Peer-victimisation in multi-cultural contexts: A structural model of the effects on self-esteem and emotions

Abstract

This study examined a structural model which integrated personal and cultural victimisation in order to identify the effects of victimisation on emotions and personal and cultural self-esteem. A sample of 1,185 adolescents from 13 secondary schools in England (n = 322) and Spain (n = 863) was recruited. Participants completed a battery of self-report questionnaires. A Confirmatory Factor Analysis (CFA) conducted regarding the initial model and then Strucutural Equations Modelling (SEM). Cultural victimisation had an indirect negative effect on cultural self-esteem and in consequence on emotional state, mediated by personal victimisation and the negative impact it produced on personal self-esteem. Only for the cultural majority did the percentage of their presence constitute a mediator variable between being a victim of cultural victimisation and the effect on cultural self-esteem. The results are discussed in relation to the extant literature and the implications for the intervention/prevention work to deal with victimisation and discrimination in multi-cultural schools.

Resumen

El estudio examina un modelo estructural que integra la victimización personal y cultural con el fin de identificar los efectos de la victimización en las emociones y la autoestima personal y cultural. Participó una muestra de 1.185 adolescentes de 13 escuelas secundarias en Inglaterra (n = 322) y España (n = 863). Los participantes cumplimentaron una batería de cuestionarios de autoinforme. Se llevó a cabo un análisis factorial confirmatorio (CFA) sobre el modelo inicial y luego un modelado de ecuaciones estructurales (SEM). La victimización cultural tuvo un efecto negativo indirecto sobre la autoestima cultural y en consecuencia sobre el estado emocional, mediado por la victimización personal y el impacto negativo que produce en la autoestima personal. Sólo para la cultura mayoritaria el porcentaje de su presencia constituye una variable mediadora entre ser una víctima de la persecución cultural y el efecto en la autoestima cultural. Los resultados se discuten en relación a la literatura existente y a las implicaciones para el trabajo de intervención/prevención para hacer frente a la victimización y a la discriminación en las escuelas multiculturales.

.

Bullying or victimisation occurs when an individual is the target of repeated aggression, an inequality of power (social, physical, or in abilities) is perceived by those involved, and there is an intention to cause harm to the victim or it is perceived as such by the victim (Hazler, Miller, Carney, & Green, 2001; Olweus, 1996).

Over the past three decades there has been a considerable amount of research into school bullying which has noted that it is experienced by a number of children and adolescents and can have damaging effects (see Smith, 2011). It has not been until recently that research into school bullying has focussed directly on the ethnic or cultural background of those involved. Many of the studies which have looked at this have focussed on whether there are differences in the experiences of victimisation between children and adolescents from cultural majority or minority groups. Most of these studies have been conducted in North America and have noted that children and adolescents from ethnic minority backgrounds are more likely than those from the ethnic majority to report having been bullied (Carlyle & Steinman, 2007; Juvonen, Graham, & Schuster, 2003; Mouttapa, Valente, Gallaher, Rohrbach, & Unger, 2004; Peskin, Tortolero, & Markham, 2006; Sawyer, Bradshaw, & O’Brennan, 2008; Spriggs, Iannotti, Nansel, & Haynie, 2007). Several European studies have reported similar findings (Strohmeier, Kärnä, & Salmivalli, 2011; Strohmeier, Spiel, & Gradinger, 2008). However, some research within Europe has concluded that there are no significant differences in the prevalence of victimisation as a function of the cultural background of the participants (Fandrem, Strohmeier, & Roland, 2009; Monks, Ortega-Ruiz, & Rodríguez-Hidalgo, 2008; Rodríguez-Hidalgo, Ortega-Ruiz, & Zych, 2014; Sweeting & West, 2001).

Other studies have directly examined victimisation with a cultural component, content or motive. Monks et al. (2008) used the term ‘cultural victimisation’ to include victimisation that appears to have a racial, ethnic, or cultural motive. Most studies of cultural victimisation have focussed on racist insults. Racist name-calling has been assessed in a number of studies as a subtype of victimisation along with other forms which are not explicitly racist in focus, such as experiencing physical aggression, having property stolen or damaged on purpose, and being the target of insults or threats (Strohmeier et al., 2011).

In contrast, some researchers have suggested that cultural victimisation is somewhat independent of personal victimisation. They argue that cultural victimisation can be carried out in a variety of ways, not only by racist insults, but also by other means such as social exclusion and the spreading of rumours, and that these forms of victimisation are experienced more often by individuals from cultural minority groups than those in the cultural majority (McKenney, Pepler, Craig, & Connolly, 2006; Monks et al., 2008; Rodríguez-Hidalgo et al., 2014; Verkuyten & Thijs, 2001, 2002, 2006). In this way, a victim of cultural victimisation may experience distinct types of racist aggression. ‘Multi-victimisation’ has been used to refer to situations in which the victim experiences several different types of aggression (Blaya, Debarbieux, del Rey, & Ortega, 2006). We suggest using the term ‘cultural multi-victimisation’ to refer to experiencing a variety of different forms of cultural victimisation.

Adolescents tend to rate cultural victimisation as being worse than personal victimisation (Monks et al., 2008). Cultural victimisation may have a negative impact on the recipient. Verkuyten and Thijs (2001) and Monks, Ortega, and Rodríguez (2010) found a negative relationship between the experience of cultural victimisation and cultural self-esteem (although this was non-significant in Verkuyten & Thijs’ study). It is possible that cultural victimisation damages feelings of cultural self-esteem, or that those with poorer cultural self-esteem are more at risk of cultural victimisation.

Recently, Rodríguez-Hidalgo et al. (2014) found negative relations between ethnic-cultural victimisation and: a) personal self-esteem (not specifically cultural) and b) social adjustment. In addition, they observed that pupils who were frequent victims of ethnic-cultural victimisation had fewer friends than non-victims. Whereas, there was a positive relationship between personal multi-victimisation and ethnic-cultural multi-victimisation. However, it was not possible to describe the meaning of these relations. Based on these findings, the researchers concluded that the integrated study of these variables is necessary by means of the development of new theoretical models supported by empirical data.

To date, very few studies of cultural victimisation and discrimination have proposed theoretical models based on empirical data regarding the effects of these experiences. The research of Thijs and Verkuyten (2008) and Verkuyten and Thijs (2006) is a notable exception, although they examined victimisation and discrimination beyond the educational context and not only caused by bullying among pupils. Verkuyten and Thijs found that personal victimisation was negatively related to global self-esteem, but not cultural self-esteem. Ethnic discrimination was negatively related to global self-esteem, but was mediated by cultural self-esteem. In order to explain the mediating role of cultural self-esteem, the researchers suggested a structural conception of global self-esteem which is made up of various components, including cultural self-esteem.

Kim, Wang, Deng, Alvarez, and Li (2011) and Shin, D’Antonio, Son, Kim, and Park (2011) produced structural models in which cultural victimisation and discrimination had the effect of a positive mediator between other variables such as the level of comprehension of English, being the object of cultural stereotypes, or the percentage of the cultural majority within their school, and symptoms of depression.

The current study attempts to test a structural equation model of victimisation in multi-cultural settings which integrates personal and cultural forms of victimisation. A model of the effects of personal and cultural victimisation is proposed, examining the impact on global self-esteem, cultural self-esteem and emotions, including possible mediator variables such as the percentage of the cultural majority within the school. From this model it may be possible to suggest some important considerations for the design of more effective intervention and prevention programmes for victimisation and discrimination in multi-cultural school contexts.

Method

Participants

The initial sample was composed of 1,185 students from 13 schools in the metropolitan areas of South-East England (n = 322) and Andalucía, Spain (n = 863) from compulsory secondary school education establishments. Participants with missing data were excluded leaving a final sample of 1,042 (England = 224 and Spain = 818).

The ages of the students varied from 11 to 18 years (mean age = 13.65, SD = 1.90). There were 47.3% (n = 493) in early adolescence aged 11 to 13 years and 52.1% (n = 573) in mid-late adolescence aged 14-18 years (n = 6, 0.6%, did not give their age). Just over half were male, 51.1%, n = 532 (n = 3, 0.3%, did not give their gender).

In terms of cultural group, 89.3% (n = 930) were from the cultural majority (in England, Caucasian 17.8%, n = 185; in Spain, Payo 71.8%, n = 748). Just over a tenth (10.5%, n = 109) identified themselves as being from a cultural minority group. In England, Mixed 1.2%, n = 13, Asian/British Asian 1.1%, n = 11, Black/British Black 0.5%, n = 5, Chinese 0.1%, n = 1, Other 0.9%, n = 9. In Spain, Mestizo Payo-Gitano 1.6%, n = 17, Mestizo Gitano 2.0%, n = 21, Latin American 2.1%, n = 22, African 0.4%, n = 4, Asian 0.5%, n = 5, Other 0.1%, n = 1 (0.3%, n = 3 did not give their cultural background).

Directly Observed Variables

Participant variables. Data regarding the age, gender and cultural group of participants were obtained from self-reports. In relation to age, participants were grouped into early adolescence (aged 11-13) or mid-late adolescence (14-18 years). Due to the small Ns for some of the cultural minority groups, participants were assigned to either the cultural majority (Caucasian or Payo) or cultural minority. Two additional variables were calculated: 1) the percentage of the cultural majority group within the school and 2) the percentage of their own cultural group within the school.

Personal self-esteem. The Rosenberg self-esteem questionnaire (1965) assessed personal self-esteem. This has been widely used and validated in England and Spain (Olaya, Tarragona, Osa, & Ezpeleta, 2008; Vázquez, Jiménez, & Vázquez, 2004). The questionnaire consists of 10 items and participants are asked to indicate the degree to which they agree with the statements on a four-point Likert scale. Cronbach's alpha was acceptable, .72 (cultural majority group = .73, cultural minority group = .68).

Cultural self-esteem. The Collective Self-Esteem Scale (CSE) (Luhtanen & Crocker, 1992) was employed. It consists of 16 items on a five-point Likert scale. The items are grouped into four subscales. One of the subscales (private subscale) has similar properties to the global CSE and has been used on its own with some minor adaptations by Verkuyten in order to assess adolescents’ cultural self-esteem (Verkuyten & Nekuee, 2001; Verkuyten & Thijs, 2001, 2002, 2006). This shortened version was used in the current study and Cronbach's alpha for the four-item scale was acceptable, .73 (cultural majority group = .72, cultural Mminority group = .74).

Emotions. Each participant was asked how they felt at that moment. Participants were asked to respond by choosing one of seven cartoon faces. The cartoons depicted faces changing progressively from ‘very sad’ to ‘very happy’. This type of questioning has been used by Yee and Brown (1992) and was coded on a seven point Likert scale.

Latent Variables

In order to find out about experiences of personal and cultural victimisation, seven scenarios were presented to the participant, four on personal victimisation, physical (violence, hitting, taking possessions), verbal (name calling), direct relational (social exclusion), and indirect relational (rumour spreading) and three described cultural victimisation (verbal cultural, direct relational cultural, and indirect relational cultural). The personal and cultural scenarios only differed in the focus of the victimisation (see Monks et al., 2008 for a description). It was possible to obtain an indication of the directly observable variables which created the latent variables, personal victimisation and cultural victimisation.

For each scenario, participants were asked whether it had ever happened to them. The number of times they answered in the affirmative was summed in order to obtain an indication of personal multi-victimisation and cultural multi-victimisation. Participants were also asked about the frequency with which they had experienced each form of victimisation ranging from never (0) to all the time (4). These scores were summed in order to obtain an indication of the frequency of personal victimisation and the frequency of cultural victimisation. Furthermore, students were asked about their feelings when they experienced the victimisation. Their responses were on a Likert scale from 0 (it has never happened to me) to 1 (I did not mind) to 5 (I was angry). These responses were summed separately for personal victimisation and cultural victimisation in order to obtain an indication of the emotions caused by the victimisation.

Procedure

The research was approved by the relevant ethics committees (University of Córdoba and Kingston University). Consent was obtained from the head teacher and parents/guardians of those pupils involved. Participants were invited to complete the self-report questionnaire in class individually and in silence. The questionnaire was administered in the following order: 1) participant variables, 2) emotional state, 3) personal self-esteem, 4) cultural self-esteem, 5) personal victimisation, and 6) cultural victimisation. A member of the research team was present to answer any questions. Participants were told that participation was voluntary and responses were anonymous and confidentiality of data was assured.

The data were coded and descriptive analyses were performed using PASW 18. Then, a Confirmatory Factor Analysis (CFA) of the model and various Structural Equations Modelling analyses (SEM) using AMOS 16 employing the maximum likelihood method were conducted. Missing data was eliminated. In order to assess the adjustment indices of the models, the index of absolute adjustment was considered (χ2(df), p < .05). For the sample size, the recommendations of Bagozzi and Yi (1989) and Bollen (1989) were followed. The recommendations of Hu and Bentler (1999) were considered in relation to normality and kurtosis. The following indices were also considered: χ2/df (CMIN/DF), Goodness of Fit Index (GFI), Adjusted Goodness of Fit Index (AGFI), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error for Approximation (RMSEA), and Parsimonious Comparative Fit Index (PCFI).

Results

Descriptive Statistics

Table 1 shows Pearson's correlations between each of the observed variables. Table 2 shows the means and standard deviations of each of these variables as a function of cultural group (majority and minority). In order to compare the mean scores of the cultural groups, independent t-tests were performed. There were no significant differences between the cultural majority and minority groups in personal victimisation (multi-victimisation, frequency, and feeling), personal self-esteem, and emotional state. There were significant differences between the 2 cultural groups in cultural victimisation (multi-victimisation, frequency and feeling) and cultural self-esteem. Those in the cultural minority scored higher than the cultural majority on each of the cultural victimisation variables, but scored lower than the cultural majority on cultural self-esteem.

Table 1.

Bivariate Pearson's Correlations of the Observed Variables.

Variable	1	2	3	4	5	6	7	8	9	10	11
1. Personal multi-victimisation	-
2. Emotions about personal victimisation	.91^**	-
3. Frequency of personal victimisation	.89^**	.86^**	-
4. Culturalmulti-victimisation	.25^**	.24^**	.21^**	-
5. Emotions about cultural victimisation	.23^**	.26^**	.22^**	.88^**	-
6. Frequency of cultural victimisation	.23^**	.22^**	.22^**	.91^**	.84^**	-
7. % of cultural majority in school	-.06	-.06*	-.06	.15^**	-.19^**	-.06	-
8. % own cultural group in school	.00	-.00	-.01	.33^**	-.33^**	-.01	.35^**	-
9. Personalself-esteem	-.21^**	-.22^**	-.25^**	-.07*	-.08^**	-.25^**	.14^**	.04	-
10. Culturalself-esteem	-.07*	-.06	-.08^**	-.11^**	-.11^**	-.08^**	.37^**	.13^**	.36^**	-
11. Emotionalstate	-.15^**	-.16^**	-.16^**	-.08*	-.09^**	-.16^**	.14^**	.04	.40^**	.25^**	-

^*p < .05, ^**p < .01

Table 2.

Means and SDs of the Observed Variables in the Cultural Groups.

		Cultural Group
Variable	Total n = 1,042 M (SD)	Majority n = 930 M (SD)	Minority n = 109 M (SD)	t
1. Personal multi-victimisation	2.22 (1.13)	2.23 (1.14)	2.15 (1.06)	0.72
2. Emotions about personal victimisation	8.04 (4.41)	8.07 (4.47)	7.88 (3.91)	0.40
3. Frequency of personal victimisation	2.10 (2.34)	2.11 (2.36)	2.03 (2.19)	0.34
4. Cultural multi-victimisation	1.14 (0.46)	1.10 (0.36)	1.57 (.81)	-10.83^***
5. Emotions about cultural victimisation	3.51 (1.74)	3.33 (1.43)	5.07 (3.00)	-10.83^***
6. Frequency of cultural victimisation	0.23 (0.80)	0.15 (0.62)	0.95 (1.51)	-10.31^***
7. % of cultural majority in school	89.18 (5.71)	89.59 (5.29)	5.41 (3.86)	4.84^***
8. % of their own cultural group in school	80.75 (26.28)	89.60 (5.27)	85.97 (7.61)	161.13^***
9. Personals self-esteem	29.75 (4.36)	29.77 (4.36)	29.58 (4.40)	0.44
10. Cultural self-esteem	17.41 (2.92)	17.48 (2.84)	16.79 (3.48)	2.00^**
11. Emotionalstate	5.36 (1.30)	5.36 (1.29)	5.29 (1.40)	0.53

^**p < .01, ^***p < .001

Measurement Model

In the first phase, two latent variables were constructed: personal victimisation and cultural victimisation. Personal victimisation was made up of three observed variables: a) personal multi-victimisation, b) frequency of personal victimisation experienced, and c) the scale of emotions caused by the personal victimisation. Cultural victimisation was comprised of: a) cultural multi-victimisation, b) frequency of cultural victimisation experienced, and c) the scale of emotions caused by cultural victimisation.

In order to carry out the CFA on the two factors, a covariance was introduced between them. The global adjustment indices showed that the model was not sufficiently robust, χ2(8 df) = 65.04, p < .01; CMIN/DF = 8.13; GFI = .98; AGFI = .94; CFI = .99; TLI = .98; RMSEA = .08. In order to obtain a better adjustment, theoretically justifiable modifications were made, keeping in mind the modification indices (MI). Three covariances between the measurement errors were introduced: a) between emotions caused by personal aggression and emotions caused by cultural aggression (MI = 29.15), b) between personal multi-victimisation and cultural multi-victimisation (MI = 28.34), and c) between frequency of personal victimisation and frequency of cultural victimisation (MI = 16.06). The fit of the model to the data was excellent, χ2(5 df) = 9.95, p = .07; CMIN/DF = 1.99; GFI = .99; AGFI = .98; CFI = .99; TLI = .99; RMSEA = .03.

Structural Model

In the second step, an initial hypothetical model was constructed which assumed that the percentage of the cultural majority in the school and the percentage of their own cultural group in the school were moderators of the possible effect of personal victimisation and cultural victimisation on cultural self-esteem, self-esteem, and emotional state (three socio-affective variables). In relation to these, the model included six direct paths, three which came from each victimisation variable tested and went to each one of the socio-affective variables mentioned. The initial model included other paths in which the links between the victimisation and socio-affective variables were moderated by the percentage of the presence of the cultural majority in the school and by the percentage of their own cultural group in the school. This initial model did not fit the data well, χ2(26 df) = 151.61, p = .00; CMIN/DF = 5.83; GFI = .97; AGFI = .93; CFI = .98; TLI = .96; RMSEA = .06 (see Figure 1).

Figure 1.

The initial model for predicting the direct and indirect effects of personal victimisation and ethnic-cultural victimisation on self-esteem and emotions. Pers Multiv = Personal multi-victimisation; Pers Emoti = Emotions about personal victimisation; Pers Frecu = Frequency of personal victimisation; Cultur Multiv = Cultural multi-victimisation; Cultur Emoti = Emotions about cultural victimisation; Cultur Frecu = Frequency of cultural victimisation.

In the search for an adjusted model, all of the paths in which the regression weights were not significant at the 1% level were eliminated (see the earlier table). Although the global adjustment indices demonstrated an improvement with respect to the initial model, the fit was still unacceptable, χ2(38 df) = 184.17, p = .00; CMIN/DF = 4.84; GFI = .97; AGFI = .94; CFI = .98; TLI = .97; RMSEA = .06. In order to finish adjusting the model, the endogenous variable, ‘percentage of their own cultural group in the school’, was eliminated, given that in the final proposed model it did not appear to act as a mediator or moderator between victimisation and the socio-emotional variables assessed in this study. The redefined model maintained the following paths: a) personal victimisation and self-esteem (β = -.24, p = .00); b) self-esteem and emotional state (β = .11; p = .00); c) self-esteem and cultural self-esteem (β = .22, p = .00); d) cultural victimisation and percentage of cultural majority in the school (β = -1.23; p = .00); e) percentage of cultural majority in the school and cultural self-esteem (β = .17, p =.00); and f) cultural self-esteem and emotional state (β = .05; p = .00). The redefined model also has a good fit with the data, χ2(29 df) = 60.93, p = .00; CMIN/DF = 2.10; GFI = .98; AGFI = .97; CFI = .99; TLI = .99; RMSEA = .03, and at the same time parsimonious (PCFI = .64). In this final model the dependent variables “cultural self-esteem and emotional” state explained 23% and 17% of the variance respectively (see Figure 2).

Figure 2.

The final model displays standardized coefficients predicting the direct and indirect effects of personal victimisation and ethnic-cultural victimisation on self-esteem and emotions. Pers Multiv = Personal multi-victimisation; Pers Emoti = Emotions about personal victimisation; Pers Frecu = Frequency of personal victimisation; Cultur multiv = Cultural Multi-victimisation; Cultur Emoti = Emotions about cultural victimisation; Cultur Frecu = Frequency of cultural victimisation. x² (29 df) = 60.93, p = .00; CMIN/DF = 2.10; GFI = .98; AGFI = .97; CFI = .99; TLI = .99; RMSEA = .03.

Several distinct multi-group analyses were carried out on this model in relation to the different participant variables. In relation to the two age groups, early adolescence (11-13 years, n = 493) and mid-late adolescence (14-18 years, n = 543), the model was well-adjusted, χ2(87 df) = 188.29, p = .00; CMIN/DF = 2.16; GFI = .98; AGFI = .96; CFI = .99; TLI = .99; RMSEA = .02. The regression weights of the model for each age group were significant (p < .05), and there were no differences between the two age groups. The same was found for gender. The model was well-adjusted for males and females, χ2(87 df) = 178.58, p = .00; CMIN/DF = 2.05; GFI = .98; AGFI = .96; CFI = .99; TLI = .99; RMSEA = .02, and there were no significant differences between the two.

In relation to the cultural group of the pupils (either cultural majority or cultural minority), the adjusted model showed a good fit, χ2(87 df) = 173.87, p = .00; CMIN/DF = 1.99; GFI = .98; AGFI = .96; CFI = .99; TLI = .99; RMSEA = .02. In the case of the cultural majority, all of the regression weights between the variables that made up the model were significant. However, for the cultural minority, two were not significant: 1) between cultural victimisation and percentage of cultural majority in the school (β = .04, p = .939) and 2) between cultural self-esteem and emotional state (β = .00, p = .964).

Discussion

When focussing on personal victimisation, personal self-esteem and the emotional state of adolescents from distinct cultural groups, it appears that being from a cultural majority or minority may not be a particular risk factor. Several recent studies have concluded that there is no significant difference between cultural groups in the prevalence levels of victimisation (not specified as having a cultural content) as a function of cultural group (Fandrem, et al., 2009; Monks et al., 2008; Rodríguez-Hidalgo et al., 2014; Sweeting & West, 2001). The results from the current study are in accord with this conclusion, expanding on this topic to demonstrate that is holds true for 1) personal multi-victimisation, 2) the frequency of personal victimisation experienced, and 3) the negativity of the feelings caused by being the target of this victimisation. In addition, it was noted that the levels of personal self-esteem and emotional state were relatively homogenous between adolescents regardless of their cultural group.

When focussing on victimisation which is cultural in nature and cultural self-esteem, we find clear evidence to support the suggestion that those adolescents from cultural minority groups are at heightened risk. For cultural victimisation, multi-victimisation, frequency, and level of negative feelings were significantly higher among those from cultural minority groups than those who were from the cultural majority. Furthermore, the cultural self-esteem of pupils in the minority groups was lower than that shown by those from cultural majority groups.

The results obtained reinforce the idea of studying both personal victimisation and cultural victimisation in order to be able to consider the range of victimisation among pupils in multi-cultural contexts. This led us to propose a model which integrates personal and cultural victimisation, which is something that has been noted in other studies (e.g., Rodríguez-Hidalgo et al., 2014). Furthermore, the current study demonstrates a step forward in the way in which the latent victimisation variables are formed, by incorporating three observed variables: multi-victimisation experienced, the frequency of being a victim, and the negativity of the feelings of the victim. These three variables are faithful to those proposed theoretically in the literature on peer-victimisation (Olweus, 1996).

The confirmation of the model integrating personal and cultural victimisation has enabled us to propose a structural equation model which is more complete in that it shows the effects of these variables over others. This model is successful for adolescents in this sample, independent of gender, age group, and whether they are from the cultural majority or minority.

Verkuyten and Thijs (2006) proposed 2 models based on which victimisation and ethnic discrimination (not exclusively bullying) were considered separately. In contrast, the present study integrated personal and cultural victimisation into the same model. This enabled us to contemplate other possible mediator and moderator variables and their effects on self-esteem, cultural self-esteem, and emotional state. The clearest agreement between the models of Verkuyten and Thijs and the model proposed by the current study is that an increase in personal victimisation has a direct negative effect on personal self-esteem.

In contrast to the models of Verkuyten and Thijs (2006), we did not find that cultural victimisation produces a negative effect on cultural self-esteem and that this in turn negatively impacts on general self-esteem. This new model considers cultural self-esteem not as a mediator, but as a variable indirectly affected by two routes from personal and cultural victimisation. The first route is common for participants from the cultural majority and cultural minority groups: personal victimisation produces a negative effect on personal self-esteem, which then impacts negatively on cultural self-esteem. This route is reinforced when adolescents recognise themselves not only as being not only victims of personal, but also targets of cultural aggression. The first route can be explained by considering the structural concept of global self-esteem: personal victimisation produces a negative effect on personal self-esteem (global), which has a negative effect on one part of its components, cultural self-esteem. The second route is more consistent among members of the cultural majority: the increase in the experience of cultural victimisation, when it takes place in contexts where the cultural majority is less numerous, produces a negative indirect effect on the cultural self-esteem of the victims. Considering the two routes described and the differences detected between the cultural groups, it is noteworthy that for adolescents from the cultural minority who suffer cultural victimisation, it is only when they also suffer personal victimisation that we see a negative impact on their general self-esteem and in consequence their cultural self-esteem.

Kim et al., (2011) and Shin, D’Antonio, Son, Kim, & Park (2011) show how discrimination and/or victimisation experienced by adolescents from Asian backgrounds in the USA has a negative effect on their psychological well-being. The model from our study is consistent with this, showing how the emotional state is indirectly and negatively affected by personal and cultural victimisation, mediated by personal and cultural self-esteem and the percentage of the presence of the cultural majority. The results of the present study enable us to see this relationship not only among adolescents from the cultural minority, but also among adolescents from the cultural majority, although there are some differences between these groups in the weight of these intermediate variables. This study suggests that there is a strong relationship between the experience of victimisation, self-esteem, and emotional adjustment. It also indicates that prevention and intervention programmes should include not only those pupils who feel excluded for being part of a cultural minority, but all pupils. Education programmes which focus on the development of self-esteem and healthy emotional adjustment can benefit not only more vulnerable pupils, such as those who come from minority groups, but also the student group as a whole. If levels of self-esteem and emotional adjustment are improved for all pupils, this may have the effect on the minority group among whom we may see a reduction in vulnerability to victimisation.

It is worth noting that the current study was cross-sectional in nature. In this study we consider the effects of victimisation on self-esteem. However, it is important to note that low self-esteem may predate victimisation and may act as both a risk factor and consequence of victimisation (e.g., Egan & Perry, 1998). Longitudinal studies would aid in making this distinction.

The model tested in the present study indicated the need to consider personal and cultural victimisation together if we want to address the effects of peer bullying. This model may have implications for the development of intervention programmes aimed at preventing and dealing with victimisation among peers in educational contexts in multi-cultural societies. Until now, psycho-educational programmes have tended to focus solely on personal victimisation, or on cultural discrimination as a wider social phenomenon (not just among peers). On the basis of the present model it is possible to infer that although these two lines of intervention may have positive effects for prevention and intervention, the effects may be limited. The model presented here opens the possibility of a third way of designing programmes to prevent/intervene with bullying and victimisation in multi-cultural contexts. These programmes could be made more effective by considering the personal and cultural characteristics of violence between peers as well as the effects on self-esteem, cultural self-esteem and emotional state. They could be seen as an alternative to those currently available and would consider personal victimisation and cultural victimisation together.

Future studies should focus on the search for the precursors and risk factors for interpersonal and intercultural violence among children and adolescents. This will enable the design of more precise strategies for prevention and palliative work.

Conflict of Interest

The authors of this article declare no conflict of interest.

Acknowledgements

The current work is a part of two research projects: 1) Violencia Escolar y Juvenil: los riesgos del Cortejo Violento, la Agresión Sexual y el Cyberbullying, VIEJUCO from I+D+i National Plan (Code PSI 2010-17246), granted by Ministerio de Educación y Ciencia, Spain; and 2) Parejas y Redes de Iguales en la Adolescencia: Un programa de Prevención de la Violencia en las Relaciones sentimentales from I+D National Plan (Code PSI2013-45118-R), granted by Ministerio de Economía y Competitividad, Spain.