Journal of Educational Psychology © 2014 American Psychological Association

2014, Vol. 106, No. 4, 1135–1143 0022-0663/14/$12.00 DOI: 10.1037/a0036110

The Role of Teachers in Bullying: The Relation Between Antibullying
Attitudes, Efficacy, and Efforts to Reduce Bullying

René Veenstra Siegwart Lindenberg
University of Groningen University of Groningen and Tilburg University

Gijs Huitsing Miia Sainio
University of Groningen University of Turku
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Christina Salmivalli
This document is copyrighted by the American Psychological Association or one of its allied publishers.

University of Turku and Edith Cowan University

In order to battle bullying, it can be important for students to have teachers whom they see as taking an
active stand against bullying in terms of propagating antibullying norms and having an efficacious
approach to decreasing bullying. This expectation was tested with data from the control schools of the
Finnish evaluation of the KiVa antibullying program. Multilevel analyses of data from 2,776 fourth- to
sixth-graders showed that students’ perceptions of their teachers’ efficacy in decreasing bullying was
related to a lower level of peer-reported bullying. Students’ perceptions of their teachers’ efforts to
decrease bullying, however, was cross-sectionally related to a higher level of peer-reported bullying, but
over time was related to a reduction in peer-reported bullying. In classes where teachers were not
perceived as efficacious and had to exert a great deal of effort to reduce bullying, students with
probullying attitudes and without antibullying effort had the highest level of bullying. The current
findings show that teachers can play an important role in antibullying programs and should be seen as
targets of intervention.

Keywords: bullying, goal-framing, longitudinal, significant others, teachers

Bullying is a common phenomenon in primary and secondary and victims but also on bystanders (Polanin, Espelage, & Pigott,
schools. Because bullying is a highly undesirable kind of behavior, 2012; Smith, Pepler, & Rigby, 2004). The advantage of going
the most obvious kind of intervention against it would seem to be beyond the scope of the bully and the victim is that a whole-group
sanctioning bullies through loss of privileges, detentions, and intervention can make all persons responsible for everyone’s well-
suspensions (Anderson & Kincaid, 2005). Alternatively, one may being, teach teachers and students safe strategies to support and
focus on empowering the victim (Fox & Boulton, 2003) or on help victims, and change classrooms norms in such a way that
mediation between bullies and victims (Pikas, 2002). Some of bullying behavior becomes associated with low status and low
these interventions have indeed been successful. However, the affection.
insight that the social context of bullying is also important A great deal is known about the role that the peer group plays in
(Salmivalli, Lagerspetz, Björkqvist, Osterman, & Kaukiainen, bullying (Salmivalli, 2010), but less is known about the role of the
1996) has inspired new interventions that focus not only on bullies teacher. Understanding how teachers’ responses to bullying influ-

This article was published Online First February 24, 2014. Ministry of Education (Onderwijs Bewijs) to René Veenstra, and by a
René Veenstra, Department of Sociology and Interuniversity Center for Toptalent Grant (021.002.022) from the Netherlands Organization for
Social Science Theory and Methodology, University of Groningen, Gro- Scientific Research to Gijs Huitsing. In addition, the study is part of the
ningen, the Netherlands; Siegwart Lindenberg, Department of Sociology KiVa project, funded by the Finnish Ministry of Education and Culture, for
and Interuniversity Center for Social Science Theory and Methodology, developing an antibullying intervention program for the comprehensive
University of Groningen, and Department of Psychology, Tilburg Univer- schools. The project is led by Christina Salmivalli (Department of Psy-
sity, Tilburg, the Netherlands; Gijs Huitsing, Department of Sociology and chology) and Elisa Poskiparta (Centre of Learning Research) at the Uni-
Interuniversity Center for Social Science Theory and Methodology, Uni- versity of Turku. We are grateful to all children, their parents, and teachers
versity of Groningen; Miia Sainio, Department of Psychology, University who made the study possible, and all the colleagues working to realize the
of Turku, Turku, Finland; Christina Salmivalli, Department of Psychology, KiVa project.
University of Turku, and Department of Psychology, Edith Cowan Uni- Correspondence concerning this article should be addressed to René
versity, Joondalup, Australia. Veenstra, Interuniversity Center for Social Science Theory and Methodol-
This research was partly supported by grants from the Academy of ogy, University of Groningen, Grote Rozenstraat 31, 9712 TG Groningen,
Finland (121091, 134843) to Christina Salmivalli and from the Dutch the Netherlands. E-mail:


& Lussier. activate this norm. as the status goal is strong in potential bullies (Sijtsema often do not support victims because bullying incidents tend to et al. Volk. In an observational study of playground activity. Frank. 2009. 1136 VEENSTRA ET AL. causing students to take the norm seriously. Given a norm that bullying is process. 2007. For the normative goal to inhibit bullying. As attitudes. in lunch. which could result in motivations are intricately intertwined (Kruglanski & Köpetz. In our previous research on bullying. the temptation to that peers are often not willing to intervene (Espelage. Lindenberg 2013). superior (status) without feeling bad as a result of losing the Another possible reason that teachers are ineffective at reducing affection of other peers (Juvonen & Galvan. Goal-framing theory (Lindenberg. Albiero. For example. On the basis of goal-framing theory (Lindenberg 2008. & Steg.. 2005. their reports may be dismissed as particularly important mental constructs in which cognitions and unbelievable. ence the likelihood of bullying occurring may be critical for the have multiple witnesses (Salmivalli. 2008. is less likely to be offered when many individuals witness a 2013). The aim of “mild. and they expect victims to handle it on their own (Hektner definitely not the right thing to do. Lindenberg.” such as verbal abuse (Rivers & Smith. For one thing. 2007). & O’Brennan. Goal-framing theory suggests ple. 2014). in hallways. and given strong cues to & Swenson. 1996). 2013) applies the perceive teachers as unable to protect them (Novick & Isaacs. their efficacy. 1996). 1994). 2008. The harm this study was to test this conjecture by analyzing how the teach. peers watched without intervening. 2003). lying norm arises. because defenders have to confront powerful The Role of Teachers This document is copyrighted by the American Psychological Association or one of its allied publishers. 2008). and keep to the norm themselves. Teachers also “normative” goal makes students focus on relevant norms that often believe that bullying is part of a normative developmental apply to the situation they are in. Lindenberg. and other places where teachers are not around. especially overarching (Atlas & Pepler. 2012). it has been found that teachers tially important role of the teacher in reducing bullying. Dane. react negatively to infractions by others. & Salmivalli. quarter of the episodes did they help victims. Camilleri. we hypothesized that teachers who stand up against bully. & Phelps. whereas in another In the case of bullying. Newman. 2005. Students might monitor each other ing create an atmosphere in which more students find it easier to and expect that someone else will intervene. Below we derive hypotheses for this poten- aware of bullying. and Altoe (2008) argued that defending is a risky type of This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. or infer that as the abstain from bullying. a feeling of oughtness concerning this antibul- Gordon & Ladd. Salmivalli et al. Salmivalli et that the potential for peers to curb bullying may have been over- al. 2001. 2006.. and Craig (1999) found that in half of the ongoing social support is necessary. strong and O’Connell. 2005. Troop.. 2010) and the likelihood of development and successful implementation of an antibullying intervention might be reduced by the classic bystander effect: Help program. 1997. or that peers will find out. & Verloove-Vanhorick. teachers may not identify relational aggression as bullying that goals that are focused on the way one feels (like gaining status (Boulton. we found that it is aware of bullying. intervene in only 4% of bullying incidents on the playground (Craig & Pepler. To gain a better understanding of what may or may not help rooms. however. they did not intervene in one out of four cases especially useful to look at the role of goals. bullies and their assistants. most teachers who clearly stood for antibullying norms would be likely bullying consists of attacks that might appear to be relatively to strengthen their students’ goal to act appropriately. Henderson. (Kärnä et al. Kochenderfer-Ladd & Pelletier. 2008. 2007). of skill and courage to put a stop to their behavior. potentially harmful situation. 1998). Teachers curb bullying. 2011. 2012).. it also requires a great deal bullying in their class. insight from (social) cognition research that mental constructs 2010). and because peers may provide support for self- . Craig. Students nounced. But there is also evidence suggesting that teachers might estimated in the literature and that teachers’ potential may have be less effective in this regard. as students (Bradshaw. Only in a precarious (Keizer.. given a strongly activated normative goal. Thus it is no surprise that victims often goals. Oliver & Bullying has been found to be associated with a goal to feel Candappa. Newman 2009). 1998) and in only 18% of bullying incidents in The Role of Overarching Goals and Significant Others the classroom (Atlas & Pepler. or they may and affection) can be inhibited by the activation of the overarching perceive it as being less serious than physical and verbal bullying goal to act socially appropriately in a particular situation. 2012. Veen- bullying is that they often do not perceive bullying in the same way stra. There is a great deal of evidence for the effectiveness more like the bullies (Juvonen & Galvan. 2008. this precariousness is especially pro- quarter of the episodes peers even assisted the bullies. 2011. Lindenberg & Steg. Yoon & Kerber. Pijpers. Sijtsema. 1998). it cannot be so serious. This (Bauman & Del Rio. it might seem adaptive for The conjecture that teachers can do much to curb bullying is not students to avoid the company of low-status victims and appear at all trivial. & Murphy. 2007). Rather. 2000). there is also evidence Thus. and that goals are they report bullying incidents. and their effort was related to the level of bullies are often perceived as popular. All this suggests of peers in preventing bullying (Kärnä et al. Benelli. we expected that others do nothing. & bully is greatly reduced. Sawyer. prosocial behavior.. Gini. & Murray. The strength of this oughtness consists in the degree to The Role of the Peer Group which the activated overarching goal inhibits conflicting goals Despite evidence that peers can be effective at reducing bullying (Delgado. Polanin. As we explicate below. & Marini. For exam. they may not be been underestimated. Pepler. reprisals (Fekkes. The normative goal is quite bullying episodes. 2009). Research has also shown that students are concerned that if have to be activated in order to affect behavior. caused is mostly psychological and thus easy to explain away or ers’ stand against bullying in terms of their perceived antibullying construe as “only joking” (Teräsahjo & Salmivalli. Murray. In addition. This may be because bul- lying often takes place on the playground. Veenstra et al. Green. it is useful to take a closer look at possible mech- may also fail to take action: Even when they were judged to be anisms. 2003).

teachers are significant others (Veen. Simply reiterating the antibulllying norm is thus not speaking schools but excluding schools for special education. Caucasian). It stands to reason that teachers cruit students. They distributed the passwords to the students. those imple. Several classes where students perceive their teachers as taking an active examples covering different forms of bullying were given. All students in a intervention program and sampling procedures). we expected that the effects of students’ own formulated in the Revised Olweus Bully/Victim Questionnaire antibullying attitudes. who used into decreasing bullying. we were able to use all 2. The students were assured that of arguments derived from goal-framing theory that teachers who their answers would remain strictly confidential and would not be clearly and actively stand for antibullying norms are likely to revealed to teachers or parents. Lindenberg. These schools were stratified by province of the normative goal and its activation are significant others who in the mainland of Finland (five provinces). efficacy. Because boosting their role as such significant others with regard to an the KiVa intervention would (theoretically) fundamentally change antibullying norm. So far. Sijtsema. 2011). whereas teachers expect students to deal with bullies ditionally. Outcome variable: Peer-reported bullying. we focused on stu- tested empirically. this is only likely to help if the teachers take an active characteristics was at most 15. THE ROLE OF TEACHERS IN BULLYING 1137 justification of aggressive behavior (Caravita. this emphasizes the repetitive nature of bullying be less closely related to the peer-reported levels of bullying in and the power imbalance between the bully and the victim. The process was ad- The Present Study ministered by the teachers. The number of program evaluation (see Kärnä et al. The enough. If these expectations bear out. this and an enrolment form. who were provided with detailed in- structions concerning the procedure about 2 weeks prior to the data We expected that the level of bullying would be lower when collection. as a by-product. To re- Lindenberg. ages 9 –11 After the numbers of nominations students received had been years) and May 2008 (Grades 4 – 6. To recruit added up. and bullying. we hypothesized on the basis them to log in to the questionnaire.7% of the target sample menting an antibullying intervention would do well to focus on received active consent to participate (Kärnä et al. this can solve the playful way. fighting between students of equal strength) was also problem of each party shifting responsibility to the other party. more likely Measures to signal lowering of status and affection for behavior that goes against the norm. ages 10 –12 years).5% boys.8%. Details concerning the percent- stand against bullying by demonstrating antibullying beliefs. 1996). and if this is true. to remind the students of the meaning of the term by themselves (Crothers & Kolbert. 2.444 schools.. 2010). for more details on the nominations students could make was unlimited. 1996). including Swedish- & Gini. Teachers read the definition out loud. and age of missing data and common missing data patterns in the data efficacy as well as effort in the enforcement of rules against set are discussed extensively elsewhere (Kärnä et al. bullying. and 78 schools were stand for situationally relevant norms (Baldwin & Holmes. in provided. Ad- victimization. Tinga. they are likely to create the required strong multiple imputation (Royston. Out of 3. What is important in this significant-other effect is that students mechanisms before any intervention. 1987. the proportion of immigrants was status and affection. and effort to battle bullying would (Olweus. while students that students expect teachers to ensure their safety against peer could read the same definition from their computer screens. randomly assigned to intervention and control conditions. Students were Sample asked to indicate which classmates (a) start bullying. in the classroom.5% were girls and 50. 2006 to all schools in mainland Finland.2%. including nonparticipating classmates. In addition. activating the normative goal of the bystanders. In this way. this study were collected in May 2007 (Grades 3–5. These norms need to be strongly and continuously acti. and thus the temptation to bully) and by 0. The proportion of missing data for student bullying. but in the case of bullying at both waves. strengthen the normative goal of potential bullies and bystanders. Among the most important supports participate in the study. (b) make others join in the bullying. After bullying. Students filled out Internet-based questionnaires in the schools’ computer labs during regular school hours. Luckily.. For most students. students who had obtained parental permission to participate in the had a high degree of efficacy in battling bullying. dents from the control schools only to investigate the “pure” This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.e..776 and ongoing support by continuously activating the normative goal cases. 2005). Shah. see teachers as figures of authority who clearly stand up against Our final sample included 31 schools.. 2013. who are then more likely to support potential victims and. 2004). yielding scores from . 2014). however.. and (c) always find new ways of We used two waves of data collected for the KiVa antibullying harassing the victim (Salmivalli et al. The term bullying was defined to the students in the way In other words. a shortened version of the definition appeared on the upper part of the computer screen while they responded to Method bullying-related questions. 49. and this takes an active agent. this proposition has not been the associations between our study variables.776 students. students and explanation of what is not bullying (teasing in a friendly and teachers can work together at reducing bullying. 2011). 91. letters describing the KiVa project were sent in the fall of number of respondents per class into account. This document is copyrighted by the American Psychological Association or one of its allied publishers. We had data for all students on peer-reported stra. 2003). their guardians were sent letters containing infor- could be such significant others. 144 classrooms. The data used in class could be nominated. 279 volunteered to agent is potentially available. proportions were calculated to take differences in the schools. & Ormel. Most students were of potential bullies (thereby also inhibiting the competing goals of native Finns (i. In total. Rambaran. letters included information about the goals and content of KiVa vated. mation and a consent form. The teachers received individual passwords for all teachers as well as students had a stronger antibullying attitude. and put effort study. 2011. An stand against bullying.

06 3. on a scale from 0 to 3. decrease bullying (“How much has the teacher done in order teachers’ antibullying attitudes were 3.81 and 1.34 .00 0. The number of nominations students could make was tions between students’ and teachers’ characteristics. students were asked to indicate which classmates (a) com. Efficacy children . family breakup.23 0.26 ⫺.05 2.34 . To measure students’ effort to decrease bul.26 0.09 . four 2.46 2. Efficacy teachers . Internal consistency analyses were necessary to control for the nested structure of the (nine items) was .13 ⫺.14 .07 0. not effects changed over time (from the pretest to the posttest).01 .18 . grade. we controlled for Swedish- .20 0. all continuous variables including the dependent vari- from 0 to 1.03 . class level). and (c) effort to .66 3.45 5.36 0. Students’ perceptions of their teacher’s attitudes. To derive students’ scores on antibullying attitudes. & Salmivalli. highly reliable.67 ⫺. The teacher variables as perceived by students were Table 1 Descriptives of and Correlations Among the Study Variables at the Pretest (Above the Diagonal) and Posttest (Below the Diagonal).06 .12 .39 at the pretest and 3. This scale was developed by Salmivalli et al.93. items regarding attitudes to bullying (“I feel bad seeing a child bullied”).72 1.79 at the posttest.12 Note. and teachers’ effort to decrease bullying was a teacher was calculated for each of these three measures.03 1.00 does your teacher think of bullying?” ranging from 0 ⫽ a good at the posttest on antibullying attitudes. proportions were calculated to take differences in regression analyses and to obtain standard errors of the same the number of respondents per class into account.07 .39 0. age.37 0. The was the dependent variable. and for that reason were Student’s attitudes. and (c) We started the multilevel analyses with a model of the main asking others to stop bullying or saying that bullying is stupid.19 0.81 0. Attitude children . Students were asked to evaluate how easy within schools (Snijders & Bosker. .66 ⫺. Results and effort.12 ⫺. Browne. The diagonal is in bold.57 . efficacy. The efficacy of teachers was 2. efficacy. behavior (Pöyhönen.43 . 2012). and Stability Coefficients at the Diagonal Pretest Posttest Variable 1 2 3 4 5 6 7 M SD M SD 1.14 .14 .23 at the posttest. Effort children .09 ⫺.12 .29 0. & Goldstein. encouraging him or her to report the bullying to the teacher.31 .82 at the posttest.06 . effects of students’ and teacher’s attitudes.01 ⫺.91 and .20 at the pretest and 3.38 at the pretest and to 4 ⫽ a great deal). Juvonen. ables were z standardized (M ⫽ 0.05| are significant at p ⬍ . victimization (“Kids who are weak are just asking for Analyses trouble” [reversed]). . we also included at the individual level interac- bullying.50 7.10 .26. efficacy. efficacy. data: Two assessments among individuals nested in classrooms Students’ efficacy. 3 ⫽ very directed at peer-reported bullying. This scale was based on Rigby and Slee We performed multilevel regression analyses using MLwiN (1991). language schools (at the school level). gender.14 4. 4 ⫽ agree completely) to of the study. The predictor variables were at the This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.65 . (“How much can the teacher do in order to decrease bullying?” The proportion score of students’ effort to decrease bullying was ranging from 0 ⫽ nothing to 4 ⫽ a great deal).77 at the pretest and .78 2.09 . To control for possible compo- easy). and (c) try to make others stop the third model. class size (both at the Students’ effort. 2010) were (a) trying and effort) and individual level (students’ attitude.04 . In the second model we tested whether or fort victims or encourage them to tell the teacher about bullying. three items used to measure self-efficacy beliefs for defending classroom (students’ perceptions of the teacher’s attitude.09 . This document is copyrighted by the American Psychological Association or one of its allied publishers. Peer-reported bullying or difficult it would be for them to defend and support victims.65 .29 at to decrease bullying since last autumn?” ranging from 0 ⫽ nothing the posttest.46 2.82 on efficacy. the students’ perceptions of teachers were based on the was . and effort Students responded on a 4-point scale (0 ⫽ very difficult. yielding scores magnitude.17 0.001.76 0.32 ⫺.66 .74 . The average score of students’ perceptions of 2.12 . and immigrant status (all lying.28 . Steele.12 0. Internal consistency (three items) at the pretest and posttest average.07 0. and to make others stop bullying. 2012).05 . Three single items were used to measure students’ perceptions of their teachers’ (a) attitude toward bullying (“What Table 1 reveals that students scored 3.21 .82 0. perceptions of 19 students per class. (b) comforting the bullied person or effort).07 ⫺. at the individual level).72 3. Attitude teachers . In (b) tell others to stop bullying. on a scale from 0 to 4.44 6. (b) efficacy in decreasing bullying Students scored 1.17 at the posttest. All correlations larger than |.39 3. According to the students. 1138 VEENSTRA ET AL. (1996). These questions were developed for the purposes type scale (0 ⫽ disagree completely. Effort teachers .02 2.91. 0 to 1.23 (Rasbash. and defending (“It irritates me when nobody defends a bullied child”).19 at the pretest and .69 at the pretest and sition differences between schools. SD ⫽ 1) across the whole Internal consistencies (three items) at the pretest and posttest were sample before being entered into the multilevel model.01 ⫺.22 3. Peer-reported bullying ⫺.51 . unlimited.37 and 2.38 0.48 . Students responded on a 5-point Likert. The multilevel negatively keyed items were reverse coded. After the numbers of nominations students received had To facilitate the interpretation of the results of the multilevel been added up.08 . Internal consistency (three items) was .14 0. thing to 4 ⫽ absolutely wrong). On 2.

The predictor variables were at the classroom (students’ perceptions of the teachers’ attitude. respect to the relation between peer-reported bullying and stu- was positively related (b ⫽ 0.025 (.031).4 (2)ⴱⴱⴱ 30.054 (.011)ⴱⴱⴱ Decrease in deviance (df) 295.000) . For the students’ perceptions of their teachers.006 (.003 (. dents’ efficacy.013)ⴱⴱ Teachers’ Effort ⫻ Students’ Effort ⫺. The proportion score of related to lower levels of peer-reported bullying (b ⫽ ⫺0. and effort) and individual level (students’ attitude. Teachers’ effort at the pretest This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.023 (. grade (both at the school level).210 (. efficacy. but it was related to a lower level of peer-reported bullying effort.051).008)ⴱⴱ .039 (.024)ⴱ ⫺.093 (.217 (.088 ⫺ 0. Table 2 Multilevel Analyses: Fixed and Random Effects for Peer-Reported Bullying (N ⫽ 31 Schools. and immigrant status (all at the individual level).104 (.211 (.05.402 (.014)ⴱⴱⴱ ⫺. and pretest.207 (.022 (.053). 144 Classrooms.021)ⴱⴱⴱ . For all other lying attitudes.009)ⴱ . gender.006 (.056 (. efficacy.076).008)ⴱⴱ ⫺.015)ⴱⴱⴱ .023 (.024)ⴱ Effort .750 (.021)ⴱⴱⴱ Random part for change School variance . at the posttest (b ⫽ 0. This is not high enough to indicate multicol- linearity.76 for reported bullying (b ⫽ ⫺0.076 (. and effort to decrease bullying as well as students’ perceptions of their teachers’ attitude. Teachers’ effort to decrease peer-reported bullying.000 (.014)ⴱⴱⴱ ⫺.000) .07 at both assessments.0 (6)ⴱⴱⴱ 29.217 (.097 (.009) ⫺.31 for students’ efficacy to .013)ⴱⴱⴱ Random part for intercept School variance . bullying was . On the diagonal it is shown that the stability of antibullying attitudes) was related to a lower level of peer- the variables ranges from .096 (.000 (. however.030)ⴱⴱⴱ ⫺.008)ⴱⴱ Individual variance .014)ⴱⴱⴱ Students’ perceptions of teachers’ Attitude .012)ⴱⴱⴱ ⫺.000) . ⴱⴱⴱ p ⬍ .000 (.001.028)ⴱⴱⴱ Change at postassessment Teacher’s effort ⫺.000 (.000) .000) Class variance .74 with each other.030)ⴱⴱⴱ Interactions Teachers’ Effort ⫻ Students’ Attitude . The effect for Grade 3 students differed over time. we found that The bivariate associations between the study variables are teachers’ perceived efficacy (which correlates highly with their shown in Table 1. ⴱⴱ p ⬍ .011)ⴱⴱⴱ Intercept-change covariance School variance . Standard deviations are in parentheses except where noted.008)ⴱⴱ ⫺.028)ⴱⴱⴱ .022 (.009)ⴱⴱⴱ Teachers’ Efficacy ⫻ Students’ Effort .009) Effort ⫺.003 (. family breakup.020 (.399 (.023) ⫺.022 (. we found no difference between the effect at the (b ⫽ ⫺0.025 (.023) Efficacy ⫺. The teacher variables correlate from .404 (.009)ⴱ Class variance .000 (.000 (.119 (. and 2.011)ⴱⴱⴱ Efficacy ⫺. 55% of the variance.9 (3)ⴱⴱⴱ Note. Also.01.758 (.008 (. .034 (.011)ⴱⴱⴱ .006 (. whereas all other control variables did not vary. The findings reveal that the stronger students’ antibul.051 (. indicating that these variables share 32% to reported bullying (b ⫽ 0.216 (.023) ⫺.57 to bullying.024)ⴱⴱ ⫺.008)ⴱ . the less they were reported by peers as bullies characteristics.026)ⴱⴱⴱ . students’ effort to decrease bullying was pretest and the effect over time. efficacy.128 (.012)ⴱⴱⴱ ⫺.011)ⴱⴱⴱ . class size (at the class level). In all models.021)ⴱⴱⴱ .088) to peer-reported bullying at the This document is copyrighted by the American Psychological Association or one of its allied publishers.014)ⴱⴱⴱ Individual variance .757 (. and effort). was related to a higher level of peer- .053 (.217).119 ⫽ ⫺0. Modeling Change Main Effects In the second model in Table 2 we tested whether or not effects Table 2 shows the results of the multilevel analyses with changed over time.011)ⴱⴱⴱ ⫺.014)ⴱⴱⴱ .009) ⫺.008)ⴱⴱ Individual variance ⫺. THE ROLE OF TEACHERS IN BULLYING 1139 all measured on a scale from 0 to 4.055 (.011)ⴱⴱⴱ ⫺. The only change effect we found was for teachers’ effort to decrease bullying.776 Children) Variable Main-effects model Change model Interactions model Fixed part Students’ Attitude ⫺.000) Class variance ⫺.060 (.088 (. age. ⴱ p ⬍ .051 (.021 (.008)ⴱⴱ . we also controlled for Swedish-language schools.

Graphical presentation of the interaction between students’ exerted a low level of effort (b ⫽ ⫺0. see Figure 2).4 Children's Effort Children's Effort Figure 2. peer-reported bullying. deviation). compared with 0.001).4 reported bullying was stronger in classes where the teachers had a Peer-Reported Bullying low level of efficacy (b ⫽ ⫺0. By contrast. when teachers 0 were low on efficacy (⫺1 standard deviation) and students’ effort was also low (⫺1 standard deviation). p ⬍ .16 at the class level and ⫺.38 at the student level.6% at the class level.2 0. .001). Graphical presentation of the interaction between students’ Figure 3. model in Table 2 was 20. We see in Figure 1 that students’ antibullying attitude was less closely related to peer.1 0 0 -0.001). p ⬍ Teacher (-1 SD) Teacher (+1 SD) . The interactions are depicted in several figures and are based on the interaction model shown in Table 2. of effort to stop bullying (b ⫽ ⫺0.1 ing effect on peer-reported bullying.001. The effect of students’ effort to decrease bullying on peer- 0.4 0.2 -0. teacher efficacy appeared to have a dampen- 0. ns). At low levels of students’ effort (⫺1 standard deviation)..6% at the individual level and 1. teachers’ efficacy made no -0.3 0. p ⬍ . The intercept-change covariance can be used to calculate Interaction Effects the correlation at the class and student level: These correlations are ⫺. 1140 VEENSTRA ET AL.1 -1 SD +1SD -0. Smith Low Efficacy High Teacher (-1 SD) EfficacyTeacher Low Effort High Effort (+1 SD) Teacher (-1 SD) Teacher (+1 SD) 0.4 students’ effort on peer-reported bullying was stronger in classes Children's Attitudes where the teachers exerted a high level of effort to decrease bullying (b ⫽ ⫺0.273. 0.090. when teachers’ efficacy was strong (1 standard This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.3 classes in which teachers had a high level of efficacy (b ⫽ ⫺0. indicating We next tested interactions between teachers’ and students’ that a high level of peer-reported bullying at the pretest is associ- characteristics.3 0. bullying was lower. p ⬍ . Three out of the nine interactions were significant ated with a change toward less bullying over time (regression to (using the Holm–Bonferroni method to correct for multiple test.1 0. Discussion reported bullying in classes where the teachers exerted a great deal The social context of bullies and victims has increasingly been the target of antibullying interventions (Polanin et al.3 difference to peer-reported bullying.3 -0.001) than in classes where the teachers Figure 1.251. The decrease in deviance was significant for each model. 2012.1 -1 SD +1SD -0.012. compared with classes Low Effort High Effort in which teachers exerted a low level of effort (b ⫽ ⫺0.2 p ⬍ .161.2 -0. the mean). antibullying attitudes and students’ perceptions of their teachers’ effort in The explained variance in peer-reported bullying in the full relation to peer-reported bullying. Graphical presentation of the interaction between students’ effort and students’ perceptions of their teachers’ efficacy in relation to effort and students’ perceptions of their teachers’ effort in relation to peer-reported bullying.4 Peer-Reported Bullying Peer-Reported Bullying 0.4 -0. bullying effort (1 standard deviation). Figure 3 shows that the effect of -0.2 0.2 This document is copyrighted by the American Psychological Association or one of its allied publishers. peer-reported bullying was -0. At higher levels of students’ anti- -0. ing).3 -0. Contrary to our expectations.183.1 -1 SD +1SD at its highest levels.

Bradshaw. 86 –99.98. Our to decrease bullying was more strongly related to peer-reported results show that students and teachers can work together at bullying in classes where students perceived teachers to have low reducing bullying. If participate in the program is not likely to be a disturbing factor. THE ROLE OF TEACHERS IN BULLYING 1141 et al. doi:10. all persons in a classroom are made respon- battle it. C. For students. 2004). teachers are seen to be efficacious. correlates highly with their antibullying attitudes. (2007). 52. Journal of Educational Psychology. goal-framing theory (Lindenberg. 0663. Atlas. 36. The finding for responsibility to the other party. L. (2005). ensure their safety against peer victimization. R. and we found In sum. Yet. teachers norm. probullying attitudes and without antibullying effort had the high- ening their goal to act appropriately. M. attitudes Strengths and Limitations and ability to cope.1037/0022-3514. 92. The strengthen antibullying norms and their activation in students. Journal of Educational Research. they can also help to exert a great deal of effort to stop bullying.2044-8279. in that way help inhibit goals that encourage bullying. but ing bullying. (2006). A. G. Observations of bullying in the These results imply that with regard to bullying.1997. Understanding how teachers’ responses to bullying (in the eyes motivated to implement an antibullying program. Our study also has some limitations.1080/ for students is a class in which the teacher is perceived by the 00220679809597580 students as having a high degree of efficacy in battling bullying Baldwin. efficacy. This study to make teachers take an antibullying stand? On the basis of has several strengths. showed high efficacy in battling bullying and low effort in reduc- sectionally related to a higher level of peer-reported bullying. the ideal class classroom. S. It would have been peer-reported bullying and how these perceptions interact with ideal to have direct observations of actual bullying. This motivation to attitudes teachers display can signal ways to act appropriately. (1987). Journal of Personality and Social Psychology. over time. students with students to keep up the antibullying classroom norm by strength. S.52. In addition. significant-other effects... . Preservice teachers’ responses to but also those who are in favor of it. 2008. Another limitation is that the descriptive find- theory. D. Peers are an important target. verbal. & O’Brennan. and effort to decrease peer reports of bullying. With teachers as significant others standing clearly perceived that their teachers exerted a great deal of effort to up against bullying. They Anderson. Sawyer. The interaction between teachers’ and students’ efforts did need to be seen as important targets of antibullying interventions. and proven valid was possible for teachers to be significant others who are able to and reliable measures. Bullying and Future studies are needed to reveal whether these findings can be peer victimization at school: Perceptual differences between students replicated among younger and older students in Finland. 223–233. A.219 Boulton. ings of this study can only be generalized to schools that are This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 2004). Salient private audiences and and does not have to exert much effort to solve bullying incidents. British Journal of Educational Psychology. 2004). not work as expected in this study. teachers’ effort is related to a reduction in bullying. 28. Bauman. Teachers’ effort was related to a reduction in bullying over time it was related to a reduction in peer-reported bullying. dispersion of the schools. impossible to collect sured in this study. J. 67. It is also and school staff. of students) affect the likelihood of bullying occurring may be than 90% of schools in Finland are registered as KiVa schools. M. D.x We used data from fourth. 98. they are likely to prevent however. 361–382. 2010. & Pepler.1111/j. Our findings are in line with this interpretation.1. J. however.6. Yoon. in addition to students own antibullying attitudes. & Holmes. Nowadays more This document is copyrighted by the American Psychological Association or one of its allied publishers. whereas teachers istics are less closely related to bullying in classes where teachers expect students to deal with bullies by themselves (Crothers & strengthen the goal to act appropriately concerning the antibullying Kolbert. (1997). L. there is also much to say for using stu- dents’ perceptions of their teachers antibullying stand is related to dents’ perceptions of teacher characteristics. M. Teachers’ views on bullying: Definitions. 2009) and that over time support. the wanted to implement the program early on. doi:10. C. was cross. we were able to test the implications of the such observations. J. (1998). our findings revealed that the level of bullying was that it was indeed related to a lower level of peer-reported bullying. 219 –231.1087 terms of attitude and effort) show relatively low levels of bullying. and that later this effort leads to a reduction References in bullying. Given the large number and geographic bullying in classrooms. Even though goals were not directly mea. and relational bullying. School Psychology Review. we predicted that it multimethod and multi-informant assessments. W. doi:10.. including the use of a proper sample size. awareness of the self. it was. This is exactly what one bullying scenarios: Comparing physical. doi:10. however.. In classes where students est level of bullying. because the focus of this study was not on descriptives bullying (Novick & Isaacs. P. we found that students’ own effort bullying is associated with low status and low affection. & Kincaid. & Del Rio. Applying behavior analysis to suggest that teachers’ effort to reduce bullying is a response to school violence and discipline problems: Schoolwide positive behavior more bullying (see also Totura et al. students’ own antibullying attitude was less strongly sible for a climate in which bullying is not accepted and in which related to bullying. Teachers’ efficacy but on the role of teachers in bullying. this can solve the problem of each party shifting efficacy and to exert a great deal of effort to stop it. We tested This limitation may not be so grave because when it comes to the direct implications of this conjecture by examining how stu. it stands to reason that teachers put in more effort to reduce bullying where there is (at first) more bullying.1037/0022- would expect from a significant-other effect of teachers. Behavior Analyst. In classes where teachers were not efficacious and had If teachers exert effort to decrease bullying.. and lack of a direct measure of teacher characteristics is of concern. In such classes not only the students who are against bullying (in 1087–1098. lowest in classes in which the teacher (in the eyes of students) Teachers’ effort to reduce bullying.. For this cooperative solution to happen.. sixth-graders from control schools. in that students expect teachers to teachers’ efficacy is in line with the idea that students’ character. 49 – 63. but how important is it important to replicate the findings in other countries.. 2013). But critical for the development and successful implementation of a the control schools in our study were pioneering schools that whole-school antibullying prevention program.tb01239.

M.. J. 33.1016/j.. B. B. M.00499. (2008)..1002/ab. M. Journal of School Psy. 29. Angeles. W. 57– 67. doi:10. CA: Sage. (2012).. K. C. Multiple imputation of missing values: Update... A. J. (2003). (1996).. E. K. F. Los The regulatory significance of needs. J. P. Benelli. doi:10. 231–247. CA: Stanford University Press.. & Craig. A. V. doi:10. Bullying and victimization . Multilevel analysis: An intro- Lindenberg. (2010). & Kolbert. J. 311–330. Grant (Eds. H. (2010). discourse of harassment in student groups.). S. doi:10. (2013). and prestige. Lindenberg. and the self.0238 This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. & Salmivalli. Journal of Early Adolescence. A. 20. Aggressive Behavior. 93–105.. Kaljonen.. M. (2011). 131.x K. W. J. R. Educa- social factors. G. MacKinnon-Lewis. C. doi:10.x relations to social status within the group. 776 – 801. L. R. & Salmivalli. D. M. special) about goals? A view from the cognitive perspective. J. doi:10. R. 32.1999. The Revised Olweus Bully/Victim Questionnaire. & Slee. & Galvan. Espelage. when and where? Involvement of children. P. (2010). Totura. doi:10. & Pigott.. F.. Teachers’ views and Shah.1300/ bullying: Insights and challenges for intervention. J. Prinstein & K. 10... Evaluating the effectiveness of a stand up for the victim of bullying? The interplay between personal and social skills training (SST) programme for victims of bullying.. doi:10. 71– 86.. England: Centre for Multilevel Modelling. S. A user’s adolescents’ active defending and passive bystanding behavior in bul- guide to MLwiN. 81–91. Rambaran. Peer Newman. C. 322. D. 41–59. A large-scale evaluation of the KiVa antibullying Salmivalli. & Rigby. C. University Press. Journal of peer victimization and bystander intervention tests of mediating pro- Social Psychology. Understanding peer influence in children and adolescents Journal. D. group influences.. C. C.1037/0022-3514. A meta-analysis of and parents in bullying behavior. S. & Murray. 283–296. E. 669 – 687. Peer involvement in tions.. Psychology. 30. Willingness to intervene in Olweus.97. O’Connell. Journal of Youth and Adolescence. J.0. Pijpers. & Steg. & Lussier. E. & Pelletier. (2004). Normative. & Gini. 23.1177/ Rivers. M.. (2001). A. 0143034302023003234 ing: Who does what..1177/0143034300211001 teacher. W. Steele. 1–14. R. Veenstra. 13. P. Wittek. J. (2002). Royston. A. (2005). In G.1177/0272431611423017 School Psychology International. doi:10. Perceptions of moral Oliver. doi:10.. Stata Dodge (Eds. 398 – 410. doi:10..0. doi:10. doi:10. Osterman. & Verloove-Vanhorick. Child Development.CO. (2008). R. Aggression Kärnä. doi:10. Journal of School Violence. R. (2005). P. (pp. J. (1994). J.1037/ Craig. P. 225–244). Frank. & Candappa. E. (2004). 424 – 439. A. Prospective teach. teachers Polanin. 93. New developments of the shared concern method..347 Craig. (1998).07. S.. S. (2003). Comparing middle school ogy. M.. Determinants of Rasbash. J. 112–120. aggression. Journal of Early Adolescence. Smith. A.3. & Altoe. B. S.. J.. doi:10. D. J. Murray. 85. (2009). (2007).. L. Divine. P.2010. 15. Stanford. 22.adolescence University of Bristol. Peer influence in involuntary social 2337(1994)20:5⬍359::AID-AB2480200503⬎3.).20282 Moskowitz & H. C. J.1006/jado. doi: Pikas.1467-8624.. 431– 453. Sijtsema.1007/ Journal of Educational Psychology. 1611–1618.1038/nn1575 03054980601094594 Espelage. G. Science.” Oxford Review of Education. gen. Lindenberg.. I. Journal of Social Issues. C. self-regulation. (2014). B. 97. D. G. I. 32. 5–21. Aggressive Behavior. Pepler. & Polanin. doi:10.424 chology. C. Bully. & Salmivalli. Albiero. (2012).1111/j ainen. 143–163... NY: Guilford Press. (2008). doi:10.. K. 45. Health Education Research. Canadian Journal of School Psychol. 193–207. Nee (Eds. J.1093/her/cyg100 Pöyhönen.1016/j. gain and hedonic goal 134 –154. doi:10. S. P. B.1017/CBO9780511584466 framing: What is it all about? Analyse & Kritik. M.9924646 cesses. (2012). school-based bullying prevention programs’ effects on bystander inter- vention behavior. C.08. D. 20. (2005). 27–55). Social rationality.. K. doi:10. Journal of Adoles- J202v03n01_03 cence.. Lindenberg.2007. (2009).1002/ab.1126/science.1080/0013188032000137238 . Poskiparta. 437– 452.1002/1098- Juvonen. & Smith.1002/(SICI)1098-2337(1996)22:1⬍1::AID-AB1⬎3. doi:10. 72–112). T. School Psychology aggressive peers at school: Students’ reluctance to seek help from the International. M. R. S. L. doi:10. How successful can interventions be? Cambridge. & Swenson.05. C.. J.01557. 30. (2008). Journal of Adolescence. (2012). D. & and Violent Behavior. The motivational looking glass: How significant others beliefs about bullying: Influences on classroom management strategies implicitly affect goal appraisals. T.. R. & Pepler. doi:10.. M. C. NY: Guilford Press. Salmivalli. (2009). .. L. Caravita. B.1037/0022- s10964-013-9953-1 0663.. What does it take to Fox. 47– 65. (2003). Novick. K. Links from teacher beliefs to children: Reported behavior and attitudes toward victims. D. & Kauki- program: Grade 4 – 6. (1999).002 Rigby. (2007). The spreading of disorder. Bullying among Australian school- Hektner. K. T. A.. Telling is compelling: The impact of ogy. lying. & V. 5. J.1080/00224545. S. & Steg. D. Bullying as a group process: Participant roles and their . & Bosker.jsp. & Goldstein.1080/01443410903573123 teachers’ and students’ views on bullying and anti-bullying interven. W. & Boulton. Ber- bullying episodes among middle school students: Individual and peer.. In R. What is so special (and non.. G. (2008). 22. (2005). Lagerspetz. J.1161405 2-T Kochenderfer-Ladd. B.2009. How students and teachers view influences on moral disengagement in late childhood and early adoles. Empirical test of bullies’ status goals: Assessing direct goals. Snijders. T. A.1540-4560.005 Sijtsema.10045 frames guiding environmental behavior. Dunham. Gadd. doi:10. Norway: University of Bergen.398 victimization in the school yard. New York. 43. The psychology of goals (pp..). Björkqvist... Handbook of rational choice social Teräsahjo. England: Cambridge Lindenberg. & Isaacs. J. the seriousness of peer harassment: When is it appropriate to seek help? cence. A. 615– 627. 3. 41. doi:10. 21. Journal of Personality and Social and students’ coping with peer victimization. In M.. L. M. goals. 46. B. Bristol. (1996). Gesten. 359 –368. Bullying and the peer group: A review.85.. S. “telling.3. 17–32. Confrontation with ers’ attitudes toward bullying and victimization. & Kamboukos. R.93. & Murphy.. M. 8..2.). & Köpetz..1177/082957359801300205 student reports of bullying on teacher intervention. C.2007. doi:10. Little. (2000). 82. doi:10. C.... 516 –536.. 56. & Phelps. 117–137. Keizer. Newman..). Browne. T. and well-being: duction to basic and advanced multilevel modeling (2nd ed. 347–365. (Eds. 1142 VEENSTRA ET AL.0046 Gini.1111/j. Henderson.007 Salmivalli. 31.avb.0. Voeten. Observations of bullying and 0022-0663. K. Social rationality. L.2-J groups: Lessons from research on bullying.. A. A. (2012). Aggressive Behavior. Bullying in schools: New York. character modulate the neural systems of reward during the trust game. H. Juvonen. Aggressive Behavior.. Delgado. W. H.1080/ Nature Neuroscience. School Psychology Review. 35. J. doi:10.2007.CO. M. Journal of Educational Psychology. P. S. Bullying and the politics of This document is copyrighted by the American Psychological Association or one of its allied publishers.. K. semi-modularity and goal. K. 307–326.1353/mpq tional Research. 63..1991. Kruglanski. Merrill-Palmer Quarterly. Pepler. W.1177/ Fekkes.. Snijders. “She is not actually bullied”: The research (pp. 1–15. Green. (1991). 1681–1685. doi:10. Educational Psychol- Crothers. T.1016/j.. Types of bullying behaviour and their 0272431611402502 correlates.

. adolescent bullying an evolutionary adaptation? Aggressive Behavior. Tinga.. 27–35.21418 Troop-Gordon. and peer victimization. J.. S. Teachers’ victimization-related Yoon. A. A. Child Development. H.. 2013 Accepted January 17. J.x This document is copyrighted by the American Psychological Association or one of its allied publishers. W. 1843– Revision received December 18. 78.69.7227/RIE. S. (2003). R.. Lindenberg.1111/j. Verhulst. S.. S. The TRAILS study. Lindenberg. doi:10. A. 37– 45. doi: Veenstra.1467-8624. R. Zijlstra.. & Ormel. Truancy in late 10. F.1002/ab. Journal of Abnormal Child Psychology. Predicting teacher interventions in bullying situations. 1854. Z. C. De Winter.. Research in Education. 29. 2014 䡲 This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. beliefs and strategies: Associations with students’ aggressive behavior Education and Treatment of Children. International Journal of Behav- ioral Development.. 2012 tion: Testing a dual perspective theory. A. doi:10. (2007). 27. (2012).. & Ormel. & Marini.1177/0165025409347987 Veenstra. (2014).3 elementary and early secondary education: The influence of social bonds and self-control. doi:10. G. The dyadic nature of bullying and victimiza. Bullying: Elementary teachers’ attitudes online publication. 34. 222–238. F. K. THE ROLE OF TEACHERS IN BULLYING 1143 among boys and girls in middle school: The influence of perceived Volk. Camilleri. (2004).. J.1007/s10802-013–9840-y and intervention strategies. doi:10. F.01102. 69. J. Journal of Early Adolescence. B. Is family and school contexts. Advance Yoon.1177/0272431608324190 38. & Ladd. doi:10.2007. Dane. W. A. J.. A. . 302–310. Received July 11. J.. & Kerber. V. (2010). 571– 609.