SCHOOL-BASED PREVENTIVE INTERVENTIONS

As indicated in the previous chapter, bullying is a particular form of aggressive behavior, but it also is part of a broader set of problem behaviors seen within schools. Programs designed to prevent both bullying and these other problem behaviors have been studied, and the results can be applied either narrowly to bullying or else more broadly, observed Catherine Bradshaw, a professor and the associate dean for research and faculty development at the University of Virginia Curry School of Education.

Ttofi and Farrington (2011) carried out a meta-analysis of 53 rigorous evaluations and randomized trials of programs aimed at preventing bullying, two-thirds of which were conducted outside of the United States and Canada. They found overall a 23 percent decrease in the perpetration of bullying and a 20 percent decrease in victimization. The most effective elements of programs that they identified were:

• use of parent training activities, meetings, and information
• high levels of playground supervision
• use of consistent disciplinary methods
• classroom management strategies
• classroom and school-wide rules related to bullying
• training of teachers (including aspects of that training and the amount of time and intensity of training)
• multicomponent prevention approaches

Some caveats should be noted concerning this meta-analysis, Bradshaw said. First, the effects generally were stronger in the nonrandomized controlled trial designs. Impacts were also larger among older children, and the programs generally were more effective in European than in North American sites.

Some observers, such as Merrell et al. (2008), have argued that there are relatively few effective universal bullying prevention programs. To examine this issue, Bradshaw and her colleagues have been looking more broadly at violence prevention efforts. School-wide efforts that involve all school staff and are implemented across all school settings show the most promise in reducing bullying and rejection, she said. For example, a model called KiVa developed in Finland shares some elements of school-wide approaches (Salmivalli et al., 2011). It uses a videogame that students can use to practice different strategies in resolving bullying situations. That model is now being studied to determine whether cultural and contextual changes need to be made to it in order for it to be imported successfully into the United States, Bradshaw said.

Universal school-wide prevention models that are broadly focused on violence and disruptive behaviors may also affect bullying, she said. For example, frameworks based on social–emotional learning have demonstrated effectiveness in some cases, although not in others. One example of a successful program is Second Step, which was originally developed for elementary school but more recently has been extended to middle school; it has had some promising effects on precursors to bullying behavior. Similarly, the PATHS (Promoting Alternative Thinking Strategies) program has been shown to affect a wide range of behavioral and mental health outcomes, Bradshaw said.

Programs that focus on classroom management may also be effective in reducing problem behaviors, including bullying, Bradshaw said. For example, studies in Baltimore by Sheppard Kellam and Nick Ialongo have shown that just 1 year of implementation of a classroom management strategy called the Good Behavior Game can produce long-term effects across a range of outcomes, including substance use, violent and aggressive behaviors, and bullying (Kellam et al., 2011). This research has also found effects on academic performance, high school completion, and the number of students who need special education services. “That is one model of classroom management that we could think about more broadly as it relates to impacting bullying and other outcomes,” Bradshaw said.

A public health approach to prevention is one way to build a multi-tiered system of support, Bradshaw said. This approach combines universal prevention for all students with indicated or intensive intervention for a few students and selective or targeted interventions for some additional students. Research has demonstrated that these approaches can have significant impacts on the school environment, Bradshaw said, including significant improvements in school climate and systems changes that are sustainable over multiple years. For example, data from randomized, controlled trials of a model known as the Positive Behavior Support Framework found significant reductions of suspensions as well as bullying behavior, she said.

Bradshaw concluded by recommending that bullying prevention programs contain the following core elements:

• Teacher training
• Activities for students
• Parent activities
• Multi-component programs
• School-wide scope
• Continuum of positive supports
• Data-driven process

Bradshaw also touched on several approaches that are not recommended. One example is a zero-tolerance policy in which students are automatically suspended for bullying. Another is to group students who bully together, because that can make their aggression worse rather than better, she said. Brief assemblies or 1-day awareness campaigns have little effect in terms of sustained outcomes for youth. Conflict resolution and peer mediation can be effective for other forms of aggressive behavior but raise concerns in the context of bullying. Finally, focusing excessively on issues of mental health and suicide can contribute to contagion processes rather than addressing the problem directly, Bradshaw said.

In multi-component programs, it is important to provide training to classroom teachers concerning classroom management and what to do in bullying situations. “It sounds like a no-brainer,” she said, “but there are programs that don't actually provide any training to teachers.” Similarly, while parent engagement is a clear challenge for nearly all schools, school–home communication about bullying is particularly important in order to let parents and caregivers know about the strategies and lessons being taught in school, she said. Multi-component programs are characterized by a continuum of responses, so that suspending a student is not the only option, but rather an array of positively oriented activities exists from which to choose, Bradshaw said.

In response to a question about the use of conflict-resolution centers to provide free confidential mediations to people in the community who need it, including students and teachers, Bradshaw responded that schools need to have many tools in their toolkits. The same tools used to deal with conflicts or fights may not be useful in a bullying situation. Many youths who are victimized by their peers do not want to sit across the table and talk with their aggressors, even in a well-structured environment. But many other strategies exist, such as restorative practices where youth can make up for past transgressions. Systematic research is needed to determine which of these interventions work best in different circumstances, she said.

In response to a later question about playground supervision and whether the lack of unstructured play may be inhibiting the development of conflict-resolution skills, Bradshaw pointed to the importance of youths learning how to resolve small day-to-day conflicts as practice for larger conflicts. “That way when it does get big, you have some kind of skill set to draw upon.” Similarly, social–emotional curricula seek to give students strategies and skills for learning how to label and regulate their own emotions, she said.

SCHOOL CLIMATE AND BULLYING

Schools are more than the sum of the individuals inhabiting those schools, said Denise Gottfredson, professor in the Department of Criminal Justice and Criminology at the University of Maryland. Each school has its own personality, and that personality influences student outcomes, including the amount and nature of bullying in a school.

Gottfredson pointed to research dating back to the 1970s that demonstrated the importance of three aspects of the school climate in predicting a variety of problem behaviors among students:

• Student-to-teacher ratio
• School norms
• Consistent discipline (Cook et al., 2010b)

Regarding student-to-teacher ratios, in the 1970s the Safe School Study, which looked at a national sample of 642 secondary schools, identified several school climate predictors of victimization, one of which was large schools (Gottfredson and Gottfredson, 1985). More recently, Gottfredson and DiPietro (2011) used data from the National Study of Delinquency Prevention in Schools to try to replicate that finding and also to explore other measures related to school size. They found that higher student-to-teacher ratios and a higher number of different students taught by typical teachers increased student victimization, replicating the findings from the Safe School Study. However, they found that higher overall student enrollment was related to lower victimization rates, Gottfredson said, suggesting that it is not the total number of students in a school that matters but the ratio of students to teachers and the ways that students are organized for instruction. Similarly, a recent study of 95 elementary schools found that students in schools with higher student-to-teacher ratios reported greater frequencies of bully victimization and reduced perceptions of safety (Bradshaw et al., 2009).

The second factor involves a sense of community or of belonging in a school, part of which involves a shared belief about acceptable behaviors. In their study of schools, Gottfredson and DiPietro (2011) found that one reason why students in schools with higher ratios of students to teachers experience more victimization is that these schools have less consensus regarding norms for behavior. Gottfredson's colleague Allison Payne, using the National Study Data, further demonstrated the importance of a communal social organization, in which students and adults know, care about, and support one another, have common goals and a sense of shared purpose, and actively contribute and feel personally committed to the school. Payne found that a communal social organization reduces student delinquency and that the effects of this factor on delinquency are mediated by student bonding (Payne et al., 2003). Students in schools with communal social organizations are more attached to school, and that attachment serves to inhibit their offending behavior. Similarly, a recent study of nearly 300 schools found that students in schools with high levels of perceived teacher and school staff support are more willing to seek help for bullying and aggressive behavior (Eliot et al., 2010).

The third factor Gottfredson highlighted is consistent discipline management that supports school norms. A study done in the 1990s by the U.S. Department of Education and the National Institute of Justice found that consistent discipline management was related to lower levels of student victimization and delinquency (Gottfredson et al., 2005). That study also found that a more positive social climate—as measured by organizational focus, teacher morale, and strong administrator leadership—was related to lower levels of teacher victimization. A more recent analysis similarly demonstrated that teacher support and firm and consistent discipline management are related to lower levels of both bullying and other forms of victimization (Gregory et al., 2010).

To her discussion of these three factors, Gottfredson added a brief description of a study undertaken to understand what factors might have led to the spate of school shootings experienced in the 1990s. This study sent teams of ethnographers into six different communities that had experienced lethal school shootings to interview people and to collect records relevant to the incidents. The study is relevant to bullying because many of the shooters reported feeling bullied at school, Gottfredson said. The most relevant finding from this study for purposes of understanding bullying was that there was a large gulf between youth and adults in these communities (IOM and NRC, 2003). The study found that the adults had a very poor understanding of the children's experiences. The study also found that shooters reported that they felt they had nowhere to turn, that they were intensely concerned about their status and protecting themselves, and that specific warnings had been given and missed. All of this research demonstrates the importance of meaningful relationships between students and the adults in a school, Gottfredson said.

Gottfredson summarized the research on the several interventions to improve the aspects of school climate that she had discussed earlier. Although the quality of the research on these interventions is generally low by current standards (e.g., most of the studies were not randomized controlled trials [ERCTs]), these trials nevertheless provide promising results for reasonable interventions that could be tested more rigorously today, she said.

BULLYING AND SCHOOL SAFETY

School policies on bullying are not carried out in a vacuum, observed Dewey Cornell, a forensic clinical psychologist and the Linda Bunker Professor of Education at the University of Virginia Curry School of Education. The policies are implemented in the context of broader school safety concerns that can be quite pressing and that can impede efforts to implement bullying prevention programs.

Since the 1990s broad school safety policies have been shaped extensively by fears of school violence, Cornell said. Today, school administrators have to consider building security, school shooting drills, zero-tolerance practices, and the pressure for high-stakes testing, all of which they see as competing priorities.

Recent school shootings have brought bullying to national attention because of a perceived link between bullying and these shootings, which has provided an important context for bullying prevention programs, Cornell said. A study of school shootings by the U.S. Secret Service and the U.S. Department of Education (2002) found that the majority of attackers felt bullied or persecuted. School shootings have seemingly transformed school safety and discipline, primarily through the expansion of zero-tolerance policies to not only firearms but also all sorts of toys and otherwise innocuous behaviors (Cornell, 2006). Today, schools are suspending students for much less serious behaviors than in the past, raising concerns about possible discriminatory practices, Cornell said. In general, events such as the Sandy Hook shootings have exposed people throughout society to terrible images, thoughts, and fears, which have strongly influenced school practices, he said. Since those shootings, some schools have rushed to institute security measures such as bulletproof building entrances, metal detectors, x-ray screening, cameras, and increased school security personnel, Cornell said. Other responses that have been discussed include giving teachers firearms training or putting additional locks on doors to keep out intruders with a gun, Cornell said.

These steps can cost millions of dollars, which can deprive schools of resources that could be allocated to anti-bullying programs and counseling services (DeAngelis et al., 2011). “It is a terrible dilemma to have to choose between security and anti-bullying,” Cornell said. These security measures also have an impact on school climate, he said. It is unclear how these drills affect children's sense of security and their impact on school climate.

Increases in School Safety

The new concern with school safety may be an overreaction, Cornell said. Homicides in which school-age youths are the victims rarely occur at school (see Figure 4-1). Instead, he said, most homicides occur in residences and at other locations, and school violence has actually declined over time. According to data from the Centers for Disease Control and Prevention, 98 to 99 percent of homicides of school-age children occur outside of school (CDC, 2014a,b; Modzeleski et al., 2008). But school shootings are so traumatic that they skew perceptions of school safety and convince the public and policy makers that there are urgent needs for security measures, Cornell said.

FIGURE 4-1

Homicides at school are a tiny fraction of all homicides of youth ages 5 through 18. SOURCE: Cornell presentation, 2014. Data from CDC, 2014a,.

Overall, school-associated violent deaths—not only among students, but also staff members and adults as well—have been declining, Cornell said. One recent study estimated that the average school in the United States can expect a student homicide about once every 6,000 years (Borum et al., 2010). School violence has also declined, with aggravated assault, robbery, and forcible rape dropping substantially since the 1990s, Cornell added.

School policies on bullying need to be disentangled from concerns about school shootings and school safety, Cornell said. “Those policies need to be based on the recognized harms associated with bullying, not on the fear of school shootings,” he said. Instead of thinking about security as the first line of defense against a school shooting or other serious act of violence, he said, the focus should be on prevention and mental health services in schools, threat assessment, and bullying prevention programs.

Almost all of the states have passed legislation related to bullying (see Chapter 9), but the legislation varies greatly from state to state, Cornell said. Usually, legislatures direct schools to come up with their own policies about bullying, sometimes with guidance to address things like defining and prohibiting bullying and mandating staff reporting for bullying, Cornell said.

Policy recommendations for schools also exist. For example, he suggested the following guidelines:

• Clarify the definition of bullying for the school community (students, staff, parents).
• Detect and intervene to stop bullying, but do not use zero tolerance.¹
• Use valid measures to assess bullying.
• Use evidence-based strategies to reduce bullying and intervene with victims, perpetrators, and bystanders.
• Recognize when bullying is criminal or discriminatory.

School policies also face an array of challenges, Cornell said. First, state definitions of bullying tend to be very inclusive. They include almost any form of intentional peer aggression and rarely mention the power imbalances that are featured in most research definitions of bullying. As a result, he said, the more narrow definitions of bullying used in evidence-based programs tend not to line up with the definitions of bullying used in schools.

Second, students are often unwilling to report bullying, Cornell said. A more supportive school climate and change in peer culture could help with this code of silence, but schools also need more systematic and effective ways to identify victims, he said, including tip lines, peer nomination surveys, and active staff monitoring and inquiry.

Third, Cornell said, schools tend to rely on anonymous self-report surveys, but anonymous surveys cannot be validated against independent criteria. Furthermore, self-reports depend on student knowledge and understanding of the complex concept of bullying (Cornell and Cole, 2011).

Fourth, many bullying prevention programs have little or no scientific support, Cornell said, so policies need to encourage greater use of evidence-based programs. Unfortunately, he said, some legislators mistrust the idea of relying on scientific research rather than on deeply held values, and the programs most often used in schools often feature motivational speakers and 1-day programs that are of unknown effectiveness rather than programs that have been studied by researchers and have empirical support for their use.

Fifth, bullying and harassment are often confused and used interchangeably. Harassment has legal significance, whereas bullying does not, Cornell noted, and harassment does not require a power imbalance.

The U.S. Department of Education (2013) has made what Cornell called “excellent recommendations” to address the bullying of students with disabilities, which can become severe enough to deny such students a free and appropriate public education. In fact, Cornell suggested that these recommendations could apply equally well for all students. These recommendations are to:

• Use a comprehensive multi-tiered behavioral framework
• Implement clear policies on bullying
• Collect data on bullying, such as frequency, types, and location
• of bullying behavior, and adult and peer responses
• Notify parents or guardians of both the student who was the
• target of the bullying behavior and the student who engaged in the bullying behavior of any report of bullying that directly relates to their child when bullying occurs
• Address ongoing concerns about a student's behavior that is not safe, responsible, or consistent with established school expectations through specific feedback on behavior, increased adult engagement, or more focused skills instruction
• Sustain prevention efforts

These recommendations were created partly in response to concerns about harassment and legal action, but they also reflect lawsuits in which parents in the community have successfully sued schools for victimization of their children, Cornell said.

In concluding, Cornell noted that federal protections from bullying are largely limited and piecemeal. They are concerned with bullying when it targets someone based on sex, race, color, national origin, or disability status, he said, but “all students are entitled to protection from unlawful discrimination and harassment.” The overarching concern should be the protection of all children and youth from bullying, he concluded.

DIFFERENCES BETWEEN THE UNITED STATES AND OTHER COUNTRIES

During the discussion period, an interesting exchange centered on why some bullying prevention programs are less effective in the United States than in other countries, as Bradshaw had mentioned in her presentation. U.S. schools emphasize testing and high-stakes accountability, Bradshaw observed, which puts tight constraints on how the schools can spend their time. She also pointed to cultural differences between societies in the levels of violence and exposure to violence in the media. Such differences also exist among U.S. schools, she said. For example, many bullying prevention programs have been studied in suburban communities rather than in urban settings, which may influence the effect sizes of the interventions. Students in urban communities tend not to even use the word “bullying,” and they may have strong cultural norms against reporting on others, Bradshaw said.

Given this variability, adapting interventions to the particular culture is critical, Bradshaw said. “We need to work with our community partners to figure out adaptation that is strategic and appropriate,” she said. “We have to make sure that we are thinking culturally and contextually about taking models not only from Europe but even within the United States. What might work in a suburban community may not work very well in inner-city Chicago or DC or Baltimore.”

Gottfredson added that another possible explanation for the lesser effectiveness of bullying prevention programs in the United States is the quality of the implementation of the programs. When she and her colleagues collected information about the number of different programs that schools are implementing, she said, “we were astounded to see that the average school is implementing 17 different programs at the same time. When we looked at the quality indicators, they were very extremely low.…I think that is one possible reason why we see differences in the U.S.”

Footnotes

1

Zero tolerance refers to “a philosophy or policy that mandates the application of predetermined consequences, most often severe and punitive in nature, that are intended to be applied regardless of the gravity of behavior, mitigating circumstances, or situational context” (APA Zero Tolerance Task Force, 2008, p. 852).