See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225942538

Promoting Children’s Prosocial Behaviors in School: Impact of the “Roots of

Empathy” Program on the Social and Emotional Competence of School-Aged
Children

Article in School Mental Health · March 2011

DOI: 10.1007/s12310-011-9064-7

CITATIONS READS

178 2,386

4 authors, including:

Kimberly A. Schonert-Reichl Veronica Smith

University of Illinois at Chicago University of Alberta
139 PUBLICATIONS 8,067 CITATIONS 33 PUBLICATIONS 1,329 CITATIONS

SEE PROFILE SEE PROFILE

Anat Zaidman-Zait
Tel Aviv University
113 PUBLICATIONS 2,671 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kimberly A. Schonert-Reichl on 26 March 2015.

The user has requested enhancement of the downloaded file.

School Mental Health (2012) 4:1–21
DOI 10.1007/s12310-011-9064-7

ORIGINAL PAPER

Promoting Children’s Prosocial Behaviors in School: Impact

of the ‘‘Roots of Empathy’’ Program on the Social and Emotional
Competence of School-Aged Children
Kimberly A. Schonert-Reichl • Veronica Smith •

Anat Zaidman-Zait • Clyde Hertzman

Published online: 4 October 2011

 Springer Science + Business Media, LLC 2011

Abstract This study examines the effects of the Roots of crying, peer nominations of prosocial behaviors, and tea-
Empathy (ROE) program on children’s social-emotional cher reports of proactive and relational aggression. Self-
competence. ROE is a theoretically derived universal pre- reported empathy and perspective taking showed no sig-
ventive program that focuses on decreasing children’s nificant changes. According to ROE instructors’ diaries
aggression and facilitating the development of their social- assessing implementation, students in the experimental
emotional understanding and prosocial behaviors. The condition were exposed to all or most of the ROE curric-
program has as its cornerstone monthly visits by an infant ulum. These findings support and extend recent research
and his/her parent(s) that serve as a springboard for lessons examining the positive impacts of classroom-based social
on emotion understanding, perspective taking, caring for and emotional learning (SEL) programs on children’s
others, and infant development. The study included a social development and behavioral adjustment.
quasi-experimental control-group pretest–posttest, multi-
informant design with 585 4th- to 7th-grade children from Keywords Primary prevention
Social and emotional
28 classrooms. Outcome measures included self-reports of competence
Prosocial and aggressive behaviors

understanding of infant distress, empathy, and perspective School-aged children
taking, and peer and teacher reports of prosocial and
aggressive behaviors. Measures assessing implementation
were also included. Children in the ROE intervention Introduction and Background
classrooms showed significant improvement across several
of the domains assessed: self-reports of causes for infant Identifying the factors that predict children’s social and
emotional competence has garnered considerable empirical
attention in recent decades, in part, because such knowl-
K. A. Schonert-Reichl (&) edge affords an understanding of the mechanisms and
Department of Educational and Counselling Psychology, processes that may propel children away from or toward
and Special Education, University of British Columbia, maladaptation: strategic prerequisites for preventive efforts
2125 Main Mall, Vancouver, BC V6T 1Z4, Canada
to guide children onto successful developmental paths
e-mail: kimberly.schonert-reichl@ubc.ca
(Izard, 2002; Staub, 2003). Most of the studies conducted
V. Smith in this domain have focused on the identification of factors
Department of Educational Psychology, University of Alberta, that predict negative outcomes, such as aggression and
Edmonton, AB, Canada
mental illness, with relatively little attention given to
A. Zaidman-Zait delineating the ways in which children’s prosocial behav-
Human Early Learning Partnership, College of Interdisciplinary iors and characteristics can be advanced. However, recent
Studies, University of British Columbia, Vancouver, BC, Canada years have witnessed a shift from a preoccupation with
repairing weaknesses to the enhancement of positive
C. Hertzman
School of Population and Public Health, University of British qualities and in preventing or heading off problems before
Columbia, Vancouver, BC, Canada they arise (Damon, 2004; Damon & Gregory, 2003;

123
2 School Mental Health (2012) 4:1–21

Huebner & Gilman, 2003; Seligman & Csikszentmihalyi, Empathy’’ (ROE)—on a sample of 4th- to 7th-grade chil-
2000). Implicit in this trend is the assumption that educa- dren’s social-emotional understanding and social behaviors
tional interventions can be designed not to only deter after exposure to the intervention.
children’s problem behaviors but to foster children’s
strengths, positive behaviors, and resiliency (e.g., Masten
& Motti-Stefanidi, 2009). Description of the Intervention
School-based prevention efforts have been heralded as
an effective and cost-effective way to promote children’s ROE Curriculum
positive development and stave off an upward trajectory of
mental illness and aggressive behaviors (Greenberg, 2010; ROE is a program for children in the kindergarten through
Weissberg & Greenberg, 1998)—problems that have been 8th grade (Gordon, 2005). It was developed in 1996 and
shown to increase over the course of the elementary school has been implemented to children in urban, rural, and
years (e.g., Aber, Brown, & Jones, 2003; Farmer & Xie, Aboriginal communities in Canada, internationally in the
2007; Murray-Close, Ostrov, & Crick, 2007; Neal, 2010). United States, New Zealand, the Isle of Man, Northern
Bolstered by evidence that empirically based curricula can Ireland, Republic of Ireland, and Scotland with pilot
prevent the onset of problem behaviors, such as anxiety and implementations elsewhere (Kendall et al. 2006; Yau,
aggression, and decrease antisocial behavior (e.g., Ang & 1999). Since the program began, it has reached over
Hughes, 2001; Cook, Murphy, & Hunt 2000; Durlak, 325,000 children worldwide.
Weissberg, Dymnicki, Taylor, & Schellinger, 2011; The primary goals of ROE are to (1) develop children’s
Fonagy et al., 2009; Gresham, Cook, Crews, & Kern, 2004; social and emotional understanding, (2) promote children’s
Jones, Brown, & Aber, 2011; Lösel & Beelmann, 2003), prosocial behaviors and decrease their aggressive behaviors,
many school districts throughout North America have and (3) increase children’s knowledge about infant devel-
strengthened their efforts to include programs that promote opment and effective parenting practices. ROE is a 9-month
students’ social and emotional competence. Because vir- program that has as its cornerstone a monthly classroom visit
tually all children go to school during key developmental by an infant and his/her parent(s) whom the class ‘‘adopts’’ at
years, this context is important when fostering children’s the beginning of the school year. It is during these monthly
positive social and emotional development. Due to many visits that children learn about the baby’s growth and
competing demands and time constraints, however, schools development via interactions and observations with the
need to prioritize their efforts and implement curricular baby. Each month, an ROE program instructor1 visits his/her
approaches that are evidence-based and both time- and participating ROE classroom three times, once for a pre-
cost-effective. family visit, another time for the visit with the parent and
Although there has been increased theoretical and infant, and finally, a postfamily visit. The lessons for the
research attention aimed at delimiting the social and bimonthly visits from the instructor are designed to foster
emotional factors that significantly impact children’s psy- empathy, emotional understanding, and problem-solving
chological well-being and behavioral adjustment (Denham, skills through discussion and activities in which the parent–
1998; Hymel, Schonert-Reichl, & Miller, 2007; Saarni, infant visit serves as a springboard for discussions about
1999), to date, few studies exist that examine the effec- perspective taking, caring for others, infant development,
tiveness of interventions specifically targeted to promoting and effective parenting practices. Each lesson is designed to
children’s prosocial behaviors—those behaviors and char- capitalize on shared observations from the family visit.
acteristics such as cooperativeness, helpfulness, trustwor- Lesson plans and accompanying activities are scripted to
thiness, and kindness (for exceptions, see Battistich, match the age of the baby and are calibrated to the students’
Solomon, Watson, Solomon, & Schaps, 1989; Conduct level of development. (Four different curriculums are
Problems Research Group, 1999; Denham & Burton, 1996; available: kindergarten, 1st to 3rd grade, 4th to 6th grade,
Greenberg, Kusché, Cook, & Quamma, 1995). As noted by and 7th and 8th grade). In addition to the three classroom
Izard (2002), ‘‘Although arousal of any negative emotion visits per month by the instructor, each classroom teacher
may provide the basis for strong learning, the induction and utilizes the lessons and ideas presented during the specific
utilization of the emotions relating to empathy, sympathy, ROE lesson within the general education program.
and prosocial behavior have the greatest relevance for Each ROE lesson provides opportunities to discuss and
preventive interventions’’ (p. 801). This present study learn about different dimensions of empathy, in particular
contributes to ongoing scholarship in the school-based emotion identification and explanation, perspective taking,
promotion of children’s social and emotional competence
by reporting empirical impacts of a unique social and 1
See the Method section for a full description of instructor training
emotional learning (SEL) intervention—the ‘‘Roots of and selection.

123
School Mental Health (2012) 4:1–21 3

and emotional sensitivity. Across various lessons, children draws on the functionalist approach to emotions (Campos,
are invited to identify the emotions of the baby and to Mumme, Kermoian, & Campos, 1994), wherein emotion
provide explanations for those emotions. Children then understanding and expressivity are seen as playing central
become engaged in lessons through stories, art projects, roles in the establishment and maintenance of children’s
and general classroom activities in which they reflect and interpersonal relationships (Saarni, 1999). Empathy, in
discuss their own emotions and the emotions of others. For particular, is core to the ROE curriculum because it is
theme 3 (Caring and Planning for the baby), for instance, in central to the genesis and enactment of prosocial behavior.
the pre-family visit, the instructor reads the book Sasha and Indeed, there is a great deal of empirical evidence indi-
the Wriggly Tooth (by R. Tregebov, 1993) to the children. cating that not only empathy leads individual to desist from
Afterward, the ROE instructor leads a discussion about the aggression but also the ability to empathize is central to the
mixed feelings that can ensue when one looses a tooth development of positive social relationships and caring and
(e.g., ‘‘happy to be getting a visit by the tooth fairy,’’ kind behaviors (Eisenberg & Miller, 1987; Eisenberg, Fa-
‘‘embarrassed because you may look or talk funny with a bes, & Spinrad, 2006). Feshbach (1979) has put forth a
missing tooth’’). In the subsequent parent and infant visit, three-component model in which empathy is conceptual-
children are provided opportunities to perspective-take via ized as comprising two cognitive components and one
asking questions of the parents about their feelings about affective component: the ability to discriminate and label
their infant’s teething experience (e.g., ‘‘how does it feel to affective states in others; the ability to assume the per-
see your baby in pain,’’ ‘‘what do you do to help your baby spective and role of another person; and emotional
feel better?’’). responsiveness (i.e., the affective ability to experience
In keeping with other comprehensive social and emo- emotions). This framework provides a conceptual basis for
tional competence promotion programs, embedded within the content of the ROE lessons and an organizing scheme
the ROE program are explicit components aimed at creating for the measures utilized in this evaluation.
an ecology in the classroom environment in which belong-
ing, caring, collaboration, and understanding others are
emphasized (Cohen, 2001; Noddings, 1992). Throughout The Present Study
each of the lessons are opportunities for children to be
engaged in activities that benefit their baby. For example, in The current research was designed to extend understanding
one lesson, children create a class recording of nursery of the school-based promotion of children’s social-emo-
rhymes and songs for their baby, and in another lesson, at the tional competence in several ways. First, we evaluated the
end of the school year, the children create a ‘‘wishing tree’’ effectiveness of ROE on the development of multiple
whereby each child in the classroom writes a wish that she or dimensions of children’s emotional understanding. More
he has for the baby’s life. Such activities, specifically those in specifically, we included assessments of children’s self-
which individuals work collectively on projects that benefit reported empathy and perspective taking along with
others, have been shown elsewhere to promote altruism and a assessments of their knowledge of emotion causes and
prosocial value orientation (Battistich, Solomon, Watson, & knowledge of emotion management/regulation strategies
Schaps, 1997; Noddings, 1992; Staub, 1988, 2003). using the crying infant paradigm—a paradigm that has pre-
viously demonstrated significant associations with empathy
The ROE Program’s Theory of Change in school-aged children (Catherine & Schonert-Reichl,
2011). According to the functionalist theory of emotion
ROE is based on empirical understandings of the processes (Barrett & Campos, 1987; Campos et al., 1994), the devel-
and mechanisms that elicit both aggression and prosocial opment of emotion regulation skills enables children to adapt
behaviors, as well as theoretical models that guide related successfully within their social environment because such
programs targeting a range of social behaviors (see Con- skills assist children in their ability to cope with their own
duct Problems Research Group, 1999; Consortium on the emotions as well as the emotional displays of others (e.g.,
School-Based Promotion of Social Competence, 1996). Rubin, Coplan, Fox, & Calkins, 1995). The development of
The ROE curriculum model draws on research and theory emotional regulation/management is posited to arise through
in developmental psychology, suggesting that emotion socialization practices that give children opportunities to
processes and social understanding play critical roles in learn how to label and interpret emotions, when emotion
children’s interpersonal relationships and social behaviors expression is warranted, and how to manage emotion arousal
(e.g., Izard, 2002; Shipman, Zeman, Penza, & Champion, (Lewis & Michalson, 1983)—skills that are specifically
2000), such that ‘‘emotions form the motivational bases for targeted in the ROE program curriculum. To our knowledge,
empathy and prosocial behavior’’ (Izard, Fine, Mostow, no previous studies have examined the effects of an inter-
Trentacosta, & Campbell, 2002, p. 761). In particular, ROE vention on children’s understanding of infant emotions.

123
4 School Mental Health (2012) 4:1–21

Second, cognizant of recent research indicating that In the present study, we also assessed intervention
children’s prosocial behaviors are important correlates of effects on another form of aggression that has risen to
empathy (e.g., Eisenberg, Fabes, & Spinrad, 2006; prominence in the theoretical and empirical literature in
Schonert-Reichl, 1993; see Schonert-Reichl & Oberle, recent years—a form of aggression that includes behaviors
2011 for a review) and intellectual competencies such as that are more covert and indirect in nature and are aimed at
academic performance (e.g., Caprara, Barbanelli, Pastorelli, damaging the target’s social relationships as opposed to
Bandura, & Zimbardo, 2000; Wentzel, 1993), we also inflicting bodily harm. Distinct from physical aggression,
included outcome measures that assessed a range of pro- different terms have been used to describe this more
social behaviors. These behaviors such as sharing, helping, subtle form of aggression, including indirect aggression
cooperation, and kindness are particularly important to (Lagerspetz, Bjorkvist, & Peltonen, 1988), social aggres-
examine in the context of an evaluation of a school-based sion (Cairns, Cairns, Neckerman, Gest, & Gariépy, 1988),
social-emotional competence promotion program because and relational aggression (Crick & Grotpeter, 1995). In the
these behaviors are considered to be hallmarks of social present investigation, we use Crick’s (1996) definition in
competence in childhood and adolescence (Wentzel, which relational aggression refers ‘‘… to harming others
Filisetti, & Looney, 2007). In addition, it is well estab- through purposeful manipulation or damage to their peer
lished that these behaviors have been related theoretically relationships (e.g., using social exclusion as a form of
and empirically to mental health and well-being (Hay, retaliation).’’ (p. 2317). Several recent studies have found
1994) and other forms of social competence such as peer that this form of aggression becomes increasingly common
acceptance and positive social relationships (Bukowski & over the course of the elementary school years (for a
Sippola, 1996; Oberle, Schonert-Reichl, & Thomson, 2010; review, see Neal, 2010), with one study showing a linear
Schonert-Reichl, 1999). increase in relational aggression among 4th-grade girls
Third, in addition to assessing the effects of the ROE over one calendar year (Murray-Close et al., 2007; see
program on children’s prosocial behaviors, we also exam- Neal, 2010 for a review). A recent meta-analysis of 148
ined a broader corpus of aggressive behaviors than typically studies on child and adolescent direct and relational
assessed in evaluations of interventions, going beyond the aggression found gender effects for direct aggression
usual exclusive focus on physical aggression. Historically, (favoring boys) and negligible gender effects for relational
researchers have often focused on overt forms of aggression, aggression (Card, Stucky, Sawalani, & Little, 2008). This
such as hitting, pushing, and verbal threats. However, recent finding of trivial gender differences in relational aggression
theoretical and methodological advances in the study of challenges current portrayals of this form of aggression
aggression have revealed the existence of a wider range of being more enacted by girls. Unfortunately, few studies to
aggressive behaviors that can be distinguished with regard to date have examined how a classroom-based intervention
their nature (e.g., overt, covert), function (e.g., social dom- can reduce relational aggression. To our knowledge, this is
inance), and underlying social-cognitive processes (e.g., one of the first evaluations of a primary preventive inter-
Arsenio & Lemerise, 2001; Crick & Dodge, 1999; Sutton, vention aimed at promoting social and emotional compe-
Smith, & Swettenham, 1999). Dodge and Coie (1987) and tence in which the dimensions of physical, proactive,
others (e.g., Vitaro, Brendgen, & Tremblay, 2002), for reactive, and relational aggression were examined con-
instance, have suggested two types of aggressive behavior comitantly in one study.
that can be distinguished by their underlying function and Fourth, we assessed children’s prosocial and aggressive
motivation: proactive and reactive aggression. Proactive behaviors via multi-informants. Specifically, teachers and
aggressive behavior is instrumental and involves attempts to peers were selected as informants as they are typically used
influence or coerce others through aversive means in an as the main informants in research on children’s social
unprovoked situation, requiring neither provocation nor behaviors in school (e.g., Ladd & Troop-Gordon, 2003).
anger. Proactive aggression has frequently been described as Moreover, Cappella and Weinstein (2006), in their evalu-
‘‘cold-blooded’’ because proactively aggressive individuals ation of a social aggression prevention program (SAPP) for
typically ‘‘… view aggression as an effective and viable 5th-grade girls, noted disparate intervention effects
means for obtaining social goals’’ (Crick & Dodge, 1999, depending on whether the informant was the teacher or a
p. 998). In contrast, reactive aggressive behavior has been peer, rendering multiple informants important when
described as affective, impulsive, defensive, and ‘‘hot- investigating the effectiveness of an SEL intervention. For
headed,’’ involving angry outbursts in response to actual or our peer assessments, we utilized a peer behavioral
perceived threats or provocations. Distinguishing between assessment method whereby children nominate their
proactive and reactive aggression is important on both the- classmates who fit particular behavioral characteristics.
oretical and empirical grounds and may also be important for Following the methodology outlined by Parkhurst and
the design of prevention and intervention programs. Asher (1992), we used single, precisely focused peer

123
School Mental Health (2012) 4:1–21 5

assessment items to measure specific prosocial and developers (Domitrovich & Greenberg, 2000; Durlak &
aggressive behaviors. Using single item for each behavioral DuPre, 2008). Central to our question was the degree to
dimension is an approach that has been frequently used in which the ROE program’s objectives and procedures were
research on children’s peer relationships (Coie & Dodge, put into everyday practice in the classroom. Hence, program
1988). Peer behavioral assessments provide data from implementation was monitored with lesson diaries and a
multiple raters of a child’s behavior and may also provide self-report survey of lesson implementation to determine
information regarding a wider corpus of observations in the degree to which the program lessons were delivered
settings beyond the classroom. Peer reports increase the with fidelity.
power of assessment because many peers are involved in
the evaluation rather than just a single teacher. Moreover,
because peers can act as ‘‘participant observers,’’ they are Method
in a particularly significant position to provide assessment
information to the researcher about the social behaviors of Participants
their classmates (Younger, Schneider, Wadeson, Guirguis,
& Bergeron, 2000). For the present investigation, children Children
provided ratings of their classmates on dimensions of
prosocial and antisocial/aggressive behaviors. Participating children were drawn from regular education
Finally, recognizing that there are no ‘‘true control classrooms in the 4th through 7th grades of 28 public
groups’’ in school-based experimental or quasi-experi- elementary schools across two Canadian cities (18 schools
mental research (Oakley et al., 2006), and following rec- in Vancouver, BC; 10 schools in Toronto, Ontario).
ommendations by Greenberg (2004) and Durlak and DuPre Schools were chosen based on their representativeness of
(2008), we obtained information in order to document the the range of socioeconomic and racial/ethnic diversity of
experiences and interventions received by children in both students in the school district. Fourteen of the schools had a
the intervention and control classrooms, other than the classroom in which the ROE program was implemented.
ROE program. To avoid contamination, 14 control classrooms were
selected from 14 schools where no ROE program was
currently being implemented. The program classrooms
Aims and Hypotheses were matched as closely as possible with respect to gender
of teacher, class size, grade level, and child characteristics
This evaluation study is a quasi-experimental control-group (e.g., race/ethnic and gender composition of children,
pretest–posttest, multi-informant design. We matched inter- percent of children whose home language was different
vention with control classes based on average age, gender, than English). Active consent for the assessment phase of
and English as a second language (ESL). We hypothesized the intervention was obtained using parental permission
that when compared to control children, ROE program chil- forms that were translated into seven languages (i.e., Chi-
dren would show significant and positive changes from pre- nese, Vietnamese, Punjabi, Spanish, Polish, Russian, and
test to posttest on our measures of understanding of infant Korean). All children recruited were from regular class-
crying, empathy, perspective taking, and teacher- and peer- rooms and hence were deemed to be competent to read and
rated social behavior. Our dual focus on increasing social and write English. Prior to providing children with the parental
emotional competence and reducing aggression was war- permission slips, either a trained research assistant or the
ranted due to empirical evidence indicating that prognosis for principal investigator of the research project provided a
children is poorest when they demonstrate a combination 15-min presentation to each participating class, describing
of low social-emotional competence alongside aggressive the study in age-appropriate language and answering chil-
behavior (CPPRG, 1999; Tolan, Guerra, & Kendall, 1995). dren’s questions. Permission rates were high, with an
Moreover, cognizant of recent research indicating increases average consent rate of 92%.
in children’s aggression during the late elementary school The recruited sample included 638 children attending
years (e.g., Aber et al., 2003; Murray-Close et al., 2007), we school in the fall of 4th to 7th grades. Fifty-three children
examined both the magnitude and direction of changes in (8%) were lost to the study because they moved away prior
aggressive and prosocial behaviors across the school year to posttesting. Attrition was comparable across the ROE
(Wilson, Lipsey, & Derzon, 2003). program and control groups [ns = 34 and 19, respectively;
A second purpose of our study was to examine the X2 (1, N = 53) = 2.74, ns]. Further analyses revealed that
fidelity of the ROE program when implemented in a ‘‘real- there were no significant differences in any of the pretest
world’’ setting delivered in regular elementary school scores between those children who attrited and those who
classrooms with minimal support from the program remained, Fs \ 1.

123
6 School Mental Health (2012) 4:1–21

The final sample was comprised of 585 children (280 The Intervention
girls, 305 boys) in the 4th to 7th grades who remained
in their respective intervention (n = 306) or control The ROE program began in late fall, and all 26 lessons
(n = 279) classrooms throughout the one school-year across the nine themes (i.e., Meeting the Baby, Crying,
intervention and were available for pre- and posttesting. Caring and Planning for the Baby, Emotions, Safety, Sleep,
Ages ranged from 8 years, 8 months to 12 years, 8 months Communication, Who Am I?, Goodbye, and Good Wishes)
at pretest, with a mean of 10 years, 7 months (SD = .87). were completed by the end of the school year across all of
With regard to first language learned, 40% of the partici- the participating classrooms. The time for each classroom
pants reported English as their first language, 39% reported lesson ranged from 30 to 45 min. Each instructor utilized
Chinese, 8% reported Vietnamese, and the remaining 13% the ROE intermediate-grade curriculum (Gordon, 2001) to
reported other languages (e.g., Punjabi, Spanish, Polish, implement the program and followed the guidelines
Russian, Tamil, and Filipino). These percentages reflect the explicated in each lesson. Each theme involved three les-
ethnic diversity of the cities in which the research took sons: (1) a pre-family visit that introduced the theme; (2) a
place. Because the design of this study was quasi-experi- class visit by the ‘‘classroom baby’’ and his/her parent(s)/
mental, it was important to demonstrate the similarity of caregiver(s), where the instructor directed interactions and
the groups across background and potentially contaminat- observations with baby and parent(s) to the students to
ing variables (Cook & Campbell, 1979). To examine for enhance learning about infant development; and (3) a
potential differences between our two groups, T tests were postfamily visit where the students were guided to discuss
utilized for continuous variables and chi-square analyses the baby visit and conclude the theme. In consonance with
for categorical variables. Analyses revealed no differences the goals of the ROE stated earlier, each lesson fostered
between intervention and control classrooms with regard to these goals through discussion and activities in which the
age, gender, or first language learned (ps [ .25), suggest- parent–infant visit served as a springboard for discussions
ing that our matching process was successful. about understanding feelings, infant development, and
effective parenting practices.
Teachers Because ‘‘pre-implementation training and on-site
coaching enable service providers to implement programs
Teachers in both the ROE and control group completed a with greater integrity and to have more positive effects on
questionnaire that assessed the following dimensions: sex, the children they serve’’ (Weissberg & Greenberg 1998,
race/ethnicity, and teaching background (years of teaching p. 925), ROE instructors underwent an intensive three-day
experience). Of the 28 teachers who participated in the training session, prior to program implementation, wherein
research, 27 completed background questionnaires (14 they were provided with program goals, as well as its
program and 13 control). Analyses of background charac- theoretical and empirical underpinnings. Instructors were
teristics of the intervention and control teachers revealed provided with a training manual (Gordon, 2000) delineat-
no significant differences. The majority of the teachers ing the theory and research guiding each ROE lesson, and a
were women (71% program and, 69% control), most curriculum manual in which plans for each lesson were
identified themselves as European Canadian (86% program thoroughly described (Gordon, 2001). The training session
and 84% control), and the majority of teachers in both included discussions on infant and child social and emo-
groups had 10 or more years of teaching experience (66% tional competence as well as lecture, video, and readings,
program and 70% control). and role-plays of curriculum instructional techniques. At
the end of the three-day training session, all instructors
ROE Program Instructors were tested on the program and its components, and those
who passed became certified ROE program instructors.
ROE instructors also completed a background question- ROE instructors were mentored throughout the course of
naire that allowed for comparisons between instructors on the program implementation. Mentors, individuals with
the following dimensions: sex, ethnicity, and experience extensive experience implementing the ROE program,
teaching the ROE program. The majority of ROE instruc- were trained and certified as ‘‘Roots of Empathy’’ leaders.
tors were women (93%) and included one counselor, three They visited the classrooms during program implementa-
youth and family workers, one community school coordi- tion to provide support and feedback to the instructor, as
nator, four community health nurses, one parent, and four well as helping trouble-shoot special situations and facili-
non-enrolling teachers. Years of experience in education or tating instructors’ learning from their programs and each
a similar position ranged from 5 to 33 with a mean of 17.5 other. All instructors received a ‘‘booster session’’ mid-way
(SD = 9.7). Exactly half of the instructors had previous through the school year. This session allowed an opportu-
experience of delivering the ROE program. nity for instructors to share their experiences with one

123
School Mental Health (2012) 4:1–21 7

another, ask questions, and obtain assistance for any issues frequency with which they implemented SEL activities
that they confronted with regard to program implementa- (other than those associated with the ROE program) in their
tion. Finally, all school principals and teachers were pro- classroom across the school year on a 5-point scale
vided with explicit guidelines to follow in order to maintain (1 = not at all to 5 = a lot). Additionally, we asked
the integrity of ROE program delivery. teachers in both ROE and control classroom to report on
As can be surmised, intervention fidelity was ensured by their training (e.g., workshops) on SEL as well as their
manualization of all components, mid-way booster session, beliefs about the importance of teaching activities to pro-
and ongoing supervision of ROE instructors via the ROE mote students’ social and emotional learning (SEL). With
mentor. Outside interventions were neither encouraged nor regard to this latter dimension, we asked teachers to rate on
discouraged and were assumed to occur at the same rate for a 5-point scale (1 = not very important to 5 = very
intervention and control groups. The control condition was important) ‘‘What importance do you give to activities or
a ‘‘treatment as usual’’ comparison that included regular teaching strategies that influence students’ social and
school prevention programs to the extent that schools chose emotional learning in your classroom?’’
to use them.
Direct Child Assessments: Demographic Information
Procedure
At pretest, children provided information about their sex,
Data were collected two to four weeks prior to exposure to ages, grade levels, first language learned, and family
the intervention in October (time 1) and two to four weeks composition.
following the end of the intervention in June (time 2). At
both time periods, self-report, peer-report, and teacher- Direct Child Assessments: Understanding of Infant
report instruments were used to assess participants’ emo- Crying
tion understanding, empathy and perspective taking, and
prosocial and aggressive/antisocial behaviors. Child self- Children’s understanding of infants’ crying was assessed
report and peer behavioral assessment measures were via a task adapted from the infant facial expression of
group-administered to children in their classrooms by emotion (IFEEL) pictures (Emde, Osofsky, & Butterfield,
research assistants, while a second research assistant cir- 1993; modified by Catherine & Schonert-Reichl, 2011),
culated to monitor children’s survey completion and designed to assess an individual’s understanding of the
answer any of their questions. Children who were non- causes of infants’ cries and knowledge of behavioral
consented worked on an alternative activity related to the strategies for helping a crying infant. Data were collected
regular school curriculum. Teachers completed ratings of using a booklet of self-report questionnaires. On one
children’s social behaviors at pretest and posttest. They page, children were shown a black-and-white photograph
were given a $200.00 honorarium for their participation in of an infant crying. The picture was cropped so that only
the study. the face of the infant was shown and was gender-neutral,
with no apparent clothing or identifiers depicting either a
male or female infant. On a separate page, children were
Measures presented with two questions designed to address two
conceptual categories, tapping their thoughts about
Assessment of Implementation (a) causes for an infant’s distress (i.e., ‘‘Why does this
baby cry?’’) and (b) emotion regulation strategies or
To assess program adherence and participant responsive- action responses to the display of emotions (i.e., ‘‘What
ness, ROE instructors were asked to complete a monthly are some ways to help a crying baby?’’). Our assessment
‘‘diary’’ in which they indicated the total number of lessons focused on two dimensions of emotion understanding—
completed (exposure), the percentage of curricular com- causality of emotions and emotion regulation strategies—
ponents completed within each lesson (adherence), and a that, based on the ROE curriculum, we believed would be
rating of the quality of student engagement (participant of particular importance in determining the ROE pro-
responsiveness) during each lesson on a 5-point scale gram’s effectiveness in advancing children’s understand-
ranging from 1 (‘‘not engaged’’) to 5 (‘‘very engaged’’). ing of infant emotions. Children were not limited with
In order to garner information about program differen- regard to the number of causes and strategies that they
tiation, we also collected data assessing the degree to could offer. In some cases, several responses produced by
which both ROE and control teachers implemented SEL the children were infrequently used, did not fit the defi-
programs and practices outside of those associated with the nitions of the other categories, or did not address the
ROE program. Specifically, we asked teachers to rate the question properly (e.g., responding ‘‘baby is wrestling/

123
8 School Mental Health (2012) 4:1–21

fighting’’ as a cause). These responses were excluded used. The empathic concern scale assesses the tendency to
from further analyses. feel compassion and concern for other individuals (e.g., ‘‘I
Responses to the question assessing children’s under- often feel sorry for people who don’t have the things I
standing for causes for infants’ cries were coded following have’’), whereas the perspective-taking subscale measures
a modified version of emotion casual explanations’ coding the tendency to consider things from others’ viewpoints
developed by Fabes, Eisenberg, Nyman, and Michealieu (e.g., ‘‘Sometimes I try to understand my friends better by
(1991). Each response was first assigned to two superor- imagining how they think about things’’). Participants rated
dinate categories depending on whether the child’s each item on a 5-point rating scale (ranging from
responses indicated references to internal states (e.g., ‘‘the 1 = never to 5 = very often). Scores were computed by
baby is sad because he wants his mom’’) or references to averaging item scores within subscales so that higher
external behaviors or actions (e.g., ‘‘the baby is crying scores signified greater empathic concern and perspective
because somebody took his toy’’). References to internal taking, respectively. Supportive evidence for the construct
states were then further subdivided into those references validity of the empathic concern and perspective taking
that referred to (a) an emotional state (e.g., feelings—‘‘the subscales of the IRI has been obtained in previous research
baby is upset,’’ ‘‘the baby is scared’’) or (b) a physical state (Davis, 1983), including significant correlations with rela-
(e.g., being sick or hurt, physical needs—‘‘the baby is ted constructs in expected directions (Schonert-Reichl,
thirsty,’’ ‘‘the baby is tired,’’ ‘‘the baby has an ear 1993). For the present study, both scales demonstrated
infection’’). adequate internal consistency both at pretest and at post-
Coding of responses to the question assessing emotion test. Specifically, Cronbach alphas were .80 at pretest and
regulation strategies (i.e., ‘‘What are some ways to help a .86 at posttest for the empathic concern subscale and .74 at
crying baby?’’) was based on past research regarding the pretest and .80 at posttest for perspective-taking subscale.
categorization of emotion management processes for reg-
ulating emotional distress (Shipman, Zeman, Penza, & Peer Report of Child: Peer-Rated Prosocial
Champion, 2000). Specifically, children’s responses were and Aggressive/Antisocial Behaviors
coded into one of the following three categories: 1) provide
physical support (e.g., ‘‘give it a bottle,’’ ‘‘change the We used a peer nomination and behavioral assessment
baby’s diaper’’), 2) provide comfort object/environment method to assess children’s prosocial and antisocial/
(e.g., ‘‘give the baby a pacifier/soother,’’ ‘‘put the baby in aggressive behaviors. Specifically, unlimited and cross-
the swing’’), and 3) provide social/relational support (e.g., gender peer nominations were used to obtain independent
‘‘put the baby in his/her mother’s lap,’’ ‘‘make it feel assessments of children’s social behavior. Six types of
loved’’). The children’s answers were scored from the prosocial behaviors (‘‘shares,’’ ‘‘cooperates,’’ ‘‘helps other
questionnaires. One coder with no additional information kids when they have a problem,’’ ‘‘kind,’’ ‘‘understands
about the child (i.e., the child’s responses to questions other kids’ point of view,’’ and ‘‘fair’’), and two types of
about other emotions or the child’s gender or group aggressive/antisocial behaviors (‘‘starts fights’’ and ‘‘breaks
assignment) coded all of the children’s responses. A second rules’’) were assessed. For each behavior, children were
coder, also with no information about the child or other provided with a list of all of their classmates participating in
responses, independently coded a random selection of 15% the research and were asked to circle the names of their
of children’s responses. Inter-rater reliability was fairly classmates on each list who fit the behavior description;
high for both the emotion causes category (K = .82) and children could circle as many or as few names as they
the emotion regulation strategies category (K = .92). All wanted. For each question, the percentage of nominations
scores in data analyses were based on adjudication of two each child received was computed by dividing the number of
coders’ decisions where there had been disagreements. nominations received by the total number of participating
children in the classroom. For each of the prosocial and
Direct Child Assessments: Empathic Concern aggressive/antisocial characteristics and behaviors, chil-
and Perspective Taking dren’s nominations were standardized within each class-
room, and a proportional nominations score was calculated
Participants’ empathy-related responding was assessed via per child for each of the behaviors/characteristics.
the Interpersonal Reactivity Index (IRI; Davis, 1983), a
self-report measure comprising four 7-item subscales Teacher Report of Child: Children’s Prosocial
(empathic concern, perspective taking, fantasy, personal and Aggressive Behaviors
distress), each of which taps a separate dimension of
empathy. For the purposes of the present study, only the Teachers rated children’s aggressive and prosocial behav-
empathic concern and perspective-taking subscales were iors on corresponding subscales of the Child Behavior

123
School Mental Health (2012) 4:1–21 9

Scale (CBS; Ladd & Profilet, 1996), an instrument Results

designed to assess children’s behaviors with peers at
school. Both of these subscales have evidence supporting The results will be described in two parts. First, pre-
their reliability and validity with school-age children (see liminary analyses are presented. In this section, we
Ladd & Profilet, 1996). The aggressive and prosocial describe results of analyses examining baseline differences
behavior subscales from the CBS demonstrated adequate between intervention and control groups along with anal-
internal reliability in the present study at pretest and yses examining implementation fidelity and dosage. Fol-
posttest, as reported below. The CBS Aggressive subscale lowing, we present results from a series of multivariate
contained seven items that referred to physical (e.g., analyses of covariance (MANOVAs) and analyses of
‘‘Kicks, bites, or hits other children’’) and verbal (e.g., covariance (ANCOVAs) employed to address the study’s
‘‘Threatens other children’’) aggression, and the CBS primary questions.
Prosocial Behavior subscale was composed of seven items
tapping e.g., cooperative (e.g., ‘‘Cooperative with Peers’’), Preliminary Analyses
and empathic and prosocial (e.g., ‘‘Shows a recognition of
the feelings of others; is empathic,’’ ‘‘Offers help or com- Baseline Comparisons
fort when other children are upset’’) behavior. An addi-
tional CBS subscale included six ‘‘filler’’ items that are not A series of ANOVAs was performed to check for mean
scored. These items refer to other types of classroom pretest differences between ROE program and control
behaviors (e.g., ‘‘Listens to classmates’’) and were included children on all outcome measures. As expected, owing to
as a way to discourage teachers from forming specific the matching procedures we employed, there were no
hypotheses about the specific constructs being assessed. intervention-control differences at the preintervention
Teachers rated children’s behavior with scale points baseline for any outcomes examined (Fs \ 1).
labeled and defined as follows: 1 = doesn’t apply (‘‘child
seldom displays the behavior’’), 2 = applies sometimes Intervention Implementation
(‘‘child occasionally displays the behavior’’); 3 = certainly
applies (‘‘child often displays the behavior’’). Implementation adherence, as assessed via instructors’
To assess proactive, reactive, and relational aggression, diaries, revealed that the average number of ROE lessons
items were taken from existing measures of proactive taught was 25.21 (SD = 1.25) out of the total 26 lessons
aggression (9 items; e.g., ‘‘Gets other children to gang up (range: 23 to 26 lessons), indicating an implementation
on a peer that he/she does not like’’), reactive aggression level of 96%. Of the 14 ROE instructors, 11 (80%) of them
(11 items; e.g., ‘‘When a peer accidentally hurts this child, reported that they implemented all 26 lessons, and the
such as by bumping into him/her, this child assumes that remaining three instructors reported implementing 23 of
the peer meant to do it, and them overreacts with anger or the 26 lessons. In addition to reporting the total number of
fighting’’) (Dodge & Coie, 1987), and relational aggression lessons completed and to obtaining more in-depth infor-
(4 items; e.g., ‘‘When mad at someone, tries to get others to mation about program adherence, ROE instructors also
dislike that person’’) (Crick, 1996) and were added to the reported the percentage of content/material that they
CBS to complement the range of behaviors that teachers implemented within each of the 26 lessons. Overall, ROE
rated. As with the CBS aggressive and prosocial subscales, instructors reported implementing, on average, 95% (range
teachers rated children’s proactive, reactive, and relational from 89% to 100%) of the lesson plan content across the 26
aggression on a 3-point scale ranging from 1 (doesn’t lessons. Given the structure of the ROE program in which
apply) to 3 (certainly applies). Items for each subscale each of the nine themes is broken down into three com-
were averaged with higher scores signifying higher levels ponents (pre-family visit, baby visit, postfamily visit), we
of each construct. Supportive evidence for the reliability also examined the percentage of lesson content presented
and validity of the scales assessing proactive, reactive, and for each of the three lesson types. Our findings revealed
relational aggression has been obtained (e.g., Crick, 1996; that, on average, ROE instructors implemented 93% of
Dodge & Coie, 1987). For the present investigation, reli- lesson content across the nine pre-family visit lessons
ability, assessed via Cronbach’s alpha, was found to be (range from 85% to 100%), 95% of the content for the nine
adequate at both pretest and posttest: Aggressive Behavior, baby-visit lessons (range from 77% to 100%), and 91% of
a = .91 and .86, respectively; Prosocial Behavior, a = .90 content for the 8 postfamily visit lessons (range from 79%
and .91, respectively; Proactive Aggression, a = .77 and to 100%). Thus, it appears that the implementation fidelity,
.76, respectively; Reactive Aggression, a = .92 and .93, assessed via ROE instructors’ self-report on number of
respectively; and Relational Aggression, a = .90 and .87, lessons completed and in the percentage of lesson content
respectively. presented, was high across instructors and classrooms.

123
10 School Mental Health (2012) 4:1–21

Because of the high level of implementation adherence, no order to control for potential confounds. In order to provide
interaction effects were tested with these as potential information about the magnitude of program effects beyond
moderators. statistical significance, we calculated effect sizes (Cohen’s
ROE instructors also rated students’ level of engagement d) for each of our analyses. According to the criteria pro-
for each of the ROE lessons on a 5-point scale ranging from posed by Cohen (1988), an effect size of .2 is considered
1 (‘‘not engaged’’) to 5 (‘‘very engaged’’). Because we were small, .5 is considered to be a medium effect, and .8 is
interested in determining students’ level of engagement for considered a large effect.2 To control for family wise error,
lessons when the baby was present and when the baby was we set the alpha level at .01 for all tests of significance.
absent, we calculated separate means for each of these lesson
types. Average level of student engagement for lessons in Understanding of Infant Crying
which the baby was present ranged from 4.72 to 5.00
(M = 4.93, SD = .10), and the average level of engagement Table 1 provides a summary of the outcome analyses on
for lessons in which the baby was absent ranged from 3.72 to children’s understanding of the causes for infant crying.
4.95 (M = 4.42, SD = .41). The effect of condition on difference scores for internal
We next tested for differences between ROE and control (i.e., emotion state, physical state) and external causes was
teachers on their training in SEL, beliefs about SE, and tested using a series of 2 (group) 9 2 (gender) ANCOVAs,
frequency with which they implemented SEL activities in with age and ESL status as covariates. As can be seen, a
their classrooms throughout the school year (other than the significant group effect was found for one of the three
ROE program). With regard to training in SEL and SEL categories: ROE children showed significant improvements
beliefs, teachers in both groups were equivalent in terms of in their ability to spontaneously generate causes for infants’
having received coursework or workshops in SEL (71% crying that reflected attention to the emotions of the baby
program and 70% control), and all (100%) teachers rated (i.e., making reference to the infant’s feelings such as ‘‘the
SEL as ‘‘important’’ or ‘‘very important.’’ An independent- baby is crying because he is upset’’ or ‘‘misses his
samples t test revealed that, on a scale ranging from 1 (not mommy’’) in contrast to children in control classrooms.
at all) to 5 (a lot), control teachers implemented signifi- The effect for this finding was d = .26, or a small effect.
cantly more SEL activities in their classrooms (M = 3.17, Neither interaction effects nor other main effects emerged
SD = .94) than did ROE teachers (M = 2.21, SD = 1.05), for the other causes dimensions for infant crying.
t(24) = -2.42, p = .023. We next conducted a series of 2 (group) 9 2 (gender)
ANCOVAs, with age and ESL status as covariates to
examine intervention effects on the three dimensions of
Main Analyses children’s strategy knowledge regarding infant crying.
Analyses revealed a significant group 9 gender interaction
Analytic Strategy for providing social/relational support dimension (e.g.,
‘‘put the baby in his/her mother’s arms or lap’’). Girls in
The effect of condition on child outcomes was tested using ROE classrooms, in contrast to girls in control classrooms,
either analysis of covariance (ANOVA) or multiple analysis increased significantly in the number of times they reported
of variance (MANCOVA), where appropriate. A 2 (group: that providing social/relational support is a viable strategy
ROE program, control condition) 9 2 (gender) model was to utilize to help a crying infant (Mean difference scores
utilized and included changes from pretest to posttest (i.e., .16 and -.14, respectively), F(3, 276) = 5.44, p \ .02.
difference scores) as the dependent variable (Zumbo, 1999; The effect size for this interaction is d = .28, or a small
for a recent discussion regarding the utility of difference effect. There was no significant difference between boys in
scores, see Thomas & Zumbo, 2011). Statistically compa-
rable to performing a repeated-measures analysis, difference 2
The use of multilevel modeling (MLM)—analyzing data at the
or ‘‘change’’ scores (posttest–pretest) provide an unbiased level of the classroom—was inappropriate due to the small classroom
estimate of true change regardless of baseline value. Dif- sample size in this study. As recommended by Raudenbusch (1997), a
ference scores are seen as an alternative to ANCOVA (in sample size of at least 40 clusters (e.g., classrooms, schools) needs to
be given in order to achieve satisfactory statistical power in
which pretest scores are statistically controlled) when the
classroom-based program evaluation using MLM. Hence, analyses
researcher is interested in examining the direction of change were conducted at the level of the individual (i.e., child), rather than
and determining whether improvement in scores from pre- at the classroom level. Although methodological research has
test to posttest is greater in the treatment group than it is indicated that significance levels resulting from individual level
analyses where a program was implemented at the level of the
for the control group, as was the case in the present study
classroom may be overstated (e.g., Donner & Klar, 2000), it has also
(Tabachnick & Fidell, 2001). In all analyses, first language supported the notion that effect sizes remain unbiased (Raudenbusch
learned (ESL status) and age were included as covariates in & Bryk, 2002).

123
School Mental Health (2012) 4:1–21 11

Table 1 Casual explanations for infant crying: adjusted means at pretest and posttest for ROE program and control groups and adjusted
difference scores, with age and ESL status as covariates
Casual explanation Intervention (n = 306) Control (n = 279) Group difference in change
Pretest Posttest Difference Pretest Posttest Difference F value Effect size
score score Cohen’s d

Internal causes
Emotion state .52 (.81) .82 (1.13) .30 (1.22) .45 (.72) .47 (.72) .03 (.85) 9.82** .26
Physical state 2.44 (1.53) 2.84 (1.53) .40 (1.56) 2.12 (1.50) 2.43 (1.58) .32 (1.58) .36ns .05
External causes .15 (.46) .12 (.35) -.03 (.59) .22 (.51) .20 (.50) -.02 (.59) .25ns .01
Numbers in parentheses are standard deviations
** p \ .01

Table 2 Strategy knowledge for infant crying: adjusted means at pretest and posttest for ROE program and comparison groups and adjusted
difference scores, with age and ESL status as covariates
Strategy knowledge Intervention (n = 306) Control (n = 279) Group difference in change
Pretest Posttest Difference Pretest Posttest Difference F value Effect size
score score Cohen’s d

Physical strategies 1.34 (1.10) 1.54 (1.09) .20 (1.22) 1.33 (1.10) 1.49 (1.14) .16 (1.23) .00ns .03
External comfort object .17 (.43) .07 (.28) -.10 (.49) .11 (.32) .08 (.27) -.03 (.34) 4.37** -.16
Social/relational support .72 (.81) .79 (.82) .07 (1.02) .64 (.89) .59 (.75) -.05 (.95) 2.15ns .11
Numbers in parentheses are standard deviations
** p \ .01

Table 3 Empathy and perspective taking: adjusted means at pretest and posttest for ROE program and comparison groups and adjusted
difference scores, with age and ESL status as covariates
Variable Intervention (n = 306) Control (n = 279) Group difference in change
Pretest Posttest Difference Pretest Posttest Difference F value Effect size
score score Cohen’s d

Empathy 3.69 (.79) 3.62 (.81) -.08 (.75) 3.54 (.72) 3.52 (.82) -.02 (.65) 1.07ns .08
Perspective taking 3.25 (.72) 3.25 (.78) .00 (.70) 3.08 (.75) 3.14 (.76) .06 (.66) 1.18ns .08
Numbers in parentheses are standard deviations

ROE classrooms and boys in control classrooms on this Empathy and Perspective Taking
dimension (Mean differences scores -.02 and .03,
respectively). No other interactions emerged for the other To test for intervention effects on empathy and perspective
two strategy knowledge dimensions. However, as can be taking, we conducted a series of 2 (group) 9 2 (gender)
seen in Table 2, the ANCOVA analyses revealed a sig- ANCOVAs for empathy and for perspective taking,
nificant main effect for group assignment for one of the respectively, with age and ESL status as covariates. As can
three strategies for helping a crying infant: in contrast to be seen in Table 3, there were no significant intervention
children in control classrooms, ROE program children effects for either empathy or perspective taking. For both
significantly decreased from pretest to posttest in the fre- of these outcome variables, the analyses of potential
quency with which they mentioned providing external interactions were non-significant.
comfort object/environment strategies. This indicates that
ROE children were less likely than control children to Peer-Nominated Prosocial and Aggressive/Antisocial
suggest that providing an external comfort object/envi- Behaviors
ronment (e.g., ‘‘give the baby a pacifier,’’ ‘‘put the baby in
a swing’’) as a strategy to help a baby who is crying. The We next examined changes from pretest to posttest on
effect for this finding was small. dimensions of peer-nominated prosocial and aggressive/

123
12 School Mental Health (2012) 4:1–21

antisocial behaviors. A MANCOVA analysis revealed a aggression was -.53 (a medium effect), and the effect size
significant multivariate main effect for intervention status, for relational aggression was -.36 (a small–medium
F(8, 570) = 12.29, p \ .001. As can be seen in Table 4, effect). None of the remaining effects were significant.
follow-up univariate analyses revealed that ROE children, Because we noted that the difference scores for proac-
relative to control children, increased significantly from tive and relational aggression were in opposite directions
pretest to posttest on all of the six dimensions of peer- for the ROE and control groups, reflecting a decrease in
nominated prosocial behaviors. In other words, children in these aggressive behaviors among ROE children and an
ROE classrooms were more likely than children in control increase in these aggressive behaviors among control
classrooms to increase in the frequency with which they children, we conducted paired t tests to determine whether
were nominated by their peers as being prosocial with the within-group change was significant from pretest to
respect to ‘‘shares,’’ ‘‘cooperates,’’ ‘‘helps other kids when posttest for proactive and relational aggression for the ROE
they have a problem,’’ ‘‘kind,’’ ‘‘understands other kids’ program children and control children, separately. Analyses
point of view,’’ and ‘‘fair.’’ The average effect size across revealed that, whereas children in the ROE program group
all six dimensions was .48, or a moderate effect. Effect evidenced significant decreases in their proactive and
sizes ranged from small (i.e., .28 for the dimension relational aggression from pretest to posttest, (t(296) =
‘‘cooperates) to large (i.e., .79 for the dimension of ‘‘fair’’). 5.24 and t(296) = 2.88, respectively, ps \ .01, d = .30),
No significant gender-by-intervention interactions emerged control children evidenced significant increases in their
for prosocial behavior, and there were no significant main proactive and relational aggression, (t(274) = -3.83 and
effects or interaction effects for peer-nominated aggres- t(274) = -2.47, respectively, ps \ .01, d = .19).
sive/antisocial behaviors.

Teacher-Rated Prosocial and Aggressive Behaviors Discussion

For teacher-reported outcomes, a 2 (group) by 2 (gender) This study represents one of the first systematic attempts to
MANCOVA was conducted to determine intervention examine the effectiveness of the ROE program—an SEL
effects on teachers’ ratings of children’s prosocial and program in which a caregiver and an infant serve as a
aggressive behaviors (i.e., physical, reactive, proactive, and springboard for lessons about emotions, caring for others,
relational aggression). A significant main effect for inter- and infant development. Using a rigorous quasi-experi-
vention status was found, F(5, 566) = 12.53, p \ .001. As mental design, we found that children in ROE program
shown in Table 5, follow-up univariate analyses revealed classrooms, in contrast to children in control classrooms,
that ROE program children, relative to control children, improved in their understanding of infant crying; signifi-
decreased significantly in teacher-rated proactive and cantly increased in prosocial behaviors as obtained via peer
relational aggression. The effect size for proactive reports; and significantly decreased in proactive and

Table 4 Peer behavioral assessment: adjusted means at pretest and posttest for ROE program and control groups and adjusted difference scores,
with age, gender, and ESL status as covariates
Intervention (n = 306) Control (n = 279) Group difference in change
Pretest Posttest Difference Pretest Posttest Difference F value Effect size
score score Cohen’s d

Prosocial behavior
Shares .43 (.19) .54 (.18) .11 (.15) .42 (.19) .47 (.20) .05 (.14) 20.87*** .41
Cooperates .52 (.22) .59 (.24) .07 (.15) .49 (.23) .52 (.25) .03 (.13) 7.22** .28
Helps .38 (.19) .44 (.21) .06 (.15) .37 (.21) .38 (.20) .01 (.14) 15.35*** .34
Kind .53 (.20) .59 (.21) .05 (.14) .51 (.22) .48 (.23) -.03 (.13) 40.23*** .59
Takes others’ view .35 (.18) .40 (.18) .04 (.13) .34 (.19) .32 (.19) -.02 (.13) 29.62*** .46
Fair .48 (.19) .53 (.19) .05 (.14) .49 (.22) .43 (.20) -.06 (.14) 68.75*** .79
Antisocial behavior
Starts fights .13 (.19) .16 (.22) .03 (.13) .15 (.19) .19 (.21) .03 (.11) .32ns .00
Breaks rules .14 (.20) .19 (.24) .05 (.13) .19 (.20) .23 (.26) .04 (.11) 5.03ns .08
Numbers in parentheses are standard deviations
** p \ .01; *** p \ .001

123
School Mental Health (2012) 4:1–21 13

Table 5 Teacher reports of child behaviors: adjusted means at pretest and posttest for ROE program and control groups and adjusted difference
scores, with gender, age, and ESL status as covariates
Intervention (n = 297) Control (n = 275) Group difference in change
Pretest Posttest Difference Pretest Posttest Difference F value Effect size
score score Cohen’s d

Aggressive with peers 1.22 (.34) 1.25 (.37) .03 (.32) 1.21 (.36) 1.24 (.38) .03 (.30) .06ns .00
Prosocial with peers 2.45 (.49) 2.46 (.46) .01 (.49) 2.35 (.58) 2.40 (.60) .05 (.43) 1.85ns .00
Reactive aggression 1.20 (.37) 1.22 (.37) .02 (.32) 1.20 (.34) 1.23 (.38) .03 (.25) .09ns .00
Proactive aggression 1.25 (.29) 1.17 (.25) -.08 (.26) 1.15 (.24) 1.20 (.28) .05 (.23) 39.93** -.53
Relational aggression 1.17 (.35) 1.11 (.31) -.06 (.34) 1.10 (.27) 1.16 (.35) .06 (.32) 14.77** -.36
Numbers in parentheses are standard deviations
** p \ .001

relational aggression as obtained through teacher reports. trials are low in external validity and tell us little about the
Additionally, a review of ROE instructors’ diaries indi- program’s effectiveness under typical conditions of com-
cated that the program was implemented as intended (i.e., munity ownership. In their review, Wilson et al. empha-
adherence), with instructors reporting that students were sized the relative dearth of evaluations of routine practice
exposed to all or most of the components of the ROE programs and concluded that there was insufficient evi-
curriculum. dence about the effectiveness of routine practice programs
Our findings both replicate and extend what is known to offer any firm conclusions about their efficacy in
about the effectiveness of school-based prevention pro- reducing aggression. Hence, the findings of the present
grams in several specific ways. First, one of the strengths of study make an important addition to a small, but growing
the present study is that it provides support for the effec- corpus of evidence for routine practice programs.
tiveness of an SEL program implemented under ‘‘real- Second, because in evaluating our results, we not only
world’’ conditions in classrooms that reflect the diversity of utilized the traditional benchmark tests of statistical sig-
many large urban school districts. Our study, although not nificance to examine the effects of the ROE program on a
a randomized controlled trial, is high in ecological validity range of child outcomes but also calculated effect sizes
because it is relevant for contexts where teachers are (Wilkinson and the Task Force on Statistical Inference,
expected to address the needs of an increasingly diverse 1999), the findings add to what is known about the practical
group of typical children in regular classrooms. Research effects of a classroom-based SEL program. Following
has indicated that within one school district, school-age Cohen’s (1988) criteria, our intervention effects on behav-
populations vary considerably in terms of their family iors revealed an average effect size of .48 (range: .28 to .79)
experiences, cultural experiences, and socioeconomic sta- for peer-nominated prosocial behaviors and an average
tus. Determining the effectiveness of a social-emotional effect size of .45 for teacher-rated proactive and relational
competence promotion program within such a diverse aggression. Applying Cohen’s (1988) criteria for small,
routine practice condition provides high external validity to medium, and large effects, our average effect sizes would
the ROE program in these conditions. Wilson et al. (2003), be considered to be in the medium or moderate range. These
in their meta-analysis of the effects of school-based inter- effect sizes are in accord those found in other meta-analyses
ventions on aggressive behavior, cogently argue for a dif- of school-based violence prevention programs (e.g.,
ferentiation between evaluations of demonstration Wilson, Gottfredson, & Najaka, 2001; Wilson et al., 2003)
programs and routine practice programs. According to and are slightly higher than those found in a recent meta-
them, demonstration programs are those set up by analysis conducted by Durlak et al. (2011) of 213 school-
researchers whose primary mandate is to determine pro- based universal social and emotional learning programs
gram efficacy under controlled conditions. Such programs involving 270,034 students. Our moderate effect sizes may
are generally implemented by the researchers and are often be due, in part, to the high fidelity of program implemen-
delivered by research staff and/or teachers who are closely tation that we found (as discussed below). Indeed, as noted
supervised by the researchers. In contrast, routine practice by Greenberg (2010) ‘‘quality of implementation has a
programs, such as the one presented herein, are those significant impact on outcome, with higher quality imple-
programs that are already being implemented in schools on mentation related to a greater effect size’’ (p. 31).
a routine, ongoing basis. Although evaluations of demon- Third, evidence from this investigation increases both
stration programs have high internal validity, such efficacy researchers’ and educators’ understanding of program

123
14 School Mental Health (2012) 4:1–21

implementation by showing that a relatively easy-to-use accommodate the parents’ schedule, hence reducing the
and cost-effective SEL program that incorporates an infant likelihood for the parent and his/her infant to miss a lesson.
and his/her caregiver can be implemented as intended, or in This finding is provocative because it raises questions
other words, with high implementation integrity. Specifi- about the availability of a parent and infant to participate in
cally, our analyses of ROE instructors’ diaries indicated the ROE program. As described earlier, a parent and his/
that 96% of ROE lessons were implemented as intended, her infant play a central role in the ROE program lesson
with an average of 95% of lesson plan content being content and delivery. Indeed, the entire delivery of the
delivered. This rate of implementation fidelity is higher ROE program is dependent on this dimension of the pro-
than that typically reported in other evaluations of class- gram. The finding in the present study was that program
room-based SEL program. As noted by Durlak and DuPre implementation was high for the parent/infant visit lesson
(2008) in their meta-analysis examining the influence of content, with an average of 95% of content being covered
implementation on program outcomes, ‘‘Expecting perfect for each of the parent/infant visits. Perhaps this high degree
or near-perfect implementation is unrealistic. Positive of implementation fidelity for the parent/infant visits is
results have often been obtained with levels around 60%; attributable, in part, to the larger cultural and political
few studies have attained levels greater than 80%. No study context in which the program was implemented. Specifi-
has documented 100% implementation for all providers.’’ cally, perhaps high implementation of the parent/infant
(p. 331). portion of the ROE curriculum was due, in part, to Can-
What can account for the high level of implementation ada’s parental leave policy. In 2000, Canada greatly
integrity as reported by the ROE instructors? Perhaps the expanded its parental leave from 10 to 35 weeks that can
successful implementation of the ROE program was due, in be divided as desired between two parents. This time is in
part, to the implementation supports that were provided to addition to 15-week paid maternity leave. Moreover, in
the ROE instructors. Dane and Schneider (1998), in their most situations, maternity leave can be combined with
review of studies of school-based interventions, posit that parental leave to yield a maximum of 50 weeks (note also
one important dimension of implementation integrity is that parental leave and maternity leave benefits are paid by
program ‘‘promotion.’’ For them, program promotion the Canadian Employment Insurance system; http://www.
includes uses of a program manual, formal training servicecanada.gc.ca/eng/ei/types/special.shtml). Hence, it
opportunities, and ongoing consultation and support for may be that because many parents are at home with their
program providers. According to these criteria, the ROE infant for almost an entire year, many of them are willing
program would be considered to be high on this dimension and available to participate in the ROE program. Future
of implementation. Specifically, the ROE curriculum is research examining the implementation of the ROE pro-
manualized. Moreover, each of the ROE instructors gram in countries other than Canada should examine fur-
received an intensive 3-day training along with a 1-day ther this relation between implementation fidelity and
booster session mid-way through the school year. More- parental leave policies.
over, each ROE instructor was provided with an expert Fourth, because proactive, reactive, and relational
ROE mentor or ‘‘coach’’ to provide support and feedback aggression were examined prospectively across a school
to the instructor, as well as helping trouble-shoot special year in both ROE and control classrooms, the findings from
situations and facilitating instructors’ learning. Joyce and the current study add to what is known about stability and
Showers (2002) have emphasized the importance of cou- change in these forms of aggression during the elementary
pling training with coaching. They found that when train- school years and how a classroom-based SEL program can
ing was combined with coaching, 95% of teachers acquired interrupt the ‘‘normal’’ trajectory. Previous research has
knowledge and developed skills for applying that knowl- demonstrated a linear increase in aggression over the
edge in the classroom. In cases in which coaching was course of the elementary school years (Aber, Brown, &
absent, a small percentage (i.e., 5%) of teachers applied the Jones, 2003; Fonagy et al., 2009). What our data have to
skills in the classroom. say about aggression in schools and the role of preventive
Or perhaps high program fidelity was maintained interventions in interrupting a trajectory characterized by
because of the inclusion of an infant and his/her caregiver increased aggression is somewhat unique to what has been
into the curriculum. That is, because a commitment was learned previously in several respects.
made by the ROE parent to participate in the program and One of these advances was attained by examining
attend the monthly ROE lessons, all of the ROE lessons change over time rather than simply conducting analyses
were completed and thus the program was implemented as comparing posttest differences between intervention and
intended. High fidelity may also have been maintained control children. Examining change from pretest to post-
because the ROE instructor regularly contacted the ROE test, our findings revealed increases in proactive aggression
parent and scheduled the parent/infant classroom visits to among control-group children and decreases in these forms

123
School Mental Health (2012) 4:1–21 15

of aggression among ROE program children. The effect to reduce aggression, they can also exacerbate aggression
size for this difference was in the moderate range, d = .53. among peer group members when children feel that their
This finding is noteworthy given the insidious and dis- position in the social system is vulnerable and their roles
turbing nature of these forms of aggression. Indeed, rela- and responsibilities in the group are threatened (Farmer &
tive to children in the ROE program group, it may not be Xie, 2007). More specifically, when social hierarchies are
too surprising that proactive aggression increased during seen as uncertain, as may be the case among children
the school year among children in the control since it is an whose self-esteem and identity are in flux due to the bio-
instrumental form of aggression that is more likely to be logical, cognitive, social, and emotional changes that are
acquired and reinforced than other forms of aggression occurring as they make transition from childhood to ado-
(Bandura, 1983). lescence, the expression of aggression may become more
Our results indicated no intervention effects for reactive endemic to the peer culture and occur more frequently
aggression. As noted by Poulin and Boivin (2000), ‘‘Dis- (Adler & Adler, 1998; Pelligrini & Long, 2002). It is likely
tinguishing reactive aggressive behaviors from proactive that as children vie for top positions in the peer hierarchy,
aggressive behaviors at the morphological level is an even ‘‘ordinary’’ children may rely on relational aggression
important first step in understanding the nature and func- to either maintain or increase their status in the peer group.
tional value of both behaviors’’ (p. 115). Indeed, while a Hence, the mechanism that may underlie the increased
number of previous empirical investigations exist lending aggression in the control group classrooms may be the
support to the distinction of these two dimensions of putative behavioral and psychological stressors that ensue
aggression in both laboratory and school settings (Boivin, as children seek to maintain or improve their status in the
Dodge, & Coie, 1995; Dodge & Coie, 1987), the present peer group.
investigation is one of the first to provide evidence of the Our findings suggest that the ROE program can serve as
importance of distinguishing between reactive and proac- a catalyst to redirect the downward spiral toward increased
tive aggression in an evaluation of primary preventive aggression that evolves during a typical elementary school
efforts. year. The findings from our research examining the
A similar, but less robust pattern (d = .36) was found effectiveness of the ROE program are not unique however.
for relational aggression. Specifically, we found increases Grossman et al. (1997), in their evaluation of the school-
in relational aggression among control children and based social and emotional promotion program Second
decreases in this form of aggression among ROE program Step, found a pattern similar to the one in the current
children. This increase in relational aggression among study. That is, in their study of 790 2nd- and 3rd-graders,
children in the control group mirrors those findings of they found that those children who received the Second
recent investigations examining the developmental trajec- Step program decreased in physical aggression over the
tories of relational aggression during the elementary school course of the school year, whereas the children who did not
years. Murray-Close et al. (2007), for example, examined receive the curriculum showed increases in both verbal and
trajectories of relational aggression and found a rise in physically violent behaviors during the same time period.
relational aggression over the course of one calendar Our research findings extend those of Grossman et al. by
among a large, diverse sample of fourth-grade girls. Sim- showing that a classroom-based social and emotional
ilarly, Tremblay (2000) and others (e.g., Cote, Vaillan- competence promotion program can also influence other
court, Farhat, LeBlanc, Nagin, & Tremblay, 2006) have forms of aggression.
noted that although physical aggression decreases from the Fifth, the results of this investigation extend knowledge
time children begin school until the end of high school, about trajectories of prosocial behavior in school-aged
the middle childhood years are characterized as a time in children and demonstrate that a classroom-based SEL pro-
the life span in which significant linear increases in rela- gram can lead to improvements in children’s sharing,
tional and other forms of indirect aggression occur. helping, and cooperative behaviors. Similar to what we
Our findings indicating a rise in relational aggression in found for teacher-rated relational and proactive aggression,
control classrooms may be due, in part, to the nature of our data indicated that peer-nominated prosocial behaviors
classroom dynamics and group formation in the elementary worsened (i.e., decreased) among children in control
school years. As noted by Farmer and Xie (2007), ‘‘When classrooms over the course of the school year and improved
same-age children or youth are aggregated together they among children who received the ROE intervention.
tend to selectively and systematically synchronize their These findings challenge a commonly held belief in
behavior with each other…. This leads to the formation of developmental psychology—that prosocial behaviors,
(1) differential preferences for peer interactions, (2) distinct including sharing, helping, comforting, and cooperation
social groups, and (3) hierarchical social structures’’ increase in frequency as children grow older. Nantel-
(p. 462). Although social hierarchies can function as a way Vivier et al. (2009), for instance, in their longitudinal,

123
16 School Mental Health (2012) 4:1–21

multi-informant study examining trajectories of children’s in evaluations of SEL programs and suggest that a range of
prosocial behaviors in Canada and Italy, found that, con- informants, including peers and teachers, are in a position
trary to commonly held beliefs, prosocial behavior to supply common and unique information about the
remained stable or declined among children between the behavioral outcomes of children and, together, produce a
ages of 10 and 15. better understanding of program impacts.
The increases in prosocial behavior among children in Finally, our data provide important information about
the ROE program relative to controls may reflect the ROE the development of children’s perceptions and comforting
program emphasis. That is, because ROE lessons focus on strategies to infant crying. We found that children who
caring for the baby and there are multiple opportunities received the ROE curriculum, in contrast to control chil-
provided throughout the curriculum for children to engage dren, significantly increased in their understanding of
in activities that benefit the baby (e.g., creating a CD of infant crying in two important ways. First, ROE children
lullabies for the baby, creating a book for the baby that showed significant improvements in their ability to spon-
documents his/her month by month growth and develop- taneously generate causes for infants’ crying that reflected
mental milestones), the ROE program creates a classroom attention to the emotions of the baby (i.e., making reference
environment in which prosocial behaviors become the to the infant’s feelings such as ‘‘the baby is crying because
norm. As noted by Staub (1988): ‘‘Acting to benefit others he is upset’’ or ‘‘misses his mommy’’). Second, with regard
can result in personal changes that lead to more significant to strategies for helping a crying baby, we found a signif-
helping. It can promote a more positive evaluation of the icant group-by-gender interaction for the dimension that
welfare of people who have benefited from one’s actions focused on providing social/relational support dimension
and of people in general and a perception of oneself as a (e.g., ‘‘put the baby in his/her mother’s arms or lap’’). Girls
caring person willing to extend effort and make sacrifices in ROE classrooms, in contrast to boys, increased signifi-
for others’ sake’’ (p. 95). Moreover, perhaps teaching cantly in the number of times they reported that providing
emotional understanding skills in conjunction with group social/relational support was a viable strategy for helping a
activities that benefit others, such as helping the baby in the crying infant. Third, ROE program children significantly
ROE program, can lead to increases in children’s prosocial decreased from pretest to posttest in the frequency with
behaviors with one another. which they mentioned providing external comfort object/
As with other studies with multiple reporters assessing environment strategies. Collectively, these findings extend
comparable constructs, we found differential outcomes by previous work on children’s understanding of infant dis-
reporter for similar behaviors (Achenbach, 2006). As tress and prosocial responding.
Achenbach has noted, when diverse perspectives are nee- For decades now, researchers have studied infant crying
ded to illustrate a construct, as in the case of children’s and its many functions. Responding sensitively to infant
aggressive and prosocial behaviors, it is important to take crying is central to optimal infant development and secure
into consideration which set of windows are most likely to attachment (e.g., Ainsworth, Blehar, Waters, & Wall, 1978;
maximize one’s view of the behavior. When teachers, not Fish, Stifter, & Belsky, 1991). Until now, however, very
peers, reported on children’s aggressive behaviors, a little research has examined the manner in which under-
positive intervention impact was found for proactive and standing of and responsiveness to infant crying emerge
relational aggression. In contrast, when peer nominations during the childhood years and how a classroom-based
were used to assess children’s social behaviors, we found SEL intervention can advance children’s understanding of
positive intervention effects for prosocial behaviors, infant distress and their responsiveness to it. Murray (1979)
including sharing, helpfulness, cooperation, kindness, per- proposed an ethological framework where cries were
spective taking, and being seen as fair. The discrepancy by viewed as elicitors of emotions that promote prosocial
informer may reflect the different contexts in which the behaviors such as helping, comforting, and protecting the
behavior is being observed as well as the differential sal- infant (Hoffman, 1978). Researchers examining children’s
ience of the behavior for the informant. In the case of understanding and responses to infant crying have shown
teachers, their observations of children’s social behaviors that caregiving responses to infant crying increase with age
may be restricted primarily to the classroom, and teachers, (e.g., Zahn-Waxler, Friedman, & Cummings, 1983) and
in contrast to peers, may perceive aggressive behaviors as that such increases are due, in part, to increased sophisti-
more salient than other types of social behaviors. In con- cation in understanding others’ needs. That is, one mech-
trast, children may display different prosocial behaviors in anism underlying developmental changes in the nature and
their peer groups, and these behaviors may take place type of responses generated may reflect advanced social-
across a range of contexts, including the playground and cognitive abilities. A possible explanation for the greater
the lunchroom. Taken together, our findings provide strong frequency with which ROE program children identified
evidence for the utility of collecting multi-informant data emotions as a cause of infant crying may be a result of the

123
School Mental Health (2012) 4:1–21 17

program focus on the emotional life of infants and that dimensions of the ROE curriculum, the characteristics of the
these discussions occurred in situations in which a real ROE instructor/classroom teacher, or some combination of
infant was present (and sometimes crying). Such experi- the two. Nonetheless, there are several reasons for confi-
ences may have been central for advancing children’s dence in our findings. First, children were not assigned to
social-cognitive understanding of infant crying. It may be teachers on the basis of the ROE program, and thus, there
that such changes would not be evident had it not been for was no reason to expect that children who received the
the presence of the ROE infant. Further research is clearly program were different from children who did not receive it.
needed to shed further light on this supposition. We tested baseline differences between children in the ROE
The findings on empathy are puzzling. Contrary to program and control groups on demographic and outcome
expectations, there were no significant differences in self- measures at baseline and found no significant differences
reported empathy and perspective taking between ROE and between groups, suggesting that our matching process was
control children. Because the bases for our null findings are successful. Second, the pattern of change across several of
indeterminant, we have no way of knowing whether the our dimensions of social behaviors are patterns that Shadish,
failure to find significant improvements in empathy-related Cook, and Campbell (2002) have identified as tenable for
responding among children who received the ROE program asserting causal claims in quasi-experimental designs. For
was due to the investigation’s methodology, an invalid instance, our findings with regard to proactive aggression
theoretical premise, or to other unknown factors that we did indicated a different pattern of change from pretest to post-
not consider. However, the following observations are test for ROE program and comparison groups, respectively.
offered to assist investigators who may wish to examine the More specifically, whereas children in the ROE program
development of empathy in future investigations into decreased in proactive aggression, control children
classroom-based SEL program. First, it may be that because increased in proactive aggression. This pattern is consistent
the measure we utilized in our study to assess empathy and with a pattern in which other plausible explanations of
perspective taking (i.e., Davis’ Interpersonal Reactivity threats to internal validity are low. Finally, it is significant
Index) is a measure that has been used primarily to measure that the teachers who volunteered for ROE in their class-
empathy in adolescents and adults (for a review, see Zhou, rooms did not rate social and emotional factors as more
Valiente, & Eisenberg, 2003), it may not have been sensi- important than control teachers, suggesting that the teachers
tive enough in detecting developmental changes in chil- differed only according to the opportunity to have the pro-
dren’s empathy and perspective taking across middle gram and not according to key attitudinal factors. Future
childhood. Beyond this consideration, it is also plausible research on ROE could be extended by using an RCT design
that because there was such an explicit teaching of to determine the extent to which our positive findings can
‘‘empathy’’ in the ROE curriculum, the children in the ROE be replicated.
program became sensitized to the meaning of empathy and Secondly, because we relied solely on ROE instructors’
hence, at posttest, approached the items on the measure self-reports of implementation fidelity/adherence, we cannot
differently. Future research is clearly needed to compare the be certain that the ROE program was truly implemented as
measurement of empathy using different approaches. intended. That is, because we have no objective data on
implementation integrity, we have no way of determining
Limitations and Directions for Future Research whether the ROE was implemented at the high level con-
veyed via the ROE instructors’ monthly diaries. As noted by
The data reported here are encouraging of continued inves- Schulte, Easton, and Parker (2009), ‘‘Research has consis-
tigation into the effectiveness of the ROE program. None- tently shown that implementers overestimate their own level
theless, while evaluating the results, several limitations of treatment integrity (Dusenbury, Brannigan, Hansen,
should be noted. Firstly, one limitation derives from the Walsh, & Falco, 2005; Lillehoj, Griffin, & Spoth, 2004;
utilization of a quasi-experimental design rather than the Resnicow et al., 1998)’’ p. 469. Hence, some caution should
‘‘gold-standard’’ (see Biglan, Mrazek, Carnine, & Flay, 2003 be taken when interpreting our implementation findings. In
for a discussion about the utility of other designs in adding to future research on the ROE program, it is important to include
the evidence base) and not the recommended ‘‘gold-stan- multiple measures for monitoring implementation, including
dard’’ randomized clinical trial (RCT). That is, classrooms observer-report and direct observations, in addition to self-
were not randomly assigned to receive the ROE program. reports (see Leff, Hoffman, & Lakin Gullan, 2009). For
Instead, teachers volunteered to have the program imple- instance, one way to objectively monitor implementation
mented in their classroom. This makes causal inferences may be through videotaping lessons that can then be coded
between the ROE program and children’s social and emo- later by independent observers (Schulte et al., 2009).
tional competence difficult to make in the absence of further Thirdly, neither teachers nor peers were blind to
evidence. The observed effects may be due to specific experimental condition when providing their ratings of

123
18 School Mental Health (2012) 4:1–21

children’s behaviors. We readily acknowledge that teachers characteristics interact with classroom contexts to deter-
can be a critical source of information due to the fact that mine outcomes. Finally, because the ROE curriculum is
they have many opportunities to observe children for sev- multifaceted with many lessons and foci, it is difficult to
eral hours each day. Nonetheless, although our teacher- determine which aspects of the program are responsible for
report measure of child behaviors included filler items to the effects in social-emotional understanding and behav-
lessen response bias in our teachers’ reports, teachers were ioral change.
aware of intervention status. With regard to our peer Despite these limitations, the results of this study pro-
behavioral assessments, peers were also not blind to vide empirical evidence that a relatively easy-to-implement
treatment condition, and they would have easily been able universal classroom-based intervention that does not focus
to discern which condition their classroom had been exclusively on disturbed or atypical children substantially
assigned. Although peers as participant observers can improves children’s cooperative and kind behavior to one
provide important sources of information about their another and decreases their aggression. Such an evaluation
classmates’ behaviors both inside and outside of the not only provides support for the continued implementation
classroom, our peer behavioral assessment measure of and evaluation of preventive efforts that are grounded in
prosocial and aggressive/antisocial behaviors may have the scientific study of children’s development but also
been influenced by peers’ knowledge about experimental contributes to the ongoing scholarship at the intersection of
condition. We speculate here that peers’ ratings of class- translational research, developmental and prevention sci-
mates’ behaviors would be less likely than teachers to be ences, and educational reform.
influenced by knowledge of the intervention status given
that it is unlikely that children would be able to generate Acknowledgments This research was supported by grants from
the University of British Columbia Hampton Research Fund and the
specific hypothesis of the study. However, we have no data Human Early Learning Partnership (HELP) to the first author. The
to support such a claim, and future investigations into the authors wish to express their thanks to the numerous school staff and
ROE program would benefit from collecting data from administrators who helped make this project possible and especially
observers blind to intervention status in order to allow for a to the children and their teachers for their participation in this project.
more objective measure of children’s behaviors.
Thirdly, our analyses were conducted at the individual
child’s level even though selection was by classroom.
Unfortunately, the small number of classrooms did not References
provide sufficient statistical power to use multilevel mod-
eling. The clustering of children within classrooms results Aber, J. L., Brown, J. L., & Jones, S. M. (2003). Developmental
trajectories toward violence in middle childhood: Course,
in the non-independence of subjects, which could bias the demographic differences, and response to school-based inter-
statistical tests used to identify intervention effects. This is vention. Developmental Psychology, 39, 324–348.
a major challenge to evaluations of universal school-based Achenbach, T. M. (2006). As others see us: Clinical and research
interventions when insufficient resources exist to recruit implications of cross-informant correlations for psychopathol-
ogy. Psychological Science, 15, 94–98.
large numbers of classrooms or schools (Stoolmiller, Eddy, Adler, P. A., & Adler, P. (1998). Peer power: Preadolescent culture
& Reid, 2000). Nonetheless, as noted by Slavin (2008), and identity. New Brunswick, NJ: Rutgers University Press.
although analyzing data at the child level when randomi- Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978).
zation was done at the classroom level is discouraged by Patterns of attachment: A psychological study of the Strange
Situation. Hillsdale, NJ: Erlbaum.
methodologists (Donnor & Klar, 2000) because the find- Ang, R., & Hughes, J. (2001). Differential benefits of skills training
ings can exaggerate statistical significance, ‘‘… their effect with antisocial youth based on group composition: A meta-
sizes are unbiased (Raudenbush & Bryk, 2002) and there- analytic investigation. School Psychology Review, 31, 164–185.
fore are of value…’’ (p. 9). Arsenio, W. F., & Lemerise, E. A. (2001). Varieties of childhood
bullying: Values, emotion processes, and social competence.
A final limitation is that our analysis did not address Social Development, 10, 59–73.
questions regarding differential program effects for indi- Bandura, A. (1983). Psychological mechanisms of aggression. In
vidual children. Although we did control statistically for R. G. Green & E. I. Donnerstein (Eds.), Aggression: Theoretical
important child demographic characteristics, namely age and empirical views (Vol. 1, pp. 1–40). New York: Academic.
Barrett, K., & Campos, J. (1987). Perspectives on emotional
and ESL status, such analyses does not allow an explora- development II: A functionalist approach to emotions. In J.
tion of differential program effects for children of different Osofsky (Ed.), Handbook of infant development (2nd ed.,
ages and/or differing ethnic and cultural backgrounds. pp. 555–578). New York: Wiley.
Given the wide-ranging ethnic and cultural diversity of the Battistich, V., Solomon, D., Watson, M., & Schaps, E. (1997). Caring
school communities. Educational Psychologist, 32, 137–151.
children in our sample, including a larger number of par- Battistich, V., Solomon, D., Watson, M., Solomon, J., & Schaps, E.
ticipants might have allowed for a more specific exami- (1989). Effects of an elementary school program to
nation of the manner in which children’s background enhance prosocial behavior on children’s cognitive social

123
School Mental Health (2012) 4:1–21 19

problem-solving skills and strategies. Journal of Applied Devel- Crick, N. R., & Dodge, K. A. (1999). ‘Superiority’ is in the eye of the
opmental Psychology, 10, 147–169. beholder: A comment on Sutton, Smith, and Swettenham. Social
Biglan, A., Mrazek, P. J., Carnine, D., & Flay, B. R. (2003). The Development, 8, 128–131.
integration of research and practice in the prevention of youth Crick, N. R., & Grotpeter, J. K. (1995). Social aggression, gender, and
problem behaviors. American Psychologist, 58, 433–440. social psychological adjustment. Child Development, 66,
Boivin, M., Dodge, K. A., & Coie, J. D. (1995). Individual-group 710–722.
behavioral similarity and peer status in experimental play groups Damon, W. (2004). What is positive youth development? Annals of the
of boys: The social misfit revisited. Journal of Personality and American Academy of Political and Social Science, 591, 13–24.
Social Psychology, 69, 269–279. Damon, W., & Gregory, A. (2003). Bringing in a new era in the field
Bukowski, W. M., & Sippola, L. K. (1996). Friendship and morality. of youth development. In R. M. Lerner & P. L. Benson (Eds.),
In W. M. Bukowski & A. F. Newcomb (Eds.), The company they Developmental assets and asset-building communities: Implica-
keep: Friendship in childhood and adolescence (pp. 238–261). tions for research, policy, and practice (pp. 47–64). New York:
Cambridge: Cambridge University Press. Kluwer/Plenum.
Cairns, R. B., Cairns, B. D., Neckerman, H. J., Gest, S. D., & Gariépy, J. Dane, A. V., & Schneider, B. H. (1998). Program integrity in primary
L. (1988). Social networks and aggressive behavior: Peer support and early secondary prevention: Are implementation effects out
or peer rejection? Developmental Psychology, 24, 815–823. of control? Clinical Psychology Review, 18, 23–45.
Campos, J., Mumme, D., Kermoian, R., & Campos, R. (1994). A Davis, M. H. (1983). Measuring individual differences in empathy:
functionalist perspective on the nature of emotion. Monographs Evidence for a multidimensional approach. Journal of Person-
of the Society for Research in Child Development, 59(2–3), ality and Social Psychology, 44, 113–126.
284–303. Denham, S. A. (1998). Emotional development in young children.
Cappella, E., & Weinstein, R. (2006). The prevention of social New York: Guilford.
aggression among girls. Social Development, 15, 434–462. Denham, S. A., & Burton, R. (1996). A social-emotional intervention
Caprara, G. V., Barbanelli, C., Pastorelli, C., Bandura, A., & program for at-risk four-year-olds. Journal of School Psychol-
Zimbardo, P. G. (2000). Prosocial foundations of children’s ogy, 34, 224–245.
academic achievement. Psychological Science, 11, 302–306. Dodge, K. A., & Coie, J. D. (1987). Social-information processing factors
Card, N. A., Stucky, B. D., Sawalani, G. M., & Little, T. D. (2008). Direct in reactive and proactive aggression in children’s peer groups.
and indirect aggression during childhood and adolescence: A meta- Journal of Personality and Social Psychology, 53, 1146–1158.
analytic review of gender differences, intercorrelations, and rela- Domitrovich, C. E., & Greenberg, M. T. (2000). The study of
tions to maladjustment. Child Development, 79, 1185–1229. implementation: Current findings from effective programs that
Catherine, N., & Schonert-Reichl, K. A. (2011). Children’s perceptions prevent mental disorders in school-age children. Journal of
and comforting strategies to infant crying: Relations to age, sex, Educational and Psychological Consultation, 11, 193–221.
and empathy-related responding. British Journal of Developmen- Donner, A., & Klar, N. (2000). Design and analysis of cluster
tal Psychology, 29, 524–551. doi:10.1348/026151010X521475. randomized trials in health research. New York: Oxford
Cohen, J. (1988). Statistical power analysis (2nd ed.). Hillsdale, NJ: University Press.
Erlbaum. Durlak, J. A., & DuPre, E. P. (2008). Implementation matters: A
Cohen, J. (Ed.). (2001). Caring classrooms/intelligent schools: The review of research on the influence of implementation on
social emotional education of young children. New York: program outcomes and the factors affecting implementation.
Teachers College Press. American Journal of Community Psychology, 41, 327–350.
Coie, J. D., & Dodge, K. A. (1988). Multiple sources of data on social Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., &
behavior and social status in the school: A cross-age comparison. Schellinger, K. B. (2011). Enhancing students’ social and
Child Development, 59, 815–829. emotional development promotes success in school: Results of
Conduct Problems Prevention Research Group. (1999). Initial impact a meta-analysis. Child Development, 82, 474–501.
of the Fast Track Prevention Trial for conduct problems: II. Dusenbury, L., Brannigan, R., Hansen, W. B., Walsh, J., & Falco, M.
Classroom effects. Journal of Consulting and Clinical Psychol- (2005). Quality of implementation: Developing measures crucial
ogy, 67, 648–657. to understanding the diffusion of preventive interventions.
Consortium on the School-Based Promotion of Social Competence. Health Education Research, 20, 308–313.
(1996). The school-based promotion of social competence: Eisenberg, N., Fabes, R. A., & Spinrad, T. (2006). Prosocial
Theory, research, practice, and policy. In R. J. Haggerty, development. In N. Eisenberg (Ed.), Handbook of child
L. R. Sherrod, N. Garmezy, & M. Rutter (Eds.), Stress, risk, psychology: Social, emotional, and personality development
and resilience in children and adolescents: Processes, mecha- (pp. 646–718). Hoboken, NJ: Wiley.
nisms, and interventions (pp. 268–316). Cambridge, MA: Eisenberg, N., & Miller, P. A. (1987). The relation of empathy to
Cambridge University Press. prosocial and related behaviors. Psychological Bulletin, 101,
Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: 91–119.
Design and analysis issues in field settings. Chicago: Rand Emde, R. N., Osofsky, J. D., & Butterfield, P. M. (1993). The IFEEL
McNally. pictures: A new instrument for interpreting emotions. Madison,
Cook, T. D., Murphy, R. F., & Hunt, H. D. (2000). Comer’s School WI: International Universities Press.
Development Program in Chicago: A theory-based evaluation. Fabes, R., Eisenberg, N., Nyman, M., & Michelieu, Q. (1991).
American Educational Research Journal, 37, 535–597. Children’s appraisals of others’ spontaneous emotional reactions.
Cote, S., Vaillancourt, T., Farhat, A., LeBlanc, J. C., Nagin, D., & Developmental Psychology, 27, 858–866.
Tremblay, R. (2006). The development of physical aggression Farmer, T. W., & Xie, H. (2007). Aggression and school social
from toddlerhood to pre-adolescence: A nation wide longitudinal dynamics: The good, the bad, and the ordinary. Journal of
study of Canadian children. Journal of Abnormal Child Psychol- School Psychology, 45, 461–478.
ogy, 34, 71–85. Feshbach, N. D. (1979). Empathy training: A field study in affective
Crick, N. R. (1996). The role of overt aggression, relational education. In S. Feshbach & A. Fraczek (Eds.), Aggression and
aggression, and prosocial behavior in the prediction of children’s behaviors change: Biological and social processes (pp.
future social adjustment. Child Development, 67, 2317–2327. 234–249). New York: Praeger.

123
20 School Mental Health (2012) 4:1–21

Fish, M., Stifter, C. A., & Belsky, J. (1991). Conditions of continuity Ladd, G. W., & Troop-Gordon, W. (2003). The role of chronic peer
and discontinuity in infant negative emotionality: Newborn to difficulties in the development of children’s psychological
five months. Child Development, 62, 1525–1537. adjustment problems. Child Development, 74, 1344–1367.
Fonagy, P., Twemlow, S. W., Vernberg, E. M., Nelson, J. M., Dill, E. Lagerspetz, K. M. J., Bjorkqvist, K., & Peltonen, T. (1988). Is indirect
J., Little, T. D., et al. (2009). A cluster randomized controlled aggression typical of females? Gender differences in aggres-
trial of child-focused psychiatric consultation and a school siveness in 11 to 12 year old children. Aggressive Behavior, 14,
systems-focused intervention to reduce aggression. Journal of 403–414.
Child Psychology and Psychiatry, 50, 607–616. Leff, S. S., Hoffman, J. A., & Lakin Gullan, R. (2009). Intervention
Gordon, M. (2000). Roots of Empathy: Training manual. Available integrity: New paradigms and applications. School Mental
from the Roots of Empathy, www.Rootsofempathy.org Health, 1, 103–106.
Gordon, M. (2001). Roots of Empathy: Training manual. Available Lewis, M., & Michalson, L. (1983). Children’s emotions and moods.
from the Roots of Empathy, www.Rootsofempathy.org New York: Plenum Press.
Gordon, M. (2005). Roots of Empathy: Changing the world child by Lillehoj, C. J., Griffin, K. W., & Spoth, R. (2004). Program provider
child. Toronto, ON: Thomas Allen. and observer ratings of school-based preventive intervention
Greenberg, M. T. (2004). Current and future challenges in school- implementation: Agreement and relation to youth outcomes.
based prevention: The researcher perspective. Prevention Sci- Health Education & Behavior, 31, 242–257.
ence, 5, 5–13. Lösel, F., & Beelmann, A. (2003). Effects of child skills training in
Greenberg, M. T. (2010). School-based prevention: Current status and preventing antisocial behavior: A systematic review of random-
future challenges. Effective Education, 2, 27–52. ized evaluations. Annals of the American Academy of Political
Greenberg, M. T., Kusché, C. A., Cook, E. T., & Quamma, J. P. and Social Sciences, 857, 177–186.
(1995). Promoting emotional competence in school-aged chil- Masten, A. S., & Motti-Stefanidi, F. (2009). Understanding and
dren: The effects of the PATHS curriculum. Development and promoting resilience in children: Promotive and protective
Psychopathology, 7, 117–136. processes in schools. In B. Gutkin & C. R. Reynolds (Eds.),
Gresham, F. M., Cook, C. R., Crews, S. D., & Kern, L. (2004). Social The handbook of school psychology. New York: Wiley.
skills training for children and youth with emotional and Murray, A. D. (1979). Infant crying as an elicitor of parental behavior: An
behavioral disorders: Validity considerations and future direc- examination of two models. Psychological Bulletin, 86, 191–215.
tions. Behavioral Disorders, 30, 32–46. Murray-Close, D., Ostrov, J. M., & Crick, N. R. (2007). A short-term
Grossman, D. C., Neckerman, H. J., Koepsell, T. D., Liu, P. Y., longitudinal study of growth of relational aggression during
Asher, K. N., Beland, K., et al. (1997). The effectiveness of a middle childhood: Associations with gender, friendship inti-
violence prevention curriculum among children in elementary macy, and internalizing problems. Development and Psychopa-
school. Journal of the American Medical Association, 277, thology, 19, 187–203. doi:10.1017/S0954579407070101.
1605–1611. Nantel-Vivier, A., Kokko, K., Caprara, G. V., Pastorelli, C., Gerbino, M.
Hay, D. F. (1994). Prosocial development. Journal of Child G., Paciello, M., et al. (2009). Prosocial development from
Psychology and Psychiatry, 35, 29–71. childhood to adolescence: A multi-informant perspective with
Hoffman, M. L. (1978). Empathy, its development and prosocial Canadian and Italian longitudinal studies. Journal of Child
implications. In C. B. Keasy (Ed.), Nebraska symposium on Psychology and Psychiatry, 50, 590–598. doi:10.1111/j.1469-7610.
motivation, Vol. 25, pp. 169–218. 2008.02039.
Huebner, E., & Gilman, R. (2003). Toward a focus on positive Neal, J. W. (2010). Social aggression and social position in middle
psychology in school psychology. School Psychology Quarterly, childhood and early adolescence: Burning bridges or building
18, 99–102. them? Journal of Early Adolescence, 30, 122–137.
Hymel, S., Schonert-Reichl, K. A., & Miller, L. D. (2007). Reading, Noddings, N. (1992). The challenge to care in schools: An alternative
‘Riting, ‘Rithmetic and relationships: Considering the social side approach to education. New York: Teachers College Press.
of education. Exceptionality Education Canada, 16, 149–191. Oakley, A., Strange, V., Bonnell, C., Allen, E., & Stephenson, J.
Izard, C. (2002). Translating emotion theory and research into (2006). Health services research: Process evaluations in ran-
preventive interventions. Psychological Bulletin, 128, 796–824. domised control trials of complex interventions. British Medical
Izard, C. E., Fine, S., Mostow, A., Trentacosta, C., & Campbell, J. Journal, 332, 413–416.
(2002). Emotional processes in normal and abnormal develop- Oberle, E., Schonert-Reichl, K. A., & Thomson, K. (2010). Understand-
ment and preventive intervention. Developmental and Psycho- ing the link between social and emotional well-being and peer
pathology, 14, 761–787. relations in early adolescence: Gender-specific predictors of peer
Jones, S. M., Brown, J. L., & Aber, J. L. (2011). Two-year impacts of a acceptance. Journal of Youth and Adolescence, 39, 1330–1342.
universal school-based social-emotional and literacy intervention: Parkhurst, J., & Asher, S. R. (1992). Peer rejection in middle school:
An experiment in translational developmental research. Child Subgroup differences in behavior, loneliness, and interpersonal
Development, 82, 533–554. doi:10.1111/j.1467-8624.2010.01560. concerns. Developmental Psychology, 28, 231–241.
Joyce, B., & Showers, B. (2002). Student achievement through staff Pelligrini, A. D., & Long, J. D. (2002). A longitudinal study of
development (3rd ed.). Alexandria, VA: Association for Super- bullying, dominance, and victimization during the transition
vision and Curriculum Development. from primary school through secondary school. British Journal
Kendall, G., Schonert-Reichl, K., Smith, V., Jacoby, P., Austin, R., of Developmental Psychology, 20, 259–280.
Stanley, F., et al. (2006). The evaluation of the ‘‘Roots of Poulin, F., & Boivin, M. (2000). Reactive and proactive aggression:
Empathy’’ in Western Australia 2005. A report written for the Evidence of a two-factor model. Psychological Assessment, 12,
Department of Education and Training, Western Australia, Perth: 115–122. doi:10.1037/1040-3590.12.2.115.
Telethon Health Research Institute. Raudenbush, S. W. (1997). Statistical analysis and optimal design for
Ladd, G. W., & Profilet, S. M. (1996). The Child Behavior Scale: A cluster randomized trials. Psychological Methods, 2, 173–185.
teacher-report measure of young children’s aggressive, with- Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear
drawn, and prosocial behaviors. Developmental Psychology, 32, models: Applications and data analysis methods. Thousand
1008–1024. Oaks, CA: Sage.

123
School Mental Health (2012) 4:1–21 21

Resnicow, K., Davis, M., Smith, M., Lazarus-Yaroch, A., Baranow- children and adolescents: Toward a unified risk and intervention
ski, T., Baranowskii, J., et al. (1998). How best to measure framework. Special Section: Prediction and prevention of child
implementation of school health curricula: A comparison of and adolescent antisocial behavior. Journal of Consulting and
three measures. Health Education Research, 13, 239–250. Clinical Psychology, 63, 579–584.
Rubin, K. H., Coplan, R. J., Fox, N. A., & Calkins, S. D. (1995). Tregebov, R. (1993). Sasha and the wiggly tooth. Toronto, ON:
Emotionality, emotion regulation, and preschoolers’ social Second Story Press.
adaptation. Development and Psychopathology, 7, 49–62. Tremblay, R. E. (2000). The development of aggressive behavior
Saarni, C. (1999). The development of emotional competence. New during childhood: What have we learned in the past century?
York: Guilford. International Journal of Behavioral Development, 24, 129–141.
Schonert-Reichl, K. A. (1993). Empathy and social relationships in Vitaro, F., Brendgen, M., & Tremblay, R. E. (2002). Reactively and
adolescents with behavioral disorders. Behavioral Disorders, 18, proactively aggressive children: Antecedent and subsequent
189–204. characteristics. Journal of Child Psychology and Psychiatry, 43,
Schonert-Reichl, K. A. (1999). Moral reasoning during early adoles- 495–505.
cence: Links with peer acceptance, friendship, and social Weissberg, R. P., & Greenberg, M. T. (1998). School and community
behaviors. Journal of Early Adolescence, 19, 249–279. competence-enhancement and prevention programs. In I. E. Sie-
Schonert-Reichl, K. A., & Oberle, E. (2011). Teaching empathy to gel & K. A. Renninger (Eds.), Handbook of child psychology:
children: Theoretical and empirical considerations and implica- Vol. 5. Child psychology in practice (5th ed., pp. 877–954). New
tions for practice. In B. Weber & E. Marsal (Eds.), ‘‘The politics York: Wiley.
of empathy’’ New interdisciplinary perspectives on an ancient Wentzel, K. R. (1993). Does being good make the grade? Social
phenomenon. Berlin, Germany: Lit Verlag. behavior and academic competence in middle school. Journal of
Schulte, A. C., Easton, J. E., & Parker, J. (2009). Advances in Educational Psychology, 85, 357–364.
treatment integrity research: Multidisciplinary perspectives on Wentzel, K. R., Filisetti, L., & Looney, L. (2007). Adolescent
the conceptualization, measurement, and enhancement of treat- prosocial behavior: The role of self-processes and contextual
ment integrity. School Psychology Review, 38, 460–475. cues. Child Development, 78, 895–910.
Seligman, M. E. P., & Csikszentmihalyi, M. (2000). Positive Wilkinson, L., & Task Force on Statistical Inference. (1999).
psychology. American Psychologist, 55, 5–14. Statistical methods in psychology journals: Guidelines and
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental explanations. American Psychologist, 54, 594–604.
and quasi-experimental designs for generalized causal inference. Wilson, D., Gottfredson, D., & Najaka, S. (2001). School based
Boston, MA: Houghton Mifflin. prevention of problem behaviors: A meta analysis. Journal of
Shipman, K., Zeman, J., Penza, S., & Champion, K. (2000). Emotion Quantitative Criminology, 17, 247–272.
management skills in sexually maltreated and nonmaltreated Wilson, S. J., Lipsey, M. W., & Derzon, J. H. (2003). The effects of
girls: A developmental psychopathology perspective. Develop- school-based intervention programs on aggressive behavior: A
ment and Psychopathology, 12, 47–62. meta-analysis. Journal of Consulting and Clinical Psychology,
Slavin, R. E. (2008). Perspectives on evidence-based research in 71, 136–149.
education. Educational Researcher, 37, 5–14. Yau, M. (1999). An evaluation of the Roots of Empathy program:
Staub, E. (1988). The evolution of caring and nonaggressive persons Some preliminary findings, 1997–99. Toronto District School
and societies. Journal of Social Issues, 44, 81–100. Board: Academic Accountability Department.
Staub, E. (2003). Notes on cultures of violence, cultures of caring and Younger, A. J., Schneider, B. H., Wadeson, R., Guirguis, M., &
peace, and the fulfillment of basic human needs. Political Bergeron, N. (2000). A behavior-based peer-nomination measure
Psychology, 24, 1–21. of social withdrawal in children. Social Development, 9,
Stoolmiller, M., Eddy, M. J., & Reid, J. B. (2000). Detecting and 544–564.
describing preventive intervention effects in a universal school- Zahn-Waxler, C., Friedman, S. L., & Cummings, E. M. (1983).
based randomized trial targeting delinquent and violent youth. Children’s emotions and behaviors in response to infants’ cries.
Journal of Consulting and Clinical Psychology, 68, 296–306. Child Development, 54, 1522–1528. doi:10.2307/1129815.
Sutton, J., Smith, P. K., & Swettenham, J. (1999). Social cognition Zhou, Q., Valiente, C., & Eisenberg, N. (2003). Empathy and its
and bullying: Social inadequacy or skilled manipulation? British measurement. In S. J. Lopez & C. R. Snyder (Eds.), Positive
Journal of Developmental Psychology, 17, 435–450. psychological assessment: A handbook of models and measures
Tabachnick, B. G., & Fidell, L. S. (2001). Computer-assisted (pp. 269–284). Washington, DC: American Psychological
research design and analysis. Boston, MA: Allyn & Bacon. Association.
Thomas, D. R., & Zumbo, B. D. (2011). Difference scores from the Zumbo, B. D. (1999). The simple difference score as an inherently
point of view of reliability and repeated-measures ANOVA: In poor measure of change: Some reality, much mythology.
defence of difference scores for data analysis. Educational and Advances in Social Science Methodology, 5, 269–304.
Psychological Measurement. doi:10.1177/0013164411409929.
Tolan, P. H., Guerra, N. G., & Kendall, P. C. (1995). A
developmental-ecological perspective on antisocial behavior in

123

View publication stats