An Evaluation of Second Step

Abstract

The impact of a school-based violence prevention program, Second Step, on peer victimization and aggression, and emotion regulation was evaluated among 457 sixth graders. A cluster-randomized trial was conducted with classrooms randomly assigned to intervention (n = 14) or control (n = 14) conditions. A repeated measures analysis of covariance on each measure was conducted using random effects. Several intervention effects were moderated by gender and disability status. Teacher-rated relational victimization at posttest decreased for students with disabilities in intervention but not control classrooms. Students without disabilities in intervention classrooms reported greater decreases from pretest to 6-month follow-up in overt aggression. Boys in the intervention classrooms had smaller increases in teacher-rated overt aggression at posttest than boys in control classrooms. Girls in intervention classrooms reported greater decreases from pretest to 6-month follow-up in relational aggression than girls in control classrooms. Study implications and directions for future research are discussed.

Keywords

special education middle/junior high program evaluation

Negative consequences of peer victimization are well documented and include difficulties in psychosocial adjustment and academic achievement (Buhs, Ladd, & Herald, 2006; Reijntjes et al., 2011). Peer victimization and aggression are connected in that victimized youth may retaliate, which can perpetuate a sequence of maladaptive behaviors (Farrell, Mehari, Kramer-Kuhn, & Goncy, 2014). The prevalence of peer victimization and aggression ranges from 20% to 35% among youth (Centers for Disease Control and Prevention [CDC], 2012), highlighting the need for violence prevention programs.

Although many schools implement violence prevention programs, few studies have evaluated their impact for adolescents with disabilities. Rose, Monda-Amaya, and Espelage (2011) reviewed studies that showed higher rates of peer victimization among youth with versus without disabilities (e.g., Norwich & Kelly, 2004) and, in a few studies, higher rates of peer-based aggression (e.g., Svetaz, Ireland, & Blum, 2000). These findings underscore the need for violence prevention programs that are responsive to the needs of youth with disabilities.

Youth with disabilities may be particularly vulnerable during early adolescence in the middle school setting. Contextual dynamics of middle school (i.e., larger student body, multiple teachers and classes) create increased demands for independence and reduced social support (Seidman, Allen, Aber, Mitchell, & Feinman, 1994). Some studies found escalating trajectories of peer-based aggression during middle school (Farrell, Sullivan, Esposito, Meyer, & Valois, 2005). Social cognitive and emotional development leads to increased sophistication in peer communications in terms of prosocial skills such as empathy but also negative interactions including relational and overt victimization (Yoon, Barton, & Taiariol, 2004). Alternatively, growth in social cognition (e.g., abstract reasoning and mutual perspective-taking skills) during early adolescence makes middle school an opportune time to address risk factors for peer-based aggression and to form prosocial behavior patterns.

The Second Step program (Committee for Children, 1997) follows a person-centered framework and addresses social and emotional competencies including the promotion of self and social awareness, effective decision making, and self and relationship management (Zins, Weissberg, Wang, & Walberg, 2004). Thus, it encompasses general characteristics of social and emotional learning programs that have been shown to increase positive attitudes toward school, prosocial behavior, and academic achievement (Zins et al., 2004). Specific mechanisms of change for Second Step include increased self-awareness and self-management in recognizing and managing anger, increased social awareness based on perspective taking and empathetic responding, and enhanced decision-making and relationship-management skills in generating and applying problem-solving skills to prevent or effectively address interpersonal conflict (Committee for Children, 1997). These positive changes at the individual level would hopefully translate into more positive student-student relationships at the peer level.

Second Step developers note that their programs have been implemented in 25,000 schools reaching nine million children in the United States and other countries (Committee for Children, 2014). However, few studies have evaluated middle school implementations of Second Step. One study of 223 sixth graders examined the effect of Second Step implemented with or without the assistance of peer leaders. For the “teacher plus peer leader” group, students in the intervention versus control classrooms reported higher levels of negative attitudes toward violent responses (Orpinas, Parcel, McAlister, & Frankowski, 1995). Among 714 students in middle and junior high schools, analyses of intervention effects were conducted separately for students based on their receipt of the Year 1 or 2 Second Step curricula (Van Schoiack-Edstrom, Frey, & Beland, 2002). For the Years 1 and 2 curricula, students in the intervention versus control classrooms reported less support for social exclusion at posttest. For Year 2, students in the intervention condition also reported less support for aggression and fewer challenges in using social skills at posttest (Van Schoiack-Edstrom et al., 2002). In another study of 156 fifth through eighth graders attending two inner-city schools who received Second Step, increases at posttest were found in curriculum knowledge, empathy, and prosocial behavior (McMahon & Washburn, 2003). Using a between-school design, Sprague et al. (2001) paired Second Step with a school-level Effective Behavior Support program and found that students in the three intervention versus three comparison middle schools reported fewer disciplinary code referrals 1 year later (Sprague et al., 2001). In another evaluation including 36 middle schools, Espelage, Low, Polanin, and Brown (2013) found greater decreases in physical aggression for sixth graders in intervention versus control schools at posttest.

A better understanding of both general and subgroup effects for Second Step is needed in the middle school context, including the effectiveness of this program for youth with and without disabilities and for boys and girls. More generally, Second Step may be helpful because it offers opportunities to generate and practice problem-solving skills for conflict situations that arise in middle school, and builds skills in anger management, perspective taking, and empathetic responding that may foster inclusion based on a better understanding of students’ viewpoints, strengths, and challenges. As some youth with disabilities may be at risk for peer victimization based on difficulties in areas of social and emotional competence (Vallance, Cummings, & Humphries, 1998), Second Step may be particularly relevant. For example, Vallance et al. (1998) found that youth with disabilities were more likely than their typically developing peers to use responses that may result in peer victimization. This may contribute to higher levels of marginalization and rejection from normative peer groups, and result in isolation that may prompt chronic victimization and/or affiliation with deviant peers (Mishna, 2003). These outcomes may also restrict opportunities for supportive peer interactions that offer a forum to develop adaptive coping and social and emotion regulation skills (Yoon et al., 2004). Universal programs such as Second Step may benefit youth with disabilities because not only are they learning skills that may contribute to greater social and emotional competence, but their peers without disabilities are also simultaneously learning similar skills. The importance of social dynamics in peer interactions (Gest, Madill, Zadzora, Miller, & Rodkin, 2014) highlights the need to address the social-interactional context within which all youth interact. Therefore, Second Step may be particularly helpful as it specifically addresses areas of social and emotional competence and offers opportunities for modeling, practice, and application of these skills.

Some prior evaluations of Second Step for middle school students have addressed the potential influence of gender on outcomes (Orpinas et al., 1995; McMahon & Washburn, 2003; Van Schoiack-Edstrom et al., 2002). Orpinas et al. (1995) found a trend for decreased aggressive behavior among boys only in two of six intervention classrooms as compared with boys in the control condition. McMahon and Washburn (2003) assessed but found no significant differences in the effect of Second Step on violence prevention knowledge and skills, empathy, or prosocial behaviors by gender. Finally, Van Schoiack-Edstrom et al. (2002) tested for potential gender differences in intervention effects for Years 1 and 2 of the Second Step curricula for the endorsement of physical, verbal, and relational forms of aggression and perceived difficulty in using social skills. None were found for the Year 1 Second Step curriculum, but for the Year 2 curriculum, girls in the comparison versus intervention group increased their endorsement of relational aggression from pretest to posttest. Notably, this study was the first to examine the impact of Second Step on relational aggression. Relational (i.e., behaviors intended to harm social relationships) and overt (i.e., behaviors that involve direct verbal teasing or physical harm) aggression and victimization are important intervention outcomes to examine as some researchers suggest that the way these forms of victimization are perceived and their relevance varies for boys and girls, with relational and physical victimization viewed as more damaging by girls and boys, respectively (Paquette & Underwood, 1999).

Overview of the Current Study

Although Second Step has the potential to enhance youths’ social, emotional, and behavioral competencies, more research is needed on its effects for subgroups of students including youth with disabilities and boys and girls. Our study used a randomized control trial to evaluate the impact of Second Step on aggression, victimization, and emotion regulation and extends the literature in several ways. First, effects were examined at posttest and two follow-up points. Only one published Second Step evaluation study included follow-up data (Orpinas et al., 1995). We also tested the moderating effects of disability status and gender. We anticipated that sixth graders receiving Second Step would report greater decreases in student- and teacher-reported aggression and victimization and increases in teacher-reported emotion regulation as compared with those in the control condition. As little research has examined differential effects of Second Step on these outcomes by disability status or gender, analyses were exploratory.

Method

Setting and Participants

Participants included sixth graders who attended three public middle schools in the Southeastern United States: two urban schools and one rural school located in a county school district bordering the urban schools’ district. At the time of the study, 83% of students in the urban schools and 22% of students in the rural school were eligible for the federally subsidized school lunch program. Of the 457 study participants, 63% were Black, 22% were White, 1% were Hispanic/Latino, 11% were Multiracial, and 3% endorsed Other race/ethnicities. Participants (50% male) ranged in age from 10 to 14 years (M = 11.3 years, SD = 0.6 years). A total of 105 students received special education services, and included youth with speech language impairments (4%), learning disabilities (11%), intellectual disability (2%), emotional disturbance (3%), and other health impairments (e.g., medical diagnoses and attention problems, 4%). The percentages for the specific disability statuses are reported based on the entire sample. Students reported a variety of family structures including living with (a) both biological parents (36%), (b) mother and another adult such as stepfather or grandparent (23%), (c) mother only (22%), (d) father only (3%), and (e) another living arrangement (16%). No significant differences in gender, race/ethnicity, or disability status were found for youth in the intervention versus control group.

Design

A cluster-randomized design was used in which 28 classrooms were randomly assigned to intervention or control conditions, including 16 health and physical education or enrichment classrooms in the urban schools and 12 health and physical education classrooms in the rural school. Equal numbers of classrooms were assigned to each condition at each school. Students in the control condition participated in typical activities for either health and physical education or the enrichment class, and not a violence prevention curriculum. At all participating schools, students generally transitioned classes multiple times a day, such that classes were taught by different teachers and included a slightly different composition of students. The degree to which students with disabilities split their day between general and special education classes varied as function of their individual needs (e.g., as specified by Individual Education Plans).

Procedures

All study procedures were approved by the University’s Institutional Review Board. Written parental permission (including permission to collect disability status and teacher report data) and adolescent assent were obtained for all participants. Of 592 eligible sixth graders, the consent rate was 77% (n = 457). See Figure 1 for the flow of participants through the study. Data were available for 456 students at Wave 1 (pretest—fall 2008), 439 at Wave 2 (posttest—spring 2009), 396 at Wave 3 (follow-up—fall 2009), and 382 at Wave 4 (follow-up—spring 2010). Students received a US$10 gift card for participating in each survey. Teacher ratings of student behavior were completed concurrently with student assessments, and included the teacher who taught a core academic subject (i.e., math, science, history, or English) and had the student in at least one class. Because physical education and enrichment (not core academic) classes were selected as intervention and control, no teachers in the control or intervention classrooms completed ratings of student behavior. Informed consent was obtained from eligible teachers, who then completed a 15- to 20-min survey for each student they were assigned and received US$20 per completed survey.

Figure 1.

Consort diagram.

Second Step Intervention

Second Step (Committee for Children, 1997) was delivered during the 2008–2009 school year at each school (in the first or second semester for the urban schools and from December 2008 to May 2009 for the county school). It included 15 lessons addressing interpersonal conflicts and problem solving, empathy and perspective taking, anger management, and the application of skills related to making a complaint, diffusing fights, dealing with bullying situations, and resisting peer and gang pressure. We incorporated relevant problem situations from qualitative studies with middle school students for role plays within the lessons (e.g., Farrell et al., 2006). Before Second Step started, an introductory session was conducted on prosocial behaviors and social networks in middle school. Five interventionists delivered Second Step (4 women, 1 man; 4 Black, 1 White) and were study staff. The two lead interventionists completed the Committee for Children’s Second Step training, and a train-the-trainer model was used to train the other interventionists. The two lead interventionists were former special education teachers of adolescents, both of whom were graduate students. Another interventionist was a retired middle school teacher. The two other interventionists were graduate students with experience working with and/or teaching adolescents.

Four independent observers assessed treatment fidelity in intervention classrooms. A 2-hr observer training was conducted at the start of the school year by the second author, with an additional re-calibration training held in January. Observers assessed the adherence and competence of seven items related to lesson delivery (e.g., reviews key concepts from previous lesson; checks for student understanding). Adherence was rated on a 4-point Likert-type scale: 1 = never, 2 = some of the time, 3 = most of the time, and 4 = always. Competence was also rated on a 4-point Likert-type scale: 1 = poor, 2 = fair, 3 = good, and 4 = very good. Alphas for the adherence and competence scales were .65 and .60, respectively. Fifty-six observations were conducted out of 117 total lessons (48%) across eight classrooms at the three schools. Mean adherence was 3.51 (SD = 0.59) and mean competence was 3.71 (SD = 0.30).

Measures

Student-reported aggression and victimization

Aggression was measured using three Problem Behavior Frequency Scales–Student Form (PBFS-S; Farrell, Kung, White, & Valois, 2000). They included the 6-item Relational Aggression scale (e.g., Spread a false rumor about someone) and a 12-item Overt Aggression scale, created by combining the Physical (e.g., Hit or slapped someone) and Nonphysical (e.g., Put someone down to their face) scales. The Overt Aggression scale was created to parallel the Overt Victimization scale. Victimization was assessed using 6-item Overt (e.g., been hit by another kid) and 6-item Relational (e.g., someone made fun of you to make others laugh) Victimization scales. Students reported how frequently they engaged in each behavior in the past 30 days using a 6-point Likert-type scale: 1 = never, 2 = 1 to 2 times, 3 = 3 to 5 times, 4 = 6 to 9 times, 5 = 10 to 19 times, and 6 = 20 or more times. Alphas at Wave 1 for the Overt and Relational Aggression scales were .87 and .69, respectively, and .85 and .84 for the Overt and Relational Victimization scales, respectively.

Teacher-reported aggression and victimization

Aggression was measured using three scales from the PBFS–Teacher Form (PBFS-T; Farrell et al., 2000). These scales included the 6-item Relational Aggression scale (e.g., This student left another student out on purpose when it was time to do an activity) and the combined Physical (e.g., This student shoved or pushed somebody) and Nonphysical (e.g., This student insulted someone’s family) Aggression scales that formed a 12-item Overt Aggression scale. Victimization was assessed using two PBFS-T scales including the 6-item Overt (e.g., This student had another kid threatened to hit or physically harm him or her) and 6-item Relational (e.g., This student has been left out on purpose by other kids when it was time to do an activity) Victimization scales. The frequency that each student engaged in each behavior in the past 30 days was rated using a 4-point Likert-type scale: 1 = never, 2 = sometimes, 3 = often, and 4 = almost always. Alphas at Wave 1 for the Overt and Relational Aggression scales were .96 and .85, respectively, and .89 and .86 for the Overt and Relational Victimization scales, respectively.

Teacher ratings of students’ aggressive behavior were also assessed using the 20-item Aggression scale from the Child Behavior Checklist (CBCL; Achenbach, 1991). Teachers rated how frequently each student engaged in each behavior over the past 3 months using a 3-point Likert-type scale: 0 = not true, 1 = somewhat or sometimes true, and 3 = very true or often true. The alpha for the Aggression scale at Wave 1 was .97. As is often the case with scores representing problem behaviors, scores on the aggression and victimization measures had a high degree of skewness and kurtosis. This was reduced by using a log transformation with the resulting scores then multiplied by 10 to facilitate reporting of parameter estimates.

Emotion regulation

Teachers rated students’ levels of emotion regulation using the Emotion Regulation scale from the Emotion Regulation Checklist (ERC; Shields & Cicchetti, 1997). This 8-item scale assesses the child’s ability to show situational-appropriate affective responses, empathy, and self-awareness of his or her own emotions (e.g., Can this student say when s/he is feeling sad, angry or mad, fearful, or afraid?). Teachers rated how well each item described the child using a 4-point Likert-type scale: 1 = never, 2 = sometimes, 3 = often, and 4 = always. The alpha for the Emotion Regulation scale at Wave 1 was .82.

Data Analysis

A repeated measures analysis of covariance was conducted on each outcome measure using SAS Proc Mixed to take into account the clustering of students within classrooms. Missing data were addressed using full-information maximum likelihood estimates that included data from all participants who participated at one or more waves. Wave was treated as a dummy-coded within-person variable with baseline outcome as the reference group such that scores at each posttest wave represented the change from pretest. Intervention condition was treated as a classroom-level variable and was dummy-coded with control classrooms representing the reference group. Analyses were based on intent-to-treat with participants classified based on the condition to which their classrooms were randomized regardless of their actual degree of intervention exposure. Covariates were all grand-mean centered and included age and dummy-coded variables representing male gender, disability status, and school with the county school serving as the reference. Main effects for intervention condition and covariates represent their relation with pretest scores on the outcome measure, and intervention and covariate interactions with wave represent their relation with pre-to-post intervention changes at each posttest wave. The model specified random intercepts at the classroom level and random effects for intercepts and wave at the student level. Degrees of freedom were based on Kenward–Roger estimates (Kenward & Roger, 1997). Effect size estimates (d-coefficients; Cohen, 1988) were calculated by dividing unstandardized intervention effects at each wave by the pretest standard deviation.

Subgroup analyses were conducted using procedures recommended by Farrell, Henry, and Bettencourt (2013). More specifically, the extent to which intervention effects were moderated by gender was tested by adding Gender × Intervention condition and Gender × Intervention condition × Wave interactions into the main effects model. For significant interaction effects, the estimate function in Proc Mixed was used to assess the significance of intervention effects within each subgroup. A similar approach was used to test for differences in intervention effects by disability status.

Results

Descriptive Analyses

Descriptive statistics are reported in Table 1. Means and standard deviations were calculated for all untransformed outcomes at Wave 1 and correlations were calculated for transformed variables. Student and teacher ratings of overt and relational aggression were positively correlated with each other (rs = .23 to .90). Student and teacher ratings of overt and relational victimization were also positively correlated with each other (rs = .14 to .64). Teacher-rated emotion regulation was negatively correlated with teacher-reported aggression (rs = −.33 to −.43) and victimization (rs = −.27 to −.29), and student-reported aggression (rs = −.15 to −.16).

Table 1.

Means, Standard Deviations, and Intercorrelations for Wave 1 Outcomes.

Variable	1	2	3	4	5	6	7	8	9	M	SD
1. Overt Agg (S)	—									1.64	0.69
2. Overt Agg (T)	.32**	—								1.31	0.54
3. CBCL Agg (T)	.33**	.87**	—							0.29	0.43
4. Relational Agg (S)	.71**	.26**	.31**	—						1.49	0.60
5. Relational Agg (T)	.29**	.90**	.83**	.23**	—					1.26	0.44
6. Overt Vict (S)	.48**	.22**	.20**	.33**	.18**	—				1.66	0.85
7. Overt Vict (T)	.23**	.76**	.67**	.19**	.66**	.24**	—			1.20	0.38
8. Relational Vict (S)	.36**	.17**	.19**	.42**	.10	.60**	.15*	—		1.57	0.78
9. Relational Vict (T)	.16*	.64**	.62**	.15*	.65**	.19**	.64**	.14*	—	1.13	0.32
10. Emotion Reg (T)	−.16*	−.36**	−.43*	−.15*	−.33**	−.09	−.27**	−.08	−.29**	2.88	0.56

Note. N = 456. Agg = aggression; Vict = victimization; Reg = regulation; S = student report; T = teacher report.

p < .01. **p < .001.

Intervention Effects on Outcomes

Aggression

Across aggression measures, no significant pretest differences were found between intervention and control classrooms (see Table 2). At pretest, students with disabilities had higher frequencies of student-reported overt aggression (d = .25) and teacher-reported overt (d = .58) and relational aggression (d = .47) on the PBFS-T and aggression on the CBCL (d = .53) than those without disabilities. There were no significant gender differences on any aggression measure. Age was positively related to teacher-rated aggression (ds = .15 to .16). Compared with students in the county school, those in the urban schools had higher rates of overt and relational aggression based on student (ds = .41 to .52) and teacher (ds = .45 to .74) reports.

Table 2.

Unstandardized Parameter Estimates From Repeated Measures Analysis of Covariance Examining Intervention Effects on Aggression at Baseline and Changes Over Time.

Variable	Overt Agg (S)	Overt Agg (T)	CBCL Agg (T)	Rel Agg (S)	Rel Agg (T)
Intercept	1.82 (0.11)	0.94 (0.11)	0.94 (0.09)	1.47 (0.10)	0.80 (0.09)
Effects on pretest
Intervention	0.04 (0.15)	0.00 (0.15)	−0.02 (0.13)	0.02 (0.14)	0.02 (0.13)
Male gender	−0.09 (0.16)	0.19 (0.12)	0.14 (0.11)	0.07 (0.14)	0.13 (0.12)
Disability status	0.40 (0.19)*	0.79 (0.15)***	0.65 (0.14)***	0.30 (0.17)	0.57 (0.14)***
Age	0.15 (0.12)	0.22 (0.10)*	0.20 (0.09)*	0.09 (0.11)	0.18 (0.09)*
School 1 vs. 3	0.66 (0.19)***	1.01 (0.18)***	0.80 (0.16)***	0.61 (0.17)**	0.76 (0.15)***
School 2 vs. 3	0.84 (0.20)***	0.67 (0.19)**	0.55 (0.16)**	0.69 (0.17)***	0.59 (0.16)***
Pre-to-post change
Wave 1 to Wave 2 change	0.31 (0.11)**	0.38 (0.08)***	0.20 (0.07)**	0.28 (0.10)**	0.32 (0.09)***
Wave 1 to Wave 3 change	0.18 (0.12)	0.07 (0.09)	0.08 (0.07)	−0.17 (0.11)	0.06 (0.09)
Wave 1 to Wave 4 change	0.19 (0.11)	0.49 (0.09)***	0.33 (0.08)***	−0.07 (0.11)	0.40 (0.09)***
Covariate effects on pre-to-post change
Male × Wave 2	−0.06 (0.15)	−0.17 (0.12)	−0.13 (0.10)	−0.03 (0.15)	−0.34 (0.12)**
Male × Wave 3	−0.26 (0.17)	−0.06 (0.13)	−0.01 (0.10)	0.05 (0.16)	−0.01 (0.13)
Covariate effects on pre-to-post change
Male × Wave 4	−0.39 (0.16)*	−0.08 (0.13)	−0.08 (0.11)	−0.16 (0.15)	−0.15 (0.13)
Disability × Wave 2	−0.47 (0.19)*	−0.09 (0.14)	0.02 (0.12)	−0.28 (0.18)	−0.21 (0.15)
Disability × Wave 3	−0.21 (0.20)	−0.39 (0.15)*	−0.27 (0.12)*	−0.16 (0.20)	−0.26 (0.16)
Disability × Wave 4	−0.55 (0.20)**	−0.32 (0.16)*	−0.26 (0.13)*	−0.52 (0.19)**	−0.28 (0.16)
Intervention effects on pre-to-post change
Intervention × Wave 2 change	0.08 (0.15)	−0.09 (0.11)	0.00 (0.10)	−0.22 (0.14)	−0.05 (0.12)
Intervention × Wave 3 change	−0.29 (0.16)	−0.10 (0.13)	−0.15 (0.10)	−0.10 (0.16)	−0.10 (0.13)
Intervention × Wave 4 change	0.05 (0.16)	−0.15 (0.13)	−0.20 (0.10)	−0.01 (0.15)	−0.09 (0.13)

Note. Results based on multilevel model in which students were nested within classrooms. Agg = aggression; Rel = relational; S = student report; T = teacher report. All covariates were centered.

p < .05. **p < .01. ***p < .001.

Analyses of changes across waves revealed higher scores relative to pretest on all five measures of aggression at Wave 2 (ds = .17 to .28) and on teacher ratings at Wave 4 (ds = .27 to .36). This pattern suggests an increase in aggression during the sixth-grade school year for students’ ratings, and during both school years for teachers’ ratings rather than an overall increase across waves. Students in intervention and control classrooms did not differ in their patterns of change over time. Disability status × Wave interactions were significant at several waves showing that although scores on aggression measures were higher at pretest for students with disabilities, differences decreased on several measures at subsequent waves.

Victimization

Intervention and control classrooms did not differ on pretest measures of victimization (see Table 3). Victimization frequencies were higher for students with disabilities based on student and teacher reports of overt (ds = .24 and .50, respectively) and relational victimization (ds = .23 and .27, respectively). Teacher-rated overt victimization was higher for boys (d = .34). Victimization frequencies in the two urban schools were higher than the county school for student ratings of overt victimization (d = .24 to .34), and teacher ratings of overt (ds = .48 to .65) and relational (both ds =.37) victimization.

Table 3.

Unstandardized Parameter Estimates From Repeated Measures Analysis of Covariance Examining Intervention Effects on Victimization at Baseline and Changes Over Time.

Variable	Overt Vict (S)	Overt Vict (T)	Rel Vict (S)	Rel Vict (T)	Emotion Reg (T)
Intercept	1.71 (0.11)	0.70 (0.08)	1.52 (0.11)	0.36 (0.07)	23.16 (0.30)
Effects on pretest
Intervention	0.16 (0.16)	−0.09 (0.11)	0.13 (0.16)	0.13 (0.09)	−0.23 (0.42)
Male gender	0.19 (0.16)	0.37 (0.1)***	−0.11 (0.15)	0.00 (0.09)	−0.76 (0.41)
Disability status	0.42 (0.20)*	0.55 (0.12)***	0.40 (0.19)*	0.24 (0.11)*	−1.58 (0.50)**
Age	−0.05 (0.13)	0.12 (0.08)	0.00 (0.12)	0.15 (0.07)*	−1.09 (0.32)***
School 1 vs. 3	0.43 (0.19)*	0.70 (0.13)***	0.30 (0.19)	0.34 (0.11)**	0.06 (0.50)
School 2 vs. 3	0.61 (0.20)**	0.52 (0.13)***	0.28 (0.19)	0.34 (0.12)**	−2.36 (0.53)***
Pre-to-post change
Wave 1 to Wave 2 change	−0.05 (0.11)	0.15 (0.08)	−0.02 (0.11)	0.24 (0.08)**	0.38 (0.44)
Wave 1 to Wave 3 change	−0.33 (0.13)**	−0.13 (0.08)	−0.44 (0.12)***	0.04 (0.08)	−0.63 (0.53)
Wave 1 to Wave 4 change	−0.35 (0.12)**	0.07 (0.09)	−0.63 (0.12)***	0.15 (0.08)	0.27 (0.51)
Covariate effects on pre-to-post change
Male × Wave 2	−0.07 (0.16)	−0.15 (0.12)	−0.12 (0.15)	−0.15 (0.11)	0.47 (0.65)
Male × Wave 3	−0.20 (0.18)	−0.21 (0.12)	0.12 (0.17)	0.07 (0.12)	0.95 (0.63)
Male × Wave 4	−0.22 (0.18)	−0.06 (0.12)	−0.10 (0.17)	−0.01 (0.11)	0.38 (0.44)
Disability × Wave 2	−0.34 (0.20)	−0.01 (0.14)	−0.06 (0.19)	−0.09 (0.14)	−0.63 (0.53)
Disability × Wave 3	−0.20 (0.22)	0.12 (0.15)	−0.08 (0.21)	0.00 (0.14)	0.27 (0.51)
Disability × Wave 4	−0.68 (0.22)**	0.06 (0.15)	−0.52 (0.21)*	−0.04 (0.14)	0.32 (0.53)
Intervention effects on pre-to-post change
Intervention × Wave 2 change	0.08 (0.16)	0.01 (0.12)	−0.01 (0.15)	−0.04 (0.11)	−0.02 (0.43)
Intervention × Wave 3 change	−0.09 (0.18)	0.18 (0.12)	−0.02 (0.17)	−0.09 (0.11)	−0.05 (0.52)
Intervention × Wave 4 change	0.02 (0.17)	0.13 (0.12)	0.22 (0.16)	−0.08 (0.11)	−0.09 (0.50)

Note. Results based on multilevel model in which students were nested within classrooms. Vict = victimization; Reg = regulation; Rel = relational; S = student report; T = teacher report. All covariates were centered.

p < .05. **p < .01. ***p < .001.

Analyses of patterns of change over time indicated that relative to pretest scores, scores were higher on teacher ratings of relational victimization at Wave 2 (d = .27). In contrast, students reported lower frequencies of both overt and relational victimization at Waves 3 (ds = −.19 and −.26, respectively) and 4 (ds = −.20 and −.37, respectively). Students in intervention and control classrooms did not significantly differ in their patterns of change over time. Analyses of Disability status × Wave interactions suggested that the difference on student-reported victimization for youth with and without disabilities was less evident by Wave 4 (ds = −.30 to −.38). Intervention condition × Wave interactions indicated no significant intervention effects.

Emotion regulation

No significant pretest differences were found between intervention and control classrooms on teacher-rated emotion regulation (see Table 3). Students with disabilities had lower teacher ratings of emotion regulation (d = −.35). Age was related to teacher ratings of emotion regulation (d = −.24). Lower ratings of emotion regulation were found at one urban school compared with the county school (d = −.53). Analyses of changes from pretest revealed increases in ratings of emotion regulation at Waves 2 (d =.17) and 4 (d = .16) but not at Wave 3. No significant intervention effects were found for emotion regulation.

Disability Status Differences in Intervention Effects on Outcomes

Analyses tested the extent to which intervention effects differed for students with and without disabilities (see Table 4). As a higher percentage of boys were classified as having a disability, these models controlled for Gender × Intervention condition and Gender × Intervention condition × Wave interactions. Intervention effects on pretest to Wave 3 changes in student reports of overt aggression differed across subgroups, t(772) = 2.74, p = .006 (see Figure 2). A moderate intervention effect was found on overt aggression such that the frequency decreased for students without disabilities in intervention versus control classrooms from Wave 1 to Wave 3 (d = −.33, p = .004). A significant difference across subgroups in intervention effects was found for teacher-rated relational victimization at Wave 2, t(570) = −2.36, p = .018 (see Figure 2). A large intervention effect was found on relational victimization such that scores decreased for students with disabilities in intervention classrooms but increased for those in control classrooms (d = −.58, p = .025). Significant interaction effects were found at Wave 3 for student-rated overt, t(781) = 2.11, p = .035, and relational, t(763) = 2.06, p = .040, victimization, but these effects were not significant for either subgroup.

Table 4.

Subgroup Analyses of Intervention Effects by Disability Status.

Variable	Effect	Posttest	6 months	12 months
Overt aggression (S)	Interaction effect	0.25 (0.23)	0.66 (0.24)**	0.22 (0.24)
	No disability	−0.01 (0.11)	−0.33 (0.11)**	−0.02 (0.11)
	Disability	0.24 (0.20)	0.33 (0.21)	0.20 (0.21)
Overt aggression (T)	Interaction effect	−0.31 (0.20)	−0.13 (0.22)	−0.25 (0.23)
	No disability	0.00 (0.09)	−0.05 (0.10)	−0.06 (0.11)
	Disability	−0.31 (0.18)	−0.18 (0.19)	−0.31 (0.20)
CBCL aggression (T)	Interaction effect	−0.02 (0.19)	0.32 (0.20)	0.21 (0.21)
	No disability	0.00 (0.09)	−0.20 (0.09)*	−0.21 (0.10)*
	Disability	−0.02 (0.17)	0.12 (0.17)	0.00 (0.18)
Relational aggression (S)	Interaction effect	0.34 (0.24)	0.41 (0.26)	0.44 (0.25)
	No disability	−0.22 (0.11)*	−0.16 (0.12)	−0.09 (0.12)
	Disability	0.11 (0.21)	0.25 (0.23)	0.35 (0.22)
Relational aggression (T)	Interaction effect	−0.31 (0.24)	−0.45 (0.25)*	−0.24 (0.26)
	No disability	0.03 (0.11)	0.02 (0.12)	−0.02 (0.12)
	Disability	−0.28 (0.21)	−0.43 (0.22)*	−0.25 (0.23)
Overt victimization (S)	Interaction effect	0.37 (0.21)	0.51 (0.24)*	0.44 (0.24)
	No disability	−0.04 (0.10)	−0.17 (0.11)	−0.08 (0.11)
	Disability	0.33 (0.19)	0.34 (0.21)	0.36 (0.21)
Overt victimization (T)	Interaction effect	−0.26 (0.26)	0.24 (0.26)	0.23 (0.26)
	No disability	0.06 (0.12)	0.11 (0.12)	0.07 (0.12)
	Disability	−0.20 (0.22)	0.35 (0.23)	0.30 (0.23)
Relational victimization (S)	Interaction effect	0.23 (0.21)	0.48 (0.23)*	0.37 (0.23)
	No disability	−0.06 (0.10)	−0.12 (0.11)	0.04 (0.11)
	Disability	0.17 (0.18)	0.36 (0.20)	0.41 (0.21)*
Relational victimization (T)	Interaction effect	−0.69 (0.29)*	−0.08 (0.3)	−0.22 (0.29)
	No disability	0.11 (0.14)	−0.08 (0.14)	−0.04 (0.14)
	Disability	−0.58 (0.26)*	−0.17 (0.26)	−0.26 (0.26)
Emotion regulation (T)	Interaction effect	0.20 (0.23)	0.31 (0.28)	0.19 (0.27)
	No disability	−0.05 (0.11)	−0.08 (0.13)	−0.06 (0.13)
	Disability	0.15 (0.20)	0.23 (0.25)	0.12 (0.24)

Note. S = student report; T = teacher report.

p < .05. **p < .01. ***p < .001.

Figure 2.

Moderating role of disability status on the effect of the intervention on student ratings of overt aggression and teacher ratings of relational victimization, respectively.

Gender Differences in Intervention Effects on Outcomes

Analyses testing the degree to which intervention effects were moderated by gender revealed that gender moderated intervention effects on teacher-rated overt aggression at Wave 2, t(475) = −2.17, p = .030. Subgroup estimates revealed that boys in the intervention versus control classrooms (d = −.25, p = .036) had smaller increases in overt aggression at Wave 2 (see Figure 3). Gender also moderated intervention effects for student-rated relational aggression at Wave 3, t(738) = 2.44, p = .015, and Wave 4, t(882) = 2.67, p = .008. Estimates of subgroup effects revealed several moderate gender-specific effects. At Wave 3 (see Figure 3), girls in the intervention classrooms showed a greater decrease from Wave 1 in relational aggression than those in control classrooms (d =−.33, p = .028). At Wave 4, there was a similar trend for girls (d = −.27, p = .055) and a trend in the opposite direction for boys (d = .28, p = .062).

Figure 3.

Moderating role of gender on the effect of the intervention on teacher ratings of overt aggression and student ratings of relational aggression, respectively.

Discussion

This study examined the impact of Second Step on middle school students’ aggression, victimization, and emotion regulation using a randomized control design, with specific interest in the moderating role of disability status and gender. Although no main intervention effects were evident, several significant subgroup effects emerged. Teachers reported greater decreases in relational victimization for students with disabilities in intervention versus control classrooms at posttest. Students without disabilities in intervention classrooms reported greater decreases in overt aggression from pretest to Wave 3 than did those in control classrooms. Teacher-rated overt aggression showed smaller increases for boys in intervention versus control classrooms at posttest. For girls in the intervention as compared with control classrooms, a significant decrease in relational aggression was found at Wave 3 with a similar trend at Wave 4. In contrast, a trend was found for boys in intervention classrooms for increased frequencies of relational aggression at Wave 4. No significant subgroup effects were found for teacher-rated emotion regulation.

Analyses of pretest differences indicated that teachers and students generally reported higher rates of aggression and victimization and teachers reported lower rates of emotion regulation for students with disabilities as compared with those without disabilities. Findings that students with disabilities were more likely to be exposed to peer violence, whether as a victim or aggressor, are consistent with other descriptive work (Rose et al., 2011) and highlight the need for prevention efforts that are relevant and effective for middle school youth with disabilities.

Although we did not find main effects for the intervention, subgroup analyses suggest that the effects differed based on gender and disability status. These analyses heed calls in the literature (e.g., Farrell et al., 2013) to investigate how interventions might work, or not work, across different contexts and subgroups. Our study is one of the first to explore subgroup differences in intervention effects for violence prevention programs by disability status. Although not consistent across waves or informants, Second Step produced a couple of subgroup effects based on disability status that are noteworthy. Specific effects were found for decreased relational victimization for students with disabilities in intervention versus control classrooms at posttest, and for greater decreases in overt aggression from pretest to Wave 3 for students without disabilities in intervention classrooms. Decreases in relational victimization for students with disabilities in the intervention classrooms may be related to information learned in Second Step on communication and problem solving and also to chances to practice the application of skills within a context where their peers were learning similar skills. Greater decreases in overt aggression for students without disabilities may reflect similar learned skills from Second Step in addition to perspective taking and empathy. Only a lesson or two was allocated to teaching some skills. As some youth with disabilities struggle in areas of social and emotional competencies, expansion of the curriculum to provide more in-depth coverage of lessons in these areas may be beneficial and may lead to more consistent and sustained intervention effects.

Although intervention effects were not found consistently at each wave or across student and teacher ratings, findings related to gender differences suggest that Second Step was more beneficial for boys in curtailing overt aggression and for girls in decreasing relational aggression. In contrast, a trend for boys in intervention versus control classrooms to exhibit more relational aggression emerged at Wave 4. Thus, Second Step produced some changes that were consistent with the types of aggression that boys and girls may experience more often and view as more salient (Paquette & Underwood, 1999).

Conceptual models (e.g., Zins et al., 2004) linking social and emotional learning to academic and social outcomes highlight the role of individual factors but acknowledge that individuals are embedded within the broader social ecology. Recent qualitative work (Sullivan et al., 2012) highlighted the role of peer group dynamics in problem situations faced by urban middle school students with disabilities. Effective teacher management of social dynamics (i.e., related to status extremes, aggression, positive behavior, and isolation) were also associated with a more positive and prosocial classroom climate (Gest et al., 2014). These findings, taken in sum, suggest the importance of the school context in reducing the risk of violence exposure for students with disabilities. Combining universal prevention programs that address individual-level skills with those that focus on school environment change such as the Olweus Bullying Prevention Program (OBPP; Olweus & Limber, 2010) or Bully Prevention in Positive Behavior Support (BP-PBS; Ross & Horner, 2014) holds promise for addressing peer factors associated with school-based violence exposure by creating a school environment conducive to using the skills taught in Second Step. In addition, findings that Second Step was effective in reducing some problem behaviors but not others for youth with disabilities indicate that some youth may need additional instruction and practice in certain areas and suggest the potential success of a multitiered approach. In multitiered prevention efforts such as Positive Behavior Interventions and Supports (PBIS; Sprague & Horner, 2006), universal prevention programs (Tier 1) form a foundation for targeted Tier 2 (e.g., small group) and intensive Tier 3 (e.g., one-on-one) interventions (Bruhn, Lane, & Hirsh, 2014). Thus, our findings suggest that Second Step may be an appropriate Tier 1 intervention for youth with and without disabilities.

It is important to note several study limitations and directions for future research. First, meaningful differences in intervention effects may exist based on specific types of disabilities (i.e., emotional and behavioral disorders); however, our sample size for students with disabilities precluded such explorations. Next steps for future research include testing the effect of specific components that may be most beneficial for youth with various types of disabilities. Also, randomization at the classroom level may have limited the intervention’s strength as some students in each school received Second Step while others did not. It is possible that this weakened our ability to detect main effects. Intervention effects may have been stronger with randomization at the school level and if all grades had participated (Henry, Farrell, & The Multisite Violence Prevention Project, 2004). This highlights a key direction for future research in that neither Second Step evaluation study that included school-level randomization delivered the intervention to all grades (Espelage et al., 2013; Sprague et al., 2001). In addition, the alpha coefficient for student-reported overt aggression was relatively weak at .69. Also, although the direct observation of adherence and competence in this study reflects recommendations on treatment fidelity measurement (Durlak & DuPre, 2008), the lack of observational data on student behavior as well as from comparison classrooms is a limitation of our study and provides important goals for future research. Finally, our study results offered some evidence that Second Step resulted in reductions in aggression for youth without disabilities while reducing rates of relational victimization experienced by youth with disabilities; it is plausible that these outcomes worked together to create a more positive learning environment for youth with disabilities. Given the influence of peer and school norms on aggression, a school environment approach to violence prevention may be a necessary underlying component to support the application and generalizability of individual social and emotional competencies learned in programs such as Second Step.

Footnotes

Authors’ Note

The research and interpretations reported are the sole responsibility of the authors and are not necessarily endorsed by the CDC or represent the views, opinions, or policies of the CDC or their staff.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Grant R01-DD000110 from the Centers for Disease Control and Prevention (CDC).

References

Achenbach

T. M.

(1991). Manual for the Teacher’s Report Form and 1991 profile. Burlington: Department of Psychiatry, University of Vermont.

Bruhn

A. L.

Lane

K. L.

Hirsch

S. E.

(2014). A review of Tier 2 interventions conducted within Multitiered models of behavioral prevention. Journal of Emotional and Behavioral Disorders, 22, 171–189. doi:10.1177/1063426613476092

Buhs

E. S.

Ladd

G. W.

Herald

S. L.

(2006). Peer exclusion and victimization: Processes that mediate the relation between peer group rejection and children’s classroom engagement and achievement? Journal of Educational Psychology, 98, 1–13. doi:10.1037/0022-0663.98.1.1

Centers for Disease Control and Prevention. (2012). Youth risk behavior surveillance—United States, 2011. MMWR Surveillance Summaries, 61(SS-4). Retrieved from www.cdc.gov/mmwr/pdf/ss/ss6104.pdf

Cohen

(1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum.

Committee for Children. (1997). Second step: A violence prevention curriculum: Middle school/junior high. Level 1 foundation lessons. Seattle, WA: Committee for Children.

Committee for Children. (2014). Our story. Retrieved from http://www.cfchildren.org/advocacy/about-us/our-story.aspx

Durlak

J. A.

DuPre

E. P.

(2008). Implementation matters: A review of research on the influence of implementation on program outcomes and the factors affecting implementation. American Journal of Community Psychology, 41, 327–350. doi:10.1007/s10464-008-9165-0

Espelage

D. L.

Low

Polanin

J. R.

Brown

E. C.

(2013). The impact of a middle school program to reduce aggression, victimization, and sexual violence. Journal of Adolescent Health, 53, 180–186. doi:10.1016/j.jadohealth.2013.02.021

10.

Farrell

A. D.

Henry

D. B.

Bettencourt

(2013). Methodological challenges examining subgroup differences: Examples from universal school-based youth violence prevention trials. Prevention Science, 14, 121–133. doi:10.1007/s11121-011-0200-2

11.

Farrell

A. D.

Kung

E. M.

White

K. S.

Valois

R. F.

(2000). The structure of self-reported aggression, drug use, and delinquent behaviors during early adolescence. Journal of Clinical Child & Adolescent Psychology, 29, 282–292. doi:10.1207/S15374424jccp2902_13

12.

Farrell

A. D.

Mehari

K. R.

Kramer-Kuhn

Goncy

E. A.

(2014). The impact of victimization and witnessing violence among high risk adolescents. Child Development, 85, 1694–1710. doi:10.1111/cdev.12215

13.

Farrell

A. D.

Sullivan

T. N.

Esposito

L. E.

Meyer

A. L.

Valois

R. F.

(2005). A latent growth curve analysis of the structure of aggression, drug use, and delinquent behaviors and their interrelations over time in urban and rural adolescents. Journal of Research on Adolescence, 15, 179–204. doi:10.1111/j.1532-7795.2005.00091.x

14.

Farrell

A. D.

Sullivan

T. N.

Kliewer

Allison

Erwin

E. H.

Meyer

A. L.

Esposito

L. E.

(2006). Peer and school problems in the lives of urban adolescents: Frequency, difficulty, and relation to adjustment. Journal of School Psychology, 44, 169–190. doi:10.1016/j.jsp.2006.02.001

15.

Gest

S. D.

Madill

R. A.

Zadzora

K. M.

Miller

A. M.

Rodkin

P. C.

(2014). Teacher management of elementary classroom social dynamics: Associations with changes in student adjustment. Journal of Emotional and Behavioral Disorders, 22, 107–118. doi:10.1177/1063426613512677

16.

Henry

D. B.

Farrell

A. D.

, The Multisite Violence Prevention Project. (2004). The study designed by a committee: Design of the Multisite Violence Prevention Project. American Journal of Preventive Medicine, 26, 12–19.

17.

Kenward

M. G.

Roger

J. H.

(1997). Small sample inference for fixed effects from restricted maximum likelihood. Biometrics, 53, 983–997. doi:10.2307/2533558

18.

McMahon

S. D.

Washburn

(2003). Violence prevention: An evaluation of program effects with urban African American students. The Journal of Primary Prevention, 24, 43–62. doi:10.1023/A:1025075617356

19.

Mishna

(2003). Learning disabilities and bullying: Double jeopardy. Journal of Learning Disabilities, 36, 336–347. doi:10.1177/00222194030360040501

20.

Norwich

Kelly

(2004). Pupils’ views on inclusion: Moderate learning difficulties and bullying in mainstream and special schools. British Educational Research Journal, 30, 43–65. doi:10.1080/01411920310001629965

21.

Olweus

Limber

S. P.

(2010). Bullying in school: Evaluation and dissemination of the Olweus Bullying Prevention Program. American Journal of Orthopsychiatry, 80, 124–134. doi:10.1111/j.1939-0025.2010.01015.x

22.

Orpinas

Parcel

G. S.

McAlister

Frankowski

(1995). Violence prevention in middle schools: A pilot evaluation. Journal of Adolescent Health, 17, 360–371. doi:10.1016/1054-139X(95)00194-W

23.

Paquette

J. A.

Underwood

M. K.

(1999). Gender differences in young adolescents’ experiences of peer victimization: Social and physical aggression. Merrill-Palmer Quarterly, 45, 242–266.

24.

Reijntjes

Kamphuis

J. H.

Prinzie

Boelen

P. A.

van der Schoot

Telch

M. J.

(2011). Prospective linkages between peer victimization and externalizing problems in children: A meta-analysis. Aggressive Behavior, 37, 215–222. doi:10.1002/ab.20374

25.

Rose

C. A.

Monda-Amaya

L. E.

Espelage

D. L.

(2011). Bullying perpetration and victimization in special education: A review of the literature. Remedial and Special Education, 32, 114–130. doi:10.1177/0741932510361247

26.

Ross

S. W.

Horner

R. H.

(2014). Bully prevention in positive behavior support: Preliminary evaluation of third-, fourth-, and fifth-grade attitudes toward bullying. Journal of Emotional and Behavioral Disorders, 22, 225–236. doi:10.1177/1063426613491429

27.

Seidman

Allen

Aber

J. L.

Mitchell

Feinman

(1994). The impact of school transition in early adolescence on the self-system and perceived social context of poor urban youth. Child Development, 65, 507–522. doi:10.2307/1131399

28.

Shields

Cicchetti

(1997). Emotion regulation among school age children: The development and validation of a new criterion Q-sort scale. Developmental Psychology, 33, 906–916. doi:10.1037/0012-1649.33.6.906.

29.

Sprague

J. R.

Horner

R. H.

(2006). School wide positive behavioral supports. In Jimerson

S. R.

Furlong

(Eds.), Handbook of school violence and school safety: From research to practice (pp. 413–427). Mahwah, NJ: Lawrence Erlbaum.

30.

Sprague

Walker

Golly

White

Myers

Shannon

. (2001). Translating research into effective practice: The effects of a universal staff and student intervention on indicators of discipline and school safety. Education & Treatment of Children, 24, 495–511.

31.

Sullivan

T. N.

Helms

S. W.

Bettencourt

A. F.

Sutherland

K. S.

Lotze

Mays

Farrell

A. D.

(2012). A qualitative study of individual and peer factors related to effective nonviolent versus aggressive responses to problem situations among adolescents with high incidence disabilities. Behavioral Disorders, 37, 163–178.

32.

Svetaz

M. V.

Ireland

Blum

(2000). Adolescents with learning disabilities: Risk and protective factors associated with emotional well-being: Findings from the national longitudinal study of adolescent health. Journal of Adolescent Health, 27, 340–348. doi:10.1016/S1054-139X(00)00170-1

33.

Vallance

D. D.

Cummings

R. L.

Humphries

(1998). Mediators of the risk for problem behavior in children with language learning disabilities. Journal of Learning Disabilities, 31, 160–171. doi:10.1177/002221949803100206

34.

Van Schoiack-Edstrom

Frey

K. S.

Beland

. (2002). Changing adolescents’ attitudes about relational and physical aggression: An early evaluation of a school based intervention. School Psychology Review, 31, 201–216.

35.

Yoon

J. S.

Barton

Taiariol

(2004). Relational aggression in middle school: Educational implications of developmental research. Journal of Early Adolescence, 24, 303–318. doi:10.1177/0272431604265681

36.

Zins

J. E.

Weissberg

R. P.

Wang

M. C.

Walberg

H. J.

(2004). Building academic success on social and emotional learning: What does the research say? New York, NY: Teachers College Press.