School Belongingness and Family Support as Predictors of School Bullying Perpetration and Victimization in Adolescents: Are Relations the Same for Students with an Immigrant Background?

ABSTRACT

The aim of the study was to examine school belongingness and family support as predictors of bullying behavior and victimization, and the moderating role of students’ gender and immigrant background in a large representative sample of Slovenian adolescents (N = 1925; 42.50% male). The results indicated that school belongingness significantly predicted both bullying and victimization. Family support did not contribute to the incremental validity beyond school belongingness in predicting victimization. Gender was found to moderate the relationship between school belongingness and both bullying outcomes. The relationship between school belongingness and victimization was stronger for students with an immigrant background.

KEYWORDS:

• Bullying
• victimization
• school belongingness
• family support
• immigrant background
• adolescents

Introduction

School bullying is a phenomenon with diverse risk and protective factors, including those which are societal, social, and individual. Previous studies have indicated that differences between schools explain the negligible amount of variance in bullying and victimization (around 2–3%; Bradshaw et al., 2009; Košir et al., 2020; Saarento et al., 2013; for an exception see, Khoury-Kassabri et al., 2004), and that classroom-level factors account for around 10% of the variance in performing and experiencing bullying (Garandeau et al., 2014; Košir et al., 2020; Pan et al., 2020; Saarento et al., 2013). However, these data probably underestimate the role of the school and classroom context factors of bullying and victimization as some characteristics of the classroom’s social environment are reflected in those traditionally viewed as students’ individual characteristics even though they are determined by the classroom’s functioning. Thus, students’ individual characteristics that are connected to their peer relations should be regarded as partly reflecting relationships and social processes in the peer group. Although assessed as individual-level variables, measures such as school belongingness, perceived school and classroom climate, and peer and teacher support are established as an interaction between the characteristics of an individual and a group, as they reflect both characteristics of group social processes and students’ perceptions of it.

Strong research evidence indicates that perceptions of supportive classroom ecology are related to lower levels of both bullying and victimization (Bacchini et al., 2009; Cook et al., 2010; Saarento et al., 2013). Thus, an inclusive classroom environment can constitute a protective factor against bullying. Empowering and educating teachers as crucial agents and promoters of inclusive classroom environments seems especially important, as teachers’ beliefs and attitudes about their role in bullying are related to the victimization rate in their classrooms (Oldenburg et al., 2015; Troop-Gordon & Ladd, 2015). More specifically, a higher victimization rate was found in classrooms of teachers with more external and uncontrollable attributions of bullying (Oldenburg et al., 2015). In school discourses of responsibility and blame, which are still predominant (Herne, 2016), bullying behavior is often attributed to parental behavior and practices, and this is related to teachers feeling little personal responsibility for coping with bullying, thus limiting their potential to prevent it.

Nevertheless, the role of family factors in bullying and victimization should not be neglected as some studies (for a review see, Nocentini et al., 2019) found these factors to be related to students’ role in the dynamics of bullying. Following the socio-ecological conceptualization of bullying as a systemic problem (Swearer & Espelage, 2011) that results from complex interactions between individuals and influences in their environment (family, peer group, school, community, and culture), it is important to investigate various aspects of students’ social environments that could co-determine their bullying behavior and victimization in the classroom. However, the roles of school and family factors in students’ involvement with bullying were seldom investigated simultaneously. Thus, the main aim of the present study is to investigate whether students’ perceptions of their family environment as either supportive or unsupportive explain an additional amount of variance in bullying and victimization beyond school belongingness. In addition, we aimed to investigate whether the relationships between students’ perceptions of their school and family environments (operationalized as school belongingness and family support) and bullying involvement (either as a bully or as a victim) are moderated by students’ gender and their immigrant background.

Bullying in the classroom as a group phenomenon

Bullying, characterized by repeated and intentional aggressive behaviors and an imbalance of power between the perpetrator and the victim, is a complex phenomenon that is inevitably embedded in a group context – it reflects the nature of peer relationships within the classroom. The term bullying refers to both bullying behavior and victimization that are conceptualized as orthogonal dimensions (Menesini & Salmivalli, 2017). The importance of peer ecology and school climate, as well as family variables for both bullying behavior and victimization, are recognized by the socio-ecological framework (Espelage & Swearer, 2010) that conceptualizes bullying as being influenced by family, school, and peers in overlapping layers surrounding the participants in the bullying.

Understanding the social dynamics of bullying in the peer group and the role of adults in these dynamics is crucial for the effective prevention of bullying. These dynamics become particularly complex during the transition into adolescence (Yeager et al., 2015). In adolescence, bullying is mostly manifested in ways that are more hidden from adults (by an increase in relational and cyberbullying, and a decrease in physical bullying). Moreover, the reasons for bullying behavior, as well as for bystanders’ responses that maintain and enhance the dynamics of bullying, can usually be found in students’ social goals – bullying becomes a tool for gaining social power and popularity within the peer group (Caravita & Cillessen, 2012; Garandeau & Lansu, 2019). Furthermore, students who experience bullying become increasingly socially isolated due to the growing influence that bullies have on bystanders (Juvonen & Schacter, 2017).

Teachers have a crucial role in bullying prevention and coping with bullying events, as they are responsible for ensuring students’ physical and psychological safety. They serve as active agents who – by their responses to bullying events – consistently activate the classroom and school norms related to students’ social behavior (Veenstra et al., 2014). Their responses are important not only for bullies and victims, but also for the whole peer ecology (Grusec & Hastings, 2015) as they strengthen (or weaken) the expectations of how to interact with others, thus influencing students’ social and emotional learning in school.

Supportive relationships with peers and teachers thus represent a significant protective factor against involvement in bullying (Košir et al., 2020; Thornberg et al., 2022). Students’ perception of their relations in school as supportive is conceptualized as a construct of school belongingness, defined as the extent to which students feel personally accepted, respected, included, and supported by others in the school social environment (Goodenow & Grady, 1993). It is thus a multidimensional construct that encompasses emotional and behavioral aspects (Allen et al., 2021). As a measure that reflects students’ feelings of being connected with and attached to one’s school, in empirical research, it is often conceptualized using other terms (e.g., school connectedness, school bonding, school attachment, school identification, sense of community; for a review see, Slaten et al., 2016). Students’ school belongingness was found to be positively related to student academic functioning, mental health, and well-being (Allen & Bowles, 2012), and negatively related to peer victimization and bullying (Arslan, 2021; Goldweber et al., 2013). A meta-analysis by K. Allen et al. (2016) found that among contextual factors, teacher support is the strongest predictor of school belonging.

The role of the family in bullying and victimization

Although there are many challenges for teachers in terms of their effective entry into peer dynamics to prevent bullying (see, e.g., Salmivalli et al., 2021), their power lies in the fact that they can influence the whole peer ecology – if they adopt a sensitive approach and take adolescents’ need for autonomy into account. Unlike teachers, parents do not have this option, and are also not qualified to do so. Nevertheless, students’ families play an important role in the development of bullying and victimization as well as in prevention and intervention efforts (Navarro et al., 2021). Both teachers and parents need to be aware of what is going on in the classroom or in their child’s peer interactions. There are many ways in which parents can influence students’ role in classroom bullying dynamics; Espelage and Swearer (2010) propose that parents influence bullying processes through attachment processes, parenting styles and practices, and family social support. As the present study investigates the role of the latter, we focus on studies related to the role of family support in preventing both bullying perpetration and victimization.

Communicating support, involvement, and care in the family can work as a protective factor that both prevents bullying and victimization, as well as helps to cope with it more effectively. One of the first studies that investigated parents’ social support of students with different bullying roles was performed by Demaray and Malecki (2003). They found that students classified as bullies and bully-victims reported lower levels of social support from parents. With regard to victimization, in their systematic review Lereya et al. (2013) found that high parental involvement and support, and warm and affectionate relationships, protected children and adolescents against peer victimization, independent of whether victimization was reported by children, parents, teachers, or mixed-method. In a more recent systematic review, Nocentini et al. (2019) found that among studies that have investigated the relationship between parental involvement and support (mostly conceptualized as emotional support) and bullying and victimization, the majority of studies confirmed the protective role of parental involvement and support in relation to bullying and victimization. Thus, higher levels of perceived parental support were associated with lower levels of both bullying and victimization. In addition, Yeung and Leadbeater (2010) found parental support to be related to a lower probability of future re-victimization. However, all reported effects of parental support were small.

As emphasized by the social-ecological model of bullying (Espelage & Swearer, 2010), peers, teachers, and parents are all central agents of the child’s microsystem. However, the connections between parts of the microsystem, conceptualized as a mesosystem in Bronfenbrenner’s ecological systems theory (1979), should also be investigated to comprehensively address bullying processes. However, only a handful of studies have simultaneously investigated students’ perceptions of social support in school and family. In a sample of adolescents, Yeung and Leadbeater (2010) investigated emotional support from teachers, father, and mother in relation to victimization. They found that teacher support was more strongly negatively related to victimization compared to paternal and maternal support. Similarly, Espelage and Swearer (2010) reported students’ perceptions of school engagement as being more strongly related to bullying perpetration and victimization than positive parenting – the latter did not predict either bullying or victimization significantly. In a recent study, Shah et al. (2021) found both school belongingness and family cohesion to be negatively related to victimization; however, the correlations were stronger for school belongingness in a sample of both immigrant and non-immigrant youth. Nevertheless, to our knowledge, none of the existing studies has addressed the question of whether parental support explains additional variance in bullying perpetration and victimization that is not explained by school support. Thus, in the present study, we have aimed to investigate whether students’ perceptions of family support predict bullying and victimization beyond their perceptions of their relations in school as supportive, operationalized as school belongingness.

Gender and immigrant background in relation to bullying

In the present study, students’ educational grade was used as a control variable. Previous studies mostly indicate an overall decline in the prevalence of bullying involvement from childhood to late adolescence (for a review see, Zych & Farrington, 2021). However, some authors (e.g., Yeager et al., 2015) suggest that this holds for direct types of bullying (e.g., physical), whereas indirect types of bullying increase in adolescence.

Students’ gender and immigrant background were used as control variables as well as moderators in the current study. With regard to gender, many previous studies (e.g., De Bolle & Tackett, 2013; De Bruyn et al., 2010; Košir et al., 2020; Ma, 2002; Nansel et al., 2001; Thornberg et al., 2022), as well as meta-analyses (Cook et al., 2010), are quite consistent in their findings indicating that boys are more engaged in bullying perpetration, whereas results regarding bullying victimization are more mixed (for a review see, Felix & Greif Green, 2010). However, in a recent meta-analysis by Kljakovic and Hunt (2016), gender was not identified as a significant predictor of either victimization or bullying. Nevertheless, as emphasized by Felix and Greif Green (2010), the role of gender in bullying has to be conceptualized in interaction with the characteristics of students’ social environment. Therefore, in the present study, we investigated whether the relationships between social support (assessed as school belongingness and family support) and bullying and victimization may differ for boys and girls.

Similarly, studies that have investigated the relationships between immigrant background and bullying perpetration and victimization have also yielded inconsistent results (e.g., Almeida et al., 2011; Bjereld et al., 2015; Feinstein et al., 2019; Kahle & Peguero, 2017; Llorent et al., 2016; Plenty & Jonsson, 2017). The recent meta-analysis that has investigated the role of ethnicity in bullying perpetration (Vitoroulis & Vallancourt, 2018) concluded that the effects are very small and non-significant. Similarly, in a recent study, Thornberg et al. (2022) found that immigrant background demonstrated negligible relationships with both bullying and victimization. However, social support (in both school and family contexts) may be especially important for students with immigrant backgrounds. In a sample of adolescents, Shah et al. (2021) found that family cohesion (but not school belongingness) moderated the association between bullying experience and internalizing problems only for immigrant youth. In the present study, we aimed to investigate whether school belongingness and family support represent a stronger protective factor against bullying and victimization for students with immigrant backgrounds.

Thus, the present study aimed to investigate students’ gender, immigrant background, school belongingness, and family support as predictors of bullying perpetration and victimization in a sample of adolescents. Specifically, we aimed to investigate whether perceived family support explains a variance in bullying and victimization beyond the variance explained by school belongingness. In addition, we aimed to investigate whether students’ gender and immigrant background moderate the relationships between school belongingness and bullying outcomes and family support and bullying outcomes.

Method

Participants and procedure

We conducted a study on 1984 adolescents (42.40% male) from 21 Slovenian elementary and 20 secondary schools.¹ Some of the participants failed to complete the questionnaires, so we analyzed the sample of 1925 adolescents. The mean age of participants was 15.35 years (SD = 1.21). The majority of the participants did not report having an immigrant background (82.40%). Overall, 29.10% of participants attended 9th grade of elementary schools (M_age = 13.96 years; SD = 0.38), and out of 70.90% who attended secondary schools, 23.30% were 1st-year students (M_age = 15.04 years; SD = 0.60), 25.30% were 2nd-year students (M_age = 16.02 years; SD = 0.65), and 20.30% reported that they were in the 3rd year (M_age = 16.90 years; SD = 0.48) of secondary school.

The current study is part of a more extensive one investigating longitudinal pathways for positive youth development: Positive Youth Development in Slovenia: Developmental Pathways in the Context of Migration. The research was approved by the local ethical committee. In accordance with the research aims in the sampling plan, all upper secondary school types were included according to the proportion of students that attend each kind in Slovenia, and all lower secondary schools were included in the sampling procedure. When schools agreed to participate, additional sampling of the classes of upper secondary schools was performed. We randomly assigned the two classes per upper secondary school that were participating in the study. All classes from each participating lower secondary school were included in the study. After obtaining informed consent from their parents, the questionnaires were administered in the classrooms. The time was not limited, and students were supervised by the school coordinator (teacher or school counselor) who was available to help students if needed. In this study, we use data from the first wave (of four waves altogether) of data collection that took place in October and December 2020.

Measures

Self-reported bullying and victimization

The Adolescent Peer Relations Instrument: Bully/Target (APRI-BT; Marsh et al., 2011) was used to measure bullying and victimization. This questionnaire measures bullying (e.g., “I got other students to start a rumor about a student.”) and victimization (e.g., “I was hit or kicked hard.”) in verbal, physical, and social domains. The instrument consists of 36 items (half for bullying and half for victimization; six for each subdomain). Participants were asked to rate the frequency of the described items for the current school year using a 6-point rating scale (1 = never, 6 = every day). Previous data on the use of the Slovenian adaptation (Košir et al., 2020) indicate excellent reliability for both bullying (α = .91) and victimization (α = .92).

School belongingness

Students’ sense of belonging at school was measured with a scale designed and used in PISA (OECD, 2019). The measure reflects how accepted, respected, and supported students feel in their social context at school. Students reported their sense of belonging in seven items (e.g., “I feel like I belong at school”) using a 4-point rating scale (1 = strongly disagree, 4 = strongly agree). Previous studies (Allen et al., 2022) indicate adequate scale reliability (α = .83).

Family support

Five items from the Developmental Assets Profile (DAP; Scales, 2011) were used to assess students’ perceived family support (“I have a family that gives me love and support.” “I ask my parents for advice.” “I have parents/guardians who are good at talking to me about things.” “I feel safe and secure at home,” and “I am spending quality time at home with my parent(s) when we do things together.”) Participants were asked to use the 4-point rating scale (1 = not at all or rarely, 4 = extremely or almost always). Confirmatory factor analyses (CFA) were conducted to establish the goodness of fit for a measure of family support. The chi-square of the measurement model was not significant, χ2(5) = 8.71, n.s. The RMSEA assessing the amount of model misfit was 0.02. The TLI was 0.97 and the CFI was 0.99, which means that the model is a very good fit. The factor loadings of five included items are DAP1 M(λ) = 0.03, DAP2 M(λ) = 0.77, DAP7 M(λ) = 0.06, DAP13 M(λ) = 0.02 and DAP26a M(λ) = 0.47.

Demographic characteristics

Students’ gender was evidenced by an open-ended question: What is your gender? Students’ answers were recoded into 0 = male, 1 = female, 2 = non-binary. Students’ educational grade was coded as 1 = final year lower-secondary, 2 = first-year upper-secondary; 3 = second year upper-secondary; 4 = third year upper-secondary. Students’ immigrant background was assessed using three open-ended questions: In which country were you born? In which country was your mother born? In which country was your father born? If the answers to all three questions were “Slovenia,” students were coded as not having an immigrant background (0). If the answer to one or more of the three questions was any other country except Slovenia, students were classified as having an immigrant background (1).

Data analysis

All statistical analyses were carried out in IBM SPSS 22. The research questions were addressed using hierarchical moderated regression with enter method. Bullying perpetration and victimization were analyzed as outcomes in separate analyses. First, all predictor variables were mean centered in order to avoid problems with multicollinearity. Following the recommendations by Aiken and West (1991), interaction terms were constructed by multiplying the predictors that indicated students’ demographic characteristics (gender, immigrant background) with those reflecting their perceived social support (school belongingness, family support). When entering predictor variables into the regression model, control variables (i.e., gender, grade, immigrant background) were entered in the first step. Measures of perceived social support (i.e., school belongingness, family support) were entered in the second and third steps, respectively. Interaction effects were entered in the last step of the regression model. We inspected significant interaction effects using plots where the relation between moderators (i.e., gender, immigrant background) and outcomes (bullying, victimization) was represented with two lines, one for each level of the social support variables (e.g., low school belongingness involves the mean score of school belongingness minus one standard deviation, whereas high school belongingness involves the mean score of school belongingness plus one standard deviation; Aiken & West, 1991).

Preliminary analysis showed that all assumptions for hierarchical regression were met. Based on the VIF (variance inflation factor) indicators, we can conclude that there are no problems with multicollinearity in our data (almost all VIF indicators were between 1 and 4). We also checked for the existence of influential points that could distort the regression model, but they were not detectable in our analysis (Cook’s distance values were between 0.00 and 0.04). A further examination of the assumptions showed that for the selected models we do not significantly violate the assumptions of homoscedasticity, linearity, and approximately normal distribution of residuals (standardized coefficients were between −2.74 and 5.95).

Like the majority of studies in the school context, our study included a three-level nested data structure with students nested in classrooms from different schools. However, our research questions do not involve multilevel relations. Therefore, multilevel modeling was not applied.

Results

Descriptive statistics, correlations, and reliability coefficients were examined (see, Table 1). The correlations show that bullying and victimization were positively related; the correlation was moderate to high. Students’ sense of belongingness and family support were moderately positively related. Boys reported more bullying and victimization, and students at higher educational grades reported less bullying and victimization.

Table 1. Descriptive statistics, correlations, and reliability coefficients for all variables.

	M	SD	Bullying	Victimization	Gender	Grade	Immigrant background	School belongingness	Family support
Bullying	1.33	0.57	.92
Victimization	1.33	0.61	0.55**	.94
Gender			−0.24**	−0.10**
Grade			−0.05*	−0.12**	0.02
Immigrant background			0.05*	0.03	0.01	−0.10**
School belongingness	3.13	0.53	−0.05*	−0.33**	−0.09**	−0.02	0.00	.81
Family support	16.80	2.75	−0.14**	−0.13**	−0.05*	−0.14**	0.02	0.31**	.82

Reliability coefficients for interval variables are presented in the diagonal.

A hierarchical regression analysis was conducted with school bullying perpetration as the dependent variable. As shown in Table 2, demographic variables – specifically gender, grade, and immigrant background – that were entered in the first step did account for a significant portion of the variance in predicting self-reported bullying perpetration. Boys, students at lower educational grades, and students with an immigrant background reported higher involvement in bullying perpetration. School belongingness entered in step 2 accounted for a small yet significant portion of variance (1.6%), with all three – gender, grade, and school belongingness remaining negative predictors of school bullying perpetration. Thus, while school belongingness declined, the level of school bullying perpetration increased. In the third step, perceived family support was entered. It accounted for an even smaller but significant increase in explained variance (0.5%), where perceived support of a family was a significant predictor of bullying; adolescents with a higher level of family support reported more classroom bullying perpetration.

Table 2. Hierarchical regression analysis for school bullying perpetration as an outcome variable.

	School bullying perpetration (N = 1981)
	Model 1		Model 2		Model 3		Model 4
	B	β	B	β	B	β	B	β
Step 1: Control variables
Gender	−0.28	−0.24***	−0.29	−0.26***	−0.30	−0.26***	−0.86	−0.75***
Grade	−0.02	−0.04	−0.02	−0.05*	−0.03	−0.06*	−0.03	−0.05*
Immigrant background	0.08	0.05*	0.08	0.05*	0.08	0.05*	0.12	0.08
ΔR^2		0.06***
Step 2: Contribution of school
School belongingness			−0.14	−0.13***	−0.11	−0.10***	−0.24	−0.22***
ΔR^2				0.02***
Step 3: Contribution of family
Family support					−0.02	−0.08**	−0.01	−0.07
ΔR^2						0.01**
Step 4: Interaction
Gender and SB							0.21	0.59***
Gender and FS							0.01	−0.09
Immigrant background and SB							−0.03	−0.11
Immigrant background and FS							0.00	0.08
ΔR^2								0.01**
R^2	0.06***		0.08***		0.09**		0.10**
F for change R^2	43.24		32.73***		11.12***		4.75

*p < .05**p < .01, ***p < .001.

In the fourth step, the interactions of school belongingness and family support with gender and immigrant background were entered. They accounted for a relatively small portion of variance (1.0%). However, while the effect sizes in psychological research usually range from 1% to 3% (Aguinis et al., 2005; McClelland & Judd, 1993), this may still be considered to be of practical significance. The interaction between gender and school belongingness was a significant predictor of bullying, which is presented in Figure 1. We found that both relationships between gender and school bullying perpetration in the condition of low and high school belongingness were significant (simple slope test for high school belongingness, B = −0.20, p < .001, simple slope test for low school belongingness B = −0.41, p < .001). This means that although boys exhibit more school bullying perpetration in both conditions in comparison to girls, the difference is two times higher in a condition of low school belongingness.

Figure 1. The interaction between gender and school belongingness as a predictor of school bullying perpetration.

Notes. SB – school belongingness.

Altogether, in the full model, 10% of the variance in bullying perpetration was explained.

Further, we conducted a hierarchical regression analysis with victimization as a dependent variable. Results, as shown in Table 3, are similar to the previous regression analysis. Demographic variables – gender, grade, and immigrant background – entered in the first step, and did account for a significant portion of the variance in predicting school bullying perpetration with gender and grade being significant predictors of victimization. Boys and students at lower educational grades reported more victimization. In the second step school belongingness was entered. It accounted for a significant portion of variance (15.6%) and negatively predicted victimization in school, meaning that while school belongingness declined, the level of victimization between peers increased. In step 3 family support was entered, where the model nor family support as a predictor were not found significant.

Table 3. Hierarchical regression analysis for victimization as an outcome variable.

	Victimization (N = 1891)
	Model 1		Model 2		Model 3		Model 4
	B	β	B	β	B	β	B	β
Step 1: Control variables
Gender	−0.12	−0.10***	−0.17	−0.13***	−0.17	−0.14***	−0.35	−0.28
Age	−0.06	−0.11***	−0.06	−0.12***	−0.07	−0.12***	−0.07	−0.12***
Immigrant background	0.04	0.03	0.04	0.03	0.04	0.03	0.12	0.08
ΔR^2		0.02***
Step 2: Contribution of school
School belongingness			−0.43	−0.37***	−0.42	−0.36***	−0.48	−0.42***
ΔR^2				0.14***
Step 3: Contribution of family
Family support					−0.01	−0.04	0.00	−0.01
ΔR^2						0.00
Step 4: Interaction
Gender and SB							0.13	0.34*
Gender and FS							−0.01	−0.19
Immigrant background and SB							−0.09	−0.29*
Immigrant background and FS							0.01	0.24
ΔR^2								0.01**
R^2	0.02***		0.16***		0.16		0.17**
F for change R^2	15.29***		311.77***		3.00		3.57**

*p < .05**p < .01, ***p < .001.

In the fourth step, the interactions between gender and immigrant background as demographic variables and school belongingness as well as family support were entered. They accounted for a relatively small yet significant portion of variance (0.01%). The interactions between gender and school belongingness and between immigrant background and school belongingness were found significant predictors of victimization which is presented in Figure 2 (gender and school belongingness) and Figure 3 (immigrant background and school belongingness).

Figure 2. The interaction between gender and school belongingness as a predictor of victimization.

Notes. SB – school belongingness.

Figure 3. The interaction between immigrant background and school belongingness as a predictor of victimization.

Notes. SB – school belongingness.

Additionally, we found that both relations between gender and school victimization in the condition of low and high school belongingness were significant (simple slope test for high school belongingness, B = −0.10, p < .001, simple slope test for low school belongingness B = −0.24, p < .001). This means that boys exhibit more victimization in low and high conditions of school belongingness in comparison with girls, and the difference is more than two times higher in a low condition of belongingness. In the interaction between immigrant background and school belongingness results show significant differences in the condition of low school belongingness (simple slope test B = 0.15, p = .002), whereas in the condition of high school belongingness no significant results were found (simple slope test B = −0.03, n.s.). This means that we can only argue that the victimization is higher for migrants than non-migrants when participants reported low school belongingness.

Altogether, all predictors accounted for 17% of the variance in victimization.

Discussion

Students’ involvement in bullying is a complex phenomenon that is, among other factors, related to their feelings of social inclusion and perceived support in their significant social contexts. In the present study, the role of support that adolescents perceived in school and family contexts was investigated to simultaneously assess both sources of support as protective factors in students’ involvement in bullying, reflected as either bullying perpetration or victimization. As emphasized by Yoon et al. (2020), teachers and parents’ role in bullying dynamics is in a way similar to peers’ roles as bystanders; however, their role is unique as they have clear responsibilities and authority that enable them to have an impact on the broader context of bullying. Family and school contexts represent the most important developmental contexts for students’ academic as well as social-emotional development (Eccles & Roeser, 2011; Osher et al., 2018). In addition, the role of school in preventing bullying is unique as it is possible for school professionals to at least partly influence the peer system and shape students’ peer relations which are crucial in the development of bullying dynamics. However, the role of support in the family context in students’ bullying involvement should not be neglected as it represents both a context for social learning and an important support resource when coping with emotional and social distress that can be behind both bullying behavior and victimization. Therefore, the role of the present study was to examine whether family support works as a predictor of bullying perpetration and victimization beyond perceived acceptance and support in the school context. Additionally, we were interested in whether the role of perceived support in school and family contexts depends on students’ gender and immigrant background.

We found that school belongingness significantly negatively predicted both bullying and victimization. The relation was stronger for victimization which is consistent with extensive research support that victimization is related to social exclusion (e.g., Cantin et al., 2019; Sentse et al., 2015). In addition, the act of victimization, especially relational victimization that prevails in adolescence, is an act of social exclusion per se. Thus, it is not surprising that students who perceived that they are less accepted, respected, included, and supported by others in their school context reported higher victimization. The finding that school belongingness negatively predicted bullying behavior in adolescents seems more intriguing as we know from previous studies that students with a higher level of bullying behavior are usually perceived as highly popular by their peers, and that this relation becomes stronger in adolescence (Caravita & Cillessen, 2012). However, students who bully are typically not well-liked by their peers (De Bruyn et al., 2010; Caravita et al., 2009). Nevertheless, they usually have reciprocal friendships (Sentse et al., 2014) and more dating opportunities (Dane et al., 2017). In a recent study, Košir et al. (2020) found that both self- and peer-reported bullying was related to higher levels of perceived peer support, but lower levels of perceived teacher support in adolescents. These findings could explain why in the current study school belongingness as a measure that includes perceptions of different relational aspects (feelings of acceptance, respect, support, and inclusion by both peers and teachers) was weakly negatively related to bullying. Thus, adolescents who experienced higher levels of bullying reported feeling lower school belongingness. The correlation between both constructs is low which might reflect the fact that the construct of school belongingness is probably saturated with various relational measures, and that we did not differentiate between popular and unpopular bullies in our study. However, we can conclude that despite the peer-perceived popularity found in many previous studies, students who bully generally report lower feelings of acceptance, respect, inclusion, and support at school. As our finding raises questions about the relationship between social inclusion, social power, acceptance, and support in adolescence, we recommend addressing this issue in future studies, especially as school belongingness is a very frequent measure of students’ social adjustment in international assessment studies (e.g., PISA, TIMSS) which are frequently used to inform policy and practice.

Perceived family support was weakly negatively related to both bullying perpetration and victimization, meaning that adolescents who perceived higher support in the family context reported lower levels of bullying behavior and victimization. However, family support was found to explain a negligible amount of variance in bullying, and no additional variance in victimization beyond school belongingness. Thus, it seems that in adolescence, the potential of family support to protect students from involvement in bullying either as a bully or as a victim is limited or at least intertwined with protective mechanisms in the school context. The correlation between school belongingness and family support found in this study was low to moderate. The findings justify the need to plan bullying prevention and intervention aligned with the characteristics of developmental stages. As adolescents lean toward out-of-family contexts, these are the contexts that can be more influential in prevention and intervention efforts.

Among demographic variables, students’ gender was found to predict bullying but not victimization. However, the correlations with gender were significant for both bullying and victimization, although the relation with victimization was weaker. Consistent with some previous studies (Cook et al., 2010; Košir et al., 2020), boys reported more bullying perpetration and victimization. Students’ gender moderated the relationship between school belongingness and bullying as well as victimization. Male adolescents that perceived lower school belongingness reported higher levels of bullying and victimization. Thus, it seems that not feeling accepted, included, and respected in school is a stronger risk factor for bullying involvement for adolescent boys than for girls. Less bullying involvement (bullying and victimization) was found in older adolescents. Immigrant background was not found to predict bullying and victimization; however, it moderated the relationship between school belongingness and victimization. In a condition of high school belongingness, immigrant background was not found to significantly predict victimization, whereas in the condition of low school belongingness students with an immigrant background reported being more victimized. Thus, feeling included, accepted, and supported in school is especially important for students with immigrant backgrounds; perceiving the school environment as unsupportive places these students at higher risk of victimization. This finding is consistent with conceptualizations of vulnerability to victimization as being inevitably understood in terms of interactions between the individual and contextual characteristics (see, e.g., Garandeau et al., 2014; Juvonen & Schacter, 2017; Schwartz et al., 2020), as immigrant background was not found to be a risk factor for victimization per se, but only when combined with an unsupportive environment. Thus, being an ethnic minority does not necessarily enhance the risk of being victimized; it can, however, become a risk factor in an unsupportive school or classroom environment. However, we should keep in mind that our research design does not permit conclusions about the direction of the relationships between constructs; therefore, we do not know whether students with immigrant backgrounds were more victimized because of their unsupportive school environment, or that they started to perceive their school environment as less supportive because of their experiences of victimization.

Neither students’ gender nor their immigrant background moderated the relationship between perceived family support and bullying outcomes. Family support thus represents a protective factor for bullying perpetration for both genders, independent of a student’s immigrant status, and does not represent a significant protective factor for victimization independent of students’ gender and immigrant background. We see this finding as an additional indicator that the potential of the school context in preventing bullying outcomes overpowers the potential of the family context, as it can influence the entirety of the schools’ peer dynamics.

Limitations and further directions

Several limitations have to be considered when interpreting and generalizing the results of the present study. Cross-sectional research design represents a major limitation in interpreting the nature of the relatedness between constructs, as it does not allow conclusions about the direction of relations. Social support in school and family contexts and bullying outcomes are very likely in a reciprocal relationship; our research design does not enable the conclusions about the strength of directions in both ways. In addition, all variables were operationalized as self-reported measures; therefore, the results are affected by common method bias. Finally, school belongingness as a global measure of students’ feelings of acceptance, respect, inclusion, and support by others in the school social environment was used in this study. Although the construct of school belongingness was found to be more strongly determined by teacher support compared to support from peers or parents (K. Allen et al., 2016), we suggest that future studies differentiate between teacher and peer support when identifying the role of teachers, peers, and parents in bullying outcomes. It would also be relevant to simultaneously investigate the potential of these support contexts as possible determinants of bullying duration, the concurrent and long-term distress related to bullying, and the effectiveness of coping with victimization.

Conclusions and implications

The results of our study indicated that school belongingness and family support as measures that mainly reflect the support from adults are more strongly related to victimization compared to bullying. School belongingness was found to be more strongly related to bullying outcomes compared to family support which accounted for a negligible amount of explained variance in bullying perpetration, and did not account for a significant amount of explained variance in victimization beyond school belongingness. For male students and for students with immigrant backgrounds, school belongingness was a stronger protective factor against victimization. Male adolescents that perceived lower school belongingness reported higher bullying perpetration, which was not the case for girls. School belongingness thus seems to be of special importance for adolescent boys. We emphasize the relevance of our finding that immigrant background was found to be a significant predictor of victimization only in a condition of low school belongingness. Thus, our findings confirm the importance of creating supportive school contexts that enhance students’ positive development. Such inclusive classroom and school environments are likely to prevent the development of aggressive popularity norms (see, e.g., Laninga-Wijnen et al., 2019), and thus help adolescents to develop positive social behavior. Moreover, interventions should empower teachers to support adolescents’ social and emotional learning and moderate their peer relationships.

Ethics approval statement

The study was approved by the ethics committee of the Faculty of Arts, University of Maribor.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Slovenian Research Agency J5-1781, J5-3114.

Notes

1. Compulsory basic education in Slovenia is organized into a single structure (ISCED 1 and ISCED 2) and consists of 9 years of basic school attended by pupils aged 6 to 15 (Taštanoska, 2019). Upper secondary education takes between 2 and 5 years (but is mostly 4 years in duration). Educational programs include vocational, professional, and general (“gimnazija”) programs.