University of Southern California Dissertations and Theses

School social dynamics as mediators of students personal traits and family factors on the perpetration of school violence in Taiwan

(USC Thesis Other)

School social dynamics as mediators of students personal traits and family factors on the perpetration of school violence in Taiwan

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content SCHOOL SOCIAL DYNAMICS AS MEDIATORS OF STUDENTS’ PERSONAL
TRAITS AND FAMILY FACTORS ON THE PERPETRATION OF SCHOOL
VIOLENCE IN TAIWAN

by

Ji-Kang Chen

________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(SOCIAL WORK)

August 2008

Copyright 2008 Ji-Kang Chen

ii
Dedication
To my family, especially my paternal grandparents (Chen Tan and Huang Mei-Cheng),
maternal grandparents (Deng Tian-Yang and Deng Lee Jin-Lian), parents (Chen
Huan-Hui and Deng Lee-Ho), and brother (Chen Chi-Chian). Without you, this would not
be possible.

iii
Acknowledgements
My doctoral career has been one of the most challenging, enriching, and joyful
experiences of my life. I would like to thank the members of my dissertation committee
for their thoughtful and valuable guidance and continuous support: Dr. Ron Astor, Dr. Iris
Chi, and Dr. C. Anderson Johnson. I am grateful to each and every one of my committee
members for believing in me and for their mentorship and generosity. I am honored to
have had the opportunity to learn from such remarkable scholars. Special thanks to Dr.
Ron Astor, my advisor, for his valued mentorship and for allowing me to learn by
exposure to and observation of his distinguished research and writing abilities. Dr. Astor
has always been an excellent and trustworthy mentor in helping me achieve my research
and professional goals. His diligence, insight into academic research, and personality will
always be an exemplary role model in my life. I would also like to especially thank Dr.
Robert Nishimoto and Dr. Penelope K. Trickett for their continued support and interest in
my professional development.
My friends have provided me with a sense of stability and sanity throughout this
process. I would like to thank my best friends, Tseng-Ping Liu and Jennifer Du, for their
unconditional support and endless encouragement. I truly appreciate Yi-Ling Yeh’s for
sharing the joys and sorrows of a Ph.D. student life. A special thanks to Tai-Cheng Wu for
accompanying me and her love during these challenging years.
Most important, from my heart, I would like to thank my parents and brother for
their love and encouragement. They have been inspired me to always strive for
excellence and furnished me the curiosity of learning. The thoughts, support, love, ideas,

iv
and company of family and friends were the key ingredients in allowing me to pursue and
be successful in obtaining my Ph.D degree.
Finally, I would like to recognize the generous funds from Chiang Ching-Kuo
Foundation for International Scholarly Exchange (USA) for this research. I wish to note
that data analyzed in this dissertation were collected by the research project “Prevention
and Control of School Violence in Taiwan” sponsored by the National Science Council:
NSC 89-2420-H-006-001-QBS. This research project was carried out by National Cheng
Kung University, and directed by Dr. Jin Wu. The Center for Survey Research of
Academia Sinica is responsible for the data distribution. The author appreciates the
assistance in providing data by the institutes and individuals aforementioned. The views
expressed herein are the authors’ own.

v
Table of Contents

DEDICATION ii

ACKNOWLEDGEMENTS

iii
LISTS OF TABLES

vii
LISTS OF FIGURES

viii
ABSTRACT

ix
CHAPTER ONE: INTRODUCTION, RATIONALE AND OVERVIEW OF
FIVE STUDIES
Structure of the Dissertation
Overview of Taiwanese Education System

1
3
10
CHAPTER TWO (STUDY 1): AN EPIDEMIOLOGICAL PRESENTATION
AND ANALYSES OF THE PERPETRTAION OF STUDENT RELATED
VIOLENCE IN TAIWAN
Introduction
Methods
Results
Discussion
Implications

16
16
22
24
27
31
CHAPTER THREE (STUDY 2): THE PERPETRATION OF STUDENT
VIOLENCE AGAINST TEACHERS IN TAIW AN
Introduction
Methods
Results
Discussion
Implications

37
37
42
45
47
52
CHAPTER FOUR (STUDY 3): SCHOOL V ARIABLES AS MEDIATORS OF
PERSONAL AND FAMILY FACTORS ON SCHOOL VIOLENCE IN
ELEMENTARY SCHOOLS
Introduction
Methods
Results
Discussion
Implications

57
57
64
69
71
74

vi
CHAPTER FIVE (STUDY 4): SCHOOL V ARIABLES AS MEDIATORS OF
PERSONAL AND FAMILY FACTORS ON SCHOOL VIOLENCE IN JUNIOR
HIGH SCHOOLS
Introduction
Methods
Results
Discussion

84
84
90
95
97
CHAPTER SIX (STUDY 5): SCHOOL V ARIABLES AS MEDIATORS OF
PERSONAL AND FAMILY FACTORS ON SCHOOL VIOLENCE IN HIGH
SCHOOLS
Introduction
Methods
Results
Discussion

109
109
116
122
125
CHAPTER SEVEN: INTERGRATION AND IMPLICATIONS FROM
FINDINGS ACROSS ALL FIVE STUDIES

139
REFERENCES

153
APPENDIX A: STUDENT QUESTIONNAIRE 161

vii

List of Tables

Table 1. Distribution of Students Participating in the Study 33

Table 2. Percentage of Perpetration in Violent Behavior by Gender and School
Type

34
Table 3. Percentage of Perpetrators’ Reasons for Violence by Gender and
School Type

35
Table 4. Percentage of Students’ Violence Against Teachers by Genders and
School Type

54
Table 5. The Percentage of Students’ Reasons for Violence against Teachers

55
Table 6. Means and Standard Deviations of Subscales by Gender

76
Table 7. Matrix of Correlation among School Violence Subscales for
Elementary School

77
Table 8. Theoretical Domains, Subscale, and Items for Elementary School

78
Table 9. Means and Standard Deviations of Subscales by Gender

101
Table 10. Matrix of Correlations among School Violence Subscales for Junior
High School

102
Table 11. Theoretical Domains, Subscale, and Items for Junior High School

103
Table 12. Means and Standard Deviations of Subscales by Genders and School
Types

130
Table 13. Matrix of Correlations among School Violence Subscales for High
School

131
Table 14. Theoretical Domains, Subscales, and Items for High School

132

viii
List of Figures

Figure 1. Theoretical Model for Perpetration of School Violence

15
Figure 2. Distribution of Percentage of Perpetration among Taiwanese Students,
by Gender, Grade, and School Type

36
Figure 3. Percentage of Perpetration in at Least One Act of Violence Against
Teachers by Gender, Grade Level, and School Type

56
Figure 4. Theoretical Model Results Based on Total Sample of Elementary
School Students

82
Figure 5. Gender Comparison for Elementary School Students

83
Figure 6. Total Sample of Junior High School Students’ Violence Against
Students and Teachers

107
Figure 7. Gender Comparison in Junior High School

108
Figure 8. Total Sample of High School Students’ Violence Against Students and
Teacher

136
Figure 9. School Type Comparison for High School Students

137
Figure 10. Gender Comparison in High School

138

ix
Abstract
To date, very few national studies have been conducted in Asia on school violence issues.
In addition, few studies explored how school dynamics, family factors, personal traits
work together to contribute to perpetration of school violence. Using a nationally
representative sample of 14,042 Taiwanese students from elementary schools (grades 4 to
6), junior high schools (grades 7 to 9), academic high schools and vocational high schools
(grades 10 to 12), this dissertation describes the perpetration of school violence in Taiwan.
These are the first nationally representative studies on these issues in Taiwan. The first
study of this dissertation focuses on the prevalence of student violence against students.
The second study focuses on student violence against teachers. The remaindering three
studies of the dissertation examine a theoretical model of how school engagement, school
risky peers, and student-teacher relationships mediate the effects of personal traits and
family factors on school violence committed by students against other students and
teachers. Structural equation modeling was used to examine the theoretical model. Three
separate studies explored this model on elementary (chapter 3), junior high (chapter 4),
and high school (chapter 5) samples respectively. The study on high school students
(chapter 5) also explores the model in vocational and academic schools. Compared with
Western countries, the prevalence of school violence in Taiwan is high. The findings
across all three structural equation model studies suggest that the theoretical model was a
good fit for the elementary school, junior high school and high school samples. Moreover,
the theoretical models developed in Western cultures explained larger amounts of the
explained variance for violence against student and teachers in Asian cultures. The
x
overall findings suggested that school factors such as school engagement, school risky
peers, and student-teacher relationships mediate family and personal factors. However,
each school variable plays a different role in mediating the relationship for each school
type and across development. Implications for theory, policy and practice, and
recommendation for future research are discussed.

1
Chapter One
Introduction, Rationale and Overview of Five Studies
Over the past three decades, a majority of studies on school violence have been
mainly conducted in North American or European countries (Nansel et al., 2001;
Pellegrini & Long, 2002; Pepler et al., 2006). By contrast, large-scale representative data
in Asia remain limited. There is little evidence either supporting or discounting the
proposition that theoretical assumptions from school violence studies in Western
countries are relevant to Asian cultures. This is problematic, since influential social
psychologists have cited theories and nonrepresentative evidence suggesting significant
differences exist between East and West in beliefs and perceptions (Nisbett, 2003). Many
popular views of East and West also assume great social and structural differences
between Eastern and Western societies. In contrast to these social psychological theories
predicting differences, some findings suggest that structural aspects of school violence
may be cross cultural, because basic school structures are more common than different
across most developed countries (Benbenishty & Astor, 2005). Nevertheless, to date,
school violence studies have not explored theoretical similarities or differences between
Western and Asian schools using empirical methods. This dissertation aims to expand the
literature on school violence by exploring the applicability of Western generated
theoretical relationships and risk factors in Asian schools.
Theoretically, this dissertation explores the patterns and influence of risk factors that
should have similar meanings in both Western and Eastern cultures. For example, most
school violence studies have explored how school violence associated with context
2
factors such as gender, age, and school type (elementary, junior high school and high
school) separately (Nansel et al., 2001; Pellegrini & Long, 2002; Pepler et al., 2006).
There is limited knowledge of how the interplay of gender, age and school type affects
prevalence of student perpetration of school violence (for and exception see Benbenisthy
and Astor, 2005). Many studies have examined how students’ personal, family, and
school experience directly affect school violence outcomes (e.g., Ando, Asakura &
Simons-Morton, 2005; Baldry, 2003). However, little is known about organizational or
school variables that could be mediators between personal traits, family factors, and
school violence (Astor, Benbenishty, Vinokur & Zeira, 2006; Astor, Benbenishty, Zeira,
& Vinokur, 2002; Benbenishty & Astor, 2005; Benbenishty, Astor, Zeira, & Vinokur,
2002). In addition, there is a paucity of empirical evidence that investigates how
age-related school contexts (e.g., elementary, junior high and high schools) influence
relationship patterns between risk factors and school violence. Using the first nationally
representative data from an Asian culture, this dissertation aims to:
1. Describe the prevalence of perpetration of school violence in Taiwan and to
examine how gender, age, and school type relate to students’ perpetration of
school violence against students and teachers in an Asian culture context.
2. Examine the pattern of relationships among risk factors (i.e., negative
personal traits, parental monitoring, prior victimization, school engagement,
at-risky peers, and student-teacher relationship) and school violence
perpetration by students against students and teachers in four distinct school

3
contexts in Taiwan: elementary schools, junior high schools, academic high
schools, and vocational high schools.
3. Determine if school variables mediate between personal traits and family
factors, and school violence perpetration committed by students against peers
and teachers in four distinct school contexts in Taiwan.
4. Evaluate if the overarching patterns of relationships and mediating effects
are similar or dissimilar in academic and vocational high schools, where
educational tracking based on academic ability accounts for different
students’ personal characteristics, family backgrounds, and school dynamics.
5. Explore how theoretical patterns of relationships between risk factors
theorized and tested in Western cultures apply to an Asian culture, namely
Taiwan.
Structure of the dissertation
This dissertation is presented in a multiple manuscript format as five distinct but
conceptually related manuscripts. The data for the empirical studies are drawn from a
comprehensive, nationally representative (N=14,044) study of school violence in Taiwan
(Project of Prevention and Control of School Violence in Taiwan; Wu, Lee, Yin, & Hu,
2000). In this project, students were given a structured questionnaire which included over
150 items in eight domains regarding students’ basic demographic background and other
information in their personal, family, and school experience. This study was supported by
Taiwan’s National Science Council (NSC).

4
This sample was designed to represent all students from 4th to 12th grades in
Taiwan. More importantly, this sample contains the first nationally representative data on
Taiwanese school violence and is the first comprehensive studies devoted exclusively to
school violence in Asia. Because the five empirical inquires presented in this dissertation
are similar in design, portions of the method sections in these five studies will also be
similar, but each will be specifically geared towards the study conducted. The goal of the
multiple manuscripts is to create separate studies that are closer to peer-reviewed journal
manuscripts. These manuscripts are intended to be of publication or near publication
quality to facilitate rapid dissemination of the findings. As such, three original theory
testing studies using multivariate structural equation modeling were originally presented
at the dissertation proposal meeting. These three studies were included in this dissertation.
However, the dissertation committee recommended that two more descriptive studies be
included as a way to situate and frame the more complex studies. Hence, two descriptive
studies of student violence against other students and student violence against teachers
were added to this multiple format dissertation. These are not theoretical but
epidemiological and are intended to give the reader the first representative picture of an
Asian society on issues of school violence. Since each chapter is also meant to be a stand
alone manuscript, parts of the literature review, descriptions of the school system in
Taiwan, and cross cultural studies are similar in some of the manuscripts. Even so, the
studies and literatures aim towards the goals of each manuscript.
This first chapter of this dissertation serves as a brief introduction to the five
manuscript studies. Study 1 (Chapter 2) and Study 2 (Chapter 3) are the first two studies
5
that describe the prevalence of student violence against students and teachers in
Taiwanese schools. These studies examine how students’ gender, grade level and school
type relate to violence against students and teachers. Study 1 focuses on violence against
students. Study 2 focuses on violence against teachers. These two studies also provide
quantitative information regarding the reasons students participate in violent acts against
other students and teachers. This information should be valuable for developing effective
programs aimed at preventing and managing violence against students and teachers. It
will also help situate the theoretically oriented studies within a national context so that
the findings can be understood within this cultural and national perspective.
In Study 3 (Chapter 4), Study 4 (Chapter 5), and Study 5 (Chapter 6) theoretical
models are presented that depict the relationships between variables that ultimately
impact or mediate perpetration in schools. Studies 3-5 present the findings surrounding a
theoretical model that examines how personal traits, prior victimization, family factors,
and school dynamics influence school violence committed by students against students
and teachers. In contrast with earlier psychological theories which stress single risk factor
contributions of prior victimization, family function, and psychological characteristics to
perpetration, this dissertation proposes that school variables act as mediators between
personal and family factors and school violence. Figure 1 presents the theoretical model
that will be tested with different populations for Studies 3-5.
Variables in the theoretical model
Negative personal traits. Students who are lack of impulsive control, have poor
anger traits, and endorse higher aggressive attitudes are more likely to engage in violent
6
behavior in school (Connolly & O’Moore, 2003; McConville & Cornell, 2003). These
associations between school violence and specific personal characteristics are also true
for Taiwanese students (Hu & Lin, 2001).
Parental monitoring. Western researchers have observed that low level of parental
monitoring is associated with school violence (Baldry & Farrington, 2000). Confucianism
strongly encourages close parental monitoring and acts of stringent discipline as means of
preventing or punishing negative behaviors on the part of children (Hu & Lin, 2001).
This implies that lack of parental monitoring as a factor in school violence in Asian
countries.
Victimization. Exposure to various kinds of violence (in the form of direct
involvement or as witnesses) is related to the development of aggressive behavior in
school (Baldry, 2003). This association is also true for Taiwanese students (Hu and Lin,
2001).
Involvement in at-risk peers and low level of school engagement. Western studies
suggest an association between friendship characteristics and bullying (Huttunen,
Salmivalli, & Lagerspetz, 1996). East Asian culture has been characterized as collectivist
orientation, which stresses that individual interests are subordinate to group norms (Chen,
2000; Triandis, 1995). Accordingly, East Asian children are perhaps more likely to be
influenced by their peers and to mimic their behaviors, including violence. Hu and Lin
(2001) have observed and reported on this association in Taiwan. In addition, studies in
West (Natvig, Albrektsen, & Qvarnstrom, 2001) and in Taiwan (Lai, 1998) indicate a
connection between low levels of school engagement and school violence.
7
Poor student-teacher relationships. Western studies suggest that quality of
student-teacher interactions is a factor in school violence (Junger-Tas, 1999; Olweus,
1999). Teachers in Chinese culture enjoy a higher social status compared to their western
counterparts (Fwu & Wang, 2002). Schoenhals (1994) describes the position and role of
teachers in East Asian culture as parallel to those of parents. Inharmonious relationships
between teachers and students may result in the latter feeling frustrated or experiencing
low self-worth, both of which can lead to student involvement in school violence (Wong,
2004; Yoneyama & Naito, 2003).
Theoretical model in the dissertation
The theoretical model in this dissertation is highly influenced by Benbenishty and
Astor’s (2005) model of social-ecological and school influences on school violence. In
this social-ecological model, Benbenishty and Astor propose that school violence is
influenced by numerous within-school variables (such as teacher child relationships,
school goals, and social organization) and other external variables such as students’
personal and family characteristics, culture, religion, and school neighborhood. In
contrast with psychological models that center around the individual, they stress that the
school itself as a social context should be the center of the theoretical model.
In addition, the theoretical model set forth by Benbenishty and Astor (2005)
suggests that the influence of external variables could be direct or, at times, mediated by
school factors. This model has been examined across different cultures and contexts with
Jewish secular, Jewish Orthodox, Arabic, and American students (Astor et al., 2002;
Astor, Benbenishty, Zeira, & Vinokur, 2002; Benbenishty, Astor, Zeira, & Vinokur, 2002;
8
Benbenishty & Astor, 2005). However, studies conducted with this theoretical orientation
have not examined how these complex patterns within and outside of the school affect
perpetration in Asian culture contexts. Moreover, school violence research has not yet
addressed how school factors can mediate between individual and family factors, and
school-based perpetration. This is important as students’ family experience may have an
indirect association with school violence that is mediated through variables on the school
grounds. For example, students possessing more positive attitudes toward violence will
be more likely to have more at-risk school peers; in turn, they will be more likely to
involve in school violence.
The theoretical model in this dissertation depicts the three school variables in the
center (low level of school involvement, risk peers, and poor student-teacher relationship),
external school variable in the left (negative personal traits, parent monitoring, and
victimization) and school violence perpetration outcomes in the right (student violence
against students and student violence against teachers). These within-school and
external-school variables in this dissertation are mostly cited in the school violence
literature. In a broad sense, the proposed theoretical model will examine how these
school variables act as mediators between personal and family factors and school
violence. Particularly, this dissertation will focus on examining this theoretical model in
different age groups and school types.
The examination of the same theoretical model in different school types and age
groups is important for gaining a comprehensive understanding of how school variables
mediate personal and family factors, and school violence. For example, many
9
developmental theories suggest that teachers and peers are influential to student behavior
in pre-adolescence. When students enter middle school or early adolescent age, the peer
influence becomes more important than the influence of school adults such as teachers
(e.g., Benbenishy & Astor, 2005; Wong, 2004). In this sense, involvement with at-risk
peers in this dissertation may play more important role than poor student-teacher
relationships in mediating school violence as student develop in middle and high schools.
The three studies will examine these questions empirically. Specifically, each will
examine if different aspects of the school social environment mediate different violence
outcomes with students in the different developmental social and academic contexts of
elementary, junior high and high school.
This dissertation (Study 5) will also explore the effects of educational tracking on
perpetration outcomes. For example, some studies suggest that educational tracking based
on academic ability accounts for different school structures and social dynamics between
vocationally versus academically-oriented high schools in Taiwan (Gang & Klaus, 2000).
Some educators predict that environmental differences between vocationally and
academically-oriented schools mediate school violence in different ways (Chang, 1992;
Gang & Klaus, 2000; Hsieh, 2003). Alternatively, some researchers argue that the
mediating mechanism may not vary significantly between the different school types in
different cultures (Benbenishy & Astor, 2005; Turiel, 2002). Empirical evidence has not
yet emerged to suggest how school variables mediate personal and family variables in
academic vs. vocational school settings. If there are differences in the influences and
patterns between academically vs. vocational high schools, it may suggest that new
10
school violence intervention programs could be developed to address the specific needs
of each school type. Alternatively, if the school mediating mechanisms are similar
between school types, different interventions for different school types may not be
necessary. Policymakers, practitioners, and researchers might have more empirical
evidence to justify using similar or separate types of interventions or for different types of
schools.
Thus, this dissertation will separately examine the same theoretical model in
elementary school, junior high school, and high school. Study 3 will present the results
examining the theoretical model based on data from elementary school students. Study 4
will examine the same theoretical model based on the data of junior high school students.
Study 5 will first examine the theoretical model based on all high school students. The
primary analysis method in Studies 3-5 for testing the strength of the proposed theoretical
model is structural equation modeling (SEM).
The final chapter of this dissertation will elaborate on the implications of the major
findings across all five empirical studies, with an emphasis on how they relate to each
other and the international school violence literature. In addition, this chapter provides an
analysis of broad cultural and theoretical issues with suggestion for future research.
Overview of Taiwanese education system
The findings reported in this dissertation were from studies conducted in Taiwan.
Thus, it is important to describe the Taiwanese educational system. Generally, there are
three major systems of basic and intermediate education in Taiwan: elementary school
(grades 1 to 6), junior high school (grades 7 to 9), and high school (grades 10 to 12).
11
High schools could be further divided into academic and vocational school tracks.
Attending a nine-year compulsory education program from elementary to junior high
school is mandatory. In 2006, there were about 3.8 million students in 3,858 schools in
Taiwan.
Overall, the immediate educational purpose and school structure appear to differ
between these four types of schools. Elementary school is the first formal education
children that receive, and the paramount aims are literacy and basic science. Each
classroom has a homeroom teacher who is responsible for taking care of students and
teaching most of the classes during the day. In contrast to other school types, elementary
school students have the least competitive learning environment, because they can enter
into junior high school without any entrance exam. Overall, the goals and structures of
elementary schools in Taiwan appear very similar to those of elementary schools in
Western countries. In fact, the role of the teacher, the focus of the classroom, peer
relations, and daily structure are nearly identical to most other Western industrial
countries.
Academic achievement is highly emphasized in junior high school. In this respect,
junior high schools in Taiwan function more similarly to private preparatory junior high
schools (or some affluent suburban schools across the USA) in the West. In general, the
goal is to prepare the student for the best high school and, subsequently the best college.
Parents in Taiwan spend a great deal of time monitoring students’ homework and
communicating about current school experiences, especially academic performance. To
ensure top grades, students are often sent by their parents to “cram school” for advanced
12
academic training. This is similar to the role private tutoring and test preparation
companies play in the USA. The goal is to maximize students’ performance on
standardized tests and gain admission to the highest track academic units (e.g., entrance
to AP, Honors, or college level classes in USA schools).
There are some distinct differences in the structure of Taiwanese junior high
classrooms. While Western secondary students often have multiple teachers with different
students in each class, Taiwanese students remain in the same class with the same cohort
of students for three years, and a homeroom teacher is assigned to each class. The main
responsibility of homeroom teachers is to help students achieve higher academic
performance. Teachers are perceived as intellectual experts, and each subject is taught by
one specialized teacher in that field. Thus, the teachers in academic fields rotate from
class to class. Most students arrive before school for tests. Students spend about ten hours
in school per day and additional two to three hours in cram schools. To achieve the
highest status, most junior high schools encourage teachers to design innovative
programs or conduct mock tests to help students succeed. Outstanding teachers are
recognized publicly if their class has a high admissions rate for academic high schools.
A competitive joint high school entrance exam must be taken if junior high school
graduates seek higher education. Students are assigned to academic or vocational high
schools based on their exam results. This is similar to how middle school students in the
West who attend elite private schools apply to elite private high schools, competitive
academic magnet schools, or charter national blue ribbon college preparatory high
schools. Generally, students with higher scores are encouraged and even expected to
13
attend academic high schools, while those with lower scores have less choice and are
tracked towards attending vocational or less rigorous high schools. This separation by
academic ability based on test scores and academic school performance could account for
different school structures and social climates in academic and vocational school settings.
Three-year academic high school programs aim to prepare students for specialized
learning as well as for college study, focusing on training for the joint university entrance
examination. As a result, academic high school students offer intense coursework in
mathematics, the sciences, and academic scholarship. The goal of vocational high schools,
however, is to provide students with professional competencies and technical skills to
help vocational high school students enter the workforce as competent entrepreneurs or
workers in companies. Most vocational high school graduates start a business or take up
employment in companies based on the specific types of technical skills that they learn in
high school. V ocational high schools in Taiwan are similar to those in Europe and the
Middle East. In the United States, most high schools offer only a few vocational
components. Except for some alternative schools for “at risk” students, it is rare to have
high schools exclusively devoted to vocational issues. Moreover, the separation of
students based on academic ability does not occur as frequently in the U.S. The
boundaries and distinctions between academic and vocational high schools are much
stronger in Taiwan.
It remains unclear from an empirical perspective how prevalence of school violence
and the patterns of relationships differ in each of these settings. From one perspective, the
goals and student population are entirely different for vocational and academic schools.
14
Alternatively, class structures, peer group influence, teaching styles, and other major
variables appear to be similar in both types of schools. In addition, previous studies
suggest that the structural relationships between risk factors and school victimization are
similar for primary, secondary, and high school students (Astor et al., 2002; Astor,
Benbenishty, Zeira, & Vinokur, 2002; Benbenishty, Astor, Zeira, & Vinokur, 2002;
Benbenishty & Astor, 2005). However, it is unknown if patterns of relationships between
risk factors and school perpetration are similar across school types. Hence, an exploration
of these settings would be revealing.

15
Figure 1. Theoretical Model for Perpetration of School Violence

16
Chapter Two
Study 1: An Epidemiological Presentation and Analysis of The
Perpetration of Student Related Violence in Taiwan
Introduction
Over the past several decades, the majority of research studies on school violence
have been conducted in the Western countries (Nansel et al., 2001; Pellegrini & Long,
2002; Pepler et al., 2006). Studies on violence in schools in non-Western countries
remain relatively unexplored. A handful of published studies have implied that school
violence is widespread among school children in East Asian countries (e.g., Ando,
Asakura & Simons-Morton, 2005; Yang, Kim, Kim, Shin, & Yoon, 2006). Yet these
studies were conducted using small convenience samples or non-representative data.
Akiba, LeTendre, Baker, and Goesling (2002) utilized nationally representative samples
(provided by the Third International Math and Science Study --TIMSS) to explore
amounts of school violence victimization among 37 nations including countries in East
Asia, such as Hong Kong, Singapore, and Korea. However, the questionnaire in that
study has only two items concerning students’ experience of school violence. In addition,
the main purpose of TIMSS is to explore mathematics and science. Hence, the only
representative sample of Asian cultures to date provides only a minimal glimpse of
school violence in those countries.
Furthermore, there is limited knowledge of how gender, school context, or
developmental findings from Western school violence studies may (or may not) apply to
Asian culture context. Some influential theorists cite empirical evidence from social

17
psychology that suggest cultural differences exist between East and West (Nisbett, 2003).
Even so, some recent international school violence research posits that many seemingly
different cultures have almost identical patterns and structures of school violence even
though they may differ on frequency base rate for any specific form of violence
(Benbenishty & Astor, 2005). Consequently, nationally representative studies are needed
in Asian cultures to gain a clearer understanding of similarities and differences within and
between Asian societies.
Taiwan is typical of many Asian cultures. Although Taiwan has experienced rapid
democratization and industrialization over the past four decades, Taiwanese are still
guided by values stemming from a blend of Confucianism, Taoism, and Buddhism
(Nisbett, 2003). School violence issues have been a major public concern for many years
in Taiwan, but no large-scale cross-national studies on this topic has been conducted (see
for example Akiba, LeTendre, Baker, & Goesling, 2002 with the exclusion of Taiwan in
their sample). In part, this is due to Taiwan’s political stalemate with the People Republic
of China.
Thus, this study aims to use a nationally representative sample in Taiwan to explore
the issues of school violence in an industrialized and highly educated Asian cultural
context. This study includes a comprehensive examination of how school violence is
related to gender and age in Taiwan. The prevalence of school violence in each school
type and between academic and vocational high schools (school contexts are distinct and
age groups are naturally controlled) were also explored. Previous studies have seldom

18
provided information on why perpetrators commit violent acts. The current study
presented students’ attribution to involve in school violence.
The term “school violence” is used in the current study to refer to in-school student
behavior intended to harm other students or to cause damage to other students’
belongings or school property. The definition includes physical and verbal violence,
threatening behavior, and property damage (Astor, Benbenishty, Pitner, & Zeira, 2004;
see Benbenishty & Astor, 2005 for a critical discussion).
The Influence of gender, age and school type on school violence: West vs. East
Gender. Western studies have reported that boys engage in more aggressive behavior
in school than girls (Baldry, 2003; Bosworth, Espelage, & Simon, 1999; Kumpulainen et
al., 1998; Nansel et al., 2001; Rigby, 2005). Consistently, previous studies in Taiwan have
suggested that male students are more violent in school than females (Hu & Lin, 2001).
Grade level and age. Recent studies have shown that prevalence of perpetration of
school violence increases and peaks during early adolescence and then decreases during
high school (Espelage & Swearer, 2003; Nansel et al., 2001; Pellegrini & Long, 2002;
Pepler et al., 2006). It is unclear whether this age trend related to school violence
perpetration holds true for Taiwanese students, but recent reports have shown that rates of
school violence victimization in Taiwan increase in junior high school and decrease
during high school (Wei, Chang, & Chen, 2007).
School types and the social context of Taiwan. Every school context (e.g., primary
school, middle school, and high school) has its own dynamics, mission, and structure.
Many practitioners believe that the prevalence of school violence is dissimilar from

19
school to school because of distinct dynamics in each individual school context (Astor et
al., 2002; Astor, Benbenishty, Zeira, & Vinokur, 2002; Benbenishity & Astor, 2005;
Benbenishty, Astor, Zeira, & Vinokur, 2002). Studies have shown that the prevalence of
school violence is dependent on school type (e.g., Zeira, Astor, & Benbenishty, 2003).
However, few of these studies have controlled the effect of age when investigating the
rates of school violence. It is also unknown whether the prevalence of school violence is
similar when the age structure of students is controlled. This could be looked at easily by
examining different school structures or types that are geared for similar aged students
(e.g., vocational high schools vs. academic high schools)
There are three major education systems in Taiwan: elementary school (grades 1 to
6), junior high school (grades 7 to 9), and high school (grades 10 to 12). High schools
could be further divided into academic and vocational (Taiwan Ministry of Education,
2006).
The educational purpose and school structure appear to differ between these four
types of schools. Elementary school is the first formal education children that receive,
and the paramount aims are literacy and basic science. Each classroom has a homeroom
teacher who is responsible for taking care of students and teaching most of the classes
during the day. In contrast to other school types, elementary school students have the
least competitive learning environment, because they can enter into junior high school
without any entrance exam (Taiwan Ministry of Education, 2006). Overall, the goals and
structures of elementary schools in Taiwan appear very similar to those of elementary

20
schools in the West: the role of the teacher, the focus of the classroom, peer relations, and
daily structure are nearly identical.
Academic achievement is highly emphasized in junior high school. In this respect,
junior high schools in Taiwan function more similarly to private preparatory schools in
the West. In general, the goal is to prepare the student for the best high school and,
subsequently, college. While Western students often have multiple teachers with
different students in each class, Taiwanese students remain in the same class with the
same cohort of students for three years, and a homeroom teacher is assigned to each class.
The main responsibility of homeroom teachers is to help students achieve higher
academic performance. Teachers are perceived as intellectual experts, and each subject is
taught by one specialized teacher in that field (Taiwan Ministry of Education, 2006).
Thus, the teachers in academic fields rotate from class to class. Most students arrive
before school for tests. Students spend about ten hours in school per day and additional
two to three hours in cram schools. To achieve the highest status, most junior high
schools encourage teachers to design innovative programs or conduct mock tests to help
students succeed. Outstanding teachers are recognized publicly if their class has a high
admissions rate for academic high schools.
A competitive joint high school entrance exam must be taken if junior high school
graduates seek higher education. Students are assigned to academic or vocational high
schools based on their exam results (Taiwan Ministry of Education, 2006). This is similar
to how middle school students in the West apply to elite private schools, competitive
magnet schools, or national blue ribbon college preparatory high schools. Generally,

21
students with higher scores are encouraged to attend academic high schools, while those
with lower scores attend vocational or less rigorous high schools (Taiwan Ministry of
Education, 2006). This separation by academic ability could account for different school
structures and social climates. Three-year academic high school programs aim to prepare
students for specialized learning as well as for college study, focusing on training for the
joint university entrance examination. As a result, academic high school students offer
intense coursework and concentrate on academic development. The goal of vocational
high schools, however, is to provide students with professional competencies and
technical skills to help vocational high school students enter the workforce. Most
vocational high school graduates start a business or take up employment in companies
based on the technical skills that they learn in high school. V ocational high schools in
Taiwan are similar to those in Europe and the Middle East. In the United States, most
high schools offer only a few vocational components. Except for some alternative schools
for “at risk” students, it is rare to have high schools exclusively devoted to vocational
issues. Moreover, the separation of students based on academic ability does not occur as
frequently in the U.S. The boundaries and distinctions between academic and vocational
high schools are much stronger in Taiwan.
It remains unclear how the rates of perpetration differ in each of these settings. The
school goals and student populations are entirely different for vocational and academic
schools; however, class structure, peer group influence, teaching style, and other major
variables appear to be similar in both types of schools. Hence, an exploration of these
seemingly diverse settings will be helpful in understanding how school dynamics

22
contribute to violence in settings which serve high academic functioning students differ
from those serving low academic functioning students.
Methods
Dataset and Sample
The data used in this study was part of a large-scale project of “Prevention and
Control of School Violence in Taiwan” (Wu, Lee, Yin, & Hu, 2000). This project
included qualitative and quantitative information. This study mainly focuses on the
quantitative data. The survey was conducted throughout Taiwan among over 14,000
students from elementary schools (grades 4 to 6), junior high schools (grades 7 to 9),
vocational high schools and academic high schools (grades 10 to 12). Students were
given a structural questionnaire in classrooms under the guidance of professionally
trained survey monitors. Respondents were assured of anonymity and were encouraged to
respond truthfully. Participants were free to withdraw from the study at any time and for
any reasons. This study was supported by Taiwan’s National Science Council (NSC). The
questionnaires, procedures, informed consent forms and other ethical concerns were
reviewed and supervised by NSC (see Appendix A for detail items).
This sample was designed to represent all students from 4th to 12th grades in
Taiwan. Table 1 shows the distribution of students in Taiwanese schools participating in
this study. The students’ response rate was over 98 percent. The probability sampling
method was a two-stage stratified cluster sample. The strata were
northern/central/southern/eastern, urban/rural, and elementary/junior/technical/academic.
In first stage, schools were randomly selected from the sampling frame according to those

23
appropriate strata. In the next stage, two classes were randomly selected according to
each grade in selected schools. All students in that class were included in sample.
The questionnaire was developed on the basis of current school violence studies and
theories from both Taiwan and Western countries, and validated by Wu and colleague
(2000). It included over 150 items in eight domains regarding students’ basic
demographic background and other information in their personal, family, and school
experience such as students’ attitude towards aggression, family climate, interpersonal
relationships in school, and students’ self-reported violent behavior towards peers and
teachers. Each of the student questionnaires took approximately 30 minutes to complete.
Before this survey was conducted, the Mandarin Chinese questionnaire was adjusted and
adapted based on two pilot studies conducted in the Tainan metropolitan areas in Taiwan
(Wu et al., 2000).
Measurement
School violence ( α=.80). This domain involved asking students to indicate how
many times they exhibited violent behavior in school toward other students as well as
how many times they damage school public and other students’ belongings during the last
year. Students could check off one of four categories: never, one or two times, three to
five times, or more than five times. The variables in this domain were then recorded into
“never” and “at least once”. These items included, for example: “Students beat or kick
other students in order to hurt him/her/them”.
Self reported reasons for the perpetration violence. Twelve reasons for the students
self reported school violence perpetration was provided, followed by each violent

24
behavior. After they indicated that they had been involved in a specific violent behavior,
students were asked to indicate one or more reasons for why they did so. These twelve
items were the major reasons for violence based on the previous qualitative analysis in
Taiwan (Wu et al., 2000), including “I acted this way because I was provoked”.
Gender. Students were asked to indicate whether they were male or female.
Grade levels. Student grade was identified in terms of grade years, ranging from 4th
to 12th grades.
School types. Students were asked to indicate their school type: elementary school,
junior high school, academic high school, or vocational high school.
Results
Table 2 presents data on the prevalence of certain types of violent behavior by
gender and school type. Overall, 59.7 percent of Taiwanese students reported that they
involved as perpetrators in at least one act of school violence during the preceding year.
This result suggests that the school climate in Taiwan is not peaceful. Among all acts of
violence, cursing or verbally insulting was the most common violent act among all
students (43.8%). This suggests that verbal violence is more prevalent than other types of
violence.
Gender
About 70 percent (71.2%) of male students and 48.5 percent of female students
reported involving in at least one violent act (see table 2). In addition, boys reported
higher rates of each violent behavior compared to girls (range of 3.3% to 53.2% for boys
and range of 0.8% to 34.7%). For example, almost one quarter (24.9%) of the boys

25
reported that they had been involved in beating or kicking others one-to-one, compared
with 6.1 percent of girls. These results suggested that male students are more violent than
female students in Taiwan.
School type
Overall, junior high school students (68.0%) reported higher rates of involving in at
least one violent act compared to elementary schools (58.8%), academic high schools
(53.0%), and vocational high schools (60.4%) (see table 2). In addition, students in junior
high school generally reported higher perpetration rates for almost all violent behaviors
(range from 2.0% to 44.1% for elementary school, range from 3.3% to 53.6% for junior
high school, range from 1.2% to 35.6% for academic high school, and range from 2.0%
to 43.8% for vocational high school). Compared to academic high schools, students in
vocational high schools reported higher rates of perpetration for each violent behavior.
These results suggest that there are different rates of perpetration of school violence
between school types.
Grade level
Figure 1 presents the rates of school violence perpetration by gender, age/grade, and
school type. Overall, the rates of perpetration increased from 4th grade (52.3%), peaked
at the point of 8th grade (71.4%), and declined with increasing grades. However, different
patterns exist between the two types of high schools. The rate of perpetration continually
declined from 10th (64.8%) to 12th (57.4%) grades in vocational high schools, but
declined from 10th (60.7%) to 11th (47.5%) grades and increased from 11th to 12th
(49.9%) grades in academic high schools.

26
For both genders, the rates of perpetration increased with age, from 4th grade
(58.5% for male and 45.8% for female) to 8th grade (78.2% for male and 65.0% for
female). There was a decrease from 8th grade to 9th grade (44.2%) for females, but
almost no change from 8th to 9th grades (78.4%) for males. In vocational high schools,
both male and female students in 10th grade (79.1% for males and 50.8% for females)
reported higher rates of perpetration than those in 9th grade, after which the rates
declined with increasing grades. In academic high schools, rates of perpetration declined
from 10th grade (73.0% for males and 51.5% for females) to 11th grade (63.5% for males
and 33.8% for females), and slightly increased from 11th to 12th grades for both males
(65.7%) and females (37.4%).
Student self reported reasons for the perpetration of violence
Table 3 shows the distribution of students’ reasons for involvement with violent
activities. The most common reasons cited were having fun, disagreement of opinion, and
being provoked. Over half (53.3%) of perpetrators reported involvement in school
violence perpetration because of having fun, and 41.1 percent reported involvement as a
result of a disagreement of opinion or being provoked by others. These three common
reasons were consistently given across gender (ranging from 42.3% to 56.2% for males
and from 32.7% to 49.2% for females) and school type (ranging from 34.4% to 37.6% for
elementary schools, from 47.6% to 54.3% for junior high schools, from 38.5% to 62.0%
for academic high schools, and from 39.1% to 58.3% for academic high schools). Some
perpetrators reported involving in violence due to being emotionally upset (28.2%), being
beaten or victimized first (16.6%), object’s weakness (15.0%), someone tattle tailing or

27
snitching (14.8%), being punished (13.8%), and someone doing things wrong (12.6%).
The least common reasons cited were money issues (3.5%) and boyfriend/girlfriend
relationships or love affairs among students (6.2%). These results indicate that students’
reasons for violence in Taiwan are varied and most of their reasons for violence are
related to provocation.
Discussion
This study is the first nationally representative sample to provide empirical evidence
related to the prevalence and scope of school violence perpetration in Taiwan. The sample
is representative of the entire student population from 4th to 12th grades in Taiwanese
schools. Students’ reports were compared by gender, school type, and grade levels, and
provided quantitative information about students’ reasons for violent behaviors.
Overall, 59.7 percent of the Taiwanese students in the sample reported that they
had perpetrated in at least one violent act of behavior in the preceding twelve months. In
comparison, the reported prevalence of school violence perpetration in other countries
was 48.3 percent in Italy (Baldry, 2003), 44.3 percent in the U.S. (Nansel et al., 2001), 38
percent in England (Boulton & Underwood, 1992), 25 percent in Australia (Slee, 1995),
24.4 percent in South Korea (Yang et al., 2006), 2.7 to 40.7 percent in Japan (Ando,
Asakura & Simons-Morton, 2005). In addition, the results from a large-scale international
study on health behavior in school-aged children (HBSC; Craig & Harel, 2004) indicated
that the percentage of students who bullied others in thirty participating countries ranges
from 9 percent in Sweden to 73 percent in Lithuania. Although the methodologies used to
measure school violence perpetration differ between previous studies and the present

28
study, the data indicates a relatively high rate of violence among Taiwanese students
compared to many other countries.
These findings do not support contemporary and popular theories asserting that
certain Asian cultural values such as emphasizing harmony in social relationships may
account for lower prevalence of school violence than Western cultures (Nisbett, 2003).
The findings suggest that the different cultural values between East and West may be not
so prominent in influencing the level of severity of school violence. However, the
inability to validate those cultural theories may be due to sample or different definitions
of school violence between studies. Future researchers must carefully consider their use
of expanded methodologies to accurately reflect differences across multiple cultures.
This study indicates that cursing, insulting, teasing, and mocking are the most
common violent acts among students in Taiwanese schools just as they are in almost all
other school violence studies conducted across the globe. Fewer students (but still a
significant number) reported physically assaulting others, damaging property, threatening,
and coercing. The least common violent act reported is extorting or blackmailing. These
findings are generally similar to results of perpetration of school violence in Western
countries, which show that cursing, insulting, and mocking are the most common violent
acts (Wolke, Woods, Standford & Schulz, 2001; Pateraki & Houndoumadi, 2001), and
extorting is one of the least common types of school violence (Wolke et al., 2001).
This study indicates that the prevalence of school violence perpetration increases
from 4th grade and peaks in 8th grade, after which it declines with increasing grades.
This developmental pattern is similar to another finding in this study, that junior high

29
school students have higher rates of school violence perpetration compared to elementary
and high schools. These findings imply that development may play an important role in
school violence. In addition, the development-related findings in this study are similar
with the findings of recent Western studies (Espelage & Swearer, 2003; Nansel et al.,
2001; Pepler et al., 2006), which have shown that violent behavior in school increases
and peaks during early adolescence, and then decreases during high school years.
This study indicates that students in vocational high schools in Taiwan have higher
rates of school violence perpetration compared to those in academic high schools. In
addition, there are different patterns of trends between the two types of high schools. For
example, this study found that the rates of perpetration continually decline with
increasing grades in vocational high schools; however, the rates decline from 10th to 11th
grades, and increase from 11th to 12th grades in academic high schools. The trends in
academic and vocational high schools also occur across gender in each school type.
These findings provide strong empirical evidence that different school dynamics or
mission of the contexts may affect the rates and patterns of school violence perpetration.
In this study, Taiwanese male students reported much higher rates of school violence
perpetration than female students. This finding of a gender difference is consistent with
studies in several countries (Baldry, 2003; Bosworth et al., 1999; Kumpulainen et al.,
1998; Nansel et al., 2001; Rigby, 2005). In this study, the prevalence of school violence
was highest for girls in 8th grade and for boys in 10th grade in vocational high schools,
and in 8th grade and 9th grade, respectively, in junior high schools. These findings are
somewhat different from those in a U.S. large-scale study by Nansel et al. (2001), which

30
showed that the highest prevalence for boys was in 8th grade and was in 9th grade for
girls. However, in the current study, one of the lowest rates of perpetration for girls is in
9th grade. In addition, the biggest difference of the perpetration rates between male and
female students appears in 9th grade, and the differences are bigger in both types of high
schools than in elementary schools. These findings may imply that preadolescent boys
and girls are more similar in school violence involvement compared to adolescent boys
and girls.
This study indicates that the majority of Taiwanese school violence perpetrators
involved in violence because of having fun, being provoked, or disagreement with others.
In addition, some perpetrators engaged in violence because they reasoned that the object
is easier to be bullied, the object did something wrong, someone is snitching, or because
perpetrators were punished, beaten, and upset at others. Although some perpetrators
indicated that their reasons for violence were not listed in this study, the findings of the
present paper suggest that perpetrators make different judgments and differ in reasoning
when committing violent acts in schools, and most of their reasons for violence are
related to provocation (Astor, 1994; Astor & Behre, 1997; Pitner, Astor, Benbenishty,
Haj-Yahia, & Zeira, 2003a, 2003b). These findings regarding perpetrators’ reasoning are
similar across gender and school type, which suggests that perpetrators’ reasons for
violence are similar between male and female students, as well as between elementary,
junior high, academic high, and vocational high schools.
Several study limitations should be considered. First, this study was based on
cross-sectional data, so the results cannot be used to establish a developmental

31
progression of violent behavior, which requires a longitudinal study. Second, sexual
harassment (Zeira, Astor, & Benbenishty, 2002) and relational aggression (Crick &
Grotpeter, 1995) were not investigated in this study. In order to comprehensively
understand the scope of school violence in Taiwan, future research needs to address these
additional types of violent behavior. Third, the self-reporting characteristic of the data
and the twelve-month reporting window may have resulted in students under-reporting
events due to poor memory or over-reporting violent behavior that they assume to be
commonplace (Benbenishty & Astor, 2005). Future researchers will benefit from
collecting information from additional sources, such as teacher or peer reports.
Implications
This study showed that school violence in Taiwan is a serious problem that needs the
immediate attention of educators and social workers. The study also suggested that the
systematic planning of an intervention program is necessary in order to prevent serious
school violence problems. However, there is no clear policy or complete intervention
program regarding school violence in Taiwan. These findings based on the nationally
representative data including all students from 4th to 12th grades in Taiwan provide
valued information for establishing policy on school violence. Potential school violence
intervention programs may need to target male and junior high school students, and
should pay more attention to decreasing cursing, insulting, teasing, and mocking behavior
among students. These intervention programs should be designed to address students’
specific needs (Astor, Meyer, Benbenishty, Marachi, & Rosemond, 2005; Zeira et al.,
2003), because the current study has shown that the prevalence of school violence may

32
differ among ages and school types. In addition, school administrators need to advocate
the severity of school violence, and intervention may also focus on resolving peer conflict
and increasing positive peer interactions, because many perpetrators reported that they
were involved in violence due to negative peer interaction, such as being provoked and
disagreement.
The similarities in gender, school type, and development-related findings between
Taiwan and Western countries may suggest more similarities than differences in school
violence between the East and West. Future studies should be conducted in other Asian
countries in order to further confirm the similarities between the East and West in school
violence findings. It is quite possible that there is great variation in rates of school
violence for Asian countries (within and between them) just as there is great variation in
rates of school violence in European and American cultures or countries. In addition,
similar findings between Taiwan and Western countries may provide potential
consideration for developing internationally-based intervention. In order to design more
effective internationally-based school violence intervention programs, future researchers
need to address how psychosocial factors related to school violence perpetration in
Western countries apply to non-Western countries.

33
Table 1: Distribution of Students Participating in the Study
Overall Male Female
Elementary
Grade 4
Grade 5
Grade 6
Total

907
1,114
1,099
3,123

477
589
547
1,613

426
516
548
1,492
Junior High
Grade 7
Grade 8
Grade 9
Total

943
1,038
1,070
3,058

489
510
473
1,475

437
500
577
1,514
Academic high
Grade 10
Grade 11
Grade 12
Total

1,345
1,210
1,211
3,769

574
550
534
1,659

769
654
674
2,098
V ocational high
Grade 10
Grade 11
Grade 12
Total

1,428
1,287
1,351
4,072

708
693
731
2,135

719
589
617
1,928
Total 14,022 6,882 7,032

34
Table 2. Percentage of Perpetration in Violent Behavior by Gender and School Type.
Overall Gender School type
Male Female ElementaryJunior Academic Vocational
Overall violence 59.7 71.2 48.5 58.8 68.0 53.0 60.4
Curse or verbally insult 43.8 53.2 34.7 44.1 53.6 35.6 43.8
Teasing or mocking 32.3 41.8 23.1 24.6 36.5 32.3 35.0
Destroy personal belongings 16.4 20.9 11.9 16.3 22.0 12.1 16.1
Beating or kicking (group to one) 15.5 22.9 8.2 19.5 18.8 9.2 15.9
Beating or kicking (one to one) 15.4 24.9 6.1 20.8 20.1 9.5 13.3
Destroy school’s public belongings 15.3 20.0 10.5 11.1 21.1 13.5 15.8
Coerce into buying things by force 12.7 17.2 8.3 8.0 16.7 11.5 14.5
Verbally threaten or humiliate 12.7 17.2 8.2 6.9 16.1 11.9 15.4
Use of objects/instruments to hurt 10.4 13.9 6.9 11.4 13.5 7.2 10.1
Extort or blackmail 2.1 3.3 0.8 2.0 3.3 1.2 2.0

35
Table 3: Percentage of Perpetrators’ Reasons for Violence by Gender and School Type
Gender School types
Overall Male Female Elementary Junior Academic Vocational
For fun 53.3 56.2 49.2 35.7 54.3 62.0 58.3
Being provoked 41.1 46.9 32.7 34.4 47.6 38.5 42.8
Disagreement 41.1 42.3 39.4 37.6 48.1 39.6 39.1
Upset at someone 28.2 29.6 26.1 15.4 35.6 27.3 32.1
Others 26.8 27.8 25.4 27.6 26.1 25.9 27.6
Being beaten 16.6 19.3 12.7 20.7 21.5 11.3 13.8
Object is weak 15.0 17.8 10.8 9.1 22.0 11.9 16.0
Someone snitch or sneak 14.8 15.4 13.8 15.2 19.8 10.6 13.7
Being punished 13.8 15.6 11.3 16.8 20.5 8.8 10.1
He/she does things wrong 12.6 14.8 9.6 7.9 12.6 14.9 14.3
Love affairs 6.2 6.5 5.7 4.1 7.6 5.6 6.5
Money issues 3.5 4.4 2.0 4.2 5.4 1.7 2.7

36
Figure 2. Distribution of Percentage of Perpetration among Taiwanese Students, by Gender, Grade, and School Type.
0
10
20
30
40
50
60
70
80
90
04 56 78 9 10 11 12
Elementary and Junior Male Elementary and Junior Female Overall Elemnetary and Junior
Academic Male Academic Female Overall Academic
Vocational Male Vocational Female Overall Vocational

37
Chapter Three:
Study 2: The Perpetration of Student Violence Against Teachers in
Taiwan
Introduction
To date, research on issues surrounding school violence against teachers have
focused on teachers’ reports on victimization (Dinkes, Cataldi, Kena, & Baum, 2006;
Lyon & Douglas, 1999). There have been relatively few studies on students’ reports on
perpetration of violence against school teachers (Khoury-Kassabri, Astor & Benbenishty,
2008). In addition, no prior studies have been published on violence against teachers
using nationally representative samples in Asian cultures.
In order to expand the research on violence against teachers, the present study
reports on the first nationally representative inquiry on violence against teachers from the
perspective of school students in an Asian culture context. The main goal of this study is
to describe the prevalence of violence against teachers in Taiwan, and to examine how
students’ gender, grade level and school type relate to violence against teachers. The
prevalence of specific acts of student violence against teachers will be examined. In
addition, this study provides quantitative information on why student perpetrators
participate in these violent acts against teachers. This information should be valuable for
developing sound programs aimed at preventing and managing violence against teachers.
Teachers are the significant adults providing students with knowledge and a safe
environment for social and psychological development. They are expected to be role
models for school children and responsible for protecting students from harm as well as

38
promoting student well-being. Inexplicably, sometimes teachers are targets of violence
committed by their students during school hours (Lyon & Douglas, 1999).
According to a national school crime survey, 7 percent of U.S. primary and
secondary school teachers were threatened with injury, while 3 percent were physically
attacked by a student from their own school between 2003 and 2004 (Dinkes et al., 2006).
Other members of the school staff are also potential victims. For example, another U.S.
national study also showed over three quarters of physical assaults or threats against
school social workers were perpetrated by students (Astor, Behre, Fravil & Wallace, 1997;
Astor, Behre, Wallace & Fravil, 1998; Astor, Meyer, Benbenishty, Marachi & Rosemond,
2005).
Traditionally, in Chinese culture teachers are highly revered professionals. Even now,
teachers in Taiwan enjoy a relatively higher prestige with job satisfaction than
international counterparts (Fwu & Wang, 2002). Students in Taiwan are also expected to
bow to teachers to show respect. Nevertheless, one study reported that 65.5 percent of
surveyed junior high school teachers had been verbally assaulted by their school students
(Chen, 1999).
Studies in Western cultures have shown that violence negatively impacted teachers’
mental health and their quality of teaching (Lyon & Douglas, 1999). Furthermore,
teachers’ victimization may influence students’ psychosocial well-beings and academic
performance. Violence against teachers increases teachers’ fears of personal safety and
often causes them to leave the profession of teaching altogether (Astor et al., 2005; Lyon

39
& Douglas, 1999). However, only few school violence intervention programs are
designed to educate or protect teachers from violence (Astor et al., 2005).
Studies on teacher victimization in Western studies showed the vast majority of
perpetrators were students (Dinkes et al., 2006; Lyon & Douglas, 1999). Unfortunately,
there is no empirical evidence concerning the demographic, contextual, and situational
characteristics associated with specific forms of student violence against teachers. There
is limited knowledge about how students’ gender, age, and school type relate to teacher
victimization. There is even less information about the reasons why students take
aggressive acts against teachers. How student perpetrator characteristics, in addition to
contextual, and situational characteristics relate to teacher victimization in school is
important because without a clear understanding of these variables it is difficult to
formulate effective prevention strategies.
Gender, grade level, school type and violence against teachers
Gender. Literature consistently indicated that male students were involved in more
violent acts against peers in schools than females (Baldry, 2003; Bosworth, Espelage, &
Simon, 1999; Kumpulainen et al., 1998; Nansel et al., 2001; Rigby, 2005). Prior studies
in the Middle East and Western societies also indicate that male students involved in
more violent acts against teachers in schools than females (Khoury-Kassabri, Astor &
Benbenishty, 2008; Lyon & Douglas, 1999).
Grade level/age. Recent studies have shown that prevalence of school violence
among students increase and peak during early adolescence then decrease during high
school (Espelage & Swearer, 2003; Nansel et al., 2001; Pellegrini & Long, 2002; Pepler

40
et al., 2006). It is unclear whether this age trend of violence against students holds true
for student violence against teachers, but teachers’ reports have shown that the
distribution of perpetrators’ ages are mostly concentrated at the age of 15, with the second
highest proportion centered on age 10 (Lyon & Douglas, 1999).
School type. Every school context (e.g., primary school, middle school, and high
school) has its own dynamics, mission, and structure, and these differences may impact
on the prevalence of school violence (Astor et al., 2002; Astor, Benbenishty, Zeira, &
Vinokur, 2002; Benbenishity & Astor, 2005; Benbenishty, Astor, Zeira, & Vinokur, 2002).
Studies on school violence among students have shown that violence was higher in junior
high school than elementary and high schools (Zeira, Astor, & Benbenishty, 2003).
Consistently, studies indicate the prevalence of violence against teachers is associated
with school type. For example, studies show that junior high school teachers are more
likely to be victims of violent crimes in school than elementary and senior high school
(Lyon & Douglas, 1999). In addition, students threatening teachers with injury is more
prevalent in secondary school than elementary school but physical attacks are more
prevalent in elementary schools (Dinkes et al., 2006).
Background of Taiwanese educational systems
The findings reported in this paper were from study conducted in Taiwan. Thus, it is
important to describe the Taiwanese education system. Generally, the educational purpose,
school structure, and teachers’ roles differ between elementary school (grades 1 to 6),
junior high school (grades 7 to 9), academic high school (grades 10 to 12) and vocational
high school (grades 10 to 12) in Taiwan (Taiwan Ministry of Education, 2006).

41
The paramount aims of elementary school are literacy and basic science (Taiwan
Ministry of Education, 2006). In contrast to other school types, elementary school
students have the least competitive learning environment, because they can enter into
junior high school without any entrance exam. Each classroom has a homeroom teacher
who is responsible for taking care of students and teaching most of the classes during the
day.
Academic achievement is highly emphasized in junior high school. In this respect,
junior high schools in Taiwan function more similarly to private preparatory schools in
the West. In general, the goal is to prepare the student for the best high school and,
subsequently, college (Taiwan Ministry of Education, 2006). In contrast to Western
students, Taiwanese students remain in the same class with the same cohort of students
for three years, and a homeroom teacher is assigned to each class. The main
responsibility of homeroom teachers is to help students achieve higher academic
performance. Teachers are perceived as intellectual experts, and each subject is taught by
one specialized teacher in that field. Thus, the teachers in academic fields rotate from
class to class. Outstanding teachers are recognized publicly if their class has a high
admissions rate for academic high schools.
Students are assigned to academic or vocational high schools based on their join
high school entrance exam results (Taiwan Ministry of Education, 2006). Generally,
students with higher scores are encouraged to attend academic high schools, while those
with lower scores attend vocational or less rigorous high schools. Thus, vocational high
school students are more likely to be recognized as failure in academic performance or

42
even life. Three-year academic high school programs aim to prepare students for
specialized learning as well as for college study, focusing on training for the joint
university entrance examination. The goal of vocational high schools, however, is to
provide students with professional competencies and technical skills to help vocational
high school students enter the workforce (Taiwan Ministry of Education, 2006).
It remains unclear how the rates of student perpetration of violence against teachers
differ in each of these settings. On one hand, the school goals, student populations, role
of teachers are entirely different for these types of schools; on the other hand, class
structure, peer group influence, and other major variables appear to be similar. Hence, an
exploration of these settings will be helpful in understanding how the school dynamic
affects the prevalence of student perpetration of violence against teachers.
Methods
Dataset and Sample
The data used in this study was part of a large-scale project of “Prevention and
Control of School Violence in Taiwan” (Wu, Lee, Yin & Hu, 2000). The survey was
conducted throughout Taiwan among over 14,000 students from elementary schools
(grades 4 to 6), junior high schools (grades 7 to 9), vocational high schools and academic
high schools (grades 10 to 12). Students were given a structured questionnaire in
classrooms under the guidance of professionally trained survey monitors. Respondents
were assured of anonymity and were encouraged to respond truthfully. Participants were
free to withdraw from the study at any time and for any reasons. This study was

43
supported by Taiwan’s National Science Council (NSC). The questionnaires, procedures,
informed consent forms and ethical issues were reviewed and supervised by NSC.
This sample was designed to represent all students from 4th to 12th grades in
Taiwan. Table 1 showed the distribution of students in Taiwanese schools participating in
this study. The students’ response rate was over 98 percent. The probability sampling
method was a two-stage stratified cluster sample. The strata were
northern/central/southern/eastern, urban/rural, and elementary/junior/technical/academic.
In first stage, schools were randomly selected from the sampling frame according to those
appropriate strata. In the next stage, two classes were randomly selected according to
each grade in selected schools. All students in that class were included in sample.
The questionnaire was developed on the basis of current school violence studies and
theories from both Taiwan and Western countries, and validated by Wu et al. (2000). It
included over 150 items in eight domains regarding students’ basic demographic
background and students’ self-reported violent behavior in schools. Each of the student
questionnaires took approximately 30 minutes to complete. Before this survey was
conducted, the Mandarin Chinese questionnaire was adjusted and adapted based on two
pilot studies conducted in the Tainan metropolitan areas in Taiwan.
Measurement
Student violence against teachers ( α=.60). This domain involved asking students to
indicate how many times they exhibited violent behavior in school toward teachers
during the last year. The variables in this domain were coded into “never” and “at least
once”. These items were:

44
Have you
z Beaten or kicked a teacher(s)?
z Used dangerous objects or instruments to harm teachers?
z Cursed or insulted teacher(s) in front of him/her/them?
z Threatened or humiliated teacher(s)?
z Extorted or blackmailed teacher(s)?
z Teased, mocked, or played physically harmful tricks on teacher(s)?
z Purposely opposed teacher(s) in order to cause psychological harm?
Reasons students gave for violent perpetration against teachers. For each violent
behavior, nine reasons for school violence were provided. After they indicated that they
had been involved in a violent behavior, students were asked to indicate one or more
reasons why they engaged in the violent act. These nine items were the major reasons for
violence based on the previous qualitative analysis in Taiwan (Wu et al., 2000). These
items were
I had engaged in this violent behavior because (of):
z Being punished by teacher(s);
z Fighting for my friends and express my support on friends;
z Teacher’s unfair treatment;
z Disagreement with teacher’s opinions;
z Being provoked by teacher(s);
z The teacher(s) was/were easy to bully;
z The teacher(s) made unreasonable academic requirements or requests;

45
z Being upset at teachers or dislike(s) teacher(s);
z Other reasons.
Results
Overall, 30.1 percent of Taiwanese students reported involving in at least one
aggressive act against their school teachers during the year preceding the survey. Table 4
shows the percentage of perpetration against teachers by gender and school type.
Purposely opposing teachers in order to causing psychological harm is the most common
aggressive act against teachers (28.0%). Other common types of aggressive acts are
cursing or verbally insulting (6.5%) and teasing, mocking, or playing harmful tricks
(5.7%). Beating or kicking (1.2%), hurting with instruments (1.2%), threatening or
humiliating (1.0%), and blackmailing (0.6%) are the least common types of student
violence against teachers.
Overall, the highest rate of perpetration is in vocational high schools (35.8%) and
the lowest in elementary schools (16.1%). The rates of perpetration are almost the same
in junior high (33.0%) and academic high (33.3%) schools. Similar rates of beating or
kicking teachers are found between different school types (range from 0.9% in academic
high school to 1.4% in elementary and junior high schools). The rates of using dangerous
instruments to harm teachers are higher in junior high schools (1.9%) than elementary
schools (0.9%), academic high schools (1.0%), and vocational high schools (1.2%). In
addition, there were low rates between different school types in threatening or
humiliating teachers (range from 0.5% in elementary schools to 1.2% in junior and
vocational high schools) and in the act of extorting or blackmailing teachers (range from

46
0.4% in elementary schools to 0.9% in junior high schools). Junior high school students
reported three times higher rates of cursing or insulting teachers (9.6%) than elementary
school (3.0%). Academic (5.7%) and vocational high school students (7.5%) reported
lower rates than junior high school. V ocational high school students reported the highest
rates of mocking, teasing, or playing harmful tricks on teachers (6.6%), and
psychologically harming teachers by opposing (34.4%) than other school types.
Table 5 represents the distribution of reasons students cited for violence against
teachers. A majority of perpetrators reported they involved in violence against teachers
because of a teacher’s unreasonable requirements (55.7%), teacher’s unfair treatment
(48.6%), and disagreements with teachers (41.6%). Some perpetrators took aggressive
action because of being punished (23.2%), being provoked by teachers (11.5%), fighting
for friends (13.2%) and being upset (16.1%). A small number of perpetrators involved in
violence because the teachers on their target were easily bullied (5.9%).
Figure 3 shows the percentage of perpetrating of at least one aggressive act against
teachers by gender, grade and school type. Rates of perpetration are higher for male
students (range from 13.2% at 4th

grade to 46.9% at 10th grade in vocational high school)
than females (range from 7.7% at 4th grade to 37.7% at 10th grade in vocational high
school). For both genders, the rates of perpetration increase from 4th grade to 9th grade
(46.1% for males and 34.5% for females). Male students at 10th grade in vocational high
school reported slightly higher rate of perpetration than those in academic high schools
(44.4% for academic high schools and 46.9% for vocational high schools). The rates of
perpetration for male students in both high schools decreased from 10th to 11th grades

47
(37.0% for 11th grade in academic high schools and 32.9% for 11th grade in vocational
high school), however, male vocational school students at 11th grade have a sharper drop
in rates than academic schools resulting in lower rates of perpetration for vocational
students in both 11th and 12th grades (40.6% for 12th grade in academic high schools
and 34.7% for 12th grade in vocational high schools). By contrast, female students have
similar patterns of rates between both types of high schools, showing decreases from 10th
grade (30.9% for academic high schools and 37.7% for vocational high schools) to 11th
grade (23.5% for academic high schools and 29.7% for vocational high schools) and
increases from 11th grade to 12th grade (34.7% for 12th grade academic high schools and
31.0% for 12th grade vocational high schools).
Discussion
This study is the first nationally representative sample to provide empirical evidence
related to the prevalence and scope of student violence against teachers in Asian culture.
The sample is representative of the entire student population from 4th through 12th
grades in Taiwanese schools. Students’ reports were compared by gender, school type,
and grade level, and provided quantitative information about students’ reasons for
violence against teachers. Overall, 30.1 percent of students in the sample reported being
involved in at least one aggressive behavior against teachers during the preceding twelve
months.
This study indicates that opposition in order to cause psychological harm is the most
common type of aggression against teachers. Some students reported cursing or verbally
insulting and teasing, mocking, or playing harmful tricks on teachers. Beating, kicking,

48
using instruments to physically hurt, threatening, humiliating, and blackmailing are the
least common of violence against teachers. These findings suggest that students tend to
be involved in more non-physical violence toward teachers than physical. In addition, this
study shows that most perpetrators involved in violence against teachers because they
perceived unfair treatment from teachers, teachers’ unreasonable requirements or requests,
and differing opinions from teachers. Some perpetrators reported they took aggressive
acts toward teachers because they were punished or provoked by teachers. These findings
suggest the majority of student violence against teachers tend to be impulsive or
emotional reactions to some perceived provocation, frustration, and unfair treatments
(Astor, 1994; Astor & Behre, 1997; Lyon & Douglas, 1999; Pitner, Astor, Benbenishty,
Haj-Yahia, & Zeira, 2003a, 2003b; Turiel, 2002).
Consistent with previous studies on violence against teachers (Dinkes et al., 2006;
Khoury-Kassabri, Astor & Benbenishty, 2008; Lyon & Douglas, 1999), Taiwanese male
students are more likely than females to perpetrate violence against teachers. However,
both boys and girls reported low and similar rates of perpetration in physical violence
(beating, kicking, or instrumentally hurting) and threats (threatening, humiliating,
extorting or blackmailing). By contrast, boys reported higher rates of than girls in cursing
or insulting, teasing or mocking, and opposition. These findings suggest that boys and
girls are more similar in physical violence and threats than other forms of violence in this
study. In addition, the differences of the perpetration rates between male and female
students are smaller in elementary schools than both types of high schools. This may

49
imply that preadolescent boys and girls are more similar in student violence against
teachers compared to adolescent boys and girls.
This study indicates that the prevalence of student violence against teachers
increases from 4th grade to 9th grade and continually increases to 10th grade in
vocational high schools; but declines to 10th grade in academic high schools. After this,
the perpetration declines to 11th grade and then increases to 12th grade in both high
schools. This finding may suggest the important roles of development in violence against
teachers. This developmental pattern is similar to the findings of Western reports (Lyon &
Douglas, 1999), which suggest that perpetrators of student violence against teachers
increased from the age of 5, peaking at age 15.
Overall, vocational high school students reported the highest rate of violence against
teachers while elementary school students reported the lowest among all school types in
this study. Although the prevalence of student violence against teachers is almost the
same between academic high school and junior high school, junior high school students
reported higher rates in cursing, physically attack, and threat, and less in opposing
teachers than academic high school. These findings provide the evidence that how
different school types or school structures influence the prevalence of student violence
against teachers. In addition, notwithstanding the same rates of beating or kicking, this
study indicates that the rates of each type of violent behavior in elementary school are
lower than half of the rates in junior high school. This finding is somewhat different from
the results of teachers’ reports on violence against teachers in the United States (Dinkes et
al., 2006), which showed that threatening with injury is more prevalent in secondary

50
school than elementary school while physical attack is more prevalent in elementary
school than secondary school. However, the inconsistent results may be due to different
methodologies between these studies.
Results in this study show that male students at 10th grade in vocational high
schools reported slightly higher rate of perpetration than those in academic high schools.
For males, the rates decline from 10th to 11th grade in both types of schools, however,
male students at 11th grade in vocational schools have a sharper drop in rates than those
in academic high schools resulting in lower rates of perpetration for vocational students
at both 11th and 12th grades. By contrast, female students in vocational high schools
reported consistently higher rates than those in academic high schools. In addition, the
results show that boys and girls reported more similar rates in vocational high schools
than those in academic high schools. These findings suggest that male and female
students have different patterns of rates between vocational and academic settings.
Moreover, academic settings produce boys with higher rates of perpetration and girls with
lower rates. V ocational settings have similar outcomes for male and female students.
Although the rates of all types of violence against teachers in this paper differ from
those in previous reports, the overall patterns of age, gender, and school type are similar
(Dinkes et al., 2006; Khoury-Kassabri, Astor & Benbenishty, 2008; Lyon & Douglas,
1999). This finding suggests that the prevalence of violence may vary from setting to
setting but the influences of age, gender, and school type on violence against teachers
tend to be similar in the Asian, Western, and Middle East cultures. This may be useful
information for school policy makers or clinicians interested in developing new

51
international school violence prevention and intervention strategies. However, caution is
required when interpreting these cross-cultural findings, since most of the cited studies
mainly focused on teachers’ reports on victimization (Dinkes et al., 2006; Lyon &
Douglas, 1999) and current study addresses students’ reports on perpetration with more
comprehensive information about types of violence. For clarifying the cross-culture
similarities or differences on violence against teachers, future researchers must carefully
consider their use of expanded methodologies to accurately reflect differences across
multiple cultures. Future studies are also needed to provide more evidence on the
prevalence and patterns of student violence against teachers in other cultures.
Several study limitations should be considered. First, this study was based on
cross-sectional data, so the results cannot be used to establish a developmental
progression of violent behavior, which requires a longitudinal study. Second, sexual
harassment (Zeira, Astor, & Benbenishty, 2002) and relational aggression (Crick &
Grotpeter, 1995) were not involved in this study. In order to fully understand the scope of
student violence against teachers in Taiwan, future research needs to address these
additional types of violent behavior. Third, the self-reporting characteristics of the data
and the twelve-month reporting window may have resulted in students under-reporting
events due to poor memory or over-reporting violent behavior they assume to be
commonplace (Benbenishty & Astor, 2005). Future researchers will benefit from
collecting information from additional sources, such as school social worker reports
(Astor et al., 1997; Astor et al., 1998; Astor et al., 2005).

52
Implications
This study shows that student violence against teachers in Taiwan is a serious
problem requiring the immediate attention of school administrators. Yet there is no clear
intervention program regarding student violence against teachers in Taiwan. These
findings based on the representative data with all students from 4th to 12th grade in
Taiwan provide important evidence that can be used to establish policy on prevention of
student violence against teachers. Potential intervention programs in Taiwan may need to
target male and junior high school and high school students. Intervention programs
should give more attention to decreasing social and emotional opposition to teachers in
order to psychologically harm (i.e., cursing or insulting teachers and mocking, teasing, or
playing harmful tricks). These intervention programs should be designed to address
specific needs in different school types, because the current study has shown that the
prevalence and patterns of student violence against teachers differs by age, gender, and
school type. In addition, interventions should focus on promoting the quality of
teacher-student interaction. Many perpetrators reported that they involved in violence
against teachers due to perception of teachers’ unfairness, unreasonable requirements or
requests, and punishments.
Findings in this study suggests that the prevalence of violence may vary from
setting to setting but the influences of age, gender, and school type on violence against
teachers tend to be similar in the Asian, Western, and Middle East cultures (Dinkes et al.,
2006; Khoury-Kassabri, Astor & Benbenishty, 2008; Lyon & Douglas, 1999). More
studies are needed to fully understand how academic vs. vocational school settings

53
produce different gender rates. This may be useful information for school policy makers
or clinicians interested in developing new international school violence prevention and
intervention strategies.

54
Table 4. Percentage of Students’ Violence Against Teachers by Genders and School Type
Overall Elementary Junior Academic Vocational
Overall violence
Total
Boys
Girls

30.1
33.5
26.8

16.1
18.2
14.0

33.0
35.1
31.0

33.3
40.8
27.2

35.8
38.2
33.1
Psychologically harm
Total
Boys
Girls

28.0
31.0
25.1

13.5
15.1
11.8

30.3
32.1
28.6

31.3
38.4
25.5

34.4
36.6
32.0
Curse or verbally insult
Total
Boys
Girls

6.5
7.9
5.1

3.0
3.8
2.2

9.6
12.7
6.5

5.7
7.3
4.3

7.5
8.0
7.1
Teasing or mocking
Total
Boys
Girls

5.7
6.4
5.0

3.0
3.1
2.8

6.4
7.4
5.3

6.4
8.0
5.1

6.6
6.9
6.3
Beating or kicking
Total
Boys
Girls

1.2
1.7
0.7

1.4
1.7
1.1

1.4
2.0
0.7

0.9
1.4
0.5

1.3
1.9
0.6
Use instruments to hurt
Total
Boys
Girls

1.2
1.7
0.7

0.9
1.1
0.8

1.9
2.8
0.9

1.0
1.3
0.6

1.2
1.7
0.6
Threaten or humiliate
Total
Boys
Girls

1.0
1.4
0.5

0.5
0.5
0.5

1.2
2.1
0.3

0.9
1.4
0.4

1.2
1.7
0.6
Extort or blackmail
Total
Boys
Girls

0.6
0.9
0.2

0.4
0.5
0.2

0.9
1.5
0.2

0.5
0.7
0.3

0.5
0.8
0.2

55
Table 5. The Percentage of Students’ Reasons for Violence against Teachers
Overall Beat/kick Use objects
to harm
Curse/insult Threaten/
humiliate
Extort/
blackmail
Tease/mock Oppose/rebel
Unreasonable 55.7 36.6 39.4 56.3 36.3 26.6 37.6 54.9
Unfairness 48.6 37.8 34.7 50.9 33.3 26.6 36.5 46.5
Disagreement 41.6 25.6 30.0 32.8 29.6 21.5 21.1 37.2
Punishment 23.2 27.9 28.8 26.6 23.7 30.4 19.2 20.3
Upset or dislike 16.1 32.6 24.7 20.4 35.6 25.3 21.1 13.7
For friends 13.2 23.8 20.6 16.7 18.5 22.8 16.4 10.9
Others 13.2 18.6 16.5 9.3 14.1 16.5 29.2 8.4
Provocation 11.5 22.7 21.8 17.5 25.2 22.8 10.2 9.8
Easy bullied 5.9 14.0 12.9 5.6 16.3 27.8 12.0 4.5

56
Figure 3. Percentage of Students reporting Perpetration Against Teachers by Gender, Grade Level, and School Type
0
10
20
30
40
50
456789 10 11 12
Elementary and Junior Male Elementary and Junior Female
Academic Male Academic Female
Vocational Male Vocational Female

57
Chapter Four:
Study 3: School Variables as Mediators of Personal and Family Factors
on School Violence in Elementary Schools
Introduction
Over the past decades, violence in schools has become a major social problem
affecting elementary school students’ personal, family, and social well-being (Astor,
Benbenishty, Vinokur & Zeira, 2006). Thus far, most of the research on school violence
has focused on dealing with the data involving adolescents (Bosworth, Espelage, &
Simon, 1999; Natvig, Albrektsen & Qvarnstrom, 2001). Less attention has been paid to
the population of primary school children (Pateraki & Houndoumadi, 2001). In addition,
many studies on risk factors of school violence have been conducted in Western countries
(Connolly & O’Moore, 2003; Craig & Harel, 2004). There is a lack of a nationally
representative sample showing that these risk factors are applicable to Asian cultures,
although significant differences appear to exist between East and West in beliefs,
perceptions and worldviews (Nisbett, 2003).
Furthermore, empirical studies on school violence in elementary settings have
examined how students’ personal, family and school experience “risk factors” affect
perpetration of school violence committed by students against students (Marie-Alsana,
Haj-Yahia, & Greenbaum, 2006). International studies suggest that teachers are also the
targets of school violence committed by students, but less is known about how students’
personal, family and school experience contributes to students’ violence against teachers
in school (Khoury-Kassabri, Astor & Benbenishty, 2008). To date, few studies have

58
examined organizational or school variables that could be mediators between personal
and family factors and school violence in the context of elementary schools (Benbenishty
& Astor, 2005). There is also a paucity of empirical evidence that investigates how
gender influence patterns of relationships between risk factors and school violence in
elementary school settings. The present study is the first inquiry using nationally
representative data to examine if school factors could mediate the effect between personal
and family factors and school violence. It is also the first major national study of school
violence in Taiwanese elementary schools.
Taiwan is typical of many Asian cultures. Despite growing influence of globalization,
the Taiwanese people are still guided by Confucianism, Taoism, and Buddhism (Nisbett,
2003). While school violence issues have been a major public concerns for many years,
for political reasons (China’s lack of recognition of Taiwan as an independent country)
Taiwan has never been part of any large-scale cross-national studies on these topics
(Akiba, LeTendre, Baker, & Goesling, 2002). Relationship patterns of risk factors and
school violence have also never been explored in Taiwan.
School violence in this study refers to student perpetration intending to harm other
students and teachers or to cause damage to belongings or school property. The definition
includes physical and verbal violence, psychological harm, threatening behavior, and
property damage (Astor, Benbenishty, Pitner, & Zeira, 2004; see Benbenishty & Astor,
2005 for a critical discussion).
School variables as mediators between risk factors and perpetration
Most theories or studies on school violence have examined in separate studies how

59
students’ personal traits, family, and school experiences independently affect perpetration
of school violence (e.g., Marie-Alsana et al., 2006). However, these different social
contextual influences or risk factors are rarely examined together either in the United
States or other regions (Benbenishty & Astor, 2005). An examination of how multiple
ecological risk factors and nested contexts influence school violence could be
theoretically and practically fruitful. This makes sense since theories and studies have
traditionally postulated that child violent behavior is highly influenced by transaction of
their experiences in various social environments or contexts (Benbenishity & Astor, 2005;
Bronfenbrenner, 1979; Johnson, 1995). Surprisingly, although this call for contextual
research emerged nearly 30 years ago, until recently there have been limit studies
examining how these multiple ecological risk factors interactively influence perpetration
of school violence (Khoury-Kassabri et al., 2008).
Recently, Benbenishty and Astor (2005) addressed this lack of theoretical integration
surrounding the nested contexts of school violence. They propose a theoretical model of
school violence that is influenced by a combination of numerous within-school variables
(e.g., teacher child relationships) as well as other external variables (e.g., students’
personal and family characteristics). In their heuristic model, schools mediate, moderate,
and attenuate contributions from external contexts. School subcontexts also generate their
own direct contributions to different forms of violence. In contrast with earlier
psychological models that center exclusively around the individual, Benbenishty and
Astor stress that when looking at “school violence” the school itself as a social context
should be the center of the theoretical model.

60
Thus far, this model has been examined school violence victimization across
different cultures and school contexts with Jewish secular, Jewish Orthodox, Arabic, and
American students (Astor et al., 2006; Astor, Benbenishty, Zeira, & Vinokur, 2002;
Benbenishty & Astor, 2005; Benbenishty, Astor, Zeira, & Vinokur, 2002). However,
little research has been done with this theoretical orientation on perpetration of school
violence (Khoury-Kassabri et al., 2008). This paper is the first study that examines how
these complex patterns between social contexts affect school violence perpetration in
Asian elementary schools. Empirical evidence that establishes the degree to which school
variables contribute to perpetration in Asian elementary schools is currently nonexistent.
Background of elementary schools in Taiwan
Because the findings reported in this paper were from studies conducted in Taiwan,
it is important to describe the Taiwanese elementary school system. Generally, there are
three major systems of basic and intermediate education in Taiwan: elementary school
(grades 1 to 6), junior high school (grades 7 to 9), and high school (grades 10 to 12).
Attending a nine-year compulsory education program from elementary to junior high
school is mandatory. A competitive joint high school entrance exam must be taken if
junior high school graduates seek higher education. In 2006, there were about 1.8 million
students in 2,655 elementary schools in Taiwan (Taiwan Ministry of Education, 2006).
Elementary school in Taiwan has different school dynamic from junior high school
and high school in philosophy, structure and mission (Taiwan Ministry of Education,
2006). In Taiwanese elementary schools, the paramount academic aims are literacy, basic
natural and social science, and arts. However, creating a cohesive caring classroom is

61
another goal. Each classroom has a homeroom teacher who is responsible for taking care
of students and teaching most of the classes during the day. In contrast to other school
types, elementary school students have a more nurturing and less competitive learning
environment, because they can enter into junior high school without any entrance exam.
High schools, on the other hand, require entrance exams making both junior high and
high schools highly competitive social environments. In the west, magnet schools and
private preparatory schools would have similar goals and structures as those in Taiwan.
Overall, the general goals and structures of elementary schools in Taiwan are very similar
to those of elementary schools in the West (i.e., the role of the teacher, the focus of the
classroom, peer relations, and daily structure are nearly identical).
Factors associated with school violence: East vs. West
Western studies suggest that elementary school students who endorse certain
negative personal traits (Bentley & Li, 1995; Connolly & O’Moore, 2003), who perceive
of low level of parental monitoring and involvement (Smith, Twemlow, & Hoover, 1999),
and who have been exposure to violence (Marie-Alsana et. al., 2006) are more likely to
involve in school violence. In addition, empirical evidence shows students’ relationships
with peers and school staff/teachers in elementary school settings have salient and
influential long-term effect on children’s behavior outcome, such as aggressive behavior
in schools (Anderson, Christenson, Sinclair, & Lehr, 2004; Huttunen, Salmivalli, &
Lagerspetz, 1996; Junger-Tas, 1999; Olweus, 1999). Low level of school engagement is
also an important factor contributing to anti-social behavior in children (Fraser, 1996)
and school violence (Natvig et al., 2001).

62
Similarly, few Asian studies show risk factors of school violence perpetration such
as negative personal traits, victimization, poor school engagement, involvement in at-risk
peers (Ando, Asakura & Simons-Morton, 2005; Hu & Lin, 2001), low level of parental
monitoring, and poor student-teacher relationships for student violence against students
(Hu & Lin, 2001; Ma, Shek, Cheung, & Tam, 2002; Yoneyama & Naito, 2003; Wong,
2004). In addition, some smaller scale studies in Taiwan show students’ negative personal
traits, parental involvement, victimization, risky behavior in school (Chen, 1999), school
climate, school engagement (Chen, 1989, 1999), and poor student-teacher relationship
(Huang & Hsien, 1987; Lin, 1986) are related to student violence against teachers.
However, most of these studies were analyzed, based on teachers’ reports and were
conducted using nonrepresentative samples from junior high schools, so it is not clear
how these findings in Western countries generalize to Asian cultures, particularly to
Taiwanese elementary schools .
Theoretical model for this study
Based on the above analysis, a model is developed for examining if school variables
mediate between personal traits, family experiences and school violence. Figure 1 shows
the theoretical model used in this study. Based on previous literature and theory, this
study proposes that the perpetration of school violence by students against peers and/or
teachers is directly associated with negative personal traits, parental monitoring,
victimization, low level of school engagement, risky peers, and poor student-teacher
relationships. Please note that the lines with plus and minus in Figure 1 represent the
hypothesis in this study. The plus signs indicate positive relationships between the two

63
variables and minus signs indicate negative. For example, the line with plus sign between
risky peers and student violence against students means this model hypothesizes students
who have risky peers are more likely to be involved in student violence against students.
The line with minus sign between parental monitoring and risky peers means that
students who have good parental monitoring are less likely to have risky peers.
Most importantly, different from prior studies, this model proposes that external
school variables (i.e., negative personal traits, parental monitoring, and victimization)
will have indirect effects on school violence mediated through school variables (i.e., low
level of school involvement, at-risk peers, and poor student-teacher relationship). For
example, this model proposes that good parental monitoring will indirectly influence
student violence against students by affecting risky peers. That is students who have good
parental monitoring are less likely to have risky peers, and in turn, their chance to involve
in school violence is decreased. Other hypotheses in the model are also indicated by
Figure 1.
Many international studies suggest boys engage in more aggressive behavior in
schools than girls (Rigby, 2005; Baldry, 2003; Nansel et al., 2001). However, based only
on these studies, it is not clear if the patterns of relationships between factors and school
violence differ across gender. Perhaps what causes perpetration behaviors in boys is the
same pattern that causes it in girls, only that girls have lower base rates of violence. Thus,
current study examines how gender differences impact the pattern of relationship between
factors and school violence committed by students against students and teachers. Current

64
study expects that the model fit male and female students, but certain paths within the
model may differ.
Method
The data used in this study was part of a large-scale project of “Prevention and
Control of School Violence in Taiwan” (Wu, Lee, Yin, & Hu, 2000). The survey was
conducted throughout Taiwan among over 14,000 students from elementary schools
(Grade 4 to 6), junior high schools (Grade 7 to 9), vocational high schools and academic
high schools (Grade 10 to 12). Students were given a structured questionnaire in
classrooms under the guidance of professionally trained survey monitors. Respondents
were assured of anonymity and were encouraged to respond truthfully. Participants were
free to withdraw from the study at any time and for any reasons. This study was
supported by Taiwan’s National Science Council (NSC). The questionnaires, procedures,
informed consent forms and other ethical concerns were reviewed and supervised by
NSC.
This sample was designed to represent all students from 4th to 12th grades in
Taiwan. The students’ response rate was over 98 percent. The probability sampling
method was a two-stage stratified cluster sample. The strata were
northern/central/southern/eastern, urban/rural, and elementary/junior/technical/academic.
In this study, only elementary school students, grades four to six, were selected for this
study. In first stage, schools were randomly selected from the sampling frame according
to those appropriate strata. In the next stage, two classes were randomly selected

65
according to each grade in selected schools. All students in that class were included in
sample.
This study examined 3,122 students from 16 elementary schools (grades 4 to 6).
51.6 percent of the students were boys, 47.8 percent were girls, and 0.6 percent did not
indicate gender; the grade level distribution was: 29.1 percent of the students were in 4th
grade, 35.7 percent in 5th grade, 35.2 percent in 6th grade, and 0.1 percent did not
indicate grade.
The questionnaire was developed on the basis of current school violence studies and
theories from both Taiwan and Western countries, and validated by Wu and colleagues
(2000). It included over 150 items in eight domains regarding students’ basic
demographic background and other information in their personal, family, and school
experience. Each of the student questionnaires took approximately 30 minutes to
complete. Before this survey was conducted, the Mandarin Chinese questionnaire was
adjusted and adapted based on two pilot studies conducted in the Tainan metropolitan
areas in Taiwan.
Measurements
In order to measure the latent variables in our model, several scales containing a
number of subscales were constructed. Table 6 represents the means and standard
deviations of the variables included in the model, broken down by genders. Table 7
represents the intercorrelations matrix among the variables. For the purpose of structural
equation modeling analyses, several subscales were constructed based on conceptual and
theoretical constructs. Each of these subscales creates the overall factor composites that

66
represents of the more general theoretical concepts discussed in literature review of this
study. Table 8 lists the domains, questions items, factor loadings, and alphas for the
theoretically created subscales.
Dependent Variables
Student violence against students. This domain involved asking students how many
times they perpetrated violent behavior in school against other students during the last
academic year. The variables in this domain were recorded into “never” and “at least
once”. This domain included three subscales on the basis of their content according to
type of violence. It included physical violence, vandalism, and verbal
violence/threat/harassment. The score of each subscale was the sum of the items included
in the subscale.
Student violence against teachers. This domain involved asking students to indicate
how many times they perpetrated violent behavior in school against teachers during the
last year. The variables in this domain were recorded as “never” and “at least once”. This
domain included three subscales on the basis of their content according to type of
violence. It included physical violence, verbal violence/threat, and emotional
violence/harassment. The score of each subscale was the sum of the items included in the
subscale.
Independent variables
Negative personal traits. This latent variable asked students about their personal
characteristics related to aggression. Three subscales were constructed on the basis of
their content. These subscales were students’ attitude towards violence, impulsive control,

67
and trait anger temperament. The rating for each item in this scale ranged from 1=
strongly disagree to 4= strongly agree. The score of each subscale was the sum of the
items included in the subscale.
Parental monitoring. This latent variable consisted of two subscales: father
monitoring and mother monitoring. Respondents were asked about parental monitoring of
them in daily life. The rating for each item ranged from 1= strongly disagree to 4=
strongly agree. Both subscales of father and mother monitoring consisted of five items,
and each subscale score was the sum of these five items.
Victimization. This latent variable asked students if they had experienced
victimization surrounding their life in the past year. The rating for each item ranged from
1=never to 4= almost every day. This latent variable consisted of two subscales on the
basis of their content. The two subscales were direct victimization and witness
victimization. The score of each subscale was the sum of the items included in the
subscale.
School engagement. This domain consisted of four items, and students were asked
questions in order to measure their level of school engagement. The ratings were
provided on a four-point scale and were coded from 1= strongly disagree to 4= strongly
agree.
Risky peers. In this domain, students were asked questions to determine the quality
of their group of friends. The ratings ranged from 1= strongly disagree to 4= strongly
agree. The score of each subscale was the sum of the items included in the subscale. This

68
domain was divided into two subscales based on preliminary factor analysis. The two
subscales were risky friendships and risky acts.
Poor student-teacher relationship. This domain consisted of five items about
whether the respondents experienced a poor relationship between teachers and
themselves. The ratings for each item were provided on a four-point scale from 1=
strongly disagree to 4= strongly agree.
Analytical plan
The primary analysis method in this study is latent variables structural equations
modeling (SEM) with maximum likelihood (ML) estimation using the AMOs program.
SEM is a statistical methodology that takes a confirmatory approach to the analysis of
structural theory describing relationships among endogenous factors (Bentler, 1988). The
structure/hypothesized model can be tested statistically in a simultaneous analysis of an
entire system of variables to determine the extent to which it is consistent with the data.
Confirmative factor analysis (CFAs) was first to be conducted to ensure the measurement
model as a good fit. Following the CFAs, the final SEM model including the full dataset
of all elementary school students was tested. Next, comparative analyses were conducted
in order to determine whether patterns of relationships and mediating effects are different
between male and female students. In this comparative analysis, all the factor loadings,
the paths, and the covariances were constrained to be equal simultaneously to the same
model across genders. Then, the model was tested by releasing path constraints one at a
time in order to find out if releasing equality constraints could significantly improve the
fit.

69
There are various indicators of the goodness of fit for a specific model. The
chi-square coefficient is used to assess the size of discrepancies between the relationships
in the original data matrix from those implied by the model. A low chi-square measure
reveals non-significant discrepancies, and means that the data “fit” the theoretical model.
However, due to the sensitivity of the chi-square coefficient to sample size, it is not a
preferred fit index for large samples such as those in this study. Indeed, researchers have
addressed the chi-square limitation by developing goodness-of-fit indices that take a
more pragmatic approach to the evaluation process. More commonly used fit indices
includes Bentler and Bonnett’s (1980) Normed Fit Index (NFI), Bollen’s (1989)
Incremental Fit Index (IFI), and Bentler’s (1990) Compared Fit Index (CFI). Typically,
these three fit indices consider a model to be a good fit when the value is above .90
(Bentler, 1992), and a superior fit when it is close to .95 (Hu & Bentler, 1999). A
common misfit measure, the Root Mean Square Error (RMSEA), is also reported in the
SEM analysis. The RMSEA considers a mediocre fit to range from .08 to .10 and a good
fit to be below .06 (Hu & Bentler, 1999).
Results
Overall model
The results of the analysis based on total sample provided a good fit to the data [ χ
2

(224, N=3,122)=1,600.31, p<.001, and with NFI=..94, IFI=.95, CFI=.95, and
RMSEA=.04]. This suggested that the model was a good one. Figure 4 presents the paths
in this model.

70
Figure 4 shows that negative personal traits, parental monitoring, victimization had
direct effect on student violence against students ( β= .33, β= -.12, and β= .13,
respectively). Negative personal traits, parental monitoring, and victimization also
affected student violence against students indirectly mediated through risky peers. Low
level of school engagement and poor student-teacher relationship had weaker mediation
effects for negative personal traits, parental monitoring, and victimization.
In contrast with student violence against students, negative personal traits, parental
monitoring, and victimization had almost no direct effect on student violence against
teachers ( β= -.03, β= -.04, and β=.02, respectively). However, these variables affected
student violence against teachers indirectly mediated through risky peers and poor
student-teacher relationship. Negative personal traits had greatest indirect effect on
violence against teachers through risky peers and poor student-teacher relationships. The
variable of risky peers had strongest mediation effect on student violence teachers.
Together, all of the variables in this model accounted for around one-third of the
explained variance for student violence against students (R
2
=.32) and around one-fifth for
student violence against teachers (R
2
=.21).
Gender analysis
The gender analysis of this study focused on the inquiry of whether the same
theoretical model was applicable to the samples of male and female students in
elementary schools. In this analysis, factor loadings, the paths, and the covariances were
constrained to be equal in order to fit the covariance matrices of the male and female
subgroups simultaneously to the same model. The analysis provided a good fit to the data

71
[ χ
2
(492, N: male=1,612, female=1,492)=2,041.18, p<.001 and with NFI= .93, IFI=.94,
CFI=.94, and RMSEA= .03]. This suggested the same theoretical model fit the data
from both genders well.
Next, the model was tested to find out if releasing equality constraints on the paths
could significantly improve the fit. After releasing path constraints one at a time, it was
determined that the release of the constraints between negative personal traits and low
level of school engagement as well as between parental monitoring and student violence
against student yield a significantly better fit to the model; thus, the results for the final
model with the two constraints released together were [ χ
2
(490, N: male=1,612,
female=1,492)= 2,021.74, p<.001 and with NFI= .93, IFI=.94, CFI=.94, and
RMSEA= .03]. This suggested that the data from the male and female students fit the
same model quite well and there were some important and significant differences in
several paths. Figure 5 represents the results of the gender analysis.
The beta coefficients for both male and female models exhibited remarkably similar.
However, there was a significant difference in the relationship between parental
monitoring and student violence against students. Parental monitoring had stronger
magnitude of effect on student violence against student for males ( β=-.16) compared with
females ( β=-.06). In addition, the effect of negative personal traits on low level of school
engagement was stronger for males than females ( β=.44 for boys vs. β=.35 for girls).
Discussion
Using a nationally representative sample in Taiwan, this study examines how school
engagement, school risky peers and poor student-teacher relationships mediate the effects

72
from students’ personal traits and family experience on student violence against students
and teachers in Asian elementary school settings. This study proposed that students’
negative personal traits, parental monitoring and victimization have indirect effects on
school violence mediated through low levels of school engagement, risky peers and poor
student-teacher relationships. This study expected that the model fit male and female
students, but certain paths within the model may differ. The findings of this study provide
support for the proposed theoretical model.
Results of this study show that elementary school students who possess negative
personal traits, perceive low level of parental monitoring, and experienced victimization
are more likely to involve in violence against students. In addition, students who possess
negative personal traits, who perceive low level of parental monitoring, and who
experience victimization are also more likely to involve with at-risk peers, and
consequently, their chance to be violent against students will be increased. By contrast,
students’ negative personal traits, parental monitoring, and victimization have limit direct
effects on student violence against teachers, but these variables indirectly affect student
violence against teachers through risky peers and poor student-teacher relationships.
These findings support the theoretical model that school violence is influenced by a
combination of numerous within-school and external-school variables, and within-school
variables can mediate contributions from external contexts (Benbenishty & Astor, 2005).
These findings provide empirical evidence to support the school environment as a unique
developmental setting to school violence compared with other normative environments
such as home (Astor et al., 2002; Benbenishty et al., 2002).

73
The results show that the variable of negative personal traits has the greatest effects,
both direct and indirect, on student violence against students. This variable also has
strongest indirect effects on student violence against teachers. Results also show that
involvement with at-risk peers has strongest mediating effect on both types of school
violence. The variable of poor student-teacher relationship has strong mediating effect on
violence against teachers. Surprisingly, school engagement has very slight or zero
mediation effects on student violence against students and teachers. This is unexpected
because previous studies suggest that school engagement should have strong influence on
perpetration of school violence (Natvig et al., 2001). Overall, these findings suggest that
intervention focusing on enhancing the quality of students’ peers and their interaction
with teachers as well as improving students’ negative personal traits may lead to a greater
reduction of school violence, because the findings show that personal traits, risky peers
and student-teacher relationships are powerful factors affecting students’ perpetration of
school violence.
The results show that the overall theoretical model is applicable across both genders.
However, it is interesting to note that the influence of negative personal traits on low
levels of school engagement is stronger for male students than female. This means that
male students who have negative personal traits are more likely than female to have low
level of school engagement. In addition, although parental monitoring is an important
factor affecting student violence against students, this impact is stronger for male students
than female. The finding implies that intervention aiming at increasing parental

74
monitoring would decrease students’ violence against other students and would be more
likely to affect males than females.
Overall, the findings suggest that most of the risk factors of school violence
perpetration reported in studies from Western countries are relevant to Asian culture
context, especially in Taiwanese elementary school. This finding may imply that
structural patterns of relationships between risk factors and school violence could share
universal patterns in the way they relate to each other in seemingly diverse cultures
(Benbenishty & Astor, 2005).
It is important to acknowledge that there are several limitations in this study. First,
the investigation in this study is cross-sectional. The results of this study do not enable
one to establish casual-effect relationships. Second, the data is only based on student
self-reporting, and school violence was measured in terms of behavior occurring during
the prior twelve months. This may lead students to under- or over- report violent behavior
because some of these events are so common that students might not recall the entire
year’s events (Benbenishty & Astor, 2005). Future studies should collect information
from multiple sources, such as teachers’ reports, parents, or peer nomination.
Implications
This study shows that within school factors mediate the relationship between
external school factors and school violence. This finding implies that improving student
school experience could show a strong effective response to violence. However, this does
not mean that interventions should solely focus on school level. To maximize
effectiveness of school violence intervention programs, practitioners must integrate

75
personal, family, and school level approaches (Ahmed & Braithwaite, 2004; Astor, Meyer,
Benbenishty, Marachi, & Rosemond, 2005), because the results of this study show that
the influence of the personal, family and school variables in this model account for a
relatively large amount of the explained variance for student violence against students
and teachers.
In addition, the overall patterns of relationships between the variables are quite
similar for male and female students. This implies school violence intervention
incorporated at a national level could be affective across genders in elementary school.
Finally, this study implies the risk factors of school violence perpetration are similar
between East and West. This empirical evidence supports school policy makers or
practitioners in developing new international-based school violence prevention and
intervention.

76
Table 6. Means and Standard Deviations of Subscales by Overall Students and by Gender
Overall Male Female Scale/subscale
Mean SD Mean SD Mean SD
Violence against students
a

Physical
Vandalism
Verbal/threat/harassment
1.65
0.52
0.27
0.86
2.01
0.82
0.56
1.05
2.02
0.69
0.34
1.00
2.18
0.91
0.61
1.11
1.26
0.34
0.21
0.71
1.71
0.68
0.49
0.95
Violence against teachers
a

Physical
Verbal/threat
Emotional/harassment
0.23
0.02
0.04
0.16
0.62
0.18
0.22
0.43
0.26
0.03
0.05
0.18
0.66
0.19
0.24
0.44
0.19
0.02
0.03
0.15
0.58
0.17
0.19
0.41
Negative personal traits
b

Attitude toward violence
Impulsive control
Trait anger temperament
54.09
18.36
12.99
22.77
15.02
6.38
3.85
6.79
56.15
19.55
13.37
23.22
15.34
6.68
3.87
6.85
51.92
17.11
12.67
22.28
14.36
5.80
3.79
6.70
Parental monitoring
b

Father monitoring
Mother monitoring
32.33
15.94
16.36
5.31
2.89
2.70
31.75
15.66
16.05
5.47
2.99
2.78
32.95
16.24
16.70
5.09
2.77
2.57
Victimization
c

Direct victimization
Witness victimization
2.67
1.53
1.17
3.84
2.23
2.03
3.01
1.77
1.28
4.01
2.36
2.10
2.31
1.29
1.05
3.63
2.06
1.94
Low school engagement
b

Doze off in class or skip class
Forget bringing material
Seldom turn in homework
Bring prohibited material
6.30
1.24
1.96
1.62
1.50
2.24
0.57
0.91
0.83
0.82
6.68
1.27
2.05
1.74
1.66
2.40
0.61
0.94
0.88
0.91
5.90
1.20
1.87
1.50
1.34
1.99
0.52
0.87
0.74
0.67
Risky peers
b

Risky friendship
Risky acts
9.15
6.13
3.05
3.36
2.41
1.88
9.35
6.22
3.16
3.53
2.52
1.89
8.92
6.02
2.94
3.15
2.27
1.86
Poor student-teacher relationship
b

Punish for no reason
Teachers do not like me
Teachers mock at me
Teachers never trust me
Observe me and snitch
7.20
1.47
1.45
1.38
1.55
1.39
3.02
0.75
0.72
0.67
0.83
0.70
7.56
1.53
1.49
1.42
1.66
1.48
3.16
0.79
0.74
0.71
0.89
0.77
6.81
1.40
1.40
1.33
1.43
1.29
2.80
0.69
0.70
0.63
0.75
0.61
Note.
a
On a scale: 0= never, 1= at least once in the past year.
b
On a scale: 1= strongly
disagree to 4= strongly agree.
c
On a scale: 1=never, 2= sometimes, 3= often, and 4=
almost every day.

77
Table 7. Matrix of Correlation among School Violence Subscales for Elementary School
A1A2A3B1B2B3C1C2C3D1D2E1 E2F1F2F3F4G1G2H1H2H3H4H5
A1 S-S Physical
A2 S-S Vandalism
A3 S-S Verbal/threat
B1 S-T Physical
B2 S-T Verbal/threat
B3 S-T Emotional
C1 Violence attitude
C2 Impulsive control
C3 Trait anger temperament
D1 Father monitoring
D2 Mother monitoring
E1 Direct victimization
E2 Witness victimization
F1 Dose off
F2 Forget bringing material
F3 Seldom turn in homework
F4 Bring prohibited material
G1 Risky friendship
G2 Risky acts
H1 Punish for no reason
H2 Teachers do not like me
H3 Teachers mock at me
H4 Teachers never trust me
H5 observe me and snitch
-- .46
--
.53
.52
--
.18
.21
.21
--
.22
.23
.26
.42
--
.31
.33
.39
.17
.34
--
.37
.26
.29
.07
.10
.20
--
.32
.27
.32
.05
.09
.21
.57
--
.32
.25
.33
.06
.08
.20
.63
.76
--
-.19
-.18
-.16
-.04
-.06
-.11
-.23
-.24
-.19
--
-.20
-.19
-.17
-.06
-.07
-.11
-.24
-.23
-.16
.83
--
.16
.14
.18
.01
.04
.10
.14
.14
.12
-.08
-.09
--
.13
.11
.12
.00
.03
.08
.09
.11
.10
-.03
-.03
.65
--
.15
.14
.11
.13
.14
.06
.20
.19
.15
-.13
-.16
.10
.05
--
.17
.13
.16
.06
.06
.11
.21
.29
.27
-.18
-.19
.05
.03
.20
--
.19
.13
.15
.12
.11
.09
.24
.30
.24
-.22
-.22
.10
.06
.30
.52
--
.25
.20
.20
.13
.13
.14
.31
.29
.25
-.22
-.24
.13
.11
.33
.31
.41
--
.26
.20
.20
.15
.16
.17
.37
.33
.31
-.17
-.17
.13
.12
.29
.18
.26
.31
--
.12
.07
.07
.07
.09
.09
.18
.14
.12
-.06
-.08
.06
.04
.19
.07
.13
.16
.23
--
.19
.19
.20
.14
.16
.25
.31
.28
.28
-.17
-.18
.09
.04
.28
.21
.26
.30
.29
.16
--
.20
.18
.19
.12
.18
.25
.29
.30
.29
-.17
-.16
.09
.04
.27
.19
.27
.26
.29
.16
.72
--
.16
.13
.16
.10
.16
.18
.29
.26
.27
-.15
-.14
.08
.05
.31
.20
.26
.28
.27
.17
.67
.68
--
.28
.17
.23
.13
.16
.25
.29
.31
.29
-.17
-.16
.10
.07
.26
.20
.30
.26
.26
.15
.60
.63
.58
--
.17
.14
.16
.12
.18
.17
.28
.26
.23
-.15
-.15
.09
.04
.30
.19
.29
.31
.29
.18
.54
.52
.58
.53
--

78
Table 8. Theoretical Domains, Subscale, and Items for Elementary School
Domain
(Alpha’s)
Subscales
(Loadings)
Items
Violence against
students
(Alpha= .77)
Physical
(loading=.70)

Students beat or kick other students (by group) in order to hurt him/her/them.
Students beat or kick other students (by individual) in order to hurt him/her/them.
Students use dangerous objects or instruments in order to harm students.
Vandalism
(loading=.67)

Students intentionally destroy or break school public belongings.
Students intentionally destroy or break other students’ belongings.
Verbal/threat
(loading=.76)

Students verbally threaten or intimidate other students.
Students curse or insult other students.
Students threaten or blackmail other students.
Students tease, mock, or play physically harmful tricks on other students on purpose.
Students threaten or force others to buy things.
Violence against
teachers
(Alpha= .55)
Physical
(loading= .48)

Students beat or kick teacher(s).
Students use dangerous objects or instruments to harm teacher(s).
Verbal/threat
(loading=.63)

Students curse or insult teacher(s).
Students threaten or intimidate teacher(s).
Students extort or blackmail teacher(s).
Emotional
(loading= .57)

Students tease, mock, or play physically harmful tricks on teacher(s).
Students oppose teacher(s) in order to cause them psychological harm.
(Table 8 continues)

79
Table 8 continued
Domain Subscales Items
Negative personal
traits
(Alpha=. 93)

Attitude toward
violence
(loading= .71)

Impulsive control
(loading= .88)

Trait anger
temperament
(loading= .85)
If someone insulted me or my family, beating him/her will make me feel better.
If someone disrespects me, I will beat him/her to regain honor
I will feel unhappy if I do not beat someone who cursed me.
A coward is a person who never retaliates when he/she is insulted.
If someone impedes my plans, he/she will pay for it.
If someone tries to hurt me, I will take vengeance on him/her.
Violence is the best way to resolve any problem.
Violence is justice.
If someone makes me unhappy, beating him is what he/she deserves.
Violence is better than negotiation.

I often make mistakes due to uncontrolled anger.
I often feel regretful about the things I do.
When I feel angry, no one can control me.
I have some bad habit that I always fail to change.
I can not help violating school rules.
I can not tolerate when others look down on me.

I loose my temper easily.
I am an irritable person.
I am easily agitated.
I feel anger if someone’s mistakes disturb my work.
I often can not control my anger.
When I feel angry, I will curse or use dirty words, such as fuck.
I will feel angry if someone criticizes me.
Beating others will make me feel better if I feel unhappy.
I feel upset if the work is done worse than I expect.
I feel angry if my work is criticized.
(Table 8 continues)

80
Table 8 continued
Domain Subscales Items
Parent monitoring
(Alpha=.84)

Father
(loading= .92)

My father (or father figure) knows my friends.
My father (or father figure) knows my conduct.
If I am not at home, my father (or father figure) will know where I am.
My father knows my schedules.
My father (or father figure) often takes activities with me.
Mother
(loading= .90)

My mother (or mother figure) knows my friends.
My mother (or mother figure) knows my personality and general conduct.
If I am not at home, my mother (or mother figure) knows where I am.
My mother knows my schedule.
My mother (or mother figure) often does activities with me.
Victimization
(Alpha= .93)

Direct
(loading=.96)

Witness
(loading= .68)

Have you been beaten or kicked by others?
Has someone hit and hurt you with an object of any kind?
Have you been beaten or kicked by groups of people after class or school?
Have you been blackmailed by others?
Have you been verbally threatened by others?
Have you been intimidated by others?
Has someone threatened or intimidated you by passing a slip of paper?

Have you seen anyone verbally threatened by others?
Have you seen anyone threatened by weapons?
Have you seen anyone destroy public belongings on purpose?
Have you seen anyone beaten by others or participating in group fights?
Have you seen anyone robbed by others?
Have you seen anyone insulted by others?
Have you seen anyone blackmailed by others?
(Table 8 continues)

81
Table 8 continued
Domain Subscales Items
Low school
engagement
(Alpha= .68)

Doze off/skip
(loading=.48)

Forget bringing
(loading=.59)

Homework
(loading= .71)

Prohibited material
(loading=.62)

I often doze off in class or skip class.

I often forget to bring required materials (textbooks/homework) to school.

I seldom turn in homework.

I often bring prohibited materials to school.
Risky peers
(Alpha= .70)

Risky friendship
(loading= .70)

I have friends who are school gang members.
I have friends who dropped out due to problem behavior in school.
When I have conflicts with others, my friends will help me beat them or fight with them.
My friends are always on my side no matter what bad things I do.
Risky acts
(loading=.38)

My friends and I often stay up all nights to drink, to gamble, or to do illegal activities.
My friends and I are often involved in fights or use our fists to protect our territory.

Poor
student-teacher
relationship
(Alpha= .88)
Punish (loading=.83)

Dislike (loading=.84)

Mock (loading=.80)

Distrust(loading=.74)

Snitch (loading=.67)
My teacher often punishes me for no reason.

I feel that my teacher does not like me.

My teachers like to mock me.

Even though I tell the truth, my teacher still distrusts me.

Teachers often ask students to observe on what I am doing and snitch on me.

82
Figure 4. Results Based on Total Sample of Elementary School Students

83
Figure 5. Gender Comparison for Elementary School Students (Male Students in Bold and Oblique)

84
Chapter Five:
Study 4: School Variables as Mediators of Personal and Family Factors
on School Violence in Junior High Schools
Introduction
During the past decade, there has been a dramatic increase in the number of
publications on risk factors for school violence in Western countries (Bosworth, Espelage
& Simon, 1999; Connolly & O’Moore, 2003). By contrast, large-scale representative data
on risk factors in Asia remain limited. There is almost no evidence that the theoretical
assumptions from school violence studies in Western countries are relevant to Asian
cultures, although significant differences exist between the East and West regarding
beliefs and perceptions (Nisbett, 2003).
Furthermore, many studies have examined how students’ personal, family, and
school experiences “risk factors” directly affect student violence against their school
peers (e.g., Baldry, 2003). Few studies have focused on how risk factors contribute to
school violence against teachers committed by students (Khoury-Kassabri, Astor &
Benbenishty, 2008). Even less is known about organizational or school variables that
could be mediators between personal traits, victimization, family factors, and school
violence. There is a paucity of empirical evidence that investigates how gender influences
patterns of relationships between risk factors and school violence. Using the first
nationally representative data from an Asian culture, this study expands the literature on
school violence by exploring the influence of factors reported in Western empirical
journals in Asian cultures. This inquiry will examine if school factors mediate between

85
personal traits and family factors, and school violence perpetration committed by students
against peers and teachers in Taiwan.
In many ways, Taiwan is representative of Asian cultures. It is a society where
long-standing cultural values and outside influences must be considered simultaneously
when examining the relationships between risk factors and school violence. The nation
has experienced democratization, rapid industrialization and urbanization, a growing
population of immigrants from other Asian countries, and an expanding economy over
the past four decades. These factors obviously influence social values and norms. Yet, the
Taiwanese are still partially guided by values stemming from a blend of Confucianism,
Taoism, and Buddhism (Nisbett, 2003). While school violence issues have been a major
public concerns for many years, for political reasons (China’s lack of recognition of
Taiwan as an independent country) Taiwan has never been part of any large-scale
cross-national studies on these topics (Akiba, LeTendre, Baker, & Goesling, 2002). Yet
from an empirical perspective, Taiwan represents a modern, industrial, capitalistic, and
pro Western democracy that is currently and historically part of Chinese culture.
School violence in this study refers to student perpetration intending to harm other
students and teachers or to cause damage to belongings or school property. The definition
includes physical and verbal violence, psychological harm, threatening behavior, and
property damage. (Astor, Benbenishty, Pitner, & Zeira, 2004; see Benbenishty & Astor,
2005 for a critical discussion).
Factors associated with school violence: East and West
Western studies on school violence clearly indicate that students’ negative personal

86
traits (i.e., aggressive attitudes, poor anger traits, and lack of impulsive control--Byrne,
1994; Connolly & O’Moore, 2003; McConville & Cornell, 2003), low parental
monitoring and involvement (Baldry & Farrington, 2000; Rigby, 1993, 1994), prior
victimization (Baldry, 2003; Flannery, Wester, & Singer, 2004), poor school engagement
(Haynie et al., 2001; Nansel et al., 2001; Natvig, Albrektsen, & Qvarnstrom, 2001; Slee,
1995; Slee & Rigby, 1993), risky peers (Huttunen, Salmivalli, & Lagerspetz, 1996), and
poor interaction with teachers (Junger-Tas, 1999; Olweus, 1999) are strongly associated
with school violence.
A few non-representative studies in Taiwan or other Asian countries have shown that
student violence against students in junior high school was related to students’ negative
personal traits, parental monitoring and victimization (Chen, 1999; Hu & Lin, 2001),
poor school engagement (Ando, Asakura, Simons-Morton, 2005; Hu & Lin, 2001), risky
peers (Ando et al., 2005; Hu & Lin, 2001), and poor interaction with teachers (Hu & Lin,
2001; Ma, Shek, Cheung, & Tam, 2002; Yoneyama & Naito, 2003; Wong, 2004). In
addition, some smaller scale studies in Taiwan have shown that negative personal traits,
parental involvement, victimization and risky behavior in school (Chen, 1999), school
climate and school engagement (Chen, 1989, 1999), and poor student-teacher
relationships (Huang & Hsien, 1987; Lin, 1986) were related to student violence against
teachers in junior high schools. Because these studies are convenience samples, it is not
clear how generalizable these findings are for Asian culture contexts such as Taiwan.
Furthermore, a review of the literature indicates that most previous studies on school
violence have examined, only in separate studies, how students’ experiences in each

87
ecological context independently affect perpetration of school violence. How multiple
ecological risk factors and nested contexts influence student violence against students and
teachers has not yet been explored much in either the West or the East.
School variables as mediators
Benbenishty and Astor (2005) recently addressed the lack of theoretical integration
surrounding the contexts of school violence. They propose a theoretical model of school
violence that is influenced by numerous within-school variables (e.g., teacher-child
relationships) and other external variables (e.g., personal and family characteristics). In
their heuristic model, schools mediate, moderate, and attenuate contributions from
external contexts. School subcontexts also mediate and generate their own direct
contributions to different forms of violence. In contrast with earlier psychological models
that center exclusively around the individual, Benbenishty and Astor stress that when
looking at “school violence” the school itself as a social context should be the center of
the theoretical model.
This model has been examined in case of school victimization across different
cultures with Jewish Secular, Jewish Orthodox, Arabic, and American students (Astor,
Benbenishty, Vinokur & Zeira, 2006; Astor, Benbenishty, Zeira, & Vinokur, 2002;
Benbenishty & Astor, 2005; Benbenishty, Astor, Zeira, & Vinokur, 2002). However,
little research has been published with this theoretical orientation on perpetration of
school violence (for an example of one see Khoury-Kassabri et al., 2008). This is the first
study that examines how these complex patterns between social contexts affect
perpetration of school violence in Asian junior high schools. Empirical evidence that

88
establish the degree to which school variables mediate or contribute to perpetration in
Asian junior high schools is sorely needed.
Background of junior high school in Taiwan
Generally, there are three major systems of basic and intermediate education in
Taiwan: elementary school (grades 1 to 6), junior high school (grades 7 to 9), and high
school (grades 10 to 12). Attending a nine-year compulsory education program from
elementary to junior high school is mandatory. A competitive joint high school entrance
exam must be taken if junior high school graduates seek higher education. In 2006, there
were about 1.0 million students in 732 junior high schools in Taiwan (Taiwan Ministry of
Education, 2006).
Overall, the educational purpose and school structure in junior high school appear to
differ from other types of schools. In general, the main goal in junior high school is to
prepare the student for a competitive joint high school entrance exam in order to enter the
best high school and, subsequently, college. Thus, academic achievement is highly
emphasized in junior high school. In this respect, junior high schools in Taiwan function
more similarly to private preparatory schools in the West. While Western students often
have multiple teachers with different students in each class, Taiwanese students remain in
the same class with the same cohort of students for three years, and a homeroom teacher
is assigned to each class. The main responsibility of the homeroom teacher is to help
students achieve higher academic performance. Teachers are perceived as intellectual
experts, and each subject is taught by one specialized teacher in that field. Thus, teachers
in academic fields rotate from class to class. To achieve the highest status, most junior

89
high schools encourage teachers to design innovative programs or conduct mock tests to
help students succeed. Outstanding teachers are recognized publicly if their class has a
high admissions rate for academic high schools.
Theoretical model for current study
Based on the above analysis, a model is developed for examining if school variables
mediate between personal traits, family experiences and school violence. Figure 1 shows
the theoretical model used in this study. Based on previous literature and theory, this
study proposes that the perpetration of school violence by students against peers and/or
teachers is directly associated with negative personal traits, parental monitoring,
victimization, low level of school engagement, risky peers, and poor student-teacher
relationships. Please note that the lines with plus and minus signs in Figure 1 represent
the hypothesis in this study. The plus signs indicate positive relationships between the
two variables based on prior empirical and theoretical formulations and minus signs
indicate negative relationships between those variables in the research literature. For
example, the line with plus sign between poor student-teacher relationships and student
violence against teachers means this model hypothesizes students who have poor
interaction with teachers are more likely to involve in student violence against teachers.
The line with minus sign between parental monitoring and poor student-teacher
relationships would predict that students who have good parental monitoring are less
likely to have poor interaction with teachers.
Most importantly, in contrast with prior theoretical conceptions of school violence,
this model proposes that variables external to the school (i.e., negative personal traits,

90
parental monitoring, and prior victimization experiences) will have indirect effects on
school violence perpetration mediated through school social environment variables (i.e.,
low level of school involvement, at-risk peers, and poor student-teacher relationship). For
example, this model proposes that good parental monitoring will indirectly influence
student perpetration of violence against teachers by affecting poor student-teacher
relationships. That is, this model predicts that students who have good parental
monitoring will be less likely to have poor interaction with teachers, and in turn, their
chance to involve in student violence against teachers is decreased. Other hypotheses in
the model are also indicated by Figure 1.
Method
The data used in this study was part of a large-scale project of “Prevention and
Control of School Violence in Taiwan” (Wu, Lee, Yin, & Hu, 2000). The survey was
conducted throughout Taiwan among over 14,000 students from elementary schools
(grades 4 to 6), junior high schools (grades 7 to 9), vocational high schools and academic
high schools (grades 10 to 12). Students were given a structured questionnaire in
classrooms under the guidance of professionally trained survey monitors. Respondents
were assured of anonymity and were encouraged to respond truthfully. Participants were
free to withdraw from the study at any time and for any reasons. This study was
supported by Taiwan’s National Science Council (NSC). The questionnaires, procedures,
informed consent forms and other ethical concerns were reviewed and supervised by
NSC.

91
This sample was designed to represent all students from 4th to 12th grades in
Taiwan. The students’ response rate was over 98 percent. The probability sampling
method was a two-stage stratified cluster sample. The strata were
northern/central/southern/eastern, urban/rural, and elementary/junior/technical/academic.
In this article, only junior high school students, grades seven to nine, were selected for
this study. In first stage, schools were randomly selected from the sampling frame
according to those appropriate strata. In the next stage, two classes were randomly
selected according to each grade in selected schools. All students in that class were
included in sample.
This study examined 3,058 students from 16 junior high schools (grades 7 to 9).
48.2 percent of the students were boys, 49.5 percent were girls, and 2.3 percent did not
indicate gender; the grade level distribution was: 30.8 percent of the students were in 7th
grade, 33.9 percent in 8th grade, 35.0 percent in 9th grade, and 0.2 percent did not
indicate grade.
The questionnaire was developed on the basis of current school violence studies and
theories from both Taiwan and Western countries, and validated by Wu and colleague
(2000). It included over 150 items in eight domains regarding students’ basic
demographic background and other information in their personal, family, and school
experience. Each of the student questionnaires took approximately 30 minutes to
complete. Before this survey was conducted, the Mandarin Chinese questionnaire was
adjusted and adapted based on two pilot studies conducted in the Tainan metropolitan
areas in Taiwan.

92
Measurement
In order to measure the latent variables in our model, several scales containing a
number of subscales were constructed. Table 9 represents the means and standard
deviations of the variables included in the model, broken down by genders. Table 10
represents the intercorrelations matrix among the variables. For the purpose of structural
equation modeling analyses, several subscales were constructed based on conceptual and
theoretical constructs. Table 11 lists the domains, questions items, factor loadings, and
alphas for the theoretically created subscales. Each of these subscales creates the overall
factor composites that represents of the more general theoretical concepts discussed in
literature review of this study.
Dependent Variables
Student violence against students. This domain involves asking students how many
times they perpetrated violent behavior in school against other students during the last
academic year. The variables in this domain were recorded into “never” and “at least
once”. This domain included three subscales on the basis of their content according to
type of violence. It included physical violence, vandalism, and verbal
violence/threat/harassment. The score of each subscale was the sum of the items included
in the subscale.
Student violence against teachers. This domain involved asking students to indicate
how many times they perpetrated violent behavior in school against teachers during the
last year. The variables in this domain were recorded as “never” and “at least once”. This
domain included three subscales on the basis of their content according to type of

93
violence. It included physical violence, verbal violence/threat, and emotional
violence/harassment. The score of each subscale was the sum of the items included in the
subscale.
Independent variables
Negative personal traits. This latent variable asked students about their personal
characteristics related to aggression. Three subscales were constructed on the basis of
their content. These subscales were students’ attitude towards violence, impulsive control,
and trait anger temperament. The rating for each item in this scale ranged from 1=
strongly disagree to 4= strongly agree. The score of each subscale was the sum of the
items included in the subscale.
Parental monitoring. This latent variable consisted of two subscales: father
monitoring and mother monitoring. Respondents were asked about parental monitoring of
them in daily life. The rating for each item ranged from 1= strongly disagree to 4=
strongly agree. Both subscales of father and mother monitoring consisted of five items,
and each subscale score was the sum of these five items.
Victimization. This latent variable asked students if they had experienced
victimization surrounding their life in the past year. The rating for each item ranged from
1=never to 4= almost every day. This latent variable consisted of two subscales on the
basis of their content. The two subscales were direct victimization and witness
victimization. The score of each subscale was the sum of the items included in the
subscale.

94
School engagement. This domain consisted of four items, and students were asked
questions in order to measure their level of school engagement. The ratings were
provided on a four-point scale and were coded from 1= strongly disagree to 4= strongly
agree.
Risky peers. In this domain, students were asked questions to determine the quality
of their group of friends. The ratings ranged from 1= strongly disagree to 4= strongly
agree. The score of each subscale was the sum of the items included in the subscale. This
domain was divided into two subscales based on preliminary factor analysis. The two
subscales were risky friendships and risky acts.
Poor student-teacher relationship. This domain consisted of five items about
whether the respondents experienced a poor relationship between teachers and
themselves. The ratings for each item were provided on a four-point scale from 1=
strongly disagree to 4= strongly agree.
Analytical plan
The primary analysis method in this study is latent variables structural equations
modeling (SEM) with maximum likelihood estimation using AMOs program.
Confirmative factor analysis (CFAs) was first to be conducted to ensure the measurement
model as a good fit. Following the CFAs, the final SEM model including the full dataset
of all junior high school students was tested. In addition, comparative analyses were
conducted in order to determine whether patterns of relationships and mediating effects
are different between male and female students.

95
The chi-square coefficient has often been used to assess if the data fit the theoretical
model. However, due to the sensitivity of the chi-square coefficient to sample size, it is
not a preferred fit index for large samples such as those in this study. Indeed, researchers
have addressed the chi-square limitation by developing goodness-of-fit indices that take a
more pragmatic approach to the evaluation process. More commonly used fit indices
includes Bentler and Bonnett’s (1980) Normed Fit Index (NFI), Bollen’s (1989)
Incremental Fit Index (IFI), and Bentler’s (1990) Compared Fit Index (CFI). Typically,
these three fit indices consider a model to be a good fit when the value is above .90
(Bentler, 1992), and a superior fit when it is close to .95 (Hu & Bentler, 1999). A
common misfit measure, the Root Mean Square Error (RMSEA), is also reported in the
SEM analysis. The RMSEA considers a mediocre fit to range from .08 to .10 and a good
fit to be below .06 (Hu & Bentler, 1999).
Results
Overall model
The results of the analysis based on total sample provides a good fit to the data [ χ
2

(224, N=3,058)= 1827.31, p<.001, and with NFI= .94, IFI=.95, CFI=.95, and
RMSEA= .05]. This suggested that the model is a good one. Figure 6 presents the paths
in this model.
Figure 6 shows that the variable of risky peers has stronger mediating effects than
student-teacher relationships and school engagement on student violence against students
for negative personal traits and victimization. In addition, the variable of risky peers is
the strongest mediator between negative personal traits and victimization and student

96
violence against teachers. Poor student-teacher relationship has a small but important
mediating effects on violence against teachers for negative personal traits, parental
monitoring, and victimization. Together, all of these variables accounted for around
one-half of the explained variance for student violence against students (R
2
=.49) and
around one-third for student violence against teachers (R
2
=.32).
Gender analysis
The gender analysis of this study focused on the inquiry of whether the same
theoretical model was applicable to the samples of male and female students. In this
analysis, factor loadings, the paths, and the covariances are constrained to be equal in
order to fit the covariance matrices of the male and female subgroups simultaneously to
the same model. The analysis provided a good fit to the data [ χ
2
(492, N: male=1,475,
female=1,514)= 2,303.46, p<.001 and with NFI= .92, IFI=.93, CFI=.93, and
RMSEA= .04]. Thus, the same theoretical model fit the data from both genders well.
Next, the model was tested to find out if releasing equality constraints on the paths
could significantly improve the fit. After releasing path constraints one at a time, it was
determined that the release of the constraints between negative personal traits and risky
peers, between victimization and low level of school engagement, and between
victimization and student violence against students yield a significantly better fit to the
model; thus, the results for the final model with the three constraints released together
were [ χ
2
(489, N: male=1,475, female=1,514)= 2259.95, p<.001 and with NFI= .92,
IFI=.94, CFI=.94, and RMSEA= .04].

97
Figure 7 represents the results of the gender analysis. The beta coefficients for both
male and female models exhibit remarkably similar. However, victimization has stronger
direct effect on student violence against students for female students than males ( β=.21
for boys vs. β=.29 for girls). By contrast, the strength of the relationship between
victimization and low level of school engagement ( β=.10 for boys vs. β=.02 for girls) as
well as relationship between negative personal traits and risky peers ( β=.57 for boys vs.
β=.46 for girls) are stronger for male students. Among all factors, the factor of risky peers
is the best predictor of both types of school violence for male and female students.
Overall, the theoretical model explains a similar proportion of the variance of student
violence against students (44% for boys and 49% for girls) and student violence against
teachers (31% for boys and 33% for girls) for both genders.
Discussion
Based on the nationally representative sample in Taiwan, this is the first study
conducted using Benbenishty and Astor’s (2005) theoretical orientation to examine how
the complex patterns of factors within and outside of the school affect school violence
perpetration committed by students against teachers and students in Asian junior high
schools. This study proposed that within school variables would mediate the effects from
external school variables on school violence. This study also expected that the model fit
male and female students, but certain paths within the model may differ.
The results of this study show that the variable of risky peers mediates the effects
from negative personal traits and victimization on student violence against students. In
addition, risky peers and poor student-teacher relationships mediate the effects from

98
negative personal traits, parental monitoring and victimization on violence against
teachers. These findings support Benbenishty and Astor’s (2005) theoretical model that
school violence is influenced by a combination of numerous within-school and
external-school variables, and within-school variables can mediate contributions from
external contexts.
Furthermore, these mediation findings provide empirical evidence to support the
school environment as a unique developmental setting to school violence compared with
other normative environments such as home (Astor et al., 2002; Benbenishty et al., 2002).
However, this does not mean that interventions should solely focus on school level. To
maximize effectiveness of school violence intervention programs, interventions must
integrate personal, family, and school level approaches (Ahmed & Braithwaite, 2004),
because the results of this study show that the influence of the personal, family and
school variables in this model account for a relatively large amount of the explained
variance for student violence against students and student violence against teachers. The
findings suggest that interventions for junior high school students in Taiwan should be
targeted toward decrease in negative personal traits and victimization as well as increase
in the quality of peers and student-teacher relationships. Most importantly, interventions
focusing on decreasing students’ involvement in at-risk peers may lead to a greater
reduction of student violence against students and teachers, because the results show that
the variable of risky peers has strongest mediating effect on student violence against
students and teachers.

99
The results show that the overall theoretical model is applicable across both genders.
This finding suggests that although male students have higher rates of perpetration in
school violence than females, how male and female students’ personal, family, and school
experiences affect student violence against students and teachers are quite similar. This
finding also indicates that school violence interventions or policy incorporated at a
national level could be affective across genders. However, it is interesting to note that
students’ negative personal traits have stronger influence on risky peers for males than
females. This suggests the variable of negative personal traits has stronger indirect effects
for male students on student violence against students and teachers mediated through
risky peers. The findings imply that interventions aimed at improving students’ negative
personal traits would reduce student violence against students and teachers and would be
more likely to affect male students. In addition, female prior victimization has a stronger
impact on student violence against students. The finding suggests that interventions
aimed at decreasing victimization would reduce student violence against students and
would be more effective for female students than for male students.
The overall results of this study show that most of the same relationships between
risk factors and school violence reported in studies from Western countries are also found
in Asian cultural contexts, especially in Taiwan. The findings provide useful information
for school policy makers or clinicians, who are interested in developing new
internationally-based school violence prevention and intervention strategies.
Surprisingly, the association between poor student-teacher relationships and student
violence against students is very weak. The factor of negative personal traits has weak

100
influence on student violence against teachers. Low level of school engagement and
parental monitoring have no direct link to both types of school violence. These are
unexpected findings because these variables are important predictors of school violence
in previous literature and theory. An explanation for these inconsistent findings may be
due to comprehensive model, sample, and data analysis approach in the present study.
Another possible explanation may be the conflicting cultural values between East and
West, contributing to these inconsistent findings. Future researchers must carefully
consider their use of expanded methodologies to accurately reflect differences across
multiple cultures. Also, future researchers may apply this theoretical model in other
countries in order to further confirm the similarities or differences across cultures in
school violence findings.
Finally, there are some limitations need to be considered. First, the investigation in
this study is cross-sectional. The results of this study may not be used to establish
cause-effect inference, which may require longitudinal data. Second, the self-reporting
characteristic of the data and the twelve-month reporting window may have resulted in
students under-reporting events due to poor memory or over-reporting violent behavior
that they assume to be commonplace (Benbenishty & Astor, 2005). Future researchers
will benefit from collecting information from additional sources such as teacher, parents,
or peer reports.

101
Table 9. Means and Standard Deviations of Subscales by Gender (Junior High School)
Overall Male Female Scale/subscale
Mean SD Mean SD Mean SD
Violence against students
a

Physical
Vandalism
Verbal/threat/harassment
2.22
0.52
0.43
1.26
2.35
0.86
0.68
1.28
2.89
0.80
0.54
1.55
2.54
0.98
0.74
1.35
1.56
0.26
0.32
0.98
1.92
0.61
0.60
1.13
Violence against teachers
a

Physical
Verbal/threat
Emotional/harassment
0.52
0.03
0.12
0.37
0.94
0.22
0.38
0.59
0.61
0.05
0.16
0.39
1.07
0.28
0.45
0.61
0.42
0.02
0.07
0.34
0.75
0.13
0.27
0.56
Negative personal traits
b

Attitude toward violence
Impulsive control
Trait anger temperament
51.00
16.95
12.26
21.92
14.48
5.96
4.16
6.50
53.37
18.65
12.51
22.32
15.24
6.45
4.25
6.68
48.74
15.23
12.00
21.49
13.26
4.90
4.03
6.26
Parental monitoring
b

Father monitoring
Mother monitoring
29.52
14.50
14.99
5.53
2.99
2.82
29.27
14.49
14.75
5.39
2.86
2.72
29.74
14.49
15.22
5.64
3.12
2.90
Victimization
c

Direct victimization
Witness victimization
2.17
1.04
1.14
2.83
1.68
1.72
2.48
1.30
1.19
3.09
1.85
1.81
1.86
0.78
1.08
2.52
1.45
1.63
Low school engagement
b

Doze off in class or skip class
Forget bringing material
Seldom turn in homework
Bring prohibited material
7.14
1.47
2.09
1.89
1.70
2.47
0.76
0.86
0.86
0.84
7.61
1.56
2.18
2.07
1.81
2.61
0.82
0.86
0.89
0.90
6.68
1.39
1.99
1.72
1.58
2.25
0.68
0.84
0.79
0.77
Risky peers
b

Risky friendship
Risky acts
9.35
6.81
2.55
3.61
2.88
1.13
9.91
7.14
2.78
4.07
3.11
1.35
8.83
6.51
2.32
3.04
2.61
0.81
Poor student-teacher relationship
b

Punish for no reason
Teachers do not like me
Teachers mock at me
Teachers never trust me
Observe me and snitch
9.11
1.79
1.75
1.65
1.84
2.09
3.48
0.83
0.81
0.78
0.89
1.00
9.50
1.89
1.82
1.73
1.93
2.14
3.61
0.87
0.85
0.82
0.93
1.00
8.71
1.68
1.68
1.57
1.74
2.05
3.31
0.77
0.77
0.72
0.84
1.00
a. On a scale: 0= never, 1= at least once in the past year.
b. On a scale: 1= strongly disagree to 4= strongly agree.
c. On a scale: 1=never, 2= sometimes, 3= often, and 4= almost every day.

102
Table 10.Matrix of Correlations among School Violence Subscales for Junior High School
A1A2A3B1B2B3C1C2C3D1D2E1 E2F1F2F3F4G1G2H1H2H3H4H5
A1 Student physical
A2 Student vandalism
A3 Student verbal/threat
B1 Teacher physical
B2 Teacher verbal/threat
B3 Teacher emotional
C1 Attitude toward violence
C2 Impulsive control
C3 Trait anger temperament
D1 Father monitoring
D2 Mother monitoring
E1 Direct victimization
E2 Witness victimization
F1 Dose off or skip
F2 Forget bringing material
F3 Seldom turn homework
F4 Bring prohibited material
G1 Risky friendship
G2 Risky acts
H1 Teacher punish
H2 Teacher dislike
H3 Teacher mock
H4 Teachers distrust
H5 Teacher observe
-- .45
--
.54
.55
--
.24
.14
.20
--
.36
.28
.34
.47
--
.37
.39
.46
.24
.44
--
.46
.34
.38
.11
.23
.26
--
.36
.33
.37
.12
.22
.30
.59
--
.33
.27
.34
.08
.19
.26
.59
.75
--
-.13
-.18
-.17
-.03
-.05
-.14
-.22
-.28
-.21
--
-.17
-.18
-.19
-.04
-.08
-.15
-.26
-.30
-.22
.80
--
.27
.18
.22
.08
.12
.13
.20
.20
.17
-.08
-.10
--
.22
.18
.24
.06
.12
.18
.23
.23
.21
-.08
-.06
.39
--
.30
.24
.20
.11
.22
.23
.33
.34
.26
-.22
-.26
.11
.12
--
.19
.16
.18
.07
.11
.15
.23
.29
.25
-.19
-.16
.10
.09
.33
--
.26
.22
.22
.07
.14
.19
.31
.36
.28
-.25
-.25
.15
.11
.43
.53
--
.23
.23
.23
.06
.16
.20
.32
.33
.25
-.23
-.22
.12
.13
.46
.31
.42
--
.40
.29
.34
.15
.28
.32
.48
.44
.36
-.17
-.21
.17
.23
.37
.18
.27
.35
--
.37
.26
.28
.20
.28
.22
.42
.34
.27
-.12
-.15
.16
.13
.38
.19
.27
.29
.54
--
.20
.21
.19
.08
.15
.24
.29
.28
.23
-.18
-.21
.13
.12
.31
.22
.30
.30
.25
.22
--
.22
.21
.20
.07
.16
.22
.29
.29
.25
-.19
-.22
.15
.13
.30
.21
.29
.31
.25
.21
.73
--
.19
.19
.21
.08
.17
.23
.26
.25
.21
-.16
-.19
.14
.13
.29
.23
.29
.32
.25
.20
.66
.71
--
.23
.24
.24
.09
.20
.25
.30
.31
.26
-.18
-.22
.15
.13
.32
.24
.32
.32
.27
.23
.63
.64
.62
--
.18
.19
.22
.04
.14
.22
.26
.26
.24
-.13
-.15
.10
.17
.23
.18
.22
.24
.22
.13
.43
.44
.44
.46
--

103
Table 11. Theoretical Domains, Subscale, and Items for Junior High School
Domain
(Alpha’s)
Subscales
(Loadings)
Items
Violence against
students
(Alpha= .80)
Physical
(loading=.72)

Students beat or kick other students (by group) in order to hurt him/her/them.
Students beat or kick other students (by individual) in order to hurt him/her/them.
Students use dangerous objects or instruments in order to harm students.
Vandalism
(loading=.66)

Students intentionally destroy or break school public belongings.
Students intentionally destroy or break other students’ belongings.
Verbal/threat
(loading=.77)

Students verbally threaten or intimidate other students.
Students curse or insult other students.
Students threaten or blackmail other students.
Students tease, mock, or play physically harmful tricks on other students on purpose.
Students threaten or force others to buy things.
Violence against
teachers
(Alpha= .62)
Physical
(loading= .50)

Students beat or kick teacher(s).
Students use dangerous objects or instruments to harm teacher(s).
Verbal/threat
(loading=.71)

Students curse or insult teacher(s).
Students threaten or intimidate teacher(s).
Students extort or blackmail teacher(s).
Emotional
(loading= .66)

Students tease, mock, or play physically harmful tricks on teacher(s).
Students oppose teacher(s) in order to cause them psychological harm.
(Table 11 continues)

104 Table 11 continued
Domain Subscales Items
Negative personal
traits
(Alpha=. 93)

Attitude toward
violence
(loading= .72)

Impulsive control
(loading= .88)

Trait anger
temperament
(loading= .83)

If someone insulted me or my family, beating him/her will make me feel better.
If someone disrespects me, I will beat him/her to regain honor
I will feel unhappy if I do not beat someone who cursed me.
A coward is a person who never retaliates when he/she is insulted.
If someone impedes my plans, he/she will pay for it.
If someone tries to hurt me, I will take vengeance on him/her.
Violence is the best way to resolve any problem.
Violence is justice.
If someone makes me unhappy, beating him is what he/she deserves.
Violence is better than negotiation.

I often make mistakes due to uncontrolled anger.
I often feel regretful about the things I do.
When I feel angry, no one can control me.
I have some bad habit that I always fail to change.
I can not help violating school rules.
I can not tolerate when others look down on me.

I loose my temper easily.
I am an irritable person.
I am easily agitated.
I feel anger if someone’s mistakes disturb my work.
I often can not control my anger.
When I feel angry, I will curse or use dirty words, such as fuck.
I will feel angry if someone criticizes me.
Beating others will make me feel better if I feel unhappy.
I feel upset if the work is done worse than I expect.
I feel angry if my work is criticized.
(Table 11 continues)

105 Table 11 continued
Domain Subscales Items
Parent monitoring
(Alpha=.87)

Father
(loading= .92)

My father (or father figure) knows my friends.
My father (or father figure) knows my conduct.
If I am not at home, my father (or father figure) will know where I am.
My father knows my schedules.
My father (or father figure) often takes activities with me.
Mother
(loading= .87)

My mother (or mother figure) knows my friends.
My mother (or mother figure) knows my personality and general conduct.
If I am not at home, my mother (or mother figure) knows where I am.
My mother knows my schedule.
My mother (or mother figure) often does activities with me.
Victimization
(Alpha= .84)

Direct
(loading=.63)

Have you been beaten or kicked by others?
Has someone hit and hurt you with an object of any kind?
Have you been beaten or kicked by groups of people after class or school?
Have you been blackmailed by others?
Have you been verbally threatened by others?
Have you been intimidated by others?
Has someone threatened or intimidated you by passing a slip of paper?
Witness
(loading= .62)

Have you seen anyone verbally threatened by others?
Have you seen anyone threatened by weapons?
Have you seen anyone destroy public belongings on purpose?
Have you seen anyone beaten by others or participating in group fights?
Have you seen anyone robbed by others?
Have you seen anyone insulted by others?
Have you seen anyone blackmailed by others?
(Table 11 continues)

106 Table 11 continued
Domain Subscales Items
Low school
engagement
(Alpha= .74)
Doze off or skip
(loading=.66)

Forget bringing
(loading=.57)

Homework
(loading= .70)

Prohibited material
(loading=.64)

I often doze off in class or skip class.

I often forget to bring required materials (textbooks/homework) to school.

I seldom turn in homework.

I often bring prohibited materials to school.
Risky peers
(Alpha= .79)

Risky friendship
(loading= .79)

I have friends who are school gang members.
I have friends who dropped out due to problem behavior in school.
When I have conflicts with others, my friends will help me beat them or fight with them.
My friends are always on my side no matter what bad things I do.
Risky acts
(loading=.68)

My friends and I often stay up all nights to drink, to gamble, or to do illegal activities.
My friends and I are often involved in fights or use our fists to protect our territory.
Poor student-teacher
relationship
(Alpha= .86)

Punish (loading=.83)

Dislike (loading=.86)

Mock (loading=.81)

Distrust(loading=.77)

Snitch (loading=.54)

My teacher often punishes me for no reason.

I feel that my teacher does not like me.

My teachers like to mock me.

Even though I tell the truth, my teacher still distrusts me.

Teachers often ask students to observe on what I am doing and snitch on me.

107
Figure 6. Total Sample of Junior High School Students’ Violence Against Students and Teachers.

108
Figure 7. Gender Comparison in Junior High School (Male Students in Bold and Oblique)

109
Chapter Six:
Study 6: School Variables as Mediators of Personal and Family Factors
on School Violence in High Schools
Introduction
Little is known about how school variables mediate between personal and family
factors and school violence, and even less is known about the patterns of relationships
between these factors in Asian cultures (Benbenishty & Astor, 2005). Furthermore, it is
not clear how the patterns of relationships differ between school types and between
genders. This study expands the literature on school violence by examining how school
experiences mediate the influence of personal and family factors on school violence in
Taiwanese vocational and academically oriented high schools.
There are few nationally-representative studies on school violence in Asian cultures.
Research on school violence conducted in Western cultures has demonstrated how
students’ personal, family, and school experiences independently affect students’
perpetration of violence against students (Baldry & Farrington, 2005; Brockenbrough,
Cornell & Loper, 2002). On the whole, these studies have examined each risk factor
separately rather than together. In addition, international studies have shown the vast
majority of perpetrators of violence against teachers are students (Dinkes, Cataldi, Kena,
& Baum, 2006; Lyon & Douglas, 1999). However, there has been far less research on
how risk factors contribute to student perpetration of violence against teachers
(Khoury-Kassabri, Astor & Benbenishty, 2008).

110
Findings from this study may have implications for other Asian cultures. Taiwan is
typical of many Asian cultures. Despite growing influence of globalization, the
Taiwanese people are still guided by Confucianism, Taoism, and Buddhism (Nisbett,
2003). While school violence issues have been a major public concerns for many years,
for political reasons (China’s lack of recognition of Taiwan as an independent country)
Taiwan has never been part of any large-scale cross-national studies on these topics
(Akiba, LeTendre, Baker, & Goesling, 2002). Relationship patterns of risk factors and
school violence have also never been explored in Taiwan or other Asian societies.
Overview of Taiwanese education system
There are three major systems of basic and intermediate education in Taiwan:
elementary school (grades 1 to 6), junior high school (grades 7 to 9), and high school
(grades 10 to 12). High schools could be further divided into academic and vocational. In
the academic year of 2006, there were about 0.42 million students in 314 academic high
schools and about 0.33 million in 157 vocational high schools in Taiwan (Taiwan
Ministry of Education [TME], 2006).
Attending a nine-year compulsory education program from elementary to junior
high school is mandatory. A competitive joint high school entrance exam must be taken if
junior high school graduates seek higher education. Students are assigned to academic or
vocational high schools based on their exam results (TME, 2006). This is similar to how
middle school students in the West apply to elite private schools, competitive magnet
schools, or national blue ribbon college preparatory high schools. Generally, students
with higher scores are encouraged to attend academic high schools, while those with

111
lower scores attend vocational or less rigorous high schools. This separation by academic
ability could account for different school structures and social climates.
Three-year academic high school programs aim to prepare students for specialized
learning as well as for college study, focusing on training for the joint university entrance
examination. As a result, academic high school students offer intense coursework and
concentrate on academic development. The goal of vocational high schools, however, is
to provide students with professional competencies and technical skills to help vocational
high school students enter the workforce. Most vocational high school graduates start a
business or take up employment in companies based on the technical skills that they learn
in high school (TME, 2006). V ocational high schools in Taiwan are similar to those in
Europe and the Middle East. In the United States, most high schools offer only a few
vocational components. Except for some alternative schools for “at risk” students, it is
rare to have high schools exclusively devoted to vocational issues. Moreover, the
separation of students based on academic ability does not occur as frequently in the U.S.
The boundaries and distinctions between academic and vocational high schools are much
stronger in Taiwan.
School variables as mediators of school violence
Literature on school violence indicates that most previous studies have examined, in
separate studies, how students’ experiences in each ecological context independently
affect perpetration of school violence (Alikasifoglu et al., 2004; Brockenbrough et al.,
2002; Chapell et al, 2006; O’Keefe, 1997). How multiple ecological risk factors and

112
nested contexts mediate student violence against students and teachers have not yet been
explored much in either the West or the East.
Benbenishty and Astor (2005) addressed this lack of theoretical integration
surrounding the nested contexts of school violence. They proposed a theoretical model in
which school violence is influenced by a combination of numerous within-school
variables (e.g., teacher child relationships) as well as other external variables (e.g.,
students’ personal and family characteristics). In their heuristic model, schools mediate,
moderate, and attenuate contributions from external contexts. School subcontexts also
generate their own direct contributions to different forms of violence. In contrast with
earlier psychological models that center exclusively around the individual, Benbenishty
and Astor stressed that when looking at “school violence” the school itself as a social
context should be the center of the theoretical model.
This model has been examined school violence across different cultures with Jewish
secular, Jewish Orthodox, Arabic, and American students (Astor, Benbenishty, Vinokur &
Zeira, 2006; Astor, Benbenishty, Zeira, & Vinokur, 2002; Benbenishty & Astor, 2005;
Benbenishty, Astor, Zeira, & Vinokur, 2002; Khoury-Kassabri et al., 2008). Using this
theoretical model, this is the first study examining how these complex patterns between
social contexts affect school violence in Taiwanese high schools.
Patterns of relationships between school types and genders
School type. Junior high school students in Taiwan are systemically assigned to
academic high school or vocational high school based on their results of comprehensive
academic exams. Some educational researcher would predict that the patterns of

113
relationships and school mediating mechanism should be different between academic and
vocational schools, because this kind of educational tracking based on academic ability
accounts for different students’ personal characteristics, family backgrounds, and school
dynamics (Chang, 1992; Gang & Klaus, 2000; Hsieh, 2003). However, some have argued
that the patterns of relationships may be not so different due to similar class structures,
peer group influence, and teaching styles in both types of schools (Astor & Meyer, 2001;
Benbenishty & Astor, 2005).
To date, several empirical studies suggested that the patterns of relationships are
similar in victimization of school violence among primary, middle, and high schools
(Astor et al., 2006; Astor et al., 2002; Benbenishty & Astor, 2005; Benbenishty et al.,
2002). However, no strong empirical evidence has emerged to suggest similarities or
differences in the kinds of perpetration. It is unclear how overarching patterns of
relationships differ between academic and vocational high schools, where personal,
family, and school variables are presumed to be entirely different.
Clarifying the similarities or differences in these patterns of relationships in
perpetration is meaningful to intervention or prevention. If patterns of relationships differ
between school types, this may suggest new school violence intervention programs must
be developed to address specific needs of each school type. Alternatively, if these patterns
are very similar between school types, it may mean that the “face value” of school type
differences is not so prominent in influencing patterns of relationships.
Gender. Many international studies suggest boys engage in more aggressive
behavior in schools than girls (Rigby, 2005; Baldry, 2003; Nansel et al., 2001; Zeira,

114
Astor, & Benbenishty, 2003). However, based only on these studies, it is not clear if the
patterns of relationships between risk factors and school violence perpetration differ
across gender. Perhaps what causes perpetration behaviors in boys is the same pattern that
causes it in girls, only that girls have lower base rates of violence. Thus, current study
examines how gender differences impact the pattern of relationship between factors and
school violence committed by students against students and teachers.
Risk factors of school violence in high school
Student violence against students. Western studies have shown that negative
personal traits (Brockenbrough et al., 2002), parental monitoring and involvement
(Baldry & Farrington, 2005), victimization (Alikasifoglu et al., 2004; Brockenbrough et
al., 2002; Chapell et al., 2006; O’Keefe, 1997), low school engagement, poor
relationships with peers and school teachers (Alikasifoglu et al., 2004; Desouza &
Ribeiro, 2005) are important factors in influencing high school students’ violence against
school students.
Few Asian studies report risk factors of school violence such as negative personal
traits, victimization, poor school engagement, involvement in at-risk peers (Ando,
Asakura & Simons-Morton, 2005; Hu & Lin, 2001), low level of parental monitoring,
and poor student-teacher relationships (Hu & Lin, 2001; Ma, Shek, Cheung, & Tam, 2002;
Yoneyama & Naito, 2003; Wong, 2004). Most of these studies were conducted using
nonrepresentative samples, so it is not clear how generalizable these Western findings are
for Asian cultures.

115
Student violence against teachers. Students’ perpetration of violence against teachers
has negatively impacted teachers’ personal safety, mental health and their quality of
teaching (Astor, Meyer, Benbenishty, Marachi & Rosemond, 2005; Lyon & Douglas,
1999). To date, there have been far fewer reports on how risk factors contributed to
students’ perpetration of violence against teachers. Recently, Khoury-Kassabri et al.
(2008) indicated that student violence against teachers is influenced by the students’
individual factors (i.e., gender, age, perception of school climate, and intervention) and
school contextual factors (i.e., cultural affiliation, socioeconomic status, family, school,
class size, school climate, and intervention). However, it is still unclear how students’
personal traits, parental monitoring, victimization, school engagement and school social
dynamics contribute to students’ perpetration of violence against teachers.
Some smaller scale studies in Taiwan have shown students’ negative personal traits,
parental involvement, victimization, risky behavior in school (Chen, 1999), school
climate, school engagement (Chen, 1989, 1999), and poor student-teacher relationships
(Huang & Hsien, 1987; Lin, 1986) were related to student violence against teachers.
However, most of these studies in Taiwan were analyzed, based on teachers’ reports and
were conducted using nonrepresentative samples. It is unknown how students’ personal,
family and school experiences contributed to students’ perpetration of violence against
teachers in Asian cultures.
Theoretical model for this study
Figure 1 shows the theoretical model used in this study. Based on previous literature
and theory, this study proposes that the perpetration of school violence by students

116
against peers and/or teachers is directly associated with negative personal traits, parental
monitoring, victimization, low level of school engagement, risky peers, and poor
student-teacher relationships. Please note that the lines with plus and minus in Figure 1
represent the hypothesis in this study. The plus sign indicates, based on prior studies, a
positive relationship between the two variables and minus sign indicates negative. For
example, the line with plus sign between risky peers and student violence against students
means this model hypothesizes students who have risky peers are more likely to be
involved in student violence against students. The line with minus sign between parental
monitoring and risky peers means that students who have good parental monitoring are
less likely to have risky peers.
Most importantly, different from some prior studies, this model proposes that
external school variables will have indirect effects on school violence mediated through
school variables (see Benbenishty and Astor, 2005 for an exception). For example, this
model proposes that good parental monitoring will indirectly influence student violence
against students by affecting risky peers. That is, students who have good parental
monitoring are less likely to have risky peers, and in turn, their chance to involve in
school violence is decreased.
This study expected that the model fit male and female students as well as academic
and vocational high school students, but certain paths within the model may differ.
Method
The data used in this study was part of a large-scale project of “Prevention and
Control of School Violence in Taiwan” (Wu, Lee, Yin, & Hu, 2000). The survey was

117
conducted throughout Taiwan among over 14,000 students from elementary schools
(grades 4 to 6), junior high schools (grades 7 to 9), vocational high schools and academic
high schools (grades 10 to 12). Students were given a structured questionnaire in
classrooms under the guidance of professionally trained survey monitors. Respondents
were assured of anonymity and were encouraged to respond truthfully. Participants were
free to withdraw from the study at any time and for any reasons. This study was
supported by Taiwan’s National Science Council (NSC). The questionnaires, procedures,
informed consent forms and other ethical concerns were reviewed and supervised by
NSC.
This sample was designed to represent all students from 4th to 12th grades in
Taiwan. The students’ response rate was over 98 percent. The probability sampling
method was a two-stage stratified cluster sample. The strata were
northern/central/southern/eastern, urban/rural, and elementary/junior/vocational/academic.
Current study only examined students from academic and vocational high school. In first
stage, schools were randomly selected from the sampling frame according to those
appropriate strata. In the next stage, two classes were randomly selected according to
each grade in selected schools. All students in that class were included in sample.
7,841 students participated in this study. 48.4 percent of the students were boys, 51.3
percent were girls, and 0.3 percent did not indicate gender; 48.1 percent of the students
were in the academic high school and 51.9 percent were in the vocational high school.
The questionnaire was developed on the basis of current school violence studies and
theories from both Taiwan and Western countries, and validated by Wu et al. (2000). It

118
included over 150 items in eight domains regarding students’ basic demographic
background and other information in their personal, family, and school experience. Each
of the student questionnaires took approximately 30 minutes to complete. Before this
survey was conducted, the Mandarin Chinese questionnaire was adjusted and adapted
based on two pilot studies conducted in the Tainan metropolitan areas in Taiwan (Wu et
al., 2000).
Measurement
In order to measure the latent variables in our model, several scales containing a
number of subscales were constructed. Table 12 represents the means and standard
deviations of the variables included in the model, broken down by genders and school
types. Table 13 represents the intercorrelations matrix among the variables. For the
purpose of structural equation modeling analyses, several subscales were constructed
based on conceptual and theoretical constructs. Each of these subscales creates the
overall factor composites that represents of the more general theoretical concepts
discussed in literature review of this study. Table 14 lists the domains, questions items,
factor loadings, and alphas for the theoretically created subscales.
Dependent Variables
Student violence against students. This domain involves asking students how many
times they perpetrated violent behavior in school against other students during the last
academic year. The variables in this domain were recorded into “never” and “at least
once”. This domain included three subscales on the basis of their content according to
type of violence. It included physical violence, vandalism, and verbal

119
violence/threat/harassment. The score of each subscale was the sum of the items included
in the subscale.
Student violence against teachers. This domain involved asking students to indicate
how many times they perpetrated violent behavior in school against teachers during the
last year. The variables in this domain were recorded as “never” and “at least once”. This
domain included three subscales on the basis of their content according to type of
violence. It included physical violence, verbal violence/threat, and emotional
violence/harassment. The score of each subscale was the sum of the items included in the
subscale.
Independent variables
Negative personal traits. This latent variable asked students about their personal
characteristics related to aggression. Three subscales were constructed on the basis of
their content. These subscales were students’ attitude towards violence, impulsive control,
and trait anger temperament. The rating for each item in this scale ranged from 1=
strongly disagree to 4= strongly agree. The score of each subscale was the sum of the
items included in the subscale.
Parental monitoring. This latent variable consisted of two subscales: father
monitoring and mother monitoring. Respondents were asked about parental monitoring of
them in daily life. The rating for each item ranged from 1= strongly disagree to 4=
strongly agree. Both subscales of father and mother monitoring consisted of five items,
and each subscale score was the sum of these five items.

120
Victimization. This latent variable asked students if they had experienced
victimization surrounding their life in the past year. The rating for each item ranged from
1=never to 4= almost every day. This latent variable consisted of two subscales on the
basis of their content. The two subscales were direct victimization and witness
victimization. The score of each subscale was the sum of the items included in the
subscale.
School engagement. This domain consisted of four items, and students were asked
questions in order to measure their level of school engagement. The ratings were
provided on a four-point scale and were coded from 1= strongly disagree to 4= strongly
agree.
Risky peers. In this domain, students were asked questions to determine the quality
of their group of friends. The ratings ranged from 1= strongly disagree to 4= strongly
agree. The score of each subscale was the sum of the items included in the subscale. This
domain was divided into two subscales based on preliminary factor analysis. The two
subscales were risky friendships and risky acts.
Poor student-teacher relationships. This domain consisted of five items about
whether the respondents experienced a poor relationship between teachers and
themselves. The ratings for each item were provided on a four-point scale from 1=
strongly disagree to 4= strongly agree.
Analytical plan
The primary analysis method in this study is latent variables structural equations
modeling (SEM) with maximum likelihood (ML) estimation using AMOs program. SEM

121
is a statistical methodology that takes a confirmatory approach to the analysis of
structural theory describing relationships among endogenous factors (Bentler, 1988). The
structure/hypothesized model can be tested statistically in a simultaneous analysis of an
entire system of variables to determine the extent to which it is consistent with the data.
Confirmative factor analysis (CFAs) was first to be conducted to ensure the measurement
model as a good fit. Following the CFAs, the final SEM model including the full dataset
of all elementary school students was tested. Next, comparative analyses were conducted
in order to determine whether patterns of relationships and mediating effects are different
between male and female students. In this comparative analysis, all the factor loadings,
the paths, and the covariances are constrained to be equal simultaneously to the same
model across genders. Then, the model was tested by releasing path constraints one at a
time in order to find out if releasing equality constraints could significantly improve the
fit.
There are various indicators of the goodness of fit for a specific model. The
chi-square coefficient is used to assess the size of discrepancies between the relationships
in the original data matrix from those implied by the model. A low chi-square measure
reveals non-significant discrepancies, and means that the data “fit” the theoretical model.
However, due to the sensitivity of the chi-square coefficient to sample size, it is not a
preferred fit index for large samples such as those in this study. Indeed, researchers have
addressed the chi-square limitation by developing goodness-of-fit indices that take a
more pragmatic approach to the evaluation process. More commonly used fit indices
includes Bentler and Bonnett’s (1980) Normed Fit Index (NFI), Bollen’s (1989)

122
Incremental Fit Index (IFI), and Bentler’s (1990) Compared Fit Index (CFI). Typically,
these three fit indices consider a model to be a good fit when the value is above .90
(Bentler, 1992), and a superior fit when it is close to .95 (Hu & Bentler, 1999). A
common misfit measure, the Root Mean Square Error (RMSEA), is also reported in the
SEM analysis. The RMSEA considers a mediocre fit to range from .08 to .10 and a good
fit to be below .06 (Hu & Bentler, 1999).
Results
Overall model
The results of the analysis, based on the total sample, provided a good fit to the data
[ χ
2
(224, N=7,841) =5,405.04, p<.001, and with NFI=.93, IFI=.93, CFI=.93, and
RMSEA=.05]. This suggested that the model is a good one. Figure 8 presents the paths in
this model.
Figure 8 shows that risky peers and school engagement had mediating effects on
student violence against students for negative personal traits, parental monitoring and
victimization. This mediating effect was stronger for risky peers. Contrary to previous
literature on school violence, the risk factor of poor student-teacher relationships was a
poor predictor of student violence against students ( β=-.02).
For student violence against teachers, low level of school engagement, risky peers,
and poor student-teacher relationships had a mediating effect on negative personal traits,
parental monitoring, and victimization. The mediating effect was stronger for at-risk
peers. Overall, the variable of risky peers was the best predictor of student violence
against students ( β=.40) and student violence against teachers ( β=.39). Together, all of

123
these variables accounted for one-half of the explained variance for student violence
against students (R
2
=.50) and around one-third for student violence against teachers
(R
2
=.34).
School-type comparison analysis
In this analysis, factor loadings, the paths, and the covariances were constrained to
be equal, in order to fit the covariance matrices of the academic and vocational subgroups
simultaneously to the same model. The analysis provided a good fit to the data [ χ
2
(492,
N: academic=3,769, vocational=4,072) =5857.28, p<.001 and with NFI= .92, IFI=.93,
CFI=.93, and RMSEA= .04].
Next, the model was tested to find out if releasing equality constraints on the paths
could significantly improve the fit. After releasing path constraints one at a time, it was
determined that the release of the constraints between negative personal traits and risky
peers, as well as between risky peers and student violence against students, yield a
significantly better fit to the model; thus, the results for the final model with the two
constraints released together were [ χ
2
(490, N: academic=3,769,
vocational=4,072)=5,834.48, p<.001 and with NFI=.92, IFI=.93, CFI=.93, and
RMSEA= .04]. Figure 9 represents the results of this analysis.
Figure 9 shows that the beta coefficients for both academic and vocational models
exhibit remarkable similarities. However, negative personal traits had a stronger
magnitude of influence on risky peers for vocational settings ( β=.54) as compared with
academic ones ( β=.50). In addition, the effect of risky peers on poor student-teacher
relationships was stronger for vocational schools than for academic schools ( β=.35 for

124
academic vs. β=.42 for vocational). The overall model explained similar proportion of
variance on violence against students for academic schools (R
2
=.48) and vocational
schools (R
2
=.51). For violence against teachers, the overall model explained around one
third of variance for both academic schools (R
2
=.34) and vocational schools (R
2
=.35).
Gender analysis
The gender analysis was conducted similar to the procedure for the school
comparison analysis. First, factor loadings, the paths, and the covariances were
constrained to be equal, in order to fit the covariance matrices of the male and female
subgroups simultaneously to the same model. The analysis provided a good fit to the data
[ χ
2
(492, N: male=3,794, female=4,026)= 5807.87, p<.001 and with NFI= .92, IFI=.93,
CFI=.93, and RMSEA= .04].
Next, path constraints were released, one at a time. It was determined that the
release of the constraints between negative personal traits and risky peers, between
negative personal traits and poor student-teacher relationships, between risky peers and
student violence against students, and between victimization and student violence against
teachers, yield a significantly better fit to the model. Thus, the results for the final
model, with the four constraints released together, were [ χ
2
(488, N: male=3,794,
female=4,026)= 5736.07, p<.001 and with NFI= .92, IFI=.93, CFI=.93, and
RMSEA=.04]. Figure 10 represents the results of the gender analysis.
There are differences between genders in several paths. The influences of negative
personal traits on risky peers and on student-teacher relationships are stronger for males
( β=.54 and β=.35, respectively) than for females ( β=.51 and β=.27, respectively). For

125
both genders, risky peers had the strongest impact on student violence against students;
however, the impact was stronger for males ( β=.44) than for females ( β=.35). Finally,
prior victimization has around two times stronger impact on student violence against
teachers for males than for females ( β=.24 for males vs. β=.10 for females). The overall
model explained the same proportion of variance on violence against students for males
and females (R
2
=.45). For violence against teachers, the overall model explained 43
percent of variance for males and 35 percent for females.
Discussion
This study, using a nationally representative sample in Taiwan, examined how
school engagement, risky peers and poor student-teacher relationships mediate school
violence perpetration by students against students and teachers in Asian high school
settings. The study proposed that low level of school involvement, risky peers, and poor
student-teacher relationships mediated the effects from negative personal traits, parental
monitoring, and victimization. The study also examined if the overarching patterns of
relationships between risk factors and perpetration of school violence are similar or
different between school types as well as between genders. This study expected that the
model fit males and females as well as academic and vocational high school students.
Overall model
Results of this study showed good indices of model fit. This suggested that the
theoretical model of how patterns of relationships affect school violence perpetration is
supported. Furthermore, the variables in this model account for a relatively large amount
of the explained variance for student violence against students and student violence

126
against teachers. This suggested that the pattern of relationships between students’
personal traits, parental monitoring, personal victimization, school engagement, risky
peers, and student-teacher relationships greatly contribute to student violence against
students and teachers.
The results showed that school engagement and risky peers mediate student violence
against students for negative personal traits, parental monitoring and students’ prior
victimization. School engagement, risky peers and student-teacher relationship mediated
student violence against teachers for negative personal traits, parental monitoring and
students’ prior victimization. These findings support Benbenishty and Astor’s (2005)
theoretical model that school violence is influenced by a combination of numerous
within-school and external-school variables, and within-school variables can mediate
contributions from external contexts. In addition, these mediation findings provide
empirical evidence to support the idea that school environment is a unique and
independent developmental setting that contributes to school violence directly and
indirectly (Astor et al., 2002; Benbenishty et al., 2002).
The results show that compared to the low level of school engagement and poor
student-teacher relationships, risky peers had much stronger mediating effect on student
violence against students and teachers. This finding suggests that improving student
school experiences could show a strong effective response to violence. Most importantly,
interventions focusing on decreasing students’ involvement in at-risk peers may lead to a
greater reduction of student violence against students and teachers. In addition to school
factors, interventions strategies geared toward decreasing school violence for high school

127
students should also target on students’ negative personal traits and their overall
experiences of victimization, because the findings suggest that these two variables have
important direct and indirect influence that fuel both types of violence.
The overall results of this study show that most of the same relationships between
risk factors and school violence reported in studies from western countries are also found
in Asian cultural contexts, especially in Taiwan. The findings provide useful information
for school policy makers or clinicians in Asia, who are interested in developing or
adapting school-based school violence prevention and intervention strategies.
However, student violence against students is not strongly predicted by parental
monitoring and poor student-teacher relationships. Negative personal traits and parental
monitoring are also weak predictors of student violence against teachers. These findings
were unexpected, because previous studies and theories suggested that those variables
should be strong predictors of school violence perpetration. An explanation for these
inconsistent findings may be due to our more comprehensive model, type of sample, and/
or data analysis approach. Future research should include a more detailed contextual
analysis to clarify whether these results are due to the sample or the methodology (or
perhaps a new finding). Future researchers may apply this theoretical model in other
countries in order to further confirm the similarities or differences across cultures in
school violence findings.
School type differences
The results show that the overall theoretical model is applicable across school types.
For academic and vocational high schools, low level of school engagement, risky peers,

128
and poor student-teacher relationships had similar mediating effects on both student
violence against students and teachers. These findings suggest that educational tracking
by academic ability does not account for different students’ perpetration patterns in
vocational and academic schools. In fact, the same school variables mediate the same
external factors in the same way.
However, it is interesting to note that both the links from negative personal traits to
risky peers and from risky peers to student violence against students are slightly stronger
for vocational students. This suggested that risky peers had a stronger mediating effect
from negative personal traits to student violence against students for vocational high
school students. The findings imply that intervention aimed at eliminating negative
personal traits and reducing risky peers would lessen student violence against students
and would be more likely to affect vocational high school students.
Gender differences
The results show that the overall theoretical model is applicable across both genders.
These findings suggested that although many international studies suggest that male
students have higher rates of perpetration in school violence than females (Rigby, 2005;
Baldry, 2003; Nansel et al., 2001), how male and female students’ personal, family, and
school experiences affect student violence against students and teachers are quite similar.
This finding also indicates that school violence interventions or policy incorporated at a
national level could be affective across genders.
However, it is interesting to note that students’ negative personal traits have stronger
influence on risky peers for male students. The link between risky peers and student

129
violence against students are also stronger for male students. These findings suggested
that risky peers had a stronger mediating effect from negative personal traits to student
violence against students for male students. The findings imply that intervention aimed at
negative personal traits and reducing risky peers would lessen student violence against
students and would be more likely to affect male students. However, risky peer groups at
school are a key mediator in this process. In addition, male prior victimization had a
slightly stronger impact on student violence against teachers. This finding suggested that
intervention aimed at decreasing or better dealing with student victimization could reduce
student violence against teachers. Our model suggests this strategy would be more
effective for male students than for female students.
Overall, this study should be replicated with a longitudinal sample that can better
determine the direction of effects and causal relationships. It suggests that interventions
aimed both at the students (one’s traditionally used) and those aimed at the school site
(mediational) could be key at disrupting the trajectory towards violence at school towards
teachers and students. Risky peer groups on school groups appear to be the strongest
mediator for all school types and across gender. Finally, the patterns of relationships that
exist in Taiwan appear theoretically and empirically similar to those found in the Western
countries. This raises theoretical questions that need to be explored surrounding the
commonality between school variables and family or personality variables in different
cultures across the globe.

130
Table 12: Means and Standard Deviations of Subscales by Genders and School Types
Scale/subscale Overall Male Female Academic Vocational
Mean SD Mean SD Mean SD Mean SD Mean SD
Violence toward students
a

Physical
Vandalism
Verbal/threat
1.64
0.33
0.29
1.02
2.09
0.73
0.59
1.21
2.28
0.52
0.38
1.38
2.33
0.87
0.67
1.31
1.02
0.15
0.19
0.68
1.61
0.49
0.49
1.00
1.44
0.26
0.26
0.92
1.94
0.65
0.56
1.16
1.82
0.39
0.32
1.11
2.21
0.79
0.62
1.25
Violence toward teachers
a

Physical
Verbal/threat
Emotional
0.50
0.02
0.08
0.39
0.87
0.18
0.33
0.59
0.58
0.03
0.10
0.45
0.95
0.22
0.37
0.61
0.42
0.01
0.06
0.34
0.77
0.13
0.27
0.57
0.47
0.02
0.07
0.38
0.83
0.17
0.30
0.58
0.53
0.02
0.09
0.41
0.90
0.19
0.35
0.60
Negative personal traits
b

Attitude toward violence
Impulsive control
Trait anger temperament
51.33
17.04
12.66
21.63
13.69
6.12
3.88
5.92
54.46
19.25
13.06
22.14
14.18
6.45
3.97
6.00
48.40
14.96
12.28
21.15
12.52
4.95
3.75
5.80
50.53
16.60
12.59
21.35
13.26
5.93
3.78
5.77
52.07
17.44
12.71
21.87
14.04
6.26
3.96
6.05
Parental monitoring
b

Father monitoring
Mother monitoring
28.38
13.88
14.48
5.68
3.09
2.88
27.67
13.62
14.03
5.53
2.98
2.80
29.05
14.12
14.90
5.73
3.17
2.89
28.67
13.99
14.65
5.66
3.11
2.87
28.11
13.77
14.31
5.68
3.06
2.88
Victimization
c

Direct victimization
Witness victimization
1.68
0.64
1.04
2.61
1.42
1.70
2.17
0.89
1.28
2.85
1.63
1.81
1.21
0.40
0.81
2.25
1.13
1.55
1.51
0.59
0.93
2.48
1.35
1.59
1.84
0.69
1.15
2.71
1.48
1.78
Low school engagement
b

Doze off or skip
Forget bringing material
Seldom turn in
Bring prohibited material
7.36
1.70
2.03
1.91
1.72
2.41
0.82
0.80
0.81
0.79
7.76
1.79
2.07
2.02
1.89
2.48
0.86
0.80
0.84
0.83
6.98
1.62
2.00
1.81
1.56
2.28
0.78
0.80
0.78
0.72
7.28
1.63
2.04
1.94
1.68
2.36
0.79
0.80
0.82
0.77
7.43
1.77
2.03
1.89
1.76
2.45
0.85
0.80
0.80
0.81
Risky peers
b

Risky friendship
Risky acts
9.89
7.36
2.54
3.42
2.82
1.02
10.64
7.90
2.47
3.68
2.99
1.17
9.19
6.85
2.34
2.96
2.54
0.81
9.65
7.20
2.45
3.24
2.72
0.95
10.12
7.51
2.61
3.56
2.90
1.08
Poor studentteacher relation
b

Punish for no reason
Teachers do not like me
Teachers mock at me
Teachers never trust me
Observe me and snitch
8.82
1.68
1.69
1.64
1.75
2.07
3.27
0.73
0.75
0.74
0.81
0.97
9.24
1.77
1.78
1.73
1.84
2.12
3.40
0.77
0.79
0.78
0.84
0.97
8.43
1.58
1.60
1.56
1.66
2.03
3.09
0.67
0.71
0.70
0.76
0.96
8.68
1.63
1.65
1.62
1.71
2.07
3.21
0.70
0.74
0.74
0.78
0.97
8.95
1.72
1.72
1.66
1.78
2.07
3.33
0.75
0.77
0.75
0.83
0.97
a. On a scale: 0= never, 1= at least once in the last year.
b. On a scale: 1= strongly disagree to 4= strongly agree.
c. On a scale: 1=never, 2= sometimes, 3= often, and 4= almost every day.

131
Table 13. Matrix of Correlations among School Violence Subscales for High School
A1A2A3B1B2B3C1C2C3D1D2E1 E2 F1F2F3F4G1G2H1H2H3H4H5
A1 S-S Physical violence
A2 S-S Vandalism
A3 S-S Verbal/threat
B1 S-T Physical violence
B2 S-T Verbal/threat
B3 S-T Emotional
C1 Attitude toward violence
C2 Impulsive control
C3 Trait anger temperament
D1 Father monitoring
D2 Mother monitoring
E1 Direct victimization
E2 Witness victimization
F1 Dose off
F2 Forget bringing material
F3 Seldom turn in homework
F4 Bring prohibited material
G1 Risky friendship
G2 Risky acts
H1 Punish for no reason
H2 Teachers do not like me
H3 Teachers mock at me
H4 Teachers never trust me
H5 observe me and snitch
-- .43
--
.51
.53
--
.27
.16
.21
--
.36
.28
.32
.52
--
.36
.39
.47
.20
.40
--
.40
.28
.35
.16
.23
.26
--
.27
.23
.30
.11
.19
.26
.54
--
.25
.22
.31
.10
.18
.25
.57
.71
--
-.13
-.12
-.15
-.03
-.06
-.12
-.17
-.22
-.17
--
-.16
-.13
-.16
-.03
-.09
-.13
-.21
-.22
-.17
.81
--
.23
.17
.22
.13
.16
.14
.15
.10
.10
-.07
-.09
--
.26
.23
.27
.11
.19
.23
.24
.19
.18
-.08
-.09
.39
--
.24
.18
.20
.12
.20
.24
.28
.32
.25
-.20
-.22
.07
.11
--
.13
.13
.14
.07
.12
.15
.18
.26
.20
-.14
-.14
.07
.08
.36
--
.16
.16
.16
.08
.14
.18
.23
.27
.19
-.19
-.21
.09
.10
.40
.55
--
.26
.23
.26
.13
.20
.25
.32
.32
.24
-.18
-.20
.11
.15
.42
.33
.40
--
.39
.26
.34
.16
.25
.29
.45
.37
.32
-.12
-.14
.12
.25
.28
.15
.17
.30
--
.36
.21
.25
.23
.28
.20
.40
.30
.26
-.13
-.17
.17
.18
.32
.17
.21
.30
.46
--
.19
.16
.18
.11
.20
.24
.30
.26
.22
-.12
-.16
.11
.12
.30
.23
.26
.33
.23
.26
--
.18
.15
.18
.10
.19
.23
.30
.26
.23
-.14
-.17
.10
.12
.29
.23
.25
.32
.22
.26
.78
--
.16
.15
.17
.10
.16
.21
.27
.25
.22
-.14
-.17
.11
.11
.28
.22
.25
.32
.19
.25
.68
.75
--
.18
.16
.19
.11
.18
.22
.30
.27
.23
-.15
-.19
.10
.14
.30
.22
.27
.32
.22
.25
.64
.69
.67
--
.15
.17
.19
.07
.12
.24
.26
.26
.23
.-.11
.-.11
.07
.14
.22
.17
.19
.26
.26
.18
.44
.45
.42
.49
--

132
Table 14. Theoretical Domains, Subscales, and Items for High School
Domain
(Alpha’s)
Subscales
(Loadings)
Items
Violence against
students
(Alpha=.79)
Physical
(loading=.70)

Students beat or kick other students (by group) in order to hurt him/her/them.
Students beat or kick other students (by individual) in order to hurt him/her/them.
Students use dangerous objects or instruments in order to harm students.

Vandalism
(loading=.64)

Students intentionally destroy or break school public belongings.
Students intentionally destroy or break other students’ belongings.
Verbal/threat
(loading=.75)

Students verbally threaten or intimidate other students.
Students curse or insult other students.
Students threaten or blackmail other students.
Students tease, mock, or play physically harmful tricks on other students on purpose.
Students threaten or force others to buy things.
Violence against
teachers
(Alpha=.58)
Physical
(loading=.55)

Students beat or kick teacher(s).
Students use dangerous objects or instruments to harm teacher(s).
Verbal/threat
(loading=.75)

Students curse or insult teacher(s).
Students threaten or intimidate teacher(s).
Students extort or blackmail teacher(s).
Emotional
(loading=.60)

Students tease, mock, or play physically harmful tricks on teacher(s).
Students oppose teacher(s) in order to cause them psychological harm.
(Table 14 continues)

133 Table 14 continued
Domain Subscales Items
Negative personal
traits
(Alpha=.92)

Attitude toward
violence
(loading=.70)

If someone insulted me or my family, beating him/her will make me feel better.
If someone disrespects me, I will beat him/her to regain honor
I will feel unhappy if I do not beat someone who cursed me.
A coward is a person who never retaliates when he/she is insulted.
If someone impedes my plans, he/she will pay for it.
If someone tries to hurt me, I will take vengeance on him/her.
Violence is the best way to resolve any problem.
Violence is justice.
If someone makes me unhappy, beating him is what he/she deserves.
Violence is better than negotiation.

Impulsive control
(loading=.82)

I often make mistakes due to uncontrolled anger.
I often feel regretful about the things I do.
When I feel angry, no one can control me.
I have some bad habit that I always fail to change.
I can not help violating school rules.
I can not tolerate when others look down on me.

Trait anger
temperament
(loading=.83)

I loose my temper easily.
I am an irritable person.
I am easily agitated.
I feel anger if someone’s mistakes disturb my work.
I often can not control my anger.
When I feel angry, I will curse or use dirty words, such as fuck.
I will feel angry if someone criticizes me.
Beating others will make me feel better if I feel unhappy.
I feel upset if the work is done worse than I expect.
I feel angry if my work is criticized.
(Table 14 continues)

134 Table 14 continued
Domain Subscales Items
Parent monitoring
(Alpha=.89)

Father
(loading=.95)

My father (or father figure) knows my friends.
My father (or father figure) knows my conduct.
If I am not at home, my father (or father figure) will know where I am.
My father knows my schedules.
My father (or father figure) often takes activities with me.
Mother
(loading=.86)

My mother (or mother figure) knows my friends.
My mother (or mother figure) knows my personality and general conduct.
If I am not at home, my mother (or mother figure) knows where I am.
My mother knows my schedule.
My mother (or mother figure) often does activities with me.
Victimization
(Alpha=.86)

Direct
(loading=.55)

Have you been beaten or kicked by others?
Has someone hit and hurt you with an object of any kind?
Have you been beaten or kicked by groups of people after class or school?
Have you been blackmailed by others?
Have you been verbally threatened by others?
Have you been intimidated by others?
Has someone threatened or intimidated you by passing a slip of paper?
Witness
(loading=.72)

Have you seen anyone verbally threatened by others?
Have you seen anyone threatened by weapons?
Have you seen anyone destroy public belongings on purpose?
Have you seen anyone beaten by others or participating in group fights?
Have you seen anyone robbed by others?
Have you seen anyone insulted by others?
Have you seen anyone blackmailed by others?
(Table 14 continues)

135 Table 14 continued
Domain Subscales Items
Low school
engagement
(Alpha=.74)

Doze off or skip
(loading=.64)

Forget bringing
(loading=.62)

Homework
(loading=.68)

Prohibited material
(loading=.64)

I often doze off in class or skip class.

I often forget to bring required materials (textbooks/homework) to school.

I seldom turn in homework.

I often bring prohibited materials to school.

Risky peers
(Alpha=.76)

Risky friendship
(loading=.71)

I have friends who are school gang members.
I have friends who dropped out due to problem behavior in school.
When I have conflicts with others, my friends will help me beat them or fight with them.
My friends are always on my side no matter what bad things I do.
Risky acts
(loading=.65)

My friends and I often stay up all nights to drink, to gamble, or to do illegal activities.
My friends and I are often involved in fights or use our fists to protect our territory.

Poor student-teacher
relationship
(Alpha=.87)
Punish (loading=.84)

Dislike (loading=.90)

Mock (loading=.83)

Distrust(loading=.78)

Snitch (loading=.53)
My teacher often punishes me for no reason.

I feel that my teacher does not like me.

My teachers like to mock me.

Even though I tell the truth, my teacher still distrusts me.

Teachers often ask students to observe on what I am doing and snitch on me.

136
Figure 8. Total Sample of High School Students’ Violence Against Students and Teachers.

137
Figure 9. School Type Comparison for High School Students (Academic High School Students are in Bold and Oblique)

138
Figure 10. Gender Comparison in High school (Male students in bold and oblique)

139
Chapter 7: Integration and Implications from Findings Across All Five
Studies
Using a nationally representative sample in Taiwan, this dissertation explores the
prevalence of school violence perpetration committed by students against other students
and teachers. It also examines how students’ personal, family, and school factors
influence school violence in Asian cultures. In contrast to earlier psychological theories
which stress the single risk factor contributions of prior victimization, family factors, and
psychological characteristics to perpetration, this dissertation proposes a theoretical
model hypothesizing that low levels of school engagement, exposure to risky peers, and
poor student-teacher relationships mediate the effects of students’ negative personal traits,
parental monitoring, and prior victimization to student violence against students and
teachers. This model is highly influenced by Benbenishty and Astor’s (2005)
social-ecological model, which stressed that when looking at “school violence”, the
school itself, as a social context, should be the center of the theoretical model.
The sample in this dissertation is representative of the entire student population from
4th to 12th grades in Taiwanese schools. Students’ reports are compared by both gender
and school type. The first manuscript focuses specifically on examining the prevalence of
student violence against other students, while the second focuses on the prevalence of
student violence against teachers. These two studies are the first national estimates of
student violence against students and teachers in Asian culture contexts. The third inquiry
examines how school factors mediate family and individual factors among elementary
school students, while the fourth examines how school factors mediate family and

140
individual factors in junior high school. The fifth manuscript compares the overarching
mediational patterns within different types of high schools (academic vs. vocational),
where educational tracking based on academic ability accounts for different students’
personal characteristics, family backgrounds, and school dynamics. The studies in
combination show that different school environment aspects mediate family and personal
factors in different ways in elementary, junior high, and high schools.
The major findings of this study, implications for theory, policy, and practice, and
recommendations for future research are discussed in this chapter.
Findings of prevalence of school violence: Consistencies and inconsistencies with
Western literature and theories
Studies 1 and 2 in this dissertation indicate a relatively high rate of school violence
against students and teachers in Taiwan (59.7 percent for student violence against
students and 30.1 percent for student violence against teachers) compared to many other
countries (e.g., Baldry, 2003; Dinkes, Cataldi, Kena, & Baum, 2006; Khoury-Kassabri,
Astor & Benbenishty, 2008; Nansel et al., 2001). These findings do not support
contemporary and popular theories asserting that certain Asian cultural values, such as
emphasizing harmony in social relationships, may account for a lower prevalence of
school violence compared to Western cultures (Nisbett, 2003).
In general, the findings in the Studies 1 and 2 regarding different types of school
violence perpetration follow similar patterns as those published in Western and Middle
Eastern countries (e.g., Benbenishty & Astor, 2005; Nansel et al., 2001). The findings do
not support most culture theories stating that different cultures experience types of

141
violence in different rank orders based on the values and circumstances which are unique
to each culture, ethnicity, and country (Benbenishty & Astor, 2005; Nisbett, 2003).
Furthermore, the results indicate that gender, school type, and development-related
patterns of school violence base rates are similar to findings in Western countries (e.g.,
Benbenishty & Astor, 2005; Nansel et al., 2001). This may suggest that gender and age,
rather than culture, are the most important variables in understanding overall prevalence
of school violence perpetration.
Gender Findings
The results show a clear and consistent trend where Taiwanese male students are
more likely than females to involve in violent behavior against students and teachers.
This gender difference is consistent with studies conducted in several Western countries
(Akiba, LeTendre, Baker, & Goesling, 2002; Baldry, 2003; Benbenishty & Astor, 2005;
Dinkes, Cataldi, Kena, & Baum, 2006; Khoury-Kassabri, Astor, & Benbenishty, 2008;
Nansel et al., 2001; Rigby, 2005; Zeira, Astor, & Benbenishty, 2002; 2003).
In addition, the school violence victimization literature shows that gender
differences in rates of victimization increase as students age (Astor et al., 2002). A similar
developmental trend is also found in this dissertation. Studies 1 and 2 indicate that gender
differences in school violence perpetration rates are smaller for elementary schools than
for junior high and high schools. This may imply that preadolescent boys and girls are
more similar in regards to the prevalence of student violence against students and
teachers compared to adolescent boys and girls. These findings may also imply that male
and female students involve in school violence during childhood, but as their ages

142
increase, the size of the group involve in violent acts decreases, with boys more likely
than girls to display violent behaviors (Benbenishty & Astor, 2005).
To further explore gender differences in types of perpetration, the results of Studies
1 and 2 suggest that the ratio of reporting physical perpetration for boys compared to
reporting for girls is greater than the ratios for other types of perpetration. This finding is
consistent with many school violence perpetration and victimization studies, which
suggest that gender patterns distinguish physical forms of violence from verbal and threat
forms (e.g., Benbenishty & Astor, 2005; Knight, Guthrie, Page, & Fabes, 2002). Some
theories point to differences in the friendship patterns of boys and girls as the underlying
reasons for these distinctive forms of violence (Olwens, 2000). For example, boys tend to
form bigger and more amorphous friendship groups, where indirect forms of violence are
not as effective as direct method aggression. Another possibility is that physical forms of
violence are much easier to identify than non-physical types. Thus, girls appear to use
less physical violence than boys (e.g., Benbenishty & Astor, 2005).
However, for student violence against teachers, males and females reported more
similar rates of physical violence compared to other forms of violence. It is possible that
teachers are highly respected in Asian societies, and male and female students perceive
that teachers have much more power than students. Overt or physical forms of violence
against teachers are much easier to identify and are much more likely to result in heavy
punishment. Male and female students shift to involve themselves in covert or
non-physical forms of violence against teachers.

143
Developmental-Trend Findings
The results of Study 1 show that the perpetration rates of student violence against
students increase and peak during early adolescence and then decrease during high school
years. This trend is consistent with reports in the bullying literature and youth violence
studies in Western countries (Espelage & Swearer, 2003; Nansel et al., 2001; Pepler et al.,
2006). The development-related trend is also found in student violence against teachers in
this dissertation. The developmental trend of student violence against teachers is also
consistent with teachers’ reports on victimization in Western countries (Dinkes, Cataldi,
Kena, & Baum, 2006; Lyon & Douglas, 1999). Similar developmental trends between
violence against students and violence against teachers may be related to the stress
associated with the transition to middle school (Benbenishty & Astor, 2005; Pellegrini &
Long, 2002). It is, therefore, important for prevention and intervention efforts to be
directed at the transition from elementary to middle school.
Many empirical studies on bullying suggest that there are developmental trends in
the forms of violence that children use as a function of their advanced skills. For example,
with the development of students’ verbal expression, the proportion of children who use
verbal and indirect forms of aggression increases during early adolescence, but the
proportion of children who use physical aggression declines (Benbenishty & Astor, 2005;
Knight, Guthrie, Page, & Fabes, 2002; Nansel et al., 2001; Sullivan, 2000). The results of
this dissertation indicate that elementary school students report higher rates of physical
violence perpetration compared to junior high and high school students. In contrast,
junior high and high school students report higher rates of verbal and psychological

144
forms of violence compared to elementary school students. The current findings mirror
reports in the bullying perpetration literature.
This dissertation provides information on why student perpetrators commit violent
acts against students and teachers. The results of Studies 1 and 2 suggest that student
perpetrators make different judgments and differ in reasoning when committing violent
acts at school. Most of the reasons given for committing violent acts against students and
teachers are related to provocation, frustration, and teachers’ unfair treatment or
punishment. These findings are similar across gender and school type, and suggest that,
although the prevalence of student violence against students differs from student violence
against teachers, student perpetrators’ major reasons for involvement in school violence
are similar for males and females as well as among the different school types.
Findings regarding the theoretical model
In general, the results of Studies 3, 4, and 5 suggested that within-school variables
(school engagement, school at-risk peers, and student-teacher relationships) mediate the
relation between external-school variables (negative personal traits, parental monitoring,
and prior victimization) and school violence committed by student against other students
and teachers. That is, as students’ negative personal and family experiences increase, they
have lower levels of school engagement and poorer social interaction with students and
teachers on school grounds. In turn, their likelihood of involving in school violence will
increase. These findings support Benbenishty and Astor’s (2005) theoretical model,
which states that school violence is influenced by a combination of numerous within- and
external-school variables, and within-school variables can mediate contributions from

145
external contexts. In contrast with earlier psychological theories which stress single risk
factor contributions of prior victimization, family function, and psychological
characteristics to perpetration, this finding provides strong ecological evidence that
school environment is a unique developmental setting, contributing to school violence
both directly and indirectly (Astor et al., 2002; Benbenishty et al., 2002).
Although school factors can mediate between external-school variables and school
violence in different types of schools, the results of Studies 3, 4, and 5 show that each
school variable plays a different role in mediating the relationship for each school type
and across development. For example, in Study 3, involvement with at-risk peers and
student-teacher relationships can strongly mediate the effects of personal and family
variables on school violence in elementary school. In contrast, school engagement has
almost no mediational effect on school violence in the elementary school setting. In study
4, the mediational effect of involvement with at-risk peers in junior high school becomes
much stronger than in elementary school. However, the mediational effect of the
student-teacher relationship becomes weaker. The mediational effect of school
engagement is not significant in junior high school. In high school, the mediational effect
of involvement with at-risk peers remains as strong as in junior high school, but the
mediational effect of student-teacher relationships is weaker than in junior high schools.
In contrast to in elementary and junior high school, the mediational effect of school
engagement becomes significant in high school. These findings indicate that school peers
play a more important role in mediating school violence as students get older. In contrast,
as students get older, the mediating effects of poor student-teacher relationships become

146
weaker. These findings mirror the current developmental theory which states that school
adults and peers influence student behavior at the stage of pre-adolescence. When
students enter middle school or early adolescence, peer influence becomes more
important than relationships with school adults, such as teachers.
Academic vs. Vocational Schools
Educational tracking based on academic ability accounts for different school
structures and social dynamics between vocationally versus academically-oriented high
schools in Taiwan (Chang, 1992; Gang & Klaus, 2000; Hsieh, 2003). Many educational
practitioners have predicted that the settings of vocational and academic schools mediate
school violence in different ways. Alternatively, some researchers have argued that the
actual mediating mechanism may not vary significantly between the different school
types in different cultures (Astor & Meyer, 2001; Benbenishty & Astor, 2005).
The results of Study 5 show that the within-school variables in vocational and
academic high schools mediate the same external factors in the same way. These findings
suggest that Taiwan’s educational tracking system, which is based on academic ability,
does not account for different students’ patterns of perpetration. These findings may also
imply that the school mediating mechanism may not vary significantly between the
different school types in the same age structure.
Males vs. Females
Many international studies suggest that boys engage in more aggressive behavior in
schools than girls (Baldry, 2003; Nansel et al., 2001; Rigby, 2005). This raises the
question of whether the patterns of relationships between factors and school violence

147
differ across gender. Perhaps what causes perpetration behaviors among boys is the same
as what causes such behaviors among girls, but girls have lower base rates of violence
(e.g., Astor, Benbenishty, Vinokur, & Zeira, 2006; Astor, Benbenishty, Zeira, & Vinokur,
2002; Benbenishty & Astor, 2005; Benbenishty, Astor, Zeira, & Vinokur, 2002).
The results of this dissertation suggest that the overall theoretical model is
applicable for both male and female students. The findings imply that male students
engage in more violence than female students, but the mediating effects of level of school
engagement, school peers, and student-teacher relationships on external-school variables
and school violence are similar across gender.
Implications
This dissertation showed that school violence in Taiwan is a serious problem that
needs the immediate attention of educators and social workers. The study also suggested
that the systematic planning of an intervention program is necessary in order to prevent
serious school violence problems. However, there is no clear policy or complete
intervention program regarding school violence in Taiwan. These findings based on the
nationally representative data including all students from 4th to 12th grades in Taiwan
provide valued information for establishing policy on school violence. Potential school
violence intervention programs may need to target male and junior high school students,
and should pay more attention to decreasing cursing, insulting, teasing, and mocking
behavior among students. School administrators need to advocate the severity of school
violence, and intervention may also focus on promoting the quality of student interaction
with their peers and teachers, because many perpetrators reported that they were involved

148
in violence due to negative social relationships with peers and teachers such as being
provoked by peers and teachers’ punishment.
Furthermore, the finding indicates school environment as a unique developmental
setting to school violence compared with other normative environments such as home
(Astor et al., 2002; Benbenishty, Astor, Zeira & Vinokur, 2002). It implies that improving
student school experience could show a strong effective response to violence. Most
importantly, interventions focusing on decreasing students’ involvement in at-risk peers
may lead to a greater reduction of student violence against students and teachers, because
the results show that the variable of risky peers is the most powerful factor influencing
student violence against students and teachers. However, this does not mean that
interventions should solely focus on school level. To maximize effectiveness of school
violence intervention programs, interventions must integrate personal, family, and school
level approaches (Ahmed & Braithwaite, 2004), because the results of this study show
that the influence of the personal, family and school variables in this model account for a
relatively large amount of the explained variance for student violence against students
and student violence against teachers.
This finding suggests that the prevalence of violence may vary from setting to
setting but the influences of personal, family, and school factors on violence against
students and teachers tend to be similar between genders and across school types and
cultures. This implies school violence intervention incorporated at a national level could
be affective across genders and different types of schools. In addition, the findings

149
support school policy makers or practitioners in developing new international-based
school violence prevention and intervention.
Limitation and Future Studies
The data in this study is only based on student self-reporting, and school violence
was measured in terms of behavior occurring during the prior twelve months. This may
lead students to under- or over- report violent behavior because some of these events are
so common that students might not recall the entire year’s events (Benbenishty & Astor,
2005). Future studies should collect information from multiple sources, such as teachers’
reports, parents, or peer nomination. Hierarchical Linear Model (HLM) may be used to
examine how nested contexts influence perpetration of school violence.
This dissertation explored the prevalence of school violence in Taiwan. It also
examined how within-school variables mediate school violence based on these
within-school dynamics across all schools in Taiwan. However, schools across an entire
country are likely to vary in levels school dynamics and school climate. Future research
need to explore the variation between schools on prevalence of school violence and on
how school variable mediate school violence. This would require different methods of
sampling such as a random sample of schools rather than a random sample of students
(e.g., Benbenishty & Astor, 2005). Such future studies could then examine if school
cultures vary widely or are similar with the variables explored in this study.
This dissertation shows that the age-related trends of the results of the model in this
study were based on cross-sectional data. This data may not be used to establish
developmental progressions and cause-effect among variables. In order to understand

150
more important information on developmental progression of violent behavior in school
and the casual relationships between variables in this model, longitudinal design should
be conducted in the future. In addition, we may use a nested design and longitudinal
studies of students and parents over time to examine the effects of the interaction between
students, family functions and school characteristics as well as the effects of changes in
the school and family overtime (Benbenishity & Astor, 2005).
This dissertation examines different forms of school violence in Taiwan. However,
sexual harassment (Zeira, Astor, & Benbenishty, 2002) and relational aggression (Crick &
Grotpeter, 1995) were not investigated in this study. In order to comprehensively
understand the scope of school violence in Taiwan, future research needs to address these
additional types of violent behavior.
The similarities in gender, school type, and development-related findings between
Taiwan and Western countries may suggest more similarities than differences in school
violence between the East and West. Future studies should be conducted in other Asian
countries in order to further confirm the similarities between the East and West in school
violence findings. For example, it is quite possible that there is great variation in rates of
school violence for Asian countries (within and between them) just as there is great
variation in rates of school violence in European and American cultures or countries.
Future researchers may also apply this theoretical model in this dissertation to other
countries in order to further confirm the similarities or differences across cultures in
school violence findings. In addition, it should be noted that cultural variables often
covary with socioeconomic issues (e.g., Benbenishty & Astor, 2005). In particular

151
poverty and low levels of education have the potential of changing the pattern of
relationships. This sample did not have SES data so these questions could not be pursued.
It is possible that the similar findings in Taiwan and in the West in this dissertation
are attributed to similarities stemming mainly from westernization, industrialization and
globalization rather than cultural variables alone. For example, rural or isolated areas in
Taiwan may have different patterns than urban areas. These differences may be evident in
other Asian cultures as well. The within country and culture variation along economic,
westernization, and socioeconomic dimensions could be an alternative explanation for
similarities between East and West found in this study. Future research should try to
distinguish between direct impact of cultural aspects and the impact of socioeconomic
variables.
Consistent with most Western studies (e.g., Ando, Asakura & Simons-Morton, 2005;
Benbenishty & Astor, 2005), these dissertation showed that involvement in at-risk peers
plays the most important role in school violence in different age groups. Future studies
need to focus on why peer influence is so powerful factor in influencing school violence.
This study did not delve into the reasons why peers influence violent behavior in schools
to the degree that the findings suggest. Are peers modeling behaviors? Are rejected peers
gravitating towards other rejected peers as suggested in some psychological literatures?
These kinds of issues would be important to explore in order to better understand the
mechanisms behind the peer influence. If future interventions are to be developed based
on peer influence, the core reasons explaining the developmental shifts in schools over
time need to be better understood.

152
The findings of how patterns of relationships surrounding the at-risk peer groups’
influences on school violence in this dissertation follow the outcomes of other risky
behaviors such as substance use among students (e.g., Islam & Johnson, 2005; Ritt-Olson
et al., 2005). This may suggest that how the patterns of relationships influence other types
of risky behavior may be similar along a wide array of risk behaviors. Future studies may
include school violence, substance abuse and other types of risky behaviors associated
with at-risk peer groups in the same model. It could help us further understand how
personal, family, and school work together to influence risky behavior and school
violence. In addition, this dissertation examined student personal, family and school
information on school violence. Biological factors such as weight and height should be
included in school violence model in the future research.

153
References

Ahmed, E., & Braithwaite, V. (2004). Bullying and victimization: cause for concern for
both families and schools. Social Psychology of Education, 7, 35-54.

Akiba, M., LeTendre, G . K., Baker, D. P., & Goesling, B. (2002). Student victimization:
National and school system effects on school violence in 37 nations. American
Educational Research Journal, 39, 829–853.

Alikasifoglu, M., Erginoz, E., Ercan, O., Uysal, U., Kaymak, D.A., & Ilter, O.(2004).
Violent behaviour among Turkish high school students and correlate of physical
fighting. European Journal of Public Health, 14 (2), 173-177.

Anderson, A.R., Christenson, S.L., Sinclair, M.F., & Lehr, C.A. (2004). Check & connect:
The importance of relationships for promoting engagement with school. Journal of
School Psychology, 42, 95-113.

Ando, M., Asakura, T., & Simons-Morton, B. (2005). Psychosocial influences on
psychical, verbal, and indirect bullying among Japanese early adolescents. Journal
of Early Adolescence, 25 (3), 268-297.

Astor, R.A. (1994). Children's reasoning about family violence and peer violence: The
role of provocation and retribution. Child Development, 65, 1054-1067.

Astor, R.A. & Behre, W.J. (1997). Violent and nonviolent children’s’ and parents'
reasoning about family and peer violence. Behavioral Disorders, 22, 231-245.

Astor, R.A., Behre, W.J., Fravil, K.A. & Wallace, J.M. (1997). Perceptions of school
violence as a problem and reports of violent events: A national survey of school
social workers. Social Work, 42, 55-68.

Astor, R.A., Benbenishty, R., Haj-Yahia, M. M., Zeira, A., Perkins-Hart, S., Marachi, R.
& Pitner, R.O. (2002). The awareness of risky peer group behaviors on school
grounds as a predictor of students’ victimization on school grounds: Part II: Junior
high schools. Journal of School Violence, 1(3), 57-66.

Astor, R.A., Benbenishty, R., Pitner, R., & Zeira, A. (2004). Bullying and peer
victimization in schools. In P.A. Meares & M.W. Fraser (Eds.), Intervention with
children and adolescents: An interdisciplinary perspective (pp. 471-448). Boston:
Pearson.

Astor, R.A., Benbenishty, R., Vinokur, A. & Zeira, A. (2006). Arab and Jewish
elementary school students’ perceptions of fear and school violence: Understanding
the influence of school context. British Journal of Educational Psychology, 76,
91-118.

154
Astor, R.A., Benbenishty, R., Zeira, A. & Vinokur, A. (2002). School climate, observed
risky behaviors, and victimization as predictors of high school students’ fear and
judgments of school violence as a problem. Health Education and Behavior, 29 (6),
716-736.

Astor, R. A. & Meyer, H.A. (2001). The conceptualization of violence-prone school
subcontexts: Is the sum of parts greater than whole? Urban Education, 36, 374-399.

Astor, R.A., Meyer, H.A., Benbenishty, R., Marachi, R. & Rosemond, M. (2005). School
safety interventions: Best practices and programs. Children & Schools, 24 (1),
17-32.

Baldry, A.C. (2003). Bullying in schools and exposure to domestic violence. Child Abuse
and Neglect, 27, 713-732.

Baldry, A.C. & Farrington, D.P. (2000). Bullies and delinquents: Personal characteristics
and parental styles. Journal of Community and Applied Social Psychology, 10,
17–31.

Baldry, A. & Farrington, D.P. (2005). Protective factors as moderators of risk factors in
adolescence bullying. Social Psychology of Education, 8 (3), 263-284.

Benbenishy, R. & Astor, R.A. (2005). School violence in context: Cultural, neighborhood,
family, school, and gender. New York: Oxford University Press.

Benbenishty, R., Astor, R.A., Zeira, A. & Vinokur, A. (2002). Perceptions of violence and
fear of school attendance among junior high school students’ in Israel. Social Work
Research.26 (2), 71-88.

Bentler, P.M. (1988). Casual modeling via structural equation systems. In J.R. Nesselroad
& R. B. Cattell (Eds.), Handbook of multivariate experimental psychology (2
nd

edition, pp. 317-335). New York: Plenum.

Bentler, P.M. (1992). On the fit of models to covariances and methodology to the Bulletin.
Psychological Bulletin, 112, 400-404.

Bentler, P.M. (1990). Comparative fit indexes in structural models. Psychological
Bulletin, 107, 238-246.

Bentler, P.M., & Bonnett, D.G., (1980). Significance tests and goodness of fit in the
analysis of covariance structures. Psychological Bulletin, 88, 588-606.

Bentley, K.M. & Li, A.K.F. (1995). Bully and victims problems in elementary schools
and students’ beliefs about aggression. Canadian Journal of School Psychology,
11(2), 153-165.

155
Bollen, K. A. (1989). A new incremental fit index for general structural models.
Sociological Methods & Research, 17, 303-316.

Bosworth, K., Espelage, D. L. & Simon, T. R. (1999). Factors associated with bullying
behavior in middle school students. Journal of Early Adolescence, 19, 341-362.

Boulton, M. J., & Underwood, K.(1992). Bully/victim problems among middle school
children. British Journal of Educational Psychology, 62, 73-87.

Brockenbrough, K.K., Cornel, D.G., & Loper, A.B. (2002). Aggressive attitudes among
victims of violence at school. Education and Treatment of Children, 25 (3), 273-287.

Bronfenbrenner, U. (1979). Basic concepts. In U. Bronfenbrenner (Ed.), The ecology of
human development (pp.3-15). Cambridge, MA: Harvard University.

Byrne, B. (1994). Bullies and victims in a school setting with reference to some Dublin
schools. The Irish journal of Psychology, 15, 574-586.

Chang, C.H. (1992). Historical trends in the equality of educational opportunity in
Taiwan. Taiwan Economic Review, 20(1), 23-50.

Chapell, M.S., Hasselman, S.L., Kitchin, T., Lomon, S.N., MacIver, K.W., & Sarullo,
P.L.(2006). Bullying in elementary school, high school, and college. Adolescence,
41 (164), 643-648.

Chen, L.H. (1999). Guo Zhong Xiao Yuan Bao Xing Jiao Shi Bei Hai Zhi Yan Jiu
[Teacher victimization in Taiwanese junior high schools]. Proceedings of Taiwanese
Symposium on Social Problems, Taiwan National Science Council, 1, 493-526.

Chen, L.H. (1989). Guo Zhong Xiao Yuan Xue Sheng Bao Xing Zhi Yan Jiu. [School
violence in Taiwanese junior high schools]. Unpublished master thesis, National
Cheng-Chi University, Taipei, Taiwan.

Chen, X. (2000). Growing up in a collectivistic culture: Socialization and socioemotional
development in Chinese children. In A. L. Comunian & U. P. Gielen (Eds.), Human
development in cross-cultural perspective (pp. 331-353). Lengerich, Germany :
Pabst Science.

Connolly, I. & O’Moore, M. (2003). Personality and family relation of children who bully.
Personality & Individual Differences, 35, 559-567.

Craig,W. M., & Harel, Y. (2004). Bullying, physical fighting and victimization. In C.
Currie, C. Roberts, A. Morgan, R. Smith,W. Settertobulte, O. Samdal,&V. B.
Rasmussen (Eds.), Young people’s health in context: Health behavior in
school-aged children (HBSC) study: International report from the 2001/2002 survey
(pp. 133–144). Copenhagen, Sweden: World Health Organization.

156
Crick, N.R., & Grotpeter, J.K. (1995). Relational aggression, gender, and
social-psychological adjustment. Child Development, 66, 710-722.

Desouza, E.R. & Ribeiro, J. (2005). Bullying and sexual harassment among Brazilian
high school students. Journal of Interpersonal Violence, 20 (9), 1018-1038.

Dinkes, R., Cataldi, E.F., Kena,G., & Baum, K. (2006). Indicators of school crime and
safety: 2006 (NCES 2007-003/NCJ 214262). U.S. Departments of Education and
Justice. Washington, DC: U.S. Government Printing Office.

Espelage, D. L. & Swearer, S. M. (2003). Research on school bullying and victimization:
What have we learned and where do we go from here. School Psychology Review,
32, 365-383.

Flannery, D., Wester, K., & Singer, M. (2004). Impact of Exposure to Violence in School
on Child and Adolescent Mental Health and Behavior. Journal of Community
Psychology, 32, 559-573.

Fraser, M. W. (1996). Aggressive behavior in childhood and early adolescence: An
ecological-developmental perspective on youth violence. Social Work, 41, 347–361.

Fwu, B.J., & Wang, H.H. (2002). The social status of teachers in Taiwan. Comparative
Education, 38(2), 211-224.

Gang, I. N. & Klaus F. Z. (2000). Is child like parent? Educational attainment and ethnic
origin. Journal of Human Recourses, 35(3), 550-569.

Haynie, D.L., Nansel, T., Eitel, P., Crump, A. D., Saylor, K., Yu, K., et al. (2001). Bullies,
victims, and bully/victims: Distinct groups of at-risk youth. Journal of Early
Adolescence, 21, 29-49.

Hsieh, M.Y . (2003). Jia Zhang She Jing Bei Jing Yu Xue Sheng Xue Ye Cheng Jiu Guan
Lian Xing Zhi Yan Jiu [The association between family background and academic
attainment in Taiwan]. Jiao Yu Yan Jiu Ji Kan, 49(2), 255-287.

Hu, L.T., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling: A Multidisciplinary Journal, 6, 1-55.

Hu, S.C. & Lin, K.S. (2001). Xiao Yuan Bao Li Sheng Xing Lv Ji Wei Xian Yin Zi Yan
Jiu: Yi Tainan Shi Guo Zhong Wei Li [Prevalence and risk factors of school violence:
a survey from junior high school students in Tainan]. Xue Xiao Wei Sheng, 38,
32-55.

157
Huang, F.Y. & Hsien, W. Y . (1987). Xiao Yuan Bao Xing Zhi Cheng Yin Yu Fang Zhi Dui
Ce. [Risk factors and intervention for school violence]. Journal of youth delinquency,
5(12), 155-198. Hu, L.T., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in
covariance structure analysis: Conventional criteria versus new alternatives.
Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55.

Huttunen, A., Salmivalli, C., & Lagerspetz, K. (1996). Friendship networks and bullying
in schools. Annals of the New York Academy of Sciences, 794, 355-359.

Islam, S. M. S., & Johnson, C. A. (2005). Influence of known psychosocial smoking risk
factors on Egyptian adolescents' cigarette smoking behavior. Health Promotion
International, 20(2), 135-145.

Johnson, L.C. (1995). Social work practice: A generalist approach. Boston: Allyn and
Bacon.

Junger-Tas, J. (1999) The Netherlands. In P. K. Smith, Y. Morita, J. Junger-Tas, D.
Olweus, R. Catalano, & P. Slee (Eds.), The nature of school bullying: a
cross-national perspective. (pp. 205-223). London: Routledge.

Khoury-Kassabri, M., Astor, R.A. & Benbenishty, R., (in press). Middle Eastern
adolescents’ perpetration of school violence against peers and teachers: A cross
cultural and ecological analysis. Journal of Interpersonal Violence.

Knight, G.P., Guthrie, I.K., Page, M.C., & Fabes, R.A. (2002). Emotional arousal and
gender differences in aggression: A meta-analysis. Aggressive Behavior, 28,
366-393.

Kumpulainen, K., Räsänen, E., Henttonen, I., Almqvist, F., Kresanov, K., Linna, S.L.,
Moilanen, I., Piha, J., Puura, K. & Tamminen, T. (1998). Bullying and psychiatric
symptoms among elementary school-age children. Child Abuse & Neglect, 22,
705-717.

Lin (1986). Guo Zhong Xiao Yuan Bao Xing Zhi Yan Jiu: Taipei shi shi zheng fen zi. [Student
violent behavior in junior high school: empirical evidence in Taipei]. Unpublished
master thesis, National Central Police College, Taipei, Taiwan.

Lyon, D.R. & Douglas, K.S. (1999). Violence against British Columbia teachers.
Vancouver, Canada: Simon Fraser University, Mental Health, Law, and Policy
Institute.

Ma, H. K., Shek, D. T. L., Cheung, P. C., & Tam, K. K. (2002). A longitudinal study of
peer and teacher influences on prosocial and antisocial behavior of Hong Kong
Chinese adolescents. Social Behavior & Personality, 30(2),157-168.

158
Marie-Alsana, W., Haj-Yahia, M.M., & Greenbaum, C.W. (2006). Violence among Arab
elementary school pupils in Israel. Journal of Interpersonal Violence, 21(1), 58-88.

McConville, D.W. & Cornell, D.G. (2003). Aggressive attitudes predict aggressive
behavior in middle school students. Journal of Emotional & Behavioral Disorders,
11(3), 179-187

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. Journal of the American Medical Association, 285,
2094-2100.

Natvig, G. K., Albrektsen, G., & Qvarnstrom, U. (2001). School-related stress experience
as a risk factor for bullying behavior. Journal of Youth & Adolescence, 30, 561-575.

Nisbett, R.E. (2003). The geography of though: How Asian and Westerners think
differently…and why. New York: The Free Press.

O’Keefe, M. (1997). Adolescents’ exposure to community and school violence:
Prevalence and behavioral correlates. Journal of Adolescent Health, 20(5), 368-376.

Olweus, D. (1999). Norway. In P.K. Smith, Y . Morita, J. Junger-Tas, D. Olweus, R.
Catalano, & P. Slee (Eds.), The nature of school bullying: a cross-national
perspective (pp. 28-48). New York, NY: Routledge.

Pateraki, L., & Houndoumadi, A. (2001). Bullying among primary school children in
Athens, Greece. Educational Psychology, 21(2), 167-175.

Pellegrini, A. D. & Long, J. D. (2002). A longitudinal study of bullying, dominance, and
victimization during the transition from primary school through secondary school.
British Journal of Developmental Psychology, 20, 259-280.

Pepler, D.J., Craig, W.M., Conolly, J.A., Yuile, A., McMaster, L., & Jiang, D. (2006). A
developmental perspective on bullying. Aggressive Behavior, 32, 376-384.

Pitner, R., Astor, R.A., Benbenishty, R., Haj-Yahia, M.M. & Zeira, A. (2003a).
Adolescents' approval of peer and spousal retribution in their culture versus other
cultures: The role of group stereotypes. British Journal of Developmental
Psychology, 21(2), 221-242.

Pitner, R., Astor, R.A., Benbenishty, R., Haj-Yahia, M.M. & Zeira, A. (2003b). The
effects of group stereotypes on adolescents' reasoning about peer retribution. Child
Development, 74 (2), 413- 425.

159
Rigby, K. (2005). Why do some children bully at school? The contributions of negative
attitudes towards victims and the perceived expectations of friends, parents and
teachers. School Psychology International, 26, 147-161.

Rigby, K. (1993). School children’s perception of their families and parents as a function
of peer relations. Journal of Genetic Psychology, 154, 501-513.

Rigby, K (1994). Psychosocial functioning in the families of Australian adolescent
schoolchildren involved in bully/victims problems. Journal of family therapy,
16,173-187.

Rigby, K. & Slee, P. T. (1991). Bullying among Australian school children: Reported
behavior and attitudes toward victims. Journal of Social Psychology, 131, 615-627.

Ritt-Olson, A., Unger, J. B., Valente, T., Nezami, E., Chou, C.-P., Trinidad, D. R., et al.
(2005). Exploring peers as a mediator of the association between depression and
smoking in young adolescents. Substance Use & Misuse, 40(1), 77-98.

Schoenhals, M. (1994). Encouraging Talk in Chinese Classrooms. Anthropology &
Education Quarterly, 25(4), 399-412.

Slee, P. T. (1995). Bullying in the playground: The impact of interpersonal violence on
Australian children’s perceptions of their play environment.
Children’sEnvironments, 12, 273-282.

Slee, P. T. & Rigby, K. (1993). The relationship of Eysenck's personality factors and
self-esteem to bully/victim behaviour in Australian school boys. Personality and
individual Differences, 14, 271-373.

Smith, J., Twemlow, S.W., & Hoover, D.W.(1999). Bullies, victims and bystanders: A
method of in-school intervention and possible parental contribution. Child
Psychiatry & Human Development, 30(1), 29-37.

Taiwan Ministry of Education (2006). Education in Taiwan, 2006. Taipei: Author.

Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO:Westview Press.

Turiel, E. (2002). The culture of morality: Social development, context, and conflict.
Cambridge: University Press.

Wei, H.S., Chang, S.W., & Chen, J.K. (2007). Jiu Wu Nian Du Taichung Shi Shao Nian
Sheng Huo Zhuang Kuang Tiao Cha [Youth survey in Taichung city in 2006]. Taipei,
Taiwan: Bureau of Children, Minster of Interior.

160
Wolke, D., Woods, S., Standford, K., & Schulz, H. (2001). Bullying and victimization of
primary school children in England and Germany: Prevalence and school factors.
British Journal of Psychology, 92, 673-696.

Wong, D.S. W. (2004). School Bullying and Tackling Strategies in Hong Kong.
International Journal of Offender Therapy and Comparative Criminology, 48(5),
537-553.

Wu, Jin., Lee, K.C., Yin, M.J. & Hu, S.Z. (2000). Taiwan Di Qu Qing Shao Nian Xiao
Yuan Bao Li Fang Zhi Ji Hua [Prevention and Control of School Violence in
Taiwan]. Taipei, Taiwan: National Science Council. (NSC
89-2420-H-006-001-QBS).

Yang, S.J., Kim, J.M., Kim, S.W., Shin, I.S., & Yoon, J.S. (2006). Bullying and
victimization behaviors in boys and girls at South Korean primary schools. Journal
of the American Academy of Child & Adolescent Psychiatry, 45(1), 69-77.

Yoneyama, S. & Naito, A. (2003). Problems with the Paradigm: The school as a factor in
understanding bullying (with special reference to Japan). British Journal of
Sociology of Education, 24(3), 315-330.

Zeira, A., Astor, R.A. & Benbenishty, R. (2003). School Violence in Israel: Findings of
national survey. Social Work,48, (4), 471-483.

Zeira, A., Astor, R. & Benbenishty, R. (2002). Sexual harassment in Jewish and Arab
public schools in Israel. Child Abuse & Neglect, 26(2), 149-166.

161
Appendix A
Students’ Questionnaire

Demographics
For the purpose of this study, we need a few details about you and your family. Please
note that all your information and responses are kept confidential.
Gender:
1. Male
2. Female
Grade:
1. Grade 4
2. Grade 5
3. Grade 6
4. Grade 1 (Junior high school)
5. Grade 2 (Junior high school)
6. Grade 3 (Junior high school)
7. Grade 1 (Academic high school)
8. Grade 2 (Academic high school)
9. Grade 3 (Academic high school)
10. Grade 1 (V ocational high school)
11. Grade 2 (V ocational high school)
12. Grade 3 (V ocational high school)

162
Appendix A continued

Attitude toward Violence
Please indicate how you usually think of yourself. Select one of the following options:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

If someone insulted me or my family, beating him/her will make me feel better.
If someone disrespects me, I will beat him/her to regain honor
I will feel unhappy if I do not beat someone who cursed me.
A coward is a person who never retaliates when he/she is insulted.
If someone impedes my plans, he/she will pay for it.
If someone tries to hurt me, I will take vengeance on him/her.
Violence is the best way to resolve any problem.
Violence is justice.
If someone makes me unhappy, beating him is what he/she deserves.
Violence is better than negotiation.

163
Appendix A continued

Impulsive Control
Please indicate how you usually think of yourself. Select one of the following option:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

I often make mistakes due to uncontrolled anger.
I often feel regretful about the things I do.
When I feel angry, no one can control me.
I have some bad habit that I always fail to change.
I can not help violating school rules.
I can not tolerate when others look down on me.

164
Appendix A continued

Trait Anger Temperament
Please indicate how you usually think of yourself. Select one of the following options:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

I loose my temper easily.
I am an irritable person.
I am easily agitated.
I feel anger if someone’s mistakes disturb my work.
I often can not control my anger.
When I feel angry, I will curse or use dirty words, such as fuck.
I will feel angry if someone criticizes me.
Beating others will make me feel better if I feel unhappy.
I feel upset if the work is done worse than I expect.
I feel angry if my work is criticized.

165
Appendix A continued

Parental Monitoring (Father monitoring)
Please tell us how your parents know about you. Select one of the following option:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

My father (or father figure) knows my friends.
My father (or father figure) knows my conduct.
If I am not at home, my father (or father figure) will know where I am.
My father knows my schedules.
My father (or father figure) often takes activities with me.

Parental Monitoring (Mother monitoring)
Please tell us how your parents know about you. Select one of the following option:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

My mother (or mother figure) knows my friends.
My mother (or mother figure) knows my personality and general conduct.
If I am not at home, my mother (or mother figure) knows where I am.
My mother knows my schedule.
My mother (or mother figure) often does activities with me.

166
Appendix A continued

Direct Victimization
Please tell us how often you directly experience violent events in your life. Select one of
the following options: 1. Never 2. Sometimes 3. Often 4. Almost Everyday
Have you been beaten or kicked by others?
Has someone hit and hurt you with an object of any kind?
Have you been beaten or kicked by groups of people after class or school?
Have you been blackmailed by others?
Have you been verbally threatened by others?
Have you been intimidated by others?
Has someone threatened or intimidated you by passing a slip of paper?

Witness Victimization
Please tell us how often you witness violent events in your life. Select one of the
following options: 1. Never 2. Sometimes 3. Often 4. Almost Everyday
Have you seen anyone verbally threatened by others?
Have you seen anyone threatened by weapons?
Have you seen anyone destroy public belongings on purpose?
Have you seen anyone beaten by others or participating in group fights?
Have you seen anyone robbed by others?
Have you seen anyone insulted by others?
Have you seen anyone blackmailed by others?

167
Appendix A continued

Level of School Engagement
Please tell us your school information. Select one of the following options:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

I often doze off in class or skip class.
I often forget to bring required materials (textbooks/homework) to school.
I seldom turn in homework.
I often bring prohibited materials to school.

Poor Student-teacher Relationships
Please tell us your school information. Select one of the following options:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

My teacher often punishes me for no reason.
I feel that my teacher does not like me.
My teachers like to mock me.
Even though I tell the truth, my teacher still distrusts me.
Teachers often ask students to observe on what I am doing and snitch on me.

168
Appendix A continued

Involvement in At-risk Peers
Please tell us your school information. Select one of the following options:
1. Strongly Agree 2. Agree 3. Disagree 4. Strongly disagree

I have friends who are school gang members.
I have friends who dropped out due to problem behavior in school.
When I have conflicts with others, my friends will help me beat them or fight with them.
My friends are always on my side no matter what bad things I do.
My friends and I often stay up all nights to drink, to gamble, or to do illegal activities.
My friends and I are often involved in fights or use our fists to protect our territory.

169
Appendix A continued

Student Violence Against Students
In the past twelve months, have you had following experience in school?
Select one of the following options:
1. Never 2. 1-2 times 3. 3-5 times 4. Above 5 times

Beat or kick other students (by group) in order to hurt him/her/them.
Beat or kick other students (by individual) in order to hurt him/her/them.
Use dangerous objects or instruments in order to harm students.
Intentionally destroy or break school public belongings.
Intentionally destroy or break other students’ belongings.
Verbally threaten or intimidate other students.
Curse or insult other students.
Threaten or blackmail other students.
Tease, mock, or play physically harmful tricks on other students on purpose.
Threaten or force others to buy things.

170
Appendix A continued

Student Violence Against Teachers
In the past twelve months, have you had following experience in school?
Select one of the following options:
1. Never 2. 1-2 times 3. 3-5 times 4. Above 5 times

Beat or kick teacher(s).
Use dangerous objects or instruments to harm teacher(s).
Curse or insult teacher(s).
Threaten or intimidate teacher(s).
Extort or blackmail teacher(s).
Tease, mock, or play physically harmful tricks on teacher(s).
Oppose teacher(s) in order to cause them psychological harm.

Abstract (if available)

Abstract To date, very few national studies have been conducted in Asia on school violence issues. In addition, few studies explored how school dynamics, family factors, personal traits work together to contribute to perpetration of school violence. Using a nationally representative sample of 14,042 Taiwanese students from elementary schools (grades 4 to 6), junior high schools (grades 7 to 9), academic high schools and vocational high schools (grades 10 to 12), this dissertation describes the perpetration of school violence in Taiwan. These are the first nationally representative studies on these issues in Taiwan. The first study of this dissertation focuses on the prevalence of student violence against students. The second study focuses on student violence against teachers. The remaindering three studies of the dissertation examine a theoretical model of how school engagement, school risky peers, and student-teacher relationships mediate the effects of personal traits and family factors on school violence committed by students against other students and teachers. Structural equation modeling was used to examine the theoretical model. Three separate studies explored this model on elementary (chapter 3), junior high (chapter 4), and high school (chapter 5) samples respectively. The study on high school students (chapter 5) also explores the model in vocational and academic schools. Compared with Western countries, the prevalence of school violence in Taiwan is high.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The perceptions of cross cultural student violence in an urban school setting

PDF

The role of school climate in the mental health and victimization of students in military-connected schools

PDF

The relationship between school factors and the prevalence of bullying amongst middle school students

PDF

An investigation of students' perceptions and fears before and after transitioning to middle school

PDF

The effects of peer helping on participants' perceptions of school climate, school connectedness, and school violence

PDF

High performing schools in high risk environments: a study on leadership, school safety, and student achievement at two urban middle schools in Los Angeles County

PDF

Organizational structures and systems in high-performing, high-poverty urban schools: the construct of race and teacher expectations as mediating factors in student achievement

PDF

Corporal punishment of students by teachers in elementary and middle schools in Taiwan: the relationship with school level, gender, school location, academic performance, and emotional reactions

PDF

Community violence exposure and children's subsequent rejection within the school peer group: the mediating roles of emotion dysregulation and depressive symptoms

PDF

Examining the longitudinal influence of the physical and social environments on social isolation and cognitive health: contextualizing the role of technology

PDF

Backing each other up "like in basketball": an examination of literacy and the forms of capital among peers in an elementary school classroom community of practice

PDF

Impacts of caregiving on wellbeing among older adults and their spousal caregivers in the United States

PDF

The effects of math anxiety and low self-efficacy on students’ attitudes and interest in STEM

PDF

The effects of physical functioning and public stigma on psychological distress as mediated by cognitive and social factors among Korean survivors of childhood cancer

PDF

A multiple comparison case study of Los Angeles Public High Schools: LGBTQA+ policies and facilities, student advocacy, and change in policies and facilities over time

PDF

Exploring the intersection of occupational engagement and well-being in student musicians: A multi-method study

Asset Metadata

Creator Chen, Ji-Kang (author)

Core Title School social dynamics as mediators of students personal traits and family factors on the perpetration of school violence in Taiwan

School School of Social Work

Degree Doctor of Philosophy

Degree Program Social Work

Publication Date 07/30/2008

Defense Date 05/01/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag bully,cross-culture,OAI-PMH Harvest,risk factors,school violence,Taiwan

Place Name Taiwan (countries)

Language English

Advisor Astor, Ron Avi (committee chair), Chi, Iris (committee member), Johnson, C. Anderson (committee member)

Creator Email jikangch@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1458

Unique identifier UC1167099

Identifier etd-Chen-20080730 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-99321 (legacy record id),usctheses-m1458 (legacy record id)

Legacy Identifier etd-Chen-20080730.pdf

Dmrecord 99321

Document Type Dissertation

Rights Chen, Ji-Kang

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu