Abstract
This study examined relationships between social–emotional skills and involvement in bullying and violence among young adolescents from ethnically diverse, economically disadvantaged urban neighborhoods. Data were from 171 sixth- and seventh-grade students involved in a larger intervention study. Analyses examined relationships between social–emotional skills measures (intrapersonal skills, stress management skills, interpersonal skills) and involvement in violence, physical bullying, and relational aggression. Of social–emotional skills indicators, interpersonal skills and stress management skills demonstrated significant bivariate relationships with each of the bullying and violence outcomes. In multivariate models, greater interpersonal skills and greater stress management skills were significantly associated with lower odds of violence involvement. Greater stress management skills were also significantly associated with lower levels of physical bullying and relational aggression. Findings suggest that efforts to foster development of young adolescents’ social–emotional skills may, in turn, reduce their risk for involvement in bullying and violence.
Introduction
Youth violence is a devastating public health problem that takes a profound toll on young people, their families, their communities, and society. It consists of an array of behaviors that cause emotional harm, physical injury, and even death. In 2004, the homicide rate for U.S. young people between 10 and 24 years old was 8.4 per 100,000, which translated into an average of 15 murders of young people each day. Homicide is the second leading cause of death for young people aged 10 to 24 years (Centers for Disease Control and Prevention [CDC], 2010a).
Although death is the most severe consequence of violence, nonfatal injuries are far more common. In 2006, more than 720,000 young people 10 to 24 years old were treated in U.S. emergency departments for violence-related injuries (CDC, 2010a). A study of emergency department injuries involving adolescents younger than 15 years revealed that most events occurred in school and involved physical fights between friends (Mollen, Fein, Vu, Shofer, & Datner, 2003). According to the 2009 Youth Risk Behavior Survey, 31.5% of 9th- to 12th-grade students reported having been in a physical fight at least once during the 12 months preceding the survey; 3.8% had been injured severely enough to require treatment by a doctor or nurse (CDC, 2010b).
Although no community is immune to the effects of youth violence, low-income urban neighborhoods are disproportionately affected, and certain demographic groups of young people are overrepresented in violence statistics. Both fatal and nonfatal forms of youth violence disproportionately affect males and young people of color. In a study of urban, multiethnic, economically disadvantaged young adolescents, 8% of girls and 20% of boys reported having been stabbed, cut, or shot at during the past 12 months (Stueve, O’Donnell, & Link, 2001). Findings from the Youth Risk Behavior Survey (CDC, 2008) indicated that the prevalence of physical fighting during the preceding 12 months was higher among male high school students (44.4%) than females (26.5%). Physical fighting was highest among African American students (44.7%), followed by Hispanic (40.4%) and White (31.7%) students.
Falling along the continuum of violence, bullying involves behaviors used by one person to exert power over another (Nansel et al., 2001). More traditionally defined forms of bullying—such as name calling, hitting, or shoving—have long been recognized as problem behaviors. In a 2001 study with a nationally representative sample, nearly 30% of 6th- to 10th-grade students reported having been involved in bullying as perpetrator, victim, or both during the current school year (Nansel et al., 2001). A recent study of youth from ethnically diverse and economically disadvantaged urban neighborhoods revealed a high prevalence of bullying, with 62% of 10- to 13-year-old participants reporting being victims of peer bullying and 32% reporting perpetrating bullying behaviors in the past year (Bauer et al., 2006).
Relational aggression is a form of bullying in which the aggressor seeks to damage the victim’s friendships or feelings of peer group acceptance (Crick & Grotpeter, 1995). Examples include spreading rumors, social exclusion, and manipulative friendships. In their study of bullying behaviors in youth, Bauer et al. (2006) found that 33% of 10- to 13-year-olds experienced social exclusion over the past year. Relational aggression may result in adverse health outcomes for both victim and perpetrator. Victims often experience loneliness, depression, and difficulties with emotional and social adjustment (Crick & Grotpeter, 1995; Nansel et al., 2001), whereas perpetrators are at risk for substance use, behavior problems, and physical violence (Murray-Close, Ostov, & Crick, 2007; Smokowski & Kopasz, 2005). Furthermore, bullying puts both the victim and perpetrator at increased risk of suicidal ideation and/or behavior (Kim & Leventhal, 2008; Nansel et al., 2001).
Violence results from a complex interaction of individual and environmental influences, consistent with a social–ecological framework (Bronfenbrenner, 1979). Risk factors make it more likely that an individual will be involved in violence, whereas protective factors reduce the likelihood of violence involvement either directly or by mediating the effects of risk exposure (Arthur, Hawkins, Pollard, Catalano, & Baglioni, 2002). Individual-level risk factors for violence involvement during adolescence include a history of childhood aggressive behavior, hopelessness, high levels of emotional distress, difficulty coping with anger, and substance use (Bolland, McCallum, Lian, Bailey, & Rowan, 2001; Botvin, Griffin, & Nichols, 2006; Resnick, Ireland, & Borowsky, 2004). Individual-level factors that appear to be protective against youth violence include empathy, positive social orientation, and involvement in prosocial activities (CDC, 2008; McMahon & Washburn, 2003; Resnick et al., 2004).
Social–emotional intelligence (SEI) is a concept that has gained recognition as a potential protective factor related to health outcomes including violence. Characterized as a developmental construct that can increase with age, SEI has been defined as the ability to understand and competently express oneself, to understand and connect with others, and to effectively cope with the stress and challenges of daily life (Bar-On & Parker, 2000). Prior research has found inverse relationships between SEI indicators and known correlates of bullying and violence, including school discipline referrals (Peters, Kranzler, & Rossen, 2009) and hopelessness (Ciarrochi, Wilson, Deane, & Rickwood, 2003). However, there is a noticeable lack of research exploring direct relationships between young adolescents’ social and emotional skills and their involvement in bullying and violent behaviors.
The current study examined relationships between social and emotional skills, violence involvement, and bullying perpetration among a group of urban, ethnically diverse, and economically disadvantaged young adolescents. We hypothesized that (a) greater social and emotional skills would be associated with less involvement in violence and that (b) greater social and emotional skills would be associated with lower levels of perpetrating physical and relational bullying.
Method
The Lead Peace Study
Data for the current study were from the Lead Peace demonstration study. Lead Peace was a school-based service learning program for urban sixth- to eighth-grade students that aimed to reduce violence involvement and school failure by promoting specific skills, opportunities, and supports in students’ lives. Begun in 2006, the demonstration study involved five K-8 Minneapolis public schools assigned to Lead Peace program and comparison conditions. In consultation with the Minneapolis Public School District (MPS), schools with similar ethnically diverse and economically disadvantaged student bodies were selected for the Lead Peace study. The study cohort included students from the eighth-grade class of 2009 at each of these schools. Students were surveyed at the start of sixth grade (T1), end of sixth/start of seventh grade (T2), 1 end of seventh grade (T3), and end of eighth grade (T4). All study protocols were approved by the MPS Research Department and the University of Minnesota Institutional Review Board. Details regarding the Lead Peace service learning program and evaluation are found elsewhere (Secor-Turner, Sieving, Widome, Plowman, & Vanden Berk, 2010; Sieving & Widome, 2008; Widome, Sieving, Harpin, & Hearst, 2008).
Lead Peace study schools are located in neighborhoods that are among the most economically and socially troubled in Minneapolis. Between 30% and 42% of families with children younger than 18 years in these neighborhoods lived in poverty, compared with 12% of families citywide (City of Minneapolis, 2010). In the 2006-2007 school year, 92% to 96% of students from these schools received free/reduced lunch compared with a school district average of 72%. In 2006, half of all homicides citywide occurred in these neighborhoods (City of Minneapolis, Department of Health and Family Support, 2008).
Study Sample
This study employed student self-report data from the T2 Lead Peace student survey. This survey round was completed by a total of 171 students, including 95 students who completed the survey at the end of sixth grade and 76 students who completed the survey at the start of seventh grade. The sample’s demographic characteristics are noted in Table 1. The proportion of males and females was approximately equal (48.0% male). Students’ mean age was 12.1 years. The sample was racially and ethnically diverse.
Sample Descriptors
Students surveyed in the spring of 2007 took the survey at the end of their sixth-grade year whereas students surveyed in the fall of 2007 took the survey at the beginning of their seventh-grade year.
Violence involvement is defined as reports of any of the following in the past year: involved in more than one physical fight, hurt someone badly in a fight, or used or threatened to use a weapon.
Measures
Study outcome variables included Lead Peace survey measures of violence involvement, physical bullying, and relational aggression.
Violence involvement
Three items from the National Longitudinal Study of Adolescent Health survey (Resnick et al., 2004) assessed physical fighting, using or threatening to use a weapon, and hurting someone badly enough to require medical attention in the past year. Responses to each item included “Never,” “Once,” and “More than once.” For this study, we created a dichotomous violence involvement measure contrasting students who reported any of the following behaviors in the past year: (a) being in a physical fight more than once, (b) hurting someone badly enough to require bandages or medical attention, or (c) using or threatening to use a weapon to students who reported minimal or no violence involvement.
Physical bullying perpetration
A three-item scale (α = .80) assessed how often in the past week students threatened to hit or hurt another student, asked someone to fight, or hit, shoved, or pushed someone. Responses to each item included “Never,” “Once,” and “More than once.”
Relational aggression perpetration
An eight-item scale (α = .82) assessed how often in the past week students instigated aggressive interpersonal behaviors. Sample items included leaving someone out on purpose, spreading rumors or gossip about someone, and saying something hurtful to someone in an e-mail or on the Internet. Responses to each item included “Never,” “Once,” and “More than once.”
Students’ social and emotional skills were assessed with items from intrapersonal, interpersonal, and stress management scales of the Bar-On Emotional Quotient Inventory Youth Version (Bar-On & Parker, 2000). Responses to each item included “NO!” “no,” “yes,” and “YES!”
Intrapersonal skills, a two-item composite (r = .55), assessed youth’s abilities to express their own emotions: It is easy to tell people how I feel, I can talk easily about my feelings.
Interpersonal skills, a five-item scale (α = .58), assessed empathy, social responsibility, and interpersonal relationship skills: I try not to hurt other people’s feelings, Having friends is important, I am able to respect others, I like most people I meet, I care what happens to other people.
Stress management skills, a four-item scale (α = .69), assessed stress tolerance and impulse control: I get too upset about things, I can stay calm when I’m upset, I fight with people a lot, It is hard to control my anger.
Covariates
We explored potential confounders including gender, age, race/ethnicity, and timing of survey (end of sixth grade vs. start of seventh grade).
Analysis Strategy
Preliminary analysis found no evidence of intervention effects on study outcome variables at this interim point in the Lead Peace intervention. Therefore, program and comparison groups were pooled, and the full sample of students was used for this etiological analysis. Initial analyses examined bivariate relationships between study independent variables and individual outcomes, using t tests (for violence involvement) and Pearson’s correlation coefficients (for physical bullying and relational aggression). Next, multivariate logistic regression models were used to predict violence involvement and multivariate linear regression models were used to predict perpetration of physical bullying and relational aggression. Only those independent variables with significant bivariate relationships with the outcome being modeled (p < .05) were included in multivariate analyses. We estimated the regression models in Stata 10.0 using robust standard errors to account for the clustering of students within schools (StataCorp, 2007).
Results
Participant Characteristics
Sample descriptors are found in Table 1. More than one third (36.3%) of students reported violence involvement in the past year, defined as having been involved in more than one physical fight, or having used or threatened to use a weapon, or hurting someone badly enough to require bandages or medical attention. On a scale ranging from 0 (no physical bullying) to 6 (high physical bullying), the average physical bullying perpetration score was 2.05. On a scale ranging from 0 (no relational aggression) to 16 (high relational aggression), the average relational aggression perpetration score was 4.73.
Bivariate Findings
Mean values of social and emotional skills measures among students reporting high and low violence involvement are noted in Table 2. Average interpersonal skills scores were significantly lower among students with high violence involvement than among students with low violence involvement. Students with high violence involvement also reported lower stress management skills on average than students with low violence involvement. The relationship between intrapersonal skills and violence involvement was nonsignificant.
Mean Values and Standard Deviations of Social and Emotional Skills Scales by Level of Violence Involvement: Bivariate Relationships
Note: Violence involvement is defined as reports of any of the following in the past year: involved in more than one physical fight, hurt someone badly in a fight, or used or threatened to use a weapon.
p < .001 for t tests.
Examination of bivariate relationships between social and emotional skills and bullying perpetration indicated that interpersonal skills were inversely related to perpetration of physical bullying (r = −0.23, p < .01) and relational aggression (r = −0.30, p < .001). Similarly, stress management skills were inversely related to perpetration of physical bullying (r = −0.38, p < .001) and relational aggression (r = −0.43, p < .001). Relationships between intrapersonal skills and bullying perpetration measures were nonsignificant.
In addition, the bivariate association between stress management skills and interpersonal skills (the two measures of social and emotional skills to be included in multivariate models) was examined. The magnitude of this correlation (r = 0.21, p < .01) indicated that these measures represented distinct skill sets.
Multivariate Findings
The multivariate model predicting violence is seen in Table 3. In this model, both interpersonal and stress management skills functioned as independent protective factors associated with significantly lower odds of violence involvement. Each 1-unit increase in interpersonal skills decreased the odds of violence involvement by 82%. Furthermore, each 1-unit increase in stress management skills decreased the odds of violence involvement by 47%.
Effects of Social and Emotional Skills on Violence Involvement: Multivariate Logistic Regression Model a
Model controlled for student gender, race/ethnicity, and timing of survey.
p < .01. **p < .001.
The multivariate models predicting perpetration of physical bullying and relational aggression are shown in Table 4. In both models, stress management skills functioned as a protective factor associated with significantly lower risk of instigating physical bullying and relational aggression. Interpersonal skills were not a significant protective factor for either bullying perpetration outcome, after controlling for stress management and covariates.
Effects of Social and Emotional Skills on Bullying Outcomes: Multivariate Linear Regression Models a
Models controlled for student gender, race/ethnicity, and timing of survey.
p < .01.
Discussion
The purpose of this study was to examine relationships between social–emotional skills, violence involvement, and bullying perpetration among young adolescents from urban, low-income neighborhoods. Results address a gap in understanding factors that protect young people from these forms of violence. Findings supported our first hypothesis that greater social and emotional skills would be associated with less violence involvement. Multivariate regression models indicated that interpersonal skills and stress management skills were both associated with significantly lower odds of violence involvement. Evidence also supported our second hypothesis that greater social and emotional skills would be associated with lower levels of bullying perpetration. In multivariate regression models, higher levels of stress management skills were associated with lower levels of instigating both physical bullying and relational aggression.
Findings from this study and several previous studies support the assertion that social and emotional skills are important protective factors against youth violence involvement. More specifically, results from this study corroborate earlier findings of Sieving and Widome (2008) in which high levels of interpersonal skills and stress management skills were associated with less violence involvement among sixth-grade students from low-income urban neighborhoods. In a study of African American fifth graders from low-income urban schools, Jagers et al. (2007) found that empathy was associated with violence avoidance self-efficacy beliefs, which, in turn, resulted in less violent behavior. Jagers et al.’s measure of empathy was very similar to the interpersonal skills measure employed in the current study.
Although both interpersonal skills and stress management skills had significant bivariate relationships with perpetration of physical bullying and relational aggression, only stress management skills was a significant predictor of these forms of bullying perpetration in multivariate models. By definition, stress management skills include impulse control and the ability to withstand stressful situations (Bar-On & Parker, 2000). Previous research has linked impulsivity and low levels of tolerance for stress with bullying behaviors (Smokowski & Kopasz, 2005). Thus, our findings are consistent with existing literature and suggest that fostering the development of stress management skills has the potential to protect young adolescents from perpetrating physical bullying and relational aggression.
Among this group of young adolescents, intrapersonal skills were not significantly associated with violence involvement, physical bullying, or relational aggression. It is possible that these nonsignificant relationships are an artifact of an imprecise measure of intrapersonal skills. However, the low average intrapersonal skills score in this sample relative to other SEI measures and the lack of statistically significant relationships with violence outcomes may mean that developmentally, this group of young adolescents has yet to establish a repertoire of intrapersonal skills that are protective against bullying and violence.
Although limited research to date has examined direct connections between adolescents’ social and emotional skills and their involvement in violence, it is clear that social and affective factors play a role in the phenomenon of youth violence. Prior research has identified individual-level risk factors for youth violence, including high levels of emotional distress, prior suicide attempt, poor mental health, and difficulty coping with anger (Botvin et al., 2006; Resnick et al., 2004). Conversely, positive social orientation and involvement in social activities have been found to be protective against violence and bullying (CDC, 2008). Furthermore, intervention studies that included a focus on development of social and emotional skills as a means to prevent violence involvement have met with success (e.g., Botvin et al., 2006; Hawkins, Kosterman, Catalano, Hill, & Abbott, 2005; McMahon & Washburn, 2003).
Social and emotional skills are typically learned from observing others in one’s immediate social environment. Young people involved in social relationships where violence is normative and violent behavior is routinely modeled may be less likely to develop healthy social and emotional skills, and more likely to be involved in bullying and violence (Kethineni, Blimling, Bozarth, & Gaines, 2004). Garbarino (1995) asserted that the chronic stress experienced by young people living in impoverished and violent communities often results in a lack of a positive future orientation and cultivates rage, depression, and violent behavior. A combination of factors, including limited social and emotional skills, living in contexts where violence is normative and support for alternative prosocial behaviors is limited, and poverty, may all predispose adolescents to develop unhealthy ways of managing stress, including the use of bullying and violence (Molnar, Cerda, Roberts, & Buka, 2008; Putnam & Feldstein, 2003).
Limitations
Several limitations should be noted. First, a relatively small sample size limits our ability to examine study hypotheses separately for boys and girls. Second, our interpersonal skills measure has a marginal level of internal consistency, which may act to attenuate true relationships between this construct and study outcomes (Sieving & Shrier, 2009). Third, honest self-disclosure of bullying and violent behaviors by young adolescents may be an issue. However, the outcome measures employed in this study have been widely used in research with this age-group (CDC, 2008). Reviewing the measurement literature, Brener, Billy, and Grady (2003) concluded that although adolescents’ self-reports of violent behaviors may be affected by cognitive and environmental factors, these factors do not pose a significant enough threat to the validity of self-reports to forego their use. Finally, findings from this study, conducted with an urban, multiethnic, low-income group of young adolescents, cannot be generalized to other social groups. Further research is needed to examine relationships between social and emotional skills, bullying, and violence among young people from different economic strata, older adolescents, and rural and suburban youth.
Conclusions
The cross-sectional associations found in this study should be examined in future longitudinal research. Longitudinal data could be used to explore whether high levels of social and emotional skills during early adolescence are predictive of lower levels of violence involvement and bullying perpetration during middle or later adolescence. Measurement of social and emotional skills at multiple points during adolescence would allow for empirical modeling of the development of these skills, as well as the underlying construct of SEI, over the course of the adolescent years.
Study findings provide preliminary support for intervention activities targeting young adolescents’ social and emotional skills as a means for reducing bullying and violence. The potential protective role that stress management skills may play, in particular, warrants the development, implementation, and evaluation of interventions that specifically address young adolescents’ impulse control and stress tolerance. Although interventions that emphasize the development of social and emotional skills are effective in reducing aggression and violence (e.g., Hawkins et al., 2005), few studies have explored whether intervention effects are mediated by improvements in social and emotional skills. One notable exception is the Aban Aya Youth Project, a multicomponent intervention tested with fifth- to eighth-grade students in urban public schools. This intervention increased students’ empathy, which, in turn, led to reduced risk for violent behavior (Jagers et al., 2007).
Although the future of young people living in environments where violence is normative appears bleak at first glance, identifying factors that may mitigate risk is imperative in the development of interventions to prevent or reduce violence among these young people. Because the presence of protective factors is known to reduce the probability of violence involvement even when multiple risk factors are present, efforts employing a dual strategy of building known protective factors while addressing risks may hold the most promise in preventing youth violence (Resnick et al., 2004). Furthermore, identifying specific protective factors during early adolescence is crucial to the development of effective prevention strategies during this time period (Multisite Violence Prevention Project, 2008). Findings from this study support the notion that efforts to build young adolescents’ stress management and interpersonal skills can play an important role in buffering them from involvement in violence and bullying.