The support group approach in the Dutch KiVa anti-bullying programme: effects on victimisation, defending and well-being at school

Abstract

Background: School bullying is a wide-spread problem with severe consequences for victims, bullies and bystanders. Schools are strongly encouraged to implement both schoolwide, preventive interventions and reactive measures to handle existing bullying situations. In the Dutch implementation of the KiVa anti-bullying programme, pervasive-bullying situations are addressed according to the support group approach. The support group approach is widely used for addressing bullying situations, but little is known about its effectiveness.

Purpose: We investigated the effectiveness of the support group approach in reducing victimisation, increasing defending and improving the victim’s well-being over the course of a school year, over and beyond of the effects of the universal KiVa intervention.

Programme description: The support group approach is a non-punitive, problem-solving strategy to address pervasive-bullying situations. In this intervention, trained teachers form a support group that consists of 6–8 children, including the bullies and their assistants, defenders or friends of the victim and prosocial classmates. The purpose of the support group is to create mutual concern for the well-being of the victim and to trigger the bullies’ willingness to alter their behaviour.

Sample: We used data from 66 Dutch elementary schools that participated in the KiVa intervention study. Data were collected in October 2012 and 2013, and May 2013 and 2014. The sample used in the analyses consisted of 38 victims for whom a support group intervention was organised (44.7% boy, M_age = 9.24; SD_age = 1.20).

Design and method: To get insight into the effects above and beyond those of the KiVa programme itself, victims with a support group (N = 38) were matched to similar victims without a support group (N = 571). Statistical analyses were undertaken to examine whether the changes in victimisation, defending and well-being at school differed between the two groups.

Result: Victims reported positive effects of the support group approach in reducing victimisation in the short term, but this decrease in victimisation was not lasting over the course of a school year. The intervention also did not improve the victims’ well-being at school in the longer term. Victims with a support group, however, were found to have more defenders at the end of the school year than victims without a support group.

Conclusion: The effectiveness of the support group approach in tackling bullying situations appears to fade over time. The findings of this study illustrate that for future evaluations of anti-bullying interventions it is essential to investigate longer term effects.

Keywords:

• Anti-bullying intervention
• support group
• victimisation
• defending
• well-being

Introduction

In past decades, bullying in schools has become a matter of serious concern. Bullying, commonly defined as the systematic, repetitive and intentional abuse of others (Olweus 1993), is a problem in almost all schools throughout the world (Salmivalli, Garandeau, and Veenstra 2012; Smith and Shu 2000).

Figure 1. Short-term and long-term change in victimisation for victims with a support group intervention (N = 38).

A wide range of research has substantiated that bullying is related to various forms of psychosocial maladjustment for all those involved (Ttofi, Farrington, and Lösel 2014). Victims of bullying suffer from mental and physical health issues, tend to be socially isolated and generally have a low level of well-being at school (e.g. Bouman et al. 2012; Reijntjes et al. 2010). Bullies are often more rejected by their classmates and have an increased likelihood of school drop-out, problem drinking and becoming unemployed in later life (e.g. Kokko and Pulkkinen 2000; Warden and MacKinnon 2003). Besides, children who witness bullying behaviour are likely to suffer from feelings of anxiety and depression, feel less comfortable at school and show social adjustment problems (Nishina and Juvonen 2005; Rivers et al. 2009; Werth et al. 2015).

The growing awareness of the prevalence and the negative consequences of school bullying have amplified the need for effective interventions to tackle bullying (Salmivalli, Garandeau, and Veenstra 2012). In many countries, including the Netherlands, schools are strongly encouraged to implement some form of anti-bullying programme that provides both schoolwide, preventive interventions and reactive measures to handle existing bullying situations (Wienke et al. 2014). Generally, a distinction between two reactive strategies can be made. On the one hand, there is the punitive approach in which bullies are directly confronted with their behaviour and they are sanctioned for it (Thompson and Smith 2011). On the other hand, there is the non-punitive approach that involves problem-solving strategies, such as the support group approach, the no blame method (Robinson and Maines 2008; Young 1998), and the method of shared concern (Pikas 1989, 2002). The latter approaches are mostly similar (Garandeau, Poskiparta, and Salmivalli 2014) and aim to change the behaviour of bullies and bystanders by increasing their feelings of discomfort and by raising their awareness of the victims’ suffering. Non-punitive strategies are considered key strategies in reducing bullying and victimisation (Young and Holdorf 2003). In the Netherlands, experts prefer the use of non-punitive, problem-solving strategies above punitive strategies in intervening in bullying behaviour (Wienke et al. 2014). In the Dutch implementation of the KiVa anti-bullying programme (Veenstra et al. 2013), pervasive bullying situations are addressed according to the support group approach.

Although the support group approach is widely used in several countries (Smith, Howard, and Thompson 2007), little is known about its success in reducing bullying (Rigby 2014). Evaluation of the support group intervention turned out to be difficult for various reasons. First, reactive interventions are used when bullying situations occur. As such, evaluations are rather ad hoc and cannot easily be organised in advance (i.e. there is no pre-test or control group). Second, there is a tendency to adjust the intervention to better suit the needs to the specific situation, as a consequence of which little insight into the exact methods of application exists and the comparability among interventions is relatively low. Third, most evaluations are based on victims’ or teachers’ self-reports and thus rely on subjective interpretations of outcomes (Rigby 2014; Smith, Cowie, and Sharp 1994).

Relatively few studies investigated the support group approach: from an evaluation of 30 case studies, Smith et al. (1994) concluded that most of the participants felt that the situation had improved, (i.e. there was less bullying). Teachers also indicated that they had the feeling that the bullying behaviour was reduced. Moreover, a study by Young (1998) revealed that in only three out of 50 cases, the victim reported continued bullying. In a report about the efficacy of anti-bullying strategies in England (Thompson and Smith 2011), it was stated that the support group approach was effective in three quarters of bullying situations, which was a higher success rate than other methods (Rigby 2014).

The studies described previously base their conclusions on short-term effects only, given that victims and their teachers were asked about their experiences immediately after the support group intervention took place. Moreover, effects of the reactive strategies are not isolated from other, prevention oriented, anti-bullying interventions that might be present in school. In the current study, we investigate the effectiveness of the support group approach – as part of the Dutch KiVa programme – in addressing pervasive bullying situations over the course of a school year. In order to examine the effects above and beyond those of KiVa itself, victims for whom a support group intervention was organised were matched with similar victims without a support group, using coarsened exact matching (CEM) (Iacus, King, and Porro 2011).

The KiVa anti-bullying programme

KiVa is an anti-bullying programme developed in Finland (Salmivalli, Kärnä, and Poskiparta 2010). It was evaluated in a randomised control trial in Finland during 2007–2009 and disseminated nationwide afterwards (Kärnä et al. 2011, 2013). KiVa is currently being implemented and tested in several countries, including the Netherlands.

The KiVa programme is predicated on the idea that bullying is a group phenomenon with different roles rather than an incident between a bully and its victim. Research on participant roles in bullying showed that the behaviour of bystanders, for instance assisting the bully or defending the victim, is crucial for maintaining or solving bullying situations (e.g. Salmivalli et al. 1996). A positive change in the behaviour of others is expected to reduce the social rewards gained by bullying and consequently the bullies’ motivation to bully (Salmivalli, Garandeau, and Veenstra 2012). Hence, KiVa aims to encourage bystanders to take a clear stance against bullying and to support the victim instead of assisting the bully. For that purpose, the programme contains universal actions that target all students. The core of these universal actions is 10 student lessons covering a wide range of themes (i.e. respect, group pressure, mechanisms and consequences of bullying). Additionally, there is a computer game in which children test their knowledge about bullying and can enhance their defending skills (Poskiparta et al. 2012). The universal actions are principally aimed at preventing bullying (e.g. by encouraging victim-supportive behaviours) and raising awareness about group processes. For solving existing bullying situations, the KiVa programme includes indicated actions, which are the focus of this study.

Indicated actions in the Dutch KiVa programme: the support group approach

In each KiVa school, there is a so-called KiVa team that consists of at least three teachers or other school personnel. Members of the KiVa team are trained in addressing pervasive cases of bullying, using the support group approach.

In line with the KiVa programme, the support group approach is based on the idea that bullying is a group phenomenon (Garandeau, Poskiparta, and Salmivalli 2014) and that others can alter the bullies’ motivations to bully (Rigby 2014; Robinson and Maines 2008; Young 1998). Instead of focusing only on (changing) the behaviour of the bully, bystanders and defenders are also involved in tackling bullying situations. The purpose of the support group is not to punish or blame the bullies and their assistants, but to create mutual concern for the well-being of the victim. It is emphasised that everyone has to do something to help to improve the situation. In other words, the responsibility to solve bullying is given to the support group. It is assumed that the shared distress will evoke empathy within the bullies and that the ‘group pressure’ of shared responsibility will trigger the bullies’ willingness to alter their behaviour. Moreover, assistants are expected to lose the excitement and arousal of watching bullying (Rigby 2014; Robinson and Maines 2008; Young 1998).

Discussion meetings

The support group approach as used in the Dutch KiVa programme consists of a set of individual meetings with the victim and small group meetings with the support group. KiVa team members are asked to fill in reports after each discussion meeting.

Prior to the intervention, a screening procedure is conducted to ensure that there is indeed bullying involved: an intentional and systematic abuse of power with negative consequences for the victim. When these criteria are met, victims are interviewed by a KiVa team member in the first session. They are asked to tell about the things that have been happening. Moreover, victims are supposed to specify who is involved in the bullying and to indicate who is likely to support them. During this session, victims are informed about the follow-up procedure and are assured that nobody will be punished.

Subsequent to the first meeting, a support group is formed. Preferably, the support group consists of 6–8 children, including the bullies and their assistants, defenders or friends of the victim and a few prosocial, high-status peers. The victim is not included. It is important that there is a balance between students who are involved in the bullying and prosocial students. In the small group discussion, the bully is not apportioned blame. KiVa team members instead share their concern about the victimised in order to raise empathy for the victim and all children in the support group are encouraged to elicit suggestions that could be helpful for the victim: ‘I have heard person X is having a hard time. What could you do to improve the situation?’ At the end of the meeting, the responsibility to provide practical support (e.g. helping with school tasks; trying to stop the bullying) and to make the victim more comfortable at school (e.g. greeting; playing together) is given to everyone present. After a week, two follow-up meetings – one for the victim and one for the support group – are held to see whether or not the situation has changed. If the situation has improved, the support group are complimented for their help and encouraged to continue their positive behaviour. If no progress has been made, additional steps are discussed.

The present study

Existing evaluations on the support group approach provide us little information about the use of the intervention. Thompson and Smith (2011) state that around 10% of the schools in England employ this strategy to solve bullying situations. However, there is no clear indication for which or how many victims a support group intervention is organised and what the support group looks like. The first objective of this study was therefore to gain more insight into the use of the support group approach within the KiVa programme in the Netherlands. How many times was a support group organised and for whom? What was the composition of the support group? Moreover, we wanted to investigate whether the victims’ short-term evaluation of the intervention was as positive as within other studies, in which almost all victims indicated that the victimisation had decreased or stopped (e.g. Rigby 2014).

The third objective of this study was to examine the effectiveness of the support group approach over the course of a school year. Until the present research, various studies showed positive outcomes regarding the (perceived) change in victimisation, but these conclusions were merely based on evaluation meetings during or immediately after the intervention (Smith, Cowie, and Sharp 1994; Thompson and Smith 2011). It is, however, essential to investigate whether the possible positive effects of the support group intervention are still visible at the end of the school year, so that stronger conclusions about its effectiveness can be drawn. Given that both the universal component of the KiVa programme and the support group approach aim to reduce victimisation and to enhance defending behaviour, in this study, changes in victimisation and defending were taken into account. Another – more indirect – aim of the support group approach is to improve the victims’ well-being at school (i.e. support group members are asked to make the victim feel more comfortable at school). Hence, we also investigated the changes concerning the victims’ well-being at school. In this way, a more overall image of the effectiveness of the support group approach can be provided rather than just focusing on the reduction of bullying.

We expected that victims for whom a support group was organised would be (1) less victimised and (2) more defended at the end of the school year than victims for whom no support group was organised. Additionally, we expected them to have (3) higher well-being at school in comparison with victims without a support group intervention.

Method

Sample

Data used in this study stem from the evaluation of the Dutch implementation of the KiVa anti-bullying programme. To recruit schools, letters describing the KiVa project were sent in the fall of 2011 to all 6938 Dutch elementary schools. Special elementary schools and schools for children with special educational needs could not take part in the KiVa programme and were hence not invited to participate. A total of 99 schools indicated they were willing to participate.

Prior to the pre-assessment in May 2012 – and for new students prior to the other assessments – schools sent information about the study to students’ parents. If parents did not want their child to participate in the assessment, they were asked to inform the teacher. Students were informed at school about the research and gave oral consent. Both parents and students could withdraw from participation at any time.

When the pre-assessment was finished, schools were randomly assigned by the Netherlands Bureau for Economic Policy Analysis (CPB) to either the control condition (33 schools) or one of the two intervention conditions (i.e. 34 schools KiVa intervention and 32 schools KiVa + intervention). Control schools were asked to continue their ‘care as usual’ anti-bullying approach until their participation in the KiVa programme in June 2014.

Procedure

Students filled out internet-based questionnaires in the schools’ computer laboratories during regular school hours. These questionnaires were developed for the evaluation of the KiVa programme in Finland (see Kärnä et al. 2011) and adjusted to the Dutch situation. Classroom teachers distributed individual passwords to their students, which could be used to access the questionnaire. The order of questions and scales used in this study were randomised in such a way that the order of presentation of the questions would not have any systematic effect on the results. Students read all questions by themselves; difficult topics were explained in instructional videos. In these videos, a professional actress explained the questions in such a way that all students could understand them (e.g. by talking slowly and articulating words clearly). The term bullying was defined in the way formulated in Olweus’ Bully/Victim questionnaire (Olweus 1996). Several examples covering different forms of bullying were given, followed by an explanation emphasising the intentional and repetitive nature of bullying and the power imbalance. Classroom teachers were present to answer questions and to assist students when necessary. Teachers were supplied with detailed instructions before the data collection started and were encouraged to help students in such a way that it would not affect their answers (e.g. by asking them questions such as ‘Which words are unclear to you?’)

During the process of the support group approach, the KiVa team members were asked to fill in a report form after each meeting. On the basis of these forms, we could derive information about the victimisation, the composition of the support group, the arrangements that were made and the victim’s perceived effectiveness.

Participants

In the present study, we used data from schools in the intervention conditions that were collected during two school years, respectively, in October 2012 and 2013, and in May 2013 and 2014. In this period, the support group approach was used for 56 victims in 28 schools. This target sample consisted of 30 girls (53.6%) and 26 boys (46.4%) in grades 2–6 (age range: 7–12; M_age = 9.15; SD_age = 1.23). In five of the cases, there was missing information on one of the outcome variables. Moreover, 13 students had indicated not being victimised in October. For these students, no baseline information about the level of bulling and defending was provided. Hence, they were excluded from the analyses. The sample used in the analyses consisted of 38 victims (44.7% boy, M_age = 9.24; SD_age = 1.20).

Measures

Victim’s evaluation

In the follow-up meeting with the victim, the KiVa team member asked the victim to indicate whether the bullying situation had changed and chose the best-fitting option from the following: the victimisation had increased (0); the victimisation had remained the same (1); the victimisation had decreased (2); and the victimisation had completely stopped (3). We used this information to construct a variable indicating the victim’s evaluation of effectiveness in the short term.

Change in victimisation

During the online data collection in October and May, students were asked to nominate peers by whom they were victimised. These nominations could be given to classmates as well as students from other classes. Peer nominations are widely used in bullying research (e.g. Huitsing et al. 2014; Sainio et al. 2011; Veenstra et al. 2005; Verlinden et al. 2014) and acknowledged as reliable and valid. In order to measure the change in the level of victimisation in one school year, for each victim, the number of bully nominations given in October were summed and subtracted from the total bully nominations given in May. In this way, for each victim, a score was created indicating whether the victimisation had increased (0), remained the same (1), decreased (2) or completely stopped (3). Besides, a score indicating the change in the frequency of victimisation was created using the revised Olweus’ Bully/Victim questionnaire (1996). Students were asked to indicate how many times they were victimised in the past months (0 = it did not happen, 1 = once or twice, 2 = two or three times a month, 3 = about once a week, 4 = several times per week). We calculated a difference score by subtracting the frequency of victimisation in May from the frequency of victimisation in October. A positive score indicates an improvement in the victim’s situation, that is, a decrease in the frequency with which the victimisation takes place.

Change in defending

In addition to nominating their bullies, students who indicated that they were victimised were asked to nominate their defenders. Again, nominations could be given to classmates as well as students from other classes. Based on the difference in the total nominations given in October and May, for each student, an indicator for the change in defending was created. Scores varied from defending had completely stopped (0) to defending had increased (3).

Change in well-being at school

Well-being at school was indicated by seven items concerning the perception of the classroom and school (Kärnä et al. 2011). Students responded to items such as ‘I feel accepted as I am at school’ (1 = never, 4 = always). The items formed an internally consistent scale (Cronbach’s α = 0.84 in October; 0.86 in May) and were averaged. We calculated a difference score by subtracting the self-reported well-being in May from that of October. A positive score indicates an increase in the student’s well-being at school.

Matching variables

A match was made on five victim characteristics that were expected to influence the change in victimisation, defending and well-being at school, that is, frequency of victimisation, the students’ level of depression and well-being at school, gender and grade. For the analytical procedure, it was necessary to create groups with discrete values (Iacus, King, and Porro 2011). Frequency of victimisation was measured using the revised Olweus’ Bully/Victim questionnaire (1996) as described previously.

Students’ psychosocial well-being was indicated by their levels of depressive symptoms and well-being at school in October. We used the adjusted major depression disorder scale (Chorpita et al. 2000) to measure the emergence of depressive symptoms. Students’ answers on the nine items (e.g. ‘I feel worthless’) could vary from never (1) to always (4). Together, the items formed an internally consistent scale and were averaged (Cronbach’s α = .81).

Well-being at school was indicated by the students’ self-reported well-being. Both indicators for psychosocial well-being were coded into three groups, based on quartiles. Lastly, grade (range 2–6; Dutch grades: 4–8) and gender (0 = girl; 1 = boy) were included as matching variables. Table 1 presents descriptive information on the study variables.

Table 1. Descriptive information on the study variables (weighted).

			Support group			No support group
	Min	Max	Mean	SD	%	Mean	SD	%
Victim’s evaluation	0.00	3.00	2.13	0.66
Change in victimisation
Increased	0.00	1.00			31.6			24.2
Remained the same	0.00	1.00			7.9			8.2
Decreased	0.00	1.00			34.2			26.8
Stopped	0.00	1.00			26.3			40.8
Change in frequency of victimisation	−4.00	4.00	0.26	1.62		1.24	1.59
Change in defending
Stopped	0.00	1.00			10.5			26.3
Decreased	0.00	1.00			26.3			28.2
Remained the same	0.00	1.00			10.5			10.8
Increased	0.00	1.00			52.6			34.7
Change in well-being at school	−1.57	2.57	0.00	0.72		0.22	0.61
Matching variables
Sex (1=boy)	0.00	1.00			44.7			44.7
Grade	4.00	6.00	4.16	1.13		4.16	1.11
Frequency of victimisation in October	0.00	4.00	2.92	1.44		2.92	1.42
Depressive symptoms in October
Low	0.00	1.00			5.3			5.3
Medium	0.00	1.00			44.7			44.7
High	0.00	1.00			50.0			50.0
Well-being at school in October
Low	0.00	1.00			73.7			73.7
Medium	0.00	1.00			21.1			21.1
High	0.00	1.00			5.3			5.3
N			38			571

^a This was only asked to victims who received a support group.

Analytical strategy

Coarsened exact matching

Victims, for whom a support group was organised, were statistically matched to victims who did not receive a support group. Matching is a non-parametric method that aims to balance the distribution of covariates in the treated and control group (Iacus, King, and Porro 2011; Stuart 2010).

The CEM procedure ensures that only respondents with identical scores on the covariates (e.g. age, gender, level of victimisation, well-being) are matched. In this way, victims with and victims without a support group intervention could be better compared as bias in the context was controlled for (Iacus, King, and Porro 2011) and, as such, the impact of the support group approach on changes in victimisation, defending and the victims’ well-being at school could be tested.

The matching procedure was conducted in SPSS 20, using the Python plug-in and CEM software (Iacus, King, and Porro 2009). Respondents for whom there was no exact match in the sample were removed from the analyses, leading to a sample of 38 victims with support group and a control group of 571 victims without support group. There were 30 sets of respondents (strata), indicating that a stratum existed of at least one victim with a support group (range per stratum: 1–4) and often several victims without a support group (range per stratum: 1–54). The multivariate imbalance λ₁ was 0 and the local common support was 100, which implies that there were indeed only exact matches included (Iacus, King, and Porro 2009).

We examined whether the changes in victimisation, defending and well-being at school differed between victims who had a support group and victims who had no support group. As two dependent variables were measured on an ordinal scale and the other dependent variables were not normally distributed, a Wilcoxon signed-rank test was performed on 30 sets of respondents.

Results

The use of the support group approach within the Dutch KiVa programme

The support group intervention was organised in 28 out of 66 KiVa schools. Descriptive analyses on the reports completed during the support group approach showed that the mean number of support group interventions per school was 2 (SD = 1.31) and ranged from 1 to 6. The average size of the support group was 5.96 (SD = 0.13; range: 3–8). Mostly, the support group consisted of students from the same classroom (76.8%) and both boys and girls (71.4%). A slight majority was formed by students who were in the same grade (57.1%). However, in multi-grade classrooms, which are common in the Netherlands, this was only in 20.7% of the cases. Victims reported to have 5.69 bullies and 3.05 defenders on average. In most of the cases, the support group consisted of two bullies (34.3%) and two defenders (42.4%). Moreover, in the majority of the support groups, at least one friend of the victim was present (76.9%). The chronicity of the victimisation varied across the cases. Most victims indicated that they were victimised for more than a year (33.9%). However, other victims had been victimised for only one or two weeks (21.4%).

Short-term versus long-term changes in victimisation

In Figure 1, it can be seen that in the short-term evaluation victims were very positive about the effect of the support group approach as a majority indicated that the bullying situation had improved in the past two weeks. In 11 out of the 38 cases, victims reported that the victimisation had been stopped, in 21 that it had decreased, and in only six of the cases, there was no change in the bullying situation. No one reported that the victimisation had increased after the support group approach.

Figure 1 also shows that at the end of the school year, the results from long-term evaluations were different from those of the short-term evaluation. In May, 10 victims for whom a support group was organised reported that the victimisation had stopped and in 13 of the cases the victimisation has decreased. However, three victims indicated that the situation had stayed the same and for 12 victims the situation had gone worse. On average, victims were significantly more positive about the change in their bullying situation immediately after the support group was organised than at the longer term (M_dif = .58, z(38) = −2.71, p = .01).

The effectiveness of the support group approach

Changes in victimisation, defending, and well-being at school for victims, who received a support group intervention, were compared with the situation for victims for whom no support group intervention was organised. Results are presented in Table 2. With regard to victimisation, the mean and median scores of both groups indicated that the number of bullies tends to remain the same over time. There was no difference between victims with a support group and victims without a support group (z = −.03; p = .49). In contrast, a significant difference concerning the change in frequency of victimisation was found (z = −3.27; p = 0.00). The outcomes revealed that for victims who were not involved in a support group intervention the frequency with which the victimisation took place decreased (M = 1.01; Mdn = 1.44), whereas for victims with a support no substantial change was found (M = 0.15; Mdn = 0). The findings did not support our hypothesis that victims for whom a support group was organised would be less victimised at the end of the school year than victims who were not involved in the support group intervention.

Table 2. Changes in victimisation, defending and well-being at school.

	No support group		Support group		Z	P^c	Effect size
	Mean (SD)	Median	Mean (SD)	Median
Victimisation	1.84 (0.58)	1.96	1.71 (1.12)	2.00	−0.03^a	0.49	0.00
Frequency of victimisation	1.11 (1.01)	1.44	0.15 (1.58)	0.00	−3.27^a	0.00	0.42
Defending	1.53 (0.58)	1.49	2.08 (1.01)	2.36	−2.39^b	0.01	0.31
Well-being at school	0.19 (0.31)	0.20	−0.02 (0.72)	0.07	−1.58^a	0.06	0.20

Note: N = 30 strata.

^a Based on positive ranks.

^b Based on negative ranks.

^c One sided.

As regards defending, the results showed that victims with a support group intervention had more defenders at the end of the school year than victims without a support group. The difference is significant (z = −2.39; p = .01). Hence, it appears that the support group is beneficial for victims in terms of being defended. This is in line with what we expected.

Table 2 also shows the results with regard to the victims’ well-being at school. It turned out that, for both groups, the changes in well-being over the course of a school year were rather small (M = 0.19/−0.02; Mdn = 0.20/0.07 for victims without and with a support, respectively). The difference between the two groups was not considered significant (z = −1.58; p = 0.06), which implies that our hypothesis that the support group approach would be beneficial for victims’ well-being at school is rejected by the findings.

In short, the results point out that victims tend to report an improvement in the bullying situation in the short term, but this decrease did not continue throughout the school year as at the end no changes in either the level or frequency of victimisation were found. Moreover, the findings showed that there was an additional effect of the support group intervention on changes in defending over and beyond the school-wide KiVa intervention itself, but this was not the case for victimisation and the victims’ well-being at school.

Discussion

The aim of this study was to investigate the effectiveness of the support group approach as part of the Dutch version of the KiVa anti-bullying programme. Although the support group approach is widely used as an anti-bullying intervention and recommended by local authorities (Smith, Howard, and Thompson 2007; Thompson and Smith 2011; Young and Holdorf 2003), relatively little is known about its successfulness. Previous studies focused on short-term effects of the support group approach, which are often based on the evaluation meeting that is held two weeks after the intervention started. It was found that a majority of students or teachers reported that the support group intervention was effective in tackling bullying situations. However, often it is not known whether similar effects would be obtained if no support group intervention was used (see Rigby 2014). Moreover, existing research merely focused on changes in the bullying situation, whereas the support group (implicitly) aims to increase victim-supportive behaviour and enhance the well-being of victims (Robinson and Maines 2008; Young 1998).

We argued that it is important also to look at whether the support group approach is still beneficial at the longer term as well as above and beyond proactive anti-bullying strategies. We therefore examined its effectiveness in terms of changes in victimisation, defending, and the victims’ well-being at school over the course of a school year. In order to control for changes in the victims’ situation caused by effects of the universal actions within the KiVa programme, victims who received a support group intervention were matched to similar victims for whom no support group was organised (Iacus, King, and Porro 2011). In this way, more insight into the unique contributions of the support group intervention concerning longer term changes in the victims’ situation was gained.

In line with previous research (e.g. Rigby 2014; Smith, Cowie, and Sharp 1994; Thompson and Smith 2011; Young 1998), we found that victims for whom a support group intervention was organised were positive about changes in the bullying situation in the short term. A majority of the victims indicated during the evaluation meeting that the victimisation had decreased or stopped. However, at the end of the school year, outcomes were less encouraging: for almost 40% of the victims it turned out that the victimisation did not alter or even became worse in comparison with their situation in October. It thus appears that the (perceived) effectiveness of the intervention is not lasting.

We hypothesised that the support group approach would be beneficial for victims in terms of victimisation, defending and well-being at school. More specifically, we expected that at the end of the school year victims who received a support group intervention would be less victimised, more defended and feel more comfortable at school than similar victims without a support group. With regard to the victim’s bullying situation, we did not find significant differences in the change in the level of victimisation between the two groups. For both victims with a support group and victims without a support group, the victimisation tended to remain the same over the course of a school year. However, the frequency with which the victimisation took place was found to decrease more for victims without a support group than for those victims who received a support group intervention.

As regards the victim’s well-being at school our outcomes revealed that victims do not benefit from a support group in terms of feeling more comfortable at school. A possible explanation for this unexpected finding is that the support group approach only makes the victims feel more comfortable at school during the intervention period and that this effect fades over time. Victims with a support group could for that reason be more likely to report a decrease in their well-being, as they probably have felt better in the intervention period. Unfortunately, with our data, we were not able to test short-term effects concerning the victim’s well-being at school.

In contrast, outcomes in relation to defending were in the expected direction. Victims for whom a support group intervention was organised reported having more defenders (i.e. an increase in defending) at the end of the school year than victims for whom no support group was organised. Regardless of whether this increase in defending is due to more stable victimisation, this is an important finding given that previous research has shown that being defended is related to a higher self-esteem and a higher peer status than undefended ones (Sainio et al. 2011).

In sum, we can conclude that despite the fact that in the short term victims were positive about the changes in the bullying situation, the support group approach only was successful in enhancing defending behaviour over the course of a school year. This might imply that prosocial students especially are effected by the intervention as the success rate concerning the resolution of victimization seems to feed over time and tends to even worsen the victim's situation. These findings offer more nuance to prior evaluations in which the support group approach was considered an effective anti-bullying strategy (e.g. Rigby 2014; Thompson and Smith 2011; Young 1998) and underline the importance of evaluating anti-bullying interventions over a longer time span as positive changes may disappear over time.

Limitations and directions for future studies

The outcomes of the current study should be interpreted in light of its limitations. First, the sample size was small. Although the support group approach is a standard component of the Dutch KiVa programme, it is only used for a small minority of the victims. This low usage of indicated actions in targeting bullying is a source of concern. It might be that the school staff does not want to use the support group approach and tries to solve bullying situations in other ways. Another possibility is that a high number of victims go unrecognised by parents, classroom teachers and the KiVa team. This would be in line with the evaluation of the KiVa programme in Finland, in which was found that only 23.5% of the structural victims received a targeted intervention (Haataja et al. 2015). It also is possible that the support group intervention was used, but the KiVa team did not fill in the reports or did not send them back. It thus might be that our ‘without support group sample’ also consists of victims who actually received a support group intervention. In other words, the reported differences between the two groups in this study might be overestimated. Future studies would develop knowledge if they used large-scale randomised control trials with schools that implement all programme components (i.e. preventive and reactive strategies) and schools that implement only preventive strategies. Such studies should examine whether fidelity to the intervention plays a role.

A second limitation concerns the use of different methods in establishing the short- and long-term effectiveness with regard to changes in victimisation. The short-term effect is derived from a personal interview during the support group intervention in which the victim was asked to indicate whether the bullying situation had changed. The long-term effect is based on information derived from questionnaires that were filled out anonymously. There is a chance that victims were prone to report an improvement during the follow-up meeting in order to either please the KiVa team or to discontinue an intervention they perceived ineffective (see also: Garandeau, Poskiparta, and Salmivalli 2014). This social desirability bias might have caused an overestimation of the difference between the changes in victimisation in the short and long term. In future studies, it would be useful to have information from several informants (e.g. classroom teachers, classmates, KiVa team members) so that changes in the victim’s situation can be studied from a range of perspectives.

Although we were able to distinguish short-term from long-term effects with regard to changes in victimisation, we have little insight into what is happening during the intervention itself. Data about the experiences of the victim and support group members should be collected on a frequent basis (e.g. via diary studies). In this way, more insight into the process of the support group approach would be provided and possible relapses in the victim’s situation could be detected earlier.

Lastly, from our study, it appears that the support group intervention is more beneficial for some victims than for others. Future research should examine these differences in effectiveness thoroughly. It might be that the support group approach is only successful among certain victims, in tackling specific forms of bullying, or in classrooms with strong anti-bullying norms (e.g. Rigby 2014; Thompson and Smith 2011). For instance, it is likely that among victims with lower social skills or communication difficulties the support group intervention is less effective. Moreover, it has been found that bystanders’ feelings of empathy differ between the various forms of bullying (Nishina and Juvonen 2005; Tapper and Boulton 2005; Werth et al. 2015). Another possibility is that the composition of the support group plays a role. The founders of the support group approach emphasise that the composition of the support group is important. Ideally, the support group consists of the bullies, their assistants and prosocial peers who are likely to help the victim (Robinson and Maines 2008; Young 1998). Up to now, little has been known about the composition of support groups. Are victims and practitioners able to indicate potential defenders? And are all (prosocial) bystanders competent to help their victimised classmates? Future evaluations would benefit from investigating the social position and characteristics of support group members and their consequences regarding the effectiveness of the intervention.

Practical implications

Our results suggest that only for a few victims a support group intervention is organised. It is important that victims are recognised as such by school personnel. Classroom teachers could be helped by feedback reports with information about the students who are structurally victimised, have a low sense of well-being at school and/or are highly disliked by their peers. Such feedback reports may be extended by suggestions for the composition of the support group or indicated actions that could be undertaken. In addition, students should be trained in identifying victimisation and practise prosocial strategies for tackling (pervasive) bullying situations. In this way, both school staff and students might be better prepared to improve the victim’s situation.

Disclosure statement

No potential conflict of interest was reported by the authors.

Acknowledgements

We are grateful to all students and teachers who participated in this study and all the colleagues working to realise the KiVa project.

Funding

This work was supported by the Dutch KiVa project. KiVa has been financed by a grant from the Dutch Ministry of Education (Onderwijs Bewijs) to the last author.