1 Introduction

Bullying among school children is a form of relational and interpersonal violence based on the abuse of power (Salmivalli & Peets, 2018). This power can come from physical strength (e.g., resulting from relative size, strength) and mental domination (Olweus et al., 2019; Salmivalli & Peets, 2018). Victims of bullying can experience enhanced stress and their experiences of being tormented can have significant longer-term consequences on their psychological and social development in childhood and adolescence (Mishna et al., 2012). Often, bullying and victimisation is an unfortunate consequence of early life and childhood exposure to multiple social and individual risk factors that can encourage the abuse of power or potential for victimisation (Mishna et al., 2012; Navarro et al., 2015).

In Ireland, bullying and victimisation are widespread especially in primary school contexts (e.g., Collins et al., 2004; Foody et al., 2017). Foody et al. (2017) identified that the prevalence of victimisation in primary schools (26.1%) was significantly higher than in secondary level education (12.4%), and that the prevalence of bullying others was also higher in primary schools (10.1%;) compared to secondary level education (6.9%). To help explain this widespread occurrence in childhood, the current study focuses on the role of social and emotional competences as a mechanism of victimisation and bullying.

In childhood, individuals are rapidly developing their social and emotional competencies (e.g., good relationships with peers, ability to manage one’s own emotions). These competencies play an important role in bullying prevention and in the development of wellbeing (e.g., Elipe et al., 2015; Marikutty & Joseph, 2016). Being socially competent can help children build prosocial relationships (Arsenio & Lemerise, 2001) and even though bullies can understand the intricacies of social situations they are found to be less skilled in using their social competencies to positively handle social situations (Crick & Dodge, 1999). Both bullies and victims are found to have lower levels of social and emotional competence compared to uninvolved children (Romera et al., 2016). Specifically, when compared to uninvolved children, bullies and victims have been found to be less skilled in social adaptation and managing social challenges and have lower levels of prosocial behaviour (Gómez et al., 2017).

To further our knowledge of how social and emotional competencies impact bullying and victimisation, it is necessary to extend from examining their relevance within individual children to studying their role in the classroom context as represented by the social and emotional competencies of children’s classmates. It is also important to consider how social and emotional competencies at the individual and classroom levels impact different forms of bullying and victimisation—as these experiences can be comorbid as in the case of bully-victims. To examine this possibility, the current study takes a person-oriented approach to studying bullying and victimisation. This approach aligns with a holistic-interactionist system view on children development. Children are seen as organized wholes with interacting factors operating together in a process to achieve a functioning system. The person-oriented approach provides a complex framework for research strategy and methodology, as well as for interpreting findings in children socio-emotional development. The main focus is on the children with the socio-functions about him/her regarded as an entity, an indivisible whole, and with a interact-system view, stressing process characteristics (Bergman & Magnusson, 1997).

1.1 Impact of classroom social contexts on bullying and victimisation

School is, all too often, a central context for bullying and victimisation in childhood (Saarento et al., 2013). Importantly, schools can act to regulate social norms that both prevent and promote aggressive behaviours amongst classmates (Maunder & Crafter, 2018; Wang et al., 2019). Recent studies have highlighted how complex social processes within schools can create social climates marked by peer problems (Brault et al., 2014) where victimisation and bullying frequently occur (Saarento & Salmivalli, 2015). Within classrooms, dysfunctional peer dynamics that give power to bullies can lead to increased bullying (Mazzone et al., 2018; Salmivalli, 2010). Also, bullies can use disruptive behaviour and weaken other people’s social ties to adapt to the challenges of classrooms and increase their social power (Garandeau & Cillessen, 2006).

Conversely, school climates not characterised by peer problems can help to reduce classmate conflict (Konishi et al., 2017; Mischel & Kitsantas, 2020). When groups of children, such as classmates, develop supportive relationships with each other this creates a social network where prosocial socialisation is mutually reinforced in the social dynamic (Charalampous et al., 2018; Salmivalli, 2010). Here, the internalisation of pro-social norms can help individual children’s form positive and supportive attitudes towards their peers which can consequently protect against bullying and victimisation (Acosta et al., 2019; Han et al., 2017). Experiencing care from classmates can strengthen children’s asocial and emotional resources that are important for protecting themselves against victimisation, preventing them from bullying others, and coping with victimisation if it does occur (D’Urso & Pace, 2019; Godleski et al., 2015; Houlston et al., 2011).

1.2 Social and emotional competencies of bullies, victims and bully-victims

In this study we take Coie’s (1990) perspective that children’s social and emotional competencies are the basis of victimisation and bullying. Social competence can be described as the ability to function successfully in social situations. This competence incorporates children’ emotional and cognitive skills; and the necessary behaviours for adapting well to different social situations and managing social interactions (Gettinger, 2003; Gómez-Ortiz et al., 2017; Varela et al., 2020). Bullying and victimization being atypical behaviours or relational phenomena can be explained by some deficits related to socio-emotional skills especially because they occur in group contexts (Salmivalli & Peets, 2018). Deficits in socio-emotional characteristics may be configured as risk factors connected with bullying and victimization behaviors (D’Urso & Symonds, 2021). On the contrary, effective emotional and cognitive regulation in peer relationships can prevent bullying and victimisation from occurring (Yang et al., 2020), for example by enabling children to modulate aggressive stimuli and to recognize their own emotional states, especially if they are victimised. Social and emotional competencies also allow children to have more meta-cognitive awareness of social events and ability to calibrate their emotional responses towards peers (Garner & Hinton, 2010).

Not managing emotions well, aggressive tendencies, and being inclined towards misconduct are all associated with bullying (Fink et al., 2018; Ostrov et al., 2019; Salmivalli & Nieminen, 2002). Bullies can also have limited knowledge and skill in exploiting peer support networks and establishing strong relational bonds with their peers (Charalampous et al., 2018; Murphy et al., 2017). Being a bully-victim can increase individual distress and problems with making friends, especially for girls (Kozasa et al., 2017; Marengo et al., 2018; Thakkar et al., 2020).

Studies conducted in Irish school context have identified that bullies can struggle with social perspective taking and empathy, and this makes them less proactive in establishing friendships and social support networks (Corcoran et al., 2012; Foody et al., 2019). Foody et al. (2019) also found that bullies in Irish schools were less prosocial than other children. Regarding victims, having lower levels of perceived relational skills with peers has predicted children’s chances of being victimised in Ireland (Corcoran et al., 2012).

1.3 Individual characteristics of bullies and victims

As well as social and emotional competencies, individual characteristics can predict the likelihood of children bullying and being victimised. For example, bullying (verbal and physical) behaviours have been found to increase across early (3–4 years old) and middle (5–11 years old) childhood (Cook et al., 2010; Rivers & Smith, 1994). In many countries, males are more often victims and perpetrators of bullying in childhood and adolescence (Smith et al., 2019). This has been found for both verbal and physical forms of bullying (Stubbs-Richardson et al., 2018). However, there is some evidence that this gender difference declines throughout childhood (Smith et al., 2019). Studies conducted in the Irish school context have highlighted how being male is a risk factor for bullying (Minton, 2014), but not for being victimised (D'Urso et al., 2020). In another Irish study, Cook and colleagues (2010) observed an association between being male and being a bully, a victim, and a bully-victim; although the gender difference in victimisation was small. In O’Higgins-Norman’s (2008) study bullying was identified as occurring more frequently in boys’ single sex schools and co-educational schools than in girls’ single sex-schools.

Social disadvantage (e.g., poverty, neglect) has also been studied as a risk factor for bullying and victimisation. Studies have observed that social disadvantage can increase stress and bullying can become a means to achieve psychological balance in that situation (Jansen et al., 2012; Lo et al., 2013). Being socially disadvantaged can also make children the object of discrimination in a culture that constructs social hierarchies based on wealth (Hosang & Bhui, 2018; Kingston & Webster, 2015). However, in an Irish national study no associations were found between parental economic wealth and the prevalence of victimisation in adolescence (D’Urso et al., 2020). Unexpectedly, attending lower income schools had a small negative impact on victimisation which was interpreted by the authors as demonstrating solidarity and close peer ties in those communities (D’Urso et al., 2020). Possibly, other types of disadvantages (e.g., lack of care from parents, being a member of a minority group) might be more important for bullying and victimisation than family economic wealth.

2 The current study

As outlined, in this study we take the perspective that children’s social and emotional competencies can protect against bullying and victimisation in the complex social environments of schools and classrooms. Moreover, we assume that these competencies, when shared between classmates, can augment the protective effects of the competencies within individuals. Put simply, a child in a classroom full of peers who are socially and emotionally competent, is perhaps less likely to be victimised or bully others, regardless of their own level of social and emotional competence. In this study we test these assumptions by examining the impact of individual children’s social and emotional competencies, and impact of the average social and emotional competencies of their classmates, on children’s chances of being bullies or victims. We also take a person-oriented approach to studying bullying and victimisation – assuming that there are ‘hidden’ subtypes of not only bullies and victims, but also bully-victims, existing in our sample of Irish primary school aged children. This extends prior research in Ireland where bullying and victimisation have been investigated primarily using variable oriented approaches. Two main questions are used to structure this investigation.

Research Question 1 Which are the most common profiles of bullying and victimisation in the sample? The rationale for RQ1 is to identify naturally occurring subtypes of bullying and victimisation that could include bully-victims, given the interest in these subtypes in other studies (e.g., Cook et al., 2010).

Research Question 2 To what extent do children's social and emotional competencies predict those profiles of bullying and victimisation at the individual level, and at the classroom level? The rationale for RQ2 is to identify the risk factors connected with the profiles to identify the main critical aspects within the school system on an individual and group level and inform the design of effective antibullying interventions in childhood.

3 Methods

3.1 Participants

The Children’s School Lives (CSL) study is an Irish national study of primary schooling that is following two age cohorts of children across five years (Devine et al., 2020). During this first wave of data collection, trained fieldworkers administered self-report surveys to children in 129 classrooms, within 97 schools. Children’s parents gave their written informed consent for children to participate in the study, and children gave their informed assent to participate after the researchers had explained the research process to them in the form of a story. The older cohort, Cohort B, were 8/9-years-old in 2019 (2nd Class/Grade 2). Data from 2062 Cohort B children (those with consent and assent) were included in the current study (there were no data on bullying collected with the younger cohort). Of those Cohort B children, 1020 (49.7%) were male and 1,034 (50.3%) were female. The age range is 7–10 years old (M = 8.59; SD = 0.36). The children’s teachers (N = 117) completed a teacher-on-child questionnaire for every CSL child in their class.

3.2 Measures

3.2.1 Gender

This variable was coded as female (1) and male (0) for the current analysis.

3.2.2 Bullying and victimisation

The children responded to the following two items using a 5-point Likert scale, from never (1) to always (5): In this school, how often have you been mean to someone and they said you bullied them? In this school, how often has someone been mean to you and you would consider it bullying? The items were adapted from the Growing up in Ireland national longitudinal study of children (Williams et al., 2009) and were informed by cognitive testing with the CSL Cohort B children at wave 1.

3.2.3 Strengths and difficulties questionnaire

We used the SDQ (Goodman, 2001) for evaluating the socioemotional competencies. It is a brief behavioural screening self-report questionnaire for 3–16-year-old children. There are 3 versions: teacher-report, parent-report, and child-report. In the present study we used the teacher-report version. Teachers were asked to answer on the basis of the child’s behaviour over the last six months of the current school year. The SDQ has 25-items and five subscales: emotional problems (e.g., often complains of headaches, stomach-aches, or sickness), peer problems (e.g., rather solitary, prefers to play alone), conduct problems (e.g., lies or cheats) hyperactivity (e.g., constantly fidgeting or squirming) and prosocial behaviour (e.g., often volunteers to help others (parents, teachers, other children). Each scale comprises five questions with 3-point response scales (1 = ‘not true’ to 3 = ‘certainly true’). The Cronbach alpha for the subscales indicates good to reasonable good reliability for the current sample: emotional problems α = 0.78, peer problems α = 0.67, conduct problems α = 70, hyperactivity α = 87, prosocial behavior α = 0.79.

3.2.4 Classmate caring

To evaluating how caring their classmates were, children responded to the following three items, using a 5-point Likert type scale from Never (0) to Always (4): In this class, other children care about my feelings; In this class, other children think it is important to be my friend; In this class, other children really care about me (Research Assessment Package for Schools [RAPS]; Rowe et al., 2010). The Cronbach’s alpha suggests good reliability (α = 0.80).

3.2.5 Neglect

To evaluate parental neglect of their children, we used six items adapted from Growing up in Ireland national longitudinal study of children (Williams et al., 2009) were used to capture teachers’ impressions of indicators of child neglect. Teachers were asked to comment on how they perceived each child in the study. Specifically, teachers were asked whether and how the child regularly attended school in the last academic year in these following conditions: inadequately dressed for the weather conditions; too tired to participate as he/she should in class; without a lunch/snack; hungry; with a general lack of cleanliness; late; unwell/suffering a minor ailment (cold, cough, etc.); and without homework completed. The items were measured using a 5-point scale ranging from never (0) to always (4). We conducted an exploratory factors analysis using principal axis factoring and direct oblimin rotation to search for the best combination of items to represent child neglect. The EFA identified one principal factor explaining 49.80% of the variance and containing all six items which loaded from 0.66 to 0.75 onto the factor (see Table 1). The Cronbach alpha for the neglect items suggests good reliability (α = 0.84).

Table 1 EFA of Neglect Scale

3.3 Analysis plan

First, we conducted a latent profiles analysis in Mplus version 8.6 to identify hidden subtypes of bullies and victims in the sample. Next, we designed a multilevel model to examine the impact of social and emotional competencies on the chances of belonging to the resultant profiles. At level 1, we regressed children’s individual social and emotional competencies (teacher-on-child report) and experiences of classmate caring (child report) on each of the four ‘atypical’ latent profiles (child report on bullying/victimization) that emerged in the analysis. At level 2 we regressed the aggregate scores of the children’s individual social and emotional competencies (teacher-on-child report) and experiences of classmate caring (child report) on each of the four atypical latent profiles (child report on bullying/victimization). We computed separate multilevel models for each atypical latent profile using Mplus version 8.6, using the largest ‘normative’ group of children with low levels of victimisation and bullying as the comparison group. In those models we controlled for gender (child report), age (child report), and neglect (teacher-on-child report) at each level. We retained and interpreted the standardized coefficients for the more parsimonious within-unit and between-unit structural model.

4 Results

4.1 Latent profile analysis

Models with up to 7 profiles were computed to identify subgroups of bullies and victims. The optimal number of profiles to retain was guided by several criteria (Nylund et al., 2007). Lower values of the Akaike information criteria (AIC), the Bayesian information criteria (BIC), and the sample-size adjusted BIC; significant values of the Lo-Mendel Rubin and Vong Lo-Mendel Rubin likelihood ratio tests (LMR and VLMR); and higher levels of Entropy, were used to evaluate the model fit (see Table 2 and Fig. 1). The LMR and VLMR tests were significant for all six solutions and showed limited information to determine the ideal number of profiles. Changes in information criteria were also plotted to aid the model selection.

Table 2 Latent Profile Analysis Model Fit Statistics
Fig. 1
figure 1

Patent profile analysis model fit statistics

The plots showed that the AIC, BIC and adjusted BIC bottomed out at the 4-profile solution: suggesting that the four-profile solution was a good fit to the data. This was confirmed by the high level of entropy for the four-profile solution (0.995). We then tested for conceptual utility and ecological validity by making a qualitative comparison between the latent profiles generated by the three, four, and five-profile solutions. We identified a profile of children with higher levels of victimisation (the victim’s profile) that only existed at the five-profile solution. Given that the AIC, BIC and adjusted BIC were nearly identical at the four- and five-profile solutions, that the five-profile solution also had an extremely high level of entropy (0.927) and that the five-profile solution distinguished an important hidden class, we selected the five-profile solution for further analysis. Further justification for our choice can be observed in the multilevel modelling section where we found a differential set of predictors for the victims’ profile. The five profiles were: Normative (N = 1098, 59.4%), Victims (N = 202, 10.9%), Bullies (N = 82, 4.4%), Meanies (N = 356, 19.3%) and Bully-Victims (N = 110, 6%). We selected the name Meanies to describe the group of children whose levels of bullying others were higher than normal but were not as high as those of the bullies. All the five profiles were distinct and theoretically interpretable (Corcoran et al., 2012). Descriptive statistics for bullying and victimisation for each group are shown in Fig. 2.

Fig. 2
figure 2

Distribution of profiles of bullying and victimisation

4.2 Multilevel modelling

Multilevel models (MLM) were computed to estimate the likelihood of belonging to the four profiles of bullying and victimisation compared to belonging to the normative profile. Each bullying and victimisation profile was tested in a separated multilevel model using a binary dependent variable (e.g., bullying = 1, normative = 0). Each model had the same correlational structure. Both within-unit and between-unit structural models were estimated simultaneously in each MLM. The results of the models are reported below and the models can be viewed in Table 3.

Table 3 Multilevel models

4.2.1 Victims profile

This model provided good fit to the data [χ2(18) = 72.595, p < 0.001, CFI = 1.00, RMSEA < 0.001, SRMRWithin < 0.001, SRMRBetween = 0.006]. The model demonstrated how higher levels of child neglect (β = 0.136, p < 0.001), greater conduct problems (β = 0.119, p < 0.01), and lower levels of classmate caring (β = − 0.126, p < 0.001) significantly predicted the victims’ profile at the individual level (level 1). At the classroom level (level 2), the model identified that classrooms with younger children (β = − 0.539, p < 0.01) and lower levels of classmate caring (β = − 0.519, p < 0.01) significantly predicted the victims’ profile.

4.2.2 Bully-victims’ profile

This model was a good fit to the data [χ2(18) = 111.01, p < 0.001, CFI = 1.00, RMSEA < 0.001, SRMRWithin < 0.001, SRMRBetween < 0.01]. The model demonstrated how fewer individual peer problems (β = − 0.100, p < 0.01), lower levels of classmate caring (β = − 0.103, p < 0.01) and higher levels of conduct problems (β = 0.226, p < 0.001) significant predicted the bully-victims profile at the individual level (level 1). No variables significant predicted the bully-victims profile at level 2.

4.2.3 Meanies’ profile

This model was also a good fit to the data [χ2(18) = 48.713, p < 0.001, CFI = 1.00, RMSEA < 0.001, SRMRWithin < 0.001, SRMRBetween = 0.003]. In this model, fewer individual peer problems (β = − 0.112, p < 0.01), greater hyperactivity (β = 0.081, p < 0.05) and lower levels of classmate caring (β = − 0.105, p < 0.01) significant predicted the meanies profile at the individual level (level 1). No variables significant predicted the moderate bully profile at level 2.

4.2.4 Bullies’ profile

This model provided a good fit to the data [χ2(18) = 79.346, p < 0.001, CFI = 1.00, RMSEA < 0.001, SRMRWithin < 0.001, SRMRBetween = 0.002]. The model demonstrated how greater child neglect (β = 0.116, p < 0.05), hyperactivity (β = 0.137, p < 0.01) and conduct problems (β = 0.186, p < 0.001) significant predicted the bullies profile at individual level (level 1). No variables significant predicted the bullies’ profile at level 2.

5 Discussion

The aims of this study were to identify the most common latent profiles of bullying and victimisation in an Irish national sample of primary school aged children, and to examine the impact of social and emotional competencies at the individual and classroom levels on children’s profile membership. We used latent profile analysis to uncover five distinct profiles of bullying and victimisation: bullies (high bullying), meanies (moderate bullying), bully-victims (higher bullying and victimisation), victims (high victimisation), and a normative profile with lower levels of bullying and victimisation. Then we used multilevel modelling to examine the impact of social and emotional competencies on profile membership at the individual and classroom levels.

5.1 Profiles of bullying and victimisation

The first main finding of the study is that some level of bullying and victimisation (as reported by children) was present in approximately 40% of the age 8/9-year-old nationally representative sample. This aligns with other Irish studies of bullying and victimisation in school aged children (Foody et al., 2017). Specifically, we found victimization to be the more frequent phenomenon (10.9%) than bullying (4.4%). This finding is in line with the meta-analysis conducted by Foody et al. (2017) which found primary school victimisation in Ireland to be higher than bullying (22.4% versus 9.4%), likely owing to a self-reporting bias where children are less likely to disclose their bullying actions. Our study adds to this literature by identifying two profiles of meanies (19.3%) and bully-victimsFootnote 1 (6%) children. The meanies profile includes children who tend to be mean to other children but their behaviours are not as frequent as those of bullies. Something to consider is that their behaviours could be predisposing factors to bullying across time.

5.2 Impact of individual differences

At individual level, regarding the impact of individual differences on bullying and victimisation, there were no gender differences in children’s profile membership. Our study suggests that bullying and victimisation are factors that do not depend on being male or female. In other words, in Irish primary schools these phenomena are quite heterogeneous, although other studies conducted in Ireland suggest that being male is a risk factor related to being a bully but among adolescents (Minton, 2014). Therefore, it is possible that in younger children the issue of gender does not yet have an impact as they are still in the development stage and knowledge of the stereotyped social roles connected to it.

However, we did find that children who were perceived by teachers as being higher in neglect were more likely to report being bullied or bullying others. This finding could highlight how the lack of care from parents lead to internalising psychopathology and aggression among peers (Lekunze & Strom, 2017). Parental care may be a protective factor as it helps to develop socio-relationship skills useful for structuring a greater self-confidence. On the contrary, the lack of adequate care may trigger frustration mechanisms in children that lead them to engage in aggressive behaviours or become victims to manage negative emotional states derived from this condition.

At classroom level, instead, we also found that being in classrooms where children were on average younger was associated with children being victims. In the current study, children’s ages ranged from 7 to 9-years old. In younger classrooms, children might be less capable of managing their emotions, behaviors and cognitions (Zych et al., 2018), meaning that those social groups might not be as well fortified against more aggressive children in the group who tend to bully others.

5.3 Impact of social and emotional competencies

At the individual level, having higher levels of conduct problems predicted whether a child was a bully, a bully-victim, or a victim. Here, lacking in emotional regulation and being more aggressive appears to mark children as victims – and leads to bullying. However, within individual classrooms, being around children with higher levels of conduct problems did not predict whether a child within those classrooms will bully others. This indicates that children are not bullies because they are in classrooms where there are more disruptive children (Brault et al., 2014). It is because (in part) of the difficulties they experience in regulating their own behaviour.

We also found that children with hyperactivity problems are more likely to be bullies and meanies. These results identify that a lack of ability to inhibit behavioural impulses can align with bullying others – suggesting that these children have less control over their aggressive tendencies (Timmermanis & Wiener, 2011). Together, these findings suggest how individual behavioural support is very important to install early on in children's educational careers.

Moreover, perceiving fewer caring classmates was a risk factor for being victimised, being a bully-victim, or being a meanie. This result was independent of gender, age, and child neglect-indicating the potential for a psychosocial process to underpin the finding. In line with the literature (Hong & Espelage, 2012), it may be that these children had difficulties in making friends, due to a lack of specific emotional impulses necessary for regulating and managing peer situations. In particular, the behavioral ambivalence of the bully-victim can derive from a lack of understanding of how to manage peer relationships, which leads them to identify themselves sometimes as bullies and sometimes as victims. Victims are sometimes children who do not have a social network, are isolated and do not establish fruitful relationships with their peers (Hong & Espelage, 2012).

Conversely, children who reported fewer peer problems were also more likely to be in the meanies and bully-victim groups. This data suggests, in line with a study conducted in Ireland (Sentenac et al., 2011), that sometimes children at risk of being bully or bully-victim often have adequate socio-relational skills. In line with the literature (Guy et al., 2019), bully-victims or meanies (moderate bullying) are often popular children within the classroom, they feel socially accepted and are well integrated into the peer group. Their display of tough behaviour and teasing others is done to maintain their higher social status. In comparison, children who have more extreme levels of bullying are not socially accepted, because they have more destructive and antisocial traits and lack the necessary skills to understand their social role in the group (e.g., Marini et al., 2006).

At the classroom level, we also found that children in classrooms where their classmates were on average less caring than children in other classrooms, had a higher risk of being victimised. In those classrooms, possibly children were less competent at regulating the behaviours of other children and showed less respect to other children, and this created a relational climate lacking in solidarity. In this situation, a less caring classroom acts as a risk factor because it frustrates the social-emotional need of children (e.g., Han et al., 2017). Conversely, this indicates that more caring classrooms protect children – suggesting that a curricular emphasis on teaching social and emotional skills is warranted (Habashy Hussein, 2013).

5.4 Limitations

Although the study provides new information on bullying and victimisation in the Irish school context, some limitations should be considered. The main limitation is that the data are cross-sectional. Future studies could verify the risk factors connected to bully and victim profiles using a longitudinal perspective, starting from the factors emerged from the current study and including the ethnic background and sociocultural variables also. Also, despite the sample being nationally representative of Irish primary school children, they do not represent all of childhood, meaning that the results must be interpreted in relation to the very specific developmental period of middle childhood. Finally, the questionnaire data comprised teacher-on-child report, and child self-report. In fact, it was not possible to compare the results of the scales with all informants because the information came from children and teachers respectively. Future studies could investigate any discrepancies. Although it is a strength to triangulate the two types of data, a limitation is that teachers might not be impartial observers of children’s behaviour in their classrooms.

6 Conclusion

The study has identified unique forms of bullying and victimisation in Irish primary schools. Like in other studies, we identified groups of bullies, victims, bully-victims, and uninvolved individuals (e.g., Corcoran et al., 2012; Leiner et al., 2014). In addition to these established profiles we identified a further profile of children who were “meanies” – highlighting how some children, especially younger ones, can have a weaker, yet still observable, tendency to be bullies. Therefore, unlike a fully-fledged bully, a child in the meanies profile could participate in sporadic episodes of bullying and, consequently, not be fully aware of the risks that his/her behaviour may have.

Moreover, the study highlights the importance of social and emotional competencies in predicting these different forms of bullying and victimisation. Within individual children, being a victim was predicted by child neglect, conduct problems and poor peer care; being a meanie was predicted by hyperactivity, less peer problems and poor peer care; being a bully was predicted by hyperactive and conduct problems; and being a bully-victim was predicted by poor peer problems, high conduct problems and less peer care. Within classrooms, having classmates who showed more peer support and who were older in general than children in other classrooms, reduced the risk of being a victim of bullying. This emphasises the importance of peers in the aetiology of adaptive and maladaptive behaviours during childhood, wherein peers act as a group to influence the development of individual children (D’Rutter, 1990; Urso et al., 2019). Furthermore, the influence of neglect on child outcomes highlights how abandonment and neglect can affect the social image of children who are perceived—sometimes erroneously—as economically disadvantaged (Sykes et al., 2017). Consequently, these children can be targeted as victims, or can become bullies, possibly because social neglect brings them frustration that turns into aggression and over time become chronic until a role is defined as a result of the self-determining prophecy (Merton, 2016).

The study suggests the importance of implementing intervention strategies aimed at improving the relational classroom climate by helping children to develop their social and emotional skills. Social and emotional learning, and specifically promoting emotional grammar in the classroom can be a protective factor to prevent against victimisation and bullying (Caprara et al., 2014). In addition, the reflection and understanding of children's mental states with the help of teachers may promote the processes of empathy and emotional regulation, useful for the proper management of emotions, important for this phase of the life span. If children learn to regulate and manage their emotions in the classroom as well as manage the harassment acted and suffered in class, they may certainly hope for positive developmental outcomes in adolescence.