Bullying prevention programs: the importance of peer intervention, disciplinary methods and age variations

Objectives

The aim of this paper is to respond to the commentary of Peter K. Smith, Christina Salmivalli, and Helen Cowie (Journal of Experimental Criminology, 2012), who raise concerns regarding some of the findings of our systematic review and meta-analyses on the effectiveness of bullying prevention programs. They target three findings in particular: (1) the significant association of ??Work with Peers?? with greater victimization; (2) the significant association of ??Disciplinary Methods?? with less bullying perpetration and victimization; and (3) the age variations in effectiveness, suggesting larger effect sizes for older age students.

Methods

We provide explicit information and further detailed analyses on the relationship between these features and effect sizes, including heterogeneity tests and results from weighted regression analyses. For one element in particular (work with peers), we present further research findings from evaluations conducted by Smith, Salmivalli, and Cowie (and also findings from other independent researchers) which support our previous findings. New within-program analyses to examine variations in effect sizes with the age of the students are also presented.

Results

Evaluations conducted by Smith, Salmivalli and Cowie (and by other independent researchers) indicate the same research conclusions: although peer support schemes appear effective based on attitudinal surveys, these schemes are not related to actual levels of bullying or victimization and, in fact, are quite often related to an increase in bullying and victimization. Our definition of ??disciplinary methods?? did not include the zero-tolerance approach or any type of harsh discipline as suggested in the commentary. In all relevant cases, ??disciplinary methods?? included sanctions within a warm and loving framework, following the Olweus bullying prevention guidelines. While most programs that utilized firm disciplinary methods were inspired by Olweus, the relationship between disciplinary methods and less victimization was not driven by the Olweus program (which was not related to the victimization effect size). Larger effect sizes (i.e. reductions in bullying and victimization) for programs implemented with older students is a robust result also found in a more recent systematic review regarding the effects of anti-bullying programs on bystander intervention. In within-program analyses, most results suggested that effect sizes were greater for younger students, but these results were driven by the less controlled evaluations. The most controlled evaluation (randomized experiment) provided the opposite result.

Conclusions

More research is clearly needed on the effectiveness of bullying prevention programs with students of different ages, and we also recommend randomized experiments to assess the importance of different intervention components.     

Effectiveness of school-based programs to reduce bullying: a commentary

Objectives

Recent decades have seen a focus on intervention programs to reduce school bullying, in light of the severe negative consequences of such behavior. A recent meta-analysis by Ttofi and Farrington (Journal of Experimental Criminology 7: 27?C56, 2011) provided encouraging findings in terms of some significant reductions in bullying and victimization achieved by many programs. They also report analyses of effect sizes associated with specific program elements and design features of the interventions. While this is an important step forward, we critique some of the strong policy implications which they draw from these latter analyses.

Methods

We discuss four important areas to substantiate this critique: analytical procedure, definitional issues, historical issues, and recent empirical data. As context, we use two particular program elements described by Ttofi and Farrington, namely use of disciplinary measures and work with peers, and one design feature, namely age of pupils.

Results

The findings for the program elements and design feature examined are complex and do not justify strong policy implications at this stage.

Conclusions

We conclude with suggestions for future research directions.     

Déjà vu! The effect of previewing test items on the validity of the Concealed Information polygraph Test

Abstract

This research compares resident, peer and staff reports of bullying and victimization in one open facility and one secure facility for young offenders. The prevalence of bullies and victims was higher according to resident reports (self-reports) than according to peer or staff reports. There was significant agreement between residents, peers, and staff in identifying bullies and victims. The majority of bullies and victims identified by peers and/or staff were also identified by self-reports. Making the plausible assumption that bullies and victims tend to be concealed, it is concluded that self-reports (in the context of an individual interview) provide the most reliable and valid information about bullies and victims in young offender institutions. Hence, anti-bullying strategies should be based on self-report data.     

The Impact of Self Control and Neighborhood Disorder on Bullying Victimization

PurposeWhereas past research has examined the effect of individual-level and neighborhood-level predictors of bullying victimization separately, the current study examines their effects collectively.MethodsMiddle and high school students (n = 1972) in randomly selected classes within a Southeastern school district completed a battery of self-report measures. Levels of self-control (an individual-level factor) and neighborhood disorganization (a neighborhood-level factor) were regressed onto measures of the six-week prevalence of verbal, physical, and cyber bullying victimization.ResultsLow self-control and neighborhood disorder were found to be associated with each type of bullying victimization, though the impact of self-control was partially mediated by neighborhood disorder when included in the same model. The effect of self-control was mediated when subsequently controlling for poly-victimization experiences. Net of these controls, neighborhood disorder continued to be associated with a statistically significant increase in the odds of bullying victimization.ConclusionsEconomic and social decay within neighborhoods increased the likelihood of bullying victimizations. These effects hold true across verbal, physical and cyber victimizations, suggesting a need to consider both community characteristics when staging bullying intervention campaigns. Additionally, the findings suggest a need for further research considering the relationship between self-control and neighborhood conditions on the risk of victimization generally.     

Does incarceration-based drug treatment reduce recidivism? A meta-analytic synthesis of the research

This research synthesized results from 66 published and unpublished evaluations of incarceration-based drug treatment programs using meta-analysis. Incarceration-based drug treatment programs fell into five types: therapeutic communities (TCs), residential substance abuse treatment (RSAT), group counseling, boot camps specifically for drug offenders, and narcotic maintenance programs. We examined the effectiveness of each of these types of interventions in reducing post-release offending and drug use, and we also examined whether differences in research findings can be explained by variations in methodology, sample, or program features. Our results consistently found support for the effectiveness of TC programs on both outcome measures, and this finding was robust to variations in method, sample, and program features. We also found support for the effectiveness of RSAT and group counseling programs in reducing re-offending, but these programs’ effects on drug use were ambiguous. A limited number of evaluations assessed narcotic maintenance or boot camp programs; however, the existing evaluations found mixed support for maintenance programs and no support for boot camps.     

Evaluating programs for men who batter: Problems and prospects

The uncertainty and controversy that surround batterer programs beg more systematic evaluation of program effectiveness. The current national surveys and limited evaluations of batterer programs point to several methodological improvements that include comparisons of program modalities, observation study of the group interaction, and more sensitive measures of abuse. Therefore, suggestions for future evaluations are posed, and continuing problematic areas, such as access to the victim, sensitized responses, selectivity of clients, and community context, are discussed.     

Preventing repeat incidents of family violence: analysis of data from three field experiments

Preventing repeat victimization is an area of criminology that has shown particular promise in recent years. Based on the premise that persons once victimized are at higher risk than others for future victimization, British officials developed successful programs that focus crime prevention efforts on victims. Of all crimes, family violence may have the highest repeat rate, especially in the first weeks after an incident is reported to the police. Accordingly, New York City officials developed an intervention program to reduce repeat incidents of family abuse. Three field experiments conducted during the 1990s evaluated whether or not this program, targeted at public housing residents who reported family violence to the police, reduced the rate of subsequent victimization. The findings produced within each study were not consistent across the studies; rather, these three experiments, separately analyzed, produce varying results. Since the composition of the samples varied across studies, however, one possible explanation is that this program has different effects within different populations. This paper reports outcomes from a series of analyses of pooled data from these three studies to address the inconsistencies. The results indicate that the intervention brought about greater reporting of subsequent abuse both to authorities and to research interviewers. The results are invariant across the three studies, indicating that greater reporting of abuse is not idiosyncratic to one particular population, and are consistent across the nature and source of outcome measures. These findings suggest the need for careful monitoring by the advocates and agencies that operate these types of programs and among those designing and testing future programs.     

How do Traditional Bullying and Cyberbullying Victimization Affect Fear and Coping among Students? An Application of General Strain Theory

School bullying victimization represents an important type of strain (or stressful event) experienced by individuals because it usually occurs frequently, over a long period of time, and is an intentional violation against another person. The current study utilizes the 2009 School Crime Supplement (SCS) of the National Crime Victimization Survey (NCVS) to assess the effect of bullying victimization on avoidance behaviors and carrying weapons to school. This study also examines how bullying victimization affects fear and whether this emotion mediates the relationship between victimization and outcomes. As expected, traditional bullying and cyberbullying increase the likelihood of fear. In addition, traditional bullying and cyberbullying victimization increase the likelihood of avoidance behaviors and bringing a weapon to school. Contrary to expectations, fear did not mediate the relationship between bullying and coping behaviors.     

A friend in need: the role of friendship quality as a protective factor in peer victimization and bullying

This study examined friendship quality as a possible moderator of risk factors in predicting peer victimization and bullying. Children (50 boys and 49 girls, ages 10 to 13 years) reported on the quality of their best friendship, as well as their bullying and victimization tendencies. Parents reported on their child's internalizing and externalizing behaviors, in addition to bullying and victimization tendencies. Results indicated that externalizing problems were related to bullying behavior; however, friendship quality moderated this relation such that among children with externalizing behaviors, a high-quality friendship significantly attenuated bullying behavior. Internalizing problems and low friendship quality were significantly related to victimization; however, friendship quality did not moderate the relation between internalizing problems and victimization. Implications for interventions based on these findings are discussed.     

The effectiveness of treatment for sexual offenders: A comprehensive meta-analysis

The article reports a meta-analysis on controlled outcome evaluations of sexual offender treatment. From 2,039 documents published in five languages, 69 studies containing 80 independent comparisons between treated and untreated offenders fulfilled stepwise eligibility criteria (total N = 22,181). Despite a wide range of positive and negative effect sizes, the majority confirmed the benefits of treatment. Treated offenders showed 6 percentage points or 37% less sexual recidivism than controls. Effects for violent and general recidivism were in a similar range. Organic treatments (surgical castration and hormonal medication) showed larger effects than psychosocial interventions. However, this difference was partially confounded with methodological and offender variables. Among psychological programs, cognitive–behavioral approaches revealed the most robust effect. Nonbehavioral treatments did not demonstrate a significant impact. There was no outcome difference between randomized and other designs, however, group equivalence was associated with slightly larger effects. Various other moderators had a stronger impact on effect size (e.g., small sample size, quality of outcome reporting, program completion vs. dropout, age homogeneity, outpatient treatment, and authors’ affiliation with the program). More differentiated, high-quality evaluations are needed to clarify: What works for whom under which circumstances?     

A meta-analytic review of court-mandated batterer intervention programs: Can courts affect abusers’ behavior?

Court-mandated batterer intervention programs are being implemented throughout the United States to address the problem of domestic violence. Prior reviews of research on the effectiveness of these programs have arrived at conflicting conclusions. This study is a systematic review of the extant research on this topic. Experimental and quasi-experimental studies that used matching or statistical controls were included. The results were mixed. The mean effect for official reports of domestic violence from experimental studies showed modest benefit, whereas the mean effect for victim reported outcomes was zero. Quasi-experimental studies using a no-treatment comparison had inconsistent findings indicating an overall small harmful effect. In contract, quasi-experimental studies using a treatment dropout design showed a large, positive mean effect on domestic violence outcomes. We discuss the weakness of the latter design and raise concerns regarding official reports. The findings, we believe, raise doubts about the effectiveness of court-mandated batterer intervention programs.     

METHODOLOGICAL CONSIDERATIONS IN EVALUATING FAMILY COURT PROGRAMS

Although there has been in recent years a proliferation of court programs, especially divorced parent education programs, evaluations of the effectiveness of these programs lag dangerously behind their inception, perhaps because program developers and courts lack the expertise to perform these evaluations. The present article provides a primer of the methodological considerations for evaluations of court programs. The authors discuss the following topics: formative versus summative evaluations; how to discover program goals by identifying stakeholders; how these goals translate into measures to be assessed; the data sources for these measures (exit questionnaires, archival data, and follow-up surveys); answering the “compared with what?” question through the selection of an appropriate research design; budgeting the evaluation; and disseminating the findings through appropriate write-up(s).     

Assessing problematic research: How can academic researchers help improve the quality of anti-crime program evaluations?

In this essay I examine some of the problems that prompted the National Research Council (NRC) report and consider how academic researchers might help resolve them. Many of the problems were found to be associated with research designed to assess program effects on child victimization and violence against women, areas in which research participation by subjects is particularly burdensome and difficult to obtain. Yet, program evaluations often assume that the process of subject participation is well understood and that outcome measures are reliable and valid across all subjects. A multidisciplinary, comprehensive and systematic review of victimization programs and past research is needed to advance the rigor of future evaluations. However, academics should not insist that all victim service programs warrant program evaluation as a condition of continued public support, because the decision to retain a program inevitably involves more than a scientific estimate of its effect.

The effects of a group batterer treatment program: A randomized experiment in Brooklyn

Despite numerous evaluations of batterer treatment programs, most lack sufficient methodological rigor to yield valid answers about the programs' effectiveness. This paper presents results from an experimental evaluation in which 376 adult males convicted of domestic violence were randomly assigned to either a 40-hour batterer treatment program or 40 hours of community service that did not include any therapeutic treatment. We examined both official records and victims' reports of recidivism. Those assigned to the treatment program showed significantly lower recidivism, on the basis of all outcome measures from official records. Although victims' reports also recorded fewer failures among the batterers assigned to the treatment group, the differences in failure rates were not large enough to be statistically significant. Overall results suggest that therapeutic treatment for batterers may reduce domestic violence among convicted batterers who agree to this sentence.     

Psychopathology in women arrested for domestic violence

This study examined the prevalence of psychopathology among women arrested for violence and whether the experience of intimate partner violence (IPV) was associated with Axis I psychopathology. Women who were arrested for domestic violence perpetration and court referred to violence intervention programs (N= 103) completed measures of IPV victimization, perpetration, and psychopathology. Results revealed high rates of posttraumatic stress disorder (PTSD), depression, generalized anxiety disorder (GAD), panic disorder, substance use disorders, borderline personality disorder, and antisocial personality disorder. Violence victimization was significantly associated with symptoms of psychopathology. Logistic regression analyses showed that sexual and psychological abuse by partners were associated with the presence of PTSD, depression, and GAD diagnoses. Results highlight the potential importance of the role of violence victimization in psychopathology. Results suggest that Axis I and Axis II psychopathology should routinely be assessed as part of violence intervention programs for women and that intervention programs could be improved by offering adjunct or integrated mental health treatment.     

Compliance and learning in an intervention program for partner-violent men

Although research has yielded mixed findings regarding the effectiveness of intervention programs for partner-violent men, it appears that greater participant compliance with such programs is associated with better outcomes. However, no research to date has jointly examined compliance with intervention programs and the extent to which partner-violent men learn specific information presented during the programs. The current study makes use of existing data to evaluate general and specific elements of partner-violent men's compliance with (i.e., active, appropriate participation in) an intervention program and recall of key points from the program. Results from a subsample of 22 men indicate that at program termination, those rated as having been "process conscious" during intervention group sessions, having self-disclosed during sessions, having evidenced awareness and use of techniques to avoid violence, and having used respectful language show greater recall of material taught in the program. This finding points to the potential benefit of taking steps to increase men's active participation in programs and of studying active engagement as a mediator of program effects on men's violence toward partners.     

Intensive case management as a jail diversion program for people with a serious mental illness: a review of the literature

This article reviews the research on intensive case management (ICM) programs as a jail diversion intervention for people with a serious mental illness (SMI). The review includes two types of ICM programs: (a) general ICM programs that included an assessment of arrests and incarceration rates for people with an SMI and (b) ICM programs specifically implemented as a component of a jail diversion intervention for people with an SMI. Results indicate that general ICM programs (19) rarely led to reductions in jail or arrest rates over time, and these rates were similar to those found in standard mental health services. General ICM programs that included an integrated addiction treatment component (8) had mixed results but a trend toward reductions in rates of arrests and incarceration over time for individuals with an SMI and a co-occurring substance use disorder. Results were mixed for jail diversion interventions with an ICM program, but most ICM programs (8) led to significant reductions in arrests and incarcerations over time. Specific elements of effective ICM jail diversion programs are discussed.     

The reliability and validity of self,peer and staff reports of bullying and victimisation in correctional and care institutions

Abstract

Connell and Farrington (1997) carried out the first study of the reliability and validity of reports of bullying and victimisation in correctional institutions, using individual interviews. The present research describes the second study on this topic using questionnaire measures of self, peer and staff reports of bullying and victimisation in one Children's Home (N=16 residents, 15 peers, 6 staff, making up to 1440 comparisons) and one Correctional Home (N=19 residents, 18 peers, 6 staff, making up to 2052 comparisons) in Croatia. The present study improved on the Connell and Farrington study by examining the test–retest stability of the questionnaire and using relative improvement over chance coefficients in assessing inter-rater reliability. Unlike in the Connell and Farrington study, self-reports of bullying and victimisation were collected in small groups rather than in individual interviews. Generally, there was significant agreement between residents, peers and staff in identifying bullies and victims, and the test–retest stability of the questionnaire was high. In line with the findings of Connell and Farrington, it is concluded that self-reports provide the most reliable and valid data about bullying amongst institutionalised youth.     

Reaching Out versus Lashing Out: Examining Gender Differences in Experiences with and Responses to Bullying in High School

The present study examines gender differences in bullying in high school. Unique contributions include comparisons of both victimization and perpetration rates across four subtypes of bullying: physical, verbal, relational, and cyber. Further, as we conceptualize bullying within the larger framework of literature on social rejection, we also address whether there are gender differences in experiencing social rejection–in the form of bullying–and responding with aggression, as opposed to asocial or prosocial behavior. The literature yields mixed findings across these three questions (i.e., gender differences in experiences with victimization and perpetration and responses to those experiences), suggesting sample variations (Archer Review of General Psychology, 8(4), 291–322, 2004; Archer & Coyne Personality and Social Psychology Review, 9, 212–230, 2005; Card, Stucky, Sawalani, & Little Child Development, 79, 1185–1229, 2008). Thus, we explored experiential differences in our sample, and hypothesized based on the tend and befriend model (Taylor et al., 2000) that girls would be more likely than boys to respond to bullying with prosocial behaviors. With regard to victimization and perpetration differences, we found that male students both experienced and perpetrated significantly more physical bullying. Boys were also significantly more likely to report experiencing verbal bullying than girls. No significant differences emerged for relational or cyber bullying. With regard to responses, social withdrawal was more common than aggressive responding, but consistent with the tend and befriend model, girls chose prosocial responses significantly more than boys, whereas boys were just as likely to choose antisocial responding as prosocial responding. These results suggest that gender should be considered in studies addressing the question of when experiences with rejection–in its many forms–results in antisocial versus prosocial behavior.     

About the authors

In recent years, afterschool programs have received support for their potential to reduce juvenile delinquency and victimization. This support stems largely from reports based on police incident data indicating that juvenile crime and victimization peak during the afterschool hours. However, prior studies of victimization surveys and self‐reports of crime suggest that delinquency is more elevated during school hours. Utilizing self‐report data from a sample of juveniles participating in an evaluation of afterschool programs in Maryland, this study shows that juvenile victimization and delinquency peak during the school hours, while substance use peaks during the weekend. Disaggregating by offense reveals, however, that the more serious violent offenses are elevated during the afterschool hours, while simple assault offenses are most elevated during school hours. Implications for research and practice are discussed.