Parental Socialization, Social Anxiety, and School Victimization: A Mediation Model

The aim of this study was to examine the relationship between parenting dimensions (involvement/acceptance vs. strictness/imposition) and school victimization, considering the possible mediating role of social anxiety. The sample comprised 887 adolescents (52.3% girls) aged between 12 and 16 (M = 13.84 and SD = 1.22) enrolled at three compulsory secondary education (“ESO” or “Educación Secundaria Obligatoria” in Spanish) schools located in the provinces of Valencia, Teruel and Seville (Spain). A structural equations model was developed using the Mplus 7.4 program. The results obtained indicate that social anxiety mediates the relationship between parenting dimensions (involvement/acceptance vs. strictness/imposition) and school victimization. Finally, the results and their potential theoretical and practical implications are discussed.


Introduction
School victimization is defined as a type of peer abuse in which students are subjected to violent verbal, physical, and psychological behavior by one or more peers [1,2]. This problem has serious consequences for the development and psychosocial adjustment of students [3][4][5]. As regards gender, data suggest that, in the case of school victimization, boys are victimized by their partners overtly and girls relationally [6,7], although some studies have not reported such differences [8,9]. Since the pioneering studies carried out by Olweus in the late 1970s, many studies and actions have been carried out to prevent victimization in the school context (see [6][7][8]). In recent years, a set of variables at individual, family, school, and social levels that related positively with school victimization have been routinely identified; for example, greater depressive symptomatology [9], problems in family functioning [10] and in school climate [11], and relations with social integration and relations with the community [12].
Another aspect that has aroused interest among social scientists is social anxiety, which has recently been incorporated in studies on victimization. Social anxiety is defined as the appearance of excessive fear of social situations in which adolescents feel that they may be negatively assessed by others [13]. It is seen as a problem that prevents the normal development of education and seriously affects the school climate and interpersonal relationships between peers [14,15]. Several studies have concluded that girls are more likely to suffer from this problem [15,16].
Previous research has shown the negative consequences that social anxiety has for the life of an adolescent; for example, low acceptance among peers [17,18], a negative perception of the adolescent about himself/herself [19,20], and his/her social skills [21,22]. These consequences, in turn, involve risk factors for adolescent victimization in the school context [23,24]. Indeed, these studies have associated social anxiety with a greater prevalence of this problem in classrooms [25,26]. However, it is important to note that most research has focused on clinical samples, and few studies have analyzed this problem in the school environment; hence, the lack of information in this regard [27,28].
Another aspect of demonstrated importance for adolescent psychosocial adjustment is parental socialization [29,30]. Traditionally, the relationship between parenting style and child adjustment has been analyzed following a two-dimensional orthogonal model of parental socialization, labeled as involvement/acceptance and strictness/imposition [31].
It is well known that certain parental attitudes and behavior patterns seem to be largely associated with the manifestation of difficulties in the adjustment and social competence of their children [32][33][34]. In this sense, recent research confirms the positive role played by parental acceptance and affection in the social adjustment of their children [35]. In contrast, it has been observed that parental styles characterized by excessive behavioral and psychological control [36,37], as well as the absence of support and affection, tend to be associated with emotional and behavioral problems in children [38,39]. In addition, some authors have highlighted that a lack of parental support and affection is a major risk factor for girls [40]; however, other authors have observed the opposite [41].
On the other hand, some studies in scientific literature have also explored the relationship between social anxiety and parental socialization. According to these studies, excessive behavioral and psychological control [36,37], as well as the absence of support and affection, increase the likelihood of experiencing social anxiety [42,43].
Ultimately, with respect to the abovementioned variables, both social anxiety and certain parental practices (authoritarian and negligent styles) have been described as risk factors for adolescent psychosocial adjustment [44,45]. However, there are still important issues to be resolved regarding the relationship between these variables and school victimization. Firstly, very few studies have analyzed the relationship between social anxiety and school victimization with samples of school adolescents. Secondly, it was confirmed that very few studies have investigated the influence of parental practices on victimization in schools. Lastly, previous studies have reported a relationship between certain parental practices and the development of social anxiety in adolescents [24,46]. However, as far as we have been able to review, no papers in the scientific literature have analyzed the role of social anxiety in the relationship between parental socialization and school peer victimization. It should be noted that social anxiety has been identified as a mediating construct interceding between stimuli in the social environment of adolescents and the latter's responses and reactions to these stimuli [47].
Given this background, the main aim of this study was to examine the relationship between parenting dimensions (involvement/acceptance vs. strictness/imposition) and school victimization, considering the potential mediating role of social anxiety. We also sought to determine the existence of any significant differences based on the gender of the adolescents. With this objective in mind, and as can be seen from the theoretical model proposed, the following hypotheses were proposed: H1: Parenting dimensions (involvement/acceptance vs. strictness/imposition) and social anxiety are directly related to school victimization (see Figures 1 and 2).

H2:
A direct relationship is expected to exist between parenting dimensions (involvement/acceptance vs. strictness/imposition) and social anxiety (see Figures 1 and 2).
H3: Parenting dimensions (involvement/acceptance vs. strictness/imposition) are indirectly related to school victimization through their relationships with social anxiety (mediating effect) (see Figures 1 and 2).
H4: The gender of adolescents is expected to have a moderating effect in these relationships.

Participants
The study was ex post facto transversal and descriptive. The study sample comprised 887 adolescents (52.3% girls) aged between 12 and 16 (M = 13.84 and SD = 1.22) enrolled at three compulsory secondary education (ESO) schools located in the provinces of Valencia, Teruel, and Seville (Spain). All of the participants attended state schools. Probabilistic sampling was performed. The primary units in the selected sample were urban geographic areas of the provinces of Valencia, Seville, and Teruel, and the secondary units corresponded to public centers in each area. Finally, the courses or classrooms were not used as tertiary units, since all of the students in the four courses comprising the ESO (compulsory secondary education) in all of the centers participated. The average amount data lost for the sample was 1.1%, and was never above 4% for individual measurements. Accurate estimates were obtained, in line with the expected values for the population due to the low level of data lost, which meant that the results were not likely to be biased [48].

Instruments
Parenting dimensions: The Parental Socialization Scale (ESPA29) [49] was used. This scale consists of 212 items that are measured using a response range of 1 (never) to 4 (always), in which the adolescents evaluate the performance of their parents in 29 situations representative of everyday family life in Western culture, of which 16 situations refer to the children's behaviors that fit the family norms (e.g., "I obey what s/he tells me to do"), and 13 refer to when their behaviors violate these norms (e.g., "If I don't study or don't want to do the homework they assign in high school."). The evaluation of the daily situations described in each item is based on the adolescent's perception of the responses in affection, dialogue, indifference, and dissatisfaction shown by his/her parents in each of them. The family score in acceptance/involvement was obtained by averaging the responses for affection, dialogue, indifference, and dissatisfaction (in the last two sub-scales, the responses were inverted by being negatively related with the dimension). The family severity/imposition score was obtained by averaging responses for verbal scolding, physical punishment, and revoking privileges. Both family indices ranged between 1 and 4 points; high levels of acceptance/involvement and severity/imposition corresponded to high scores. The psychometric properties of the scale are adequate (see Table 1). The confirmatory factor analysis (CFA) had a good model fit to the data (SBχ 2 = 87.8158,

Participants
The study was ex post facto transversal and descriptive. The study sample comprised 887 adolescents (52.3% girls) aged between 12 and 16 (M = 13.84 and SD = 1.22) enrolled at three compulsory secondary education (ESO) schools located in the provinces of Valencia, Teruel, and Seville (Spain). All of the participants attended state schools. Probabilistic sampling was performed. The primary units in the selected sample were urban geographic areas of the provinces of Valencia, Seville, and Teruel, and the secondary units corresponded to public centers in each area. Finally, the courses or classrooms were not used as tertiary units, since all of the students in the four courses comprising the ESO (compulsory secondary education) in all of the centers participated. The average amount data lost for the sample was 1.1%, and was never above 4% for individual measurements. Accurate estimates were obtained, in line with the expected values for the population due to the low level of data lost, which meant that the results were not likely to be biased [48].

Instruments
Parenting dimensions: The Parental Socialization Scale (ESPA29) [49] was used. This scale consists of 212 items that are measured using a response range of 1 (never) to 4 (always), in which the adolescents evaluate the performance of their parents in 29 situations representative of everyday family life in Western culture, of which 16 situations refer to the children's behaviors that fit the family norms (e.g., "I obey what s/he tells me to do"), and 13 refer to when their behaviors violate these norms (e.g., "If I don't study or don't want to do the homework they assign in high school."). The evaluation of the daily situations described in each item is based on the adolescent's perception of the responses in affection, dialogue, indifference, and dissatisfaction shown by his/her parents in each of them. The family score in acceptance/involvement was obtained by averaging the responses for affection, dialogue, indifference, and dissatisfaction (in the last two sub-scales, the responses were inverted by being negatively related with the dimension). The family severity/imposition score was obtained by averaging responses for verbal scolding, physical punishment, and revoking privileges. Both family indices ranged between 1 and 4 points; high levels of acceptance/involvement and severity/imposition corresponded to high scores. The psychometric properties of the scale are adequate (see Table 1). The confirmatory factor analysis (CFA) had a good model fit to the data (SBχ 2 = 87.8158,

Participants
The study was ex post facto transversal and descriptive. The study sample comprised 887 adolescents (52.3% girls) aged between 12 and 16 (M = 13.84 and SD = 1.22) enrolled at three compulsory secondary education (ESO) schools located in the provinces of Valencia, Teruel, and Seville (Spain). All of the participants attended state schools. Probabilistic sampling was performed. The primary units in the selected sample were urban geographic areas of the provinces of Valencia, Seville, and Teruel, and the secondary units corresponded to public centers in each area. Finally, the courses or classrooms were not used as tertiary units, since all of the students in the four courses comprising the ESO (compulsory secondary education) in all of the centers participated. The average amount data lost for the sample was 1.1%, and was never above 4% for individual measurements. Accurate estimates were obtained, in line with the expected values for the population due to the low level of data lost, which meant that the results were not likely to be biased [48].

Instruments
Parenting dimensions: The Parental Socialization Scale (ESPA29) [49] was used. This scale consists of 212 items that are measured using a response range of 1 (never) to 4 (always), in which the adolescents evaluate the performance of their parents in 29 situations representative of everyday family life in Western culture, of which 16 situations refer to the children's behaviors that fit the family norms (e.g., "I obey what s/he tells me to do"), and 13 refer to when their behaviors violate these norms (e.g., "If I don't study or don't want to do the homework they assign in high school."). The evaluation of the daily situations described in each item is based on the adolescent's perception of the responses in affection, dialogue, indifference, and dissatisfaction shown by his/her parents in each of them. The family score in acceptance/involvement was obtained by averaging the responses for affection, dialogue, indifference, and dissatisfaction (in the last two sub-scales, the responses were inverted by being negatively related with the dimension). The family severity/imposition score was obtained by averaging responses for verbal scolding, physical punishment, and revoking privileges. Both family indices ranged between 1 and 4 points; high levels of acceptance/involvement and severity/imposition corresponded to high scores. The psychometric properties of the scale are adequate (see Table 1). The confirmatory factor analysis (CFA) had a good model fit to the data (SBχ 2 = 87.8158, df = 5, p < 0.001, comparative fit index (CFI) = 0.937, root mean square error of approximation (RMSEA) = 0.047, I.C. 90 (0.036, 0.060)). Social anxiety: The Spanish version of the Social Anxiety Scale for Adolescents (SAS-A) [47], adapted by Olivares, Alcázar, and Piqueras [50], was used. This scale consists of 22 items that measure, on a response range of 1 (never) to 5 (always), the frequency with which the symptoms of social anxiety are expressed. The scale consists of three dimensions: Fear of negative evaluation (e.g., "I worry that others don't like me"), social avoidance and distress specific to new situations (e.g., "I get nervous when I talk to peers I don't know very well"), and social avoidance and distress experienced more generally (e.g., "I'm afraid to invite others to do things with me because they may say no"). The psychometric properties of the scale are adequate (see Table 1). The confirmatory factor analysis (CFA) had a good model fit to the data (SBχ 2 = 404.3731, df = 125, p < 0.001, comparative fit index (CFI) = 0.941, root mean square error of approximation (RMSEA) = 0.050, I.C. 90 (0.045, 0.056)).
School victimization: The Spanish version of the School Victimization Scale [51], adapted by Martínez-Ferrer, Moreno, and Musitu [4], was used. This scale consists of 20 items that measure, on a response range of 1 (never) to 4 (always), the frequency with which students have been subjected to violent behavior in the last year. The scale consists of three dimensions: Relational victimization (e.g., "a peer got angry with me and separated me from my group of friends to prevent me from playing or participating in any activity"), physical victimization (e.g., "a peer hit me to really harm me"), and verbal victimization (e.g., "a peer insulted me"). The psychometric properties of the scale are adequate (see Table 1). The confirmatory factor analysis (CFA) had a good model fit to the data (SBχ 2 = 439.1333, df = 163, p < 0.001, comparative fit index (CFI) = 0.938, root mean square error of approximation (RMSEA) = 0.044, I.C. 90 (0.039, 0.049)). Table 1 shows the reliability (Cronbach's alpha, composite reliability) and validity indexes (average variance extracted (AVE) and McDonald's omega), mean, and standard deviation for the scales and subscales used.

Procedure
Before carrying out the investigation, we interviewed the management teams in the schools to explain the objectives, the procedures to follow, the involvement of the teachers, tutors, and professionals, and the commitments of the research group in detail. The instruments were administered in the classroom, in two sessions lasting approximately 50 minutes each with a maximum interval of one week between them, after receiving the consent of the parents, teachers, and students. This administration of the instruments was always coordinated and supervised by previously trained researchers and without the presence of the teachers. These investigators explained the transcendence of the study to the participants, highlighting that their participation was voluntary and anonymous.

Data Analysis
Firstly, a correlation analysis was performed to analyze the relationships between the variables studied. These analyses were calculated using the SPSS 25 program. Secondly, two structural equations models were evaluated to verify the hypothetical models using the Mplus 7.4 program. Complementary analyses were also performed to determine the significance and magnitude of the potential mediating effect [52,53]. The maximum likelihood estimate with robust standard errors (MLR) was used to correct the non-normality of the variables [54]. The fixed models were evaluated using a combination of indices, including the Satorra-Bentler Chi-Square ratio/degrees of freedom (SB-χ 2 /df of approximately 2 or lower indicated a good fit), the comparative index set and the Tucker Lewis index (CFI/TLI ≥ 0.90 for a reasonable fit and CFI/TLI ≥ 0.95 for a good fit), the mean square root of the residuals (SRMR < 0.09), and the mean square root of the approximation (RMSEA < 0.06 for a proper fit [55]). Table 2 shows the means, standard deviations, and correlations between the variables studied. Significant relationships were found in the variables studied. As can be seen in Table 2, the mother's acceptance and involvement were negatively and significantly related to fear of negative evaluation, social avoidance and general anguish, relational victimization, overt physical victimization, and overt verbal victimization. The father's acceptance and involvement were negatively and significantly related to fear of negative evaluation, social avoidance and general anguish, relational victimization, overt physical victimization, and overt verbal victimization. The strictness/imposition of the mother was positively and significantly related to the strictness/imposition of the father, fear of negative evaluation, social avoidance and anguish in new situations, social avoidance and general anguish, relational victimization, overt physical victimization, and overt verbal victimization. Finally, the strictness/imposition of the father was positively and significantly related to fear of negative evaluation, social avoidance and anguish in new situations, social avoidance and general anguish, relational victimization, overt physical victimization, and overt verbal victimization.   To analyze the direct relationship between parental socialization and victimization, and indirectly between these two variables through social anxiety, two independent structural equations models (SEMs) were created based on acceptance/involvement and parental strictness/imposition. Table 3 shows the latent variables included in the models, their respective indicators, the standard error, and the associated probability for each indicator in the corresponding latent variable. Table 3. Factorial saturations, standard error, and associated probability.

Variables
Factor Loadings General Model The first model ( Figure 3) showed an adequate fit: S-B χ 2 = 60.03, gl = 16, p < 0.001, CFI = 0.99, RMSEA = 0.056 (0.041, 0.071). A direct and negative relationship was observed between parental acceptance/involvement and school victimization (β = −0.08, p < 0.05), as well as with social anxiety (β = −0.14, p < 0.001). Moreover, social anxiety was related directly and positively to school victimization (β = 0.40, p < 0.001). The percentage of variance associated with school victimization was 19.7%, which can be considered to be a size of the effect of the statistical significance of the estimated model.   The first model ( Figure 3) showed an adequate fit: S-B  2 = 60.03, gl = 16, p < 0.001, CFI = 0.99, RMSEA = 0.056 (0.041, 0.071). A direct and negative relationship was observed between parental acceptance/involvement and school victimization (β = −0.08, p < 0.05), as well as with social anxiety (β = −0.14, p < 0.001). Moreover, social anxiety was related directly and positively to school victimization (β = 0.40, p < 0.001). The percentage of variance associated with school victimization was 19.7%, which can be considered to be a size of the effect of the statistical significance of the estimated model. The second model (Figure 4) showed an adequate fit: S-B  2 = 54.30, gl = 16, p < 0.001, CFI = 0.99, RMSEA = 0.052 (0.037, 0.068). Parental strictness/imposition showed a significant direct relationship with school victimization (β = 0.10, p < 0.05) and with social anxiety (β = 0.16, p < 0.001). Social anxiety was also directly and positively related to school victimization (β = 0.42, p < 0.001). The percentage of variance associated with school victimization was 19.9%, which can be considered to be a size of the effect of the statistical significance of the estimated model. The second model (Figure 4) showed an adequate fit: S-B χ 2 = 54.30, gl = 16, p < 0.001, CFI = 0.99, RMSEA = 0.052 (0.037, 0.068). Parental strictness/imposition showed a significant direct relationship with school victimization (β = 0.10, p < 0.05) and with social anxiety (β = 0.16, p < 0.001). Social anxiety was also directly and positively related to school victimization (β = 0.42, p < 0.001). The percentage of variance associated with school victimization was 19.9%, which can be considered to be a size of the effect of the statistical significance of the estimated model. As regards the indirect relationships or mediation effects, the results showed that parental acceptance/involvement was negatively related to school victimization through social anxiety (β =−0.06, IC (−0.092 ; −0.026), p < 0.001) in the first model; in contrast, parental strictness/imposition was positively related to school victimization through social anxiety (β = 0.07, IC (0.034 − 0.103), p < 0.001) in the second model (see Table 4). Finally, a multi-group analysis was performed to analyze the moderating effect based on sex in both models. The effects of family acceptance/involvement and social anxiety in school victimization were estimated in model 1, and those of family strictness/imposition and social anxiety in school victimization were estimated in model 2. The models were restricted on the basis of gender (boys and girls). The first model (parental acceptance/involvement) presented a fit S-B 

Discussion
The aim of this study was to analyze the mediating role of social anxiety in the relationship between parenting dimensions (involvement/acceptance vs. strictness/imposition) and school victimization in school adolescents.
Firstly, the results confirmed a direct and negative relationship between the educational practices of parental acceptance and involvement and school victimization, and a direct and positive As regards the indirect relationships or mediation effects, the results showed that parental acceptance/involvement was negatively related to school victimization through social anxiety (β = −0.06, IC (−0.092; −0.026), p < 0.001) in the first model; in contrast, parental strictness/imposition was positively related to school victimization through social anxiety (β = 0.07, IC (0.034 − 0.103), p < 0.001) in the second model (see Table 4). Finally, a multi-group analysis was performed to analyze the moderating effect based on sex in both models. The effects of family acceptance/involvement and social anxiety in school victimization were estimated in model 1, and those of family strictness/imposition and social anxiety in school victimization were estimated in model 2. The models were restricted on the basis of gender (boys and girls). The first model (parental acceptance/involvement) presented a fit S-B χ

Discussion
The aim of this study was to analyze the mediating role of social anxiety in the relationship between parenting dimensions (involvement/acceptance vs. strictness/imposition) and school victimization in school adolescents.
Firstly, the results confirmed a direct and negative relationship between the educational practices of parental acceptance and involvement and school victimization, and a direct and positive relationship between parental severity and imposition and this problem, albeit with a low size effect, as evidenced in previous studies [56,57].
In other words, the positive role played by parental acceptance and affection as a protective factor against emotional and behavioral problems in children was confirmed, in this case, against the victimization of adolescents in school [35]. In parallel, a direct and positive relationship was observed between social anxiety and school victimization. This result coincides with the findings reported in other studies, which have shown that social anxiety, i.e., fear of negative evaluation, fear of the negative opinion of others, and avoidance behaviors when faced with new social situations, is an experience that seems to promote school victimization [58,59].
Secondly, social anxiety was observed to be negatively related to parental involvement/acceptance and positively to parental strictness/imposition. Thus, the results obtained in this study suggest the existence of a protective relationship with respect to the practices of responsiveness and affection, i.e., based on the frequency of eulogies and praise as well as empathy and dialogue, in social anxiety situations, while the relationship is at risk when the practices are characterized as being coercive, imposed, and based on submission, obedience, and control, as well as with a low level of affection and involvement. In other words, the feeling of being accepted, loved, and protected by parents is one of the main resources that enrolled adolescents have for not experiencing this excessive and persistent fear of being observed and judged by others in situations of social interaction. These results confirm those described by Gómez, Romera, Jiménez-Castillejo, Ortega-Ruiz, and García-López [36], who reported that parental educational practices represent a family asset that either promotes or hinders the development of social and emotional attitudes and skills in students and, through them, either favors or prevents the appearance of problems such as social anxiety in the classroom.
As regards the indirect effects of the present study, our findings support the mediating function of social anxiety and suggest that, in this relationship, parenting dimensions may either enhance or inhibit relevant domains of social anxiety (fear of negative evaluation, avoidance and social anxiety in new situations, and avoidance and social anxiety experienced in general), symptoms that are, in turn, important predictors of school victimization (relational, physical, and psychological). The observed mediation process is very new and interesting because it suggests that the characteristics of parental socialization, in terms of acceptance and involvement as well as strictness and imposition, influence school victimization, mainly because they previously influence their response in the social interaction of adolescents. Our findings could suggest that adolescents' perceptions of the socialization practices used by their parents not only influence the appearance, development, and maintenance of social anxiety [27,34,60], but also their relationships with their peers in the classroom, which would be an extension of the family system.
Lastly, for both sexes, parental practices (involvement/acceptance vs. strictness/imposition) are indirectly related to school victimization through their relationships with social anxiety (mediating effect). These results coincide with those described in previous studies, which have reported that the relationships between parental socialization styles and children's adjustment do not vary with changes in demographic variables such as gender [61,62]. One explanation for this result could be the types of victimization between boys and girls. Recent studies have indicated that boys are victimized by their peers overtly and girls relationally [63,64]. Further research into these relationships in multiple expressions of violence would provide more in-depth knowledge of the underlying explanatory mechanisms.
However, the results of the present study have certain limitations. It should be noted that a cross-sectional design was employed in this study, implying that caution should be exercised when making causal inferences based on available data. The directionality between the variables tested in this study was based on both theoretical foundations and empirical evidence found in previous studies. In spite of the foregoing, a longitudinal study would be needed to shed more light on these associations and draw more reliable conclusions about the causal directions of the relationships between the variables. Future research should incorporate different informants and take into account the perceptions of both teachers and parents. Moreover, this study was cross-sectional in nature. Therefore, it is important to highlight that a bilateral relationship may have existed between the variables studied. For this reason, a longitudinal study would be necessary to shed more light on these associations and determine with greater certainty the causal directions of the relationships between the variables. Future research should examine a broader range of information sources and obtain the perceptions of both teachers and parents. In addition, parenting dimensions (acceptance/involvement and strictness/imposition) are theoretically independent, i.e., orthogonal; the fact that parents use parenting practices from one dimension does not exclude the use of the practices defining the other dimension. In future research, it would be worthwhile to consider parenting styles (authoritative, indulgent, authoritarian vs. negligent) to test possible interaction effects between both variables in the different adjustment criteria evaluated.

Conclusions
Despite these limitations, the results of this study contribute to a better understanding of the relationship between social anxiety and school victimization in samples of school adolescents. These findings revealed a relationship between parenting dimensions and victimization in the school setting. Likewise, this study confirmed the key role played by social anxiety as a mediating construct interceding between the socialization practices of adolescents' parents and school victimization. Therefore, specific intervention programs should be promoted to address social anxiety with parents and adolescents, a problem that, as highlighted, is associated with parental socialization practices. Funding: This study was funded by the project "Bullying, cyberbullying child-parent violence in adolescence"-reference: PSI2015-65683-P (MINECO/FEDER, UE), and was subsidized by the Ministry of Economy and Competitiveness of Spain and the European Union through the European Regional Development Fund (in Spanish, ERDF)-"A way of making Europe".