Julide S. Peace
St. Cloud State University/Autism Partnership
Note: This article was originally published in the July 2016 issue of Behavior Analysis Quarterly, Vol. 2, (3). Please consult the original article for material not included here.
In a national survey, 30% of children in middle school reported being bullied, bullying others or being both a bully and a victim (Nansel, Overpeck, Pilla, Ruan, Simons-Morton, & Scheidt, 2001). The prevalence of bullying is important because bullies and those who are bullied demonstrate poorer psychosocial adjustment than youth who are not identified as bullies or victims including cigarette and alcohol use, poor peer relationships and self-reported loneliness (Nansel et al. 2001). Research is showing that bullying peers in school is correlated with criminal offending in the future even when controlling for other risk factors for criminal behavior (Ttofi and Farrington, 2011) and victims of bullying tend to have a higher than normal risk for depression and suicide (Sourander, Helstelae, Helenius & Piha, 2000). While bullying and victimization have not been identified as causal factors for criminal offending or suicidal ideation, the correlation alone should move us to consider what variables account for these relationships.
While there has been a wealth of research conducted on bullying and the outcome of bullying prevention programs, the results are often inconclusive and discouraging. Systematic reviews and meta-analyses of bullying prevention programs indicate that some studies on prevention and intervention have positive effects, but they are very small, and many programs show mixed and sometimes negative effects. Of the 14 whole schools studies reviewed by Smith, Schneider, Smith and Ananiadou (2004), the effects fell almost exclusively in the categories of small, negligible or negative. Only 7% of studies had effects that could be categorized as medium and none had large effects. In another meta-evaluation of 12 interventions conducted in different countries, bully prevention programs were generally found to have had significant but relatively small effects in reducing the proportion of children being bullied, and little or no effect in the reduction of bullying perpetration (Rigby, 2002 as cited in Rigby 2004). Merrell and Isava (2008) conducted a meta-analysis where they looked at 16 studies on bully prevention programs in schools. None of the studies reviewed showed a reduction in observed bullying behaviour. Many programs showed a change in perception of bullying behaviour but no actual behaviour change was documented. And most recently, Evans, Fraser and Cotter (2014) conducted a systematic review of school-based anti-bullying programs and found only 50% of those that looked at the perpetration of bullying behaviour saw program effects that could be considered significant, and only 67% that looked at victimization saw significant positive effects.
Not surprisingly, even when prevention programs do appear to have some positive effects, there is a lack of consensus as to why bullying occurs and what elements of change result in a positive intervention outcome. Farrington & Ttofi (2009) conducted a meta-analysis including research on anti-bullying programs from all over the world. Their results indicated that effective components of anti-bullying programs included: the presence of parent and teacher training, use of clearly defined classroom disciplinary methods, implementation of whole school anti-bullying policy, and use of instructional videos. In statistical analyses, these elements were correlated with reduced bullying and victimization. However, Evans et al. (2014), when using the same framework as Farrington & Ttofi and looking at results of school-wide interventions did not find the same results. In the meta-analysis conducted by Smith et al. (2004), the authors concluded that higher numbers of components included in the anti-bullying program correlated with a higher probability of a positive effect. This should not be surprising to us, because at this time, we have not identified the causal variables that result in a higher probability of bullying, so in essence, the more shots we take in the dark, the bigger the chance we have of hitting something.
There may be several reasons for the lack of significant positive outcomes, the lack of overall consistency in outcomes and the inconsistency in the results of the synthesis of data. These reasons include issues around the validity of the studies due to ambiguous operational definitions and subjective measures of bullying, which further confound the efforts to identify the functionally related variables that can be manipulated to decrease bullying. Behaviour analysis however, does have the technology to solve these problems. A behavioural perspective would allow for a more objective, yet functional definition of bullying, as well as a more scientific method of assessment. This could be the first step in identifying causal variables for bullying, which is required if we are to develop successful interventions to prevent or reduce bullying.
The definitions of bullying that are used in many published studies are problematic as significant disparity exists among studies as to how bullying is defined. Some experimenters, with fore-thought, have not defined what they mean by “bullying”, even on self-report measures, allowing participants to decide whether actions they have perpetrated, or been victims of, can be interpreted as bullying (Evans et al. 2014). Mostly, however, the definitions of bullying given to study participants involve complex explanations that use terms that cannot be observed and require subjective inferences on the behalf of the reporter. Table 1 provides a list of definitions of bullying provided by various published researchers. While there are numerous similarities, on closer analysis, this consistency does not prove helpful in defining the term. Consider the terms “attack”, “harassment,” “intimidation,” “dominating” “upsetting (someone)” “behaving negatively” and whether or not, by using them interchangeably, we are measuring the same class of behaviours. It is impossible to say, as it is unlikely that any of these words would mean exactly the same thing to two different individuals, which leads to a problem in construct validity. A second problematic component of many definitions includes the added ambiguity regarding how many acts constitute “repeated” acts and within what time period these acts must occur to be considered “repeated”. Some definitions include a qualifier that an act is not defined as bullying if it happens “occasionally” yet no distinction is provided between the terms “occasionally” and “repeatedly”. A third problematic component of numerous definitions is that they require the reporter to judge the relative power of the bully and the victim as a means to qualify an aggressive act as “bullying”. Some definitions imply there needs to be a physical power imbalance (Olweus, 1996) whereas others refer to the requirement of some form of domination taking place where a more powerful child oppresses a less powerful one (Farrington, 1993). Yet, other definitions require the victim to respond in a submissive manner (Pelligrini & Long, 2002) to be considered bullied. How to identify the difference in power, beyond physical size, is not indicated, and must be inferred by the reporter in a subjective, and therefore, unscientific manner. Finally, and possibly the most problematic element of the definition of bullying, is the inclusion of intent. The words “intentionally” or “deliberately” refer to some inner property that cannot be seen, and again, require that we infer something that accompanies the intimidating, coercive or negative act which somehow changes its meaning. Intent is inferred when the child behaves in a manner that is consistent with what we say he intends to do. Saying a child intends to bully does not give us any further understanding as to why he bullies. All we know is that children who engage in these problem behaviours do so at a higher than average frequency because of a reinforcement schedule that is in place for the behaviours (Skinner, 1953, p.143). If we are to find a definition of bullying useful for scientific purposes, we require observable, physical properties of behaviour that we can measure objectively (Skinner, 1953). Behaviour analysis provides a means to do this by using an operational definition of the problem behaviours that we intend to change. In a study designed to reduce bullying in six elementary school students, Ross & Horner (2009) measured playground bullying through clearly defined acts of verbal or physical aggression. By removing the requirement for those acts to be repeated, perpetrated against less powerful peers and with malicious intent, they created a definition that could garner observer reliability. Behaviour analysts need to provide more examples to the larger research community of how this can be done with regards to exclusion, cyber-bullying, coercion, and other difficult-to-define terms if we hope to increase the validity of current bullying research.
Perhaps stemming from the use of subjective definitions of bullying comes the subjective measure of bullying and victimization. A second issue that reduces the validity of much of the research on bullying is the use of self-report as the primary means to measure bullying. In a meta-analysis on school-wide anti-bullying programs, conducted by Smith et al. (2004), 13 of the 14 studies that were included used self-report of bullying or victimization to measure outcome of the intervention programs. Of the 22 studies reviewed by Evans et al. (2014) that measured change in perpetration of bullying following implementation of anti-bullying programs, 15 used only bully self-report as a measure. Of the 27 that measured changes in victimization, 19 used victims self-report as the only measure. While self-report may be an easier method of collecting data from large groups of students, and therefore increasing sample sizes, it does not often correspond with information about bullying that is obtained by observations of teachers and peers (Pellegrini & Bartini, 2000) therefore posing limits on the internal validity of the study.
Directly observing problem behaviour is time consuming, which can be limiting. However, there are other measures, more objective than self-report that have been used in behaviour analytic literature. Office discipline referrals have been used to measure the effectiveness of school wide positive behavioral supports and, while some may criticize these as indirect measures, they are created as a result of teacher’s direct observation of student behaviour (Anderson & Kincaid, 2005). In addition, empirical support exists showing a high correlation between discipline referrals and other tools used to measure student problem behavior, such as teacher and student interviews and direct observation (Irvin, Tobin, Sprague, Sugai and Vincent, 2004).
A Functional Definition of Bullying
If we are to better understand bullying and consequently increase the effectiveness of anti-bullying programs, we must consider bullying as an operant: a class of responses that effect the environment in such a way to generate consequences (Skinner, p. 65) and apply a function-based definition. It is evident from what is included in Table 1, that the shape and form of bullying may be widely diverse, but if we look at bullying as a class of responses that have a common effect on the environment, we have a more efficient way to define and understand the bullying problem. With the exception of Pelligrini & Long (200) where a consequence is described, the definitions in Table 1 only include topographical definitions. Topographical descriptions alone fall short because they describe only the physical properties of the behaviour, but not the effect that he behaviour has on the environment even though it is this effect that ultimately maintains or extinguishes the behaviour. For example, a student spitting on another student while others watch, in the context of the school play, is not an act of bullying, whereas a student spitting on another student in a isolated corner of the playground, while others watch, may well be. Topographically these responses are similar, but functionally they are different as they produce different consequences. Conversely, at times, the topography of bullying may be similar but the consequences may be different, which would result in different functional definitions of bullying. With support from current empirical evidence, it can be theorized that bullying may be maintained by at least four different classes of consequences; positive reinforcement in the form of tangible rewards, positive reinforcement in the form of social attention or interactions, negative reinforcement in the form of escape from aversive stimulation and automatic reinforcement. These will be considered below.
Positive Reinforcement Through Access to Tangible Rewards
One consequence of bullying may be the positive reinforcement associated with access to tangible items. Bullies may engage in aggressive or coercive behaviour to get what they want just like the classic image of the bully who threatens to beat up a classmate to steal his lunch money. This theory may be supported by the correlation between certain physical characteristics and bullies and victims. Research shows that bullies tend to be larger, stronger, and generally more aggressive whereas victims tend to be smaller and weaker (Rigby, 2004). Larger, stronger bullies and smaller, weaker victims makes sense when it comes to access to limited tangible resources. Bigger, stronger children may be more likely to behave aggressively to access tangible rewards because aggression is more likely to result in reinforcement and less likely to result in physical punishment (such as getting beaten up) when the aggressor is larger or stronger than the victim. Ultimately, larger children are less likely to get hurt and more likely to get what they want using physical means than a smaller child simply due to size and strength. Consistent with the above, the reinforcement available for bullying may in fact be control. Many primary reinforcers are only accessed when we successfully control our physical environment (Skinner, 1953, p. 77). This becomes a two-fold issue when we consider the reinforcement that may maintain bullying. First of all, bullying might be reinforced by control of the environment because it results in the bully having access to tangible rewards. For example, the bully who pushes another child aside to get a better place in line may get the last dessert in the school cafeteria. In this case, control is reinforcing as a means to an end because it is required to access the tangible reward. Being able to guarantee others will act in certain ways is what makes certain behaviours such as coercion reinforcing, and bullying has been theorized as a means to access valued resources such as tangible items and attention of opposite sex peers (Pelligrini, 2001). However, if control is repeatedly paired with access to other reinforcers, control itself becomes a conditioned reinforcer and being the person “in control” of the environment becomes reinforcing without needing to be a means to a tangible reward. Skinner states that “the bully is reinforced by signs of cowardice…” (p. 79), because cowardice is often paired with submission and therefore makes control more probable. Skinner suggests that the reinforcement that is generated by control may be unconditioned and may be a result of the evolutionary process. He states, “Any organism which is reinforced by its success in manipulating nature, regardless of the momentary consequences, will be in a favored position when important consequences follow.” (p. 78), so in effect, the bully may bully because control is reinforcing in and of itself and because the tendency to find control reinforcing is an adaptive tendency that would be naturally selected.
Positive Reinforcement Through Social Attention and Group Membership
Another potentially reinforcing outcome of bullying is social reinforcement, specifically attention, social interactions and group affiliation. Empirical evidence that supports this idea comes from the investigation of the role of bystanders in bully behaviour. Research supports that bullying is often reinforced by peer attention (Salmivalli, 2002) and studies that have reduced or removed peer responses to bullying have significantly reduced aggressive behaviour (Ross & Horner, 2009). Membership in different peer groups is also common in schools and it has been theorized that bullying is maintained by group affiliation or allegiance (Rigby 2004). Evans et al. (2014) reported that more non-significant effects in bully prevention programs were reported in culturally diverse or heterogeneous samples (such as racial or ethnic heterogeneity), than in homogeneous samples which lead to them concluding that design and delivery of anti-bullying programs is more complex in heterogeneous samples than in homogenous samples. Income inequality in schools is also linked with bullying (Elgar, Criag, Boyce, Morgan & Vella-Zarb, 2009). This data could support the idea that bullying is influenced by group membership and differences amongst groups. Peer groups have been shown to influence the probability of both desired and undesired behaviours within the classroom (Xie, Cairns & Cairns, 1999). From a behaviour analytic perspective, explaining the behaviour of groups is no different than explaining the behaviour of the individuals within the group. If we believe that bullying results from group affiliation then we must identify what variables are generated by the group which serve to reinforce group affiliation and reinforce the child for behaving in a manner that is condoned or encouraged by the group. This will be the only way to understand and manipulate the behaviour of the bullying group. There are several reasons why group affiliation may be maintained, even when the group condones behaviours that could be punished by parents, school administration or by conditioned punishers such as guilt and shame. Behaving as others behave is likely to be reinforced (Skinner, 1953, p. 311). For example, laughing at a child who trips is more likely to be reinforced if other people are laughing at the child that tripped. Calling a child a name is more likely to be reinforced if others have called a child the same name. In addition, by joining the group the individual increases his likelihood for reinforcement (Skinner, 1953, p. 312). The reinforcing consequences generated by the group exceed the sum of reinforcers achieved by each member individually, so in effect, the group of bullies has more power than each member would have individually. Further, the group itself defines good and bad behaviour so even punishment may not deter bullying as affiliation with the group is more reinforcing than the punishment is punishing. This occurs because as long as the individual is part of the group he can control others and access reinforcers that he is unable to access as an individual. The control exercised by the peer group of the bully is supported by studies that show that peers tend to support bullying rather than stop it (Salmivalli, 1999).
The idea that bullying is maintained by social reinforcement may also explain some of the widely varied effects of anti-bullying programs. Social reinforcement is reinforcement that is mediated by another person such as approval, attention, affection and submission (Skinner, 1953, p. 299). Unlike tangible reinforcers, social reinforcement varies from moment to moment. Different responses may achieve the same level or type of reinforcement and the same response may achieve different levels or types of reinforcement. This may help explain the diversity in results of anti-bullying programs. Social reinforcement is also more likely to be intermittent which may be why bullying appears relatively stable and resistant to extinction and programs with a longer duration show more positive effects (Fox, Farrington, & Ttofi, 2012).
Negative Reinforcement through Escape or Avoidance of Aversive Stimulation
Another function of bullying may be escape or avoidance from aversive stimulation. Research shows that most students identified as bullies tend to have poor academic skills and do not feel integrated into their communities or pride in their school (Roberts & Morotti, 2000 as cited in Smith et al. 2004). Environmental factors that are associated with problem behaviours such as violence and aggression in schools include staff treating students disrespectfully, staff talking negatively about students, teachers blaming students for failure, chaotic and crowded classrooms and hallways, and frequent teasing and put-downs amongst students (Osher, VanAcker, Morrison, Gable, Dwyer & Quinn, 2004). If the classroom or school is an aversive stimulus to the bully then the any behaviour that avoids or temporarily escapes the school environment may be reinforcing. Punishment-based interventions are often used to address behaviour problems in schools and these may consist of exclusion, expulsion, suspension, or detention (Sugai & Horner, 2002). It is possible that bullying is negatively reinforced because bullying results in a temporary escape from academic demands or an aversive school environment due to these supposedly punitive consequences. Research suggests that punishment-based interventions for students with violent or anti-social behaviour usually increase the behaviours they are trying to reduce (Mayer & Sulzer-Azaroff, 1990) which may be a result of negative reinforcement. Possibly more support exists for escape/avoidance as the function of bullying in the existence of the bully-victim. Bully-victims are students who are victims of bullies and also perpetrators of bullying against other students. Approximately 6% of middle school students are self-reported bully-victims. This group of students constitute 33% of the middle school students who report to engage in bullying on a moderate or frequent basis (Nansel et al. 2001). It is possible, for these students, perpetration may be reinforced if it results in exclusion from group activities (such as recess), suspension from school, detention (which may reduce the likelihood of interacting with their own bullies after school) as these consequences would also allow the bully-victim to temporarily escape being victimized.
In the context of peer groups, avoidance of punishment may also reinforce bullying. It has been stated above (Salmivalli, 1999) that the peer group rarely intervene to stop bullying from occurring. This may be because members of the peer group are obedient to the leaders of the group, or to each other, because of the threat of loss of group affiliation. Loss of affiliation is a threat to group members as it likely to precede an aversive condition. This aversive condition could be loneliness, diminished control or security or the possibility that the expelled member may now become a victim of the bullying group. So long as the condition is aversive, circumstances that may lead to it will be avoided and the probability of an incompatible response will increase. For example, assisting the victim of the bullying group would be considered bad behaviour by the group as it reduces the probability of reinforcement for the group. Group members who engage in bad behaviour risk a loss of affiliation in the group which would be aversive, so group members are more likely to engage in behaviour that is incompatible with assisting the victim, such as assisting the bully or cheering him on. In this manner, engaging in bullying can result in avoidance of loss of affiliation with the group, which in turn can be negatively reinforcing to the bully, thus maintaining the bullying.
While bullying is most often viewed as a social phenomenon, we cannot assume that only socially mediated reinforcement maintains it. Bullying may be perpetrated against another person, but under certain conditions, the reinforcement generated may not always be socially mediated. While the presence of another person is required to bully, there may be instances where that other person’s participation is as an object rather than as a social being.
Non-social reinforcement can be automatically positive if the behavior itself produces a rewarding consequence, or automatically negative if the behavior alleviates aversive sensory stimulation. It is not likely that verbal behaviour such as teasing, name-calling or exclusion could be automatically reinforced, however automatic reinforcement could occur if bullying takes the form of physical aggression. In a study that interviewed 30 six through eighth graders, students described experiencing a state of arousal that they described using words such as a “buzz” or a “rush” as a result of participating in or observing school violence (Kerbs & Jolley, 2007). For some students, this physiological state of arousal itself may be reinforcing, especially in situations where stimulation is limited. Osher et al. (2004) identify student boredom as being correlated with higher levels of violence and aggression in schools. One possible explanation is that engaging in bullying results in a reduction of boredom that is negatively reinforcing for the bully. Structured interviews conducted by Kerbs and Jolley (2007) lend support to this as students reported engaging or encouraging school violence as a means to reduce boredom. Another possibility is that aggression may be automatically reinforcing. This is rarely considered as aggression against another person requires the presence of the other person, however a small number of functional analyses have been conducted in which aggression appears to be automatically reinforced. (May 2011, Ringdahl, Call, Mews, Boelter & Christensen, 2007). Although most studies are limited to participants with developmental disabilities, nonsocial consequences have been identified in approximately 9% of functional behavior assessments on aggression (May, 2011). It is theorized in these cases that aggression is maintained and reinforced, to some extent, by the pleasurable sensations, such as a physiological “high” produced by an aggressive act. While this may not be the case with most instances of bullying, it is a consideration that cannot be ignored if our goal is to prevent bullying from occurring.
At the present time, even though we have hundreds of thousands of studies on bullying, we know very little about why bullying occurs or what maintains it. Behaviour analysts are equipped to start answering these questions. We must begin with operational definitions and observable measures as a means to bring the science back to the research around bullying.
We know that it is likely that bullying operates on the environment in a multitude of different ways. Considering the role of tangible reinforcement, social reinforcement, negative reinforcement and even automatic reinforcement can lead to many different theories as to what variables maintain bullying and how antecedents and consequences must be manipulated to reduce or prevent bullying. Behaviour analysis has the technology to support a reduction in bullying in our schools. We have the tools to objectively and functionally define bullying, measure it and through the use of functional analysis, identify specifically the maintaining consequences. Only when we have this information will we be able to create anti-bullying programs that are armed with the specific elements of change required in each environment. Unlike the large majority of research that has been conducted with large sample sizes and group studies, behaviour analysis can offer the single case design that can demonstrate the functional relations that are at play within the school environment that maintain these problematic behaviours. We cannot be overwhelmed by the scope of the problem or the number of students affected or believe that the answer lies in large group studies because the variables that maintain bullying may differ on an individual basis and only by beginning the process of their identification can we move towards informed and scientific intervention.
Anderson, C. M. & Kincaid, D. (2005). Applying behavior analysis to school violence and discipline problems: School-wide positive behavior support. Behavior Analyst, 28, 49-63.
Baldry,, A. C., & Farrington, D. P. (2004). Evaluation of an intervention program for the reduction of bullying and victimization in schools, Aggressive behavior, 30, 1-15.
Glew, G., Fan, M., Katon,W., Rivara, F. P., & Kernic,M.A. (2005). Bullying, psychosocial adjustment, and academic performance in elementary school. Archives of Pediatric Adolescence, 159, 1026–1031, http://dx.doi.org/10.1001/archpedi.159.11.1026.
Elgar, F. J., Criag, W., Boyce, W., Morgan, A., & Vella-Zarb, R. (2009). Income inequality and school bullying: Multilevel study of adolescents in 37 countries. Journal of Adolescent Health, 45, 351-359. doi: 10.1016/j.jadohealth.2009.04.004
Evans, C. B. R., Fraser, M. W., & Cotter, K. L. (2014). The effectiveness of school-based bullying prevention programs: A systematic review. Aggression and Violent Behavior, 19, 532-54. doi: 10.1016/j.avb.2014.07.004
Farrington, D. P. (1993). Understanding and preventing bullying. In: M Tonry (Ed) Crime and Justice, 17. Chicago: University of Chicago Press.
Farrington, D. P., & Ttofi, M. M. (2009). School-based programs to reduce bullying and victimization. Campbell Systematic Reviews, 2009(6), 1-148. doi: 10.4073/csr.2009.6
Fox, B. H., Farrington, D. P., & Ttofi, M. M. (2012). Successful bullying prevention programs: Influence of research design, implementation features, and program components. International Journal of Conflict and Violence, 6, 273-283.
Hyman, I. A. & Perone, D. C. (1998). The other side of school violence: Educator policies and practices that may contribute to student misbehavior. Journal of School Psychology, 36, 7-27.
Irvin, L. K., Tobin, T. J., Sprague, J. R., Sugai G. & Vincent, C. G. (2004). Validity of office discipline referral measures as indices of school-wide behavioral status and effects of school-wide behavioral interventions. Journal of Positive Behavioral Interventions, 6, 131-147.
Kerbs, J. J. & Jolley, J. M. (2007). The joy of violence: What about violence is fun in middle-school? American Journal of Criminal Justice, 32, 12-29. doi: 10.1007/s12103-007-9011-1
Merrell, K., Gueldner, B., Ross, S.W. & Isava, D. (2008). How effective are school bullying intervention programs? A meta-analysis of intervention research. School Psychology Quarterly. 23, 26–42. doi:10.1037/1045-38126.96.36.199
May, M. E. (2011). Aggression as positive reinforcement for people with intellectual disabilities. Research in Developmental Disabilities, 32, 2214-2224. doi: 10.1016/j.ridd.2011.05.029
Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simons-Morton, B., & Scheidt, P. (2001). Bullying behaviours among US youth: Prevalence and association with psychosocial adjustment. Journal of the American Medical Association, 285, 2094-2110. doi:10.1001/jmam.285.16.2094
Olweus, D. (1996). The revised Olweus bully/victim questionnaire. Bergen, Norway: University of Bergen, Research Center for Health Promotion (HEMIL Center).
Olweus, D. (2011). Bullying at school and later criminality: findings from three Swedish community samples of males. Criminal behavior and mental health, 21, 151-156.
Osher, D., VanAcker, R., Morrison, G. M., Gable, R., Dwyer, K., & Quinn, M. (2004). Warning signs of problems in schools: Ecological perspectives and effective practices in combating school aggression and violence. Journal of School Violence, 3, 13-37.
Pelligrini, A. (2001). The roles of dominance and bullying in the development of early heterosexual relationships. Journal of Emotional Abuse, 2, 63-73.
Pellegrini, A. D. & Long, J. D. (2002). A longitudinal study of bullying, dominance, and victimization during the transition from primary school through secondary school. British Journal of Developmental Psychology, 20, 259-280.
Pellegrini, A., & Bartini, M. (2000). An empirical comparison of methods of sampling aggression and victimization in school settings. Journal of Educational Psychology, 92, 360-366.
Rigby, K. (2004). Addressing bullying in schools: Theoretical perspectives and their implications. School Psychology International, 25, 287-300. doi:10.1177/0143034304046902
Ringdahl, J. E., Call, N. A., Mews, J. B. Boelter, E. W., & Christensen, T. J. (2007). Assessment and treatment of aggressive behavior without a clear social function. Research in Developmental Disabilities, 29, 351-362. doi: 10.1016/j.ridd.2007.06.003
Ross, S. W., & Horner, R. H. (2009). Bully prevention in positive behavior support. Journal of Applied Behavior Analysis, 42, 747-759.
Salmivalli, C. (1999). Participant role approach to school bullying: Implications for intervention. Journal of Adolescence, 22, 453-459.
Salmivalli, C. (2002). Is there an age decline in victimization by peers at school? Educational Research, 44, 269–277.
Salmivalli, C., Kaukiainen, A., & Voeten, M. (2005). Anti-bullying intervention: Implementation
and outcome. British Journal of Educational Psychology, 75, 465-487. doi:10.1348/000709905X26011
Skinner, B. F. (1953). Science and Human Behavior. New York, NY: The Free Press.
Smith, J.D., Schneider, B. H., Smith, P. K. & Anadiadou, K. (2004). The effectiveness of whole-school anti-bullying programs: A synthesis of evaluation research. School Psychology Review, 33, 547-560.
Sourander, A., Helstelae, L., Helenius, H., & Piha, J. (2000). Persistance of bullying from childhood to adolescence: A longitudinal 8-year follow-up study. Child Abuse & Neglect, 24, 873-881.
Soutter, A. & McKenzie, A. (2000). The use and effects of anti-bullying and anti-harassment policies in Australian schools. School Psychology International, 21, 96–105.
Sugai, G. & Horner, (2002). The evolution of discipline practices: School-wide positive behavior supports. Child and Family Behavior Therapy, 24, 23-50.
Ttofi, M. M., & Farrington, D. P. (2011). Effectiveness of school-based programs to reduce bullying: A systematic and meta-analytic review. Journal of Experimental Criminology, 7, 27-56.
Xie, H., Cairns, B. D., & Cairns, R. B. (1999). Social networks and configurations in inner-city schools: Aggression, popularity, and implications for students with EBD. Journal of Emotional and Behavioral Disorders, 7, 147-155.