 

 

 

LIBRARY
Michigan State
University

This is to certify that the
dissertation entitled

A Longitudinal Study of Classroom Level Aggression and
Social Competency: Effects on Students’ Aggressive Behavior

presented by

Sonia A. Patil

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Counseling, Educational
Psychology and Special
Education

 

 

 

 

him mm

fMajor [Pro7:eZor’ 3 Signature

/07

Date

MSU is an afﬁrmative-action, equal-opportunity employer

—-—I-O-l-I-O-I-O-l-.-I-I-I-I-l-C-O-O-C-U-l-

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

FEﬁ f 6 £091

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/CIRC/DateDue.indd-p.1

 

A Longitudinal Study ofClassroom Level Aggression and Social Competency: Effects
on Students’ Aggressive Behavior

By

Sonia A. Patil

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Counseling, Educational Psychology and Special Education

2007

ABSTRACT

A Longitudinal Study of Classroom Level Aggression and Social Competency: Effects
on Students’ Aggressive Behavior

By
Sonia A. Patil
This study examined the effects of the classroom social context on the
development of student aggression across a school year among elementary school
students. Moderating effects of students’ initial level of aggression, gender and
developmental level were examined. The classroom context was deﬁned as the aggregate
level of aggression and social competency. Aggression and social competency were
measured using The Behavior Assessment System for Children Teacher Rating Scales
(BASC-TRS-C; Reynolds & Kamphaus, 1992). Multiple regression analyses using
general linear modeling (GLM) were employed to assess for change in student
aggression. Findings indicate that the classroom social context did inﬂuence change in
student aggression across a school year; however, the ﬁndings were not always in the
hypothesized direction. Major results included: 1) Initial level of aggression moderated
the effects of classroom social competency on student aggression. Higher levels of
classroom social competency led to increases in student aggression at lower levels of
initial aggression and decreases in student aggression at higher levels of initial
aggression; 2) For upper elementary aged students, higher levels of classroom aggression
led to more signiﬁcant decreases in student aggression than lower elementary students,
and 3) higher levels of classroom social competency led to more signiﬁcant increases in
aggression for younger students than older elementary students. The implications of the

ﬁndings are discussed in regard to future research and practice.

ACKNOWLEDGEMENTS

First, I would like to thank my advisor and dissertation chair Jean Baker whose
feedback, mentorship and encouragement have been instrumental to the completion of
this dissertation. Secondly, I would like to thank my dissertation committee members:
Linda Anderson, for providing feedback on my dissertation and for her support while I
was teaching; Sara Bolt for her feedback during this dissertation process; and, John
Carlson for his support in both my research and clinical endeavors. Thirdly, I am very
grateful for the friendships that I made during graduate school and thank my Michigan
Statue University comrades for their laughter and support. Finally, I would like to thank
my family, in particular my parents, for their constant love and encouragement and for
always believing in me.

Data for this study were collected as part of Project A.C.T. Early, ﬁmded by
F ield-Initiated Studies grants (R306F60158, R305T990330) from the Institute for At-
Risk Children of the Office of Educational Research and Improvement, United States

Department of Education

iii

TABLE OF CONTENTS

LIST OF FIGURES ...................................................................... vi

CHAPTER 1

INTRODUCTION ........................................................................ 1

CHAPTER 2

LITERATURE REVIEW ................................................................. 6
Theoretical Framework ......................................................... 6
Research on Classroom Social Context and Children’s Aggression. 10
Research Questions and Hypotheses ............................................ 24

CHAPTER 3

METHODS ................................................................................ 26
Participants ....................................................................... 26
Procedure ......................................................................... 27
Measures .......................................................................... 28
Data Analyses .................................................................... 30

CHAPTER 4

RESULTS .................................................................................. 32
Preliminary Analyses ........................................................... 32
Results ............................................................................ 33

CHAPTER 5

DISCUSSION ............................................................................ 38
Effects of Aggressive Classroom Environments ............................ 40
Effects of Classroom Social Competency ..................................... 43
Gender Differences 47
Developmental Level Differences................................................. 48
Limitations ........................................................................ 50
Implications for Future Research and Practice ............................... 52

REFERENCES ........................................................................... 63

iv

LIST OF TABLES

Table 1
Means and Standard Deviations for Aggression Subscale and Classroom
Level Variable ............................................................................ 55

Table 2
Correlations for Study Variables ....................................................... 56

Table 3
Regression Analyses Predicting Student Aggression from Classroom Social
Context and Moderator Vanables 57

Table 4
Mixed Model Analysis Predicting Student Aggression from Classroom Social

Context and Moderator Variables ...................................................... 59

LIST OF FIGURES

Figure 1.
Interaction of Classroom Level Social Competency and Students’ Initial
Level of Aggression on Aggression at the End of the School Year ............... 60

Figure 2.
Interaction of Classroom Level Aggression and Grade Level on Aggression at
the End of the School Year .............................................................. 61

Figure 3.
Interaction of Classroom Level Social Competency and Grade Level on
Aggression at the End of the School Year ............................................ 62

vi

Chapter One
Introduction

There has been a visible increase in disruptive behavioral patterns, which includes
aggression, acting out, hyperactivity, and oppositionality, among school-age youth over
the past three decades (Achenbach, Dummenci, & Rescorla, 2003). This increase is
alarming given the well-established link between childhood problem behaviors and later
maladjustment across academic, social, and psychological functioning. Speciﬁcally,
early disruptive behaviors have been connected to later substance abuse, delinquency,
low achievement, dropping out of school, poor social relationships, and antisocial
behavior (Loeber, 1990).

Over the past 40 years, a general consensus has been built that childhood
problems fall into one of two broad categories: externalizing behavior and internalizing
behavior. This consensus has evolved from a large body of literature that has clearly
validated these two broad areas of functioning (Achenbach, 1966; Achenbach &
Edelbrock, 1978). Extemalizing behaviors are described as “undercontrolled” behaviors
that are directed toward” others. These behaviors include conduct problems,
hyperactivity, and aggression; disruptive behaviors fall into this category. Intemalizing
behaviors, in contrast, are described as “overcontrolled” behaviors that are directed
toward the self (i.e. withdrawn, depression, anxiety) (Mash & Dozois, 2003).

Traditionally, externalizing behavior has been examined ﬁ'om a medical model
perspective that focuses on within-child factors (i.e. biological functioning,
temperament). However, over the past thirty years, the work of Urie Bronfenbrenner has

highlighted the importance of understanding children’s development and adjustment in

relation to the various contexts in which youngsters reside (Brofenbrenner, 1979, 1992).
His ecological perspective challenges scholars and those working with children to
consider the multiple contextual inﬂuences on human behavior. These contexts include
the family, peers, communities, work places, schools, and larger macro-system variables
(e.g., culture and societal values).

Of the various contexts in which individuals reside, the family environment has
perhaps been the most studied contextual inﬂuence on children’s development
(Werthamer-Larsson, Kellam, & Wheeler, 1991). For example, children’s externalizing
behaviors are best understood as developing from reciprocal interactions between the
child’s temperament and behavior and negative parenting (Patterson, 1982). However,
from an ecological perspective, it is important to consider other contexts, as they are also
likely to affect children’s development. One such setting that has the potential to
signiﬁcantly shape children’s behavioral adjustment is the classroom. Starting from a
young age, children spend a large amount of time in school settings, second only to the
home environment. Schools and classrooms afford exposure to more adults and children
in settings with rules and practices that are often different from those of their home life
(Perry & Weinstein, 1998).

There is a signiﬁcant amount of research examining structures and practices
within the school and classroom environment that contribute to the shaping of children’s
adjustment. Much of this research has been within the school effectiveness and school
climate domains. Early research in these areas focused on physical, structural, and
organizational inﬂuences on student achievement. These factors are easily observable

and measurable (e. g., size of school, student body composition, amount of resources).

IQ

Scholars have since shifted focus to classroom- and school-level processes such as
instruction, staffcollaboration, teacher-student relationships, and peer relationships
(Anderson, 1982; Good & Weinstein, 1986). While some of the aforementioned
variables (i.e. relationships) reﬂect the social context, it is only in more recent years that
greater attention has been given to the social context of the school environment.

The classroom social context has been conceptualized using a variety of
constructs including teacher-student conﬂict, teacher and peer support and relationships,
peer acceptance and rejection, friendships, classroom membership, and belongingness
(Baker, 1999; Battistich & Horn, 1997; Battistich, Solomon, Kim, Watson, & Schaps,
1995; Howes, 2000; Kuperminc, Leadbeater, & Blatt, 2001; Kuperminc, Leadbeater,
Emmons, & Blatt, 1997). Less commonly, the classroom social context has also been
conceptualized as the collective competencies and problem behaviors of students in the
classroom (Hoglund & Leadbeater, 2004; Perry & Weinstein, 1998; Werthamer-Larsson
et al, 1991).

The latter line of research is valuable as it allows us to examine how the
characteristics of a group (i.e. classroom) might affect individual externalizing behaviors.
Groups have been deﬁned as “a collection of interacting individuals who have some
degree of reciprocal inﬂuence over one another” (Rubin, Bukowski, & Parker, 1998, p.
626). The interactions among group members (students) form group characteristics that
are not necessarily present in the individual relationship themselves. Social leaming
theory posits that children learn behaviors and attitudes by observing their peers.
Therefore, we can infer that the behavioral characteristics of students in a classroom

would signiﬁcantly shape individual behavior.

Early work examining the inﬂuence of group composition on children’s
development focused on academic competence, race and SES on student achievement
(Anderson, 1982; Good & Weinstein, 1986). However, more recently, scholars have
broadened their view of student outcomes to include psychological and behavioral
indicators such as externalizing and aggressive behaviors. The extant literature suggests
that an aggressive classroom climate affects students’ problem behavior. However, there
are several limitations to the existing research base. First, much of our understanding is
limited to the intervention literature which has studied the negative effects of aggregating
clinical samples of aggressive children into intervention groups. Second, these studies
focus on male youth. Even when females are included in the sample, there is often little
discussion and analysis of gender differences. An understanding of whether the social
climate is related to student aggression differently given gender seems important given
that the expression and frequency of externalizing behaviors are different among male
and female youth (Barkley, 2003; Hinshaw & Lee, 2003). Third, despite developmental
differences in the expression and frequency studies have not examined whether the
classroom social climate affects student aggression differently for older verses younger
children. Finally, classroom social climate is almost always assessed by measuring the
aggregate aggression in the classroom while the aggregate effects of positive
characteristics, such as social competency on externalizing behavior are largely ignored.
A classroom characterized by a low level of aggression does not necessarily indicate that
the group is composed of socially competent students exhibiting behaviors that allow for

positive and successful interactions.

The present study adds to the theoretical and empirical literature base by
examining how the social context of the classroom affects the development of aggressive
behavior over the course ofa school year among elementary school students. The social
context was deﬁned as the collective aggression and collective social competency of the
students in the classroom. Gender differences, students’ initial level of aggressive

behaviors, and developmental level were examined as possible moderators.

Chapter Two
Literature Review
This section will begin with an in-depth review of the theoretical framework
guiding this study. A discussion of Bandura’s social learning theory will be followed by
an examination of the existing literature on classroom aggression and social competency.
Due to the limited works on the aforementioned topics, the intervention and aggression
literature will also be examined to help delineate the empirical connection between
group-level aggression and competency and the development of aggressive behaviors
among children. In addition, the empirical and theoretical literature on gender,
individual behavior (i.e. initial levels of aggressive behavior), and developmental
differences in externalizing behaviors will be discussed to address their potential
moderating inﬂuence.
Theoretical Framework
Bandura’s social learning theory (1977, 1983, 1992) has been inﬂuential in
understanding the effects of the classroom peer environment on the development of
children’s externalizing behaviors. Social learning theory (SLT) combines the behavioral
and cognitive disciplines to provide a framework for how individuals learn and develop
within a social context. Central to social learning theory is the inﬂuence of models on
individual behavior. Modeling serves to teach observers about new behaviors, including
when these behaviors are most appropriate. In addition modeling functions as a social
prompt, facilitating already acquired behaviors. These ideas will be further explored in
the following section with connections made to how the classroom environment may

inﬂuence students’ disruptive and acting out behaviors.

Observational learning.

A core assumption of social learning theory is that learning can occur without
direct experience and trial and error (Bandura, 1973, 1977, 1983). Learning is posited to
largely occur as a result of observing others or models (i.e. observational leaming).
Observational learning provides a safe and expedited mechanism for learning.
Individuals do not have to perform complex tasks themselves, which if done inaccurately,
may lead to harmful results. Learning from direct experience entails learning through
trial and error, which may be time consuming. However, observing a competent model
provides the observer with knowledge of how and when to execute a behavior leading to
acquisition of behaviors in a timelier manner than trial and error would allow.

In order for observational learning to occur models must exist within the leamer’s
environment. However, the mere existence of others (models) does not automatically
lead to observational learning. Four processes are thought to be necessary in order for
one to learn through observation: attention, memory, motor production, and motivation.

Four necessary components for observational learning.

In order for one to learn a behavior through observation they must attend to the
model(s). There are numerous factors that affect attention including observer
characteristics (i.e. attentional capacity), environmental features, and model attributes
(e.g., age, gender, competence). One factor thought to be especially important is the
extent to which the observer perceives similarity to the model (Schunk, 1987). Perceived
similarity between model and observer lends theoretical support to the assertion that
peers are powerful models, and consequently, that the classroom serves as a powerful

context for observational learning. Peers are similar in their developmental level and

needs making them relevant sources of information to children about how to successfully
behave to achieve goals.

The second and third factors necessary for observational learning to occur are the
observer’s ability to retain the observed behavior and the ability to physically produce the
behavior. In accordance with SLT, children’s ability to retain and accurately produce
behaviors they have observed is enhanced within the classroom environment as children
are provided with ample opportunities to observe and practice behaviors. Speciﬁcally,
SLT leads us to hypothesize that within a highly aggressive classroom environment,
children are provided with many occasions to view their peers engage in aggressive acts
fostering retention. Further, an aggressive environment provides a context in which
students are more likely to be able to practice the behavior.

However, acquisition of behavior does not automatically mean the behavior will
be executed. The fourth process involved in observational learning is motivation.
Modeling an aggressive act depends largely upon whether a child believes they can
successfully execute the behavior to achieve the desired outcomes. This phenomenon is
known as self-efﬁcacy. Students’ self-efﬁcacy is fostered when children observe models
that are similar to themselves (i.e. peers) engage in disruptive behaviors that lead to
successful outcomes. It is through vicarious experiences that children form expectations
of whether they will successfully be able to execute a behavior. When peers are
aggressive and their aggressive acts result in a desired outcome (e. g., toys, attention),
observers are vicariously reinforced. Conversely, when peers are punished for their

aggressive acts, observers are vicariously punished When children see their peers obtain

some desired outcome via aggressive means, they will also be more likely to behave in
disruptive ways.

Another component ofmotivation is the extent to which one ﬁnds the observed
behavior functional. From a social learning perspective, the functional value of
aggression is higher in aggressive environments than non-aggressive environments.
Interestingly, normally passive children are posited to learn to aggress when they see that
counteraggression is successful in stopping others from victimizing them physically and
emotionally. Aggression for a normally passive child becomes functional when
approached by aggressive peers (Bandura, 1983; Warren, Schoppelrey, Moberg, &
McDonald, 2005).

Observation of models or peers does not solely function to facilitate acquisition of
new behavioral patters. The behaviors of models also serve as social prompts for
observers to exhibit preexisting behaviors (Bandura, 1973, 1977). This phenomenon is
referred to as response facilitation. Observing peers engage in aggressive behaviors and
the subsequent consequences of their behavior provides observers information about the
acceptability of aggression. In highly aggressive classrooms, children observe many acts
of aggression by their peers, informing children that aggression in the classroom is
socially acceptable, normalizing the behaviors, and subsequently decreasing their
inhibition to aggress themselves.

In sum, students’ classroom peers are powerful socialization agents, passing on
information that facilitates acquisition of new behavioral patterns, information about
when those behaviors are most likely to be successful, and the norms and values of

particular behavioral patterns in a given context. Within a social learning framework, it

then makes sense to hypothesize that children’s externalizing behavior will be affected by
their classroom peers. Specifically, individual children’s aggressive behavior would be
expected to increase over the course ofa school year in classrooms characterized by high
levels of aggression compared to classrooms with low levels of aggression. Further, in
classrooms characterized by high levels of social competency, children’s individual
externalizing behavior would be expected to decrease because the classroom context
would not support observational learning and response facilitation of aggressive,
disruptive behavior.

Research on Classroom Social Context and Children’s Aggressive Behavior

Research on the inﬂuence of classroom level aggression and social competence
on externalizing behavior is scarce for the former and virtually nonexistent for the latter.
As discussed earlier, this study is framed within social learning theory; that is peers
within group contexts inﬂuence the behaviors of individual students. Therefore, we can
draw from the aggression and intervention literature that has examined the effects of
grouping aggressive peers on individual students’ later aggressive behaviors.

Research on classroom level aggression.

The extant literature provides inconsistent and conﬂicting ﬁndings on the
relationship between classroom aggression and students’ individual aggressive behaviors.
In several longitudinal studies, classroom aggression did not predict students’
externalizing behavior among lower elementary students. Among an urban, Northeast
sample, Werthamer-Larsson, et al. (1991.) found no differences on teacher-rated
aggression between students in “poor” behaving classrooms and “non-poor” behaving

classrooms as measured by aggregating student conduct grades. However, a limitation of

10

this study includes the use ofconduct grades to measure the behavioral characteristic of
the classrooms.

Hoglund and Leadbeater (2004) also did not ﬁnd a relationship between
classroom aggression and externalizing behaviors. They examined the inﬂuence of
student-reported classroom victimization on teacher-rated disruptiveness and aggression
among a Canadian and predominantly Caucasian sample. Additionally, among a low-
income racially diverse, upper elementary sample, Henry, Guerra, Huesmann, Tolan,
VanAcker, and Eron (2000) found that teacher-rated classroom aggression did not predict
change in peer-rated individual students’ aggression.

However, other studies have found that classroom level aggression did predict
change in students’ aggression. Boxer, Edwards-Leeper, Goldstein, Musher—Eizenman,
and Dubow (2003) conducted a cross-sectional study to examine whether student’s
exposure to peer aggression as a witness and victim in school was related to their
individual student-reported aggressive behavior among an upper elementary school
sample. The authors found that an aggressive school environment was associated with
higher levels of individual aggression. While a relationship was established between
exposure to aggression and individual levels of aggression, the study’s cross-sectional
design does not allow for inferences about causation.

Using a longitudinal design, Kellam, Ling, Merisca, Brown, and Ialongo (1998)
examined whether the level of classroom aggression during 1St and 2nd grade predicted
students’ individual aggressive behavior in middle school. The authors found a
signiﬁcant effect on the level of classroom aggression to middle school aggression.

Students in high aggressive classrooms in the 1S” and 2nd grades were found to be more

11

aggressive in middle school compared to students in low aggression classrooms.
However, a limitation of this study was the lack of control for other potential inﬂuences
on student behavior between 1St and 2nd grades and middle school, making it difﬁcult to
infer that aggression in lSt grade classrooms alone contributed to students aggressive
behavior in middle school.

Using a sample of 5’h and 6’h grade students Barth, Dunlap, Dane, Loehman, and
Wells (2004) examined how variations in the classroom context as deﬁned by academic
achievement and behavioral problems inﬂuenced student behavior over the course of the
year and in subsequent years. The authors found that exposure to aggressive classroom
environments predicted change in student behavior during that school year such that
negative classroom behavior was related to increases in negative student behavior.
However, when students’ behavior was assessed during the following academic year
sustained change in student behavior was not found contradicting the ﬁndings of Kellam
et al. (1998), and raising the question about whether classroom environments necessarily
affect students’ behavior in the long-term.

One study has looked more closely at this question among a sample of 1St through
3rd grade students (Thomas, Beirman, & The Conduct Problems Prevention Research
Group, 2006). The researchers found that overall children’s exposure to aggressive
classroom contexts during their ﬁrst three years of elementary school contributed to
negative behavioral adjustment; however, they also found signiﬁcant differences on
student behavior given the number of years of exposure. Speciﬁcally, students with three
years of exposure to aggressive classrooms had signiﬁcantly higher 3rd grade aggression

scores than those exposed to two years of aggressive classrooms. In turn. those exposed

12

to aggressive classrooms for two years had higher scores in 3rd grade then those with one
year or no exposure. This research suggests strong cumulative effects of exposure to
highly aggressive classrooms.

Drawing from the aggression intervention literature may be fruitful as scholars in
this area have examined how the level of aggression in small group interventions affects
students’ externalizing behavior. At the forefront of this work are Dishion and his
colleagues who have indicated concern over the negative effects of aggregating
aggressive youth into small groups. Their concerns have stemmed from a review of
empirical work including their own investigations that suggest aggressive youths’
behaviors escalated after being involved in intervention groups that were composed of
other aggressive youths (Boxer, Guerra, Huesmann, & Morales, 2005; Dishion, McCord,
& Poulin, 1999). While their work evidences the ill effects of aggregating aggressive
youth together, these studies are limited to male, adolescent, and clinical samples.

Empirical research on classroom-level social competencies

As stated earlier, negative group level characteristics (e. g. aggression) have been
examined much more frequently than positive group level characteristics (e.g. social
competency). This pattern of research is consistent with traditional research that has
focused on the inﬂuence of negative characteristics and deﬁcits on children’s adjustment.
The widely held assumption that the lack of negative characteristics is automatically
related to the presences of positive characteristics has been called into question.
Researchers now believe that strengths such as adaptive skills and social competency are

not necessarily present when individuals exhibit low levels of negative behaviors or

13

characteristics (Seligman & C sikszentmihalyi, 2000), underling the importance of
measuring both students’ negative and positive behavioral characteristics.

From a social learning perspective, positive group level characteristics, such as
social competency, like group level aggression, affects the individual behavioral
development of students. In classrooms made up of socially competent children, the
norms of acceptable behavior may be quite different than the norms within classrooms
characterized by disruptive students. Consequently, disruptive students may ﬁnd that
their typical behavioral patterns are not effective in the presence of many socially
competent peers and subsequently decrease their use of aggressive behavior.

While a strong theoretical rationale exists for hypothesizing that classrooms with
high aggregative levels of social competency will be related to decreased student
aggression, the research base for this relationship is scant. Two studies were found that
examined the effects of positive group level characteristics on students’ behavioral
adjustment in the classroom. Among an urban, Canadian sample, Leadbeater, Hoglund,
and Woods (2003) investigated whether 1St grade classroom characteristics including
classroom levels of prosocial behaviors predicted student reports of victimization in the
2"d grade for students with emotional and behavioral problems. They found that prosocial
behaviors were not related to decreases in victimization. Interestingly they found that
among students with emotional problems, higher classroom levels of prosocial behavior
were related to increased reports of victimization. No changes were found for students
with behavioral problems.

The authors conducted another study in which they examined whether the

concentration of prosocial behaviors in the classroom was predictive of behavioral

l4

problems in a longitudinal study among a 1St grade sample of Canadian children
predominantly of Caucasian-European descent (Hoglund & Leadbeater, 2004). Student
behavior was measured by children’s self-reports of peer prosocial behaviors (e. g.,
helping, sharing, caring behaviors). The authors found that prosocial levels in the
classroom did not mitigate aggressive and disruptive behaviors in children. The authors
again found that classroom levels of prosocial behavior did not decrease reports of
victimization. Given the nature of the samples it is not possible to generalize these
ﬁndings to other populations (e. g., non-Caucasian European descent and older children).
Further, students rated their peers’ prosocial behaviors and given research suggesting
differences between how adults rate children’s behaviors, how peers rate children, and
how children rate themselves the limitations of these results should be noted (Kamphaus
& Frick, 2002).

Some research has been conducted within the aggression intervention literature.
In response to the large body of intervention research that suggests grouping aggressive
youth together exacerbates individual aggressive behavior, a small body of research has
examined whether aggregating high-risk students into groups composed of socially
competent children might inﬂuence the trajectory of externalizing behaviors. The two
studies that have examined the aforementioned relationship have indicated conﬂicting
ﬁndings. Feldman, Caplinger, and Wodarski (1983) examined whether grouping
aggressive children, ages 9 to 16-years old with other aggressive children (pure groups)
and grouping aggressive children with non-aggressive children (mixed groups) lead to

different outcomes on children’s disruptive behaviors. They found that children who

15

participated in the mixed groups displayed less disruptive behavior in and out of the
group sessions than children in the pure groups.

However, more recently Mager, Milich, Harris, and Howard (2005) found that
early adolescents in pure intervention groups exhibited greater decreases on teacher-and
parent-reported externalizing behavior than those in mixed groups at post-intervention.
However, at the 6-month follow-up, no differences existed between the two groups. In
addition, the authors found that youth in the pure group behaved more appropriately (i.e.
participation and complimenting) than those in the mixed group during the sessions.
Mixed groups consisted of a 2:1 ratio of non-aggressive to aggressive students.

Due to limited research and inconsistent ﬁndings, the existing literature on
classroom level competency does not afford adequate information to make predictions or
hypotheses about the effects of classroom-level competency on students’ externalizing
behavior. Though empirical support is lacking, strong theoretical rationale does exist for
hypothesizing that students in “competent” classrooms will exhibit fewer disruptive
behaviors.

Initial risk of aggression.

A. couple of studies have examined whether the classroom social environment
inﬂuences the development of aggressive behavior differently given students’ initial (or
baseline) levels of behavioral problems. However, these studies have only examined the
moderating inﬂuence of students’ initial risk of externalizing behavior for classroom
level aggression. Research was not found for classroom level competency.

As part of their analysis, Kellam et al. (1998) examined whether students’ initial

level of aggression interacted with the classroom level of aggression to inﬂuence

l6

students’ later aggression in middle school. They found that an interaction did exist such
that students with high baseline aggression, in highly aggressive lSt grade classrooms,
were found to be most at-risk for severe aggression in middle school compared to
students of the same level of aggression in less aggressive classrooms and to students
with lower levels of aggression in highly aggressive classrooms. These results indicate
that aggressive classroom environments are particularly detrimental for already
aggressive students.

In a similarly designed study among elementary school students, Warren et al.
(2005) found that highly aggressive classrooms were related to more aggressive
behaviors among students at a 2-year follow—up. However, this relationship was stronger
for students who had higher initial levels of aggression. Warren et al.’s ﬁndings suggest
that students’ with varying levels of aggression react to classroom aggression in a similar
pattern but with different intensities. Further, their ﬁndings are consistent with Kellam et
al. (1998) both indicating that exposure to an aggressive peer environment is particularly
harmful for children already at-risk for aggressive behavior.

Interestingly, within the aggression intervention literature, Boxer et al. (2005)
have proposed the “discrepancy—proportional peer inﬂuence” principle after ﬁnding that
the more discrepant a target child’s behavior was from that of the group level, the more
likely the child’s behavior would change in the direction of the peer group’s average.
This principal was proposed after the authors found that changes in individual student
aggression was greatest the more discrepant the initial level of student aggression was
from group level aggression among an elementary school sample of high risk students for

externalizing behavior. They found that ifa student was more aggressive than the group

17

level, the student was likely to decrease in aggression over time. However, if the student
was less aggressive than the group, the student was likely to increase in aggression over
time. The discrepancy—proportional peer inﬂuence principle does not reﬂect the ﬁndings
of Kellam et a1 (1998) and Warren et a1 (2005). They found that students with high levels
of externalizing behavior in highly aggressive classrooms were more at-risk for later
externalizing behaviors than their less aggressive peer in the highly aggressive
classrooms. However, it should be noted that their samples were not restricted to only
participants identiﬁed as high-risk for externalizing problems.

Classroom studies indicate that exposure to an aggressive classroom environment
is particularly harmful for children already at high-risk for aggressive behavior. These
studies help us to understand for whom an aggressive classroom environment is
particularly harmful by grouping students into low-and high-risk groups. However,

9“

another way to conceptualize the inﬂuence of students at-riskness” on the classroom
social environment and later aggressive behavior is to examine this relationship across
the range of initial levels of risk (aggression scores). The use of this kind of methodology
would allow for examination of the moderating effects of initial level of risk of
aggressive behavior on the effects of the classroom social environment and later
aggression.

As indicated above there is gap in the literature with regard to how socially
competent classrooms may affect students’ development of aggression given students’
level of initial “risk.” As discussed earlier, Leadbeater et al. (2003) investigated whether

1St grade classroom characteristics including classroom levels of prosocial behaviors

predicted student reports of victimization in the 2nd grade for students with emotional and

18

behavioral problems. They found that while among students with emotional problems,
higher classroom levels of prosocial behavior were related to increased reports of
victimization, students with behavioral problems were not affected by classroom levels of
prosocial behavior. Assuming that classrooms with higher levels of social competency
are related to lower levels of student aggressive behavior, does this relationship look
different across the range of students’ initial level of risk? It may be reasonable to
hypothesize that students who start the school year exhibiting lower levels of aggression
(i.e. are at low risk), may not ﬁnd more socially competent classrooms as inﬂuential on
their later aggressive behavior than students with higher initial levels of risk. For
students with higher initial levels of aggression, more socially competent classrooms may
be more meaningful to their later externalizing behavior because these students
experience that their current behavioral patterns are not functional, subsequently
changing the way they interact with others in order to get their needs met.

Gender differences.

The literature on externalizing behavior suggests signiﬁcant differences in the
prevalence of externalization among males and females. A large body of work indicates
that males tend to be more temperamental, hyperactive, aggressive and disinhibited than
their female counterparts (Barkley, 2003; Hinshaw & Lee, 2003; Keenan & Shaw, 1997).
Further, externalizing behaviors among males have been found to not simply be
expressions of antisocial behaviors but are manifested in the everyday play activities of
male children (Dipietro, 1981). Given these differences found between males and
females one could speculate that gender has a moderating inﬂuence on the relationship

between classroom level aggression and later aggressive behavior.

19

The research to date on whether gender differences exist for the relationship
between an aggressive classroom context and aggressive behavior is limited to two
studies that provide conﬂicting results. In an exploratory analysis, due to a small sample
size, of lst grade students, Kellam et al. (1998) found that an aggressive social context
did not affect female students, though the context did predict males’ aggressive behavior
in middle school. However, using a longitudinal design, Hanish, Martin, Fabes, Leonard,
and Herzog (2005) found that gender moderated the effects of an aggressive climate on
individual extemalization among low-risk (i.e. two parent household, high SES families),
Kindergarten students. The speciﬁcally found that exposure to externalizing peers within
the classroom predicted aggression, hyperactivity, and anxiety among females. However,
signiﬁcant ﬁndings were not found among males.

The ﬁndings from Hanish et al. (2005) are inconsistent with the results of Kellam
et al. (1998) and ﬁndings within the intervention aggression research that has consistently
found aggressive contexts to predict subsequent externalizing behaviors among male
youth (e.g., Boxer et al., 2005; Dishion et a1, 1999; Vitaro, Tremblay, Kerr, Pagani &
Bukowski, 1997). However, it should be noted that these studies use a clinical sample
mostly comprised of adolescent males while the ﬁndings of Kellam et al. and Hanish et
al. used early elementary aged, nonclinical male samples.

Interestingly, Hanish et al. (2005) have theorized about why gender may be a
moderator for social context and later externalizing behavior. They posit two processes
for understanding gender differences: normative process and accumulation process. The
normative process predicts that because an aggressive behavior is not typically a

normative experience for females as it is for males, exposure to an aggressive

20

environment would be more detrimental to females’ development. The accumulation
model, however, predicts that because boys are exposed to more aggressive behavior in
their social contexts, higher levels of behavioral problems will result in boys than girls.
The ﬁndings from their study support the accumulation model. Further research is
needed to better understand whether males and females are affected differently by their
classroom social context as characterized by aggression and social competency.

Unlike the externalizing literature, research on social competence fails to provide
consistent data for gender differences (Rubin et a1, 1998), which may support speculation
that gender does not act upon the relationship between classroom level social competency
and later aggression. The lack of empirical data makes it difﬁcult to hypothesize whether
gender has a moderating inﬂuence.

Developmental differences.

The literature on aggression indicates developmental differences in the frequency,
expression, and function of aggressive behavior. As children move from the preschool to
middle childhood years they engage in aggression less frequently. Furthermore, there is a
general trend of physical aggression decreasing while verbal and relational aggression
increases, paralleling the shift in function of aggressive behaviors. Young children’s
aggressive behavior is more likely to be instrumental, meaning aggressive behaviors
serve to attain some tangible or activity whereas as children get older aggressive acts are
more likely to be person-oriented and hostile. Related to the function of aggression, the
topography of aggression also changes from overt, physical acts to covert, verbally

aggressive acts (Coie & Dodge, 1998; Parker, Rubin, Price, & DeRosier, 1995).

21

While there is signiﬁcant research on the above developmental changes with
regard to aggression, less is know about how the social context of classrooms might
differentially affect students given their developmental level. The research that has
examined the effects of the social context of the classroom (i.e. classroom level
aggression and classroom level social competency) on students’ individual aggression has
rarely included study of developmental differences. The aforementioned studies (i.e.
Werthamer-Larsson et al., 1991; Hoglund and Leadbeater, 2004; Henry et al., 2000;
Boxer et al., 2003; Kellam et al., 1998) have all included samples of elementary aged
students. In contrast, research within the aggression intervention ﬁeld has primarily
focused on the grouping effects among adolescents at-risk for aggression. One exception
is a study conducted by Boxer et al. (2005), who examined the effects of aggregating
groups of students in grades 3 and 6 exhibiting high levels of aggression into small group
interventions. The authors found that for students in both grades, exposure to more
aggressive intervention groups was related to increases in individual student aggression.
However, analyses were not done to determine whether the strength of the relation
between group level aggression and student aggression was moderated by grade level.

Social learning theory has a strong cognitive basis in which abilities to attend,
retain and make social comparisons are necessary precursors for the process of
observational learning to occur (Bandura, 1973). Within the child development literature,
it is well established that as children grow older, they acquire more sophisticated and
increased cognitive capacities. While studies have shown that children’s ability to both
attend to and retain information from observation increases with age, little attention has

been given to the developmental changes in observational learning (Parke & Slaby,

22

1983). Given the cognitive abilities that are needed for observational learning to occur,
learning effects may be more pronounced for older children. However, while older
children may be more capable of learning aggressive behavior through observation than
younger children, older children are also less likely to ﬁnd aggression socially acceptable
(Coie & Dodge, 1998).

Summary

Children’s aggressive behaviors are of great concern to parents and teachers. In
fact, children are most likely referred to school and community agencies because of
disruptive and acting out behaviors. An ecological framework pushes us to consider how
the classroom social context, an environment in which children spend a tremendous
amount of time, affects the trajectory of children’s aggressive behavior. One way to
conceptualize the social context is to consider the collective aggression and social
competency of the students in the classroom. Social learning theory posits that children
are inﬂuenced by the behaviors of their peers. Thus, when peers engage in aggressive,
acting out behaviors, or in prosocial ways children are socialized to behave in the same
way.

A review of the relevant empirical literature highlights that a consistent body of
research is lacking regarding the inﬂuence of the classroom social context on students’
aggressive behavior. This problem is exacerbated by the dearth of research about for
whom the classroom context might be most important. Three speciﬁc areas that would
be important to further study include an individual’s initial level of aggression, gender,
and developmental level as moderators. Hi gh-risk students for externalizing problems

have been found to have a poorer trajectory across adjustment domains. Males have

23

been evidenced to exhibit signiﬁcantly higher levels of behavioral problems than females.
Additionally, aggression has been documented to shift in prevalence and expression
across the elementary school years. Finally, the social context, deﬁned as aggregating
student characteristics has been limited to negative indicators, namely aggression.
However, it would be incorrect to assume that non-aggressive classrooms are prosocial,
adaptive, and socially competent. If children are embedded within a classroom social
context in which their peers are collectively socially competent, will their aggressive
behaviors change across the school year? From a social learning theory, one might
hypothesize that within a socially competent classroom, children will learn more adaptive
behavior, and their aggressive behavior will subsequently decrease.

The study expands upon the literature by examining how both the collective
behavioral problems and social competency of students in the classroom affect individual
students’ aggressive behaviors. In addition, the study explored for whom the social
environment of the classroom is important. Speciﬁcally, the moderating inﬂuence of
children’s initial level of aggressive behaviors, gender, and developmental level were
examined.

Research Questions and Hypotheses

1. Is the level of classroom aggression/classroom social competency related to
change in students’ aggressive behaviors over the course of a school year after
controlling for initial levels of aggressive behavior? It was hypothesized that the
aggressive behavior of children will be affected by the classroom social context. Using a
social learning perspective it was predicted that in aggressive classrooms, children would

become more aggressive over the course of the school year. Conversely, in classrooms

24

characterized by high levels of social competency, children were expected to become less
aggressive over the course of the school year.

2. Does the relationship between classroom aggression/ social competency and
students’ development of aggressive behavior vary across the range of students’ initial
level of aggression? It was hypothesized that there would be a positive relationship
between classroom aggression and individual aggressive behaviors for all students
regardless of initial levels. However, based on previous research ﬁndings it was
predicted that the relationship between classroom aggression and later aggressive
behavior would become stronger as the initial level of risk increases. A negative
relationship was hypothesized for classroom social competency and students’ aggressive
behavior for all students. The relationship was predicted to be stronger at higher initial
levels of risk.

3. Does gender moderate the inﬂuence of classroom level aggression and
classroom level social competency on aggressive behavior? From a social learning
perspective, the author hypothesized that for both female and male students there would
be a positive relationship between classroom level aggression and aggressive behavior
and a negative relationship between classroom level social competency and aggressive
behavior. This question was exploratory.

4. Does developmental level moderate the inﬂuence of classroom
aggression/social competency on students’ aggressive behavior? It is hypothesized that
the social context will affect students’ aggressive behavior at the end of the school year.
However, given the lack of research studying the moderating inﬂuence of developmental

level this research question is exploratory.

Chapter Three
Methods
Participants

The participants in this study were drawn from Addressing the Context of .
Teaching for Behaviorally At-Risk Students (ACT EARLY), a three-year longitudinal
study (Baker, Kamphaus, & Home, 1999). Participants from the ACT Early data set
included 1,434 kindergarten through 5’h grade students from four elementary schools
located in a small city in the Southeast. The school district has a large at-risk student
population due to poverty. Approximately 70% of the students participate in the free and
reduced cost lunch program. Additionally, a large percentage of the students from each
school live in public housing units and there is less than a 50% on—time graduation rate
from high school in the district. Poverty and race are confounded in the district with
more African American students eligible for the free and reduced cost lunch program
than students of other races.

The current study drew participants from the larger three-year study. In order to
eliminate dependency in the data contributed by multi-year participants, only one year of
data was selected. There were a total of 697 participants in the data year selected for
analysis. In order to examine change in participants’ aggression over time, participants
were included in this study if they had data collected at time 1 and time 2. There were
652 participants who had complete data sets leaving 45 participants with missing data, all
for time 2 data collected in the Spring. Additionally, classrooms with fewer than 10
students were not included to ensure that classroom group level dynamics were being

measured. Two classrooms were eliminated due to insufﬁcient classroom sizes, leading

26

to an additional 12 participants being eliminated from the data set. One classroom had 9
students and the other classroom had 3 students. This left 8% of the original sample
being eliminated due to missing data or insufﬁcient classroom size for a total sample of
640.

Children. The sample included 640 students (52% female) consisting of 14%
kindergartens, 21% 1St graders, 15% 2nd graders, 13% 3rd graders, 16% 4th graders and
21% 5th graders. The racial makeup of the participants consisted of 58% African
American, 29% White, 9% Hispanic, 2% Asian, and 2% Multi-racial.

Teachers. In the current study there were 39 teacher participants. The sample
was 98% female, 83% White and 18% African American. Sixty percent of teachers had
7 or more years of teaching experience and only 5% were in their ﬁrst year of teaching.
Additionally, 69% of teachers had a degree beyond a 4-year bachelor’s degree.

Classrooms. Each of the 39 teachers had a separate classroom resulting in 39
different classrooms. Ten percent of the classrooms were Kindergarten classrooms, 25%
were 1St grade classrooms, 18% were 2nd grade classrooms, 13% were 3rd grade
classrooms, 15% were 4th grade classrooms, and 20% were 5th grade classrooms.
Classrooms ranged in size from 11 to 22 students. As indicated earlier, classrooms with
fewer than 10 students were not included to ensure that classroom group dynamics were
being measured. Two classrooms were eliminated.

Procedure
All children in the four participating schools were recruited to participate in the
study. Parental consent forms were sent home at the beginning of the school year in their

native language and only those students with active parental consent were included. The

27

participation rate was over 90% at each of the schools. Prior to participating in the study,
teachers provided informed consent and received a small stipend for their participation.
For each of the participating students, teachers completed a standardized behavior rating
scale, in the fall, about 10-12 weeks into the school year and again in the spring.
Teachers completed other study measures as part of the larger study.

Measures

Aggressive Behavior. The Behavior Assessment System for Children Teacher
Rating Scales for Children (BASC-TRS-C; Reynolds & Kamphaus, 1992) was used to
assess individual students’ aggressive behavior. The BASC-TRS-C is a nationally
standardized, reliable and valid measure. It consists of 148 items in which teachers rate
the frequency of problem and adaptive behaviors on a 4-point scale. There are nine
problem behavior scales (Aggression, Conduct Problems, Hyperactivity, Anxiety,
Depression, Somatization, Attention Problems, Learning Problems, and Atypicality) and
4 adaptive scales (Adaptability, Leadership, Social Skills, and Study Skills) as well as
ﬁve composite scores (Extemalizing Problems, Intemalizing Problems, School Problems,
Behavioral Symptoms Index, and Adaptability). Standard scores (T-scores) are used to
indicate behavioral adjustment.

Children’s aggressive behavior was measured using the Aggression Subscale.
Ratings of student aggression were collected at two time points: time 1 was during the
fall and time 2 was in the spring of the academic school year. The Aggression Subscale
measures both physical and verbal aggression such as arguing, name-calling, verbally
threatening others, breaking others’ possessions, and hitting others. The Cronbach alpha

reliability for the Aggression Subscale ranges from .93 to .95.

28

Classroom Level Aggression. The Aggression Subscale from the BASC-TRS-C
(Reynolds & Kamphaus, 1992) was also used to measure the aggregate level of
classroom aggression. Consistent with previous studies exploring similar research
questions, classroom level aggression was calculated for each child by summing the
aggression t scores of all the children in the class except that of each target child, and
dividing the sum by n-l (Henry et al., 2000; Hoglund & Leadbeater, 2004; Warren et al.,
2005; Werthamer-Larsson etal., 1991). This method ensures that a score is computed for
each child independently of his or her own aggression protecting against potential
confounding effects. Aggression scores collected during the fall were used to calculate
the classroom context.

Classroom Level Social Competency. The Social Skills Subscale from the BASC-
TRS-C (Reynolds & Kamphaus, 1992) is an adaptive scale and was used to measure the
level of social competency in the classrooms. Deﬁnitions of social competency have
typically included the general idea of effective interactions that allow one to meet their
goals and needs, and have often been operationalized by measuring social skills (Rose-
Krasnor, 1997, p. 112). The BASC social skills measure examines the “interpersonal
aspects of social adaptation (p. 51),” which is consistent with existing deﬁnitions of
social competence. Sample items include admitting mistakes, complimenting others,
encouraging others, offering assistance, beginning conversations appropriately, and
saying “please” and “thank you.” The Cronbach alpha reliability ranges from .92 to .93.
Classroom level social competency was computed in the same manner that classroom
level aggression was computed. In each classroom, the social skill t scores collected

during the fall were summed for all children except that of each child, and the total was

29

then divided by n-l to generate a score for the level of social competency in each
classroom for each child. These two scores, classroom aggression and classroom social
competency, were used to reﬂect the classroom social environment to which each child
was exposed during the course ofa school year.

Data Analyses

Multiple regression analyses using general linear modeling (GLM) were
employed to assess whether the classroom social context (class aggression; class social
competency) led to change in students’ individual aggressive behavior over the course of
a school year. The moderating effects of gender and student’s initial level of aggressive
behavior were also examined. In order to control for the shared variance within
classrooms contributed by the nested data (i.e. the same teacher rating all participating
students in the classroom), GLM was used (Baker, 2006). GLM models exhibit non-
normally distributed residuals, violating the assumption of homoscedasticity that is
required in ordinary least squares regression. The current study has a hierarchical data
structure as a result of multiple children being rated by the same teacher. GLM corrects
for the shared variance within classroom.

For each model the shared variance within classrooms was accounted for by
entering in the child’s teacher as a random factor. Main effect and interaction terms were
entered simultaneously into the regression equations. Moderation effects were tested
using procedures recommended by Baron and Kenny (1986). A moderator variable
inﬂuences the relationship (direction and/or strength) between an independent and
dependent variable by interacting with the independent variable to inﬂuence the

dependent variable (Baron & Kenny, 1986). The means of the independent (i.e.

classroom aggression and classroom social competency) and moderator variables (i.e.
time 1 aggression scores collected during the fall) were centered as recommended by
Aiken and West (1991). For the regression equations using gender as a moderator
variable, males were the reference group. For regression equations assessing the
moderating effects of students’ initial level of aggression, the moderator variable was
continuous. To interpret interaction effects that were found in this study, the
unstandardized regression coefﬁcients (B) were used to calculate and interpret the simple
slopes. Untransforrned values were used to calculate the slopes (Aiken & West, 1991).

To test the moderating effects of developmental level on the association of the
classroom social context and individual student aggression at the end of the school year, a
mixed model analysis was conducted. A mixed model analysis was used to control for the
nested nature of the teacher and developmental level (i.e. grade) variables. A mixed
model analysis combines an ANOVA and regression using GLM. The ANOVA tested
the relation of grade level on the dependent variable (i.e. student aggression at the end of
the school year). Regression using GLM assessed the effects of the classroom social
context (classroom aggression; classroom social competency) and the interaction of grade
level and classroom social context on the dependent variable. For grade level two groups
were formed: lower elementary grades (K, 15’ and 2nd grades) and upper elementary

grades (3“, 4th and 5’h grades). The upper elementary group served as the reference group.

31

Chapter 4
Results
Preliminary Anal yses

Descriptive statistics and visual inspection of the data were conducted to assess
the adequacy of the study variables and integrity of analyses. Means, standard deviations,
medians, and correlations of study variables are presented in Tables 1 and 2. Correlations
among the independent variables did not indicate problems with multicollinearity.
Multicollinearity exists when the independent variables are highly correlated at .7 or
higher (Tabachnik & Fiddell, 2001).

An examination of whether the study variables were normally distributed did
indicate problems with normality. The Kolmogrov-Smirnov statistic indicated violations
of normality for aggression at time 1 and classroom social competency. Further
inspection of normal Q-Q plots indicated signiﬁcant deviations from normality for time 1
aggression with points deviating from the line at the lower and upper ends. In addition,
signiﬁcant positive skewness for time 1 aggression and moderate negative skewness for
classroom social competency were observed in the data. Outliers were inspected for the
variables using boxplots, which indicated the existence of outliers in the scores of the
variables. To determine the extent to which the outliers were affecting the data the
original means were compared to recalculated means that excluded the top and bottom
5% of the data (i.e. eliminating the outliers). Little difference was found in the original
and recalculated means indicating that the outliers were not greatly affecting the data
(Pallant, 2001), so the outliers were retained in the data ﬁle. Due to signiﬁcant violations

of normality time 1 aggression and classroom social competency variables were

32

transformed using logarithmic transformations (Pallant, 2001; Tabachnik and F iddell,
2001)

Classroom aggression and classroom social competency were measured by
computing a mean score for the level of aggression and social competency in the
classroom for each student. Given the skewed nature of the variables and the tendency
for mean scores to be affected by outliers within the data, median scores were examined
to determine whether the mean scores accurately reﬂected the group-level of aggression
and social competency. While the median is also a measure of central tendency, median
scores represent the middle score within a data set and are less inﬂuenced by extreme
scores. Median and mean scores (see Table 1) for both classroom aggression and
classroom social competency were very similar indicating that the mean scores were not
inﬂuenced by extreme scores and accurately reﬂect the group level of aggression and
social competency.

An adequate sample size affects both power (the probability of detecting
statistical signiﬁcant results) and generalizability. Tabachnik and Fiddell (2001, p. 117)
provide a formula for calculating sample size requirements that takes into account the
number of independent variables: N > 50 + 80(IV). Using this formula, this current study
needed an N of 210 (50 + 80(2)) to test each proposed model. This study has a sample N
of640.

Results

Across all models, signiﬁcant values were found for the teacher identiﬁcation

variable indicating signiﬁcant differences between classrooms on the dependent variable

(i.e. time 2 aggression). Effect sizes indicated moderately strong differences between

33

classroom teachers on ratings of student time 2 aggression at the end of the school year.
Table 3 presents the results of all the models tested.

The ﬁrst set of hypotheses examined the contribution of the classroom social
context (aggression; social competency) to students’ time 2 aggression at the end of the
school year after controlling for students’ level of time 1 aggression at the start of the
school year. Higher levels of classroom aggression were predicted to be related to
increases in student aggression from time 1 to time 2, and a higher level of classroom
social competency was predicted to lead to decreases in student aggression from time 1 to
time 2. Two regression analyses using GLM were conducted to test these hypotheses.
The ﬁndings did not support the hypotheses. The level of classroom aggression was
found to be signiﬁcantly related to change in students’ aggression from the beginning of
the school year (time 1 aggression) to the end of the school year (time 2 aggression), F (1,
599) =167.53, p <.0005. A small effect size was indicated (1]2 =22). Effect sizes were
interpreted using Cohen’s (1988) recommendations. Results indicated a negative
relationship such that higher levels of classroom aggression were related to decreases in
student aggression from time 1 to time 2.

The level of classroom social competency was also found to be signiﬁcantly
related to change in students’ aggression, F (1,599) = 44.89, p <.0005 with a fairly small
effect size of .07. The direction of change in aggression was unexpected. Classrooms
with higher aggregate levels of social competency led to increases in time 2 aggression at

the end of the school year.

34

The second set of hypotheses examined whether students’ initial level of
aggression (time 1 aggression) moderated the relationship between the classroom social
context (aggression; social context) and change in students’ aggression.

Classroom level aggression was hypothesized to be positively related to students’
individual aggression regardless of their initial level of aggression (time 1 aggression);
however, it was predicted that the association between classroom aggression and change
in student aggression would be stronger for students who started out the school year with
higher levels of time 1 aggression. The ﬁndings do not support the hypothesis. A
statistically signiﬁcant though small main effect was found for classroom level
aggression, F (1, 598) = 142.59, p <.0005, n2 = .19. The ﬁndings indicate that higher
levels of classroom aggression are related to decreases in student time 2 aggression. A
main effect albeit very small was also found for students’ initial level of aggression (time
1 aggression), F (1, 598) = 16.07, p<.0005, n2 = .03. Again, the direction of the change
in aggression from the fall to spring was not anticipated with ﬁndings yielding a negative
relationship between time 1 aggression scores in the fall and time 2 aggression scores in
the spring. Higher levels of time 1 aggression were related to decreases in time 2
aggression at the end of the school year. An interaction effect was not found, F (1, 598) =
.918, p =.339.

Higher levels of classroom social competency were predicted to be related to
decreases in student time 2 aggression across the range of time 1 aggression scores;
however, the relationship was hypothesized to be stronger at higher initial levels of
aggression (time 1 aggression). An interaction effect was found, F (1, 598) = 11.53,

p=.001; however, the effect size was very small, n2: .02. While a signiﬁcant interaction

35

effect was found, the hypothesis was not supported. At higher levels of initial aggression
(time 1 aggression), classrooms that had higher levels of aggregate social competency
were found to be related to decreases in student time 2 aggression. However, at lower
initial levels of aggression (time 1 aggression), classrooms with higher levels of social
competency were related to increases in student time 2 aggression (see Figure l).

The third set of hypotheses examined whether gender moderated the association
between the classroom social context (aggression; social competency) and change in
students’ aggression from the beginning (time 1 aggression) to end of the school year
(time 2 aggression).

A statistically signiﬁcant main effect for classroom level aggression was found, F
(1, 598) = 1203.41, p<.0005 with a large effect size (n2 =.67). The association between
classroom level aggression and time 2 aggression at the end of the school year was
negative. A main effect for gender was also found, F (1 , 598) = 8.4, p=.004; however the
effect size was very small (rj2 =.01). Males had higher time 2 aggression scores than
females. The interaction effect did not reach statistical signiﬁcance, F (1, 598) = 3.25,
p=.07.

A signiﬁcant main effect for classroom social competency was indicated, F (1,
598) = 247.93, p<.00005 with a moderate effect size (112 = .29). This was a positive
relationship. Higher levels of classroom social competency were associated with higher
levels of time 2 aggression at the end of the school year. The main effect for gender F (1,
598) = .002, p=.96 and the interaction effect, F (1 , 598) =.001, p=.98 did not reach

statistical signiﬁcance.

36

The fourth set of hypotheses examined whether developmental level (i.e. grade)
has a moderating effect on the association between the classroom social context
(aggression; social competency) and change in students’ aggression over the course of a
school year. An interaction effect was found, F (1, 623) = 4.29, p=.04, see Figure 3.
Students’ developmental level was operationalized by forming two groups based on
grades: lower elementary (K-lSt grades and upper elementary (3“1-5’h grades). As
classrooms became more aggressive, students’ time 2 aggression scores decreased for
both lower and upper elementary students. However, as the classrooms become more
aggressive, the gap between the two lines depicting the grade levels becomes slightly
larger indicating that upper elementary school students’ time 2 aggression decreased
more signiﬁcantly than lower elementary school students. A signiﬁcant interaction for
classroom social competency was also found, F (1, 530) = 5.55, p=.02. For both upper
and lower elementary school students, higher levels of classroom social competency were
related to increases in student time 2 aggression. However, this effect was signiﬁcantly
stronger for lower elementary school students with a much steeper line upwards indicated

(see Figure 3).

Chapter 5
Discussion

This ﬁnal chapter will brieﬂy restate the research problem and review the
methods utilized in this study. The majority of this section will summarize the results
and discuss their implications. Incidents of disruptive behavior among school-age youth
have increased (Achenbach et al., 2003) which is alarming given that disruptive and
aggressive behavior is related to poor academic, behavioral and social adjustment among
students who engage in these behaviors repeatedly (Loeber, 1990). Given the signiﬁcant
amount of time children spend in schools, an important context to consider in the
development of student behavior is the classroom.

There is a growing literature base that suggests the social context, speciﬁcally the
behavioral characteristics at the classroom level, may be related to students’ behavioral
adjustment. This research is primarily focused on assessing how aggressive classrooms
contribute to students’ own externalizing behavior. Studies have shown that children’s
exposure to classrooms characterized by high rates of aggression is related to increases in
aggressive acts among individual students (e.g., Barth et al., 2004; Kellam et al., 1998;
Thomas et al., 2006; Warren et al., 2005).

From a social learning perspective, the behavioral characteristics of classrooms
affect individual student behavior because of the strong inﬂuence that classroom peers
have on “teaching” other students about behavior through observation. By observing
frequent aggression within the classroom students learn how and when to aggress, that
aggressive acts are normal, socially acceptable and functional within their classroom

setting, leading to subsequent aggression among students.

38

Continued longitudinal research is needed to better understand how exposure to
aggressive classrooms affects the development of children’s aggression. In particular,
studies that examine student characteristics such as developmental level, gender, and
initial risk of aggression are important to understanding for whom the negative classroom
context is especially harmful. These child characteristics are important to consider as the
aggression literature indicates that 1) behaviorally at-risk students have a poorer
adjustment trajectory and are more vulnerable to aggressive group contexts than their
peers exhibiting low rates of aggression, 2) males exhibit signiﬁcantly higher levels of
behavioral problems, and 3) there is a signiﬁcant change in the prevalence and expression
of aggression across the elementary school years. An additional gap in the research is
predominantly the classroom social context has been deﬁned using negative indicators
such as aggression instead of positive indicators such as social competence. This current
study employed a longitudinal design to investigate the inﬂuence of the social context of
the classroom on the development of children’s aggression across an academic school
year among elementary-aged students. The effects of gender, developmental level, and
initial levels of aggression were examined. Consistent with previous research on
aggressive classrooms, the classroom social context was deﬁned as the aggregate level of
aggression in the classroom as indicated by teacher ratings of each student’s behavior;
however, this study also examined positive characteristics of the classroom social
context, social competency.

The following four research questions were examined: 1) Is the classroom social
context (classroom aggression; classroom social competency) related to change in

individual student aggression over the course of a school year? 2) Is the relationship

39

between the classroom social context and change in student aggression different across
the range of students’ initial levels of aggression? 3) Does gender moderate the inﬂuence
of the classroom social context on students’ aggressive behavior? 4) Does developmental
level moderate the inﬂuence of the classroom social context on change in student
aggression over the course of a school year? For research questions one, two and three
multiple regression analyses using general linear modeling (GLM) were used to assess
whether the classroom social context is related to change in students’ aggressive behavior
over the course of a school year. For research question 4 a mixed model analysis was
used to control for the nested nature of the teacher and developmental level variables.
Overall, the study found that the social context of the classroom is related to change in
children’s aggression across an academic year, and that these effects are different given
student characteristics (i.e. initial levels of aggression, gender and developmental level).
However, the ﬁndings were not always in the hypothesized direction.
Effects of A ggressive Classroom Environments

In this study, the level of classroom aggression was found to be related to change
in children’s aggression; however, contrary to the study’s hypothesis, higher levels of
classroom aggression were related to decreases in children’s aggression over the course
of a school year. Additionally, all children regardless of their initial levels of aggression
experienced positive changes in their behavioral adjustment at the end of the school year
when exposed to higher levels of classroom aggression. These ﬁndings are in conﬂict
with previous studies that used similar methodologies to the current study, which found
that aggressive classrooms were related to the development of more aggressive behavior

in children (Barth et al., 2004; Duane et al., 2006; Kellam et al., 1998; Warren et al.,

40

2005). Prior research has also indicated that aggressive classrooms are particularly
harmful to students exhibiting high levels of aggression and disruptive behavior (Kellam
et al., 1998, Warren et al., 2005). This research parallels ﬁndings from the aggressive
intervention literature that has well documented the negative effects of aggregating
clinical and at-risk aggressive youth into small groups. Disruptive youth have been found
to become more aggressive as a result of being exposed to group contexts in which their
peers are also exhibiting high levels of disruptive behavior (Ang & Hughes, 2001;
Dishion et al., 1999).

The negative effect of aggressive group contexts on the development of children’s
own aggressive behavior has been conceptualized within a social learning framework.
From this perspective, classrooms are posited to serve as powerful environments in which
students observe and learn behavioral patterns from their peers. Social learning theory
asserts that observational learning is more likely to occur when the observer perceives the
model as similar to himself. Thus, peers are believed to be strong socialization agents as
they are similar in their characteristics, needs and abilities to other children making them
relevant sources of information about behavior. When children are placed in aggressive
classrooms, they are provided with frequent opportunities to witness and practice
aggressive behaviors. Furthermore, in contexts where there is a high level of aggressive
behavior, a social atmosphere is created in which aggression is typically functional and
normalized making it a socially acceptable behavioral pattern for children to exhibit
(Bandura, 1973; 1977).

Aggressive classrooms and other group contexts are reasoned to be harmful to the

behavioral adjustment of all children; however, the research has indicated more negative

41

effects for students with already high levels ofaggressive behaviors. Within the
aggression intervention literature, the particularly negative effects of aggregating
aggressive youth have been explained by a phenomenon called “deviancy training,”
which is guided by social learning theory. Deviancy training refers to how aggressive
youth within the group context are provided opportunities to become even more
aggressive and disruptive by teaching one another new disruptive behavioral patterns,
practicing new and old behaviors with one another through their informal interactions and
reinforcing negative behaviors (Dishion et al., 1999). Being around other aggressive
youth provides willing participants and an audience that allows for disruptive behavior to
remain valued and functional.

While the ﬁndings from this current study are inconsistent with previous research
and theoretical frameworks there are possible explanations for why classrooms with
higher levels of aggression led to decreases in children’s aggression across a school year.
In response to high levels of aggressive behavior at the beginning of the school year,
teachers may have altered their behavior and utilized more effective classroom
management strategies, including the use of classwide interventions to deal with
aggression. Numerous studies have indicated that implementation of classroom
management strategies and interventions can successfully change the context of the
classroom (e.g., Kellam et al., 1998; Murphy, Theodore, Aloiso, Alric-Edwards, &
Hughes, 2007). Yet, research also indicates that aggressive students tend to actually
decrease teachers’ use of effective behavior management strategies and instead teachers
respond to disruptive students using punitive and coercive strategies that can serve to

reinforce and perpetuate aggressive student behavior (Nelson & Roberts, 2000). The

42

teacher sample in the current study do represent an experienced and educated group with
60% of teachers having 7 or more years of teaching experience and 69% of teachers
having a degree beyond a 4-year bachelor’s degree. It is possible that these teachers may
have been better equipped to effectively respond to aggressive classroom climates at the
onset of the school year.
Effects of Classroom Social Competency

This study also examined the social context of the classroom by examining how
the level of classroom social competency is related to change in student aggression across
a school year. The hypothesis that higher levels of classroom social competency would
be related to decreases in student aggression at the end of the school year was not
supported. It was hypothesized that all students would beneﬁt from being in socially
competent classrooms; however, the positive effects were predicted to be more
pronounced for students with higher initial levels of aggression. The ﬁndings indicated
that at higher and mid levels of initial aggression and at higher levels of classroom social
competency, student aggression at the end of the school year decreased. Furthermore, the
decrease in aggression was more signiﬁcant at higher levels of initial aggression, which
supports the hypothesis. However, at lower levels of initial aggression and higher levels
of classroom social competency, student aggression was found to increase. Thus,
exposure to socially competent classrooms was found to be related to change in
children’s aggression quite differently, in fact in opposite ways, given initial levels of
aggression.

There is little research examining how levels of classroom social competency

affect the development of student aggression. The previous research that does exist

43

appears inconsistent with the findings of the current study. In one study, Leadbeater et al.
(2003) examined the moderating effects of behavioral and emotional problems on
classroom social competency and later reports of victimization. Interestingly, they found
that students with higher levels of emotional problems (i.e. internalizing behaviors) in
classrooms with higher levels of social competency reported increases in relational and
physical victimization. Yet, for students with behavioral problems (i.e. externalizing
behaviors) the level of classroom social competency did not change students’ reports of
victimization. This ﬁnding is inconsistent with the current study that found at higher
levels of initial aggression and higher levels of classroom social competency, decreases
in student aggression were reported. However, the two studies do use different constructs.
The construct of self-reported victimization while helpful to understand the aggressive
nature of a classroom is a different construct from the current study’s construct of teacher
reported aggression. Often students with disruptive and aggressive behaviors
inaccurately interpret the actions of peers and adults as hostile, coercive and aggressive
(Crick & Dodge, 1994). In another study, Hoglund and Leadbeater (2004) found that
classroom levels of prosocial behaviors did not decrease or mitigate student reported
aggressive and disruptive behaviors. This study did not examine variables such as initial
levels of aggression that may have had moderating effects on the development of
aggression.

The ﬁnding that at higher and medium initial levels of aggression, higher levels of
classroom social competency were related to decreases in aggressive behavior is
consistent with social learning theory. In classrooms that are characterized as more

socially competent, students have many opportunities to observe and learn prosocial

44

behaviors from their peers. Furthermore, classrooms in which students are socially
competent may not afford aggressive students with opportunities to practice their
aggressive behaviors or provide functional utility for aggressive behaviors. Students who
aggress toward those who are socially competent may ﬁnd that their aggressive acts are
not rewarded, as those who are socially competent are more likely to be skilled in dealing
with the aggressor. When placed in a context in which aggressive behavior is neither
functional nor reinforced, we would expect student aggression to decrease. In turn, new
behavioral patterns that are more appropriate might emerge as a result of observing the
behaviors of more socially competent peers.

Given the above discussion, it appears that the ﬁnding that at lower initial levels
of aggression, socially competent classrooms were related to increases in aggression at
the end of the school year conﬂicts with social learning theory. The constructs of
aggression and social competency can appear to be behavioral characteristics that are
mutually exclusive of one another or exist on opposite sides of the spectrum (Hawley,
2002). This notion is supported by research suggesting children who are aggressive lack
friends, are disliked by adults and peers, and in general have less successful social
interactions (e. g., Rubin et al., 1998), all indicators of social competence or in this case
social incompetence. Interestingly, there is some research that has shown negative and
aggressive behaviors to be related to social competence. Vaughn, Vollenweider, Bost,
Azria-Evans, and Snider (2003) examined the relationship between negative and
aggressive behaviors and social competence among preschool children. The authors
found that children characterized as socially competent also tended to have higher rates

of initiating negative interactions, described as coercive and dominant, than less socially

45

competent children. These “socially competent” children were found to be more
aggressive and yet also more prosocial then their peers. Their ﬁndings are analogous
with other research indicating similar positive relationships between aggressive
behaviors, such as dominant and coercive behavior, with social competence among
young children (Hawley, 2002; Hawley & Vaughn, 2003).

These researchers have couched their ﬁndings within social ethology and
evolutionary psychology perspectives. These perspectives posit that some children are
particularly apt at controlling resources and do so by their ability to utilize both
aggressive and prosocial behaviors to attain and control needs and desires. These
students are able to differentiate when particular behaviors will allow them to meet their
social goals, and while they may utilize aggressive tactics, they frequently exhibit
prosocial and afﬁliative interactions enabling them to maintain social acceptability
among their peers. Thus, use of aggressive behaviors have been argued by these
researchers to be an indicator of social competence when children are able to differentiate
their behavior to access desired resources and meet social goals within speciﬁc contexts
(Hawley, 2002; Hawley & Vaughn, 2003; Vaughn et al, 2003).

This aforementioned research may help to shed light on this current study’s
ﬁndings. While high classroom levels of social competence may indicate that many
peers were engaging in prosocial behaviors, it does not necessarily mean that there was
not use of aggressive behaviors within these classrooms. As discussed earlier, socially
competent children may also be quite apt at using aggression in functional ways that
allow them to also maintain their social status among peers and teachers. It is possible

that students with lower levels of aggression attended to these aggressive behaviors

46

exhibited by socially competent peers and added them to their own behavioral repertoire
via observational learning. However, for students with much higher levels of aggression,
they may not have the social competencies to be able to differentiate and use aggression
in a context-speciﬁc way. In fact, highly aggressive kids have been found to generally
interpret social situations as aggressive and also respond to social situations with
aggressive behavior (Crick & Dodge, 1994).

Another possibility for an increase in aggression among students entering the
classroom at low levels of aggression may be that they were picked on and taken
advantage of by their socially competent peers, and their subsequent aggression could
have been reactive. This theory may parallel the aforementioned ﬁndings of Leadbeater
et al. (2003) who found that students with higher levels of emotional problems (i.e.,
internalizing behaviors) in classrooms with higher levels of social competency reported
increases in relational and physical victimization. The authors hypothesized that students
with emotional problems were particularly vulnerable to manipulation and coercion by
their socially competent peers.

Gender Differences

Gender differences were studied to better understand whether the social context of
the classroom was related to change in student aggression differently given student
gender. While there is a plethora of research indicating gender differences with regard to
prevalence and expression of aggression (e.g., Barkley, 2003; Keenan & Shaw, 1997;
Hinshaw & Lee, 2003), this study did not ﬁnd gender differences for the effects of the
classroom context on the development of aggression. Similarly, gender differences were

also not found when the classroom context was deﬁned by examining the aggregate

47

levels of social competency in the classroom. The previous research is limited to two
studies (i.e. Hanish et al., 2005; Kellam et al., 1995) and the ﬁndings were inconsistent.

Interestingly, Hanish et al (1995) found that females were detrimentally affected
by aggressive classroom contexts whereas males were not. These ﬁndings conﬂicted
with another study conducted by Kellam et a1 (1998) who found that exposure to
aggressive classrooms climates was harmful for boys but not girls. The aggression
intervention literature has also well documented the deleterious effects of exposures to
aggressive peers among males. The ﬁndings from the current study suggest that for both
males and females, higher levels of classroom aggression are related to decreases in
individual aggression over the course of the school year.

While presently there is not strong theory to support gender differences, two
processes have been posited to explain why gender differences may be evident:
normative and accumulation processes. Hanish et al. (1995) proposed that aggressive
environments may be more detrimental to female development because aggressive
behavior is not the typical female experience. The accumulation process was proposed to
provide rationale for why aggressive contexts may be more detrimental for males.
According to this process, males are predicted to be more detrimentally affected by
aggressive environments because they are typically exposed to aggressive behaviors
within their social contexts, and the cumulative effects lead to more negative outcomes
for males compared to females who do not regularly experience aggression.
Developmental Level Differences

Developmental differences in the prevalence, form, and function of aggression

has also been well documented in the literature. As children move from early to middle

48

childhood, aggressive acts decrease, change from being instrumental (i.e. trying to get
something) in nature to more person-directed and hostile and become less overt (i.e.
physical) and more covert (i.e. verbal aggression, insults, threats) (Parker et al., 1995;
Coie & Dodge, 1998). Consistent with prior research, this study found that younger
students had higher reported rates of aggression than older elementary students. This
study furthered the research by ﬁnding that the classroom context was related to change
in students’ aggression across a school year differently given developmental level. More
aggressive classrooms were found to lead to decreases in student aggression for both
upper (3-5) and lower (K-2) elementary aged students. However, as classroom
aggression increased, more substantial decreases in aggression over the course of the
school year were found among older students. Previous research has not explored how
the classroom context might differentially affect the development of students’ aggression
given their developmental level.

For children to be able to learn behaviors from their peers via observational
learning certain abilities must be present including being able to attend, retain and make
social comparisons (Bandura, 1973). Given that these cognitive abilities develop and
become more sophisticated as children become older, it is possible that older elementary
school aged children would be more likely to be affected by the classroom social context.

The social context of the classroom deﬁned by social competency was also found
to be related to students’ development of aggression differently given developmental
level. Again, for both students, higher levels of social competency were related to
increases in aggression at the end of the school year; however, the increase was

signiﬁcantly more pronounced for younger students. As documented in the literature,

49

younger children use more aggressive and coercive acts and aggression among younger
children is typically instrumental in nature (Parker et al., 1995). In others words, younger
children are more likely to utilize aggression to gain access to something they want or
need. The authors of the above-cited work on the positive relationship between social
competence and aggression (i.e., Hawley, 2002; Hawley & Vaughn, 2003, Vaughn et al.,
2003) suggest the relationship between the two constructs is stronger for younger
children. These authors argue that as children become older they View aggression as less
socially acceptable than their younger peers, and they will utilize more prosocial
strategies to attain resources. It is possible that in this study younger children’s more
signiﬁcant increase in aggression as a result of exposure to socially competent classrooms
compared to older children is reﬂective of developmental differences in the relationship
between social competence and aggression. Younger children engaging in socially
competent behaviors may have been more likely to also be engaging in aggressive
behaviors than older students providing more opportunities for observational learning in
the lower elementary classrooms. Further, when exposed to aggression, older students
might have bee less likely to actually engage in the aggressive acts compared to their
younger peers given beliefs about the acceptability of aggression.

Limitations

While this study provides interesting and valuable information, a few limitations should
be noted when interpreting the results. First, the behavior rating scales were ﬁrst
administered to teachers ten to twelve weeks after the school year started. At ten weeks,
classroom level behavioral characteristics may have already been ﬁrmly established and

as a result, individual student behavior may have already been affected by the classroom

50

environment. Thus, scores measuring students’ aggression at the start of the school year
may have already reﬂected exposure to the classroom social context, compromising
interpretation of the moderating effects of students’ initial levels of aggression.

Second, this study measured classroom and student aggression with teaching
rating scales. There is the possibility that teacher biases contributed to the ﬁndings.
Safeguards were taken to minimize any potential biases by using general linear modeling,
which accounted for the fact that teachers were rating several children and by calculating
a classroom mean for each child.

Third, while this study has added to the literature on classroom social context by
examining both negative and positive indicators of group level behavioral adjustment (i.e.
aggression and social competency), this study focused on measuring negative outcomes
(i.e. aggression) and moderators (i.e. initial levels of aggression) of behavioral
adjustment. Examining how the classroom context affects the development of students’
social competency and how students’ initial level of social competency moderates the
effects of student outcomes would have provided a better understanding of how exposure
to the classroom social context affects students’ behavioral adjustment. Given that
aggression and social competency are not dichotomous variables, it is not sufﬁcient to
assume a negative relationship between the two variables.

Fourth, while this study helps to understand the inﬂuences of exposure to
particular classroom contexts on students’ adjustment across a year, long-term inﬂuences
on student adjustment was not assessed. There is some evidence indicating that while
aggressive contexts inﬂuence change student behavior, change is not necessarily carried

across school years (e. g., Barth et a1, 2004; Thomas et al., 2006). An important variable

51

is argued to be the number of years of exposure to aggressive contexts. Barth and
colleagues found that while students with one year of exposure indicated elevated
aggression, this was not necessarily evident in subsequent years. Students with multiple
years of exposure to aggressive classrooms were found to have the most signiﬁcant
increases in aggression suggesting that exposure to aggressive classrooms has cumulative
effects on child behavioral adjustment.

Finally, this study did not account for other contextual inﬂuences on the
development of student aggression over the course of the school year. While group level
behavioral characteristics have been shown to inﬂuence individual student behavior, it is
important to note that these group level characteristics occur within a context in which
there are other strong inﬂuences such as teacher behavior. It is possible that there were
mediating effects such that classroom levels of behavior led to changes in teacher
behavior (e. g., use of effective or ineffective classroom management skills, interventions)
that in turn inﬂuenced the classroom social climate and subsequently the development of
individual student aggression.

Implications for Future Research and Practice

Despite the noted limitations, the results of this study have numerous implications
for future research and school based practice and policy. As indicated above a limitation
of this study is that teachers assessing student behavior may have provided biased ratings
of student aggression. Future studies should incorporate other methods for assessing
behavior such as direct observations and/or using multiple rating sources (e. g., student

and teacher reports) to decrease any potential teacher bias.

52

Secondly, future research should also examine how negative and positive group
level behavioral characteristics affect the development of both students’ positive and
negative adjustment. The ﬁndings from this study indicate that positive and negative
classroom characteristics do not necessarily lead to decreases and increases in negative
student outcomes, respectively, as one might hypothesize. The ﬁndings in this study
challenge us to rethink the seemingly dichotomous nature of social competency and
aggression. This has signiﬁcant implications for school professionals intervening with
aggressive behavior and attempting to promote prosocial behaviors. The most commonly
used practices within schools to address aggressive and disruptive behavior are punitive
strategies (e.g., detention and suspensions) (Mayer, 1995). The reliance on such practices
is problematic because they focus only on decreasing inappropriate behavior and do not
teach students appropriate social skills to allow them to get their needs met in socially
acceptable ways. As the research is suggesting, just because a student is not engaging in
a disruptive behavior does not necessarily mean that they are engaging in prosocial
behaviors that are important for life long adjustment. Reactive strategies that focus on
reducing disruptive behavior is not sufﬁcient for promoting positive social behaviors and
foster social competency within children.

As noted above, a limitation of this current study was the lack of long-term
analysis on how the classroom environment affected students’ behavior. Change was
only assessed across a school year. Previous studies have indicated conﬂicting results on
whether change in student behavior has long-term effects. Further, studies often fail to
examine how subsequent classroom experiences may contribute to behavioral adjustment.

Research should continue to examine how classroom environments shape behavior across

53

an academic year, whether there is any signiﬁcant long-term inﬂuence on behavior, and
the mediating variables on long-term inﬂuences. Studies will need to assess the
subsequent classroom environments to which children are exposed to best understand the
inﬂuence of classroom environments on behavioral adjustment.

Finally, while this study did not account for teacher-mediated variables on the
development of aggression, future research should also include teacher-mediated
processes such as teacher use of classroom management strategies and classroom based
interventions. Additionally, assessing the academic environment including instructional
match to student skill and curriculum and instructional strategies and practices should
also be included to understand how these contextual factors may mediate the effects of
classroom behavioral characteristics on individual students’ behavioral adjustment.

In conclusion, while the ﬁndings of this study are not necessarily consistent with
prior research, social learning theory or proposed hypotheses, they do indicate that the
collective behavioral characteristics (aggression and social competency) of students
within a classroom contribute to individual student levels of aggression. This study
contributes to the literature base by examining negative and positive classroom
behavioral characteristics and ﬁnding that aggressive contexts and socially competent
classrooms may not necessarily indicate deleterious or positive effects, respectively, on
students’ adjustment. Further, this study has added to the literature by examining the
relation between the classroom context and change in student aggression given gender,

developmental level and initial levels of aggression.

54

Table 1
Means, Medians, and Standard Deviations for Aggression Subscale and Classroom Level
Variables

 

 

N Mean SD Median
Aggression Time 2 640 52.67 12.23 --
(Spﬁng)
Aggression Time 1 640 51 .25 l 1.48 --
(Fall)
Classroom Aggression 640 51.25 4.77 51.09
Classroom Social 640 50.89 6.06 51.42
Competency

 

55

Table 2
Correlationsfor Study Variables

 

 

1 2 3 4
1. Aggression Time 1 (Fall) _ .84** .27** .09*
2. Aggression Time 2 (Spring) _ .28** .1 1**
3. Classroom Aggression .08*

4. Classroom Social Competency

 

** Signiﬁcant at the .01 level
* Signiﬁcant at the .05 level

56

Table 3

Regression Analyses Predicting Student Aggression from Classroom Social Context and

illoderator Variables (Gender and Initial Risk)

 

 

 

Model Aggression at Time 2 (Spring)
F B 112
Classroom Social Context
Classroom Aggression
(df=] ,5 99 )
Student Aggression Time 1 (Fall) 16.07*** -6.59+ .03
Classroom Aggression 167.53*** -9.93 .22
Teacher (df= 38, 599) 9.69*** .38
Classroom Social Competency
(df=],599 )
Aggression Time 1 (Fall) 499.46*** -21.9+ .46
Classroom Social Competency 44.89*** 770.62+ .07
Teacher (df= 38, 599) 5.7*** .27
Initial Risk (aggression level)
Classroom Aggression
(df= 1, 599 )
Classroom Aggression 142.68*** -9.68 .19
Student Aggression Time 1 l6.07*** -7.38+ .03
Class Aggression X std agng .918 -.176+ .002
Teacher (df= 38, 598) 7.78*** .33

 

57

 

Table 3 (cont'd)

 

 

MOdCl Aggression at Time 2 (Spring)
F B n2
Classroom Social Competency
(df=] ,598)
Classroom Social Competency 19.16*** 565.03+ .03
Student Aggression Time 1 13.99*** -237.72+ .02
Class SC X std agng 11.53** -107.64+ .02
Teacher (cl/"=38, 598) 4.98*** .24
Gender
Classroom Aggression
(df=l, 598)
1203.41 ** -12.64 .67
Classroom Aggression *
Gender 8.40* * -1.49 .01
Gender X Classroom Aggression 3.25 .192 .01
Teacher (df=38, 598) 40.45*** .72
Classroom Social Competency
(df=] ,598)
247.93*** 2118.03+ .29
Classroom Social Competency
Gender .002 -2.87 .00
Gender X Classroom SC .001 .80 .00
Teacher (df=38, 598) 12.51*** .45

 

+Unstandardized Betas represent untransformed values (signs)
***p<.001 **p<.01 *p<.05

58

Table 4
Mixed Model Analysis (A nova and Regression) Predicting Student Aggression from
Classroom Social Context and Moderator Variables (Grade level)

 

Model Standard error E _ B
of estimate (estimates)

 

Classroom Aggression

(df=],623)

Grade Level 20.21 .89 19.09
Classroom Aggression .494 695.61 *** -13.04
Class Aggression X Grade .702 4.29 1.45
Classroom Social

Competency

(df = 1, 530)

Grade Level 483.81 5.87** 1172.41
Classroom Social 175.70 153.8“ 2178.95+
Competency

Class SC X Grade 241.87 5.55* -570.02+

 

*Unstandardized Betas represent untransformed values (signs)
***p<.001 **p<.01 *p<.05

59

Figure 1.

Interaction of Classroom Level Social Competency and Students ’ Initial Level of
Aggression on Aggression at the End of the School Year

20000
18000 — _ — -
16000 iv“ -7 ._ - -/

14000 2,___ __. . / _ ..

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

<2 12000 4 _L- ~eeeeeeL—— 1//,4//, -e—————-

E: 1:333 _H __ / ___ _ __ "___g +LowAggT1
= g

g 6000 ——- / -— —4 +MedAgng
:53 WOO—__.: / ' “'——‘“ +HighAgng
< 2000 /\"T -- -- -

 

 

 

 

 

 

 

-2000 - ___-ISO .. 08D \30

-4000-————u——_ _-_—_n~.._ ”______._
~6000

 

 

 

 

 

Level of Classroom Social Competency

6O

Figure 2.
Interaction of Classroom Level Aggression and Grade Level on Aggression at the End of
the School Year

 

 

140- ~ 7 , 7* L - e , e f ** ~ , - e j
12 - l
10 \\

80 i
\ ,I 1
60 ‘
\ “"upper
" “" lower
40 \
l

.131) OSD \so g

 

 

 

 

Aggression at Time 2

 

-2

 

 

-4
Classroom Aggression

61

Figure 3.

Interaction of Classroom Level Social Competency and Grade Level on Aggression at the
End of the School Year

20000", ' ' _" ' " "

l

15000 i

 

 

, “i {/4 + upper
5000 * lower

 

 

 

 

Aggression at Time 2

 

 

-1so 030 130

 

-5000 i I

-10000 b ,-

 

Classroom Social

62

REFERENCES

Achenbach, T.M. (1966). The classiﬁcation of children's psychiatric symptoms: a
factor analytic study. Psychological Monographs: General and Applied, 80, 1-37.

Achenbach, T.M. & Edelbrock, CS (1978). The classiﬁcation of child
psychopathology: A review and analysis of empirical efforts. Psychological
Bulletin, 35, 1275-1301.

Achenbach, T.M., Dummenci, L., & Rescorla, LA. (2003). Are American children's
problems still getting worse? A 23-year comparison. Journal of Abnormal Child
Psychology, 31, 1-1 1.

Aiken, L.S. & West, S.G (1991). Multiple Regression: Testing and Interpreting
Interactions. Thousand Oaks, California: Sage Publications, Inc.

Ang, R.P., & Hughes, J .N. (2001). Differential beneﬁts of skills training with antisocial
youth based on group composition: 3 meta analytic investigation. School
Psychology Review, 31, 164-185.

Anderson, CS. (1982). The search for school climate: a review of the research. Review
of Educational Research, 52, 368-420.

Baker, J .A. (1999). Teacher-student interaction in urban at-risk classrooms: Differential

behavior, relationship quality, and student satisfaction with school. Elementary
School Journal, 100, 57-70.

Baker, J .A. (2006). Contributions of teacher-child relationships to positive school
adjustment during elementary school. Journal of School Psychology, 44, 211-229.

Baker, J. A., Kamphaus, R. W., & Home, A. M. (1999). Project A.C.T. Early [data ﬁle].
Institute for At-Risk Children of the Ofﬁce of Educational Research and

Improvement, United States Department of Education.

Bandura, A. (1973). Aggression a Social Learning Analysis. Oxford, England: Prentice-
Hall.

Bandura, A. (1977). Social Learning Theory. Englewood Cliffs: Prentice-Hall.
Bandura, A. (1983). Psychological mechanisms of aggression. In Green, Russel G. &

Donnerstein, E. (Eds), Aggression: theoretical and empirical reviews (pp. 1-
40). New York: Academic Press, Inc.

63

Bandura, A. (1992). Social cognitive theory. In Ross Vasta (Ed.). Six theories of child
development: revised formulations and current issues. London: Jessica Kingsley
Publishers, Ltd.

Barkley, R.A. (2003). Attention-deﬁcit/hyperactivity disorder. In Mash, E. J. & Barkley,
R. (Eds), Child psychopathology (pp. 75-143). New York: Guilford Press.

Baron, R. M. & Kenny, D. A. (1996). The moderator-mediator variable distinction in
social psychological research: conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51, 1 173-1 182.

Barth, J .M., Dunlap, S.T., Dane, H., Loehman, J .E., & Wells, KC. (2004). Classroom
enviromnent inﬂuences on aggression, peer relations, and academic focus.
Journal of School Psychology, 42, 115-133.

Battistich, V. & Hom, A. (1997). The relationship between student's sense of
their school as a community and their involvement in problem behaviors.
American Journal of Public Health, 87, 1997-2001.

Battistich, V., Solomon, D., Kim, D., Watson, M., & Schaps, E. (1995). Schools as
communities, poverty levels of student populations, and students’ attitudes,

motives, and performance: a multilevel analysis. American Educational Research
Journal, 32, 627-658.

Boxer, P., Edwards-Leeper, L., Goldstein, S. E., Musher-Eizenman, D., & Dubow, E. E.
(2003). Exposure to "low-level" aggression in school: association with

aggressive behavior, future expectations, and perceived safety. Violence and
Victims, I 8, 691-705.

Boxer, P.L., Guerra, N.G., Huesmann, L.R., Morales, J. (2005). Proximal peer-level
effects of a small group selected prevention on aggression in elementary school
children: an investigation of the peer contagion hypothesis. Journal of A bnormal
Child Psychology, 33 (3), 325-338.

Brofenbrenner, U. (1979). The ecology of human development. Cambridge, Mass:
Harvard University Press.

Brofenbrenner, U. (1989). Ecological systems theory. Annals of Child Development, 6,
187-249.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Lawrence Earlbaum Associates.

Coie, J .K., & Dodge, K.A. (1998). Aggression and antisocial behavior. In W. Damon, &

N. Eisenberg (Eds), Handbook of child psychology (5”) Volume 3 Social,
emotional, and personality development. (pp.779-861) NY: Wiley.

64

 

Coie, J .D. & Kupersmidt, J .B. (1983). A behavioral analysis of emerging social
status in boys’ groups. Child Development, 54, 1400-1416.

Crick & Dodge (1994). A review and reformulation of social information processing
mechanisms in chidlren’s social adjustment. Psychological Bulletin, 115, 74-101.

Dodge, K.A. (1983). Behavioral antecedents of peer social status. Child Development, 54,
1386-1399.

DiPietro, J .A. (1981). Rough and tumble play: A function of gender. Developmental
Psychology, 1 7, 50-58.

Dishion, T.J., McCord, J ., & Poulin, F. (1999). When interventions harm: peer groups
and problem behavior. American Psychologist, 54, 755-764.

Dishion, T.J., Spacklen, K.M., Andrews, D.W., & Patterson, GR. (1996). Deviancy
training in male adolescent friendships. Behavior Therapy, 27, 373—390.

Feldman, R. A., Wodarski, J. S., & Flax, N. (1983). Prosocial and antisocial boys
together. Social Work, 18, 26-37.

Good, T. L. & Weinstein, R. S. (1986). Schools make a difference: evidence, criticisms,
and new directions. American Psychologist, 41, 1090-1097.

Hawley, RH. (1999). The ontogenesis of social dominance: a strategy based
evolutionary perspective. Developmental Review, 19, 37- 132.

Hawley, PH. (2002). Social dominance and prosocial and coercive strategies of resource
control in preschoolers. International Journal of Behavioral Development, 26,
167-176.

Hawley, P.H. & Vaughn, BE. (2003). Aggression and adaptive functioning: the bright
side to bad behavior. Merrill-Palmer Quarterly, 49, 239-400.

Hanish, L.D., Martin, C.L., Fabes, R.A., Leonard, S., & Herzog, M. (2005). Exposure to
externalizing peers in early childhood: homphily and peer contagion processes.
Journal of Abnormal Child Psychology, 33, 267-281.

Henry, D., Guerra, N., Huesmann, T., Patrick, V., Richard, & Eron, L. (2000). Normative
inﬂuences on aggression in urban elementary school classrooms. American
Journal of Community Psychology, 28, 59-81.

Hinshaw, S., P. & Lee, S., S. (2003). Conduct and oppositional deﬁant disorders. In

Mash, E.J. & Barkley, R. (Eds), Child psychopathology (pp. 144-199). New
York: Guilford Press.

65

Hoglund, W. L. & Leadbeater, BL. (2004). The effects of family, school, and
classroom ecologies on changes in children's social competence and emotional
and behavioral problems in ﬁrst grade. Developmental Psychology, 40, 533-544.

Howes, C. (2000). Social-emotional classroom climate in child care, child-teacher
relationships and children's second grade peer relations. Social Development, 9,
191-204.

Kamphaus, R.W. & F rick, P.J. (2002). Clinical assessment of child and adolescent
personality and behavior (2'1“ Ed). Boston: Allyn and Bacon.

Kellam, S.G., Ling, X., Merisca, R., Brown, C.H., & Ialongo, N. (1998). The effect of
the level of aggression in the ﬁrst grade classroom on the course of malleability of
aggressive behavior in middle school. Development and Psychopathology, 10,

165-185.

Keenan, K. & Shaw, D. (1997). Developmental and social inﬂuences on young girls’
early problem behavior. Psychological Bulletin, 121, 95-113.

Kuperminc, G.P., Leadbeater, B.J., & Blatt, SJ. (2001). School social
climate and individual differences in vulnerability to psychopathology among
middle school students. Journal of School Psychology, 39, 141-159.

Kuperminc, G.P., Leadbeater, B.J., Emmons, C. & Blatt, SJ. (1997). Perceived school
climate and difﬁculties in the social adjustment of middle school students.
Applied Developmental Science, I, 76-88.

Leadbeater, , B., Hoglund, W., & Woods, T. (2003). Changing contexts? The effects of a
primary prevention program on classroom levels of peer relational and physical
victimization Journal of Community Psychology, 31, 397- 418.

Loeber, R. (1990). Development and risk factors of juvenile antisocial behavior and
delinquency. Clinical Psychology Review, 10, 1-41.

Mager, W., Milich, R., Harris, M.J., & Howard, A. (2005) Intervention groups for
adolescents with conduct problems: is aggregation harmful or helpful? Journal of
Abnormal Child Psychology, 33 , 349-362.

Mash, E.J. & Dozois, David J. A. (2003). Child psychopathology: a developmental
systems perspective. In Mash, Eric J. & Barkley, Russel (Eds), Child
psychopathology (pp. 3-75). New York: Guilford Press.

Mayer, GR. (1995). Preventing antisocial behavior in schools. Journal of Applied
Behavior Analysis, 28, 467-478.

66

IV'lurphy, KA., Theodore L.A., Aloiso, D., Alric-Iﬂidwards. J.M., & Ilughes TL. (2007).

lnterdependent group contingency and mystery motivators to reduce preschool
disruptive behavior. Psychologt.’ in the Schools, 44. 53.63.

Nelson, J .R. & Roberts, ML. (2000). Ongoing reciprocal teacher-student interactions
involving disruptive behaviors in general education classrooms. Journal of
Emotional and Behavioral Disorders, 8, 27-3 7.

Pallant, J. (2001) SPSS Survival Manual: A Step by Step Guide to Data Analysis Using
SPSS Version 12 (2"d Ed).Chicago: Open University Press.

Parke, R.D., & Slaby, R.G. (1983). The development of aggression. In P. Mussen &
E.M. Hetherington (Eds), Handbook of child psychology: V01 4. Socialization,
personality and social development (pp. 547-641). New York: Wiley.

Parker, J ., Rubin, K., Price, J ., & de Rosier, M. (1995). Peer relationships, child
development, and adjustment. In D. Cicchetti, & D. Cohen (Eds), Developmental

psychopathology: Vol 2. Risk, disorder, and adaptation (pp. 96—161). New York:
Wiley.

Patterson, GR. (1982). Coercive family processes. Eugene, OR: Castalia.

Perry, K. E. & Weinstein, RS. (1998). The social context of the early schooling
and children's school adjustment. Educational Psychologist, 33, 177-194.

Reynolds, C.R. & Kamphaus, R.W. (1992). BASC: Behavior assessment system for
children manual. Circle Pines, MN: American Guidelines Services, Inc.

Rose Krasnor, L. (1997). The nature of social competence: a theoretical review. Social
Development, 6 , 111-129.

Rubin, K.H., Bukowski, W., & Parker, J .G. (1998). Peer interactions, relationships, and
groups. In W. Damon & N. Eisenberg (Eds), Handbook of child psychology vol.
3 social, emotional, and personality development (pp. 619-700). New York:
Wiley.

Schunk, DH. (1987). Peer models and children's behavioral change. Review of
Educational Research, 57, 149-174.

Seligman, M.E.P & Csikszentmihalyi, M. (2000). Positive psychology: an introduction.
American Psychologist, 55, 5-14.

Tabachnik, B.G. & Fiddell, LS. (2001 ). Using Multivariate Statistics, 4:}: Eds. Boston:
Allyn & Bacon.

67

Thomas, D.E., Beirman, K.L., & The Conduct Problems Prevention Research
Group (2006). The impact of classroom aggression on the development of aggressive
behavior problems in children. Development and Psychopathology, 18, 471-487.

Vaughn, B.E., Vollenweider, M., Bost, K.K., Azria-Evans, M.R., & Snider, J .B. (2003).
Negative interactions and social competence for preschool children in two
samples: reconsidering the interpretation of aggressive behavior for young
children. Merrill-Palmer Quarterly, 49, 245-278.

Vitaro, R., Tremblay, R.E., Kerr, M., Pagani, L., & Bukowski, W.M. (1997).
Disruptiveness, friends' characteristics, and delinquency in early adolescence: a
test of two competing models of development. Child Development, 68, 676-689.

Warren, K., Schoppelrey, S., Moberg, D.P., & McDonald, M. (2005). A model of
contagion through competition in the aggressive behaviors of elementary school
students. Journal of Abnormal Child Psychology, 33, 283-292.

Werthamer-Larsson, L., Kellam, S., & Wheeler, L. (1991). Effect of ﬁrst grade classroom

environment on shy behavior, aggressive behavior, and concentration problems.
American Journal of Community Psychology, 19, 585-602.

68

 

   

lilililiiliiliiiil