You are on page 1of 14

J Youth Adolescence (2014) 43:1890–1902

DOI 10.1007/s10964-013-0051-1

EMPIRICAL RESEARCH

Unique and Interactive Effects of Empathy, Family, and School


Factors on Early Adolescents’ Aggression
Milena Batanova • Alexandra Loukas

Received: 14 June 2013 / Accepted: 22 October 2013 / Published online: 7 November 2013
Ó Springer Science+Business Media New York 2013

Abstract Although research indicates that empathy school relationships among boys, as well as to more
inhibits youth aggression, little is known about the pro- heavily consider the role of emotion processes in the study
spective associations between different components of and prevention of early adolescents’ aggression.
empathy and aggression, as well as whether family and
school factors moderate the aforementioned associations in Keywords Empathy  Aggression  Family  School 
early adolescents. Based on prior research, the current Moderation  Early adolescence
study examined whether empathic concern and perspective
taking would contribute to subsequent overt and relational
aggression over a 1-year period in middle school. Guided Introduction
by the social development model, we also examined if
positive family relations and school connectedness would Although a number of studies demonstrate that empathy
differentially moderate the associations between both inhibits aggressive behaviors in children and adolescents
components of empathy and aggression. Participants were (Gini et al. 2007; Robinson et al. 2007; Strayer and Roberts
481 10- to 14-year old students (54 % female; 78 % 2004), researchers increasingly recognize that empathy
European American) who completed the first wave of a involves cognitive (e.g., perspective taking) and affective
survey in 6th and 7th grades. Hierarchical regression (e.g., empathic concern) components that can carry different
analyses indicated that only for girls did lower levels of implications for aggression (Caravita et al. 2009, 2010;
empathic concern, not perspective taking, contribute to Jolliffe and Farrington 2006, 2011; Espelage et al. 2004).
increases in subsequent overt, not relational, aggression. While researchers also recognize that there are multiple
Although neither positive family relations nor school forms of aggression, such as overt and relational aggression,
connectedness played protective roles for girls, results limited extant research has assessed the prospective asso-
indicated that boys’ reports of positive family relations ciations between both components of empathy and both
buffered the negative impact of low empathic concern on forms of aggression during early adolescence (see Batanova
both forms of aggression 1 year later. Over and above the and Loukas 2011), let alone how both components might
two components of empathy, school connectedness also interact to impact aggression (see Caravita et al. 2010).
contributed to a decline in boys’ subsequent overt aggres- Moreover, although belonging with peers and non-parent
sion. Recommendations are made to foster family and adults is believed to become increasingly more important
than the family during this developmental period (Eccles
and Roeser 2011; Roeser et al. 1998, 2000), no studies to our
M. Batanova  A. Loukas (&) knowledge have examined students’ connections to their
Department of Kinesiology and Health Education, The
family and school as they might differentially moderate the
University of Texas at Austin, 1 University Station D3700,
Austin, TX 78712, USA associations between both components of empathy and
e-mail: alexandra.loukas@austin.utexas.edu aggression. Thus, the current study examined the unique and
M. Batanova interactive effects of empathic concern, perspective taking,
e-mail: mbatanova@austin.utexas.edu positive family relations, and school connectedness on

123
J Youth Adolescence (2014) 43:1890–1902 1891

subsequent overt and relational aggression 1 year later in a While the inhibitory role of affective empathy in ado-
sample of early adolescent girls and boys. lescent aggression is generally well supported by research
In contrast to overt aggression, which includes direct (e.g., (see Lovett and Sheffield 2007 for a review), research on
hitting, yelling) acts meant to harm another person, relational the cognitive component of empathy has produced mixed
aggression involves more indirect, oftentimes covert strate- findings. For instance, Jolliffe and Farrington (2006, 2011)
gies (e.g., gossiping, rumor spreading) for harming another found no evidence for cognitive empathy being related to
individual by damaging or manipulating his or her friend- either form of aggression in girls or boys, while Espelage
ships and/or feelings (Bjorkqvist et al. 1992; Crick and et al. (2004) found perspective taking to be associated
Grotpeter 1995; Underwood et al. 2001). While overt negatively with bullying in middle school students. In
aggression is generally more common in boys than girls at contrast, the study by Batanova and Loukas (2011) found
any age (see Archer 2004 for a review), the evidence is mixed that elevated levels of perspective taking were associated
as to whether girls engage in more relational aggression than with increases in relational aggression across gender, even
boys during early adolescence (Batanova and Loukas 2011; after accounting for empathic concern and social anxiety.
Galen and Underwood 1997; Peets and Kikas 2006; Salm- Besides the possibility of a suppressor effect, the authors
ivalli and Kaukiainen 2004). What is known, however, is argued that early adolescents who are capable of taking
that, unlike overt aggression, the use of relational aggression others’ points of view might be socially intelligent (see
requires one to build and maintain social connections so as to Sutton et al. 1999a) and, therefore, can use that capability
manipulate others (Puckett et al. 2008; Salmivalli et al. 2000; for adopting covert aggressive strategies that entail more
Xie et al. 2002). Relational aggression also necessitates the thought and lower likelihood of getting caught than phys-
ability to read and decode social situations in order to execute ical aggression (Bjorkqvist et al. 2000).
successful manipulation (see Archer and Coyne 2005; Sutton Indeed, some researchers have found that adolescents
et al. 1999a). Given the gender differences in overt aggres- with competencies similar to perspective taking (e.g., moral
sion and the conflicting findings for relational aggression, it competence; attention to social cues) can still exhibit
is important to consider that the factors contributing to boys’ aggressive behaviors (Gini et al. 2011; Hawley 2003). In the
and girls’ use of overt and relational aggression may differ. literature on bullying, a few studies indicate that while
adolescent bullies lack moral compassion, they may never-
theless demonstrate moral competence, or a good under-
Empathy and Aggression standing of others’ perspectives (Gini 2006; Gini et al. 2011;
Sutton et al. 1999b). Moreover, Caravita et al. (2010) argued
To date, evidence indicates that the emotional or affective that emotional factors may be particularly relevant for
component of empathy plays an important role in the informing the relationships between socio-cognitive skills
inhibition of both overt and relational aggression in early like perspective taking and subsequent behavior like
adolescence. For instance, several studies found that ado- aggression or bullying (see Arsenio and Lemerise 2001).
lescent reports of affective empathy inhibit indirect Thus, the authors hypothesized that cognitive empathy
aggression as well as frequent and violent aggression in would be associated with bullying in instances of low emo-
female and male adolescents (Endresen and Olweus 2001; tional empathy. In other words, low levels of emotional
Jolliffe and Farrington 2006, 2011; Kaukiainen et al. 1999). empathy were expected to exacerbate the effects of per-
Similarly, De Kemp et al. (2007) reported that higher levels spective taking on bullying, measured as a composite vari-
of affective empathy were associated with less delinquent able encompassing general acts of aggression (e.g., ‘‘starts
and aggressive behavior 6 months later in both female and bullying’’; ‘‘convinces other children to bully’’). Although
male early adolescents. In a longitudinal sample of middle the hypothesis was not supported, the cross-sectional nature
school youth, Batanova and Loukas (2011) corroborated of the study, as well as the limited measure of bullying,
and extended these previous findings by showing that ele- warrant further investigation into the interaction between
vated levels of empathic concern were predictive of empathic concern and perspective taking as it might impact
decreases in both girls’ and boys’ self-reported overt as well subsequent forms of aggression. As Caravita et al. hypoth-
as relational aggression, over and above perspective taking esized, low levels of empathic concern may in fact exacer-
and social anxiety. Although one study found that affective bate the effects of perspective taking on overt and relational
empathy was associated negatively with bullying in ado- aggression 1 year later.
lescent boys only, the authors acknowledged that bullying
was measured with peer nominations of more overt rather The Moderating Roles of Family and School Factors
than relational forms of aggression and, therefore, affective
empathy could have been associated with other types of According to the social development model (Catalano and
bullying in females (Caravita et al. 2009). Hawkins 1996; Catalano et al. 2004), the development of

123
1892 J Youth Adolescence (2014) 43:1890–1902

aggression can be explained by students’ involvement examine the ways in which school connectedness informs
within a socializing unit, such as the family or school. the relationships between empathy and aggression among
Specifically, the theory posits that involvement is integral middle school students. Yet, just as the family context
to the socialization process, whereby commitment or shapes early adolescents’ sense of morality, youth who feel
bonding to a particular social unit carries power to affect a sense of connection to the school would most likely
behavior. Thus, adolescents who report positive family internalize and model the positive values and behaviors
interactions and connectedness to their school are likely to promoted by that school (Barr and Higgins-D’Alessandro
be less aggressive than those who do not because they 2007, 2009; Batanova and Loukas 2012).
conform to the positive norms and values associated with A recent study by Batanova and Loukas (2012) found
their family and school contexts. Indeed, evidence indi- that school connectedness contributed to higher subsequent
cates that students who report positive connections to their levels of empathic concern and perspective taking in
school also report fewer behavioral problems, including middle school boys. Also, in the study by Caravita et al.
aggression and bullying (Brookmeyer et al. 2006; Loukas (2010), theory of mind skills, or the ability to attribute
and Murphy 2007; Ma 2002). Moreover, warm family mental states to others, was especially related to boys’
interactions and positive parent–child relationships play a defending behaviors during bullying episodes when they
crucial role in the inhibition of aggressive and antisocial were well liked by other peers. Thus, boys who demon-
behaviors (Arim et al. 2011; De Kemp et al. 2007; Shetgiri strate positive connections with peers might be better sui-
et al. 2012). ted to use their cognitive skills in positive ways, thereby
Unlike the large body of literature linking familial fac- refraining from exhibiting subsequent aggression. The
tors to adolescent aggression, no known research has current study explored whether this hypothesis would be
examined the ways in which positive family relations can true for girls as well, whereby girls reporting positive
modify or moderate the empathy-aggression relationships connections to their school might be better equipped to use
during early adolescence. Rather, research shows that their perspective taking skills to refrain from aggression.
positive parenting might be more influential for some Given limited evidence that positive relationships at school
children over others due to children’s individual differ- are linked to adolescents’ emotional empathy (Barr and
ences in responding to socialization practices (Edens et al. Higgins-D’Alessandro 2007), the current study also
2008; Oxford et al. 2003). For instance, Van der Graaff examined whether high levels of school connectedness
et al. (2012) found that higher perceived parental support at would protect early adolescents low in empathic concern
age 14 predicted lower levels of aggression 1 year later, but from exhibiting both overt and relational forms of
only in instances of high affective empathy among the aggression 1 year later.
adolescent participants. Nonetheless, research shows that
warm and supportive parenting can facilitate the develop-
ment of cognitive empathy (Soenens et al. 2007) and The Current Study
emotional empathy in adolescents (Carlo et al. 2007;
Eisenberg et al. 2006). Thus, a positive family environment This study examined the unique and interactive effects of
likely sets a moral standard for early adolescents, whereby affective and cognitive empathy on subsequent overt and
positive interactions serve as cues for how to act and feel relational aggression 1 year later in a sample of middle
towards others (see Carlo et al. 2010; Hoffman 2000). school students, and whether positive family and school
Extending existing research, the current study examined factors would moderate the empathy-aggression associa-
whether a positive family environment would protect youth tions. We expected that empathic concern would be pre-
low in empathic concern and potentially high in perspec- dictive of a residualized decrease in both forms of
tive taking from subsequent aggression. aggression. Conversely, because perspective taking skills
In addition to the family context, it is reasonable to could serve as a unique tool for social manipulation, we
expect that students’ perceived school connectedness hypothesized that perspective taking would contribute to a
would also moderate their empathy levels as they con- residualized increase in subsequent relational aggression.
tribute to subsequent aggression. Used interchangeably Also because research indicates that cognitive empathy is
with constructs such as school bonding, school engage- associated with overt forms of bullying, we hypothesized
ment, and school climate (see Anderman and Freeman that perspective taking would be positively predictive of
2004, for a review), school connectedness captures mostly overt aggression. Based on the social development model
affective aspects of students’ school experiences and (Catalano and Hawkins 1996; Catalano et al. 2004), we
includes students’ sense of safety, belonging to the school, further hypothesized that positive family relations and
and perceptions of teachers’ fairness and support (Resnick students’ school connectedness would buffer the negative
et al. 1997). To our knowledge, there are no studies that impact of low affective empathy and potentially high

123
J Youth Adolescence (2014) 43:1890–1902 1893

perspective taking on subsequent overt and relational consistency reliabilities for Wave 1 empathic concern were
aggression. .82 and .85 for girls and boys, respectively.

Perspective Taking
Method
The Wave 1 7-item perspective taking subscale was also
Participants drawn from the Interpersonal Reactivity Index (Davis
1980). The perspective taking items were also slightly
Participants were 260 girls (74 % White; 18 % Hispanic; modified for middle school students (e.g., ‘‘You sometimes
3 % Black) and 221 boys (78 % White; 14 % Hispanic; try to understand your friends better by imagining how they
4 % Black) between 10 and 14 years of age involved in think about things’’). All items were scored on a scale
two study waves. At Wave 1 (fall of 2001), students were ranging from 0 (‘‘Not at all like you’’) to 4 (‘‘Very much
in the 6th and 7th grades (M age = 11.68; SD = .75), and like you’’) and were averaged so that higher scores reflect
Wave 2 occurred 1 year later when students were in the 7th higher perspective taking. Davis (1980) also showed that
and 8th grades (M age = 12.75; SD = .72). Data from an the test–retest and internal consistency reliabilities for this
additional 19 students were eliminated from the final subscale are substantial. For the current study, the internal
sample due to substantial amounts of missing information consistency reliability for Wave 1 perspective taking was
relevant to the present study. .81 for both girls and boys.

Procedure Positive Family Relations

At Wave 1, active parental consent was obtained from Quality of family relations was measured at Wave 1 with 6
76 % (N = 884) of all 6th and 7th grade students attending items adapted by Metzler et al. (1998) from the Family
three schools. Of the students participating at Wave 1 and Environment Scale (Moos and Moos 1986). Items assessed
who were eligible to participate at Wave 2, 71 % received the degree of warmth (‘I really enjoyed being with my
parental permission to participate at Wave 2. However, parents’), trust (‘My parents trusted my judgment’),
eight students refused participation and 30 students were togetherness (‘There was a feeling of togetherness in our
absent on both the day of the survey and the one make-up family’) and fun (‘The things we did together were fun and
day; consequently, 66 % of the eligible students (n = 500) interesting’) between the adolescent and his/her parents.
participated at Wave 2. A questionnaire consisting of 161 Responses ranged on a 5-point scale from 1 (‘Never True’)
items at Wave 1 and 160 items at Wave 2 was group- to 5 (‘Always True’). Higher scores reflect more positive
administered to participating students in one 40-min family relations. Internal consistency reliabilities for the
homeroom class. A member of the research team read each 6-item scale were .86 for girls and .84 for boys.
question aloud to students to maintain compliance and to
control for varying levels of reading comprehension. School Connectedness

Measures Five items from the National Longitudinal Study of Ado-


lescent Health (see Anderman 2002; McNeely et al. 2002)
Empathic Concern were used to assess school connectedness at Wave 1.
Students responded to items such as ‘‘I feel close to people
The 7-item empathic concern subscale from Davis’s (1980) at this school,’’ and ‘‘I feel like I am part of this school.’’
self-report Interpersonal Reactivity Index (IRI) was asses- The five items were scored on a 5-point scale ranging from
sed at Wave 1 in the current study. Since the IRI was 1 (‘‘Strongly Agree’’) to 5 (‘‘Strongly Disagree’’). Items
initially tested on college students, all of the empathic were reverse coded and averaged so that higher scores
concern items were slightly modified so as to be relevant reflect higher levels of school connectedness. The internal
for middle school students (e.g., ‘‘You often feel sorry for consistency reliabilities for the five items were .73 for girls
people who don’t have the things you have’’). All items and .71 for boys.
were scored on a 5-point scale ranging from 0 (‘‘Not at all
like you’’) to 4 (‘‘Very much like you’’) and were averaged Overt Aggression
so that higher scores reflect higher levels of empathic
concern. Davis (1980) has shown that test–retest and Overt aggression was assessed at Waves 1 and 2 using an
internal consistency reliabilities for the empathic concern adapted version of Crick’s (1996) 3-item peer nomination
subscale are substantial. For this study, the internal measure of overt aggression. The three items assessing

123
1894 J Youth Adolescence (2014) 43:1890–1902

physical and direct, verbal aggression (hitting, calling names, Multivariate Analyses
and starting fights) were adapted for self-report by asking
adolescents to identify how often they engaged in each of the Because research indicates that girls and boys typically
behaviors. Items were scored on a scale ranging from 1 (‘‘Not experience their family and school contexts in different
at all’’) to 5 (‘‘All the time’’) and were averaged so that higher ways (Grusec and Hastings 2007; McHale et al. 2003), and
scores reflect more overt aggression. Internal consistency of that girls report higher empathic concern and perspective
the 3-item measure was .78 for girls’ and .69 for boys’ reports taking than boys (Batanova and Loukas 2011; Espelage
of overt aggression at Wave 1. At Wave 2, the reliabilities et al. 2004; Gini et al. 2007; Mayberry and Espelage 2007),
were .71 and .82 for girls and boys, respectively. the current study examined girls’ and boys’ data separately.
To verify that separate models should be conducted, Box’s
Relational Aggression M test was used to determine the homogeneity of variance–
covariance matrices of all study variables across gender.
Relational aggression was assessed at Waves 1 and 2 using Results indicated that the variance–covariance matrices
a 6-item scale, with four adapted items from Crick and significantly varied across gender (Box’s M = 110.42,
Grotpeter’s (1995) peer-nominated measure of relational F(36) = 3.01, p \ .001), confirming that all subsequent
aggression. The scale was adapted for the present study by models should be run separately for female and male
modifying the wording of the four items so that middle students.
school students could self-report how likely they were to
engage in each of the behaviors (e.g., ‘‘When you’re mad at Zero-Order Correlations
someone, how often do you ignore them or stop talking to
them’’). Two additional items assessing the frequency of Prior to examining the hypotheses, zero-order correlations
negative facial expressions (making mean faces and rolling were examined between all study variables (see Table 1 for
eyes at peers) were added, given existing evidence that girls’ and boys’ data, separately). Examination of Table 1
such expressions are important features of girls’ aggressive showed that both components of empathy were correlated
behaviors (Galen and Underwood 1997). In fact, Paquette positively with the Wave 1 family and school factors for
and Underwood (1999) found that early adolescents girls and boys. Moreover, both components of empathy and
reported rolling eyes as the most frequently experienced both forms of aggression at each wave were positively
non-physically aggressive behavior. The six self-reported correlated at each wave. Empathic concern was associated
items were scored on a scale ranging from 1 (‘‘Not at all’’) negatively with both overt and relational aggression at both
to 5 (‘‘All the time’’) and were averaged so that higher waves, but only for girls. Girls’ perspective taking was
scores reflect more relational aggression. Internal consis- associated negatively with both forms of aggression at both
tency reliabilities for the six items at Wave 1 were .71 and waves, whereas boys’ perspective taking was inversely
.75 for girls and boys, respectively. Reports of relational related to both forms of aggression at Wave 1, only. For both
aggression at Wave 2 had internal consistency reliabilities girls and boys, positive family relations were correlated
of .80 and .67 for girls and boys, respectively. negatively with Wave 1 overt and relational aggression.
Only girls’ reports of positive family relations were corre-
lated negatively with Wave 2 overt and relational aggres-
Results sion. Conversely, girls’ school connectedness was associated
negatively with both forms of aggression at Wave 1 only,
Attrition Analyses whereas only boys’ school connectedness was associated
negatively with overt and relational aggression at both
Attrition analyses were conducted to determine if students waves.
who participated at both waves of the study differed from
their peers who participated only at Wave 1. Results Regression Analyses
indicated that in comparison to students who participated at
both waves of the study, students who did not participate at Study hypotheses were tested using 4-step hierarchical
Wave 2 reported lower levels of Wave 1 empathic concern regression analyses, with separate models conducted for
[t(875) = -2.39, p \ .05], perspective taking [t(879) = girls’ and boys’ Wave 2 overt and relational aggression
-2.63, p \ .01], school connectedness [t(880) = -3.88, outcome variables. Step 1 included early adolescent age and
p \ .001], and positive family relations [t(878) = -2.29, baseline levels of both forms of aggression. Age was
p \ .05], as well as higher levels of both Wave 1 overt included to control for the possibility that students of
[t(880) = 2.16, p \ .05] and relational [t(879) = 2.96, varying age levels had different cognitive processing or
p \ .01] aggression. understanding of the study items, and also because physical

123
J Youth Adolescence (2014) 43:1890–1902 1895

Table 1 Zero-order correlations and descriptive statistics for all study variables among female (n = 260) and male (n = 221) middle school
students
Study variables 1 2 3 4 5 6 7 8

1. Wave 1 empathic concern – .67** .29** .23** -.26** -.31** -.30** -.24**
2. Wave 1 perspective taking .63** – .30** .28** -.14* -.28** -.13* -.14*
3. Wave 1 positive family relations .18** .27** – .36** -.26** -.31** -.24** -.25**
4. Wave 1 school connectedness .14* .20** .39** – -.24** -.21** -.11 -.12
5. Wave 1 overt aggression -.13 -.26** -.21** -.19** – .52** .49** .37**
6. Wave 1 relational aggression -.05 -.21** -.39** -.38** .50** – .38** .53**
7. Wave 2 overt aggression -.13 -.12 -.08 -.22** .36** .27** – .57*
8. Wave 2 relational aggression -.11 -.10 -.13 -.22** .26** .37** .52** –
Girls’ M (SD) 3.19 (.64) 2.69 (.71) 3.89 (.84) 3.96 (.62) 1.66 (.66) 1.83 (.53) 1.60 (.62) 1.83 (.63)
Boys’ M (SD) 2.80 (.73) 2.29 (.80) 3.99 (.78) 3.77 (.66) 2.00 (.64) 1.82 (.62) 2.02 (.80) 1.68 (.55)
Values above the diagonal are for females and values below the diagonal are for males. Both empathy scales were measured on a 0–4 scale
(0 = not at all like you, 4 = very much like you). Positive family relations were measured on a 1–5 scale (1 = never true, 5 = always true).
School connectedness and both waves of overt aggression and relational aggression were measured on a 1–5 scale (1 = not at all, 5 = all the
time)
* p \ .05; ** p \ 0.01

maturation has been found to affect aggression (Celio et al. interactions were assessed separately and in the presence of
2006). Baseline levels of both forms of aggression, which all main effects.
are highly correlated and found to overlap (Rose et al. 2004;
Smith et al. 2009), were also included in step 1 so as to Girls’ Results
partial out the impact of one on change in the other. Because
the variance associated with baseline levels of aggression As shown in Table 2, girls’ Wave 2 overt aggression was
was partialled out of the Wave 2 outcomes, models were predicted by baseline levels of overt aggression in all of the
interpreted as predicting the residualized difference in the main effects models (see steps 1 through 3), as well as by
aggression outcomes over a 1-year period of time. Empathic Wave 1 relational aggression in the main effects models
concern and perspective taking were entered simultaneously including age and empathy (see steps 1 and 2). Similarly,
in step 2. Thus, any impact of either component of empathy Table 3 indicates that girls’ Wave 2 relational aggression
on subsequent aggression could be interpreted as over and was predicted by baseline levels of relational aggression in
above the other empathy subscale. In step 3, the moderator all of the main effects models (see steps 1 through 3), and
variables of positive family relations and school connect- by baseline overt aggression in step 1, only. Age, however,
edness were assessed simultaneously. Then, five two-way was not significantly associated with change in either form
interactions, one between both components of empathy of aggression. Moreover, even after baseline levels of
(empathic concern x perspective taking) and four between aggression and age were partialled out, Wave 1 empathic
each of the empathy components and moderator variables, concern was negatively predictive of residualized change
were entered separately in step 4 (empathic concern x family in overt aggression, only (see Table 2). Neither positive
relations; perspective taking x family relations, empathic family relations nor school connectedness contributed
concern x school connectedness; perspective taking x school uniquely to either form of aggression. Overall, the main
connectedness). effects models accounted for 31 % of the variance in Wave
The two-way interactions assessed if (1) empathic con- 2 overt aggression [F(7, 259) = 15.79, p \ .001] and 31 %
cern moderated the associations between perspective tak- of the variance in Wave 2 relational aggression [F(7,
ing and subsequent aggression, and (2) student reports of 259) = 16.08, p \ .001]. None of the five two-way inter-
their family environment or level of school connectedness actions predicted the Wave 2 overt and relational aggres-
moderated the relations between either component of sion outcomes for girls.
empathy and either form of aggression 1 year later. Inter-
action terms were comprised of the product term between Boys’ Results
empathic concern and perspective taking or the empathy
and moderator variables that were each grand-mean cen- Examination of the main effects models (see Tables 2 and
tered to avoid problems with multicollinearity. Following 3, steps 1–3) showed that boys’ baseline levels of overt
the guidelines of Aiken and West (1991), two-way aggression were significantly predictive of Wave 2 overt

123
1896 J Youth Adolescence (2014) 43:1890–1902

Table 2 Predicting Wave 2 overt aggression from Wave 1 empathic concern and perspective taking and Wave 1 positive family relations and
school connectedness in female (n = 260) and male (n = 221) middle school students
Girls’ results Boys’ results
Step 1 Step 2 Step 3 Step 4 Step 1 Step 2 Step 3 Step 4

Predicting Wave 2 overt aggression from Wave 1 study variables


Covariates
Overt aggression .41*** .38*** .38*** .30*** .29*** .30***
Relational aggression .16* .14* .13 .12 .13 .10
Age (in years) .06 .06 .05 .06 .02 .01
Main effects
Empathic concern -.24*** -.24*** -.13 -.12
Perspective taking .13 .13 .06 .06
Positive family relations -.10 .09
School connectedness .06 -.15*
Two-way interactions
EC 9 PT -.08 .11
EC 9 familya -.09 .15*
PT 9 familya -.09 .11
EC 9 schoola -.09 .07
PT 9 schoola -.06 -.02
R2 (change) .27*** .30 (.03)** .31 (.01) .31 (.01) .14* .15 (.01) .17 (.02)? .20 (.04)*
?
p \ .10; * p \ .05; ** p \ .01; *** p \ .001
a
Two-way interactions examined individually in the presence of all main effects

Table 3 Predicting Wave 2 relational aggression from Wave 1 empathic concern and perspective taking and Wave 1 positive family relations
and school connectedness in female (n = 260) and male (n = 221) middle school students
Girls’ results Boys’ results
Step 1 Step 2 Step 3 Step 4 Step 1 Step 2 Step 3 Step 4

Predicting Wave 2 relational aggression from Wave 1 study variables


Covariates
Relational aggression .46*** .45*** .44*** .32*** .32*** .30***
Overt aggression .13* .11 .11 .11 .11 .11
Age (in years) .04 .05 .04 -.06 -.05 -.06
Main effects
Empathic concern -.13 -.12 -.12 -.12
Perspective taking .11 .09 .07 .07
Positive family relations -.09 .06
School connectedness .04 -.11
Two-way interactions
EC 9 PT -.03 -.01
EC 9 familya -.02 .20**
PT 9 familya -.01 .03
a
EC 9 school .06 -.02
PT 9 schoola .07 -.10
R2 (change) .29*** .30 (.01) .31 (.01) .31 (.00) .15*** .15 (.01) .16 (.01) .20 (.04)**
* p \ .05; ** p \ .01; *** p \ .001
a
Two-way interactions examined individually in the presence of all main effects

123
J Youth Adolescence (2014) 43:1890–1902 1897

aggression (see steps 1 through 3), and that baseline levels high levels of positive family low levels of positive family
relations relations
of relational aggression were significantly predictive of 2.2
Wave 2 relational aggression (see steps 1 through 3). Boys’
age was not significantly associated with change in either

Wave 2 Overt Aggression


form of aggression, and neither component of empathy 2
contributed to change in the aggression outcome variables.
Similarly, positive family relations was not uniquely pre-
dictive of change in either form of aggression. Even after 1.8
partialling out baseline levels of aggression, age, and both
components of empathy, elevated levels of school con-
1.6
nectedness predicted a residualized decrease in boys’ overt
aggression (see Table 2). However, this result should be
interpreted cautiously, as the amount of variance explained
1.4
by school connectedness was only marginally significant 1 SD below mean 1 SD above mean
(p = .09). Overall, the main effects models accounted for Wave 1 Empathic Concern
17 % of the variance in Wave 2 overt aggression [F(7,
220) = 6.09, p \ .001] and 16 % of the variance in Wave 2 Fig. 1 Examining the empathic concern 9 positive family relations
interaction for boys’ Wave 2 overt aggression
relational aggression [F(7, 220) = 5.96, p \ .001].
Examination of the two-way interactions indicated that
high levels of positive family low levels of positive family
only the interaction between empathic concern and positive relations relations
family relations was significant for boys’ overt and relational 2

aggression outcomes (see Tables 2 and 3). Similar to other


Wave 2 Relational Aggression

non-experimental studies (see Chaplin 1991), small but


significant portions of the variance in boys’ Wave 2 overt 1.8
aggression [R2 change = .02, F(8, 220) = 6.15, p \ .001]
and Wave 2 relational aggression [R2 change = .04, F(8,
220) = 6.71, p \ .001] were explained by these two-way
interactions. Using the methods outlined by Aiken and West 1.6
(1991), the two-way interactions were probed by examining
the associations between empathic concern and boys’ Wave
2 aggression outcomes at high (1 standard deviation above
1.4
the mean) and low (1 standard deviation below the mean) 1 SD below mean 1 SD above mean
levels of positive family relations. Wave 1 Empathic Concern
As depicted in Fig. 1, probing the two-way interaction
for overt aggression showed that among boys reporting low Fig. 2 Examining the empathic concern 9 positive family relations
interaction for boys’ Wave 2 relational aggression
levels of positive family relations, empathic concern was
associated negatively with residualized change in overt
aggression (b = -.27, p \ .05). When boys reported high that the family context served as a protective factor,
levels of positive family relations, however, empathic whereby elevated levels of positive family relations buf-
concern was not associated with residualized change in fered or offset the negative influence of low empathic
overt aggression (b = .05, p [ .05). This finding indicates concern on boys’ subsequent relational aggression.
that elevated levels of positive family relations buffered or
offset the influence of low empathic concern on boys’
subsequent overt aggression. Discussion
Probing the two-way interaction for relational aggres-
sion in Fig. 2 showed the same pattern of findings as for Traditionally, empathy is known to inhibit aggressive
overt aggression. Among boys reporting low levels of behaviors in children and adolescents (Gini et al. 2007;
positive family relations, empathic concern was linked Robinson et al. 2007; Strayer and Roberts. 2004). Given
negatively to residualized change in relational aggression the multidimensional nature of empathy and aggression,
(b = -.28, p \ .01). Conversely, when boys reported high however, much less is known about the prospective asso-
levels of positive family relations, empathic concern no ciations between different components of empathy and
longer contributed to any residualized change in relational aggression, as well as the moderating functions of social
aggression (b = .11, p [ .05). This finding also indicates factors in the aforementioned associations. Thus, the

123
1898 J Youth Adolescence (2014) 43:1890–1902

current study examined the unique and interactive roles of impulsive and fail to consider the consequences of their
empathic concern, perspective taking, positive family actions. This hypothesis reiterates Arsenio and Lemerise’s
relations, and school connectedness in subsequent overt (2001) argument that emotion processes, including emotion
and relational aggression 1 year later in a sample of early intensity and emotion regulation, also may underlie the
adolescents. Study findings indicated that among girls only, empathy-aggression associations.
empathic concern was uniquely predictive of a residualized Interestingly, positive family relations protected boys
decrease in overt aggression. Among boys, positive family low in empathic concern from exhibiting overt and rela-
relations buffered the negative impact of low empathic tional aggression 1 year later. Supported by the social
concern on both forms of aggression 1 year later. Over and development model (Catalano and Hawkins 1996; Cata-
above both components of empathy, boys’ elevated levels lano et al. 2004), findings highlight the importance of a
of school connectedness contributed to a residualized positive socializing family environment for early adoles-
decrease in subsequent overt aggression. Unexpectedly, cent boys. That is, positive interactions and quality time
perspective taking was not associated with either form of among family members may serve as models and instruc-
aggression across gender, and there were no significant tions for appropriate behavior, even among boys with low
interactions between the two components of empathy or levels of empathic concern (see Eisenberg et al. 2006;
between either component of empathy and school con- Hastings et al. 2007). Although other studies have also
nectedness in predicting boys’ or girls’ aggression. Taken tested two-way interactions between family context vari-
together, the current study highlights the importance of ables and empathy as they relate to aggression, they con-
affective empathy in girls’ aggression and the roles of sidered empathy the moderator variable and findings from
family and school in boys’ aggression. these studies are conflicting. For instance, De Kemp et al.
In partial support of our first hypothesis and consistent (2007) found no evidence that parental support and
with existing research on affective empathy and aggression empathy made any interacting contributions to adolescents’
(De Kemp et al. 2007; Jolliffe and Farrington 2006, 2011), subsequent aggression. In a recent study by Van der Graaff
elevated levels of empathic concern contributed to lower et al. (2012), however, perceived parental support con-
levels of overt aggression in early adolescent girls, over and tributed to lower levels of aggression only for adolescents
above levels of perspective taking. Contrary to previous who reported high levels of affective empathy. Conversely,
research, empathic concern was not uniquely predictive of when low empathic adolescents reported higher levels of
boys’ overt aggression (see Caravita et al. 2009; Espelage parental support, they showed even higher levels of
et al. 2004). Although the lack of findings for boys was aggression and delinquency 1 year later. While the authors
unexpected, they are consistent with research by Jolliffe and suggested that adolescents with low levels of emotional
Farrington (2006, 2011) showing that boys’ affective empathy might wrongfully interpret their parents’ support
empathy was not significantly related to their bullying. as a permissive parenting style, our findings indicate that
Rather, these researchers found that the frequency of bully- adolescent boys with low empathic concern may in fact
ing mattered, such that males who bullied frequently showed benefit from positive relationships with family members,
significantly low levels of affective empathy. Thus, it is including parents. However, more research is needed to
possible that for boys, affective empathy plays a role only in examine specific family system dynamics (e.g., inclusion
cases of persistent or chronic forms of overt aggression. of siblings) and contextual factors (e.g., parental supervi-
Subsequent research is needed to consider the interaction sion) as they contribute to gender socialization relevant to
between affective empathy and frequency of aggression. adolescent empathy and aggression (see Jolliffe and Far-
In contrast to results for overt aggression, empathic rington 2011; McHale et al. 2003). For instance, girls low
concern was not associated with relational aggression for in empathic concern may also benefit from positive family
either boys or girls. This finding came as a surprise, given relations, but more specific measures might be needed to
that a previous study found empathic concern to reduce identify factors that are salient to early adolescent girls’
both forms of aggression in middle school students (see outcomes, such as parent-daughter communication about
Batanova and Loukas 2011). However, the study also their sense of self-worth (see Bosacki 2003).
found that the main effect of empathic concern was qual- It is notable that boys’ perceived levels of school con-
ified by a two-way interaction with fear of negative eval- nectedness uniquely contributed to a residualized decrease
uation, a form of social anxiety where one is prone to in their overt aggression, over and above their empathy
misinterpreting others’ intentions as hostile. Perhaps, then, levels and positive family relations. Consistent with limited
empathy’s effect on relational aggression is only relevant previous research indicating that boys’ school connected-
under certain conditions, such as when students experience ness leads to fewer emotional symptoms and peer problems
elevated levels of social anxiety, or, as Jolliffe and Far- (Shochet et al. 2006), the current study further highlights
rington (2011) have also pointed out, when they are highly that feelings of school belonging and closeness with teachers

123
J Youth Adolescence (2014) 43:1890–1902 1899

and peers might be especially important for reducing boys’ levels of empathic concern to exacerbate the effects of
overt aggression. Specifically, boys who feel accepted and perspective taking on aggression (see Caravita et al. 2010),
supported by people at their school might refrain from dis- it is likely that other factors related to adolescents’ sense of
playing overtly harmful behaviors (Crosnoe et al. 2002). It moral values and emotions need to be investigated as well
must also be noted, however, that although the main effect of (see Arsenio and Lemerise 2001; Hart and Carlo 2005). For
school connectedness was significant, the amount of vari- instance, Laible et al. (2008) demonstrated that both moral
ance it accounted for in the aggression outcome was only affect (comprised of guilt, shame, sympathy, and empathic
marginally significant. Thus, results should be replicated anger) and moral cognition (comprised of prosocial rea-
prior to drawing any firm conclusions regarding the role of soning and internalization of values) were linked nega-
school connectedness as a unique predictor of change in tively to adolescent bullying. Thus, empathic concern is
boys’ overt aggression. Moreover, because girls’ school likely one piece of the puzzle in which a variety of moral
connectedness did not uniquely predict declines in overt factors impact the relationship between perspective taking
aggression, it is likely that the school context is not as rel- and subsequent aggression.
evant to girls’ overt aggression once their levels of empathy There are also some limitations that warrant discussion.
are taken into consideration. That is, given that girls are First, although there was a 1-year lag between the predictor
generally socialized to be prosocial, caring, and ‘‘well- and outcome variables, all measures were self-reported by
behaving’’ (Hastings et al. 2007), factors that contribute to the adolescents, increasing the possibility that the obtained
these individual difference variables may be more important associations are inflated due to shared method variance.
and warrant further investigation in explaining their levels of Nevertheless, given the highly personal and internal nature
overt aggression than is school connectedness. of empathy, adolescents are likely the best informants of
Interestingly, school connectedness did not contribute to their own empathic concern and perspective taking. Simi-
either girls’ or boys’ reports of relational aggression. larly, because relational aggression includes acts that are
According to the social development model (Catalano and often covert and hidden (see Kaukiainen et al. 1999) and
Hawkins 1996; Catalano et al. 2004), adolescents who feel overt aggression may vary across contexts (see Achenbach
connected to their school should be less likely than their et al. 1987), self-reports likely reflect the full repertoire of
counterparts to deviate from the school’s norms and individual behaviors, whereas reports by others may be
expectations, and for this reason have a lower likelihood of limited to behaviors specific to particular contexts. Second,
acting out or showing aggression. Because relational the study is limited by two waves of data, which does not
aggression entails covert behaviors, like excluding or allow for the examination of trajectories of relational and
ignoring a peer, it might be difficult for a school to con- overt aggression across time. A longitudinal study could
sistently identify and assist with such problems (Holt and address how changes in empathy, as well as family and
Keyes 2004; Young et al. 2006). Schools might inadver- school relationships, contribute to changes in aggression
tently allow or ignore acts of relational aggression and across time. Third, the family and school moderator vari-
some might even tolerate these tactics given limited evi- ables were relatively broad, and therefore, future studies
dence that up to 30 % of students get ignored and verbally should replicate the current study with specific measures,
harassed by teachers (James et al. 2008; Twemlow et al. such as parental acceptance (e.g., Hare et al. 2011) and sense
2006). If school norms and expectations neglect to ade- of community connectedness in the school (e.g., Flanagan
quately address relational aggression, it is unlikely that et al. 2007). Finally, attrition analyses showed that students
school connectedness will contribute to decreases in either who did not participate at Wave 2 were more likely than
boys’ or girls’ relational aggression. students who did participate to report heightened levels of
While the current study adds some novel findings to the both overt and relational aggression, as well as lower levels
literature, there are two additional non-significant associ- of empathic concern, perspective taking, positive family
ations that warrant some discussion. First, in contrast to relations, and school connectedness. Retention of these
studies that found unique associations between the cogni- students may have produced even stronger associations
tive component of empathy and adolescent aggressive between variables, thereby strengthening the final models.
behaviors (Batanova and Loukas 2011; Caravita et al.
2009, 2010), the current study found no such evidence. The
lack of significant findings might be attributed to other Conclusions
related factors that either confound the effects of perspec-
tive taking or better explain adolescent aggression, such as This study extends the literature by examining the pro-
impulsivity, perceptions of the self, and social goals (see spective associations between different components of
Bosacki 2003; Ojanen et al. 2012; Jolliffe and Farrington empathy and aggression, as well as the moderating roles of
2011). Second, because the current study did not find low positive family and school factors in the aforementioned

123
1900 J Youth Adolescence (2014) 43:1890–1902

associations. While empathic concern contributed to girls’ Barr, J., & D’Alessandro, A. (2009). How adolescent empathy and
subsequent overt aggression, boys’ positive family relations prosocial behavior change in the context of school culture: A
two-year longitudinal study. Adolescence, 44(176), 751–752.
buffered the negative effects of low empathic concern on Barr, J. J., & Higgins-D’Alessandro, A. (2007). Adolescent empathy
both forms of aggression. Interestingly, school connected- and prosocial behavior in the multidimensional context of school
ness played an inhibitory role in boys’ overt aggression as culture. The Journal of Genetic Psychology, 168, 231–250.
well. Thus, despite evidence indicating that adolescent girls Batanova, M., & Loukas, A. (2011). Social anxiety and aggression in
early adolescents. Journal of Youth and Adolescence, 40(11),
are more relationally oriented than boys (Rose and Rudolph 1534–1543.
2006), current study findings underscore the need for Batanova, M., & Loukas, A. (2012). What are the unique and
intervention and prevention programs to promote school interacting contributions of school and family factors to early
connectedness and positive family relationships for boys, adolescents’ empathic concern and perspective taking? Journal
of Youth and Adolescence, 41, 1382–1391.
particularly those who demonstrate little concern or care for Bjorkqvist, K., Lagerspetz, K., & Kaukiainen, A. (1992). Do girls
others. Given that socio-cognitive processing, such as per- manipulate and boys fight? Developmental trends regarding
spective taking, has long been emphasized in children’s direct and indirect aggression. Aggressive Behavior, 18,
empathy development and social functioning (see Malti and 117–127.
Bjorkqvist, K., Osterman, K., & Kaukiainen, A. (2000). Social
Latzko 2010), it is noteworthy that neither perspective intelligence – empathy = aggression? Aggression and Violent
taking nor its interaction with empathic concern and the Behavior, 5(2), 191–200.
other social factors produced any significant changes in Bosacki, S. L. (2003). Psychological pragmatics in preadolescents:
early adolescents’ aggression. The current study highlights Sociomoral understanding, self-worth, and school behavior.
Journal of Youth and Adolescence, 32(2), 141–155.
the role of empathic concern as it varies for girls’ and boys’ Brookmeyer, K. A., Fanti, K. A., & Henrich, C. (2006). Schools,
aggression, but calls for future research to disentangle the parents, and youth violence: A multilevel, ecological analysis.
many factors related to moral affect and cognition as they Journal of Clinical Child and Adolescent Psychology, 35,
contribute to adolescents’ aggression. 504–514.
Caravita, S., Blasio, P. D., & Salmivalli, C. (2009). Unique and
interactive effects of empathy and social status on involvement
Author contributions Milena Batanova conceived of the study, in bullying. Social Development, 18(1), 141–163.
performed the statistical analyses, and coordinated the writing of the Caravita, S., Blasio, P. D., & Salmivalli, C. (2010). Early adolescents’
manuscript; A.L. provided the data necessary for the study, partici- participation in bullying: Is ToM involved? The Journal of Early
pated in the interpretation of the data, and edited all drafts of the Adolescence, 30, 138–170.
manuscript. All authors read and approved the final manuscript. Carlo, G., McGinley, M., Hayes, R., Batenhorst, C., & Wilkinson, J.
(2007). Parenting styles or practices? Parenting, sympathy, and
prosocial behaviors among adolescents. The Journal of Genetic
Psychology, 168(2), 147–176.
Carlo, G., Mestre, M. V., Samper, P., Tur, A., & Armenta, B. E.
References (2010). The longitudinal relations among parenting styles,
sympathy, prosocial moral reasoning, and prosocial behaviors.
Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). International Journal of Behavioral Development, 35(2),
Child/adolescent behavioral and emotional problems: Implica- 116–124.
tions of cross-informant correlations for situational specificity. Catalano, R. F., Haggerty, K. P., Oesterle, S., Fleming, C. B., &
Psychological Bulletin, 101, 213–232. Hawkins, J. D. (2004). The importance of bonding to school for
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and healthy development: Findings from the social development
interpreting interactions. Newbury Park, PA: Sage. research group. Journal of School Health, 74(7), 252–261.
Anderman, E. M. (2002). School effects on psychological outcomes Catalano, R. F., & Hawkins, J. D. (1996). The social development
during adolescence. Journal of Educational Psychology, 94, model: A theory of antisocial behavior. In J. D. Hawkins (Ed.),
795–809. Delinquency and crime: Current theories (pp. 149–196). New
Anderman, L. H., & Freeman, T. M. (2004). Students’ sense of York, NY: Cambridge University Press.
belonging in school. In P. R. Pintrich & M. L. Maehr (Eds.), Celio, M., Karnik, N. S., & Steiner, H. (2006). Early maturation as a
Advances in motivation and achievement (Vol. 13, pp. 27–63). risk factor for aggression and delinquency in adolescent girls: A
Oxford: Elsevier. review. International Journal of Clinical Practice, 60(10),
Archer, J. (2004). Sex differences in aggression in real-world settings: 1254–1262.
A meta-analytic review. Review of General Psychology, 8, Chaplin, W. F. (1991). The next generation of moderator research in
291–322. personality psychology. Journal of Personality, 59, 143–178.
Archer, J., & Coyne, S. M. (2005). An integrated review of indirect, Crick, N. R. (1996). The role of overt aggression, relational
relational, and social aggression. Personality and Social Psy- aggression, and prosocial behavior in the prediction of children’s
chology Review, 9(3), 212–230. future social adjustment. Child Development, 67, 2317–2327.
Arim, R. G., Dahinten, V. S., Marshall, S. K., & Shapka, J. D. (2011). Crick, N. R., & Grotpeter, J. (1995). Relational aggression, gender,
An examination of the reciprocal relationships between adoles- and social-psychological adjustment. Child Development, 67,
cents’ aggressive behaviors and their perceptions of parental 2317–2327.
nurturance. Journal of Youth and Adolescence, 40, 207–220. Crosnoe, R., Erickson, K. G., & Dornbusch, S. M. (2002). Protective
Arsenio, W. F., & Lemerise, E. A. (2001). Varieties of childhood functions of family relationships and school factors on the
bullying: Values, emotion processes, and social competence. deviant behavior of adolescent boys and girls: Reducing the
Social Development, 10, 59–73. impact of risky friendships. Youth & Society, 33(4), 515–544.

123
J Youth Adolescence (2014) 43:1890–1902 1901

Davis, M. H. (1980). A multidimensional approach to individual James, D. J., Lawlor, M., Courtney, P., Flynn, A., Henry, B., &
differences in empathy. JSAS Catalog of Selected Documents in Murphy, N. (2008). Bullying behavior in secondary schools:
Psychology, 10, 85. What roles do teachers play? Child Abuse Review, 17, 160–173.
De Kemp, R., Overbeek, G., De Wied, M., Engels, R., & Scholte, R. Jolliffe, D., & Farrington, D. P. (2006). Examining the relationship
(2007). Early adolescent empathy, parental support, and antiso- between low empathy and bullying. Aggressive Behavior, 32,
cial behavior. The Journal of Genetic Psychology, 168(1), 5–18. 540–550.
Eccles, J. S., & Roeser, R. W. (2011). Schools as developmental Jolliffe, D., & Farrington, D. P. (2011). Is low empathy related to
contexts during adolescence. Journal of Research on Adoles- bullying after controlling for individual and social background
cence, 21(1), 225–241. variables? Journal of Adolescence, 34, 59–71.
Edens, J. F., Skopp, N. A., & Cahill, M. A. (2008). Psychopathic Kaukiainen, A., Bjorkqvist, K., Lagerspetz, K., Osterman, K.,
features moderate the relationship between harsh and inconsis- Salmivalli, C., Rothberg, S., et al. (1999). The relationships
tent parental discipline and adolescent antisocial behavior. between social intelligence, empathy, and three types of
Journal of Clinical Child and Adolescent Psychology, 37, aggression. Aggressive Behavior, 25, 81–89.
472–476. Laible, D., Eye, J., & Carlo, G. (2008). Dimensions of conscience in
Eisenberg, N., Fabes, R. A., & Spinrad, T. L. (2006). Prosocial mid-adolescence: Links with social behavior, parenting, and
development. In W. Damon & R. M. Lerner (Series Eds.) & N. temperament. Journal of Youth and Adolescence, 37, 875–887.
Eisenberg (Vol. Ed.), Handbook of child psychology, Vol. 3: Loukas, A., & Murphy, J. L. (2007). Middle school student
Social, emotional, and personality development (6th ed., perceptions of school climate: Examining protective functions
pp. 646–718). New York, NY: John Wiley. on subsequent adjustment problems. Journal of School Psychol-
Endresen, I. M., & Olweus, D. (2001). Self-reported empathy in ogy, 45, 293–309.
Norwegian adolescents: Sex differences, age trends, and rela- Lovett, B. J., & Sheffield, R. A. (2007). Affecive empathy deficits in
tionship to bullying. In A. C. Bohart & D. J. Stipek (Eds.), aggressive children and adolescents: A critical review. Clinical
Constructive and destructive behavior: Implications for family, Psychology Review, 27, 1–13.
school and society (pp. 147–166). Washington, DC: American Ma, X. (2002). Bullying in middle school: Individual and school
Psychological Association. characteristics of victims and offenders. School Effectiveness
Espelage, D. L., Mebane, S. E., & Adams, R. S. (2004). Empathy, and School Improvement, 13(1), 63–89.
caring and bullying: Toward an understanding of complex Malti, T., & Latzko, B. (2010). Children’s moral emotions and moral
associations. In D. L. Espelage & S. Swearer (Eds.), Bullying in cognition: Towards an integrative perspective. In B. Latzko & T.
American schools (pp. 37–61). Mahwah, NJ: Erlbaum. Malti (Eds.), Children’s moral emotions and moral cognition:
Flanagan, C. A., Cumsille, P., Gill, S., & Gallay, L. S. (2007). School Developmental and educational perspectives. New Directions for
and community climates and civic commitments: Patterns for Child and Adolescent Development (Vol. 129, pp. 1–10). San
ethnic minority and majority students. Journal of Educational Francisco: Jossey-Bass.
Psychology, 99(2), 421–431. Mayberry, M. L., & Espelage, D. L. (2007). Associations among
Galen, B. R., & Underwood, M. K. (1997). A developmental empathy, social competence, and reactive/proactive aggression
investigation of social aggression among children. Developmen- subtypes. Journal of Youth and Adolescence, 36, 787–798.
tal Psychology, 33, 589–600. McHale, S. M., Crouter, A. C., & Whiteman, S. D. (2003). The family
Gini, G. (2006). Social cognition and moral cognition in bullying: contexts of gender development. Social Development, 12(1),
What’s wrong? Aggressive Behavior, 32, 528–539. 125–148.
Gini, G., Albiero, P., Benelli, B., & Altoe, G. (2007). Does empathy McNeely, C. A., Nonnemaker, J. M., & Blum, R. W. (2002).
predict adolescents’ bullying and defending behavior? Aggres- Promoting school connectedness: Evidence from the National
sive Behavior, 33, 467–476. Longitudinal Study of Adolescent Health. Journal of School
Gini, G., Pozzoloi, T., & Hauser, M. (2011). Bullies have enhanced Health, 72, 138–146.
moral competence to judge relative to victims, but lack moral Metzler, C. W., Biglan, A., Ary, D. V., & Li, F. (1998). The stability
compassion. Personality and Individual Differences, 50, and validity of early adolescents’ reports of parenting constructs.
603–608. Journal of Family Psychology, 12, 600–619.
Grusec, J. E., & Hastings, P. D. (2007). Handbook of socialization: Moos, R., & Moos, B. (1986). The family environment scale manual
Theory and research. New York: The Guilford Press. (2nd ed.). Palo Alto, CA: Consulting Psychologists Press.
Hare, A. L., Marston, E. G., & Allen, J. P. (2011). Maternal Ojanen, T., Findley, D., & Fuller, S. (2012). Physical and relational
acceptance and adolescents’ emotional communication: A lon- aggression in early adolescents: Associations with narcissism,
gitudinal study. Journal of Youth and Adolescence, 40, 744–751. temperament, and social goals. Aggressive Behavior, 38,
Hart, D., & Carlo, G. (2005). Moral development in adolescence. 99–107.
Journal of Research on Adolescence, 15, 223–233. Oxford, M., Cavell, T. A., & Hughes, J. N. (2003). Callous/
Hastings, P. D., Utendale, W. T., & Sullivan, C. (2007). The unemotional traits moderate the relation between ineffective
socialization of prosocial development. In J. E. Grusec & P. parenting and child externalizing problems: A partial replication
D. Hastings (Eds.), Handbook of socialization: Theory and and extension. Journal of Clinical Child and Adolescent
research (pp. 638–664). New York: The Guilford Press. Psychology, 32, 577–585.
Hawley, P. H. (2003). Prosocial and coercive configurations of Paquette, J. A., & Underwood, M. A. (1999). Young adolescents’
resource control in early adolescence: A case for the well- experiences of peer victimization: Gender differences in
adapted Machiavellian. Merrill-Palmer Quarterly, 49, 279–309. accounts of social and physical aggression. Merrill-Palmer
Hoffman, M. (2000). Empathy and moral development. New York, Quarterly, 45, 233–258.
NY: Cambridge University Press. Peets, K., & Kikas, E. (2006). Aggressive strategies and victimization
Holt, M. K., & Keyes, M. A. (2004). Teachers’ attitudes toward during adolescence: Grade and gender differences, and cross-
bullying. In D. L. Espelage & S. M. Swearer (Eds.), Bullying in informant agreement. Aggressive Behavior, 32, 68–79.
American schools: A social-ecological perspective on prevention Puckett, M. B., Aikins, J. W., & Cillessen, A. H. N. (2008).
and intervention (pp. 121–139). Mahwah, NJ: Lawrence Moderators of the associations between relational aggression and
Erlbaum. perceived popularity. Aggressive Behavior, 34, 563–576.

123
1902 J Youth Adolescence (2014) 43:1890–1902

Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. Strayer, J., & Roberts, W. (2004). Empathy and observed anger and
M., Jones, J., et al. (1997). Protecting adolescents from harm. aggression in 5-year olds. Social Development, 13(1), 1–13.
Findings from the national longitudinal study on adolescent health. Sutton, J., Smith, P. K., & Swettenham, J. (1999a). Bullying and
Journal of the American Medical Association, 278, 823–832. ‘theory of mind’: A critique of the ‘social skills deficit’ approach
Robinson, R., Roberts, W. L., Strayer, J., & Koopman, R. (2007). to anti-social behavior. Social Development, 8, 117–127.
Empathy and emotional responsiveness in delinquent and non- Sutton, J., Smith, P. K., & Swettenham, J. (1999b). Social cognition
delinquent adolescents. Social Development, 16(3), 555–579. and bullying: Social inadequacy or skilled manipulation? British
Roeser, R. W., Eccles, J. D., & Sameroff, A. J. (1998). Academic and Journal of Developmental Psychology, 17, 435–450.
emotional functioning in early adolescence: Longitudinal rela- Twemlow, S. W., Fonagy, P., Sacco, F., & Brethour, J. R. (2006).
tions, patterns, and prediction by experience in middle school. Teachers who bully students: A hidden trauma. International
Development and Psychopathology, 10, 321–352. Journal of Social Psychiatry, 52, 187–198.
Roeser, R. W., Eccles, J. S., & Sameroff, A. J. (2000). School as a Underwood, M. K., Galen, B. R., & Paquette, J. A. (2001). Top ten
context of early adolescents’ academic and social-emotional challenges for understanding gender and aggression in children:
development: A summary of research findings. Elementary Why can’t we all just get along? Social Development, 10,
School Journal, 100, 443–471. 248–266.
Rose, A. J., & Rudolph, K. D. (2006). A review of sex differences in Van der Graaff, J., Branje, S., De Wied, M., & Meeus, W. (2012). The
peer relationship processes: Potential trade-off for the emotional moderating role of empathy in the association between parental
and behavioral development of girls and boys. Psychological support and adolescent aggressive and delinquent behavior.
Bulletin, 132, 98–131. Aggressive Behavior, 38, 368–377.
Rose, A. J., Swenson, L. P., & Waller, E. M. (2004). Overt and Xie, H., Swift, D. J., Cairns, B. D., & Cairns, R. B. (2002).
relational aggression and perceived popularity: Developmental Aggressive behaviors in social interaction and developmental
differences in concurrent and prospective relations. Develop- adaptation: A narrative analysis of interpersonal conflicts during
mental Psychology, 40(3), 378–387. early adolescence. Social Development, 11, 205–224.
Salmivalli, C., & Kaukiainen, A. (2004). ‘‘Female aggression’’ Young, E. L., Boye, A., & Nelson, D. (2006). Relational aggression:
revisited: Variable- and person-centered approaches to studying Understanding, identifying, and responding in schools. Psychol-
gender differences in different types of aggression. Aggressive ogy in the Schools, 43, 297–312.
Behavior, 30(2), 158–163.
Salmivalli, C., Kaukiainen, A., & Lagerspetz, K. (2000). Aggression and
sociometric status among peers: Do gender and type of aggression
matter? Scandinavian Journal of Psychology, 41, 17–24. Author Biographies
Shetgiri, R., Lin, H., Avila, R. M., & Flores, G. (2012). Parental
characteristics associated with bullying perpetration in U.S. Milena Batanova is a doctoral candidate in Health Behavior and
children aged 10 to 17 years. American Journal of Public Health Education and a Powers Fellow at the University of Texas at
Health, 102(12), 2280–2286. Austin. She studies a variety of individual (e.g., cognitive and
Shochet, I. M., Dadds, M. R., Ham, D., & Montague, R. (2006). School affective empathy) and social factors (e.g., school climate and
connectedness is an underemphasized parameter in adolescent connectedness) relevant to aggression, bullying involvement, and
mental health: Results of a community prediction study. Journal moral development in general.
of Clinical Child and Adolescent Psychology, 35(2), 170–179.
Smith, R., Rose, A. J., & Schwartz-Mette, R. A. (2009). Relational Alexandra Loukas is a Professor in Health Behavior and Health
and overt aggression in childhood and adolescence: Clarifying Education at the University of Texas at Austin. She studies the
mean-level gender differences and associations with peer interaction between individual and contextual influences on the
acceptance. Social Development, 19(2), 243–269. development of adolescent problem behaviors, including aggression
Soenens, B., Duriez, B., Vansteenkiste, M., & Goosens, L. (2007). and substance use.
The intergenerational transmission of empathy-related respond-
ing in adolescence: The role of maternal support. Personality
and Social Psychology Bulletin, 33, 299–311.

123
Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.

You might also like