Content uploaded by Ashwin Rambaran
Author content
All content in this area was uploaded by Ashwin Rambaran on May 18, 2020
Content may be subject to copyright.
Content uploaded by Ashwin Rambaran
Author content
All content in this area was uploaded by Ashwin Rambaran on Oct 25, 2019
Content may be subject to copyright.
Journal of Educational Psychology
Stability and Change in Student Classroom Composition
and Its Impact on Peer Victimization
J. Ashwin Rambaran, Marijtje A. J. van Duijn, Jan Kornelis Dijkstra, and René Veenstra
Online First Publication, December 2, 2019. http://dx.doi.org/10.1037/edu0000438
CITATION
Rambaran, J. A., van Duijn, M. A. J., Dijkstra, J. K., & Veenstra, R. (2019, December 2). Stability and
Change in Student Classroom Composition and Its Impact on Peer Victimization. Journal of
Educational Psychology. Advance online publication. http://dx.doi.org/10.1037/edu0000438
Stability and Change in Student Classroom Composition and Its Impact on
Peer Victimization
J. Ashwin Rambaran, Marijtje A. J. van Duijn, Jan Kornelis Dijkstra, and René Veenstra
University of Groningen
Although peer victimization in school mainly takes place between children in the same classroom or
grade and bullying is generally seen as a group process, little is known about how stability and change
in classroom composition affect peer victimization. Hence, this study addressed the following questions:
(a) Are newcomers in the classroom more likely to become victims? (b) Does a stable classroom, where
children generally have the same classmates over time, lead to less change in bully nominations? To
address these questions, this article examined 3 waves of bully nominations in a sample of 3,254 children
(50% boys; age 8 –12) in 31 elementary schools, displaying three types of schools: stable or unstable
administrative or pedagogical multigrade. Both research questions were answered by longitudinal social
network analyses of the school-wide networks. The meta-analyzed results of these analyses with small
effect sizes showed that (a) although stable classrooms do not necessarily show less change in bully
nominations than in unstable classrooms, victim-bully ties are more likely to develop among students in
the same grade or same classroom and (b) newcomers were more likely to become victims, more so in
unstable schools than in stable schools.
Educational Impact and Implications Statement
This study contributes to the existing bullying literature by providing first insights into the formation
and development of bullying relationships within the school context by examining changes in
victim-bully networks in schools that do and do not combine classrooms or grades over the school
years. The findings of this study suggest that school and classroom stability and change have a minor
impact on the formation of victim-bully relationships between children. Bullying relationships were
found to develop most easily between children in the same grade, more so in stable classrooms than
in classrooms with changing classroom composition, with no clear evidence that newcomers are more
at risk of becoming victimized. The formation and development of bullying relationships among
students within the same grade was weakest in unstable pedagogical multigrade schools, after
controlling for school size. These findings may be beneficial to schools that consistently deal with
changing compositions in their student population and highlight that a context-specific approach may
be necessary to tackle bullying in stable and unstable schools.
Keywords: social networks, peer victimization, student classroom composition, stability and change,
childhood
Supplemental materials: http://dx.doi.org/10.1037/edu0000438.supp
Peer victimization is widespread in elementary schools across
the world. Although prevalence of peer victimization at school
varies between countries, figures from nationally representative
samples in Europe and North America show that on average 30%
of children are occasionally victimized by schoolmates and 10%
are chronic victims (Chester et al., 2015). The long-term effects of
peer victimization can be devastating for victims of school bully-
ing, including poor academic functioning, anxiety, depression, and
XJ. Ashwin Rambaran, Marijtje A. J. van Duijn, Jan Kornelis Dijkstra,
and René Veenstra, Department of Sociology and Interuniversity Center
for Social Science Theory and Methodology (ICS), University of Gro-
ningen.
This research was funded by the Dutch Scientific Organization (NWO)
Program Council for Fundamental Scientific Education Research (PROO)
Project 411-12-027 awarded to René Veenstra, Jan Kornelis Dijkstra,
Wilma A. M. Vollebergh, Marijtje A. J. van Duijn, Zeena Harakeh, and
Christian Steglich (2013). We thank the members of the research groups
Social Development of Adolescents and Statistical Methods for Social
Network Analysis at the University of Groningen for their useful com-
ments and suggestions on this study.
Correspondence concerning this article should be addressed to J. Ashwin
Rambaran, Department of Sociology and Interuniversity Center for
Social Science Theory and Methodology (ICS), University of Groningen,
Grote Rozenstraat 31, 9712 TG, Groningen, the Netherlands. E-mail:
j.a.rambaran@rug.nl
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Educational Psychology
© 2019 American Psychological Association 2019, Vol. 1, No. 999, 000
ISSN: 0022-0663 http://dx.doi.org/10.1037/edu0000438
1
future delinquent and aggressive behavior (Ladd, Ettekal, &
Kochenderfer-Ladd, 2017; McDougall & Vaillancourt, 2015;
Wolke & Lereya, 2015). Prevention efforts to stop bullying be-
havior has been at best moderate, making it an ongoing concern for
schools, teachers, and parents (for a review see Rivara & Le
Menestrel, 2016). It is, therefore, important to understand when
and under what conditions bullying emerges and persists.
Peer victimization is largely studied as phenomenon that
takes place between classmates (Salmivalli, 2010). Yet, little is
known about how the school or classroom context affects peer
victimization (Juvonen & Graham, 2014). A contextual factor
to consider is stability of the classroom composition across the
school years. It is reasonable to expect that classroom compo-
sition changes are likely to impact the interactions between
children because it minimizes their opportunities to connect
with each other (Valente, 2012).
Using individual-level bullying measures, previous research
found lower self-reports of bullying and victimization among
students who moved to a different location during the transition to
middle school compared with students who stayed in the same
school (Farmer, Hamm, Leung, Lambert, & Gravelle, 2011; Wang,
Brittain, McDougall, & Vaillancourt, 2016). An explanation is that
such transitions break up the dominance structures and the accom-
panying bullying. Bullying may also be the reason why children
move to a different school. Thus, changing the classroom compo-
sition may break up victim-bully relationships and help in reducing
bullying.
Researchers increasingly recognize that bullying is relational,
and that a relational approach allows for a more nuanced under-
standing of who bullies whom in the classroom (Rodkin, Espelage,
& Hanish, 2015; Veenstra et al., 2007). Bullies target specific
victims, particularly the classmates with the weakest positions, and
victim-bully ties are also subject to change over time (Huitsing,
Snijders, van Duijn, & Veenstra, 2014; Rambaran, Dijkstra, &
Veenstra, 2019). Classmates generally have a good sense of each
other’s social positions (Farmer, Lines, & Hamm, 2011), and it is
reasonable to assume that this is greater in a stable classroom
(Farmer, Hamm, et al., 2011). In addition to established classroom
members with weak positions, newcomers may also suffer from an
initially weak social status when transitioning to a new school
environment, although still relatively little is known about this
group.
We examine the dynamics in victim-bully relationships and
focus on individual effects (newcomers) and dyadic effects (num-
ber of times a child shared the same classroom with a peer) that
capture the complexity of stability and change in classroom com-
position. To this end, we examine three waves of victim-bully
relationships in a sample of elementary schools in middle to late
childhood that differ in the extent to which students are organized
in same or different classrooms over time, using longitudinal social
network analysis.
Positions in Peer Groups in Middle and
Late Childhood
Middle and late childhood is an important developmental period
in which children develop social skills that help them establish
positive peer relationships. During this period, children become
more aware of their own and other’s position in the peer group
(Kolbert & Crothers, 2003). Within the school context, peer groups
are largely formed within the classroom as children spend most of
their school time there. The way children behave and interact with
each other plays an important role in how positions and roles are
defined (Farmer, Hamm, et al., 2011). During the middle and late
childhood period, children differentiate peers who are prosocial
(referring to being nice and cooperative), from peers who are
coercive (referring to being harmful and aggressive), to obtain
social positions in the group based on social status and acceptance
from peers (Hawley, 1999). Both behaviors form the basis for
youth in defining positions and roles in the classroom. For in-
stance, children who are generally prosocial may receive more
friendships and likes, which makes others perceive them as social
leaders among peers. However, children differ in their abilities to
be prosocial, and, some may turn to coercive (aggressive) strate-
gies to obtain dominant positions among peers, most likely by
bullying others in the group. Bullies are considered to be socially
skilled children that use proactive aggressive strategies to obtain
dominance and social status among peers (Sijtsema, Veenstra,
Lindenberg, & Salmivalli, 2009). In doing so, bullies tend to
enhance their position in the peer group by targeting weaker peers
(Rodkin et al., 2015; Salmivalli, 2010; Veenstra, Lindenberg,
Munniksma, & Dijkstra, 2010). Moreover, bullies often seek social
support from peers that help them to maintain a high position, by
becoming friends with others who join their bullying (Rambaran et
al., 2019) and by receiving help against defenders of victims
(Huitsing et al., 2014). Once positions are formed, children may
settle with their group position, as bully, victim, or uninvolved.
However, children’s positions and roles in the classroom are not
necessarily stable and it is reasonable to assume that this depends
on changes in the classroom composition.
Classroom Composition Changes and
Victim-Bully Networks
Changes in the classroom composition and its impact on peer
victimization may take different forms. To clarify this, it is im-
portant to consider school networks as nested structures with
individual students nested in classrooms in schools. Of further
importance is that peer victimization is nested in dyads as it
describes a relationship between two students (e.g., student i
nominates another student jas his or her bully). Figure 1 provides
an illustration of a school-wide victim-bully network with transi-
tions across three school years. The network consists of 93 stu-
dents clustered in five classrooms in Year 1 (T1) and Year 2 (T3),
but four classrooms in Year 3 (T5) because the students in one
classroom moved on to secondary education at T5. The students
are represented with colored nodes, where each color represents a
different classroom. Their victim-bully ties to each other are rep-
resented with directed arrows. Three of the five classrooms at T1
(Grade 2, 3, and 5) remained relatively stable over time (the
students remained in the same class together over the school
years). Two other classrooms with the same grade (Grade 5A and
Grade 5B) were “mixed” at T3. In addition to changes at the
classroom (or school) level (referring to classroom mixing),
there are changes at the individual level. For instance, one
student joins the classroom (network) at T3, whereas 17 stu-
dents leave the (school) network (at either T3 or T5). As shown
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
in Figure 1, most victim-bully ties are clustered within class-
room or grade, and the number of victim-bully ties gradually
decreases over time.
To capture the above-described complexities and changes
within a school victim-bully network, we examine the individual
effect of being a newcomer, referring to students who enter a rather
stable classroom, and the dyadic effect of same classroom before,
referring to the number of times a student shared the same class-
room with a peer. The same classroom now effect refers to
students being in the same classroom only once; whereas the same
grade effect refers to children being grademates. Together, these
effects capture how changes in classroom composition affect
changes in victim-bully relationships. This is done by including
“regular” schools where children typically move classrooms in a
following school year with most of their classmates from a previ-
ous year. In our study, these schools are referred to as stable
schools as compared with other schools that (consistently) com-
bine classrooms or grades, which are referred to as unstable
schools. In addition, we take into account the multigrade class-
rooms based on administrative and pedagogical reasons. The dis-
tinction between the two types of classrooms is important as the
motivation for having multigrade classrooms may affect the rela-
tion between change in classroom composition and peer victim-
ization because of school climate (Rambaran, van Duijn, Dijkstra,
& Veenstra, 2019a).
Individual Changes
The student population of classrooms change because of chil-
dren who repeat a grade, skip a grade, or move houses. In addition,
children may have to move to another school, because of low
academic achievement, behavioral problems, special learning
needs, or parents’ request (OECD, 2013). These individual
changes affect the amount of change in classroom composition and
ultimately children’s positions in the classroom. Children who are
new in a classroom may experience more difficulties with social
adjustment than established classroom members (Geven, Weesie,
& van Tubergen, 2016; Lubbers, Snijders, & van der Werf, 2011).
In a stable classroom, children generally know each other, as they
have a shared history. Newcomers might experience more diffi-
culties to integrate within the group and establish friendships.
Therefore, we expected that in stable classrooms, newcomers are
more likely to become targets for peer victimization (H1).
Classroom Changes
In a stable classroom context children know about each other’s
positions, including whom to target (Farmer, Hamm, et al., 2011;
Farmer, Lines, et al., 2011). This implies that, in such a context,
once children have a weak position they cannot easily change that
(Evans & Eder, 1993). In this situation, victims do not have a
chance for a “fresh start” and cannot easily escape their bullies and
Figure 1. Transition of students (nodes) and their victim-bully ties (arrows) in classes in a stable (single-grade)
school (n
ⴱ
⫽93 students) with two parallel groups that are mixed at T3.
ⴱ
For practical reasons (ease of
interpretation), we removed eight isolates from the networks (students who were uninvolved in victim-bully ties
from T1 to T5). Sec. Edu. ⫽secondary education. See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
remain socially isolated. Hence, a stable classroom group is likely
to lead to persistent peer victimization. Stability in the classroom
context may contribute to ongoing bullying of the same targets
(Farmer, Lines, et al., 2011; Wang et al., 2016). In an early study,
it was found that the stability of bullying behavior was weaker in
low-stability classrooms (Salmivalli, Lappalainen, & Lagerspetz,
1998). More recent empirical findings point in the direction that
higher group stability results in higher bullying (Farmer, Lines, et
al., 2011; Wang et al., 2016). Yet, it remains unclear who the
victims are because previous research focused on general bullying
behavior.
In contrast to stable schools, some other (typically smaller)
schools (yearly) combine classrooms or grades for practical rea-
sons, for example, to deal with low enrollment and uneven class-
room sizes (Mulryan-Kyne, 2007; Veenman, 1995). Classroom
“mixing” greatly affects the amount of stability and change in
classroom composition. Entering a new classroom context offers
opportunities to reestablish group positions and hierarchies
through bullying (Farmer, Hamm, et al., 2011). Bullying research
on the effects of changing peer groups generally focused on the
transition to middle or (junior) high school (Farmer, Hamm, et al.,
2011; Pellegrini & Bartini, 2000; Pellegrini & Long, 2002; Wang
et al., 2016). School transitions are a risk period for peer victim-
ization because existing peer groups are reshuffled and new social
structures are established that are linked to bullying, as children
compete over status through bullying in a new social environment
(Farmer, Hamm, et al., 2011; Farmer, Lines, et al., 2011; Pel-
legrini, 2002; Pellegrini & Bartini, 2000; Pellegrini & Long,
2002). Reshuffling of peer groups or classrooms results in loss of
existing friendships (Neckerman, 1996), which may lead to adjust-
ment problems for victims (Hodges, Boivin, Vitaro, & Bukowski,
1999), as in a new classroom environment victims may find it
difficult to find new friends. At the same time, it may also provide
an opportunity to escape from their previous bullies. This is
probably a reason why children reported less bullying and victim-
ization if they changed classrooms (Farmer, Lines, et al., 2011;
Wang et al., 2016). Following this line of reasoning, we expected
that, in stable schools as compared with unstable schools, victim-
bully relationships are more likely to be formed among children
who are in the same classroom over multiple school years (H2).
In a stable school context, same classroom typically refers to
same grade (referring to same classmates within the same grade).
However, same-grade victim-bully relationships may occur also
outside the own classroom when there are multiple classrooms of
the same grade in the same school (see, e.g., the two Grade 4
classes 4A and 4B at Time 1 in Figure 1). In this perspective, same
grade may already capture a large proportion of the victim-bully
ties between the children who were classmates before. Hence, the
two can be seen as complementary to each other. Accordingly, we
expected that victim-bully relationships are more likely to be
formed among children who are in the same grade, particularly in
stable schools compared with unstable schools (H3).
Also in stable schools, changes occur in classroom composition
when individual children leave or join the classroom, and we
expected that victim-bully relationships are formed among chil-
dren who are together in the same classroom, even when it is only
for one school year, rather than among children in different class-
rooms (H4).
In addition, research points in the direction that in a stable
school environment, higher-grade students in school seek out
lower-grade victims because they form an easy target (Huitsing et
al., 2014). In view of this power imbalance, we expected lower-
grade students to be victimized by higher-grade students (H5).
School Climate
The hypothesized effects described above (individual new-
comer, and dyadic same classroom before, same classroom now,
and same grade) may vary between school types because of school
climate. The evolutionary model of risky child/adolescent behavior
(Ellis et al., 2012) posits that mixed-age settings, rather than
age-segregated school and peer environments, are the natural con-
text for child development. The presence of both older and
younger children in mixed-age settings provides a natural hierar-
chy based on age. In this context, both older and younger children
settle more easily with their position in the social group, which
decreases the tendency to compete for dominance and status.
Evolutionary psychologists argue that older children can serve as
positive role models, and that the positive association between
status and prosocial behavior reduces the need to gain status
through antisocial means. When older children are assigned to
younger children as caregivers, buddies, or playmates, they tend to
behave less aggressive and more prosocial toward younger chil-
dren and same-age peers in other contexts as well (Gray, 2011).
Thus, the presence of younger children in mixed-age settings
reduces aggression and promotes nurturance and compassion in
children (Gray, 2011). In contrast, age-segregated schools and peer
environments, such as stable schools (with single-grades), have
been argued to evoke aggression and conflict in children, and, in
such a stable classroom context, children may actively search for
dominance (Ellis et al., 2012).
We distinguish unstable schools with administrative multigrade
(mixed-age) classrooms from schools with pedagogical multigrade
classrooms because the latter aims to stimulate prosocial relations
among children by encouraging the provision of help across grades
within the same classroom (Gray, 2011; Lillard & Else-Quest,
2006), whereas schools with administrative multigrade classrooms
do not have such an explicit goal (Mulryan-Kyne, 2007; Veenman,
1995). By contrast, teachers in administrative multigrade class-
rooms were generally found to teach the grades separately
(Mulryan-Kyne, 2007; Veenman, 1995), which decreases oppor-
tunities for prosocial behavior between older and younger children
as such multigrade classrooms emphasize individualized work and
do not necessarily encourage interactions between children from
different grades (Juvonen, 2018). Schools with pedagogical mul-
tigrade classrooms also encourage inclusive behavior, which re-
duces the probability of (repeated) victims or students who enter a
new classroom environment (newcomers) being rejected. The find-
ings of a recent study show that schools engaged in practices to
promote inclusiveness and equity as a school program foster
positive relationships (Rivas-Drake, Saleem, Schaefer, Medina, &
Jagers, 2019). Moreover, a positive school climate, for instance
through school and student support, community building, and
cooperative learning, was shown to reduce the prevalence of peer
victimization (Cornell, Shukla, & Konold, 2015; Fink, Patalay,
Sharpe, & Wolpert, 2018; Van Ryzin & Roseth, 2018). Following
this line of reasoning, we expected to find smaller effects (indi-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
vidual, classroom, and grade) in schools that formed multigrade
classrooms based on pedagogical reasons compared with stable
schools or schools that formed multigrade classrooms based on
administrative reasons (H6).
The Present Study
We investigate the extent to which stability and change in
classroom composition affects the formation of victim-bully rela-
tionships among children. We address the following questions: (a)
Are newcomers in the classroom more likely to become victims?
(b) Does a stable classroom, where children generally have the
same classmates over time, lead to less change in bully nomina-
tions? To address our research questions, we used a large data set
from the Netherlands and selected schools that differ in the extent
to which students are organized in same or different classrooms
across three school years. We defined three types of schools in the
available data: (a) schools that generally form single-grades and
are relatively stable in terms of classroom composition, and
schools that are relatively unstable in terms of classroom compo-
sition because of generally forming multigrades either for (b)
administrative or (c) pedagogical reasons. Based on information
provided by the school office about the school’s educational phi-
losophy, only schools that mentioned to have a specific educa-
tional philosophy were considered as multigrade for pedagogical
reasons (e.g., Montessori or Jenaplan schools).
We tested our hypotheses by investigating the changes in
victim-bully relationships in three measurements of social net-
works containing students’ bully nominations (“By whom are you
victimized?”). We analyzed the longitudinal social network data
using SIENA (Simulation Investigation for Empirical Network
Analysis; Snijders, van de Bunt, & Steglich, 2010; Steglich, Sni-
jders, & Pearson, 2010). We controlled for sex and grade (age),
because it is relevant in victim-bully relationships, with boys being
more dominant and aggressive toward girls (Cook, Williams,
Guerra, Kim, & Sadek, 2010) and higher-grade (older) students in
school may be seeking out lower-grade (younger) victims because
they form an easy target (Chaux & Castellanos, 2015; Huitsing et
al., 2014).
Method
Sample
Schools were drawn from the Dutch KiVa study (Huitsing et al.,
2019; Kaufman, Kretschmer, Huitsing, & Veenstra, 2018; Ram-
baran et al., 2019a) in three consecutive years (Spring, 2012, 2013,
and 2014, corresponding to Wave 1, 3, and 5). KiVa is an inter-
vention program, originally developed in Finland (Kärnä et al.,
2011, 2013), and aimed to reduce bullying among children from
Grades 2–5 in elementary education (7–11 years) in the Nether-
lands. As part of the intervention program, the participating
schools (n⫽99) were randomly assigned by the Netherlands
Bureau for Economic Policy Analysis (CPB) to either the control
condition (33 schools) or to the intervention condition (66
schools). The 33 control schools were selected for the analysis to
avoid that differences between schools were a result of the inter-
vention (Kaufman et al., 2018; Rambaran et al., 2019a). Two
control schools were dropped: one school did not participate after
Wave 4, and in another school, 35 of the children participating in
Wave 1 (12.5%) transitioned from a control school to an interven-
tion school, which made the school a special case and unfit for
comparison.
The 31 remaining schools, with 3,254 students (school size
varies between 36 and 276) over the 3 years, were categorized as
stable (n⫽8; 1,203 students), unstable administrative multigrade
(n⫽18; 1,436 students), and unstable pedagogical multigrade
(n⫽5; 615 students). Of all students, 49.9% were boys, the
average age of the sample at T1 was 9.6 years (SD ⫽1.4), 76.1%
of students were native Dutch, 18.5% were non-Dutch (minority),
and for 7.3% information about their parent’s ethnic background
was missing.
Of the 2,607 Grade 2–5 students who were targeted to partici-
pate in the KiVa control sample in T1 (the “eligible participants”),
2,562 students (98.3%) participated at T1 (Grade 2–5 in May,
2012), 2,415 (92.6%) at T3 (Grade 3– 6 in May, 2013), and 1,734
(66.5%) at T5 (Grade 3– 6 in May, 2014). The significant drop in
participation rate at T5 is because Grade 6 students at T3 continued
their educational career in secondary education. The difference
between the total sample size of 3,254 students and the number of
“eligible participants” at T1 is explained by classroom composi-
tion changes (students who joined the school at later time points),
and inclusion of students (mainly students who were in Grades 1
or 6 at T1) who did not participate themselves but could be
nominated (as bully) by peers.
On each measurement occasion, in an instructional movie, a
professional actress explained to students what bullying means,
using the following text:
Bullying is when some children repeatedly harass another child. The
child who gets bullied has problems defending itself against this.
Bullying is not the same as having a fight between two people who are
equally strong. Bullying should also not be confused with joking
around. Bullying is treating someone repeatedly in a mean way.
Several examples of bullying were given to students, including
physical and material forms (e.g., hitting someone, kicking or
pinching; stealing or damaging someone’s belongings) and rela-
tional and verbal forms (e.g., making fun of someone, calling
names, saying mean things; gossip about someone; excluding from
social activities).
Procedure
Students filled in an Internet-based questionnaire in their class-
room during regular school hours. The process was administered
by the teachers, who were present to answer questions and to assist
the students when needed. Before the data collection, teachers
were given detailed instructions concerning the procedure. During
the data collection, support was available through phone and
e-mail.
At the beginning of the questionnaire, students received infor-
mation about the goal of the study, and how to fill in the ques-
tionnaire. They were told not to talk to each other or to discuss
their answers when they filled out the questionnaire or afterward to
ensure each other’s privacy. It was explained to students that their
answers would remain confidential. The teachers ensured that
students who could not complete the questionnaire at the day of the
data collection participated at another day within a month.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
Before the first measurement (and for students who were new in
school, after the first measurement), schools sent information
letters to students’ parents. A passive consent procedure allowed
students or parents to opt out of student participation. At the start
of data collection (2012), universities in the Netherlands did not
require Institutional Review Board (IRB) permission for this type
of research. All procedures performed in this study were in accor-
dance with the 1964 Helsinki declaration and its later amendments
or comparable ethical standards. A few students did not want to
participate; also a few parents objected to their child’s participa-
tion. Accordingly, for the first wave participation rate among the
“eligible participants” was high (98.3%).
Measures
Dependent variable. Peer victimization was measured with
network nominations for bullying following a two-stage proce-
dure. To identify the victims for the second stage (the nomination
procedure, detailed below), all participating students were first
asked to indicate how often they were victimized in general in the
previous months (T1 and T5: “since the Christmas break”; T3:
“since the summer break”), according to Olweus’ (1996) self-
reported bully/victim items, and, to indicate this for specific forms
of victimization; physical harm (e.g., kicked), verbal harm (e.g.,
name calling), relational harm (e.g., gossiping), and cyber victim-
ization. Answers were given on a five point scale: (1) “Not at all,”
(2) “Once or twice,” (3) “Two or three times a month,” (4) “Once
a week,” and (5) “Several times a week.”
If participants indicated that they were victimized by classmates
at least “Once or twice” (score 2) on any item, they were presented
with a roster showing the names of all classmates, and asked whom
of their classmates victimized them (referring to “By whom are
you victimized?”). In addition, respondents could type in the
names of other schoolmates who victimized them by typing in the
first letters of their names on the computer screen (a name gener-
ator was used). In this nomination procedure, victims nominated
their bullies as perceptions and experiences of victims particularly
matter. For that reason, we look at bullying from the point of view
of the victim. Bullying nominations were measured as present (1)
or absent (0). Students who indicated not being victimized by
classmates or other schoolmates did not fill out the nomination
question. Their “answers” were considered as absent nominations.
Based on these nominations, school-wide victim-bully (referring to
victim sender and bully receiver) networks were obtained contain-
ing all bully nominations in a school or classroom (from the
victim’s perspective). The obtained school-wide victim-bully net-
works were used in the longitudinal social network analysis.
Explanatory variables. Individual (newcomers) and dyadic
(same class before,same class now, and same grade) variables
were constructed to examine whether and to which extent stability
and change in student classroom composition affects changes in
victimization relations.
Newcomers. We determined for students to which extent they
were new in school (transferred from another school) or in a
classroom (transferred from another classroom in the same
school). This was done by calculating the proportion of classmates
with which each student remained together in the same classroom
across two consecutive school years (referring to between T1 and
T3, between T3 and T5). After subtracting the proportion scores
from 1, we obtained a continuous covariate that ranged from 0
(students share a classroom with all of their classmates from a
previous school year)to1(students do not share a classroom with
classmates from a previous school year; newcomers).
Same class before. We determined for every pair of students
in school how long they had been together in the same class-
room, by counting the shared classrooms at T3 (0 or 1) and T5
(0, 1, or 2).
Same grade and same class now. These two binary dyadic
variables indicate whether pairs of children are currently (at T3
and T5) in the same classroom or grade.
Control variables. We included sex (1 ⫽boy). Students’
grade was obtained from the school’s office.
Analytic Strategy
The stochastic actor-based model implemented in SIENA al-
lows us to examine to which extent changes in victim-bully net-
works are related to endogenous network effects (e.g., reciprocity)
and exogenous individual (newcomers and sex) and dyadic (num-
ber of shared classrooms and grade) effects that may explain the
changes of ties in these networks.
School network is the unit of analysis. In most schools, student
classroom composition changed (Figure S1 in the online supple-
mental materials), particularly in schools that mixed or combined
classrooms over time. To facilitate the composition changes in
each school, we analyzed each school network including the par-
ticipants who were present at the first observation moment as well
as students who joined or left the networks at the third or fifth
observation. In addition, students who were not part of the study
design at T1 (the “eligible participants”) but were part of a com-
bination group were also included because they could be nomi-
nated as bully by schoolmates, making them part of a school
victim-bully network. This enabled us to make full use of the
available information and to analyze the networks according to the
“method of joiners and leavers,” which uses information about
composition changes in an efficient way (Huisman & Snijders,
2003; Ripley, Snijders, Boda, Vörös, & Preciado, 2019).
Each school network was estimated with the same model spec-
ification. In some (larger) school networks, however, additional
effects were necessary to achieve acceptable model fit. Ultimately,
all models showed good convergence and fit statistics (see for
details Table S2 and Figure S3 in the online supplemental mate-
rials). Goodness of fit (GoF) was examined using four network
statistics: outdegree distribution, indegree distribution, geodesic
distance, and triad census, by investigating how well these statis-
tics are captured in a sample of networks simulated according to
the estimated model. For each of these statistics, the differences
between the values in the observed school network (summed
across the three waves of data) and the estimated values (summed
across 1,000 simulated networks) are assessed with the Mahalano-
bis distance (Ripley et al., 2019). Fit for a particular statistic is
good or acceptable when the Mahalanobis is small as expressed by
apvalue larger than .05. The violin plots in Figure S3 in the online
supplemental materials can be used for a graphical inspection of
the departure of the simulated values from the observed value of
the statistics (in red). A good network fit is essential to interpret the
effects of main interest more reliably.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
The effects were first analyzed for each school network sepa-
rately, and the parameter estimates were then summarized with a
meta-analysis using R-package metafor (in a random effects model
using the default REML function; Viechtbauer, 2010). Two anal-
yses were performed: one “simple” model (without covariates)
rendering the mean model parameter estimates, obtained in the
SIENA analysis per school with the accompanying standard errors
as weights as well as a test for between-school heterogeneity, and
a model with school type and school size as “explanatory vari-
ables” (that can best be understood together, referring to a com-
bination of school type and school size). We controlled for school
size because unstable schools are generally smaller compared with
stable schools. Thus, to understand the effects of unstable schools
school size was taken into account.
Model Specification and Effect Interpretation
To adequately capture important features of the victim-bully
networks, we followed previous research in choosing the structural
parameters in the stochastic actor-based models (Huitsing et al.,
2014; Rambaran et al., 2019). Several structural network effects
were included to account for changes in the overall network
structure. Network rate effects indicate the number of changes in
the victim-bully networks. Out-degree (density) is included to
indicate that students start sending bully nominations to school-
mates (victim inominates jas his or her bully). Out-degree isolates
is included to inversely indicate that students are not likely to start
sending bully nominations to schoolmates (nonvictimized students
or having zero out-degree). By including these three basic struc-
tural effects, it was possible to capture many of the network
properties of the school-wide victim-bully networks (as indicated
by fit statistics). For most school networks, one or two additional
effects were necessary to capture dispersion in out- or in-degrees
(referring to individual differences in sending and receiving bully
nominations).
Next to these structural effects, we included effects related to
sex and grade, to account for sex and grade differences in victim-
bully ties by including alter, ego, and similarity effects. The alter
and ego effects indicate that boys or higher-grade students are
more likely to receive and send bully nominations than girls or
lower-grade students; the same-sex and same-grade effects indi-
cate that victim-bully ties are more likely to be formed among
same-sex or same-grade students than among cross-sex and cross-
grade students. The combination of these three effects is necessary
to adequately capture selection tendencies based on sex and grade
in the victim-bully networks (Snijders et al., 2010).
Are newcomers in the classroom more likely to become
victims? To answer our first research question, we included an
ego effect for newcomers. This newcomer ego effect assessed
whether newcomers were more likely to start sending bully nom-
inations than established classroom members (H1).
Does a stable classroom lead to less change in bully
nominations? To answer our second research question, we in-
cluded three parameters. To test H2, the same class before effect
assessed whether the formation of victim-bully ties was more
likely among students who had been classmates before. To test H3,
the same grade effect was included. To test H4, the same class
now effect, tests whether changes in victim-bully ties are more
likely among same classmates being in the same classroom con-
currently.
Additional analyses. To assess effect sizes, we calculated the
relative importance of each effect on the probability of a tie change
(Indlekofer & Brandes, 2013). For this analysis, each school was
analyzed using the exact same model specification to ensure the
comparability of model parameters and relative effect sizes across
schools. The parameter estimates of these models were not very
different from those in the full models (see online supplemental
materials Table S6 and Table S8).
Results
Descriptive Analysis
Table 1 summarizes the descriptive information on the sample
and network characteristics. Table S4 in the online supplemental
materials provides information for each school separately. Table
S5 in the online supplemental materials summarizes this informa-
tion per school type (stable, unstable administrative multigrade,
and unstable pedagogical multigrade).
The density of the school-wide victim-bully networks was rel-
atively low (average density was .019 at T1; see Table 1), which
is because density takes into account network size. At the first time
point, children nominated on average 1 to 2 other schoolmates who
victimized them (average degree); this varied between schools
(minimum ⫽0.6, maximum ⫽3.5; see Table 1). Children’s
involvement in victim-bully relationships decreased over time:
whereas most children at Time 1 received a bully nomination
(sinks), sent one (sources) or both sent and received one (actives),
this was no longer the case at Time 3 and Time 5 (increasing
number of isolates).
Victimization occurred equally often among same-sex and
cross-sex peers per and between school types (Table S5 in the
online supplemental materials). Occurrence of victimization was
also similar across school types, whereas same-grade victimization
was higher in stable schools compared with the two unstable
school types. Almost three-quarters of victim-bully ties were
within classroom, and this was similar in all three school types.
Victim-bully ties were unstable from one time point to the next:
Many new victim-bully ties were created, and even more victim-
bully ties dissolved each school year. This can also be seen in
Figure 1, where the number of victim-bully ties decreases signif-
icantly over time. Only about 10% of the victim-bully ties were
stable from one time point to the next. This did not much differ
between the three school types (Table S5 in the online supplemen-
tal materials). Over 50% of the victim-bully ties at Time 3 and
Time 5 were between children who had shared the same class
before; this was higher in stable schools than in unstable (admin-
istrative and pedagogical multigrade) schools.
SIENA Analysis
Table 2 provides the summary of the SIENA findings (using
RSiena Version 1.1–307), meta-analyzed over the 31 school-wide
victim-bully networks. The first column in Table 2 shows the mean
estimates across all schools. The next column shows the mean
estimates across stable schools (as reference), the three other
columns show the degree to which administrative multigrade
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
schools and pedagogical multigrade schools deviate from this,
taking into account school size. Findings are also summarized for
the three different types of schools (stable, unstable administrative
multigrade, and unstable pedagogical multigrade), and this will be
used to discuss differences in the four effects of main interest
(newcomers,same class before,same class now, and same grade).
Figure S7 in the online supplemental materials provides forest
plots of these analyses, which are used to inspect outliers.
Network effects. The rate effects indicate that the average
number of changes in bully nominations was 18 between the
school years (T1–T2 and T3–T5), with significant variation be-
tween schools. In accordance with the low density of the victim-
bully networks, the negative outdegree (density) effect indicates a
low probability of students sending bully nominations to school-
mates. The accompanying negative outdegree-isolates effect indi-
cates that students who were not victimized tended to remain
nonvictimized. Compared with stable schools, density and isolates
effects were stronger in unstable (administrative and pedagogical
multigrade) schools. Note that density and isolates effects were
stronger (positive) in smaller schools, which typically corre-
sponded with unstable (administrative and pedagogical multi-
grade) schools.
Sex effects. The three included sex-selection parameters in
Table 2 (sex ego effect, sex alter effect, and same sex effect) are
interpreted with so-called ego-alter selection tables, represent-
ing the relative contribution to the evaluation function for the
four alter and ego sex combinations (Ripley et al., 2019). The
positive values on the diagonal in the left panel of Table 3 show
that victim-bully ties were more likely to be formed among
students of the same sex. This was similar between the three
school types. In addition, Table 3 also shows that girls were
rather victimized by boys (positive value for girl sender to boy
receiver) than vice versa, more so in pedagogical multigrade
schools than in the two other school types.
Individual effects. We expected that newcomers would be
more likely to become victims in schools where the group or
classroom is stable over time rather than unstable (H1). The
summary across all schools in Table 2 shows a positive (nonsig-
Table 1
Descriptive Statistics of Individual, Dyadic, and Network Change Characteristics Summarized
for All 31 School-Wide Victim-Bully Networks (n ⫽3,254 Students)
Sample T1 T3 T5
Network (schools)
School size 92 (27–261) 87 (20–195) 71 (20–182)
Percentage of boys 49% (35–63%) 49% (38–61%) 48% (31–64%)
Density .019 (.004–.05) .014 (.004–.042) .013 (.001–.04)
Average degree 1.5 (.6–3.5) 1.1 (.3–2.1) 1.1 (.2–3)
Number of ties 113 (11–324) 79 (8–280) 49 (3–144)
Mutual ties 9 (0–46) 4 (0–24) 2 (0–8)
Total sample (students)
Sinks
a
23.3% (12.5–36.3%) 21.6% (10.9–42.1%) 16.3% (7.5–27.5%)
Sources
a
16.4% (5.7–39.6%) 14.3% (4.5–22.6%) 11.8% (5–20.4%)
Isolates
a
41.3% (8.3–79.5%) 53.2% (15.8–81.2%) 65.4% (52.9–87.5%)
Actives
a
19% (1.9–47.5%) 10.9% (0–32%) 6.6% (0–16%)
Dyadic variables
Same sex 51% (26–71%) 54% (14–88%) 55% (21–83%)
Same grade now 61% (16–97%) 56% (12–91%) 59% (25–89%)
Same class now 74% (39–91%) 72% (34–97%) 74% (33–97%)
Sample change T1–T3 T3–T5
Totals
Joiners in school 3 (0–9) 4 (0–14)
Joiners in classroom 1 (0–9) 1 (0–9)
Leavers 8 (0–40) 25 (5–60)
Stayers 89 (18–221) 68 (18–173)
Tie change
Creating tie (0 ¡1) 57 (4–209) 30 (2–107)
Dissolving tie (1 ¡0) 78 (5–238) 39 (3–113)
Stable tie (1 ¡1) 15 (0–56) 8 (0–26)
Jaccard index
b
8.9% (0–24.4%) 10.3% (0–24.1%)
Individual variables
Newcomer
c
(corr) .14 (⫺.07–.36) .19 (⫺.02–.41)
Dyadic variables
Same grade before
d
55% (12–91%) 58% (25–89%)
Same class before
d
47% (12–73%) 50% (24–81%)
Note. Table reports averages, minimum and maximum in parentheses.
a
Sinks are actors with zero out-ties and at least one in-tie, sources are actors with at least one out-tie and zero
in-ties, isolates are actors with zero in-ties and zero out-ties, and actives are children with at least one out-tie and
at least one in-tie.
b
The Jaccard index is the fraction of stable ties relative to all new, lost and stable
ties.
c
Correlations between number of bully nominations send and continuous scores for newcomers. Corre-
lations were summarized using Fisher’s r-to-ztransformation.
d
Percentage of ties in T3 or T5.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
nificant) newcomer ego effect across all schools, which indicates
that, in some schools, newcomers are more likely to become
victimized by peers compared with students who had not changed
classrooms. A summary of the three types of schools, shows that
contrary to our expectation, this was not the case in the stable
schools, but only in the two types of unstable (multigrade) schools.
A positive newcomer ego effect was observed in half of the
administrative multigrade schools (Figure S7.K). A positive effect
was also observed in four of the five pedagogical multigrade
schools, the other school having a negative effect (Figure S7.K).
Grade effects. The three included grade-selection parameters
in Table 2 (grade ego effect, grade alter effect, and same grade
effect) are also summarized with ego-alter selection tables in Table
3. The positive values on the diagonal in the right panel of Table
3 show that victim-bully ties were more likely to be formed among
students of the same grade. We expected to find stronger effects in
schools where changes in classroom composition are small (H3).
In line with this, we found that students are more likely to be
victimized by peers in the same grade, more so in stable schools
than in the two other unstable school types (see the larger positive
values on the diagonal in Table 3). We also expected that the
higher-grade students would target their lower-grade peers, be-
cause they form easy targets (H5). The results were in line with
this expectation (positive values in the off-diagonal on the upper
right side of Table 3), but differences between the three school
types were small (see Table 2).
Classroom effects. The summary across all schools in Table
2 shows small (nonsignificant) effects of same class now and same
class before. Contrary to our expectation, no evidence was ob-
tained for an additional effect on victim-bully relationships being
formed between students who shared the same class before (H2).
The same class before effects were not larger in stable schools
compared with the two other unstable school types, when taking
into account their present classroom sharing status. The effect was
stronger in larger schools (Figure S7.M). The same class now
effect was positive in the four largest stable schools (Figure S7.L),
indicating that victim-bully ties occurred between children who
were currently (at T3 and T5) in the same classroom. This effect
was as expected (H4) and stronger in larger schools.
School climate: Administrative versus pedagogical multi-
grade schools. Finally, we expected to find smaller classroom
effects in schools that formed multigrade classrooms based on
pedagogical reasons compared with schools that formed multi-
grade classrooms based on administrative reasons (H6). Our find-
ings provided some evidence for this for same-grade victimization
(Figure S7.J), whereas the difference in parameter estimates as
shown in Table 2 is not significant. It should be noted, however,
that this finding is based on a small sample (only five pedagogical
multigrade schools were included).
Relative importance of effects. Network structure is the most
important determinant of change in the school-wide victim-bully
networks (Figure S9 in the online supplemental materials). Con-
Table 2
Meta-Analysis of School-Wide Victim-Bully Networks (31 Schools, 3,254 Students)— Results for All Three Different School Types
Controlling for School Size
b
Parameters Illustration n
All schools
Stable schools
(intercept)
Unstable
administrative
multigrade schools
Unstable pedagogical
multigrade schools School size
Est. SE Est. SE Est. SE Est. SE Est. SE
Rate effects
Network rate w1–w3 28 18.53
ⴱⴱⴱ
1.34
a
22.06
ⴱⴱⴱ
2.34
a
⫺2.25 3.31 ⫺6.90
†
3.83 .294 .242
Network rate w3–w5 27 18.37
ⴱⴱⴱ
2.35
a
22.54
ⴱⴱⴱ
3.59
a
⫺.44 5.07 3.37 5.96 1.261
ⴱⴱ
.386
Network effects
Density 31 ⫺3.88
ⴱⴱⴱ
.29
a
⫺5.28
ⴱⴱⴱ
.33
a
1.24
ⴱⴱ
.47 2.10
ⴱⴱⴱ
.58 ⫺.13
ⴱⴱⴱ
.035
Isolates 31 ⫺4.06
ⴱⴱⴱ
.18
a
⫺4.73
ⴱⴱⴱ
.30 .77
†
.41 .51 .52 ⫺.037 .030
Sex effects
Sex (boy) alter 31 .40
ⴱⴱⴱ
.08 .41
ⴱⴱ
.13 ⫺.08 .20 .19 .26 ⫺.003 .015
Sex (boy) ego 31 ⫺.18
ⴱ
.08 ⫺.16 .12 ⫺.01 .19 ⫺.03 .25 .003 .014
Same sex 31 .28
ⴱⴱⴱ
.08 .22
†
.13 .20 .20 ⫺.19 .26 .002 .015
Individual effects
Newcomer ego 31 .16 .11 ⫺.07 .19 .35 .28 .32 .37 ⫺.007 .021
Grade effects
Grade alter 31 .15
ⴱⴱ
.06 .22
ⴱ
.10 ⫺.11 .14 ⫺.10 .17 ⫺.003 .010
Grade ego 31 ⫺.15
ⴱⴱ
.05 ⫺.21
ⴱ
.10 .11 .14 .003 .17 .000 .010
Same grade 31 1.13
ⴱⴱⴱ
.11
a
1.44
ⴱⴱⴱ
.17 ⫺.45
†
.25 ⫺.80
ⴱⴱ
.28 ⫺.022 .017
Classroom effects
Same class now 31 .03 .11 .34
†
.18 ⫺.25 .25 ⫺.39 .32 .048
ⴱⴱ
.018
Same class before 31 .02 .09 ⫺.03 .14 .16 .21 .16 .28 .032
ⴱ
.016
Number of schools 31 8 18 5 31
Number of students 3,254 1,203 1,436 615 3,254
a
Significant differences between schools. In some (smaller) schools, rate parameters were unreasonably high because there were few or no stable ties from
one time point to the next. As a possible solution, the rate parameters were fixed at the observed value (see Ripley et al., 2019).
b
School size was mean
centered around the rounded mean school size of the reference category (here stable schools). Intercept represents the “baseline effect” of the reference
category. Using density as an example, a 10-unit (estimates and standard errors were multiplied by 10 for convenience) increase in school size (referring
to ⫹10 above the rounded mean school size of stable schools, which was 150) results in a ⫺.13 decrease in density in terms of the average effect estimate
for a particular school type.
†
p⬍.10.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
sidering that model fit was generally good, the victim-bully net-
works are relatively simple in terms of network structure, and were
mostly captured by the general tendency to be involved in victim-
bully relationships (referring to network density) or to remain
uninvolved in such relationships (referring to network isolates).
Overall, the density and isolates effects explain 70% of the influ-
ence choices for changes in victim-bully ties. This implies that the
other effects in the models are (relatively) small (see Figure S9).
Figure 2 summarizes the relative effect of the four parameters of
main interest in our study (individual effect: newcomers; dyadic
effects: same class before,same class now, and same grade),
where Figure S9 provides the complete information. In general,
dyadic effects were more important than individual effects in
explaining changes in victim-bully ties. In particular, of the three
dyadic effects same grade was the most important, although not for
all schools. In view of the absence of a pattern in the relative
effects in Figure 2 no indication was obtained that the strength of
effects depends on stability of the classroom composition.
Discussion
We examined the extent to which stability and change in student
classroom composition (referring to the group of students who are
in the same classroom together over the school year) affects the
formation of victim-bully relationships. Following a relational
approach, victim-bully relationships were examined with longitu-
dinal social network analysis. In doing so, our study contributes to
the existing bullying literature by providing first insights into the
formation and development of victim-bully relationships within
the school context by examining changes in victim-bully networks
in schools that do and do not combine classrooms or grades over
the school years.
Are Newcomers in the Classroom More Likely to
Become Victims?
Although based on a small number of schools and not signifi-
cant, the results indicate that newcomers in unstable (administra-
tive and pedagogical multigrade) schools are at risk to become
victimized by peers. In stable schools on the other hand, victim-
ization occurs mainly between students in the same grade, with no
indication for newcomers being more at risk. An explanation is
that a new classroom context, for example, through school transi-
tions, can be a stressful event for many children by disrupting their
friendships with peers (Benner & Graham, 2009; Parker, Rubin,
Erath, Wojslawowicz, & Buskirk, 2006). These negative experi-
ences are associated with loneliness and stress that form risk
factors for peer victimization (Farmer, Hamm, et al., 2011; Farmer,
Lines, et al., 2011; Wang et al., 2016). Research also shows that
some children are affected by classroom changes because of dis-
rupted social hierarchies, which requires them to “reestablish”
their social position by engaging in aggression or bullying (Pel-
legrini, 2002).
Table 3
Bully Nominations Based on Sex (Left) and Grade (Right) Summarized Across All 31 School-
Wide Victim-Bully Networks and for the Three Different School Types
a
Bully sex Bully grade
Victim Girl (0) Boy (1) Victim Grade 2 Grade 3 Grade 4 Grade 5
A: All schools (31 schools, 3,254 students)
Girl (0) .17 .29 Grade 2 1.12 .14 .29 .45
Boy (1) ⫺.29 .39 Grade 3 ⫺.16 1.12 .15 .30
Grade 4 ⫺.30 ⫺.15 1.13 .15
Grade 5 ⫺.45 ⫺.30 ⫺.15 1.13
B: Stable schools (8 schools, 1,203 students)
Girl (0) .10 .28 Grade 2 1.41 .20 .42 .63
Boy (1) ⫺.28 .34 Grade 3 ⫺.22 1.41 .21 .42
Grade 4 ⫺.43 ⫺.22 1.42 .22
Grade 5 ⫺.64 ⫺.42 ⫺.21 1.43
C: Unstable administrative multigrade schools (18 schools, 1,436 students)
Girl (0) .32 .27 Grade 2 1.06 .08 .21 .33
Boy (1) ⫺.27 .48 Grade 3 ⫺.15 1.09 .11 .23
Grade 4 ⫺.25 ⫺.12 1.11 .13
Grade 5 ⫺.35 ⫺.22 ⫺.09 1.14
D: Unstable pedagogical multigrade schools (5 schools, 516 students)
Girl (0) ⫺.16 .40 Grade 2 .73 .26 .38 .50
Boy (1) ⫺.38 .25 Grade 3 ⫺.07 .65 .17 .29
Grade 4 ⫺.27 ⫺.15 .56 .09
Grade 5 ⫺.48 ⫺.36 ⫺.24 .48
Note. The values on the diagonal indicate the likelihood of bully nominations when the individual and peer
have exactly the same score on sex or grade. The values in the cells in these tables can be transformed to odds
by taking the exponential function (exp.(k)); calculation based on the estimates in Table 2.
a
Table S2 in the online supplemental materials reports the model specification with all the effects.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
Classroom mixing may impact positive relationships between
students as it could lead to a loss of friendships (Neckerman,
1996), which seems especially problematic for the victims
(Hodges et al., 1999; Pellegrini, Bartini, & Brooks, 1999; Sainio,
Veenstra, Huitsing, & Salmivalli, 2011). It is plausible that the
dispositional or behavioral characteristics of newcomers, for in-
stance, being aggressive, withdrawn, or previously victimized
(Geven et al., 2016; Lubbers et al., 2011) are part of the explana-
tion for why newcomers might become a target of victimization.
Moreover, other behavioral or reputational risks for victimization,
such as low acceptance or an isolated position in the group (e.g.,
no or few friends), may increase the chance of being victimized.
Although not an easy task, taking these individual factors into
account in the study of victim-bully networks might lead to a more
nuanced understanding of the effects of newcomers in terms of
victimization in school. At the same time, classroom mixing offers
an opportunity for victims to find new friends or defenders that
help victims against a vicious cycle of victimization and its neg-
ative experiences (Hodges et al., 1999; Pellegrini et al., 1999;
Sainio et al., 2011). An avenue for further research is to understand
how victim-bully relationships develop in the context of friend-
ships within stable and unstable schools (for an example in stable
peer groups see Rambaran et al., 2019).
Does a Stable Classroom Lead to Less Change in
Bully Nominations?
To answer our second research question, we tested complementary
hypotheses. First, contrary to our expectation, the analyses provided
no evidence that students who shared the same class before (referring
to the number of shared classrooms) were more likely to be victimized
in the stable schools as compared with unstable schools. An expla-
nation might be that change in victim-bully ties is high, as shown in
our analysis, and in line with previous findings (Huitsing et al., 2014;
Rambaran et al., 2019). In light of bullying as strategic and goal-
oriented behavior (Salmivalli, 2010; Sijtsema et al., 2009; Veenstra et
al., 2010), recent research on the dynamic network interplay between
bullying and popularity reveals that bullies frequently change victims
to maintain a high social status (van der Ploeg, Steglich, & Veenstra,
2019). Other recent research also shows that, contrary to conventional
notions (Salmivalli, 2010), many children involved in bullying switch
roles: they may be a bully at one time point, a victim at another, and
uninvolved at the next time point (Zych et al., 2019).
Second, we expected that in a stable school context same
grade (typically referring to same classroom) may already cap-
ture a large proportion of the victim-bully ties between the
children who were classmates before. In line with our expec-
tation, our findings did show that the formation and develop-
ment of victim-bully relationships occurred more often among
same-grade students in stable schools than in unstable multi-
grade schools. An obvious explanation is that students in stable
schools remain together with their classmates from a previous
school year. This increases opportunities for bullying between
classmates who are in the same grade each year as their class-
room composition remains stable. The negligible effect of a
shared classroom history in our analyses might already be
accounted for by being consistently in the same grade over
time. Multigrade schools are highly unstable in terms of class-
room composition across the school years. However, in this
unstable school context bullying occurs more often among
same-grade than cross-grade students, and this was more
Figure 2. Relative importance of effects contributing to peer victimization for the four effects of main interest
(A. newcomer ego, B. same grade, C. same class now, and D. same class before). Calculated using the exact
same model specification for each school (results are reported in Table S5 in the online supplemental materials).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
strongly in the administrative multigrade schools than in ped-
agogical multigrade schools.
To summarize, our analyses showed that although stable class-
rooms do not necessarily show less change in bully nominations
than in unstable classrooms, victim-bully ties are more likely to
develop among students in the same grade or same classroom.
Does School Climate Affect the Relation Between
Change in Classroom Composition and Peer
Victimization?
Pedagogical multigrade schools share many features with ad-
ministrative multigrade schools, but there is an important distinc-
tion: classroom or grade mixing occurs purposefully to foster
positive interactions between cross-grade students in a classroom
(Gray, 2011; Lillard & Else-Quest, 2006). Our tentative findings
suggest that same-grade victim-bully relationships occur less in
unstable pedagogical multigrade schools than in unstable admin-
istrative multigrade schools and particularly stable schools, but at
the same time, there is also a relatively higher chance of cross-
grade victimization. This points at a potential trade-off effect:
while higher-grade students are encouraged to demonstrate their
value as social leaders toward their lower-grade classmates it
might also come across as social dominance for lower-grade
students, who are then at the bottom of the hierarchy.
Implications, Limitations, and Directions for
Future Research
Our findings may have implications for school policy about
classroom organization. Overall, results of this study suggest that
school and classroom stability and change have a small impact on
the formation of victim-bully relationships between children. Bul-
lying relationships were found to develop most easily between
children in the same grade, more so in stable classrooms than in
classrooms with changing classroom composition, with no clear
evidence that newcomers are more at risk of becoming victimized.
In addition, rates of bullying relationships appear to be very low in
schools with changing classrooms as compared with schools with
stable classrooms. To some extent these findings are reassuring for
schools that consistently have to deal with changing compositions
in their student population. Particularly in small cities or villages
where student enrollment declines, classroom sizes become un-
even, and the team of teachers is small (Mulryan-Kyne, 2007;
Veenman, 1995), schools are often forced to combine classrooms
or grades.
Although a direct comparison is not possible, our findings on
peer victimization extend previous research on sociocognitive out-
comes, which also found no clear indication that children in
administrative multigrade classrooms are worse off in terms of
academic achievement and social adjustment (Mulryan-Kyne,
2007; Veenman, 1995, 1996; but see Mason & Burns, 1996).
Multigrade schools may be beneficial, however, when they have
explicit goals in terms of enhancing the school climate. This
stimulates prosocial behavior and the provision of help across the
grades within a classroom (Lillard & Else-Quest, 2006; Moller,
Forbes-Jones, & Hightower, 2008), but also developing less bul-
lying between grademates.
These findings may be useful for school-wide or whole-group
interventions. Teachers should be aware of the fact that most
bullying-victimization situations occur between students in the
same grade or same classroom. As classmates or grademates
probably know each other better, the solution lies within the group,
for example by developing positive relations and helping behavior
between members of the classroom. This is supported by recent
research on the effects of intraschool dynamics among peers,
showing that a positive school climate enhances friendships and
acceptance of others (Cornell et al., 2015; Fink et al., 2018;
Rivas-Drake et al., 2019; Van Ryzin & Roseth, 2018).
Not only the potential importance of school climate and classroom
stability but also other school factors are worth considering (Juvonen,
2018; Juvonen & Graham, 2014). Where it was argued that peer
dynamics and hierarchies might differ across the different school
settings, it is not unlikely that also within the same school setting the
dynamics might differ, across classrooms, or grades. However, a
clear-cut measure of peer hierarchy, dominance or classroom norms is
not readily available and is a topic of study on its own (Rambaran, van
Duijn, Dijkstra, & Veenstra, 2019b; Salmivalli, 2010).
We measured peer victimization by asking the victims to nom-
inate their bullies (referring to by whom are you bullied?). We note
that by doing so we did not take into account the perspective of the
bullies. It is plausible that some of the bullies did not consider
themselves as a bully of (some of) the victims who nominated
them, which would lead to a discrepancy (low agreement) between
both sources of informants (Oldenburg et al., 2015). An avenue for
future research is to validate our findings by examining victim-
bullying networks using a multi-informant approach, with the
perspectives of both the victims and the bullies.
Future research would also benefit from examining victim-bully
networks in adolescence as bullying processes are likely to differ
between childhood and adolescence. The number of bullies in-
creases or remains stable in early adolescence, whereas the number
of victims decreases during this time (Nansel et al., 2001; Pel-
legrini & Long, 2002). This increase in bullying during early
adolescence is followed by a decrease during mid- and late ado-
lescence (Kretschmer, Veenstra, Dekovic´, & Oldehinkel, 2017),
which suggests that victims are especially in a vulnerable position
in early adolescence when bullying peaks. Investigation of victim-
bully networks in adolescence may be more complicated than in
childhood because adolescents have more diverse social interac-
tions with peers in school and outside of the own school context.
Within school, students share the same homeroom but switch
classes on a daily basis as they start to follow different school
programs (school courses and school subjects). Outside of school,
adolescents are often online or meet others that they know from
their neighborhood, or the sport club.
Conclusion
Our study shows that the impact of stability and change in
classroom composition on victim-bully relationships was rela-
tively small: students who shared the same class before were not
more likely to be victimized in the stable schools as compared with
students in unstable multigrade schools. Next to the general ten-
dency to form victim-bully ties or to remain uninvolved in victim-
bully ties over time, same-grade victimization appeared to be the
strongest predictor of victim-bully ties, particularly in stable
schools. In view of the strong same-grade effects, there was no
extra effect to be found of being in the same classroom before. The
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
12 RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
formation and development of victim-bully relationships among
students within the same grade was weakest in unstable peda-
gogical multigrade schools, after controlling for school size.
Taken together, the findings highlight that a context-specific
approach may be necessary to tackle bullying in stable and
unstable schools.
References
Benner, A. D., & Graham, S. (2009). The transition to high school as a
developmental process among multiethnic urban youth. Child Develop-
ment, 80, 356 –376. http://dx.doi.org/10.1111/j.1467-8624.2009.01265.x
Chaux, E., & Castellanos, M. (2015). Money and age in schools: Bullying
and power imbalances. Aggressive Behavior, 41, 280 –293. http://dx.doi
.org/10.1002/ab.21558
Chester, K. L., Callaghan, M., Cosma, A., Donnelly, P., Craig, W., Walsh,
S., & Molcho, M. (2015). Cross-national time trends in bullying victim-
ization in 33 countries among children aged 11, 13 and 15 from 2002 to
2010. European Journal of Public Health, 25(Suppl. 2), 61– 64. http://
dx.doi.org/10.1093/eurpub/ckv029
Cook, C. R., Williams, K. R., Guerra, N. G., Kim, T. E., & Sadek, S.
(2010). Predictors of bullying and victimization in childhood and ado-
lescence: A meta-analytic investigation. School Psychology Quarterly,
25, 65– 83. http://dx.doi.org/10.1037/a0020149
Cornell, D., Shukla, K., & Konold, T. (2015). Peer victimization and
authoritative school climate: A multilevel approach. Journal of Educa-
tional Psychology, 107, 1186 –1201. http://dx.doi.org/10.1037/
edu0000038
Ellis, B. J., Del Giudice, M., Dishion, T. J., Figueredo, A. J., Gray, P.,
Griskevicius, V.,...Wilson, D. S. (2012). The evolutionary basis of
risky adolescent behavior: Implications for science, policy, and practice.
Developmental Psychology, 48, 598 – 623. http://dx.doi.org/10.1037/
a0026220
Evans, C., & Eder, D. J. (1993). “No exit”: Processes of social isolation in
the middle school. Journal of Contemporary Ethnography, 22, 139 –170.
http://dx.doi.org/10.1177/089124193022002001
Farmer, T. W., Hamm, J. V., Leung, M.-C., Lambert, K., & Gravelle, M.
(2011). Early adolescent peer ecologies in rural communities: Bullying
in schools that do and do not have a transition during the middle grades.
Journal of Youth and Adolescence, 40, 1106 –1117. http://dx.doi.org/10
.1007/s10964-011-9684-0
Farmer, T. W., Lines, M. M., & Hamm, J. C. (2011). Revealing the
invisible hand: The role of teachers in children’s peer experiences.
Journal of Applied Developmental Psychology, 32, 247–256. http://dx
.doi.org/10.1016/j.appdev.2011.04.006
Fink, E., Patalay, P., Sharpe, H., & Wolpert, M. (2018). Child- and
school-level predictors of children’s bullying behavior: A multilevel
analysis in 648 primary schools. Journal of Educational Psychology,
110, 17–26. http://dx.doi.org/10.1037/edu0000204
Geven, S., Weesie, J., & van Tubergen, F. (2016). Peers in transition:
Students’ problem behavior in school across their transition to a new
school class. In S. Geven, Adolescent problem behavior in school: The
role of peer networks (Doctoral dissertation). ICS/Utrecht University.
Retrieved from https://dspace.library.uu.nl/handle/1874/338820
Gray, P. (2011). The special value of children’s age-mixed play. American
Journal of Play, 3, 500 –522.
Hawley, P. (1999). The ontogenesis of social dominance: A strategy-based
evolutionary perspective. Developmental Review, 19, 97–132. http://dx
.doi.org/10.1006/drev.1998.0470
Hodges, E. V. E., Boivin, M., Vitaro, F., & Bukowski, W. M. (1999). The
power of friendship: Protection against an escalating cycle of peer
victimization. Developmental Psychology, 35, 94 –101. http://dx.doi.org/
10.1037/0012-1649.35.1.94
Huisman, M. E., & Snijders, T. A. B. (2003). Statistical analysis of
longitudinal network data with changing composition. Sociological
Methods & Research, 32, 253–287. http://dx.doi.org/10.1177/
0049124103256096
Huitsing, G., Lodder, G. M. A., Oldenburg, B., Schacter, H. L., Salmivalli,
C., Juvonen, J., & Veenstra, R. (2019). The healthy context paradox:
Victims’ adjustment during an anti-bullying intervention. Journal of
Family Studies, 28, 2499 –2509. http://dx.doi.org/10.1007/s10826-018-
1194-1
Huitsing, G., Snijders, T. A. B., Van Duijn, M. A. J., & Veenstra, R.
(2014). Victims, bullies, and their defenders: A longitudinal study of the
coevolution of positive and negative networks. Development and Psy-
chopathology, 26, 645– 659. http://dx.doi.org/10.1017/S09545794
14000297
Indlekofer, N., & Brandes, U. (2013). Relative importance of effects in
stochastic actor-oriented models. Network Science, 1, 278 –304. http://
dx.doi.org/10.1017/nws.2013.21
Juvonen, J. (2018). The potential of schools to facilitate and constrain peer
relationships. In W. M. Bukowski, B. Laursen, & K. H. Rubin (Eds.),
Handbook of peer interactions, relationships, and groups (pp. 491–509).
New York, NY: Guilford Press.
Juvonen, J., & Graham, S. (2014). Bullying in schools: The power of
bullies and the plight of victims. Annual Review of Psychology, 65,
159 –185. http://dx.doi.org/10.1146/annurev-psych-010213-115030
Kärnä, A., Voeten, M., Little, T. D., Alanen, E., Poskiparta, E., &
Salmivalli, C. (2013). Effectiveness of the KiVa antibullying program:
Grades 1–3 and 7–9. Journal of Educational Psychology, 105, 535–551.
http://dx.doi.org/10.1037/a0030417
Kärnä, A., Voeten, M., Little, T. D., Poskiparta, E., Kaljonen, A., &
Salmivalli, C. (2011). A large-scale evaluation of the KiVa antibullying
program: Grades 4 – 6. Child Development, 82, 311–330. http://dx.doi
.org/10.1111/j.1467-8624.2010.01557.x
Kaufman, T. M. L., Kretschmer, T., Huitsing, G., & Veenstra, R. (2018).
Why does a universal anti-bullying program not help all children?
Explaining persistent victimization during an intervention. Prevention
Science, 19, 822– 832. http://dx.doi.org/10.1007/s11121-018-0906-5
Kolbert, J. B., & Crothers, L. M. (2003). Bullying and evolutionary
psychology: The dominance hierarchy among students and implications
for school personnel. Journal of School Violence, 2, 73–91. http://dx.doi
.org/10.1300/J202v02n03_05
Kretschmer, T., Veenstra, R., Dekovic´, M., & Oldehinkel, A. J. (2017).
Bullying development across adolescence, its antecedents, outcomes,
and gender-specific patterns. Development and Psychopathology, 29,
941–955. http://dx.doi.org/10.1017/S0954579416000596
Ladd, G. W., Ettekal, I., & Kochenderfer-Ladd, B. (2017). Peer victimiza-
tion trajectories from kindergarten through high-school: Differential
pathways for children’s school engagement and achievement. Journal of
Educational Psychology, 109, 826 – 841. http://dx.doi.org/10.1037/
edu0000177
Lillard, A., & Else-Quest, N. (2006). The early years. Evaluating Montes-
sori education. Science, 313, 1893–1894. http://dx.doi.org/10.1126/
science.1132362
Lubbers, M. J., Snijders, T. A. B., & van der Werf, M. P. C. (2011).
Dynamics of peer relationships across the first two years of junior high
as a function of gender and changes in classroom composition. Journal
of Research on Adolescence, 21, 488 –504. http://dx.doi.org/10.1111/j
.1532-7795.2010.00688.x
Mason, D. A., & Burns, R. B. (1996). ⬙Simply no worse and simply no
better⬙may simply be wrong: A critique of Veenman’s conclusion about
multigrade classes. Review of Educational Research, 66, 307–322.
http://dx.doi.org/10.3102/00346543066003307
McDougall, P., & Vaillancourt, T. (2015). Long-term adult outcomes of
peer victimization in childhood and adolescence: Pathways to adjust-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
13
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
ment and maladjustment. American Psychologist, 70, 300 –310. http://
dx.doi.org/10.1037/a0039174
Moller, A. C., Forbes-Jones, E., & Hightower, A. D. (2008). Classroom age
composition and developmental change in 70 urban preschool class-
rooms. Journal of Educational Psychology, 100, 741–753. http://dx.doi
.org/10.1037/a0013099
Mulryan-Kyne, C. (2007). The preparation of teachers for multigrade
teaching. Teaching and Teacher Education, 23, 501–514. http://dx.doi
.org/10.1016/j.tate.2006.12.003
Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simons-Morton, B.,
& Scheidt, P. (2001). Bullying behaviors among U.S. youth: Prevalence
and association with psychosocial adjustment. Journal of the American
Medical Association, 285, 2094 –2100. http://dx.doi.org/10.1001/jama
.285.16.2094
Neckerman, H. J. (1996). The stability of social groups in childhood and
adolescence. The role of the classroom social environment. Social De-
velopment, 5, 131–145. http://dx.doi.org/10.1111/j.1467-9507.1996
.tb00076.x
OECD. (2013), PISA 2012 results: What makes schools successful? Re-
sources, policies and practices (Volume IV). Paris: OECD Publishing.
http://dx.doi.org/10.1787/9789264201156-en
Oldenburg, B., Barrera, D., Olthof, T., Goossens, F., van der Meulen, M.,
Vermande, M.,...Veenstra, R. (2015). Peer and self-reported victim-
ization: Do non-victimized students give victimization nominations to
classmates who are self-reported victims? Journal of School Psychology,
53, 309 –321. http://dx.doi.org/10.1016/j.jsp.2015.05.003
Olweus, D. (1996). The revised Olweus bully/victim questionnaire. Bergen,
Norway: Research Center for Health Promotion (HEMIL Center), Uni-
versity of Bergen.
Parker, J., Rubin, K. H., Erath, S., Wojslawowicz, J. C., & Buskirk, A. A.
(2006). Peer relationships, child development, and adjustment: A devel-
opmental psychopathology perspective. In D. Cicchetti & D. Cohen
(Eds.) Developmental psychopathology: Risk, disorder, and adaptation
(pp. 419 – 493). New York, NY: Wiley.
Pellegrini, A. D. (2002). Bullying, victimization, and sexual harassment
during the transition to middle school. Educational Psychologist, 37,
151–163. http://dx.doi.org/10.1207/S15326985EP3703_2
Pellegrini, A. D., & Bartini, M. (2000). A longitudinal study of bullying,
victimization, and peer affiliation during the transition from primary
school to middle school. American Educational Research Journal, 37,
699 –725. http://dx.doi.org/10.3102/00028312037003699
Pellegrini, A. D., Bartini, M., & Brooks, F. (1999). School bullies, victims,
and aggressive victims: Factors relating to group affiliation and victim-
ization in early adolescence. Journal of Educational Psychology, 91,
216 –224. http://dx.doi.org/10.1037/0022-0663.91.2.216
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 Educational Psychology,
20, 259 –280. http://dx.doi.org/10.1348/026151002166442
Rambaran, J. A., Dijkstra, J. K., & Veenstra, R. (2019). Bullying as a group
process in childhood: A longitudinal social network analysis. Child
Development. Advance online publication. http://dx.doi.org/10.1111/
cdev.13298
Rambaran, J. A., van Duijn, M. A. J., Dijkstra, J. K., & Veenstra, R.
(2019a). Peer victimization in single-grade and multigrade classrooms.
Aggressive Behavior, 45, 561–570. http://dx.doi.org/10.1002/ab.21851
Rambaran, J. A., van Duijn, M. A. J., Dijkstra, J. K., & Veenstra, R.
(2019b). The relation between defending, (dis)liking and the bullying
norm in the classroom: A multilevel individual and social network
approach. In J. A. Rambaran, The classroom as context for bullying: A
social network approach (Doctoral dissertation). ICS/University of Gro-
ningen. http://dx.doi.org/10.33612/diss.96793146
Ripley, R. M., Snijders, T. A. B., Boda, Z., Vörös, A., & Preciado, P.
(2019). Manual for RSiena. Oxford, UK: University of Oxford. http://
www.stats.ox.ac.uk/~snijders/siena/RSiena_Manual.pdf
Rivara, F., & Le Menestrel, S. (2016). Preventing bullying through science,
policy, and practice. Washington, DC: The National Academies Press.
http://dx.doi.org/10.17226/23482
Rivas-Drake, D., Saleem, M., Schaefer, D. R., Medina, M., & Jagers, R.
(2019). Intergroup contact attitudes across peer networks in school:
Selection, influence and implications for cross-group friendship. Child
Development, 60, 1898 –1916. http://dx.doi.org/10.1111/cdev.13061
Rodkin, P. C., Espelage, D. L., & Hanish, L. D. (2015). A relational
framework for understanding bullying: Developmental antecedents and
outcomes. American Psychologist, 70, 311–321. http://dx.doi.org/10
.1037/a0038658
Sainio, M., Veenstra, R., Huitsing, G., & Salmivalli, C. (2011). Victims
and their defenders: A dyadic approach. International Journal of Be-
havioral Development, 35, 144 –151. http://dx.doi.org/10.1177/
0165025410378068
Salmivalli, C. (2010). Bullying and the peer group: A review. Aggression and
Violent Behavior, 15, 112–120. http://dx.doi.org/10.1016/j.avb.2009.08.007
Salmivalli, C., Lappalainen, M., & Lagerspetz, K. M. J. (1998). Stability
and change of behavior in connection with bullying in schools: A
two-year follow up. Aggressive Behavior, 24, 205–218. http://dx.doi
.org/10.1002/(SICI)1098-2337(1998)24:3⬍205::AID-AB5⬎3.0.CO;2-J
Sijtsema, J. J., Veenstra, R., Lindenberg, S., & Salmivalli, C. (2009).
Empirical test of bullies’ status goals: Assessing direct goals, aggression,
and prestige. Aggressive Behavior, 35, 57– 67. http://dx.doi.org/10.1002/
ab.20282
Snijders, T. A. B., van de Bunt, G. G., & Steglich, C. E. G. (2010).
Introduction to stochastic actor-based models for network dynamics.
Social Networks, 32, 44 – 60. http://dx.doi.org/10.1016/j.socnet.2009.02
.004
Steglich, C., Snijders, T. A. B., & Pearson, M. (2010). Dynamic networks
and behavior: Separating selection from influence. Sociological Meth-
odology, 40, 329 –393. http://dx.doi.org/10.1111/j.1467-9531.2010
.01225.x
Valente, T. W. (2012). Network interventions. Science, 337, 49 –53. http://
dx.doi.org/10.1126/science.1217330
van der Ploeg, R., Steglich, C., & Veenstra, R. (2019). The way bullying
works: How new ties facilitate the mutual reinforcement of status and
bullying in elementary schools. Social Networks. Advance online pub-
lication. http://dx.doi.org/10.1016/j.socnet.2018.12.006
Van Ryzin, M. J., & Roseth, C. J. (2018). Cooperative learning in middle
school: A means to improve peer relations and reduce victimization,
bullying, and related outcomes. Journal of Educational Psychology, 110,
1192–1201. http://dx.doi.org/10.1037/edu0000265
Veenman, S. (1995). Cognitive and noncognitive effects of multigrade
and multi-age classes: A best-evidence synthesis. Review of Educa-
tional Research, 65, 319 –381. http://dx.doi.org/10.3102/00346
543065004319
Veenman, S. (1996). Effects of multgrade and multi-age classes reconsid-
ered. Review of Educational Research, 66, 323–340. http://dx.doi.org/
10.3102/00346543066003323
Veenstra, R., Lindenberg, S., Munniksma, A., & Dijkstra, J. K. (2010). The
complex relation between bullying, victimization, acceptance, and re-
jection: Giving special attention to status, affection, and sex differences.
Child Development, 81, 480 – 486. http://dx.doi.org/10.1111/j.1467-
8624.2009.01411.x
Veenstra, R., Lindenberg, S., Zijlstra, B. J. H., De Winter, A. F., Verhulst,
F. C., & Ormel, J. (2007). The dyadic nature of bullying and victimiza-
tion: Testing a dual-perspective theory. Child Development, 78, 1843–
1854. http://dx.doi.org/10.1111/j.1467-8624.2007.01102.x
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
14 RAMBARAN, VAN DUIJN, DIJKSTRA, AND VEENSTRA
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor
package. Journal of Statistical Software, 36, 1– 48. http://dx.doi.org/10
.18637/jss.v036.i03
Wang, W., Brittain, H., McDougall, P., & Vaillancourt, T. (2016). Bullying
and school transition: Context or development? Child Abuse & Neglect,
51, 237–248. http://dx.doi.org/10.1016/j.chiabu.2015.10.004
Wolke, D., & Lereya, S. T. (2015). Long-term effects of bullying. Archives
of Disease in Childhood, 100, 879 – 885. http://dx.doi.org/10.1136/
archdischild-2014-306667
Zych, I., Ttofi, M. M., Llorent, V. J., Farington, D. P., Ribeaud, D., &
Eisner, M. P. (2019). A longitudinal study on stability and transitions
among bullying roles. Child Development. Advance online publication.
http://dx.doi.org/10.1111/cdev.13195
Received July 12, 2018
Revision received October 11, 2019
Accepted October 22, 2019 䡲
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
15
VICTIMIZATION IN STABLE AND CHANGING CLASSROOMS
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Educational Psychology
This content is subject to copyright. Terms and conditions apply.
