AGGRESSIVE BEHAVIOR
Volume 37, pages 234–247 (2011)
Getting Precise and Pragmatic About the Assessment
of Bullying: The Development of the California Bullying
Victimization Scale
Erika D. Felix1!, Jill D. Sharkey1, Jennifer Greif Green2, Michael J. Furlong1, and Diane Tanigawa1
1
2
University of California, Santa Barbara, California
Boston University, Boston, Massachusetts
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
Accurate assessment of bullying is essential to intervention planning and evaluation. Limitations to many currently available selfreport measures of bullying victimization include a lack of psychometric information, use of the emotionally laden term ‘‘bullying’’
in definition-first approaches to self-report surveys, and not assessing all components of the definition of bullying (chronicity,
intentionality, and imbalance of power) in behavioral-based self-report methods. To address these limitations, we developed the
California Bullying Victimization Scale (CBVS), which is a self-report scale that measures the three-part definition of bullying
without the use of the term bully. We examined test–retest reliability and the concurrent and predictive validity of the CBVS across
students in Grades 5–12 in four central California schools. Concurrent validity was assessed by comparing the CBVS with a
common, definition-based bullying victimization measure. Predictive validity was examined through the co-administration of
measures of psychological well-being. Analysis by grade and gender are included. Results support the test–retest reliability of the
CBVS over a 2-week period. The CBVS was significantly, positively correlated with another bullying assessment and was related
in expected directions to measures of well-being. Implications for differentiating peer victimization and bullying victimization via
r 2011 Wiley-Liss, Inc.
self-report measures are discussed. Aggr. Behav. 37:234–247, 2011.
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
Keywords: bullying; victimization; power imbalance; measurement; self-report; psychometrics
INTRODUCTION
Accurate assessment of bullying is essential to
intervention planning and evaluation [Furlong et al.,
2010]; however, there are long-standing concerns
about its measurement [Cornell et al., 2006]. Bullying
is recognized as a subset of peer victimization that is
intentional, chronic, and characterized by an imbalance of power between victim and aggressor [Olweus,
1993]. Researchers worldwide have long struggled to
measure bullying in ways that facilitate cross-national
comparisons and to accurately estimate prevalence
rates. Past efforts have produced equivocal results
with considerable differences of prevalence rates
across studies [Smith et al., 2002], which raises the
issue of whether rates of bullying differ dramatically
across samples or if differences reflect measurement
imprecision. Measurement concerns include: (a)
whether or not to provide an a priori definition of
bullying to respondents [Espelage and Swearer,
2003]; (b) variations in definitions and time frames
used; (c) whether to use self-report, peer nomination,
or teacher-report methods [Cornell et al., 2006;
r 2011 Wiley-Liss, Inc.
Solberg and Olweus, 2003]; and (d) if available
measures actually assess the subset of peer victimization that is intended to be captured by the scientific
definition of bullying [Greif and Furlong, 2006].
The need for a sophisticated bully victimization
assessment tool has prompted scholars to rely on
prominent existing measures for more advanced
purposes than originally intended. Olweus [1996]
developed one of the first and most widely used and
adapted measures. His measure, and similar bullying
assessment instruments, was developed to estimate
prevalence—examine base rates at the school level.
Grant sponsor: United States Office of Juvenile Justice and
Delinquency Prevention via George Washington University.
!Correspondence to: Erika D. Felix, Department of Counseling,
Clinical, and School Psychology, Gevirtz Graduate School of
Education, University of California, Santa Barbara, CA 93106-9490.
E-mail: [email protected]
Received 20 April 2010; Revised 16 September 2010; Accepted 13
January 2011
Published online 14 March 2011 in Wiley Online Library (wiley
onlinelibrary.com). DOI: 10.1002/ab.20389
California Bullying Victimization Scale 235
Scholars have adapted these measures from their
original purpose in order to examine individual
differences through methods of classifying children
by bullying experience and correlating antecedents
and consequences to this classification. Bully assessment tools are also needed to evaluate intervention
programs at the individual- vs. school-level [Greif
and Furlong, 2006]. It is only recently that scales
developed for epidemiological purposes have been
evaluated for the psychometrics needed for individual differences research [e.g., Kyriakides et al.,
2006; Lee and Cornell, 2010], but with limited,
unique samples. Thus, assessing the psychometrics of
self-report bullying victimization measures remains
crucial for research and applied practice to continue
to advance. In this article, we describe the development and psychometric properties of the California
Bullying Victimization Scale (CBVS). We begin by
reviewing the state of self-report bullying assessment,
different approaches within self-report, and the need
for further refinement of self-report assessment.
Self-Report Bullying Measurement Approaches
Self-report assessments are the most common
method to measure bullying victimization, in part
because they are easier and less costly for researchers
and educators to implement than other methods,
such as intensive behavioral observations of schoolyard interactions. Researchers have debated the
relative benefits of self-report vs. other methods.
Solberg and Olweus [2003] argued that self-report is
the best method for ascertaining prevalence estimates, which is information that educators and
policy makers often need. On the other hand,
Cornell and colleagues [Cornell et al., 2006; Lee
and Cornell, 2010] pointed out that many available
measures do not have adequate reliability and
validity. For example, Henry [2006] found that
self-reports of aggressive behavior lacked criterionrelated validity, whereas teacher- and peer-report
methods were significantly correlated to observations of aggressive behavior. However, research has
not demonstrated a relative lack of validity for selfreports of victimization, especially for indirect forms
of victimization that are not readily observable.
Thus, self-reports remain a viable way to assess the
prevalence of bullying victimization and further
research is clearly warranted.
Considering concerns about the reliability and
validity of bullying self-reports, there have been calls
to use other assessment procedures such as peer
nominations [e.g., Chan, 2006; Chan et al., 2005]
and behavioral observations [e.g., Cornell and
Brockenbrough, 2004; Craig et al., 2000; Hawkins
et al., 2001]. Studies using these methodologies have
provided valuable data to answer questions that
could not be addressed by self-report data alone.
However, there could be ethical and/or logistical
challenges with peer nominations and behavioral
observations that make it more difficult to obtain
research review board or school approval and
written guardian consent [Espelage and Swearer,
2003]. For example, in playground observations,
researchers face an ethical dilemma when they
witness victimization as it would be unethical to
observe a physical attack without intervening. Likewise with peer nominations, if parents do not consent
for their children to participate, researchers may not
collect peer nominations of their sociometric statuses. A single missing participant could affect the
validity of all classroom or grade-level reports.
Although researchers have successfully negotiated
these potential dilemmas to produce valuable research, these considerations may be limiting factors
in some schools, communities, and universities,
making self-reports a favorable alternative.
In fact, self-report has been the primary approach
used by researchers. Hence, it is critical to examine
the status and quality of bullying self-report
measures [see Furlong et al., 2010 for a review].
The self-report measures used by researchers generally either (a) provide a working definition of
bullying and ask a youth if they had experienced this
type of victimization or (b) present a list of
victimization-related behaviors and ask how often
the youth has experienced or committed them. We
will review both approaches.
Definition-based self-report strategy. Assessments of bullying and bullying victimization
often provide a definition of bullying to acquaint
students to the idea of bullying as defined by the
researchers, with the intent of establishing a shared
meaning of the bullying experience across all study
participants. Considerable effort has been devoted to
developing a consensus on how to define and
measure bullying [e.g., Arora, 1996; Naylor et al.,
2006; Smith et al., 2002]. Smith et al. [2002] point out
that the issue of definition is crucial for accurate
reporting of bullying prevalence cross-nationally and
that the word ‘‘bully’’ is not easy to translate. To
overcome terminology challenges, they used 25 stickfigure drawings depicting several types of aggressive
experiences to assess the words used to describe these
experiences across a 14-nation sample. They found
significant variation in the number of drawings
assigned to each term, demonstrating different
victimization scenarios implied by each term.
Aggr. Behav.
236 Felix et al.
Researchers have debated the merits of defining
bullying before asking students’ about their experiences. The definition of bullying has evolved over
the past 30 years to include a broad range of
victimization; hence, studies have varied in measurement practices. These practices have led to varying
prevalence estimates and consequently it is difficult
to evaluate the nature and extent of cross-sample
comparability. Also, studies have repeatedly shown
that different informants have different views of the
definition. Older youth have a more nuanced and
complex view of bullying than younger children
[Smith et al., 2002]. Similarly, teachers’ definitions of
bullying tend to correspond more to the three-part
definition of intentionality, repetition, and imbalance of power than students [Menesini et al., 2002].
In Europe, a workgroup reviewed self-report and
peer nomination bullying assessment methods in an
effort to improve bullying intervention [Ortega et al.,
2001]. In a comprehensive review, they assessed the
content, items, age ranges, and advantages and
disadvantages for both self-report and peer nomination methods to propose core questions for a general
survey questionnaire. They concluded that the ideal
time frame to explore is the last 3 months, and they
recommended using a definition, with cartoons to
illustrate examples of physical, verbal, and indirect
bullying. However, it is not clear the extent to which
these practices have been adopted.
Even if definitional consensus were possible,
asking a youth to label himself or herself as a victim
or a bully may provoke emotional reactions that
could influence victims or bullies to not endorse
experiences associated with the label [Cornell and
Brockenbrough, 2004; Greif and Furlong, 2006;
Hamby and Finkelhor, 2000]. One study found that
when students are provided with a definition of
bullying, and are repeatedly exposed to the term
‘‘bully,’’ they report significantly less bullying than
those not exposed to a definition or the word
‘‘bully’’ [Kert et al., 2010]. In this circumstance,
bullying prevalence would be underestimated and
intervention outcomes would be affected by imprecise measurement. Likewise, in studies of adult
sexual harassment, participants commonly endorse
the critical behavioral experiences associated with
the legal definition of sexual harassment, but at the
end of the questionnaire, when asked if they had
been ‘‘sexually harassed’’ many denied it, perhaps
not identifying with the label [Cortina et al., 1998;
Fitzgerald and Ormerod, 1991].
Others argue that using a definition-first strategy
that includes multiple forms of aggression [e.g.,
presenting a definition listing various types of
Aggr. Behav.
aggression and then asking the youth if he or she
has been bullied) may produce heterogeneous data
that mask trends and correlations among subtypes
of bullying experiences [Cornell et al., 2006].
A broad definition does not allow for assessing the
relative contribution of each form of bullying
victimization to psychosocial adjustment [Cornell
et al., 2006; Felix and McMahon, 2006]. Even if
more than one question about victimization is asked
following a bullying definition, it is unclear how well
students remember long, multipart definitions when
responding to a series of items. Smith et al. [2002]
point out that even if a student is initially given an
extended definition of bullying, the term ‘‘bullying’’
is used in most of the following questions, and
children may use their natural understanding of this
term in their culture, rather than the originally
presented definition. Menesini et al. [2002] noted
that students in their sample were able to take into
consideration repetition, intention, and imbalance of
power, but their measure used visual aides in the
form of cartoons, which most definition-first
approaches do not do. Given these concerns, some
researchers have adopted a behavior-based strategy.
Behavior-based self-report strategy. In a
behavior-based approach, specific bully-related behaviors are presented (e.g., hitting, threatening, or
spreading rumors) and youth are asked to specify if
they have committed or experienced them without
using the term bullying. The behavioral method
breaks down victimization into specific experiences
in order to avoid individual perceptions, stigma, or
bias associated with using the term ‘‘bullying’’ or
‘‘victim’’ in the measure itself, and to allow
researchers to examine the frequency of each type
of victimization. However, this approach does not
tend to assess for an imbalance of power, and thus
fighting between peers of equal strength, which is
considered to be aggression but not bullying, would
likely be included [Smith et al., 2002].
Behavioral strategies generally classify the group
with the highest level of victimization experiences as
‘‘victims’’ and those with the highest levels of
perpetration as ‘‘bullies’’ [Espelage and Swearer,
2003]. Some issues associated with this approach
are: (a) the criteria used to select extreme responders
(e.g., one standard deviation above the mean, less,
more) vary across studies [Solberg and Olweus,
2003]; (b) educators conducting school-based surveys are unlikely to be able to complete the complex
computations needed to classify students with
statistical methods; and (c) unless additional items
are coadministered, the behavioral method assesses
only one aspect of the definition of bullying, which is
California Bullying Victimization Scale 237
the repetitive nature of the experience. This neglects
intentionality and imbalance of power, which also
differentiate bullying from general peer victimization. Although it may be argued that if a child is
repeatedly victimized, there must be an imbalance of
power, as the child is unable to stop the victimization, this is not always the case. For example, two
boys of the same strength and social standing may
dislike each other and repeatedly fight, but this
would not traditionally be considered bullying.
Advantages of Self-Report Strategies
and Moving Forward
Self-reports address many complex issues of
understanding bullying behavior. First, self-report
procedures are efficient and can be conducted in
school, community, and laboratory settings with
minimal cost. Second, self-reports can assess the
diverse and important subtypes of bullying behavior, including direct, indirect, and relational
aggression. This includes behaviors that are not
readily observable by others. Third, they have the
potential to assess the perceived power imbalance
from the perspective of the bully, victim, or bullyvictim. Fourth, they do not rely on consent from
other participants (as in peer nomination procedures). Fifth, they assess the perspective of the
participant (both bully and victim), which is critical
when attempting to understand intentionality and
impacts on psychosocial functioning. Finally, selfreport can simultaneously assess and distinguish
between different forms of bullying behavior to
better understand differential impact.
Although there are advantages to using self-report,
there are also unresolved issues to be addressed in
order to move research and evaluation of prevention
programs forward. Bullying is a subset of peer
victimization [Björkvist et al., 1982] and a challenge
of bullying assessment methods is to distinguish
bullying from other types of peer victimization
including playful behavior [Cornell et al., 2006].
Indeed, some researchers have gone beyond the bully,
victim, bully-victim, and nonvictim classification to
explore the roles of all students in bullying encounters
[e.g., Salmivalli, 1998, 1999, 2010]. Salmivalli [1998]
articulated multiple participant roles, in addition to
bully and victim, which included the assistant to the
bully, the reinforcer, the defender, and the outsider.
As many have noted, friends or acquaintances can
engage in teasing and horseplay that may look like
bullying to an outside observer [Cornell et al., 2006;
Rigby, 2004]. Given various bully victimization roles,
it is critical to evaluate the importance of nuances in
types of peer victimization to accurately understand
the impact of these diverse experiences.
Research supports the important distinction
between the larger group of peer victims and the
subset of bully victims [You et al., 2008]. Thus, there
is a need to explore alternative strategies to assess
power imbalance, one of the three key elements of
bullying. Namely, youth who have experienced
repeated bullying have worse psychological health
than youth who have experienced peer victimization
that was not bullying. Thus, relying solely on the
repetitive nature of victimization may be inadequate
to truly understanding the bully experience and
research needs to implement and test methods of
assessing the power imbalance.
Lastly, researchers vary on defining chronicity.
Studies using a behavioral approach have identified
extreme responders to classify ‘‘bullies’’ or ‘‘victims’’
[Espelage and Swearer, 2003], which necessitates a
cumbersome statistical procedure. However, Solberg
and Olweus [2003] argued that using a definition of
bullying and the empirically derived frequency criteria
of ‘‘2–3 times per month’’ or more derives the best
estimate of bullying victimization based on its relation
with negative conditions such as depression. As there
are advantages and disadvantages to both definitionfirst and behavioral approaches, and there is a need for
identifying the psychometric properties of a self-report
bullying assessment specifically designed for individual
differences research, we developed the CBVS aiming
to address current needs in bullying assessment.
California Bullying Victimization Survey
The CBVS was developed as a self-report measure
of multiple forms of bullying victimization, and to
distinguish bullying from peer victimization, without
the use of a definition. The CBVS avoids the label
‘‘bully,’’ on the premise that this term may be
emotionally laden to respondents, and influence
victims or bullies not to endorse experiences associated with the label [Cornell and Brockenbrough,
2004; Greif and Furlong, 2006]. As one method to
differentiate bully victims from peer victims, we
specify in each question that the behavior must be
done on purpose in a mean and hurtful way.
Anecdotally, we observed in student interviews that
some children who initially endorsed a behavioral
description later stated that it was not done on
purpose in a mean or hurtful way. Thus, without
this specifier, prevalence rates may be overestimated.
The aggressor and victim may have different
opinions as the aggressor may deny that he or she
intended to hurt the other person or may minimize
Aggr. Behav.
238 Felix et al.
the harm done. We took the perspective of the
victims, as ultimately it is their appraisal of the
situation that will likely affect their well-being.1
Another way to differentiate bullying from peer
victimization is through the presence of a power
imbalance. Without using a definition, the CBVS
was designed to assess whether the reporter perceives a power imbalance between self and the
aggressor. We assess power imbalance in terms of
how popular, smart, and strong was the person who
did mean things on purpose to the respondent.
These power imbalance options are similar to those
in the Swearer Bullying Survey [Swearer, 2001].
An additional measurement consideration addressed by the CBVS is the time frame and
frequency scale. Given that the accuracy of recall
diminishes with time, the CBVS uses a past 30-day
time frame and the same frequency scale used in the
Olweus Questionnaire, which allows comparison
across studies using his measure. The Olweus
frequency scale classifies victims as those experiencing bullying 2–3 times a month or more, a cutoff
that is easy for schools to use and understand.
Study Purpose
The purpose of this study is to examine the
psychometric properties of the CBVS for use with
individual students to determine its functioning as a
new measure of bullying victimization. First we
examine item-level test–retest reliability of the CBVS
victimization items. We subsequently use the CBVS
to classify students as nonvictims, peer victims, or
bully victims and examine the test–retest classification
reliability of these categories. We next turn to
measure validity and compare the CBVS classification of bully victims to respondents who are classified
as bully victims using the ‘‘definition-first approach.’’
Finally, we examine the predictive validity of the
CBVS to determine the extent to which classification
on this measure is associated with theoretically
related constructs and measures of psychological
well-being. As there are gender differences in the type
of victimization girls and boys experience most
frequently [e.g., Crick and Grotpeter, 1995; Olweus,
1993], initial analysis was by gender. In addition,
given the developmental range covered in this study,
analyses were also separated by grade groupings (i.e.,
elementary, middle school, high school).
1
Research on sexual harassment notes a significant difference in
opinion between the perpetrator and victim on whether a behavior
constituted harassment and legally, courts uphold the view that it is
not the intent of the perpetrator that matters, but rather the effect on
the victim [Paludi, 1997].
Aggr. Behav.
METHOD
Preliminary Development of CBVS
In response to the need for improved self-report
assessments, as well as a request for help with a
school-wide screening by a local junior high school,
we created a bullying victimization survey that
served as the basis for the CBVS. This initial draft
of the survey was administered to 463 junior high
school students of whom 329 were in seventh grade
(71%) and 134 were in eighth grade (29%) in June
2005. The sample was 55% White, 26% Latino/a,
and 19% representing other ethnic groups, and was
evenly divided on gender (51.6% female). To further
develop the survey, we analyzed student responses to
items for any inconsistent responding and queried
students for qualitative feedback on item content.
We delete revised items to increase clarity and added
an item on sexual harassment for students in Grade 7
or higher, in recognition that it is a common form of
peer victimization starting around puberty [Felix and
McMahon, 2006, 2007]. Subsequently, our research
team thoroughly reviewed several drafts of the CBVS
for consensus on item content, wording, and layout.
To further refine the CBVS, focus groups were
subsequently conducted at junior high and elementary school levels to review a draft of the survey and
provide feedback on item wording and clarity. At
the junior high school level, we recruited two
classrooms at a midsize school with diverse socioeconomic status (37% disadvantaged) and primarily
European American (42%) and Latino/a American
(40%) ethnic background: an eighth-grade leadership class and a seventh-grade general science class.
At the elementary level, we selected the single fifth/
sixth-grade classroom at a small school that serves
primarily Latino/a American (84%) and socioeconomically disadvantaged youth (74%). Consent
forms were sent home to parents at both schools
with over 90% positive consent across classrooms.
All students present on the day of the focus group
participated for one class period, which was 50 min.
More females than males participated, and the
groups were roughly evenly divided between grades,
with the exception of a lower number of sixth
graders. One of the principal investigators and
a graduate student research assistant conducted
each focus group. Students completed an item-byitem review of the CBVS and were asked about the
clarity of instructions and item wording. The
research assistant documented feedback on each
item and notes on each focus group were reviewed
by the research team and common themes were
identified. For example, students questioned the
California Bullying Victimization Scale 239
term ‘‘washroom’’ and suggested ‘‘bathroom’’ instead. The survey wording was changed where there
was consistent feedback or a consensus between
researchers, and the research team approved a final
version.
In October 2006, two focus groups were conducted with a convenience sample of high school
students in their elective class to review the wording
of the survey to make it appropriate for an older
student population. All students present for that
class on the day of the focus group participated. The
student population for the school is 53% Latino/a,
42% Caucasian, and the remaining is a mix of other
ethnicities. Approximately 27% of students were
considered socioeconomically disadvantaged and
38% were English Learners. One of the principal
investigators and a graduate student research
assistant conducted each focus group. Students were
led on an item-by-item review of the CBVS and
asked if the instructions and item wording were
understandable. No wording changes were suggested for the items assessing victimization.
administration. For the K-8 school and the elementary school, all students who initially participated
and who were present at school on the day of the
second administration, retook the questionnaire.
For the junior high, one class period was randomly
selected for each teacher to offer the Time 2
questionnaire. At Time 2, there were 43% boys,
33% fifth graders, 27% sixth graders, 20% seventh
graders, and 20% eighth graders. The ethnic
composition was 46% White, 38% Hispanic/
Latino(a), and 16% other or mixed ethnicity.
Sample 2. In study 2, we extended our descriptive study and exploration of the validity of the CBVS
to a high school sample. One high school in the central
coast region participated, and a total of 354 students
provided valid surveys, representing 14% of the total
student population. Over half were female (53%). For
grade representation, there were 39% in ninth grade,
22% in tenth grade, 22% in eleventh grade, and 17%
in twelfth grade. The ethnic composition was 54%
White, 34% Hispanic/Latino(a), and 12% representing other or mixed ethnicities.
Participants
Measures
Criteria for retention in the study. We
reviewed the data to find cases that had such high
levels of missing data that it was not acceptable to
retain them in the study. Using the following criteria,
we maintained 92% of all participants in the final
dataset: (a) no more than five items missing from the
validity scales and (b) at least one victimization item
completed. Although the complete analysis required
that the youth complete all victimization items, we
retained cases with incomplete victimization responses for the item-by-item test–retest analysis.
For this reason, the number of cases varies in the
analyses shown in the following sections.
Sample 1. Three schools in the central coast
region of California participated in the survey. One
school was a junior high school, one was a K-8
school, and one was a K-6 elementary school.
A total of 330 students had valid responses: 46%
boys, 14% fifth graders, 12% sixth graders, 40%
seventh graders, and 34% eighth graders. The ethnic
composition of the students was 51% White, 29%
Hispanic/Latino(a), and 20% other or mixed
ethnicity. Institutional Review Board concerns for
protection of human subjects prevented us from
gathering more detailed ethnicity data for minority
students given the low numbers of students of
alternate ethnic groups.
We readministered the survey to a subsample
(n 5 131) of the students 2 weeks from the original
A single questionnaire was created and titled, My
Experiences with Classmates at School. After brief
instructions, the questionnaire consisted of the
following scales in the order described.
California Bully Victimization Sale. The final
version of the CBVS includes items asking about six
(elementary version) to seven (secondary version)
forms of victimization that respondents may have
experienced at school: been teased or called names by
another student; rumors spread or gossip behind
someone’s back; left out of a group or ignored; hit,
pushed, or physically hurt; been threatened; had
your things stolen or damaged; or had sexual
comments or gestures directed at them (this question
is only included in the secondary school versions).2
Students are asked to rate the frequency of each of
these experiences on a five-point scale (0 5 Never,
1 5 once in the past month, 2 5 2 or 3 times in the past
2
For Study 2, we also piloted a question on cyber-victimization that
was not originally included in Study 1 with the middle school
students. We found 11% of the high school students we sampled
reported some cyber-victimization. We decided not to include that
one item in this study because (a) it was only in Study 2, whereas
sexual harassment was assessed in both studies (Grade 7 or higher),
and (b) it seems to be more of a means of victimization than a type,
in our estimation. For example, texting can be considered akin to
passing mean notes; threatening a fight via a social network sites is
similar to threatening via ‘‘friends’’ or other associates. Other
research has supported that cyber-bullying is more of a modality
than a distinct type [e.g., Varjas et al., 2009].
Aggr. Behav.
240 Felix et al.
month, 3 5 about once a week, and 4 5 several times a
week). In the next question, respondents are asked to
indicate if these behaviors were intentional and done
in a mean way. They select from among the
following responses in reference to the previously
rated victimization experiences broadly: They were
almost never mean ( just joking), they were sometimes
mean, they were almost always mean. Next, the CBVS
assesses power imbalance through a series of items
asking students to compare themselves to the ‘‘main
person who did these things to you.’’ Students rated
responded to a three-point scale (less than me, same
as me, more than me) on how popular, smart in
schoolwork, and physically strong the other person is
in comparison to themselves. These choices are
similar to the ones Swearer [2001] used in her survey.
The CBVS includes several other questions not used
in this study that are designed to guide intervention
planning including types of bullying witnessed (as a
bystander), information about students doing the
bullying, where on the school campus bullying
occurs, when during the school day it is most
frequent, and who students talk with about bullying.
Swearer Bullying Survey. The CBVS was
compared with an item from the Swearer [2001]
Bullying Survey that assessed bullying by providing
a definition and asking students how frequently they
have been victimized. The definition states: ‘‘Bullying
happens when someone hurts or scares another
person on purpose and the person being bullied has
a hard time defending himself or herself. Usually,
bullying happens over and over.’’ The following are
listed as possible examples of bullying behavior:
punching, shoving and other acts that hurt people
physically; spreading bad rumors about people;
keeping certain people out of a ‘‘group’’; teasing
people in a mean way; and getting certain people to
‘‘gang up’’ on others. Students are asked to indicate
if they were bullied this month and if yes, how often
(i.e., once in the past month, 2– 3 times in the past
month, once a week, several times a week).
Students’ Life Satisfaction Scale (SLSS). This
seven-item measure uses a six-point Likert scale
(strongly disagree to strongly agree) to assess overall
well-being (e.g., ‘‘My life is going well’’ and ‘‘My life is
just right’’) for students ages 8–18 years old. According
to prior research, internal consistency ranges between
.73 and .86 [Huebner et al., 2005] and test–retest
reliability is .76 across one to two weeks [Terry and
Huebner, 1995]. Correlations of the SLSS with other
life satisfaction scales meet expectations [Huebner,
1991], and studies support its construct, discriminant,
and predictive validity [Huebner et al., 2005]. For this
study, the a coefficient was .88.
Aggr. Behav.
School Connectedness Scale (SCS). The
SCS measures the bond felt by students toward their
school and the quality of the relationship between
students and teachers [McNeely, 2005]. Responses
are measured using a five-point Likert scale (strongly
disagree to strongly agree). This scale was constructed out of items originally included in the
National Longitudinal Study of Adolescent Health.
McNeely [2005] noted that three versions of the SCS
have been used from the NLSAH Study. The version
used in this study is the one employed by McMeely
et al. [2002], and administered in the Resilience
Youth Development Module of the California
Healthy Kids Survey (see WestEd; www.wested.
org/hks). An a of .81 was obtained in this study.
Children’s Hope Scale (CHS). The six-item
CHS is designed for children between the ages of 7
and 15 years. The CHS measures two aspects of
hope—the cognitive capacity to formulate plans to
achieve set goals (three items; e.g., ‘‘I can think of
many ways to get the things in life that are most
important to me’’), and self-efficacy, the belief that
one can achieve set goals through effort (three items;
e.g., ‘‘I think I am doing well’’). Children respond to
the items using a six-point scale (1 5 none of the time
to 6 5 all of the time). Both the internal consistency
and the test–retest reliability are greater than .70
[Snyder, 2005]. The CHS possesses strong concurrent
validity and acceptable predictive and discriminant
validity [Snyder, 2005]. The a for this study was .88.
Procedures
Sample 1. Schools participated in the survey in
May and June of 2006 after obtaining Institutional
Review Board approval. At the junior high school,
two social studies teachers agreed to administer the
survey to their students in multiple classes. At the
K-8 and elementary schools, all fifth- and sixthgrade students were given the opportunity to
participate. Teachers were provided with detailed
written instructions for administering the survey by
the research team. Teachers administered the survey
in their classrooms to students with parental consent
and student assent. Students were not asked to
provide their names; rather, random numbers were
used to protect their identity. For test–retest
reliability analysis, students received an envelope
with two surveys matched by their random identification number. Students wrote their name on the
envelope and approximately 2 weeks later, teachers
readministered the survey (retest subsample). Students turned in the envelopes with the matched
surveys and the envelopes were destroyed.
California Bullying Victimization Scale 241
Sample 2. In November 2006, teachers administered the survey in a paper-and-pencil format
during one class period to all students in the high
school who provided parental consent and student
assent. A detailed instruction sheet was provided for
teachers to administer the survey.
RESULTS
Rates of Peer Victimization
Descriptive statistics were calculated to determine
the percentage of boys and girls at each school level
(elementary, junior high, high school) who experienced each type of victimization with the frequency
criteria set at 2–3 times per month or more. Chisquare tests of significance revealed if rates were
different by gender. Results indicate that peer
victimization is a common experience at school
(Table I). Teasing was the most frequent form of
victimization reported by male and female students
across grade levels. Females reported higher rates of
victimization through rumors and social exclusion,
as compared with males. Males in Grades 7–12
reported that they were hit significantly more often
than females, although the opposite trend was found
in Grades 5 and 6. In general, there was a decline in
frequency of reported victimization from elementary
to high school.
Students were also classified as nonvictims, peer
victims, and bullied victims. Nonvictims were
students who reported no victimization experiences
on the CBVS. Peer victims reported at least one
victimization experience of any frequency, but no
power imbalance. Bully victims were students who
reported at least one type of victimization at least
2–3 times per month, reported that this victimization
was intentional at least some of the time, and
reported at least one form of power imbalance in
relation to the main person bullying them (Table II).
A hierarchical log-linear analysis with backward
stepwise model selection was conducted with SPSS
18.0 to test the higher-order relations between the
three classification variables of gender, age, and
victimization status. In the backwards elimination
steps, the first interaction eliminated was for gender
by grade, w2 5 0.71, df 5 2, P 5 .70, followed by the
three-way interaction, w2 5 4.64, df 5 6, P 5 .59, and
finally the gender by victimization interaction,
w2 5 4.35, df 5 2, P 5 .11. There was a remaining
significant interaction for grade by victimization,
w2 5 163.41, df 5 4, Po.001. The percentages of
students categorized as peer and bullied victims
declined with age; this decrease was most pronounced for the general class of peer victims, which
decreased from 45.3% and 52.2% among elementary school and junior high school students,
respectively, which compared with only 11.3%
reported by high school students. Almost threefourths of high schools students (73.4%) were in the
nonvictim group, compared with 28.2% of junior
high and 31.0% of elementary school students. The
decline in bullied victims was smaller across the
grade levels: elementary school (23.8%), junior high
school (19.6%), and high school (15.3%).
Test–Retest Stability
Two-week stability of the CBVS was evaluated to
determine test–retest reliability in a subsample of
students (n 5 131). First, as one test of response
stability, a total victimization score was calculated
as a sum of the six core victimization items (all but
sexual harassment). Correlations between the total
victimiation scores at time 1 and time 2 indicated a
high degree of agreement across points in time
TABLE I. Item-Level Descriptive Statistics of Victimization Experienced 2–3 Times or More per Month by Males (M) and Females (F)
Grade
5–6
7–8
9–12
M (n 5 36) (%)
F (n 5 48) (%)
M (n 5 114) (%)
F (n 5 130) (%)
M (n 5 167) (%)
F (n 5 187) (%)
22.2
16.7
8.3
8.3
13.9
11.1
38.3
25.0
37.5!!
18.8
14.6
12.5s
a
a
26.3
9.6
7.0
14.9
14.0
8.8
20.2
26.9
19.2!
16.9!
6.2!
8.5
9.3
18.5
18.0
9.6
6.0
12.0
4.8
4.2
9.0
14.4
13.9
11.2
4.8
3.2
2.1
12.3
Teased
Rumors
Ignored
Hit
Threatened
Property
Sexual comments
!Po.05; !!Po.01 (w2).
a
Not asked of elementary school participants.
Aggr. Behav.
242 Felix et al.
TABLE II. Percentages of Males (M) and Females (F) by Grade Level in Each Victimization Category at Time 1
Grades 5–6
Grades 7–8
Grades 9–12
Victim group
M (n 5 36) (%)
F (n 5 48) (%)
M (n 5 114) (%)
F (n 5 130) (%)
M (n 5 167) (%)
F (n 5 187) (%)
Not a victim
Peer victim
Bullied victim
39
39
22
25
50
25
26
58
16
30
47
23
74
14
12
73
9
18
TABLE III. Test–Retest Reliability: Kappa Agreement
Between Victimization Items (2–3 " /month or more) at Time 1
and Time 2 (after 2 Weeks) for Grades 5–8 (n 5 131)
Scale
Teased
Rumors
Ignored
Hit
Threatened
Property
Sexual comments
Time 1
k
.582!!!
.510!!!
.553!!!
.619!!!
.636!!!
.620!!!
.461!!
!!Po.01; !!!Po.001.
(r 5 .80 for Grades 5–6, r 5 .83 for Grades 7–8, for
both Po.001). Next, Cohen’s kappa coefficient was
examined for each item (Table III). For individual
types of victimization, Cohen’s ks ranged from .46
to .64, and were all significant, indicating that
agreement exceeds chance levels across victimization
experiences. As a final analysis of test–retest
stability, we examined the consistency in classifications of students as nonbullied and bullied across
two points in time (percent agreement 5 89.6,
k 5 .71). The classification table is presented in
Table IV. The majority of students (79%) were
classified as nonbullied at both points in time and
13% of students were classified as bullied at both
assessments. A total of 8% of youth reports
indicated a change of status with two thirds of these
youth changing from the bullied group at time 1 to
the nonbullied group at time 2.
Concurrent Validity
Evidence for the concurrent validity of the CBVS
was evaluated in comparison to a ‘‘definition-first’’
approach assessing bullying using the Swearer
Survey [Swearer, 2001; Table V]. The rates of
students classified as being bullied were relatively
similar in the definition-first measure (20.2% of
fifth/sixth graders, 17.5% of seventh/eighth graders)
and the CBVS (23.8% of fifth/sixth graders, 20.0%
of seventh/eighth graders). However, there were
substantial differences in exactly which students
were identified as being bullied by the two measures
Aggr. Behav.
TABLE IV. Reliability Analysis: Test–Retest Reliability of
CBVS Bullied Victimization Classification (n 5 131)
Time 2 (2 weeks later)
Not bullied
Bullied
103
4
7
17
Not Bullied
Bullied
Percent agreement: 89.6%; k: .71; w2: 65.6, Po.001.
TABLE V. Concurrent Validity: Comparison of Bullied
Victimization Classification Using CBVS Versus
Definition-First Strategy
CBVS classification
Definition-first
bully
Grades 5–6 (n 5 84)
Grades 7–8 (n 5 245)
item
classification
Not bullied
(%)
Bullied
(%)
Not bullied
(%)
Bullied
(%)
Not Bullied
Bullied
66.7
9.5
13.1
10.7
73.5
6.5
9.0
11.0
Note. Grades 5–6: Although there was a 77.4% agreement rate
between the two measures, there was only 24% concordance about
who experienced bullying; k 5 .343; w2 5 9.97, Po.01. Grades 7–8:
Although there was an 84.5% agreement rate between the two
measures, there was only 29% concordance about who experienced
bullying; k 5 .492; w2 5 59.69, Po.001.
(k 5 .34 for fifth/sixth graders, k 5 .49 for seventh/
eighth graders). This indicates that the different
measurement strategies significantly influenced classification status of individual students (i.e., they
identified somewhat different students).
Predictive Validity
The predictive validity of the CBVS was analyzed
per grade-level grouping by examining the relation
between victimization experiences and measures of
life satisfaction, school connectedness, and hope
(Table VI). The total victimization score had a
moderately negative relationship (r 5 !.14 to !.50)
with each of these indicators across age groups,
which were statistically significant in eight out of
nine comparisons.
California Bullying Victimization Scale 243
Analysis of variance of victimization classification
with measures of life satisfaction, school connectedness, and hope indicated a significant decrease in
positive functioning among students experiencing
peer victimization and bullying (Table VII). Students who were specifically classified as bullied
victims by the CBVS had significantly lower scores
on these indicators of positive functioning than
students who were peer victims but did not meet
criteria to be classified as bullied. The association of
victimization and positive functioning were most
pronounced for junior high and high school
students. Life satisfaction and school connectedness
seemed to suffer, in particular, among students who
were classified as peer victims and bullied victims.
Power Difference and Victimization Frequency
When applying the theoretical definition of bullying
to assign groups, there is a possible confound
between frequency of victimization and reporting a
power differential. Although our objective was to
determine whether strictly defined bullied victims
differed from other types of peer victims, we
conducted a contingency table analysis to determine
the level of association between power differential
and frequency of victimization. Results are displayed in Table VIII. Although there was a
significant relation between frequency of victimization experience and endorsement of a power
TABLE VI. Validity Coefficients of Total Victimization
Score (range 0–24) and Well-being Measures
Life satisfaction
School connectedness
Hope
7–8
9–12
!.50!!!
!.37!!
!.41!!
!.29!!
!.25!!
!.34!!
!.14!!
!!!Po.001; !!Po.01; one-tailed Pearson’s correlation.
Despite the many measures that have been
developed to evaluate bullying, assessment continues to need improvement in this field of research
[Cornell et al., 2006; Lee and Cornell, 2010]. This
study evaluated the psychometric properties and
validity of a new instrument, the CBVS, which was
designed to address some of the weaknesses in
existing measures of bullying. In particular, the
CBVS differentiates bullying from other forms of
peer victimization, a distinction that is rarely studied
[Hunter et al., 2007; Schäfer et al., 2002], yet
important for identifying students with the most
negative peer experiences. The CBVS specifically
measures the hallmark characteristics of bullying
(intentional, repeated, power differential) to make
this differentiation, while aiming to increase the
accuracy of classification of peer victims and bullied
victims. Several findings from this study indicate
that the CBVS demonstrates strong reliability and
TABLE VIII. Victimization Total Score (range 0–24) and
Percent with a Power Differential Endorsement
Victimization score
5–6
!.17
!.29!!
DISCUSSION
Power differential
Grade
Well-being measures
differential, t 5 7.53, df 5 326, Po.05, practically,
there were several students who reported frequent
victimization yet did not endorse a power differential, suggesting that assessing a power differential
can more accurately identify bullied victims.
Percent
n
0
34.7
52.3
75.0
25.0
29.0
94
170
44
16
4
328
0
1–6
7–12
13–18
19–24
Total
TABLE VII. Validity of Bully Victimization Classification for Victimization Group by Grade Level
Grades 5–6
Not a victim (N)
Peer victim (P)
Bullied victim (B)
F
df
Tukey HSD
Grades 7–8
Grades 9–12
LSS
SCS
CHS
LSS
SCS
CHS
LSS
SCS
CHS
3.76 (0.9)
3.63 (1.0)
2.92 (1.2)
4.40!
2, 77
BoP, N
2.88 (0.9)
2.97 (0.7)
2.68 (0.5)
1.23
2, 77
B5P5N
3.69 (0.7)
3.48 (0.9)
3.03 (1.1)
3.51!
2, 77
BoN
4.19 (0.8)
3.55 (1.0)
3.06 (1.1)
19.01!!!
2, 236
BoPoN
3.14 (0.7)
2.96 (0.6)
2.34 (0.8)
23.01!!!
2, 236
BoP 5 N
3.79 (0.9)
3.53 (1.1)
3.08 (1.1)
6.89!!
2, 236
BoP 5 N
3.64 (1.0)
3.42 (1.1)
3.08 (1.1)
8.62!!!
2, 348
BoN
2.86 (0.8)
2.65 (0.9)
2.29 (0.9)
13.29!!!
2, 348
BoN
3.56 (1.0)
3.44 (1.0)
3.13 (1.2)
5.10!!
2, 348
BoN
Note: LSS 5 Life Satisfaction Scale, range 0–5; SCS 5 School Connectedness Scale, range 0–4; CHS 5 Children’s Hope Scale, range 5 0–5.
Standard deviations in parentheses.
!!Po.01; !!!Po.001.
Aggr. Behav.
244 Felix et al.
validity to evaluate bullying victimization at the
individual level.
First, CBVS prevalence estimates had a similar
pattern to other studies that report age-related
decreases in rates of bully victimization, as youth
transition from junior high to high school [Nansel
et al., 2001; Solberg et al., 2007]. The CBVS
additionally showed gender-related differences in
the type of bullying experienced that are similar to
previous research, with males reporting higher rates
of physical victimization and females reporting
higher rates of relational victimization [Crick and
Grotpeter, 1995; Rivers and Smith, 1994].
Second, the CBVS had good test–retest reliability.
Report stability is a key indicator of effective bullying
evaluation as bullying, by definition, occurs repeatedly. The CBVS demonstrates good reliability over a
2-week period for the total scale score, each form of
victimization, and classification of students as bully
victims. It is notable that students who reported bully
victimization consistently across the two assessments
had lower scores on indicators of well-being than non
victims or unstable bullied victims. This suggests that
consistency across test–retest conditions is not only a
methodological issue, but may also indicate the
development of stable identification as a bullied
victim, which is associated with poor psychosocial
outcomes [Rosen et al., 2007].
Third, responses to the CBVS are associated with
indicators of well-being. Students who reported
higher CBVS scores reported lower life satisfaction,
school connectedness, and hope, indicating that the
CBVS detects students who are most at-risk for poor
psychosocial development at school. Particularly
striking is our finding that the group of students who
were bullied had lower scores on positive indicators
than students who reported more general experiences of peer victimization. This result is consistent
with Hunter et al. [2007] who found when they
categorized bullied youth separately from other peer
victims, those who were bullied were at much greater
risk for depressive symptomatology. Although few
studies have isolated these groups and evaluated
differential outcomes, students who specifically meet
criteria for the definition of bullying experience
significantly more difficulty than those who report
more general peer victimization. In addition, we
found that assessing for the power differential
helped in the identification of bullied victims than
when examining frequency of victimization alone.
Although CBVS prevalence rates were similar to
those derived from a definition-first measure that
was coadministered in this study, the two measures
identified somewhat different sets of students as
Aggr. Behav.
having been bullied. The lack of concordance
between the two measures indicates that the way
self-report surveys are structured fundamentally
impacts which students will identify themselves as
having been bullied. The absence of the word
‘‘bully’’ from the CBVS may have made questions
easier to endorse for students who had been
victimized. However, the requirement that students
explicitly endorse power differential may have had
the contrary effect of limiting respondents who did
not perceive a power difference or did not identify
the particular power difference that they experienced
among the list provided by the CBVS. Our choice of
what qualities to assess as a power differential
represented only a few of potentially many sources,
and was guided by previous bullying surveys.
‘‘Smart’’ was included as one source, as some prior
research indicated that victims have lower selfconcept in terms of intelligence [Salmivalli, 1998].
However, in practicality, only a few students
identified smart as being the only power differential.
It may be that other sources of power differential are
more beneficial to assess.
In addition, prevalence rates in this study were
higher than rates reported in some of the prior
research using definition-first measures [Nansel et al.,
2001, 2004; Solberg and Olweus, 2003] and peernomination methods [Juvonen et al., 2003]. Higher
response rates on the CBVS and the Swearer Survey
raise questions about whether the sample in this
study truly experienced higher rates of bullying or
whether inflated rates were a product of the
methodology used in this evaluation. Yet, validity
analyses support the categories yielded by the CBVS
as students whose responses resulted in the bullied
status also reported the lowest levels of well-being.
In developing the CBVS, we were particularly
interested to determine whether it would be possible
to include all three definitional characteristics of
bullying in a self-report measure. We did this by
asking a series of questions, so participants did not
have to remember all three definitional components
at once. Prior research has found that very few
children naturally incorporate power differential and
repetition in self-generated definitions of ‘‘bullying’’
[Vaillancourt et al., 2008]. Although definition-first
measures commonly embed a statement about power
differential in their description of bullying and
subsequently ask students to rate the frequency of
victimization, to our knowledge, only one other study
has explicitly incorporated items measuring power
differential in bullying assessment [Hunter et al.,
2007]. This study found considerably higher rates of
depressive symptoms among students who met
California Bullying Victimization Scale 245
criteria for all components of the bullying definition.
Similarly, Solberg and Olweus [2003] found that
frequent victimization was associated with greater
internalizing and externalizing problems.
Preliminary analysis indicated a significant, but
not fully overlapping relation between repetition of
victimization and power differential endorsement. It
is possible that these two groups of victims have
different relations to outcome variables. In the
future, it will be important to determine whether
all three components of the definition of bullying are
equally associated with psychological well-being, or
whether repetition, intentionality, and power imbalance may be unequally weighted in their impact
or differentially associated with outcomes.
LIMITATIONS AND FUTURE DIRECTIONS
This study is an initial demonstration that the
CBVS appears to have promise as a psychometrically
sound measure, but some limitations deserve mention. First, the CBVS is a self-report measure and
there have not yet been any data collected from other
informants or through observation to confirm the
validity of the measure. Even though other studies of
bullying have found limited agreement between
informants [Rønning et al., 2009], it would be useful
to assess agreement between self-, teacher-, and peerinformant detection of students who were bullied
victims vs. peer victims. Self-report alone may be
subject to misattribution of aggression, which can
lead to overidentification of bullying, or shame and
fear of repercussions, which can lead to underestimation of bullying [Card and Hodges, 2008].
Second, the CBVS does not yet distinguish
students who are purely victims of bullying from
those who are also bullies. This is in development.
Identifying bullies is associated with a new set of
challenges; however, evidence suggests that students
who are both bullies and victims (bully-victims)
should be considered a separate group from pure
victims as their profiles differ and are associated
with worse outcomes [Nansel et al., 2004; Solberg
et al., 2007]. Future iterations of the CBVS will
explore identification of bullies, using similar
strategies to evaluate power differential in relationships. Likewise, bullying can be seen as a peer group
process [Salmivalli, 1998, 1999, 2010], and other
participant roles, such as reinforcer and defender,
deserve assessment and study to better inform
intervention efforts.
A third limitation of this study is a question of the
generalizability of the sample. As noted earlier, rates
of bullying among the students surveyed were higher
than is typical in bullying research, even for the
definition-first measure that was coadministered
with the CBVS. Although data were collected from
four different schools, all schools were located
within the central coastal region of California, with
a predominantly White and Latino/a student
population. We have a strong interest to expand
our administration of the CBVS to students and
schools in other locations.
CONCLUSION
Although there is a significant amount of research
dedicated to peer victimization and bullying, this
discipline is affected by a lack of measurement
precision. Measures that have been thoroughly
developed to estimate prevalence within broad
populations have been adapted for individual
difference analyses without revalidation. For bullying
scholarship to advance, it is critical to have
consistently high standards for psychometric testing
of instruments before implementation in research or
in practice. We developed the CBVS to address gaps
between the definition and theory of bullying and
the way bullying is measured. In this preliminary
study, the CBVS reliably assessed the multiple
components of bullying and validly distinguish
between nonvictims, peer victims, and bullied
victims. Results also indicated that these distinctions
are important for identifying students most impacted by victimization experiences. Overall, these
findings suggest that the CBVS has promise to be
included in a new generation of bullying assessments
that are valid for use in research and in practice to
study the individual-level impacts of bullying.
Ultimately, the aim is for the CBVS to become an
early screening in a multigating assessment, by
which students with high CBVS scores will be
referred for a more in-depth interview to determine
the extent to which they are involved in bullying and
identify mechanisms of the bullying relationship that
can inform interventions. For example, a student
who reports being bullied by a single classmate may
need an intervention directed at that dyadic relationship in the context of their classroom. However, a
student who is repeatedly bullied by multiple
individuals across different settings may benefit
from the equivalent of a functional behavioral
analysis to understand the process by which bullying
is occurring and guide intervention for that particular student [Colvin et al., 1998]. In addition, a
classroom or school-wide intervention aimed at
Aggr. Behav.
246 Felix et al.
addressing bullying behavior and the supporting
roles of assistant to the bully, or reinforcer
[Salmivalli, 1999] would be indicated. A multigating
assessment framework provides school-wide data
to derive large-scale prevention and intervention
efforts, while offering detailed information about
students who are bullied that can have practical
implications for the targeted delivery of interventions for individuals.
ACKNOWLEDGMENTS
The project was supported by the Field Initiated
Studies Program from the Hamilton Fish Institute
with funding from the United States Office of
Juvenile Justice and Delinquency Prevention via
George Washington University. Additional information about the California Bully Victimization
Scale is available from the first author.
REFERENCES
Arora CMJ. 1996. Defining bullying: Towards a clearer general
understanding and more effective intervention strategies. Sch
Psychol Int 17:317–329. DOI: 10.1177/0143034396174002.
Björkvist K, Ekman K, Lagerspetz K. 1982. Bullies and victims. Their
ego picture, ideal ego picture, and normative ego picture. Scand J
Psychol 23:307–313. DOI: 10.1111/j.1467-9450.1982.tb00445.x.
Card NA, Hodges EVE. 2008. Peer victimization among schoolchildren: Correlations, causes, consequences, and considerations
in assessment and intervention. Sch Psychol Q 23:451–461. DOI:
10.1037/a0012769.
Chan JHF. 2006. Systemic patterns in bullying and victimization.
Sch Psychol Int 27:352–369. DOI: 10.1177/0143034306067289.
Chan JHF, Myron R, Crawshaw M. 2005. The efficacy of nonanonymous measures of bullying. Sch Psychol Int 26:443–458.
DOI: 10.1177/0143034305059020.
Colvin G, Tobin T, Beard K, Hagan S, Sprague J. 1998. The school
bully: Assessing the problem, developing interventions, and
future research directions. J Behav Educ 8:293–319. DOI:
10.1023/A:1022819213727.
Cornell DG, Brockenbrough K. 2004. Identification of bullies and
victims: A comparison of methods. In: Furlong MJ, Morrison G,
Skiba R, Cornell D (eds). Issues in School Violence Research.
New York, NY: Haworth Press, pp 63–87.
Cornell DG, Sheras PL, Cole JC. 2006. Assessment of bullying. In:
Jimerson SR, Furlong MJ (eds). Handbook of School Violence
and School Safety. Mahwah, NJ: Erlbaum, pp 587–602.
Cortina LM, Swan S, Fitzgerald LF, Waldo C. 1998. Sexual
harassment and assault: Chilling the climate for women in
academia. Psychol Women Q 22:419–441.
Craig WM, Pepler D, Atlas R. 2000. Observations of bullying in the
playground and in the classroom. Sch Psychol Int 21:22–36. DOI:
10.1177/0143034300211002.
Crick NR, Grotpeter JK. 1995. Relational aggression, gender, and
social-psychological adjustment. Child Dev 66:710–722. DOI:
10.2307/1131945.
Espelage DL, Swearer SM. 2003. Research on school bullying and
victimization: What have we learned and where do we go from
here? Sch Psychol Rev 32:365–383.
Aggr. Behav.
Felix ED, McMahon SD. 2006. Gender and multiple forms of peer
victimization: How do they influence adolescent psychosocial
adjustment? Violence Vict 21:707–724. DOI: 10.1891/vv-v21i6a003.
Felix ED, McMahon SD. 2007. The role of gender in peer
victimization among youth: A study of incidence, interrelations,
and social cognitive correlates. J Sch Violence 6:3–22.
Fitzgerald LF, Ormerod AJ. 1991. Perceptions of sexual harassment:
The influence of gender and academic context. Psychol Women Q
15:281–294.
Furlong MJ, Sharkey JD, Felix ED, Tanigawa D, Green JG. 2010.
Bullying assessment: A call for increased precision of selfreporting procedures. In: Jimerson SR, Swearer SM,
Espelage DL (eds). Handbook of Bullying in Schools:
An International Perspective. New York, NY: Routledge,
pp 329–345.
Greif JL, Furlong MJ. 2006. The assessment of school bullying:
Using theory to inform practice. J Sch Violence 5:33–50. DOI:
10.1300/J202v05n03_04.
Hamby SL, Finkelhor D. 2000. The victimization of children:
Recommendations for assessment and instrument development.
J Am Acad Child Adolesc Psychiatry 39:829–840. DOI: 10.1097/
00004583-200007000-00011.
Hawkins DL, Pepler DJ, Craig WM. 2001. Naturalistic observations
of peer interventions in bullying. Soc Dev 10:512–527. DOI:
10.1111/1467-9507.00178.
Henry DB. 2006. Associations between peer nominations, teacher
ratings, self-reports, and observations of malicious and disruptive
behavior. Assessment 13:241–252. DOI: 10.1177/1073191106287668.
Huebner ES. 1991. Correlates of life satisfaction in children. Sch
Psychol Int 12:231–240. DOI: 10.1177/0143034391123010.
Huebner ES, Suldo SM, Valois RF. 2005. Children’s life satisfaction.
In: Moore KA, Lippman LH (eds). What do Children Need to
Flourish? New York, NY: Springer, pp 41–59. DOI: 10.1007/
0-387-23823-9-4.
Hunter SC, Boyle JME, Warden D. 2007. Perceptions and correlates
of peer victimization and bullying. Br J Educ Psychol 77:797–810.
DOI: 10.1348/000709906X171046.
Juvonen J, Graham S, Schuster MA. 2003. Bullying among young
adolescents: The strong, the weak, and the troubled. Pediatrics
112:231–1237. DOI: 10.1542/peds.112.6.1231.
Kert AS, Codding RS, Tryon GS, Shiyko M. 2010. Impact of the
word ‘‘bully’’ on the reported rate of bullying behavior. Psychol
Sch 47:193–204. DOI: 10.1002/pits.20464.
Kyriakides L, Kaloyirou C, Lindsay G. 2006. An analysis of the
Revised Olweus Bully/Victim Questionnaire using the Rasch
measurement model. Br J Educ Psychol 76:781–801. DOI:
10.1348/000709905X53499.
Lee T, Cornell D. 2010. Concurrent validity of the Olweus Bully/
Victim Questionnaire. J Sch Violence 9:1–18. DOI: 10.1080/
15388220903185613.
McNeely C. 2005. Connection to school. In: Moore KA, Lippman LH
(eds). What do Children Need to Flourish? New York, NY:
Springer, pp 289–303. DOI: 10.1007/0-387-23823-9-18.
McMeely C, Nonnemaker JM, Blum RW. 2002. Promoting school
connectedness: Evidence from the National Longitudinal Study
of Adolescent Health. J Sch Health 72:138–146. DOI: 10.1111/
j.1746-1561.2002.tb06533.x.
Menesini E, Fonzi A, Smith PK. 2002. Attribution of meanings to
terms related bullying: A comparison between teacher’s and
pupil’s perspectives in Italy. Eur J Psychol Educ 17:393–406.
Nansel TR, Overpeck M, Pilla RS, Ruan WJ, Simons-Morton B,
Scheidt P. 2001. Bullying behaviors among U.S. youth: Prevalence and association with psychosocial adjustment. J Am Med
Assoc 285:2094–2100. PMid:15289243 PMCid:2556236.
California Bullying Victimization Scale 247
Nansel TR, Craig W, Overpeck MD, Saluja G, Ruan WJ and the
Health Behaviour in School-aged Children Bullying Analyses
Working Group. 2004. Cross-national consistency of the relationship
between bullying behaviors and psychosocial adjustment. Arch
Pediatr Adolesc Med 158:730–736. DOI: 10.1001/archpedi.158.8.730.
Naylor P, Cowie H, Cossin F, de Bettencourt R, Lemme F. 2006.
Teachers’ and pupils’ definitions of bullying. Br J Educ Psychol
76:553–576. DOI: 10.1348/000709905X52229.
Olweus D. 1993. Bullying at School: What We Know and What We
Can Do. Malden, MA: Blackwell.
Olweus D. 1996. The Revised Olweus Bully/Victim Questionnaire.
Mimeo. Bergen, Norway: Research Center for Health Promotion
(HEMIL Center), University of Bergen.
Ortega R, Mora-Merchán JA, Singer M, Smith PK, Pereira B,
Menesini E. 2001. Final report of the working group on general
survey questionnaires and nomination methods concerning
bullying. Retrieved on September 14, 2010, from http://
www.gold.ac.uk/tmr/reports/aim2_seville1.html.
Paludi MA. 1997. Sexual harassment in schools. In: O’Donahue W
(ed.). Sexual Harassment: Theory, Research, and Treatment.
Boston, MA: Allyn and Bacon, pp 225–240.
Rigby K. 2004. What It Takes to Stop Bullying in Schools: An
Examination of the Rationale and Effectiveness of School-Based
Interventions. In: Furlong MJ, Bates MP, Smith DC, Kingery P
(eds). Appraisal and Prediction of School Violence. Hauppauge,
NY: Nova Science, pp 165–191.
Rivers I, Smith PK. 1994. Types of bullying behavior and
their correlates. Aggr Behav 20:359–368. DOI: 10.1002/10982337(1994)20:5o359::AID-AB248020050343.0.CO;2-J.
Rønning JA, Sourander A, Kumpulainen K, Tamminen T,
Niemelä S, Moilanen I, Helenius H, Piha J, Almqvist F. 2009.
Cross-informant agreement about bullying and victimization
among eight-year-olds: Whose information best predicts psychiatric caseness 10-15 years later? Soc Psychiatry Psychiatr
Epidemiol 44:15–22. DOI: 10.1007/s00127-008-0395-0.
Rosen PJ, Milich R, Harris MJ. 2007. Victims of their own
cognitions: Implicit social cognitions, emotional distress, and
peer victimization. J Appl Dev Psychol 28:211–226. DOI:
10.1016/j.appdev.2007.02.001.
Salmivalli C. 1998. Intelligent, attractive, well-behaving, unhappy:
The structure of adolescents’ self-concept and its relations to
their social behavior. J Res Adolesc 8:333–354. DOI: 10.1207/
s15327795jra0803_3.
Salmivalli C. 1999. Participant role approach to school bullying:
Implications for interventions. J Adolesc 22:453–459. DOI:
10.1006/jado.1999.0239.
Salmivalli C. 2010. Bullying and the peer group: A review. Aggr
Violent Behav 15:112–120. DOI: 10.1016/j.avb.2009.08.007.
Schäfer M, Werner NE, Crick NR. 2002. A comparison of two
approaches to the study of negative peer treatment: General
victimization and bully/victim problems among German
schoolchildren. Br J Dev Psychol 20:281–306. DOI: 10.1348/
026151002166451.
Smith PK, Cowie H, Olafsson RF, Liefooghe APD. 2002. Definitions
of bullying: A comparison of terms used, and age and gender
differences, in a fourteen-country international comparison. Child
Dev 73:1119–1133. DOI: 10.1111/1467-8624.00461.
Snyder CR. 2005. Measuring hope in children. In: Moore KA,
Lippman LH (eds). What do Children Need to Flourish?
New York, NY: Springer, DOI: 10.1007/0-387-23823-9-5.
Solberg ME, Olweus D. 2003. Prevalence estimation of school
bullying with the Olweus Bully/Victim Questionnaire. Aggr
Behav 29:239–268. DOI: 10.1002/ab.10047.
Solberg ME, Olweus D, Endresen IM. 2007. Bullies and victims at
school: Are they the same pupils? Br J Educ Psychol 77:441–464.
DOI: 10.1348/000709906X105689.
Swearer SM. 2001. Bully Survey. Unpublished survey. Lincoln, NE:
University of Nebraska-Lincoln. As retrieved on September 6,
2010, from http://www.targetbully.com/Bully_Survey.php.
Terry T, Huebner ES. 1995. The relationship between self-concept
and life satisfaction in children. Soc Indicators Res 35:39–52.
DOI: 10.1007/BF01079237.
Vaillancourt T, McDougall P, Hymel S, Krygsman A, Miller J,
Stiver K, Davis C. 2008. Bullying: Are researchers and children/
youth talking about the same thing? Int J Behav Dev 486–495.
DOI: 10.1177/0165025408095553.
Varjas K, Henrich CC, Meyers J. 2009. Urban middle school
students’ perceptions of bullying, cyberbullying, and school safety.
J Sch Violence 8:159–176. DOI: 10.1080/15388220802074165.
You S, Furlong MJ, Felix ED, Sharkey JD, Tanigawa D, Green JG.
2008. Relations among school connectedness, hope, life satisfaction, and bully victimization. Psychol Sch 45:446–460. DOI:
10.1002/pits.20308.
Aggr. Behav.