This action might not be possible to undo. Are you sure you want to continue?
Bradley Cousins, Rebecca Stewart Reviewed work(s): Source: Canadian Journal of Education / Revue canadienne de l'éducation, Vol. 28, No. 4 (2005), pp. 739-762 Published by: Canadian Society for the Study of Education Stable URL: http://www.jstor.org/stable/4126453 . Accessed: 28/03/2012 18:58
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact firstname.lastname@example.org.
Canadian Society for the Study of Education is collaborating with JSTOR to digitize, preserve and extend access to Canadian Journal of Education / Revue canadienne de l'éducation.
Antibullying Interventionsin Schools: Ingredients of Effective Programs Stewart J. David Smith,J. Bradley Cousins,& Rebecca
Because bullying is a serious problem in Canadian schools, antibullying programs have been widely implemented to redress the problem. School principals in Ontario (N=395) completed a questionnaire to document the severity of bullying, the amount of anti-bullying resources, and the variety of antibullying activities in their schools. Results reveal that reductions in bullying in previous years, sufficiency of resources for resolving bullying, and amounts of antibullying programming were all positively associated antibullying program outcomes. These data suggest that the investment of time, effort, and money in school-based antibullying initiatives can lead to safer and more peaceful schools environments. Key words: bullying, primary prevention, program evaluation Comme l'intimidation est un probleme serieux dans les &coles canadiennes, des programmes de lutte contre l'intimidation font leur apparition un peu partout. Des directeurs et directrices d'&cole en Ontario (N = 395) ont rempli un questionnaire visant a documenter la gravite du problbme, les ressources existantes et les diverses activites anti-intimidation mises en place dans les dcoles. D'apres les rdsultats du questionnaire, la rdduction de l'intimidation au cours des annees precedentes, la pertinence des ressources en place pour faire face aux incidents d'intimidation et le nombre de programmes de lutte contre l'intimidation etaient tous correlds a l'amelioration des resultats en la matiere. Ces donnies semblent indiquer que le temps, les efforts et les fonds investis dans les initiatives anti-intimidation contribuent a creer un climat de paix et rendre les &colesplus securitaires. Mots cles : intimidation, prevention, dvaluation de programmes.
Bullying is a serious social problem afflicting schools around the world, and it appears that Canada is no exception to this trend. Bullying is a subtype of aggressive behaviour in which the perpetrator exerts power over a weaker victim through various means including physical size or
CANADIAN JOURNAL EDUCATION28, 4 (2005): 739-762 OF
& McDougall. depression.g. and the roles children play within bullying relationshipsas bullies or victims consolidate over time. who are increasingly preoccupied with the problem of bullying. 2003). diminished selfesteem. it is not surprising that educators in Canada. or psychological advantages. but they can be sorted generally into two categories: direct bullying. physical methods (Crick & Nelson. rates of depression among . Although victims as a group show the highest rates of depression.740 & J. making indirect bullying difficult for school authorities to detect. with physical bullying declining following its peak in early childhood. It appears that girls tend to favour verbal aggression or indirect bullying strategies. 2002). Victimized children tend to display internalizing symptoms.2003). Methods of bullying evolve through development. SMITH. STEWART J. R. research evidence indicates clearly that bullying and victimization are toxic to children's health (Rigby. COUSINS. and social withdrawal (Nansel et al. and indirect bullying.Bullying occurs within a within a dynamic relational context. Consequently. Canada placed in the top quartile for bullying and the top third for victimization among the 36 participating countries (Craig & Pepler. Hymel. and verbal and indirect bullying increasing into adolescence before declining thereafter (Craig & Pepler. spreading rumours and excluding individuals from peer groups). Although the effects of bullying are not fully known. which typically involves covert activities intended to isolate and marginalize victims (e. With these results.The data revealed that 54 per cent of Canadian boys and 32 per cent of girls bullied others in a six-week interval. These features make bullying a particularly vicious kind of peer-on-peer aggression and render victimized children vulnerable to long-term consequences(Rigby. which involves physical and verbal attacks on victims. Indirectbullying tends to be more difficult to observe than acts of direct bullying. Results from a recent international study reveal a rather bleak picture of the scope of bullying problemsin Canada. 2003). including anxiety. 2003)..and 34 per cent of boys and 27 per cent of girls were victimized in the same interval.. whereas boys most often use direct. strength. and that is repeated over time (Vaillancourt. age.D. 2001).B.2003). are seeking effective solutions for peer aggression and harassment. Bullying comprises a variety of behaviours.
even deadly.Graham. Josephson.Sutton & P. Gough. 2004).2000). For example. Childrenwho bully also show negative health consequences from involvement in bullying. & Atlas. For example. 2001).. Smith. Pepler. identified a variety of roles that peers can play in bullying situations.g. is increasingly situated within larger social systems like peer groups. and schools.ANTIBULLYING INTERVENTIONS SCHOOLS IN 741 bullies are still significantly higher than non-involved peers. Johnson.These childrentend to display both the anti-socialbehaviour of bullies as well as the social and emotional difficulties of victims (Glover. no longer viewed simply as a dyadic interactionbetween a perpetrator and victim. 1999). & Taradash."although not directly engaged in the bullying.& Schuster. "Onlookers. displaying the highest levels of conduct. 2001). This model finds support in empirical studies indicating that bullying frequently occurs in the presence of peers and that the actions of peers more often encourage the bullying than stop it (Craig & Pepler. school.. Bullying. 2002). who studied how bullying unfolds in peer groups. 1997. & Cartwright.. adolescents who reported bullying others viewed their girlfriends or boyfriends less positively and equitably and reported more physical and social aggression with them than other adolescents (Connolly. and bully and victim groups have equally high rates of suicidal ideation (Roland. Perhaps the most significant and worrisome consequence of bullying others is susceptibility to future problems of violence and delinquency. . Nansel et al. This profile of dysfunction places these youth at high risk for violent. children in the aggressive-victimgroup were the most troubled. 2000).2003). reactionsto chronicbullying (Andersonet al. and peer relationshipproblems (Juvonen. families. "assistants" participate in secondary roles by helping the bully commit the act. K. are an audience for the bully and tacitly reinforcethe aggression with their attention (Craig. Pepler. in a study of the dating relationships of bullies.2000. Salmivalli (1999). Craig. In a recent study of mental health problems associated with involvement in bully/victim problems. "reinforcers"provide verbal encouragement to the bully as the act occurs. A subgroup of victims reacts aggressively to abuse ("aggressive victims") and displays a distinct pattern of psychosocial adjustment. Research also suggests that adults subjected to victimization as children show longterm effects on their personal relationships and mental health (e.
& Egan. ..D. lack of warmth and involvement) are common. STEWART J. 2000). Furthermore. 2001).either as perpetratorsor victims.g. Peer-led interventions. currently popular in schools. & Cowie.742 & J. and supportiveness that they need to help students involved in a bullying situation. 2005).. Parentsand home environmentsalso contributeto bully/victim problems in children. monitor their children. using interventions like social skills and assertiveness training and anger management (P. negative emotional attitudes (e. is predicated on the assumption that bullying is a systemic problem. involve teaching peer helpers the basic skills of active listening. set age-appropriatelimits.2001). empathy. The whole-school approach. COUSINS. 2000). R. Individualized interventions target children who have had significant involvement in bullying situations. ANTIBULLYING INTERVENTIONS Concurrentwith the rising awareness of bullying and its consequences has been a marked increase in school antibullying interventions that address the problem from different angles.parents who communicatelove and warmth. SMITH. and create roles and structures that encourage students to act in responsible and empathic ways (Cowie & Olafsson.B. Children who bully tend to come from homes where aggression is a favoured problem-solving method.who tend to be less accepting of adult authority and direction than younger children (Salmivalli. Smith. like befriendingor conflict resolution. problem solving. These interventionsinvolve the active participationof many students. K. These interventionstypically focus on remedying specific externalizing problems of bullies or the internalizing problems displayed by victims. and the children are encouraged to fight back when harassed (Glover et al. promote communication rather than blame among those involved in bullying. research indicates that chronically victimized children may have histories of insecure parental attachmentsin infancy and are subjectto intrusive and overprotectiveparenting(Perry. The peer-led approachmay be particularlyadvantageous for adolescents. Conversely. and. 2003). Ananiadou. and use non-physical punishment to deal with misbehaviour constitute an important protective factor against involvement in bully/victim problems (Orpinas & Home.Hodges.
Their data from fourth. and in some jurisdictionsare requiredby law.. community leaders and organizations) to become involved in their initiatives. Furthermore.For example. increased adult supervision on school grounds.IN INTERVENTIONS SCHOOLS ANTIBULLYING 743 by implication. victims. few studies exist of the effectiveness of antibullying programs implemented in schools.and fifthgrade children revealed reductions in bullying for non-aggressive children but increases for aggressive children (although these increases were smaller than increasesobserved in aggressive children in a wait-list control group). Teglasi and Rothman (2001) evaluated a 15-session. and they may be invited to participate directly in some activities. programs must address the problem at all levels of a school community.the findings of these studies do not provide conclusive evidence for the effectiveness of such programs. Baldry and Farrington(2004) evaluated a program for students aged 10 to 16 that taught social competence skills using . Whole-school programs may mobilize peer-helpers in mediation or befriending programs with appropriate adult supervision. Salmivalli (2001)examined the impact of a weeklong. most programs designated as whole-school share the core features of the original Olweus program. The Olweus Bullying Prevention Program (Olweus. schools may also solicit community stakeholders (e. Parents of all children receive information about the schools' initiatives and goals. peer-led program in Finland to reduce bullying among children in grades 7 and 8. classroom program that used peer group and story form to improve social problem solving for bullies. 1993) was the first comprehensive. Within classrooms. These features typically include activities for the entire school. In the context of these programs. whole-school intervention implemented on a large scale and systematically evaluated. and bystanders. Finally. Although antibullying programs are ubiquitous in North American schools. such as the development of an antibullying policy. Results showed a positive effect of the interventionfor girls but a negative effect for boys. targeted interventions are usually offered to children directlyinvolved in bully/victim problems. who reported more pro-bullying attitudes at the end of the program.g. and the establishmentof an antibullying committee. teachers may develop behaviour codes with the children and engage their students in a variety of curricularactivities with antibullying themes.
and the impact they have across the school environment. intermediate. activities.D.and outcomes) as a conceptual guide for the survey (Rush & Ogbourne. the way they are implemented. we designed a survey to explore the relations between various aspects related to the content and implementation of school-based antibullying programs and the perceived impact of those programs. P. Smith. Resources include human and fiscal inputs into the programs. K.744 & J. . there is a pressing need for more data from varied perspectives to advance understanding of the kinds of interventions that work for different groups of children.We solicited the perspective of school principals for this survey because they are commonly in the position of selecting or approving these programs and therefore understand the nature of the programs.B. We used a generic program logic model consisting of five basic components (i. adapted to antibullying programs. who meta-analyzed 14 studies of wholeschool antibullying programs (including two Canadian studies). evidence-based decisions about appropriate programs.who reportedless victimization. including such things as who receives the program and in what "dosage. The program was effective with older children. This generic framework.but not with younger children. D. Clearly. found that only one program yielded significant reductions in victimization and bullying. & Ananiadou (2004). videotapes and written material. while the other 13 yielded either negligible changes or increaseson these outcomes. Outputs are the necessary (but not sufficient)conditions requiredfor achieving desired outcomes. needs. This inconclusive data hampersthe efforts of school authorities who want to make sound. Needs correspond to the identified needs to which antibullying programs are expected to provide a response. who reported more victimization. STEWART J. Finally.e.. data analysis. J. resources. 1991). As this sample of findings illustrates. Schneider. Smith."Finally outcomes are associated with the immediate. guided questionnaire development. In this context. COUSINS. and data interpretationfor this exploratorystudy. R.outputs. and activities consist of program actions or implementation behaviours. it is difficult to discern clear patterns in existing literature on the effectiveness of antibullying programsin schools. SMITH.and longerterm consequences of the program that align with program objectives.
Completed survey forms were sent to a private company for electronicscanning of responses.g. participantsanswered questions . vice-principal or guidance counsellor). a blank questionnaire. beginning at a randomly selected starting point. Procedure We sent each school in the sample a survey package addressed to the school principal. and an addressed and stamped envelope to return the questionnaire.. Participationin the study was voluntary. It included a cover letter explaining the nature and purpose of the study. Two weeks after the packages were sent. All 3469 schools were entered into a single list. the set of private schools was treated as a single board. nine pages in length. and participatingschools were anonymous to the researchers.1In the first part. These 3106 schools and an additional 363 private schools not under the jurisdiction of any school board constituted the sampling frame for the study. although we provided space for participantsto add verbal comments. we mailed a remindercard to all school principalsin the sample. Measures The Antibullying Program Survey questionnaire. (Forthe purpose of this operation. comprised mainly scaled questions and was divided into four parts identified with headings. We received positive responses from 43 school boards with a total of 3106 member schools.We asked the principals to complete the questionnairethemselves (which took approximately2030 minutes) or to delegate the task to another appropriateschool official (e.) We constructed an initial sample of 1734 schools by selecting every second school on the list. Ethical clearance for this study was grantedby the University of Ottawa'sResearchEthicsBoard. which was ordered alphabeticallyfirst by school board and then by school within boards.ANTIBULLYING INTERVENTIONS SCHOOLS IN 745 METHOD Participants We contactedthe directorsof all 90 English-languagepublic and Catholic school boards in Ontario to request permission to recruit school principals in their districts to participate in a survey of school-based antibullying programs.
The fourth section probed a school's efforts to evaluate its antibullyingprograms. After we fully developed the questionnaire. comprising 17. related to the characteristicsof the school and the student body. Six additional items queried respondents about who planned. provided resources. In the second part. and individual children) inventoried current interventions and services in the school intended to deal with bullying and its effects.5 per cent of the final sample relative to 10. STEWART J. 13.making a number of minor changes to the text on the basis of their feedback. and received the programs and whose roles in creating and/or solving bullying problems were addressed in the programs. 7. Private schools were somewhat over-representedin the final sample of respondents. 30 items organized into six categories corresponding to the whole-school model (school.or stealing) and indirect and relationalbullying (e. To reduce the chances of type II error. classroom.B. RESULTS We received 395 completed and usable questionnairesfor a response rate of 22.For the purposes of this study.. we pilot-tested it with three local school principals. peers. parents. 18) where each was entered as a dependent variable in a 2 (public versus private schools) x 2 (principalsand vice-principalsversus others) fully factorial ANOVA.g.4 per cent of the initial sampling frame of 3469 schools. physical attacks. name calling. social exclusion or spreading rumours). we conducted several preliminary analyses with the data. 12 items probed the nature and severity of bullying problems at the school. 2..8 per cent.D. Three additional items queried respondents about the adequacy of resourcesdirected at solving bullying.746 J. In the third section. including incidents of direct bullying (e. For the first of these analyses.g. but excluded data from section 4. To test for equivalency of responses across school and respondentgroups and to assess the representativenessof the sample.which is a more critical . Principalsand vice-principalscompleted 88 per cent of returnedsurveys. community. we used standardized scores for the five key variablesfor this study (see Table 2: Variables1. SMITH. we analyzed and used the data collected on all items in sections 1-3 of questionnaire. delivered.& R. COUSINS. the remaining 12 per cent were completed by delegates such as guidance counsellors or teachers.
we tabulated data about characteristics of schools in the sample (see Table 1) and compared them to provincial norms for the 2000-2001 school year. 20 per cent or fewer students belonged to visible minority groups.8 per cent of schools. which were the latest available norms (Ontario Ministry of Education. the main effect for respondent on two of the five dependent variables (2 and 7) reached significance.3 per cent of participating schools. we decided to exclude data from the private schools from subsequent analyses.ANTIBULLYING INTERVENTIONS SCHOOLS IN 747 consideration in this instance than type I error. Data on private schools in Table 1 reveal a number of important deviations between sample and the and of population. they said. and in 77. These data show that the sample of public and Catholic schools is roughly representative of Ontario schools on these variables. In 84. The results showed that the main effect for school on all five dependent variables reached significance. in 75. Public and Catholic schools in the sample were located in varying geographical and socio-economic contexts: 52. including secondary proportions schools and student enrolments. and 18. and the interaction was significant on one of the five variables (7). therefore. particularly in elementary/secondary the elementary and secondary schools. their schools are very small and highly specialized and. 42. although the group of secondary schools in the sample tended to have a larger enrolment than the provincial average. inappropriate to participate in the survey.5 per cent in communities of belowaverage SES. 20 per cent or fewer students were receiving special education services for learning-related difficulties. respondents from several secondary schools explicitly declined to participate because. On the basis of these findings. To assess the representativeness of the sample. 20 per cent or fewer did not have English as their first language.2 additional information about the Respondents provided characteristics of their schools.8 per cent of schools. Similarly. 1971).4 per cent of schools were in communities of average SES. Anecdotally. 39. 2003).2 per cent in communities of above-average SES. alpha was set at .20 for this series of tests (Winer.2 per cent of secondary schools were situated in urban areas. .1 per cent of elementary schools and 58.
B. The activities component includes three variables. that from In preparation for analyses.2) 9 (13. and scale-reliability statistics are reported for these scale variables in Table 2.748 & J. COUSINS.8) 16 (23. The six different variables comprising the needs component together reveal the current levels of bullying occurring in the schools and the degree to which these levels have changed in recent months and years. The two other variables in the activities component assess the degree of change in the amount of programming offered in schools in recent years.0) 453 (61. We derived aggregate scores on six of the 18 variables listed from multiple items. The resources component includes six variables that collectively show the amounts of various resources (e.g.4) 81 (20. Within the outputs . activities.8) 210 (28.6) ulation pie 371 (173) 1005 (401) ulation pie 21 (37) 62 (26) 18 (--) 51 (--) Public and Catholic Schools 3963 (82. materials) dedicated to resolving bullying as well as respondents' opinion about the sufficiency of these resources (variable 8). outputs and outcomes) along with relevant descriptive statistics. we organized survey data according to the components of the program logic model.. the first of which (variable 13) represents the amount of antibullying programming currently offered in the schools.7) 830 (17. STEWART Table 1 School Characteristics: and Data Sample Population School type N (% of Total) SamPopEnrolment: Mean (SD) SamPopTeacher FTEs:Mean (SD) SamPopulation pie Elementary Secondary 314 (79. time. SMITH. resources. J. R.3) 209 (227) 216 (218) 252 (156) 221 (--) -154 (--) -113 (--) -- aDashes indicate dataareunavailable source.3) 362 (__)a 854 (--) Private Schools Elementary Secondary Elementary/ Secondary 44 (63.0) 80 (10.D. Table 2 displays the survey variables corresponding to each of the usual components of a logic model (needs. people. money.
Direct bullying -frequency (name calling. activities.. Finally.73) 2. the degree of bullying at our school is greater than the average) Changes in bullying over previous 3 months Changes in bullying over previous 1 year 3.63) 2 (4) . assault.88) 1 (5) 1 (5) . The objective of these analyses was to select the most parsimonious set of variables that accounts for unique Table 2 Framework Questionnaire of Conceptual Logic Model Components Survey Variables (Sample Items)a Mean (SD) Number of Scale Items (Scale Points) 1.96 (. In all regression analyses reported below. social exclusion.e.83 6 (5) .01 (. ignoring) Seriousness of bullying (bullying is a serious problem.52 (. two variables reveal how many different groups are targeted for interventions and the number of groups whose roles in creating and/or solving bullying problems are addressed in program activities.87 Scale Alpha 2. Using SPSS 13.15 (. 3. 5.73 3 (5) . resources. stealing) Indirect bullying -frequency (spreading rumours.79) 1. Needs 4. the outcomes component is comprised of a single 10-item variable that measures the impact of antibullying programming as observed by respondents. and outputs) function as predictors of a single criterion variable: impact of antibullying programs (variable 18).89) 2.0. we ran multiple regression analyses to test the statistical significance of the relationships among the variables comprising the logic model components. 2.IN ANTIBULLYING INTERVENTIONS SCHOOLS 749 component.89 (. the variables in the first four components of the logic model listed in Table 2 (i. needs.
51 (.78) 1 (5) 2 (4) . Activitiespeer-led antibullying committee.06) 2.67 (. Number of stakeholder groups involved in program delivery 11.61 (.42 (1.91 (.81) 3.75) 1.64) 1 (12) planning 10.92) 1 (5) .750 & J. SMITH. Amount of current programming compared to 1 year ago 1 (7) 1 (7) 24 (5) . 7. we commit substantial time and resources to solving bullying) Sufficiency of resources committed to antibullying initiatives to effectively deal with bullying Number of stakeholder groups involved in program 1.58 (. Number of different sources of program funding 12.71 (1. 8.D.92) 3. COUSINS. information for parents) 14.87 (1. Number of different sources of non-monetary resources 13.64) 1 (12) 4. Amount of current antibullying programming (individual counselling. R. regular classroom discussion.66) 1. Resources 9. Logic Model Components Survey Variables (Sample Items)a Mean (SD) Number of Scale Items (Scale Points) Scale Alpha 6.92 3. Changes in bullying over previous 5 years Amount of resources committed to antibullying initiatives (dedicating time and resources is highest priority. interventions. STEWART J.B.85 1 (4) 2.00 (.
1 (12) 8.ANTIBULLYING INTERVENTIONS SCHOOLS IN 751 Logic Model Components SurveyVariables(SampleItems)a Mean (SD) Number Scale of Scale Alpha Items (Scale Points) 15.school atmosphere is more positive and peaceful. Number of differentgroups whose roles in creatingor (2. families. bystanders. Impactsof antibullying 10 (4) 2. .95) programmingcompared years ago 16.) aSampleitems are truncatedto save space.70. whole grades) administrators.parents) 18.60.94 (. In the first step of the selection process. To accomplish this.84 variation in the criterion variable. Amount of current 1 (5) 4.92) targetedfor antibullying programming(students. Number of differentgroups 1 (12) 5.32 to 5 (. correlations among the predictor variables are all less than 0. Copies of the complete questionnaireare availablefrom the first author(JDS).39) programs(school personnel use more effective strategiesto Outcome stop bullying. we undertook a two-step process to select variables for a final hierarchical regression in which blocks of predictor variables associated with each of the logic model components are stepped into the regression model.numberof students who bully has decreased. we examined the zero-order correlations among the 18 study variables to assess the risk of multicollinearity among variables. As reported in Table 3.03 17.65) solving bullying are addressed in programs(bullies. teachers. and only two are above 0. leading us to conclude that the probability of violating .80 (2.
were retained for the second phase of the selection process.B.D. therefore. and all of these correlations were in the expected directions. was regressed onto the eight remaining predictors in a stepwise analysis. the criterion variable. Additionally. Table 4 displays the results of these analyses for each of the logic model components. we entered three blocks before this point was reached. proceeding one logic-model component at a time. and one within the activities component did not explain a significant proportion of variance in the criterion variable. therefore. program impacts. In the present analysis.752 & J. Therefore. COUSINS. we dropped these five variables.e. one within the resources component. Blocks were entered into the regression equation up to the point that the change in R2 did not reach statistical significance. and retained the remaining eight predictors for the final hierarchical regression analyses.. Results for two need variables show a significant negative association . p < . These 13 predictor variables. Among the six predictors included in the final equation. In the second step. Variables comprising the last block did not add significantly to R2. These results indicate that variables in the needs component account for the highest overall percentage of variance in the criterion and that outputs contributed the least. we identified 13 of the 17 predictor variables as having statistically significant correlations (i. Predictor variables that did not account for unique variance in the criterion when the regression terminated were eliminated from subsequent analyses. beta weights corresponding to four of the variables reached statistical significance. all variables within each component were stepped sequentially into a regression equation beginning with the predictor having the highest correlation with the criterion. Additionally. STEWART assumptions of multicollinearity in the regression analyses is low. We entered variables within each logic model component simultaneously as a block beginning with the block with the largest corresponding R2 value in the previous regression analyses and proceeding in order to the block with smallest R2 value. SMITH. In the final phase of our analyses.05) with the criterion variable. they are not included in the final regression model. J. one variable within the needs component. The results of the third and final step of this hierarchical regression are displayed in Table 5. R.
and. In light of the urgent need for effective solutions to bullying problems.In specific terms. However.ANTIBULLYING INTERVENTIONS SCHOOLS IN 753 between program outcomes and changes in bullying levels over the previous 1. the investment of time. on the other. we found that school principals who reported larger reductions in bullying over the previous one to five years. who have the difficult task of deciding where to invest limited. This means that predictors included in the final regression model account for a substantial42. (2004). and higher amounts of antibullying programming tended to report better outcomes for the antibullying interventions implemented in their schools. on the one hand. at least from the perspective of school principals. The results also show a strong. Smith et al. D. who found that a large majority of the documented outcomes of whole-school programs were of either of negligible size or . effort. The relationshipbetween the predictors and criterionin the final model was strong. it is important to note that these findings are generally not consistent with the findings of other studies on schoolbased antibullying and anti-violenceprograms. with a multiple correlationof 0. resources.9 per cent of variancein the criterionvariable.665. by implication.and 5-year intervals suggesting that respondents who reportedlower levels of currentbullying as well as decreases in bullying over the previous one to five years tended to report better outcomes for their antibullying programs. DISCUSSION In this exploratory study. these findings are encouraging for school officials.For example. These results lead us to tentatively conclude that. and money in schoolbased antibullying initiatives can yield valuable returns by helping to create school environmentsthat are safer and more peaceful for children. more conducive for learning and healthy development. sufficient resources currentlydedicated to resolving bullying. J. and the degree of positive outcomes reported by respondents. positive relationshipbetween the sufficiency of resources to resolve the problem and the amount of programming offered in schools. we investigated the relationship between aspects of programs implemented in schools to reduce bullying and victimization and perceptions about the outcomes of these programs. and often meagre.
05 -.04 .11 -.10 .19** 2 3 4 5 6 7 8 9 10 11 .35** -.18** .22** .26** .11 Notes: *p < 0.09 .59** -.38** .17* -.09 .40** -.45** .27** .25** .11* .16** .09 -.13* .45** -.20** .33** .12 -.01 -.07 .27** .21** .10 .01 .43** .05 **p< 0.15** -.31** .11 -.41** -.15** -.00 .12* -.03 .16** -.07 .22** -.IN INTERVENTIONS SCHOOLS ANTIBULLYING Table3 Zero-orderCorrelationsbetween Survey Variables Component Variable 2 3 Needs 4 5 6 7 8 Resources 9 10 11 12 13 Activities 14 15 16 Outputs Outcome 17 18 1 .26** .18** .36** .11 .00 -.11* .37** .05 -.23** .19** .46** .03 .27** -.12* .10 .16** -.24** .17** .18** -.06 .15* -.06 .01 .20** .15** .07 -.37** .13* -.02 -.43** .16* -.12 .13* .14* .10 -.27** .05 -.11* .27** .22** -.12* .05 .35** .18** .04 -.11* .14* .00 .52** .13* -.12* -.01 .40** .09 -.04 -.13* -.05 -.13* .07 -.00 -.06 -.45** .18** .27** .11 .17** .39** .14* .21** -.19** .29** .16** .01 .07 .20"* -.21** .21** .23** -.15** .20** -.19** -.06 -.61** .51** .34** .14* .02 .66** .
755 J.201*** .102 .219 .017 (N) Needs (223) Resources (300) Activities (320) Outputs (320) 13. Number of groups targeted for programming Notes: *p < 0.055 .01 ***p< 0. Amount of resources .00 .075*** . Amount of current programming 17. Number of groups involved in program delivery 7.135 -.229 .029** .090 .129*** . Sufficiency of resources 10.067*** .147* .028 . Component Table4 Stepwise Regressions of Predictors(within Components)on ProgramImpa Predictor Variables AR2 B 5.016* -. Changes in bullying over previous 5 years 8.049*** .05 **p< 0. Changes in bullying over previous 1 year 6. Number of groups whose roles are addressed 16.
STEWART Step 1 Table 5 Predictors Program on (N=213) of Regression Stepwise Impacts PredictorVariables AR2 B SEB 5. SMITH.042 -.241*** .013 . Howard. Amount of current programming < Notes: **p 0. J.021 .031 . only modest intervention effects.010 .. Changesin bullying over previous 1 year 6.g. at best.756 & J.032 .013 .108*** . see Orpinas.089 . however.287*** 3 13.250*** . Amount of resources .274*** -.210** 2 8.023 -.B. COUSINS.D. see Olweus. concluded that the effectiveness of this approach to solving bully/victim problems is not yet established.174 .030 -. Changesin bullying over previous 5 years -. most notably the Olweus antibullying program developed through the pioneering work of Norwegian researcher.001 < were negative. Flora. There are. isolated exceptions to this pattern.140 . Number of groups involved in program delivery 7.058 -. Sufficiencyof resources 10. .141 . R. and Griffin (1999) synthesized the research on violence prevention programs in schools and found that these programs achieved. 1993. Dan Olweus (e.048*** .01 ***p 0. and for a more recent example.
at least in terms of their experiences with bullying and antibullying programs. although our initial sampling frame was large and assembled to reduce sampling bias. survey respondents on average indicated that bullying tends not to be a serious problem in their schools and that the resources dedicated to antibullying programs are substantial and sufficient to deal with their bully/victim problems. The challenges include the lingering resistance of educators and parents about the seriousness of bully/victim problems. It is possible. There are a number of possible explanations for the inconsistency between our survey findings and the trends in the literature. Limber (2004) described a number of challenges that face educators who wish to implement the Olweus program in Americanschools. As a case in point. In the first place. This positive assessment contrasts with the bleaker picture of bullying in Canadian schools exposed by a recent international study (Craig & Pepler. they offer only one perspective on these issues. & Staniszewki. that our informants'assessments of bully/victim problems in the larger school context arise less from direct observations than from other . 2003). we had a low response rate. Therefore. whereas previous evaluations studies have used self-reportsfrom school children as the primary data source. less classroom time for bullying prevention. and these may explain why whole-school programshave not been as successful in North America as they have in Norway.we cannot be certainthat our sample is representativeof most schools in Ontario. 2003). Although principals are informed sources with knowledge about antibullyingactivities in their schools. and the detrimentaleffects of widely used group treatmentsand peer mediation programs. Therefore. we cannot rule-out the possibility that our sample is biased in favour of schools that have invested heavily in antibullying programming and have experienced success at levels that surpass provincialnorms. and our findings and conclusions are necessarily constrainedby this restrictedpoint of view. the use of simple.ANTIBULLYING INTERVENTIONS SCHOOLS IN 757 Home. piece-meal approaches. the larger size and complexity of American schools. and we consider several of the more viable of these possibilities at this juncture. A second possibility is related to the fact that our survey was addressed to school principals. for example.
B. a piece-meal approach with inadequate resources will likely have little impact on bully/victim problems and may leave students. the data support the view that solutions to bully/victim problems require adequate resources because the schools' ratings of a variety of program outcomes were directly related to the amount of programmingand the sufficiency of resources brought to bear on the problem. This finding suggests that programsneed time to penetratea school culture and influence attitudes and behaviours of students and school personnel. R. such as discipline referral patterns. Implications Practice for Our data point to several implications related to the implementation of antibullyingprogramsin schools. Fullen. Implications Research for Although many consider bullying to be a serious social problem in schools and communities and schools across Canada are implementing .2004).Otherresearchshows that perceptionsof bullying problemsvary accordingto who is asked. 2001).e. Additionally. information that may bias their assessments. parents. SMITH.or the opinions of other school personnel. but over longer intervals. This notion is consistent with substantial literature on educational programming in general (e.g. and school personnel frustrated about insufficient progress (Limber. which precludes making causal inferences with certainty about relationships among study variables. such as one to five years.Pellegrini and Bartini(2000)suggested that children have greater access to information about bully/victim problems than adults. The data show that the positive impact of programs did not appear in the short-term(i. given that all items on the questionnaire are of similar format and are answered by the same respondents. STEWART J. since bullying is usually perpetratedwhen adults are not present. the multiple demands for services in the face of limited resources place school officials in difficult positions. Undoubtedly.D.. By way of explanation. it is importantfor readers to keep in mind that our findings and conclusions are limited by design of this survey..758 & J. However. COUSINS. 3 months). Finally. The study was cross-sectional and correlational. our analyses are confounded by shared method variance. Additionally.
In the future it would be informative if researchers used layered evaluation designs to disentangle the program effects that result from implementing program in particular classrooms. The authors acknowledge Jay Poitrasfor his contributionsto data collectionfor this study.. administrators. Ottawa. P. ACKNOWLEDGEMENTS This researchwas supposed by a grantto the first authorfrom the SocialSciences and Humanities Research Council of Canada. Additionally. Therefore. K1N 6N5.. Clearly. and effective solutions to bullying will only be identified through continued research and program evaluations that are commensurately complex. bullying is a complex social problem. Stn.ca .g.g. scholars need more refined knowledge to assist school officials in developing effective antibullying interventions that they can tailor to the different needs of students. despite the research that indicates differences in bullying behaviours as a function of age and gender (e. Additionally. Send e-mail correspondenceto:jdsmith@uottowa.University of Ottawa. and students). Baldry & Farrington. David Smith. 2004. Box 450. researchers should also collect information about the contexts in which these interventions occur. Ontario. very little research addresses how antibullying interventions may differentially influence subgroups of students. Making sense of the relationships between the variety of antibullying interventions delivered in schools and their putative effects is a complex task. and it is critical that this research consider the influence of student characteristics on the outcomes of antibullying programs. our first recommendation is for much more research and program evaluations on interventions for bullying in schools. we recommend that researchers collect data from a variety of key informants in the school community (e. researchers have conducted very little research on the effectiveness of these programs. To achieve this understanding. Faculty of Education. and particular communities. A.O. teachers. 2002). and much of the existing research is not rigorous.. Clearly. particular schools. Crick & Nelson. Address correspondenceconcerningthis article to: Dr.INTERVENTIONS SCHOOLS IN ANTIBULLYING 759 anti-bullying programs.
D. Overall. 2236. School Psychology International. R. W.B.5(4). 1-15. (2001). Pepler. & Atlas. (2004). 1 2 REFERENCES Anderson. Evaluation of an intervention program for the reduction of bullying and victimization in schools. 79-95. Crick. 299-310. Canadian 13. L.. N.21. & Farrington. and benefited slightly more from antibullying programs than public schools. Craig. Psychology. The new meaning of educational change. School-associated violent deaths in the United States. D. School Psychology International.. M. W. T. M.. (2002). 1994- 1999. 2695of 2702.. L. CanadianJournalof Psychiatry . private schools indicated that they received fewer reports of bullying. et al. M. Cowie.Revue Canadiennede Psychiatri. SMITH. D. Simon. COUSINS. Identifying and targeting risk for involvement in bullying and victimization. Observations of bullying and victimization in the school yard. Journal of Abnormal Child 30.. Unfortunately. J. STEWART A copy of the questionnaire may be obtained from the first author. & Ryan. J. R. D. W. W.. Relational and physical victimization within friendships: Nobody told me there'd be friends like these.. Craig. offered fewer interventions. R. Kaufman. M. G. J. 21. P. Aggressive Behavior. the number of participating private schools (N=60) did not permit more detailed comparison between public and private schools. C. & Pepler. 41-59. The role of peer support in helping the victims of bullying in a school with high levels of aggression.D. D. H. 286(21). (1997). R. (2001)..... & Pepler. 599-607. 48. (2003). D. A. J. (2000). & Taradash. Fullan.. Observations of bullying in the playground and in the classroom. A. Craig. J. New York: Teachers College Press. . 577-582. & Olafsson. ChildMaltreatment. Journal School of Psychology. Pepler. Dating experiences of bullies in early adolescence. Paulozzi. JAMA:Journal the American MedicalAssociation. (2000).. R. dedicated fewer resources to solving bullying. M.760 NOTES & J.. (2000). & Nelson. Connolly. Barrios. 30. A. Craig.. Baldry.
G. (2003).. 431444. Ruan. . M. L. Olweus. (2005). 351-364). & Griffin. (1993). P. Pellegrini. J. A. Limber. Josephson. M. R. 360-366. S.. & Schuster.edu. Graham. & Cartwright.. G. 112. the weak. S. M. A. Mahwah. Bullying at school: What we know and what we can do. 197-215. M. D. (2003).... A model of depressive symptoms in gay men.. Quick facts: Ontario schools. P. Malden. R. M. Journal of EducationalPsychology. 285. Unpublished doctoral dissertation. (2000). University of Ottawa. In D. Swearer (Eds. Pilla. W.. 2006 from http://www. Jama:Journal of the American Medical Association. B. Orpinas. A.42.on. P. & Staniszewski. 2000-2001. 141-156. Home. 2094-2100. Howard. (2000)... 1231-1237. Washington. Flora. Bullying in 25 secondary schools: Incidence. Violence-preventions programs in schools: State of the science and implications for future research. (2004). Bullying in American schools: A social-ecological perspective on prevention and intervention (pp. Bullying among young adolescents: The strong. N. 92. P. S. 8. J. Ontario Ministry of Education (2003).html.ANTIBULLYING INTERVENTIONS SCHOOLS IN 761 Glover. Applied and Preventative Psychology.. impact and intervention. Juvonen. A. MA: Blackwell. M. NJ: Lawrence Erlbaum. & Scheidt. M. D. (2001). (1999). Overpeck.. Orpinas... K. An empirical comparison of methods of sampling aggression and victimization in school settings. Bullying prevention: Creating a positive school climate and developing social competence. Johnson. School bullying: Changing the problem by changing the school. Bullying behaviors among US youth: Prevalence and association with psychosocial adjustment. 32. D. T. DC: American Psychological Association. Simons-Morton. A. Implementation of the Olweus Bullying Prevention Program in American Schools: Lessons learned from the field. D..). M. Educational Research. Pediatrics.gov. School Psychology Review. & Home.. and the troubled. & Bartini. Retrieved January 4.ca/eng/ educationFacts. Gough. Espelage & S. (2004). Nansel. School of Psychology.
25.. P. New York: Guilford Press. (1991). L.Statistical principles experimental designs(2nd ed. Journalof SchoolPsychology. in Winer. Bullying is power: Implications for school-based intervention strategies. Educational Research. 6. H. & Egan. P. Program logic models: Expanding their role and structure for program planning and evaluation.). (2003). & Rothman.22. Salmivalli. Vaillancourt. (2004). J. Smith. 95-106. Canadian Journal of de 48. COUSINS. S. who benefited? EducationalResearch. 583-590.B. 453-459.762 & J. (2003) Interventions to reduce school de Journal Psychiatry RevueCanadienne Psychiatrie. D.. STORIES: A classroom-based program to reduce aggressive behavior. D... K. 33. & Ogbourne.157-176.. (2001). 263278. Canadian of 48. Program Salmivalli. 44(1). Schneider. K. Roland. 97-111. R. Bullying as a group process: An adaptation of the participant-role approach. & Ananiadou. Journal of Applied School 19(2). P. K. (2001). Smith. p.Peerharassment school: plightof the vulnerableand victimized (pp 73-104). A. & Cowie. P. 591-599. Canadian Journal of Evaluation.43. K. Determinants of chronic victimization by peers: A review and new model of family influence. (1971). The effectiveness of whole-school antibullying programs: A synthesis of evaluation research. Rush. T. Aggressive Behavior.D. 39. V. & McDougall. G. Bullying. S. Hymel. Rigby. & Smith. Juvonen & S. K. (2003). Teglasi. 71-94.. bullying. Graham(Eds. K. STEWART Perry. Smith. (1999). B. J. B. J.. (2002). 548-561. C. E. J. H. C..55-67. Consequences of bullying in schools. K.. Journalof Adolescence.. Sutton.. Psychology. New York: McGraw-Hill.B... Peer-led intervention campaign against school bullying: Who considered it useful. depressive symptoms and suicidal thoughts. H. (1999).). Psychiatry RevueCanadienne Psychiatrie. Participant role approach to school bullying: Implications for intervention. SMITH. Ananiadou. In in The J. SchoolPsychology Review. (2001). E. 379 . Hodges.
This action might not be possible to undo. Are you sure you want to continue?