer aggression are used in combination. Overall, somewhat lower

proportions of psychosomatic problems are captured at the lower frequency cut point for

the peer aggression measure. Further, girls reported more psychosomatic problems

Repeated peer- aggression (43.6%)

Bullying (13.1%)

Both (43.3%)

55

compared to boys. Lastly, 14.6% among non-victims report psychosomatic problems

(Figure 3).

Figure 3. Proportion (%) of adolescents reporting psychosomatic problems on or above the 75th percentile

among victims of: bullying, peer aggression, both bullying and peer aggression, and neither, distributed by

gender (N=2568). Note: PA=Peer Aggression. Occasional peer aggression includes adolescents

responding “occasional” to at least one peer aggression behavior; frequent peer aggression includes

adolescents responding “2-3 times/month”, “Once a week” or “many times a week” to at least one peer

aggression behavior.

Second, an Ordinary Least Squares regression was conducted to analyze the unique

associaton with psychosomatic problems (values on the PSP scale) among five mutually

exclusive groups of victims identified by a bullying measure and a composite measure of

peer aggression. Table 3 shows that all regressors representing peer victimization reveal

significant effects on psychosomatic health using no peer victimization as the reference

category. There are clear differences between some of the estimates of the peer

victimization regressors. To test for differences in mean values on the PSP scale between

the mutually exclusive groups of victims identified by a bullying measure and a composite

measure of peer aggression, an ANOVA was conducted. The ANOVA test with one set

of 2 (gender) × 6 (victimized groups) showed significant main effects for victimization

types (F = 75.80, p <0.001) and a main effect for gender (F=33.01, p <0.001). No

interaction effect (gender × victimization type) appeared. Table 4 presents mean-level

differences on the PSP-scale for no victims and victims identified by a measure of

22

68

44

64 59

75

8

33

14

38

19

44

15

53

32

53

44

62

0

20

40

60

80

Neither Bullyingnor PA

Bullying Occasional PA Frequent PA Bullying andoccasional PA

Bullying andfrequent PA

Girls Boys Total

%

56

bullying, a measure of peer aggression and both measures. The ANOVA showed no

difference between the bullying measure and the measure of occasional peer aggression or

between bullying and frequent peer aggression regarding the capturing of psychosomatic

problems. A measure of bullying and a measure of frequent peer aggression in

combination captured significantly higher levels of psychosomatic problems. i.e., higher

values on the PSP scale, compared to a bullying measure or a peer aggression measure

used in isolation (p<0.001). In the tentative analyses based on the mutually exclusive

categories, the number of observations was largely reduced making it hard to draw any

confirmative conclusions. The main results from paper II show that no statistically

significant differences in the capturing of psychosomatic problems among adolescents

were found between a measure of bullying and a measure of peer aggression.

Table 3. Ordinary Least Square regression showing psychosomatic health problemsa

regressed on bullying and peer aggression: Unstandardized coefficients (B) and confidence

intervals (95%) (N=2568)

B 95% C.I

Bullying only 1.017 (0.589-1.444)

Occasional peer aggression only

0.580 (0.458-0.703)

Frequent peer aggression only

0.991 (0.789-1.192)

Bullying and occasional peer aggression 0.986 (0.765-1.207)

Bullying and frequent peer aggression

1.653 (1.453-1.853)

Non-victimized (no bullying and no peer

aggression)

0

Adjusted R2=0.215

Note: gender and grade level were included as control variables

aValues on the PSP scale: maximum value = 4.417; minimum value = -4.954

57

Ta

ble

4.

Res

ult

s o

f A

NO

VA

ex

amin

ing m

ean

val

ues

on

th

e P

SP

-sca

le b

etw

een

dif

fere

nt

pee

r v

icti

miz

atio

n m

easu

res

Typ

e o

f

mea

sure

No

vic

tim

izat

ion

(n=

12

38

)

Bu

llyin

g

(n=

32

)

Occ

asio

nal

PA

(n=

56

0)

Fre

qu

ent

PA

(n=

15

9)

Bu

llyin

g +

Occ

asio

nal

PA

(n=

13

0)

Bu

llyin

g +

Fre

qu

ent

PA

(n=

16

3)

F (

df)

Mea

n

(SD

)

-1.6

3a

(1.2

9)

-0.7

1b

cd

(0.9

9)

-1.0

0b

(1.1

4)

-0.5

9cd

(1

.23)

-0.5

6d

(1.0

2)

0.0

8e

(1.7

4)

78

.93

(5)

No

te.

Mea

ns

wit

h d

iffe

ren

t su

bsc

rip

ts d

iffe

r si

gnif

icantl

y a

t p

< .

05.

Mea

ns

shar

ing o

ne

of

the

sub

scri

pts

do

no

t d

iffe

r. F

or

exam

ple

, cd

do

es n

ot

dif

fer

fro

m c

or

d,

but

cd

dif

fers

fro

m a

58

Study III

In study III, a questionnaire and four focus group interviews were used to explore

perceptions and understandings of bullying among adolescents in Sweden. The results

from the questionnaire study showed that among school year 9 students, significantly

more girls compared to boys reported the following behaviors to be bullying: ‘repeatedly

write mean things on someone’s facebook page or in a chat’ (p≤0.001), ‘sending several

mean text messages to the same person’ (p≤0.001), ‘a group of students calling someone

mean things’ (p≤0.001) and ‘writing mean things to someone online who does not have

many friends (p≤0.001). Similar results were found among school year 7 students. While

‘hitting someone for fun’ was reported as bullying twice as often among boys in school

year 7 (15%) compared to girls in school year 7 (7%), the differences were non-significant.

The results revealed that in general, students in school year 9 more often reported the

different behaviors as bullying compared to students in school year 7. Students in school

year 9 reported behaviors such as social exclusion to be bullying more often compared to

students in school year 7, e.g., ‘constantly ignoring someone or not talk to this person’,

and ‘during recess decide who can participate (or not participate) in games or other

activities’. Comments regarding their responses included circumstances under which the

adolescents were more likely to consider a behavior as bullying, namely: the effect on the

victim (e.g., “I think it’s bullying when the person being exposed think it’s bullying… If

it’s for fun it doesn’t have to be bullying, as long as no one feels bullied”); if both parties

are in on it (e.g., “posting an embarrassing photo can be okay, if the person is in on it, or

if it’s posted in a Facebook-group where similar photos are posted”); repetition (e.g.,

“much of it is bullying but it depends if the behavior is repeated”); and intent (“it depends

whether you mean it or not”).

One category and three sub-categories emerged from the analysis of the

focus group interviews. The category was, “The core of bullying” and the sub-categories

included, “Behavior descriptions”; “Self-interpretation”; and “Something hurtful” (Figure

4).

59

Figure 4. Category and sub-categories that emerged as important for understanding adolescents’ definition of

bullying.

The core of bullying

It emerged that the adolescents used different behavior descriptions to define bullying

behaviors. These included teasing, giving nasty comments, oppressions or threats with

words, hitting, pushing, talking behind someone’s back, whispering, spreading rumors,

giving glances, ignoring or leaving someone out of the peer group. Bullying also included

malicious behavior on social media sites such as facebook and Twitter. Physical

aggression was types of bullying behavior described to be more common among boys

while girls were described to more often bully with words and behind someone’s back.

Jokes that were repeated were automatically considered bullying and in contrast to

bullying, occasional arguments or fights were solved right away and all involved had an

equal share in the argument and in the chance of “winning”. The adolescents described

that it was hard to draw a line for when a behavior should be considered bullying and that

it was very much a question of self-interpretation. The adolescents mentioned that the

boundary for bullying is when someone stops laughing, which could be different for

different people. When someone takes offense and feels bad as a consequence of the

incident, it should be considered bullying. It was expressed that not being able to interpret

the tone of voice or facial expression made it harder to separate jokes from bullying,

especially over the internet. According to the adolescents, bullying is something hurtful that

Something

hurtful Self-

interpretation

Behavior

descriptions

T

The core of

bullying

60

leads to negative health consequences. Verbal bullying in particular leaves scares that lead

to low self-esteem, feelings of not being good enough and sadness. Being different and

standing out could mean that no one wants to be with you and that you sometimes have

to stand the bullying in order not to be alone. According to the adolescents, the comfort

of hiding behind a computer screen often made bullying incidents online more aggressive

and brutal compared to bullying in real life. Despite this, the adolescents thought that it

was easier to dismiss incidents happening online. In sum, the results from paper III show

that the adolescents view a negative effect on the victim as a criterion for defining

bullying and not only as a consequence of bullying.

Study IV

General level of analysis (17 items)

The psychometric testing in the study concerned the construction and evaluation of the

PESOC-PLP scale, focusing on the operating characteristics of the items. The value of

the person separation index (PSI) for the 17 items was 0.90242, indicating good

separation of persons and high power of the test statistics. Targeting was examined by

person-item threshold distribution, showing the locations of the item threshold parameter

estimates relative to the distribution of the persons for the items. The person locations

are positively skewed with a mean of 1.173, indicating positive perceptions of the

principal as a pedagogical leader. At the positive end of the scale, there are no thresholds

matching the person locations indicating that there is information missing about these

persons. None of the 17 items showed reversed thresholds which means that the

response categories work as intended and just with a few exceptions the distances

between the threshold estimates were rather similar. The spread of the item location

values corresponds to different severity between items with respect to the principal as a

pedagogical leader.

Item 14 shows misfit according to the Chi square statistics (p < 0.001) and

the residual fit values. In Figure 5, the item characteristic curves (ICC) reveal that item 14

shows lower discrimination than expected according to the Rasch Model. This implies

61

that individuals located at the beginning of the scale tend to score higher than expected,

and individuals located in the right end of the continuum tend to score lower than

expected on this particular item. Deleting item 14 increased the person separation index

value slightly (from 0.902 to 0.904) without affecting targeting, thus indicating that this

item is redundant in relation to the other 16 items. The item could be removed given no

loss in content validity. Principal component analysis (PCA) of the item residuals

indicated multidimensionality. Based on the PCA of residuals, t-tests and analysis of item

content, the entire set of items were split into two sub dimensions of PLP (table 5) that

were subjected to further analyses.

Figure 5. Item characteristic curves for item 14 “Teaching methods are first of all chosen and judged based on

whether they lead to fulfillment of students’ learning”

Finer level of analysis (two sub dimensions)

The value of the PSI for sub dimension 1 (6 items) was 0.798 and for sub dimension 2 (10

items) was 0.852. None of the items showed reversed thresholds. One item in sub

dimension 2 showed a fit residual above 2.5 (item 10). However, deleting item 10

decreased the PSI value and it was retained. Two separate DIF analyses based on gender

and years of employment were included. At the Bonferroni adjusted significance level

(0.001) one item in sub dimension 1 (item 5) and one item in sub dimension 2 (item 7)

showed DIF with respect to years of employment at current school. Teachers that had

62

been working 4 years or less at their current school scored higher on item 5 and lower on

item 7 compared to teachers that had been working 5 years or more at their current

school given the same location on the latent trait. For sub dimension 1, the person-item

threshold distribution showed that at the negative as well as at the positive end of the

scale, there were no thresholds matching the person locations. This means that there is a

lack of information about the persons in these areas (concerning their perception of the

principal as a pedagogical leader). The mean value of the person’s location values is 0.214.

For sub dimension 2, the person locations were positively skewed with a mean value of

1.814. There are areas where there are no thresholds matching the persons, indicating that

there is information missing about these persons. Although the persons were dislocated

towards the upper end of the continuum representing better pedagogical leadership, the

item thresholds were spread along the latent variable. The thresholds at the lower end of

the continuum were matching the persons well.

Table 5. Two subset of items as indicated by principal component analysis

Item Subset

1. The principal initiates regular observations of teachers’ teaching methods and feedback

discussions with the teachers

1

2. The principal takes the ultimate responsibility for teaching quality

3. The principal takes responsibility for ensuring the performance levels of instruction and

criteria for assessing students' knowledge

1

1

5. The principal place high priority on pedagogical activities over administrative tasks and

contacts outside school

1

9. The principal is often seen in school

17. The principal gives regular feedback on my work in the classroom

4. Assistant principals and other leaders in the school essentially agree with the principal

concerning basic educational issues and school goals

1

1

2

6. The principal often show strength in decision-making 2

7. The principal’s expectations of me as a teacher are high 2

8. The majority of teachers’ understanding of the school goals and ways of working are in

agreement with the principals goals and ways of working

2

10. Careful preparation is given to the planning of teacher conferences at the school and in-

service training for the staff

2

11. The principal is available for helping to solve conflicts among teachers 2

12. The principal place high priority on pedagogical activities in contacts with teachers,

students and parents

2

13. The principal is available to discuss issues related to education 2

15.The principal is actively involved in teachers' professional development 2

16. The principal is available for discussions on teaching methods 2

Note: Subset 1= “Direct pedagogical leadership”; Subset 2= “Indirect pedagogical leadership”.

63

DISCUSSION

Discussion of results

The main results of the four studies and their contributions can be summarized in the

following way: Study I and II show that while a measure of bullying and measures of peer

aggression partly capture different adolescents, both capture adolescents reporting the

same level of psychosomatic problems. The present thesis thus shows that the previously

established association in research between bullying and mental health problems also

applies to peer aggression. This indicates that using only one single measure of bullying is

likely not only to underestimate the overall prevalence rates of peer victimization but also

the number of children experiencing psychosomatic problems in combination with peer

victimization. Study III shows that the adolescents’ understanding and definition of

bullying did not only include the traditional criteria of repetition and power imbalance,

but also a criterion based on the health consequences of bullying. In other words, a single

but hurtful or harmful incident could also be considered bullying irrespective of whether

the traditional criteria were fulfilled or not. This adds to the existing literature by showing

that adolescents consider the victim’s feelings of hurt and harm to be a defining criterion

for bullying and not only a consequence of bullying. The results are important as adults’

and adolescents’ differing perception of bullying can affect the preventive and anti-

bullying work in school. Measurement methods assessing effective leadership are

important since strong school leaders are vital not only for the anti-bullying work in

schools, but also for the creation of a positive school climate and academic success. The

results in study IV show that the PESOC-PLP scale holds satisfying psychometric

properties, capturing the multidimensionality of the pedagogical leadership construct.

Since PESOC has been widely used for school improvement efforts in Sweden but have

not previously been exposed to rigorous psychometric testing, the evaluation of the

PESOC-PLP scale is a contribution to previous literature. By highlighting methodological

shortcomings in research concerning social relations in school, the objective of the thesis

is to raise awareness of how to ensure sound assessment of aspects associated with

negative experiences among adolescents in school. The results partly support previous

studies but also add to the understanding of peer victimization among children and youth,

64

and of the principal as a pedagogical school leader, as well as how these aspects can be

measured.

Overlap of bullying, peer aggression and psychosomatic problems

The prevalence rates of bullying among Swedish adolescents are relatively low. In the

current thesis 10% of the students reported being traditionally bullied, 8% reported being

cyberbullied, while 14% reported victimization by either type of bullying at least

occasionally. Further, a total of 29% of the adolescents reported being an occasional

victim of at least one type of peer aggression behavior, while 13% reported being

frequently victimized by at least one type of peer aggression behavior. Depending on cut

points for bullying, similar or somewhat higher prevalence rates have been reported in

previous Swedish studies on bullying (Erling & Hwang, 2004; e.g., Frisén et al., 2008;

Slonje, Smith, & Frisén, 2012; Thornberg & Knutsen, 2011), while studies examining

prevalence rates of peer aggression are scarce. Hence, the results in this thesis show that a

measure of peer aggression identified a greater number of adolescents compared to a

measure of bullying. Similar findings have been reported in earlier research showing that

measurement methods including specific aggressive behaviors result in higher prevalence

rates compared to single measures of bullying that include a definition of bullying (Sawyer

et al., 2008; Stockdale et al., 2002). However, in addition to previous literature, the current

thesis also shows that the two measures capture partly different pupils (Study I). When

the concordance in individual responses between a single bullying measure and specific

types of peer aggression behavior was analyzed it indicated that a measure of bullying and

a measure of peer aggression used in isolation fail to capture many adolescents victimized

by peers. These are important findings as they indicate that there are many adolescents

exposed to repeated incidents of peer aggression without considering themselves to be

victims of bullying or reporting it. This implies that many adolescents victimized by peers

will not be captured by the most commonly used tool for assessing peer victimization, i.e.,

a measure of bullying.

The importance of these findings are further highlighted by the results in

Study II showing that regardless of measurement method, victims of peer victimization

65

reported significantly more psychosomatic problems compared to non-victimized

adolescents. There were no statistically significant differences in the mean values of

psychosomatic problems between adolescents captured by a bullying measure and a peer

aggression measure, implying that a measure of bullying and a measure of peer aggression

capture the same proportion of adolescents with psychosomatic problems. While

previous research has established an association between measures of bullying and mental

ill health (Fekkes, Pijpers, Fredriks, Vogels, & Verloove-Vanhorick, 2006; Gini & Pozzoli,

2009), the current study demonstrate similar association for measures of peer aggression.

Considering that nearly four times as many victims of peer victimization were identified

by a measure of repeated peer aggression, compared to a measure of bullying, using only

one single measure of bullying is therefore likely not just to underestimate the overall

prevalence of peer victimization but also the number of children experiencing

psychosomatic problems in combination with peer victimization. Considering that

psychosomatic problems in adolescents is a matter of growing concern (Hagquist, 2010)

and also increases the risk of later problems such as anxiety and depression (SBU, 2010),

these results highlight the importance of using methods with the potential to identify all

children and adolescents experiencing peer victimization. Contrary to research claims that

it is necessary to clearly distinguish bullying from other types of peer victimization (Greif

& Furlong, 2006; Hunter et al., 2007), the results from Study I and II imply that

prevention efforts based solely on a single bullying measure may not target all children at

risk.

So, why does a measure of bullying fail to capture a great number of

adolescents that are victimized by peers and at risk of mental health problems? One

explanation could be the feelings of shame involved in admitting to being bullied

(Bosworth et al., 1999) or that one’s understanding of what it means to be bullied is

different from what is stated in the definition (Thornberg & Knutsen, 2011). In

comparison to a measure of bullying, it may be easier to admit to being victimized by

peers on a measure of peer aggression without having to label oneself as a victim of

bullying (Bosworth et al., 1999; Greif & Furlong, 2006). While a measure of peer

aggression may include single incidents of harmless “joking around” (Solberg & Olweus,

2003), the higher prevalence rates of peer aggression could also be that it captures other

forms of harassment that the current definition of bullying screens out. If clearly defined

66

and operationalized, a prevalence estimate gives a clear picture of the full extent and

distribution of the social and mental health consequences of peer victimization and is a

useful measure to evaluate mental health interventions in school (Greif & Furlong, 2006;

Solberg & Olweus, 2003). However, if definitions are inconsistent and concepts are

poorly operationalized, there is a great risk of incorrect estimates of prevalence rates.

Children’s and adult’s differing understandings

Incorrect estimates of prevalence rates could also be a consequence of children’s and

adult’s different ways to define bullying (Varjas et al., 2008). An important step to

understand bullying behavior among children is to ask them about their perceptions,

experiences and explanations (Thornberg, 2015). If children are asked to indicate whether

they have been bullied in terms of a predetermined view defined by adults, many incidents

that do not fit this definition may not get reported. The results from Study III showed

that the adolescents’ understanding and definition of bullying did not only include the

traditional criteria of repetition and power imbalance, but also a criterion based on the

health consequences of bullying. That is, a single but hurtful or harmful incident could

also be considered bullying irrespective of whether the traditional criteria were fulfilled or

not. This implies that the traditional criteria included in most definitions of bullying may

not fully reflect adolescents understanding and definition of bullying. Assessments of

bullying behaviors that ask adolescents to strictly adhere to the traditional definition of

bullying might not identify all adolescents experiencing peer victimization and therefore

not provide estimates of prevalence rates reflecting adolescents’ own understanding of

bullying. Children often judge bullying behaviors as something wrong using moral reasons

such as the harm that the actions cause (Thornberg, 2010; Thornberg et al., 2014). The

results in the current thesis add to the existing literature by showing that adolescents

consider the victim’s experiences of hurt and harm as a criterion for defining bullying and

not only as consequences of bullying. This may be of special relevance for the

identification and classification of bullying incidents on the internet where devastating

consequences have been reported from single incidents and the use of the traditional

67

criteria of intent, repetition and power imbalance may not be as relevant as for traditional

bullying.

Even though earlier studies have shown that children may not share adult’s

understanding of bullying (e.g., Varjas et al., 2008), children’s understandings have had

little impact on the definition. However, the social context in which today’s adolescents

are brought up may have changed the perception of what is considered unacceptable

behavior (Abt & Seesholtz, 1994) and this change serves to further stress the need to let

children’s voices be heard. Expressing their voice in matters that concerns them is every

child’s right, according to the UN Convention on the Rights of the Child (UNCRC)

(General Assembly, 1989). The convention, which was signed by Sweden in 1990, clearly

states that every child has the right to be taken seriously and that every adult has the

obligation to find out what children’s opinions are and to take them into account.

However, school children’s voices relating to everyday school life is often suppressed and

its value minimized (Thornberg & Elvstrand, 2012). Any type of peer victimization

reported by children need to be taken seriously and dealt with. Despite school’s legal

obligations to take action to prevent and stop degrading treatment of children, a Swedish

study showed that 35% of bullied students reported that they had not received help from

any adult in school (Frisén et al., 2008). Adults may not consider certain incidents as

bullying, which may in part explain their unwillingness to intervene in bullying situations

(Bauman & Del Rio, 2006; Varjas et al., 2008). Research has also shown that the

perceived seriousness of a behavior is a big predictor of teacher intervention (Yoon,

2004). Hence, adults’ perceived neglect to intervene could indicate that they may not

correctly perceive the harm in certain situations.

A focus on the effect on the victim rather than the intent of the bully has

been put forward by a few researchers. Griffin & Gross (2004) acknowledge that the

definition of bullying is typically focused on the actual behavior or probable intention of

the bully rather than on the perception or experience of the victim. They argue that more

comprehensive definitions and methodology that include both the behaviors of bullies

and the subjective experiences of victims could be more useful. Goldsmid & Howie

(2013) argue that the established association between distress and peer victimization may

justify considering the negative effect on the victim as a criterion for bullying. However,

the authors also acknowledge problems with including a subjective judgment in

68

measurement methods because of children’s differences in vulnerability. Rigby (2007)

illustrates this by stating that what appears to one person as playful teasing can in some

circumstances be extremely hurtful. The definition of bullying states that if the

“aggressor” knows that the behavior is perceived as distressing and is aware that the

victim wants it to stop—and still continues—that it is “bullying”. However, even if we

may think that the victim is overly sensitive, the action cannot be ignored (Rigby, 2007).

Self-reported victimization needs to be taken seriously since it has shown to associate

with emotional problems (Scholte et al., 2013). By including the victim’s experience of

hurt and harm as a criterion for defining bullying and not only as consequences of

bullying, the results in Study III confirms those of Study I and Study II indicating that a

measure of bullying based on the traditional definition of bullying may not capture all

children at risk of harm.

Alternative terms and definitions

The definition of bullying has been a topic of debate in much of the bullying literature. By

restricting bullying to include only intentional, repetitive aggression including some sort of

power balance, the definition aims to capture the most harmful incidents. The definition

implies that all three criteria need to be present for the behavior to be bullying (Greif &

Furlong, 2006). This means that many incidents causing distress in children are not

included when peer victimization assessments are restricted to the traditional definition of

bullying. Adolescents may experience repeated acts of physical violence without it being

considered bullying if the offender is of equal status or strength. Further, adolescents may

receive a very threatening message from an anonymous sender, but it would not be

defined as bullying unless it was repeated. For this reason, it has been argued that the

concept of bullying may be too narrow to use in efforts to combat peer victimization.

Finkelhor et al. (2012) suggest that the concepts of peer victimization and peer aggression

have much more openness and flexibility compared to the concept of bullying and

therefore should be used in addition to bullying. They define peer victimization as peers

acting outside of the norms of appropriate conduct causing harm and peer aggression as

acts intended or perceived as intended to cause harm. Since most schools in Sweden have

69

adopted a zero tolerance against bullying it would be surprising to find them accepting

any form of peer victimization. Williams and Stelko-Pereira (2013) further suggest that the

concept of bullying should be expanded not only to peer victimization but to the wider

notion of school violence, adopting the WHO definition (Krug, Dahlberg, Mercy, Zwi, &

Lozano, 2002):

School violence is the intentional use of physical force or power, threatened or actual,

against oneself, another person, or against a group or community, that either results in

or has a high likelihood or resulting in injury, death, psychological harm,

maldevelopment or deprivation (p. 5).

Olweus critique against using definitions that do not include the criteria of intent, power

imbalance and repetition is that they include innocent arguments. So, the traditional

definition excludes harmless brawl by excluding behaviors that are “done in a friendly or

playful way” (Olweus, 1996). However, such a restriction also excludes one-time serious

incidents. Excluding behaviors “done in a friendly and playful way” could also be used as

an excuse in favor of the bully, i.e. “I was only joking…” As the results in Study III

showed, adolescents find it very hard to separate jokes from serious behavior and they

also considered repeated jokes to be bullying. Nearly a third of all boys in school year 9

considered hitting someone for fun to be bullying. While it may be easier to assess your

own behavior towards someone else as a joke, it may be very hard for the receiver to

interpret the intent behind the behavior. Although showing indifference towards the bully

can be a strategy for handling a bullying incident, it can also reflect badly on the victim

and be interpreted as a lack of assertiveness. This dilemma is not covered by the current

definition of bullying.

Bullying as a legal term in the school context

In the work to reduce bullying, the school has a legal obligation to counteract all forms of

abusive behavior and harassment and to investigate and take action if it comes to their

knowledge that a child or a student claims to have been subjected to abusive treatment or

harassment (Skolinspektionen, 2010). However, adults’ and adolescents’ differing

understanding of what it means to be bullied may hinder adult intervention and the anti-

70

bullying work in school. Single but hurtful or harmful incidents of peer victimization may

go unnoticed as such incidents do not fit the traditional view of bullying and may not be

considered to be bullying by adults. The results of the current thesis provide empirical

support to the Swedish Education Act (2010:800) where the word bullying is considered

too narrowly defined to be included as a term on its own. Instead, bullying is included

under the term degrading treatment. One of the reasons is that the legislators believe that

incidents should not have to be repeated to qualify as a violation of the law. Once is

enough. By showing that measurement methods of peer aggression captures the same

level of psychosomatic problems as a measure of bullying, the results of the current thesis

is an indication that prevention efforts based on a single measure of bullying may not

target all children at risk.

Another reason why degrading treatment is used in the Education Act

instead of bullying is related to who defines the action. The person who violates does not

necessarily have malicious intent, but it is the offended person who determines what is

considered degrading treatment. Insults may thus take place without malice. The number

of people who define themselves as being exposed to degrading treatment has increased

rapidly in the last couple of years. In 2008, the Swedish School Inspectorate received

1,290 reports of abuse while the number had risen to 2,968 in 2012. The most common

reason for reporting abuse is that the school has not done enough in situations where

students feel that they have been exposed to degrading treatment.6 Degrading treatment is

a broad term that captures any behavior that the victim considers degrading while the

purpose of using the term bullying is to single out the most harmful types of degrading

treatment. The term degrading treatment is based on the understanding that situational

factors, rather than individual factors, are more important when describing different

behaviors (See SNAE, 2009b). Accordingly, victimization needs to be embedded within

the culture of the school where it is taking place. In order to reduce prevalence rates, the

focus needs to be on changing the system rather than the individuals within it (Monks et

al., 2009). The most important agents in the system are the principals, as they set the tone

for the anti-bullying work in schools (Dake et al., 2004). Receiving feedback on their

leadership could make principals more aware of strengths and weaknesses and what they

can contribute to creating a more positive school climate with positive social relations.

6 http://www.skolinspektionen.se/sv/Anmalningar/Att-anmala-till-Skolinspektionen/Statistik/

71

However, evaluated tools for assessing successful leadership in Swedish schools are

lacking.

Evaluation of the PESOC-PLP scale

The psychometric analyses of the PESOC-PLP scale (Study IV) showed that the

pedagogical leadership of the principal (PLP) consisted of two sub dimensions and

analysis of item content revealed two types of leadership; direct pedagogical leadership

and indirect pedagogical leadership. Direct pedagogical leadership is defined as

responsibility of teaching methods and quality, and indirect pedagogical leadership is

defined as ability to lead staff and building relationships. It is stated that the items

comprising the PESOC-PLP scale are intended to capture both the structure and the

culture of a school (Grosin, 2004). In Study IV, the subset “direct leadership” includes

leadership variables such as regular observations of teachers’ teaching methods, giving

feedback, responsibility for teaching quality and quality instruction, i.e., the principal’s

direct actions for school improvement. These are variables that have been identified as

important for the structure of the school and academic goal fulfilment (Höög, Johansson,

& Olofsson, 2009). It has been suggested that leaders make a difference on student

achievement through a clear focus on academic and learning goals (Robinson, 2007;

Stronge et al., 2008), which could be accomplished by a direct leadership style. Further,

the subset “indirect leadership” includes leadership variables such as agreement among

school personnel concerning goals and visions, high availability to resolve conflicts, and

collaboration with teachers, i.e., indirect actions for school improvement through teacher

influences. These variables rather seem to relate to school culture and could be seen as

important for social and democratic goal fulfilment (Höög et al., 2009). Successful

principals can contribute to enhancing both academic and social goals by understanding

and addressing issues related to the culture and structure of the school (Höög et al., 2005).

The results and content of the two subsets of the PESOC-PSP scale

correspond well with the aim to incorporate the structure and culture of the school and

with the Swedish Education Act, which embraces both social/democratic and academic

goals (SFS, 2010:800). This indicates that the PESOC-PLP scale will most likely measure

72

both the direct effects of school leadership on student outcomes as well as indirect effects

through teacher influence. This is of interest as some of the disparities in current

measurement methods of school leadership involve the distinction between direct and

indirect effects on student outcomes and the fact that relatively few studies consider the

connection between direct instructional leadership and student outcomes (Seashore Louis

et al., 2010). The demonstrated sound psychometric properties of the PESOC-PLP scale

make it a valuable contribution to the research on the development of pedagogical

leadership. In Sweden and other countries prioritizing comprehensive head teacher

training programs the PESOC-PLP scale may be used as a tool for evaluating these

programs.

Methodological discussion – strengths and weaknesses

Choice of methodology

Both qualitative and quantitative methods were chosen in order to address the research

questions in the different studies appropriately. Study I and II are based on questionnaire

data, collected from adolescents, designed to compare two measures of peer victimization

in terms of prevalence rates and associations with psychosomatic problems. In Study III,

both questionnaire data and data from focus group interviews were chosen to explore

adolescents’ understandings and perceptions of bullying. Study IV is based on

questionnaire data collected from teachers to explore teacher assessments of the

principal’s role as a pedagogical leader by analyzing the psychometric properties of a pre-

existing instrument. The result of each study has generated new questions for the

subsequent studies.

Response rate

The response rates for the student questionnaires were considered good. The data

collection in 2009 reached a response rate of 82% and involved 1760 students. In 2010,

73

the response rate was 90% and involved 2004 students. Further, the fact that eight out of

nine of the compulsory schools (school years 7-9) in the municipality participated in 2009

and all of the compulsory schools participated in 2010 is an indication of good

generalizability of the results to the study population. The response rate for the teacher

questionnaire was 57%. Even though this response rate would be considered low,

response rates around 50% is not uncommon for this type of research (Baruch &

Holtom, 2008). Despite numerous reminders it was hard to get the teachers to take time

to answer the questionnaire. Low response rates can undermine the external validity and

generalizability of the collected data because of nonresponse bias (Rogelberg & Stanton,

2007). Nonresponse could reflect the school climate and indirectly the leadership in

school. Based on information given by teachers, the major reason for not participating,

however, was lack of time. With a higher response rate a wider spectrum of attitudes

might have been captured.

Small number of victims

The prevalence rates of bullying and peer aggression is relatively low in Sweden when

compared internationally, as was also found in this thesis. Hence, the analyses are based

on a relatively small number of observations. Given larger samples, the analyses could

have been conducted at a finer level. For example, in Study II, the associations between

different measures of peer victimization and psychosomatic problems could have been

conducted using mutually exclusive groups of victims, bullies, traditional bullying, and

cyberbullying. Such tentative analyses were performed, but since they suffered from very

few observations in each unique category, it was hard to draw any definite conclusions.

Direction of causality

This thesis uses a cross-sectional design, which means that it is not possible to draw any

conclusions regarding the direction of causality. This is foremost an issue discussed in

74

Study II. For example, no conclusion can be drawn on whether peer victimization leads to

psychosomatic problems. It is also possible that students with psychosomatic problems

are more likely to report peer victimization (Fekkes et al., 2006). Youth with

psychosomatic problems may also suffer from other adversities such as childhood abuse

or chronic illness that have been shown to associate with peer victimization (Fekkes et al.,

2006; Jernbro, Svensson, Tindberg, & Janson, 2012).

Instruments and questions

The questions in the student questionnaire regarding bullying and peer aggression are

taken from Olweus Bully Victim Questionnaire (OBVQ) (Olweus, 1996). OBVQ is the

most commonly used instrument to measure bullying among children and it has been

validated in previous research (Kyriakides, Kaloyirou, & Lindsay, 2006). To measure

cyberbullying, a question originally developed by Smith, Mahdavi, Carvalho, & Tippett

(2006) and translated into Swedish by Slonje and Smith (2008) was used. When

constructing the student questionnaire, the research team decided that the definition of

bullying used in OBVQ took up too much space. Instead, the definition of bullying used

in the Swedish version of Health Behavior in School Aged Children (HBSC) was used

(Marklund, 1997). The definitions used in OBVQ and HBSC both include the criteria of

intention, repetition and power imbalance. Unfortunately, by using an older version of the

definition, the aspect of social exclusion was not included. Had this aspect been included

in the definition, it is possible that more adolescents would have been identified as victims

by a measure of bullying. Further, the peer aggression items did not include behaviors that

are specific to internet aggression. Even though a wide range of questions on different

behaviors were included, it is possible that internet aggression and other types of peer

aggression have been underreported. However, most of the questions on peer aggression

were worded in a way that could be applicable to traditional forms of bullying as well as

bullying online.

75

Categorization of victims and psychosomatic problems

Bullying

In Study I and II adolescents were categorized as victims of bullying if they reported

either traditional bullying or cyberbullying. Even though some might argue that these are

two different phenomena, the two forms of bullying are highly related. Many children

who are victimized online are also victims of traditional bullying (Wang, Iannotti, Luk, &

Nansel, 2010). The decision to include both types of bullying is further justified by

research showing that negative incidents online are also linked to real-world antisocial

behaviors (Sticca, Ruggieri, Alsaker, & Perren, 2013; Teräsahjo & Salmivalli, 2003). Rather

than being two separate phenomena it has been suggested that traditional bullying and

cyberbullying are rather two sides of the same coin (Beckman et al., 2012; Wang et al.,

2010). The commonly used cut point for bullying “2-3 times/month or more” were not

applied in the studies. Instead, all respondents indicating bullying victimization (even

occasionally) were classified as victims since repetition by definition is an integral part of

bullying. If the stricter cut point for bullying had been used, the prevalence rates for

bullying had been lower. The analyses in Study II were replicated using the stricter cut

point for bullying but did not have any effect on the results.

Psychosomatic problems

In study II, the linear PSP variable was analyzed both as a continuous and categorical

variable. The reason for using a categorical variable was that it allowed for a comparison

to be made among the adolescents who reported relatively severe psychosomatic

problems. Similar categorization of psychosomatic problems has been used in previous

studies (Carlerby, Viitasara, Knutsson, & Gadin, 2012; Gådin & Hammarström, 2005;

Hagquist, 2010). In addition, the linear variable was used in an ANOVA and OLS

regression analysis to include the full range of scores on the PSP scale. In the ANOVA,

differences in mean values on the PSP scale were tested for victims identified by a

measure of bullying, a measure of peer aggression and both measures. In the OLS

76

regression analysis, the PSP scale was included as the dependent variable and mutually

exclusive groups of victims were included as independent variables using dummy coding.

By treating the PSP variable as a continuous variable, more information from the data

could be utilized. The level of psychosomatic problems can, however, be seen as relative,

meaning that there are no clinical cut points of what is considered a mild or severe

problem. Whereas the use of percentiles leaves out more information, only adolescents

with relatively more severe problems were compared.

Self-reports

Self-reports were used as the data collection method in all the studies included in the

thesis. While self-report is a common and accepted method for measuring bullying,

individual perceptions of what it means to be bullied may vary (Nansel et al., 2001). In an

effort to reduce subjective interpretations, a definition of bullying was provided together

with the question on bullying. Despite the inclusion of a definition, the respondents may

have interpreted bullying differently. Research has shown that when answering questions

on bullying, children may use their own understanding of what bullying means instead of

relying on given definitions (Smith et al., 2002). Definition discrepancies are

acknowledged methodological ambiguities in peer victimization research and these were

addressed in Study I, II and III. Bullying may be considered a sensitive subject involving

shame and stigmatization. Social desirability means that respondents will answer questions

so as to present themselves in a favorable light and is a recognized problem in research

using self-report measures (Podsakoff & Organ, 1986). Hence, there is a risk that the

prevalence rates of bullying in the thesis are underreported.

Focus groups

The group setting offered by a focus group stimulates discussions and allows children to

respond to each other’s statements and share their views, thus creating a unique group

77

dynamic (Wyatt, Krauskopf, & Davidson, 2008). In qualitative research the findings are

evaluated in terms of trustworthiness (credibility, dependability, and transferability)

(Graneheim & Lundman, 2004). The decision to use separate groups for boys and girls in

study III is that girls tend to hold back in groups allowing boys to take the lead in

responding to questions (Shucksmith & Henry, 1998). Choosing girls and boys from

different schools and grade levels also enhanced the credibility of the data as it offered a

richer variation and understanding of the phenomena of bullying. Further, the credibility

was enhanced by involving three researchers in the analysis process to reach consensus

and by including quotations from the transcribed text, showing similarities within and

differences between categories (Graneheim & Lundman, 2004). Further, dependability

refers to the stability of data over time. The focus group interviews were conducted over

a two-week period which enhanced the consistency and dependability of the results. Even

though the moderator and research-colleague after each focus group discussed what the

adolescents had brought up, more time in between the focus group interviews could

perhaps have offered a better insight into adolescents’ perception of bullying allowing for

more follow-up questions to be developed. This study identifies adolescents’ perceptions

of bullying but did not ask the participants about their personal experiences with bullying.

Hence, we do not know whether the participants had been bullied or had bullied others,

which could have affected their perceptions of bullying.

CONCLUSIONS AND IMPLICATIONS

Even though bullying is a relatively rare phenomenon in Sweden, it is nonetheless an

important public health issue considering previous research results that indicate long

lasting effects on health and the negative effects associated with peer victimization

experienced also by students not directly involved. School has an obligation to meet the

needs of children and adolescents and to ensure that every child develops to their full

potential and that their integrity and human rights are respected. Ensuring a safe and

secure school experience is a relevant public health issue as well as an implementation of

the UN Convention on the Rights of the Child. To be able to work effectively to reduce

peer victimization in schools and to ensure a positive school experience for all children, a

78

basic premise is to make sure that we have measurement methods that accurately portray

the challenges facing today’s children and that will identify all children at risk of harm.

The main contribution of this thesis is that it provides a deeper understanding of the

problems associated with current measurement methods in peer victimization and school

leadership as well as insights into how to deal with some of the problems.

The questions shape the answers [“Som man mäter får man svar”].

This proverbial expression has guided my thesis work from start to finish. The

measurement methods chosen to research a problem have a great impact on the results

yielded and it is of uttermost importance to be aware of the questions asked and why.

This thesis shows that bullying interventions using a single measure of bullying, i.e. the

most commonly used measure to assess peer victimization in school will not identify all

children at risk of harm. Measurement methods that exclude other forms of peer

victimization than bullying have serious implications for the number of adolescents who

are identified as victims and may underestimate the full impact of peer victimization on

children. Rather than using a single measure of bullying that is based on the traditional

definition, which requires that the criteria of intention, repetition and power imbalance

are all met, other factors such as the hurt and harm experienced by children should also

be considered in the bullying definition. Peer victimization can also be seen as a

continuum taking into account the severity and circumstances of the different hurtful and

harmful behaviors. Peer victimization, then, is better measured as a continuous variable

rather than a dichotomous variable with only a “yes” or “no” answer.

This thesis also highlights the important role of adults in the anti-bullying

work in schools. Teachers and other school personnel need to be more in tune with

children’s view of bullying to be able to intervene and deal with bullying incidents in the

most effective way. Children’s reluctance to tell adults about being bullied suggests lack of

trust and could be a sign of a poor school climate. In the work to prevent peer

victimization and mental health among school children, the school principal has an

important role in ensuring a positive school climate and children’s wellbeing. The

evaluation of the PESOC-PLP scale is a step towards achieving valid assessments of

effective and pedagogical school leadership. Society's view of bullying has changed. From

79

seeing bullying as a matter of disorderly behavior, where the behavior of the bully was the

target of change, bullying is now treated as a breach of law (under the Education Act and

the Discrimination Act) and seen as an indication of a negative school climate and a

matter for school leaders to address. By strengthening the understanding of measurement

methods of peer victimization and school leadership the results from the thesis may

indirectly contribute to a safe and positive school experience for children and adolescents.

FUTURE STUDIES

The results of the current thesis suggest that more research is needed on how to best

identify all the children that are exposed to hurtful and harmful types of peer

victimization. The result of Study II needs to be further investigated. To refine the

analyses of the association between different measures of peer victimization and

psychosomatic problems, analyses based on unique categories (1) excluding bullies from

the victim group and (2) distinguishing between traditional bullying and cyberbullying

should be made using larger samples. To determine whether incorporating the effect on

the victim in the definition of bullying would lead to a better estimate of the number

children at risk due to peer victimization, the results of Study III should be further

examined using larger quantitative studies. In a Swedish context, to conclude that

pedagogical leadership should be measured based on the two sub dimensions of the

PSEOC- PLP scale or if the entire set of items also may be used needs to be further

examined based on larger data sets. Future studies could use the PESOC-PLP scale to

analyze the correlation between different school factors, such as peer victimization,

mental health and school climate. This could help extend the knowledge of the

importance of good school leadership for students’ well-being.

80

ACKNOWLEDGEMENTS

Throughout this six year journey, there have been moments of doubt, moments of pain

and many moments of sleepless nights. But above all, there have been moments of joy,

moments of laughter and moments of accomplishment. And without any of these

moments, good and bad, I would not be where I am today, accomplishing one of the

greatest achievements in my life: finishing my doctoral thesis. For this, there are many

people to whom I owe my gratitude.

First and foremost, I would like to thank my supervisor Curt Hagquist who has stood by

my side patiently and with endless words of encouragement. His high expectations and

precision have resulted in a lot of hard work (and sometimes frustration) but always

with a good result. Do not underestimate the power of a perfect abstract! He has an

incredible talent for turning defeat into victory which has taught me how to make

lemonade when life hands you lemons. This is a lesson for life. I would also like to thank

Mikael Svensson, my second supervisor, for his insightful comments and enriching

discussions!

Many thanks to all who have provided guidance and invaluable comments on my work

along this journey: Viveca Östberg, Lars Cernerud, Per-Anders Rydelius and Finn Nilson

for your critical eyes at my half-time and final seminar and especially P-A for your

comments on the early drafts of the thesis; Bodil Wilde-Larsson, for your valuable

comments on Paper III; Lennart Grosin for your discussions about PESOK (Paper IV);

Annelie Ekberg-Andersson for being the best and most helpful librarian in history. Thank

you to my current and former colleagues and friends at CFBUPH for providing a

stimulating research environment with a very wide and open door policy!

I would also like to thank Bo Thåqvist, Stefan Gräsberg (Karlstads kommun) and all

principals in Karlstad for good teamwork and cooperation in the project “Skolan

Förebygger”.

A special thanks to my colleagues, friends and partners in crime – Linda Beckman, Louise

Persson, Daniel Bergh, Stefan Persson and Victoria Lönnfjord. Their encouragement and

81

discussions about everything (livet) and nothing (getögon) have filled my journey with

endless laughter and wisdom. Linda – my research-twin and roomie, without whom this

journey would not have been the same. Our talks about science, bullying and life in

general have been invaluable. Louise – for being my friend through fail and foul and for

showing me what’s important in life. Daniel – for always believing that everything will be

OK, and for always finding time for my worries in his thick schedule. Stefan for his

endless positive attitude, statistical advice and for always bringing out laughter! Victoria –

for keeping it real at the end of my crazy journey!

I am also grateful to the rest of my colleagues and friends at Public Health Sciences and

Nursing Science at Karlstad University for encouragement, guidance and laughter!

I would also like to thank all my friends at home in Skåneland for their unconditional

friendship and endless encouragement, I love you!

Finally, to my family… my strongest supporters! In Swedish…

… den tacksamhet jag känner inför det stöd och den kärlek jag fått från er går inte nog att

beskriva i ord! Min älskade mamma som alltid har trott på mig och stöttat mig i alla mina

val i livet. Från skjuts till tidiga morgonträningar och USA-flytt till djupa filosofiska

diskussioner om livet. Pappa som har fått följa denna resa från himlen, du och mamma

har format mig till den människa jag är idag och för det är jag er evigt tacksamma. Du är

min regnbåge! Till mina systrar, Tina och Lotta med familjer, som har stöttat mig från

andra sidan Sverige och andra sidan Atlanten – ni betyder allt för mig! I am a crazy

donkey Mina svärföräldrar Rolf och Majvi – tack för ert outtröttliga stöd i otaliga

flyttar och för att ha tagit in mig i ert hem och som en del i Hellström-familjen!

Sist men inte minst… min kärlek, bäste vän och största stöttepelare i livet, Dan. Utan dig

hade denna resa aldrig nått sitt mål. Den stöttning, uppmuntran och förståelse jag har fått

från dig är obeskrivbar. Vi har klarat av mycket tillsammans och tack vare dig och din

aldrig sviktande kärlek har denna resa varit den bästa i mitt liv. Hittills. Vi har mycket

spännande kvar som väntar… Nu börjar livet!

Lisa Hellström

Lund den tredje oktober 2014

82

REFERENCES

Abt, V., & Seesholtz, M. (1994). The shameless world of Phil, Sally and Oprah: Television

talk shows and the deconstructing of society. The Journal of Popular Culture, 28(1), 171-

191.

Aftonbladet News: Tonåringar åtalas efter Instagram – upplopp i Göteborg [teens

charged after Instagram – riots in Gothenburg]. 2013, http://www.aftonbladet.se,

Accessed May 21, 2013.

Ahlström, B., & Höög, J. (2009). Measuring the social and civic objectives of schools. In

S. Huber (Ed.), School leadership - international perspectives. Bern: Peter Lang Publishing

Group.

Ahlström, B. (2009). Bullying and social objectives: A study of prerequisites for success in Swedish

schools. Umeå: Sociologiska institutionen, Umeå University.

Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology,

53(1), 27-51.

Andrich, D. (1988). Rasch models for measurement. Sage university paper on quantitative applications

in the social sciences, series 07-068. Beverly Hills: Sage Publications.

Andrich, D., Sheridan, B. S., & Luo, G. (2010). RUMM2030: Rasch unidimensional models for

measurement. Perth, WA: RUMM Laboratory.

83

Andrich, D. (1985). An elaboration of Guttman scaling with Rasch models for

measurement. In N. Brandon-Tuma (Ed.), Sociological methodology (pp. 33-80). San

Fransisco: Jossey-Bass.

Andrich, D. (2005). Rasch models for ordered response categories. In B. S. Everitt, & D.

C. Howell (Eds.), Encyclopedia of statistics in behavioral science. Chichester: John Wiley &

Sons.

Arora, C. M. J. (1996). Defining bullying. School Psychology International, 17(4), 317-329.

Barboza, G. E. (2015). The association between school exclusion, delinquency and

subtypes of cyber-and F2F-victimizations: Identifying and predicting risk profiles and

subtypes using latent class analysis. Child Abuse & Neglect, 39(2015), 109-122.

Baruch, Y., & Holtom, B. C. (2008). Survey response rate levels and trends in

organizational research. Human Relations, 61(8), 1139-1160.

Bauman, S., & Del Rio, A. (2006). Preservice teachers' responses to bullying scenarios:

Comparing physical, verbal, and relational bullying. Journal of Educational Psychology,

98(1), 219-231.

Beckman, L. (2013). Traditional bullying and cyberbullying among Swedish adolescents -

gender differences and associations with mental health. Doctoral Thesis, Karlstad

University.

Beckman, L., Hagquist, C., & Hellström, L. (2012). Does the association with

psychosomatic health problems differ between cyberbullying and traditional bullying?

Emotional and Behavioural Difficulties, 17(3-4), 421-434.

84

Beran, T. N. (2006). A construct validity study of bullying. Alberta Journal of Educational

Research, 52(4), 241-250.

Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. New York: Mcgraw-Hill.

Besag, V. E. (1989). Bullies and victims in school. Buckingham: Open Univeristy Press.

Björkqvist, K. (1994). Sex differences in physical, verbal, and indirect aggression: A review

of recent research. Sex Roles, 30(3), 177-188.

Bland, J. M., & Altman, D. G. (1995). Multiple significance tests: The Bonferroni method.

BMJ (Clinical Research Ed.), 310(6973), 170.

Bliding, M. (2002). Vad betyder mobbning? - om mobbningsbegreppet och den praktiska

vardagen. In B. Andersson, & S. Hägglund (Eds.), Barnets rättigheter (pp. 205-224).

Göteborgs universitet, Göteborg: Centrum för värdegrundsstudier.

Bond, L., Carlin, J. B., Thomas, L., Rubin, K., & Patton, G. (2001). Does bullying cause

emotional problems? A prospective study of young teenagers. British Medical Journal,

323(7311), 480.

Bosworth, K., Espelage, D. L., & Simon, T. R. (1999). Factors associated with bullying

behavior in middle school students. The Journal of Early Adolescence, 19(3), 341-362.

Boulton, M. J., Trueman, M., & Flemington, I. (2002). Associations between secondary

school pupils definitions of bullying, attitudes towards bullying, and tendencies to

engage in bullying: Age and sex differences. Educational Studies, 28(4), 353-370.

85

Bradshaw, C. P., Sawyer, A. L., & O'Brennan, L. M. (2007). Bullying and peer

victimization at school: Perceptual differences between students and school staff.

School Psychology Review, 36(3), 361-382.

Brenner, J., & Smith, A. (2013). 72% of online adults are social networking site users.

social networking sites remain most popular among young adults, but other age

groups cpntinue to increase their engagement. Retrieved from

http://www.pewinternet.org/~/media//Files/Reports/2013/PIP_Social_networkin

g_sites_update.pdf

Brookover, W. B., Beady, C., Flood, P., Schweitzer, J., & Wisenbaker, J. (1979). School

social systems and student achievement. New York: Praeger.

Brown, K., Atkins, M. S., Osborne, M. L., & Milnamow, M. (1996). A revised teacher

rating scale for reactive and proactive aggression. Journal of Abnormal Child Psychology,

24(4), 473-480.

Carlerby, H., Viitasara, E., Knutsson, A., & Gadin, K. G. (2012). How discrimination and

participation are associated with psychosomatic problems among boys and girls in

northern Sweden. Health, 4(10), 866-872.

Cassidy, W., Jackson, M., & Brown, K. N. (2009). Sticks and stones can break my bones,

but how can pixels hurt me? students' experiences with cyber-bullying. School

Psychology International, 30(4), 383-402.

86

Cassidy, W., Faucher, C., & Jackson, M. (2013). Cyberbullying among youth: A

comprehensive review of current international research and its implications and

application to policy and practice. School Psychology International, 34(6), 575-612.

Cohen, J., McCabe, L., Michelli, N. M., & Pickeral, T. (2009). School climate: Research,

policy, practice, and teacher education. The Teachers College Record, 111(1), 180-213.

Condon, C., & Clifford, M. (2010). Measuring principal performance: How rigorous are commonly

used principal performance assessment instruments? A quality school leadership issue brief.

Naperville, IL: Learning Point Associations.

Cornell, D. G., Sheras, P. L., & Cole, J. C. M. (2006). Assessment of bullying. In S. R.

Jimerson, & M. J. Furlong (Eds.), Handbook of school violence and school safety: From

research to practice (pp. 191–210). Mahwah, NJ: Erlbaum.

Creemers, B., & Kyriakides, L. (2008). The dynamics of educational effectiveness: A contribution to

policy, practice and theory in contemporary schools. London: Routledge.

Crick, N. R., & Bigbee, M. A. (1998). Relational and overt forms of peer victimization: A

multiinformant approach. Journal of Consulting & Clinical Psychology, 66(2), 337-347.

Dake, J. A., Price, J. H., Telljohann, S. K., & Funk, J. B. (2004). Principals’ perceptions

and practices of school bullying prevention activities. Health Education & Behavior,

31(3), 372-387.

Dooley, J. J., Pyżalski, J., & Cross, D. (2009). Cyberbullying versus face-to-face bullying.

Zeitschrift Für Psychologie/Journal of Psychology, 217(4), 182-188.

87

Edmonds, R. (1979). Effective schools for the urban poor. Educational Leadership, 37(1),

15-24.

Eisenberg, M. E., Neumark-Sztainer, D., Haines, J., & Wall, M. (2006). Weight-teasing

and emotional well-being in adolescents: Longitudinal findings from project EAT.

Journal of Adolescent Health, 38(6), 675-683.

Entergate (2009). esMaker Användarmanual. Enkäter och Enkäthantering – standard. Halmstad:

Entergate.

Erling, A., & Hwang, C. P. (2004). Swedish 10-year-old children's perceptions and

experiences of bullying. Journal of School Violence, 3(1), 33-43.

Eron, L. D. (1987). The development of aggressive behavior from the perspective of a

developing behaviorism. American Psychologist, 42(5), 435-442.

Espelage, D. L. (2014). Emerging issues in school bullying research and prevention. In E.

Emmer, & E. Sabornie (Eds.), Handbook of classroom management (2nd ed., pp. 76-93).

New York, NY: Routledge.

Espelage, D. L., Low, S. K., & Jimerson, S. R. (2014). Understanding school climate,

aggression, peer victimization, and bullying perpetration: Contemporary science,

practice and policy. School Psychology Quarterly, 29(3), 233-237.

Fekkes, M., Pijpers, F. I. M., Fredriks, A. M., Vogels, T., & Verloove-Vanhorick, S. P.

(2006). Do bullied children get ill, or do ill children get bullied? A prospective cohort

study on the relationship between bullying and health-related symptoms. Pediatrics,

117(5), 1568-1574.

88

Felix, E. D., Sharkey, J. D., Green, J. G., Furlong, M. J., & Tanigawa, D. (2011). Getting

precise and pragmatic about the assessment of bullying: The development of the

california bullying victimization scale. Aggressive Behavior, 37(3), 234-247.

Finkelhor, D., Turner, H. A., & Hamby, S. (2012). Let's prevent peer victimization, not

just bullying. Child Abuse & Neglect, 36(4), 271-274.

Fisher, E. J. (2013). From cyber bullying to cyber coping: The misuse of mobile

technology and social media and their effects on people’s lives. Business and Economic

Research, 3(2), 127-145.

Fox, C. L., & Boulton, M. J. (2005). The social skills problems of victims of bullying: Self,

peer and teacher perceptions. British Journal of Educational Psychology, 75(2), 313-328.

Fredrickson, B. L., & Roberts, T. (1997). Objectification theory: Toward understanding

women's lived experiences and mental health risks. Psychology of Women Quarterly, 21(2),

173-206.

Frisén, A., Holmqvist, K., & Oscarsson, D. (2008). 13-year-olds' perception of bullying:

Definitions, reasons for victimisation and experience of adults' response. Educational

Studies, 34(2), 105-117.

Furlong, M. J., Sharkey, J. D., Felix, E. D., Tanigawa, D., & Green, J. G. (2010). Bullying

assessment: A call for increased precision of self-reporting procedures. In S. R.

Jimerson, S. M. Swearer & D. L. Espelage (Eds.), The handbook of bullying in schools: An

international perspective (pp. 329-346). New York: Routledge.

Gadamer, H. (2003). Den gåtfulla hälsan. Ludvika: Dualis.

89

General Assembly. (1989). The United Nations Convention on the Rights of the Child.

Resolution 44/25 of November 1989.

Gini, G., & Pozzoli, T. (2009). Association between bullying and psychosomatic

problems: A meta-analysis. Pediatrics, 123(3), 1059-1065.

Glew, G. M., Fan, M. Y., Katon, W., Rivara, F. P., & Kernic, M. A. (2005). Bullying,

psychosocial adjustment, and academic performance in elementary school. Archives of

Pediatrics & Adolescent Medicine, 159(11), 1026-1031.

Glew, G. M., Fan, M., Katon, W., & Rivara, F. P. (2008). Bullying and school safety. The

Journal of Pediatrics, 152(1), 123-128.

Goldsmid, S., & Howie, P. (2013). Bullying by definition: An examination of definitional

components of bullying. Emotional and Behavioural Difficulties, (ahead-of-print), 1-16.

doi:10.1080/13632752.2013.844414

Goodenow, C., & Grady, K. E. (1993). The relationship of school belonging and friends'

values to academic motivation among urban adolescent students. The Journal of

Experimental Education, 62(1), 60-71.

Gradinger, P., Strohmeier, D., & Spiel, C. (2009). Traditional bullying and cyberbullying.

Zeitschrift Für Psychologie/Journal of Psychology, 217(4), 205-213.

Graneheim, U. H., & Lundman, B. (2004). Qualitative content analysis in nursing

research: Concepts, procedures and measures to achieve trustworthiness. Nurse

Education Today, 24(2), 105-112.

90

Green, J. G., Felix, E. D., Sharkey, J. D., Furlong, M. J., & Kras, J. E. (2013). Identifying

bully victims: Definitional versus behavioral approaches. Psychological Assessment.

Advance Online Publication, 25(2), 651-657.

Greif, J. L., & Furlong, M. J. (2006). The assessment of school bullying. Journal of School

Violence, 5(3), 33-50.

Griffin, R. S., & Gross, A. M. (2004). Childhood bullying: Current empirical findings and

future directions for research. Aggression and Violent Behavior, 9(4), 379-400.

Grosin, L. (2002). Rektorer i framgångsrika skolor. Nordisk Pedagogik, 22, 158175.

Grosin, L. (2003). Forskningen om framgångsrika skolor som grund för skolutveckling

[research on successful school as basis for school development]. In G. Berg, & H.

Scherp (Eds.), Skolutvecklingens många ansikten [the many faces of school development] (pp.

137-177). Stockholm: Myndigheten för Skolutveckling.

Grosin, L. (2004). Skolklimat, prestation och anpassning i 21 mellan- och 20 högstadieskolor.

(Research report No. 71). Stockholm: Pedagogiska institutionen, Stockholms

universitet.

Guerin, S., & Hennessy, E. (2002). Pupils’ definitions of bullying. European Journal of

Psychology of Education, 17(3), 249-261.

Gustafsson, J., Allodi Westling, M., Åkerman, A., Eriksson, C., Eriksson, L., Fischbein, S.,

. . . Ogden, T. (2010). School, learning and mental health: A systematic review. Stockholms

universitet, Samhällsvetenskapliga fakulteten, Specialpedagogiska institutionen:

Kungl. Vetenskapsakademien.

91

Gådin Gillander, K., & Hammarström, A. (2003). Do changes in the psychosocial school

environment influence pupils' health development? results from a three-year follow-

up study. Scandinavian Journal of Public Health, 31(3), 169-177.

Gådin, K. G., & Hammarström, A. (2005). A possible contributor to the higher degree of

girls reporting psychological symptoms compared with boys in grade nine? The

European Journal of Public Health, 15(4), 380-385.

Hagquist, C. (2001). Evaluating composite health measures using Rasch modelling: An

illustrative example. Sozial-Und Präventivmedizin/Social and Preventive Medicine, 46(6),

369-378.

Hagquist, C. (2008). Psychometric properties of the psychosomatic problems scale: A

Rasch analysis on adolescent data. Social Indicators Research, 86(3), 511-523.

Hagquist, C. (2010). Discrepant trends in mental health complaints among younger and

older adolescents in Sweden: An analysis of WHO data 1985-2005. Journal of

Adolescent Health, 46(3), 258-264.

Hagquist, C., & Andrich, D. (2004). Measuring subjective health among adolescents in

Sweden. Social Indicators Research, 68(2), 201-220.

Hagquist, C., Bruce, M., & Gustavsson, J. P. (2009). Using the Rasch model in nursing

research: An introduction and illustrative example. International Journal of Nursing

Studies, 46(3), 380-393.

Hallinger, P., & Murphy, J. (1985). Assessing the instructional leadership behavior of

principals. Elementary School Journal, 86(2), 217-248.

92

Hallinger, P., Wang, W., & Chen, C. (2013). Assessing the measurement properties of the

principal instructional management rating scale A meta-analysis of reliability studies.

Educational Administration Quarterly, 49(2), 272-309.

Hawker, D. S. J., & Boulton, M. J. (2000). Twenty years' research on peer victimization

and psychosocial maladjustment: A meta-analytic review of cross-sectional studies.

Journal of Child Psychology and Psychiatry, 41(4), 441-455.

Heinemann, P. P. (1972). Mobbning: Gruppvåld bland barn och vuxna. Stockholm: Natur och

Kultur.

Holt, M. K., Green, J. G., Reid, G., DiMeo, A., Espelage, D. L., Felix, E. D., . . . Sharkey,

J. D. (2014). Associations between past bullying experiences and psychosocial and

academic functioning among college students. Journal of American College Health, 62(8),

552-560.

Hopkins, L., Taylor, L., Bowen, E., & Wood, C. (2013). A qualitative study investigating

adolescents’ understanding of aggression, bullying and violence. Children and Youth

Services Review, 35(4), 685-693.

Hunter, S. C., Boyle, J. M. E., & Warden, D. (2007). Perceptions and correlates of peer-

victimization and bullying. British Journal of Educational Psychology, 77(4), 797-810.

Höög, J. (2011). Vad är en framgångsrik skola? In J. Höög, & O. Johansson (Eds.),

Struktur, kultur, ledarskap–förutsättningar för framgångsrika skolor (pp. 25-51). Lund:

Studentlitteratur.

93

Höög, J., & Johansson, O. (2011). Struktur, kultur, ledarskap - ett projekt och dess

resultat. In J. Höög, & O. Johansson (Eds.), Struktur, kultur, ledarskap - förutsättningar

för framgpngsrika skolor (pp. 15-24). Lund: Studentlitteratur.

Höög, J., Johansson, O., & Olofsson, A. (2005). Successful principalship: The Swedish

case. Journal of Educational Administration, 43(6), 595-606.

Höög, J., Johansson, O., & Olofsson, A. (2009). Swedish successful schools revisited.

Journal of Educational Administration, 47(6), 742-752.

Jernbro, C., Svensson, B., Tindberg, Y., & Janson, S. (2012). Multiple psychosomatic

symptoms can indicate child physical abuse–results from a study of swedish

schoolchildren. Acta Paediatrica, 101(3), 324-329.

Juvonen, J., Wang, Y., & Espinoza, G. (2011). Bullying experiences and compromised

academic performance across middle school grades. The Journal of Early Adolescence,

31(1), 152-173.

Karlstad municipality. (2007). Implementeringsplan skolan förebygger. plan för spridning av

universella, evidensbaserade program och metoder i barn- och ungdomsförvaltningen. Karlstad:

Karlstads komun.

Kelley, C., & Halverson, R. (2012). The comprehensive assessment of leadership for

learning: A next generation formative evaluation and feedback system. Journal of

Applied Research on Children: Informing Policy for Children at Risk, 3(2), Article 4.

Kelley, R. C., Thornton, B., & Daugherty, R. (2005). Relationships between measures of

leadership and school climate. Education, 126(1), 17-25.

94

Kert, A. S., Codding, R. S., Tryon, G. S., & Shiyko, M. (2010). Impact of the word “bully”

on the reported rate of bullying behavior. Psychology in the Schools, 47(2), 193-204.

Kochenderfer-Ladd, B., & Ladd, G. W. (2001). Variations in peer victimization: Relations

to children's maladjustment. In J. Juvonen, & S. Graham (Eds.), Peer harassment in

school: The plight of the vulnerable and victimized (pp. 25-48). New York: Guilford.

Kochenderfer-Ladd, B., & Pelletier, M. E. (2008). Teachers' views and beliefs about

bullying: Influences on classroom management strategies and students' coping with

peer victimization. Journal of School Psychology, 46(4), 431-453.

Kochenderfer, B., & Ladd, G. W. (1996). Peer victimization: Manifestations and relations

to school adjustment in kindergarten. Journal of School Psychology, 34(3), 267-283.

Krug, E. G., Dahlberg, L. L., Mercy, J. A., Zwi, A. B., & Lozano, R. (2002). World report on

violence and health. Geneva: World Health Organization.

Kyriakides, L., Kaloyirou, C., & Lindsay, G. (2006). An analysis of the revised Olweus

Bully/Victim Questionnaire using the Rasch measurement model. British Journal of

Educational Psychology, 76(4), 781-801.

Kyriakides, L., Creemers, B., Antoniou, P., & Demetriou, D. (2010). A synthesis of

studies searching for school factors: Implications for theory and research. British

Educational Research Journal, 36(5), 807-830.

Lacey, A., & Cornell, D. (2013). The impact of teasing and bullying on schoolwide

academic performance. Journal of Applied School Psychology, 29(3), 262-283.

95

Larson, J. S. (1999). The conceptualization of health. Medical Care Research and Review,

56(2), 123-136.

Larsson, A. (2008). Mobbningsbegreppets uppkomst och förhistoria: En

begreppshistorisk analys. Pedagogisk forskning i Sverige 13(1), 19-36.

LaRusso, M. D., Romer, D., & Selman, R. L. (2008). Teachers as builders of respectful

school climates: Implications for adolescent drug use norms and depressive

symptoms in high school. Journal of Youth and Adolescence, 37(4), 386-398.

Lazarus, R., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer

Publishing Company.

Leff, S. S., & Waasdorp, T. E. (2013). Effect of aggression and bullying on children and

adolescents: Implications for prevention and intervention. Current Psychiatry Reports,

15(3), 1-10.

Leithwood, K. (2001). School leadership in the context of accountability policies.

International Journal of Leadership in Education, 4(3), 217-235.

Lenhart, A., Purcell, K., Smith, A., & Zickuhr, K. (2010). Social media & mobile internet use

among teens and young adults. Washington, DC: Pew Internet & American Life Project.

Libbey, H. P. (2004). Measuring student relationships to school: Attachment, bonding,

connectedness, and engagement. Journal of School Health, 74(7), 274-283.

96

Lien, L., & Welander-Vatn, A. (2013). Factors associated with the persistence of bullying

victimization from 10th grade to 13th grade: A longitudinal study. Clinical Practice and

Epidemiology in Mental Health. 2013; 9:243-250.

Liu, M. C., Lan, C., Hsu, J., Huang, K., & Chen, Y. (2014). Bullying victimization and

conduct problems among high school students in Taiwan: Focus on fluid intelligence,

mood symptoms and associated psychosocial adjustment. Children and Youth Services

Review, 47(3), 231-238.

Livingstone, S., Haddon, L., & Ólafsson, K. (2011). EU kids online: Final report.

Lorenz, K. (1968). Aggression: Dess bakgrund och natur. Stockholm: Norstedt & Söner.

Ludvigsson, A. (2009). Samproducerat ledarskap - hur rektorer och lärare formar ledarskap i skolans

vardagsarbete [co-produced leadership - how principals and teachers shape leadership in the school's

everyday work]. (Doctoral thesis, School of Education and Communication, Jönköping

University, Sweden).

Madden, M., Lenhart, A., Cortesi, S., Gasser, U., Duggan, M., Smith, A. & Beaton, M.

(2013). Teens, social media, and privacy, May 21, Pew Research Center & Berkman Center

for Internet & Society [online] http://www.pewinternet.org/~/media//

Files/Reports/2013/PIP_TeensSocialMediaandPrivacy.pdf (accessed June 2013).

Madsen, K. (1996). Differing perceptions of bullying and their practical implications.

Educational and Child Psychology, 13(2), 14-22.

Marais, I., & Andrich, D. (2008). Formalizing dimension and response violations of local

independence in the unidimensional Rasch model. J Appl Meas, 9(3), 200-215.

97

Marklund, U. (1997). Skolbarns hälsovanor under ett decennium. Tabell-Rapport.Health

Behavior in School-Age Children. A WHO Collaborative Study. Folkhälsoinstitutet.

Marzano, R. J., Waters, T., & McNulty, B. A. (2005). School leadership that works: From

research to results. Alexandria, VA: Association for Supervision & Curriculum

Development (ASCD).

Maunder, R. E., Harrop, A., & Tattersall, A. J. (2010). Pupil and staff perceptions of

bullying in secondary schools: Comparing behavioural definitions and their perceived

seriousness. Educational Research, 52(3), 263-282.

McEvoy, A., & Welker, R. (2000). Antisocial behavior, academic failure, and school

climate A critical review. Journal of Emotional and Behavioral Disorders, 8(3), 130-140.

Mehta, S. B., Cornell, D., Fan, X., & Gregory, A. (2013). Bullying climate and school

engagement in ninth‐grade students. Journal of School Health, 83(1), 45-52.

Menesini, E., & Nocentini, A. (2009). Cyberbullying definition and measurement.

Zeitschrift Für Psychologie/Journal of Psychology, 217(4), 230-232.

Modin, B. (2012). Beteendeproblem bidrar til utsatthet för mobbning och psykisk ohälsa. I: Skolans

betydelse för barns och ungas psykiska hälsa - en studie baserad på den nationella

totalundersökningen i årskurs 6 och 9 hösten 2009 [behavior problems contribute to victimization

from bulying and poor self-related mental health. in: The role of the school for young people's mental

health. A study based on the national swedish survey of students in grades 6 and 9 in 2009]. ( No.

2012-5-15). Stockholm, Sweden: National Board of Health and Welfare and Centre

for Health Equity Studies.

98

Modin, B., & Östberg, V. (2009). School climate and psychosomatic health: A multilevel

analysis. School Effectiveness and School Improvement, 20(4), 433-455.

Monks, C. P., & Smith, P. K. (2006). Definitions of bullying: Age differences in

understanding of the term, and the role of experience. British Journal of Developmental

Psychology, 24(4), 801-821.

Monks, C. P., Smith, P. K., Naylor, P., Barter, C., Ireland, J. L., & Coyne, I. (2009).

Bullying in different contexts: Commonalities, differences and the role of theory.

Aggression and Violent Behavior, 14(2), 146-156.

Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simons-Morton, B., & Scheidt, P.

(2001). Bullying behaviors among US youth: Prevalence and association with

psychosocial adjustment. JAMA: The Journal of the American Medical Association,

285(16), 2094-2100.

Naylor, P., Cowie, H., Cossin, F., de Bettencourt, R., & Lemme, F. (2006). Teachers' and

pupils' definitions of bullying. British Journal of Educational Psychology, 76(3), 553-576.

Nestor, B. (1993). Pedagogiskt ledarskap och pedagogik ledning: Två svårfångade

begrepp. In I. B. Stålhammar (Ed.), Skolledare i en föränderlig värld. Göteborg: Gothia.

Notar, C. E., Padgett, S., & Roden, J. (2013). Cyberbullying: A review of the literature.

Universal Journal of Educational Research, 1(1), 1-9.

Olweus, D. (1978). Aggression in the schools: Bullies and whipping boys. Washington, NY:

Halsted Press.

99

Olweus, D. (1991). Bully/victim problems among schoolchildren: Basic facts and effects

of a school based intervention program. In D. Pepler, & K. Rubins (Eds.), The

development and treatment of childhood aggression (pp. 411-448). Hillsdale, NJ: Erlbaum.

Olweus, D. (1993). Bullying at school: What we know and what we can do. Malden, MA:

Blackwell Publishers.

Olweus, D. (1996). The revised Olweus Bully/Victim Questionnaire. Research Center for Health

Promotion (HEMIL): University of Bergen.

Olweus, D. (2001). Peer harassment: A critical analysis and some important issues. In J.

Juvonen, & S. Graham (Eds.), Peer harassment in school the plight of the vulnerable and

victimized (pp. 3-20). New York: The Guilford Press.

Olweus, D., & Breivik, K. (2014). Plight of victims of school bullying: The opposite of

well-being. In A. Ben-Arieh, F. Casas, I. Frones & J. E. Korbin (Eds.), Handbook of

child well-being: Theories, methods and policies in global perspective (pp. 2593-2616) Dordrecht:

Springer Netherlands.

Ostrov, J. M., Murray-Close, D., Godleski, S. A., & Hart, E. J. (2013). Prospective

associations between forms and functions of aggression and social and affective

processes during early childhood. Journal of Experimental Child Psychology,

doi:http://dx.doi.org/10.1016/j.jecp.2012.12.009

Pakaslahti, L., & Keltikangas-Jarvinen, L. (2000). Comparison of peer, teacher and self-

assessments on adolescent direct and indirect aggression. Educational Psychology, 20(2),

177-190.

100

Perneger, T. V. (1998). What's wrong with Bonferroni adjustments. BMJ (Clinical Research

Ed.), 316(7139), 1236-1238.

Perry, D. G., Kusel, S. J., & Perry, L. C. (1988). Victims of peer aggression. Developmental

Psychology, 24(6), 807-814.

Petersen, S., Brulin, C., & Bergström, E. (2006). Recurrent pain symptoms in young

schoolchildren are often multiple. Pain, 121(1), 145-150.

Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research:

Problems and prospects. Journal of Management, 12(4), 531-544.

Pozzoli, T., & Gini, G. (2010). Active defending and passive bystanding behavior in

bullying: The role of personal characteristics and perceived peer pressure. Journal of

Abnormal Child Psychology, 38(6), 815-827.

Proposition. (2007/08:110). En förnyad folkhälsopolitik. no. 2007/08:110. Stockholm:

Regeringskansliet.

Purkey, S. C., & Smith, M. S. (1983). Effective schools: A review. The Elementary School

Journal, 83(4), 427-452.

Rabiner, D. L., Keane, S. P., & MacKinnonLewis, C. (1993). Children's beliefs about

familiar and unfamiliar peers in relation to their sociometric status. Developmental

Psychology, 29(2), 236-243.

Randall, P. E. (1991). The prevention of school based bullying. Hull: University of Hull.

101

Rapp, S. (2009). Rektor och lagen: En studie av skolans pedagogiska ledare. Stockholm: Norstedts

juridik.

Rasch, G. (1960/1980). Probabilistic models for some intelligence and attainment tests (first published

1960 by the Danish institute for educational research). Chicago: MESA Press.

Reynolds, D., Sammons, P., De Fraine, B., Van Damme, J., Townsend, T., Teddlie, C., &

Stringfield, S. (2014). Educational effectiveness research (EER): A state-of-the-art

review. School Effectiveness and School Improvement, 25(2), 197-230.

Reynolds, H. T. (1984). Analysis of nominal data. Newbury Park, CA: Sage.

Reynolds, W. (2003). Reynolds bully victimization scales for schools manual. San Antonio,

TX: The Psychological Corporation.

Richard, J. F., Schneider, B. H., & Mallet, P. (2012). Revisiting the whole-school approach

to bullying: Really looking at the whole school. School Psychology International, 33(3),

263-284.

Rigby, K. (2002). New perspectives on bullying. London: Jessica Kingsley Publishers.

Rigby, K. (2006a). Towards a definition of bullying. In K. Rigby (Ed.), New perspectives on

bullying (pp. 27-51). London: Jessica Kingsley Publishers.

Rigby, K. (2006b). What international research tells us about bullying. In H. McGrath, &

T. Noble (Eds.), Bullying solutions: Evidence-based approaches to bullying in australian schools

(pp. 3-15). Crowst Nest, NSW: Pearson Education.

102

Rigby, K. (2007). Bullying in schools: And what to do about it. Australian Council for

Educational Research, Australia: ACER press.

Robinson, V. (2007). The impact of leadership on student outcomes: Making sense of the

evidence. Paper Presented at the International Conference New Imagery for School and Schooling:

Challenging, Creating, and Connecting, Sydney, Australia.

Rogelberg, S. G., & Stanton, J. M. (2007). Introduction understanding and dealing with

organizational survey nonresponse. Organizational Research Methods, 10(2), 195-209.

Rutter, M., Maugham, B., Mortimore, P., Ousten, J., & Smith, A. (1979). Fifteen thousand

hours. secondary schools and their effects on children. London: Open Books.

Sammons, P., Thomas, S. M., & Mortimore, P. (1997). Forging links: Effective schools and

effective departments. London: Paul Chapman.

Sawyer, A. L., Bradshaw, C. P., & O'Brennan, L. M. (2008). Examining ethnic, gender,

and developmental differences in the way children report being a victim of bullying

on self-report measures. Journal of Adolescent Health, 43(2), 106-114.

SBU. (2010). Program för att förebygga psykisk ohälsa hos barn. Stockholm: Statens beredning

för medicinsk utvärdering.

Scheerens, J. (2013). The use of theory in school effectiveness research revisited. School

Effectiveness and School Improvement, 24(1), 1-38.

103

Scholte, R. H., Burk, W. J., & Overbeek, G. (2013). Divergence in self-and peer-reported

victimization and its association to concurrent and prospective adjustment. Journal of

Youth and Adolescence, (2013) 42:1789-1800.

Seashore Louis, K., Dretzke, B., & Wahlstrom, K. (2010). How does leadership affect

student achievement? results from a national US survey. School Effectiveness and School

Improvement, 21(3), 315-336.

SFS. (2008:567). Discrimination Act. Stockholm: Regeringen.

SFS. (2009:1521). Act to amend the Education Act. Stockholm: Regeringen.

SFS. (2010:800). Education act. Stockholm: Regeringen.

Shucksmith, J., & Henry, L. B. (1998). Health issues and adolescents: Growing up, speaking out.

London: Routledge.

Skolinspektionen. (2010). Skolors arbete vid trakasserier och kränkande behandling. ( No.

2010:1). Stockholm: Skolinspektionen.

Slonje, R., & Smith, P. K. (2008). Cyberbullying: Another main type of bullying?

Scandinavian Journal of Psychology, 49(2), 147-154.

Slonje, R., Smith, P. K., & Frisén, A. (2012). Processes of cyberbullying, and feelings of

remorse by bullies: A pilot study. European Journal of Developmental Psychology, 9(2), 244-

259.

104

Smith, P. K., Mahdavi, J., Carvalho, M., & Tippett, N. (2006). An investigation into

cyberbullying, its forms, awareness and impact, and the relationship between age and gender in

cyberbullying. (No. Research Brief No. RBX03-06). London: DfES.

Smith, P. K., Morita, Y., Junger-Tas, J., Olweus, D., Catalano, R., Slee, P., . . . editors.

(1999). The nature of school bullying: A cross-national perspective. London and New York:

Routledge.

Smith, P. K., & Sharp, S. (1994). School bullying: Insights and perspectives. London: Routledge.

Smith, P. K., Cowie, H., Olafsson, R. F., & Liefooghe, A. P. D. (2002). Definitions of

bullying: A comparison of terms used, and age and gender differences, in a Fourteen–

Country international comparison. Child Development, 73(4), 1119-1133.

SNAE. (2009a). Diskriminerad, trakasserad, kränkt? Barns, elevers och studerandes uppfattningar

om diskriminering och trakasserier. (No. 326). Stockholm: The Swedish National Agency

for Education [Skolverket].

SNAE. (2009b). På tal om mobbning – och det som görs. Stockholm: The Swedish National

Agency of Education [Skolverket].

Solberg, M. E., & Olweus, D. (2003). Prevalence estimation of school bullying with the

Olweus Bully/Victim Questionnaire. Aggressive Behavior, 29(3), 239-268.

SOU (2006). Ungdomar, stress och psykisk ohälsa: Analyser och förslag till åtgärder:

Slutbetänkande. Utredningen om ungdomars psykiska hälsa. Swedish Government

Official Report 2006: Report No. 77.

105

Stapinski, L. A., Araya, R., Heron, J., Montgomery, A. A., & Stallard, P. (2014). Peer

victimization during adolescence; concurrent and prospective impact on symptoms

of depression and anxiety. Anxiety, Stress, & Coping, 28(1): 105-120.

Stein, M. K., & Nelson, B. S. (2003). Leadership content knowledge. Educational Evaluation

and Policy Analysis, 25(4), 423-448.

Stevens, V., De Bourdeaudhuij, I., & Van Oost, P. (2000). Bullying in flemish schools: An

evaluation of anti-bullying intervention in primary and secondary schools. The British

Journal of Educational Psychology, 70(2), 195-210.

Stewart, D. (2008). Implementing mental health promotion in schools: A process

evaluation. International Journal of Mental Health Promotion, 10(1), 32-41.

Sticca, F., Ruggieri, S., Alsaker, F., & Perren, S. (2013). Longitudinal risk factors for

cyberbullying in adolescence. Journal of Community & Applied Social Psychology, 23(1), 52-

67.

Stockdale, M. S., Hangaduambo, S., Duys, D., Larson, K., & Sarvela, P. D. (2002). Rural

elementary students', parents', and teachers' perceptions of bullying. American Journal

of Health Behaviour, 26(4), 266-277.

Stronge, J. H., Richard, H. B., & Catano, N. (2008). Qualities of effective principals.

Alexandria, VA: Association for Supervision and Curriculum Development (ASCD).

Swain, J. (1998). What does bullying really mean? Educational Research, 40(3), 358-364.

106

Swearer, S. M., Espelage, D. L., Vaillancourt, T., & Hymel, S. (2010). What can be done

about school bullying? Educational Researcher, 39(1), 38-47.

Tattum, D. P., & Tattum, E. (1992). Bullying: A whole-school response. In N. Jones, &

E. Baglin Jones (Eds.), Learning to behave. London: Kogan Page.

Tejeda, M. J., Scandura, T. A., & Pillai, R. (2001). The MLQ revisited: Psychometric

properties and recommendations. The Leadership Quarterly, 12(1), 31-52.

Tennant, A., & Conaghan, P. G. (2007). The Rasch measurement model in rheumatology:

What is it and why use it? When should it be applied, and what should one look for

in a Rasch paper? Arthritis Care & Research, 57(8), 1358-1362.

Teräsahjo, T., & Salmivalli, C. (2003). “She is not actually bullied.” the discourse of

harassment in student groups. Aggressive Behavior, 29(2), 134-154.

Thapa, A., Cohen, J., Guffey, S., & Higgins-D’Alessandro, A. (2013). A review of school

climate research. Educational Reserch, 55(4), 411-426.

The Swedish School Inspectorate. (2012). Prinicipal leadership: With responsibility for pedagogical

activities. ( No. 2012:1). Stockholm: The Swedish School Inspectorate.

Thompson, J. K., Coovert, M. D., Richards, K. J., Johnson, S., & Cattarin, J. (1995).

Development of body image, eating disturbance, and general psychological

functioning in female adolescents: Covariance structure modeling and longitudinal

investigations. International Journal of Eating Disorders, 18(3), 221-236.

107

Thornberg, R. (2010). A study of children's conceptions of school rules by investigating

their judgments of transgressions in the absence of rules. Educational Psychology, 30(5),

583-603.

Thornberg, R., & Knutsen, S. (2011). Teenagers’ explanations of bullying. Child and Youth

Care Forum, 40(3), 177-192.

Thornberg, R., Elvstrand, H. (2012). Children's experiences of democracy, participation,

and trust in school. International Journal od Educational Research, 53(2012), 44-54.

Thornberg, R., Birberg Thornberg, U., Alamaa,R., & Daud, N. (2014) Children's

conceptions of bullying and repeated conventional transgressions: moral,

conventional, structuring and personal-choice reasoning, Educational Psychology: An

International Journal of Experimental Educational Psychology, DOI:

10.1080/01443410.2014.915929.

Thornberg, R. (2015). Distressed bullies, social positioning and odd victims: Young

people's explanations of bullying. Children & Society 29(1), 15-25.

Tokunaga, R. S. (2010). Following you home from school: A critical review and synthesis

of research on cyberbullying victimization. Computers in Human Behavior, 26(3), 277-

287.

Törnsén, M. (2009). Successful principal leadership: Prerequisites, processes and outcomes. Doctoral

thesis, Faculty of social sciences, Umeå Univeristy, Sweden.

UNICEF. (2011). Child safety online: Global challenges and strategies. Florence, Italy, UNICEF

Innocenti Research Centre.

108

Vaillancourt, T., McDougall, P., Hymel, S., Krygsman, A., Miller, J., Stiver, K., & Davis,

C. (2008). Bullying: Are researchers and children/youth talking about the same thing?

International Journal of Behavioral Development, 32(6), 486-495.

van der Wal, M. F., de Wit, C. A. M., & Hirasing, R. A. (2003). Psychosocial health among

young victims and offenders of direct and indirect bullying. Pediatrics, 111(6), 1312-

1317.

Vandebosch, H., & Van Cleemput, K. (2008). Defining cyberbullying: A qualitative

research into the perceptions of youngsters. CyberPsychology & Behavior, 11(4), 499-503.

Vandenbosch, L., & Eggermont, S. (2012). Understanding sexual objectification: A

comprehensive approach toward media exposure and girls' internalization of beauty

ideals, self‐objectification, and body surveillance. Journal of Communication, 62(5), 869-

887.

Varjas, K., Meyers, J., Bellmoff, L., Lopp, E., Birckbichler, L., & Marshall, M. (2008).

Missing voices: Fourth through eighth grade urban students' perceptions of bullying.

Journal of School Violence, 7(4), 97-118.

Vitak, J., & McLaughlin, C. (2011, May). A picture's worth a thousand words, but

friendships matter more: Applying warranting theory to facebook. Paper Presented at

the 61st Annual Conference of the International Communication Association, Boston, MA.

Värmlands Folkblad News: Barnen tävlar om vem som är snyggast [the children compete

to see who is the hottest]. 2013, http://www.vf.se, Accessed July 1 2013.

109

Wadsworth, M., & Kuh, D. (1997). Childhood influences on adult health: A review of

recent work from the British 1946 national birth cohort study, the MRC national

survey of health and development. Paediatric and Perinatal Epidemiology, 11(1), 2-20.

Wang, J., Iannotti, R. J., Luk, J. W., & Nansel, T. R. (2010). Co-occurrence of

victimization from five subtypes of bullying: Physical, verbal, social exclusion,

spreading rumors, and cyber. Journal of Pediatric Psychology, 35(10), 1103-1112.

Weber, G. (1971). Inner-city children can be taught to read: Four successful schools. Washington

D.C.: Council for Basic Education.

WHO. (1948). Preamble to the constitution of the world health organization as adopted by the

international health conference, new york, 19-22 june, 1946; signed on 22 july 1946 by the

representatives of 61 states (official records of the world health organization, no. 2, p. 100) and

entered into force on 7 april 1948.

Williams, L. C. A., & Stelko-Pereira, A. C. (2013). Let's prevent school violence, not just

bullying and peer victimization: A commentary on Finkelhor, Turner, and Hamby.

Child Abuse & Neglect, 37(4), 235-236.

Woods, S., Done, J., & Kalsi, H. (2009). Peer victimisation and internalising difficulties:

The moderating role of friendship quality. Journal of Adolescence, 32(2), 293-308.

Wyatt, T. H., Krauskopf, P. B., & Davidson, R. (2008). Using focus groups for program

planning and evaluation. The Journal of School Nursing, 24(2), 71-82.

Yoon, J. S. (2004). Predicting teacher interventions in bullying situations. Education and

Treatment of Children, 27(1), 37-45.

110

Yukl, G. A. (2012). Leadership in organizations. Upper Saddle River, NJ: Prentice Hall.

Ärlestig, H. (2011). Vardagssamtal och rektors pedagogiska ledarskap. In J. Höög, & O.

Johansson (Eds.), Struktur, kultur, ledarskap–förutsättningar för framgångsrika skolor (pp.

115-133). Lund: Studentlitteratur.

Measuring peer victimization and school leadershipA study of definitions, measurement methods and associations with psychosomatic health

Lisa Hellström

Lisa Hellström

| Measuring peer victim

ization and school leadership | 2015:15

Measuring peer victimization and school leadership

The negative consequences of peer victimization on children and adolescents such as worsening academic achievement and mental ill health are major public health concerns which have been subjected to extensive research. However, there are long-standing concerns how to define, measure, and estimate prevalence rates of peer victimization and successful school leadership. The aim of this thesis is to study methods for assessing peer victimization and pedagogical leadership in school. The results show that excluding other forms of peer victimization than bullying have serious implications for the identification of victims and may underestimate the full impact of peer victimization on children. Further, the validation of the Pedagogical and Social Climate (PESOC-PLP) scale is a step towards ensuring valid assessments of pedagogical school leadership. By strengthening the understanding of measurement methods of peer victimization and school leadership the results from the thesis may indirectly contribute to a safe and positive school experience for children and adolescents.

DISSERTATION | Karlstad University Studies | 2015:15 DISSERTATION | Karlstad University Studies | 2015:15

ISSN 1403-8099

Faculty of Health, Science and TechnologyISBN 978-91-7063-626-4

Public Health Science