Empirical Research on the Relationship between Family Economic, Social and Cultural Status and Students’ Exposure to School Bullying: Mediating Effects of Parental Support and Teacher Support Liang Huang 1,2 , Decheng Zhao 3 1. The Chinese University of Hong Kong, Hong Kong, China 2. Southeast University, Nanjing, China 3. Beijing Normal University, Beijing, China Abstract. Using survey data collected by the Organization for the Eco- nomic Cooperation and Development (OECD) in its 2015 Program for International Student Assessment (PISA2015), this study explores the relationship between family economic, social and cultural status (ESCS) and students’ exposure to school bullying for students in Beijing- Shanghai-Jiangsu-Guangdong (China). Additionally, the study examines the mediating effects of parental support and teacher support on the re- lationship between family ESCS and students’ exposure to school bully- ing. Lower family ESCS led to significantly increased exposure to school bullying. Parental support and teacher support mediated the relation- ship between family ESCS and students’ exposure to school bullying, with slightly different magnitudes. The results suggest that China should continue to improve its system for school bullying prevention and treat- ment, and should especially focus on reducing bullying victimization among students from disadvantaged backgrounds. Further, parents in disadvantaged families should provide support and care to their children to enhance their abilities to tackle bullying victimization, and teachers should treat students fairly and guide students in socializing with their peers appropriately. Best Evid Chin Edu 2019; 1(1):15-27. Doi: 10.15354/bece.19.ar1006 Keywords: PISA2015; Family ESCS; School Bullying; Parental Support; Teacher Support; Mediating Effect Vol. 1, No. 1, 2019 16 RELATIVELY high prevalence of school bullying in China, with events of bullying and school violence being reported, has been observed recently (Wang, 2016). Bullying is widely recognized as a specific form of aggressive behavior that is repetitive, intentional, and based on a power imbalance (Smith et al., 1999). School bullying experienced by adolescent students, who are in a period of tran- sition, has negative effects on their physical and mental health, on their academic capa- bility and social adaptability, and can be irreversibly detrimental to the students’ long- term wellbeing (Woods et al., 2004). Research reveals that bullying victimization se- verely hampers students’ educational attainment and personality development. Students frequently exposed to school bullying have lower educational performance and are more likely to suffer from various psychological symptoms, such as anxiety, depression, loneliness, and suicidal tendencies, compared to their counterparts (Delprato et al., 2017; Kaltiala-Heino et al., 2000). In fact, the high prevalence of school bullying and its se- vere impacts has become a major policy concern in China. In 2016, a national anti- bullying policy, Guidance on the Prevention and Treatment of Bullying and School Vio- lence, that required schools’ active participation in counteracting school bullying and protecting students’ rights, was promulgated by the Ministry of Education and eight other central ministries. Consequently, it is imperative for researchers and practitioners to identify the factors associated with school bullying and to help design effective poli- cies that support and protect at-risk students (Olweus, 1994). Of the risk factors for bullying victimization, family socioeconomic background is recognized as an important but still under-researched predictor. Prior research by schol- ars in other countries has shown connections between students’ exposure to school bul- lying and their family socioeconomic backgrounds. In many cases, students with disad- vantaged family backgrounds tend to suffer peer victimization at a higher rate than stu- dents from advantaged families. For example, Due et al. (2009), using data from the international ‘Health Behavior in School-aged Children’ study involving 35 countries, examined the socioeconomic inequality in students’ exposure to school bullying and found that students from families of low affluence were more likely to be bullied. Con- sistent with this, Analitis et al. (2009) explored the factors associated with bullying vic- timization in children and adolescents aged 8 to 18 and found that students whose par- ents had a lower educational level reported more peer victimization. Empirical findings suggest that the inequality in students’ exposure to bullying that exists in various cultur- al contexts may prevent some students from equally healthy development. As a result, this has become a concern for policy-makers, researchers, and practitioners worldwide. Nonetheless, empirical research on the mechanisms underlying the inequality of About the Authors: Liang Huang is a Lecturer at Department of Public Administration, Southeast University, China; and Decheng Zhao is a Professor at Faculty of Education, Beijing Normal University, China. Correspondence to: Liang Huang, Department of Public Administration, Southeast University, China. E-mail: eliot_huang@163.com. Conflict of Interests: None. A Vol. 1, No. 1, 2019 17 students’ exposure to bullying is limited, specifically the ways in which variances in family economic, social and cultural status (ESCS) affect the likelihood of a student being bullied. An emerging line of empirical evidence suggests social support is im- portant in protecting students from being bullied. Students from families with lower ESCS may be more likely to be bullied because they receive less social support com- pared with other students (Jansen et al., 2011). As parents and teachers are the most important sources of social support for students facing possible bullying, we hypothe- size that inequality in family ESCS among students may result in an inequality of pa- rental support and teacher support and that this may further lead to inequalities in stu- dents’ exposure to school bullying. Specifically, two questions guided this research: 1. What are the effects of family ESCS on students’ exposure to school bullying? 2. Do parental support and teacher support mediate the effects of family ESCS on students’ exposure to school bullying? Data from students in China’s four provinces of Beijing, Shanghai, Jiangsu, and Guangdong were used in this study. The findings of this study can inform the design of effective anti-school bullying policies to put into practice. Methods Sample The Organization for Economic Cooperation and Development (OECD) collects data on students’ abilities from countries worldwide through its Program for International Student Assessment (PISA). In the 2015 assessment round (hereafter, PISA2015), PISA for the first time included questions to evaluate bullying in schools of participating countries and economies (OECD, 2016). The school bullying variable, together with other variables in the PISA2015 database reflecting students’ individual characteristics, family backgrounds and school characteristics, provided the data needed for this study. Listwise deletion was employed to exclude those cases with missing data; 8,671 15- year-old students from the PISA2015 Beijing-Shanghai-Jiangsu-Guangdong (China) database were identified as the sample of this study. Variables Students’ Exposure to School Bullying PISA2015 measured students’ exposure to school bullying from the perspective of the victims (OECD, 2017). Students were asked to report their experiences during the past one year with six forms of school bullying actions: 1) “Others spread rumors about me”; 2) “Being left out of things by others intentionally”; 3) “Being made fun of by others”; 4) “Being threatened by others”; 5) “Others took away or destroyed my things”; and 6) “Being hit or pushed by others” (OECD, 2017). The frequency of students experiencing any one of the six forms of school bullying was classified into four categories: 1) “never or almost never”; 2) “a few times a year”; 3) “a few times a month”; and 4) “once a Vol. 1, No. 1, 2019 18 week or more” (OECD, 2017). Based on Maynard et al. (2016), students with responses of “a few times a month” or “once a week or more” were defined as being exposed to that specific form of school bullying action. Students exposed to least one form of school bullying were defined as being exposed to school bullying and assigned a value of 1 for the bullying variable; other students were assigned a value of 0. In this study, 21.74% of students reported being exposed to school bullying during the past one year. Family ESCS PISA computed an index of family ESCS using family background information, such as parents’ highest education level, parents’ highest occupation status, and home posses- sions (OECD, 2017). The family ESCS was transformed with 0 representing the aver- age family ESCS of OECD students and 1 representing the standard deviation (SD) of the OECD students’ family ESCS. In the subset of data used for this study, student fam- ily ESCS had a mean of -0.84 and a SD of 1.12. Parental Support PISA2015 asked students to report the extent to which their parents support them in their daily living, school learning and other activities (e.g., “My parents are interested in my school activities,” and “My parents support me when I am facing difficulties at school,” and “My parents encourage me to be confident”) (OECD, 2017). From the re- sponses, PISA computed an index of parental support, which was transformed with 0 representing the average parental support of OECD students and 1 representing the SD of the OECD students’ parental support. In this study, the sample students’ parental support had a mean of -0.16 and a SD of 0.88. Teacher Support PISA2015 asked students to report the frequency of the help, care and support they got from their science teachers in classroom learning (e.g., “The teachers give extra help when students need it,” and “The teachers help students with their learning,” and “The teachers give students an opportunity to express their opinions”) (OECD, 2017) and used this to compute an index of teacher support. Once again, the index of teacher sup- port was transformed with 0 representing the average teacher support of OECD students and 1 representing the SD of the OECD students’ teacher support. For the sample subset used here, the students’ teacher support index had a mean of 0.28 and a SD of 0.92. Control variables Included control variables related to individual students were gender (52% males, 48% females) and study program (56% lower secondary school students, 35% upper second- ary school students, and 8% vocational school students). Also included were control variables related to school characteristics: school location (37% urban schools, 63% Vol. 1, No. 1, 2019 19 rural schools), school type (90% public schools, 10% private schools), school size (in hundreds of students, mean = 19.70, SD = 19.18), class size (number of students, mean = 41.66 , SD = 9.37), student-teacher ratio (mean = 12.62, SD = 8.55), proportion of all teachers fully certified (mean = 0.97, SD = 0.07), school educational material shortage (mean = 0.07, SD = 1.22), and school educational staff shortage (mean = 0.77, SD = 1.28). Data Analysis Data analysis involved three phases. In phase one, descriptive statistics were used to preliminarily examine whether exposure to school bullying differed among students with different family ESCS. In phase two, a binary logit model was constructed to ana- lyze the predictive effects of family ESCS on students’ exposure to school bullying. The binary logit model took the following form, Yi = α0 + + γ1ESCSi + ui1 (1) where Yi refers to students’ exposure to school bullying, α0 is a constant, xij is a vector of control variables, ESCSi is the student’s family ESCS, and ui1 is the error term. In phase three of the data analysis, a path analysis model was constructed to exam- ine the mediating roles of parent support and teacher support on the relationship be- tween family ESCS and students’ exposure to school bullying. The path analysis model has the advantage that it can be used to simultaneously estimate casual relationships among multiple variables (Wang, 2014). The path analysis model had the following form, Mti = α1 + + γ2ESCSi + ui2 (2) Yi = α2 + + γ3ESCSi + γ4Mti + ui3 (3) where Yi refers to students’ being exposed to school bullying, α1 and α2 are constants, Xij is a vector of control variables, Mti are the mediating variables between family ESCS and students’ exposure to school bullying, including the parental support variable (M1) and the teacher support variable (M2), and ui2 and ui3 are error terms. As parental support and teacher support are correlated, the path coefficient between the two mediat- ing variables in the model was set to be freely estimated. Since PISA2015 employed two-stage sampling methods to select sample students 1 , for phases two and three of the data analysis, the standard errors of parameter estimates were estimated using both the final student weighting and Fay’s balanced repeated rep- lication (BRR) method with a 0.5 coefficient (OECD, 2009). Moreover, in phase three data analysis, a bias-corrected bootstrap method was employed for analysis of mediat- ing variables. The bootstrap method as a resampling estimation method produces more accurate interval estimates (Hayes, 2009). In this study, resampling was set to 2,000 Vol. 1, No. 1, 2019 20 times. The software SPSS 23.0 was used to conduct descriptive statistics (phase one) and Mplus 7.0 was used to conduct data analysis (phases two and three). Results Descriptive Statistics To initially explore whether students’ exposure to school bullying was affected by fami- ly ESCS background, we grouped students into three categories based on their family’s ECSC: (1) advantaged family backgrounds – student’s family ESCS was in the top quarter of the index of the sample students’ family ESCS; (2) average family back- grounds – student’s family ESCS was between the 25th to 75th percentile of the index of the sample students’ family ESCS; and (3) disadvantaged family backgrounds – stu- dent’s family ESCS was in the bottom quarter of the index of the sample students’ fami- ly ESCS. We then used descriptive statistics for each group to look for differences among groups in students’ exposure to school bullying behavior. Table 1 shows the results of the descriptive statistics. In general, students’ exposure to any form of school bullying was less for students from advantaged family backgrounds (18.96%) than for students from average family ESCS backgrounds (21.55%), which, in turn, was less than that for students from disadvantaged backgrounds (24.91%). Students’ exposure to school bullying differed significantly among the three types of family backgrounds (χ 2 = 22.75, p < 0.001). The percentage of students exposed to each specific form of school bullying simi- larly decreased when going from disadvantaged family backgrounds to average family backgrounds to advantaged family backgrounds. The percentage of students who re- ported “Being left out of things by others intentionally,” or “Being made fun of by oth- ers,” or “Others took away or destroyed my things,” or “Being threatened by others” differed significantly among the three types of family backgrounds (Chi-squared p- value ≤ 0.01 in all tests). This further verified that differences exist in students’ expo- sure to school bullying under different family backgrounds. Binary Logit Model: Effects of Family ESCS on Students’ Exposure to School Bully- ing Given the differences observed above in students’ exposure to school bullying based on different family backgrounds, we next analyzed the predictive effects of family ESCS on students’ exposure to school bullying by employing a binary logit model (equation (1) above) to the data. Table 2 displays these results. Without controlling for other vari- ables, family ESCS had a significant negative effect on students’ exposure to school bullying (γ = -0.055, p < 0.001). The predictive effects of family ESCS on students’ exposure to school bullying remained significant after controlling for variables related to individual students (γ = -0.052, p< 0.01) or variables related to school characteristics (γ = -0.040, p < 0.05). The results of the binary logit model are consistent with the re- Vol. 1, No. 1, 2019 21 Table 1. Students’ Exposure to School Bullying under Different Family Back- grounds (%). Exposure To School Bullying Actions All Students Disadvantaged Family Back- grounds Average Fami- ly Back- grounds Advantaged Family Back- grounds x 2 Exposure to any form of school bullying 21.74 24.91 21.55 18.96 22.75*** Others spread rumors about me 6.30 6.50 6.34 6.00 0.51 Being left out of things by others intentionally 7.57 9.87 7.24 5.90 25.68*** Being made fun of by others 11.50 13.93 11.35 9.36 22.40*** Being threatened by others 3.29 4.11 3.37 2.31 11.21** Others took away or destroy my things 12.57 14.30 12.57 10.84 11.80** Being hit or pushed by others 4.29 4.57 4.41 3.78 1.90 *P<0.05，**P<0.01，***P<0.001 sults of the descriptive statistics and both show that students with higher family ESCS are less likely to be bullied. Path Analysis: Mechanisms Underlying Effect of Family ESCS on Students’ Expo- sure to School Bullying To further explore the mechanisms underlying effect of family ESCS on students’ expo- sure to school bullying, we conducted path analysis using path analysis equations (2) and (3) (see Methods). The weighted root mean-square residual (WRMR), an indicator of the path model fit, was 0.002, indicating a good model fit. 2 Table 3 shows the results of the path analysis model. Results reveal that the direct predictive effects of family ESCS on students’ exposure to school bullying was -0.019 (p > 0.05). Family ESCS affected students’ exposure to school bullying indirectly through its influence on parental support and teacher support of students. The predictive effects of family ESCS on parental support was 0.138 and the predictive effects of paren- tal support on students’ exposure to school bullying was -0.086; both these path coeffi- cients reached at least a 0.01 level of significance. The predictive effects of family ESCS on teacher support was 0.070 and the predictive effects of teacher support on students’ exposure to school bullying was -0.142; both path coefficients had p < 0.001. Vol. 1, No. 1, 2019 22 Analysis of mediating variables was conducted using the bias-corrected bootstrap method. Table 4 shows the results. The 95% bootstrap confidence interval (CI) of both the indirect effects of family ESCS on students’ exposure to school bullying through parental support and the indirect effects of family ESCS on students’ exposure to school bullying through teacher support did not include zero, indicating significant mediating effects. The magnitude of the mediating effects of parental support and teacher support on the relationship between family ESCS and students’ exposure to school bullying were slightly different. Thus, the mechanism underlying family ESCS effects on students’ exposure to school bullying has been elucidated: students with a higher family ESCS get more pa- rental support and teacher support, and students with more parental support and teacher support have a lower risk of being bullied. Conversely, students with lower family ESCS get less parental support and teacher support, and students with less parental sup- port and teacher support have a higher risk of being bullied. Table 2. The Predictive Effects of Family ESCS on Students’ Being Exposed to School Bullying. Students’ Exposure to School Bullying Coefficient SE Coefficient SE Coefficient SE Family ESCS -0.055*** 0.016 -0.052** 0.019 -0.040* 0.020 Male (Reference group: Female) 0.423*** 0.046 0.422*** 0.046 Lower secondary school students (Reference group: Vocational school students) -0.030 0.079 -0.114 0.122 Upper secondary school students (Reference group: Vocational school students) -0.105 0.087 -0.166 0.108 Urban school (Reference group: Rural school) -0.025 0.067 Public school (Reference group: Private school) -0.108 0.113 School size -0.002 0.002 Class size 0.007 0.005 Student-teacher ratio 0.003 0.003 Proportion of all teachers fully certified 0.027 0.317 School educational material shortage 0.008 0.029 School educational staff shortage 0.012 0.036 Threshold 0.833*** 0.033 1.018*** 0.076 1.148** 0.375 *P<0.05, **P<0.01, ***P<0.001 Vol. 1, No. 1, 2019 23 Discussion and Implications Discussion In this study, we found that students’ exposure to school bullying differed depending on family ESCS background. Students from advantaged family backgrounds had decreased exposure to bullying (18.96%) compared to those from disadvantaged family back- grounds (24.91%). We also verified that the inequality of family ESCS is a key factor affecting the likelihood of students being bullied in the four China provinces of Beijing, Shanghai, Jiangsu, and Guangdong. Bullying victimization can hinder students’ physi- cal and mental development, their educational attainment, and undermine their social Table 3. Results of the Path Analysis Model. 3 Parental support (M1) Teacher support (M2) Students’ Exposure to School Bullying Coefficient SE Coefficient SE Coefficient SE Family ESCS 0.138*** 0.013 0.070*** 0.015 -0.019 0.020 Parental support -0.086** 0.028 Teacher support -0.142*** 0.022 Male (Reference group: Fe- male) -0.078*** 0.023 -0.090*** 0.026 0.402*** 0.047 Lower secondary school stu- dents (Reference group: Vo- cational school students) 0.026 0.094 0.113 0.071 -0.095 0.118 Upper secondary school stu- dents (Reference group: Vo- cational school students) 0.109 0.094 -0.024 0.064 -0.161 0.103 Urban school (Reference group: Rural school) -0.045 0.034 0.011 0.040 -0.028 0.067 Public school (Reference group: Private school) -0.100 0.067 -0.086 0.081 -0.129 0.109 School size 0.000 0.001 -0.001 0.001 -0.002 0.002 Class size 0.002 0.002 -0.007* 0.003 0.006 0.005 Student-teacher ratio -0.004* 0.002 -0.008 0.005 0.001 0.003 Proportion of all teachers fully certified 0.850*** 0.221 0.698 0.385 0.194 0.284 School educational material shortage -0.017 0.016 -0.054* 0.027 -0.001 0.029 School educational staff shortage -0.027 0.015 -0.011 0.024 0.008 0.035 Intercept -0.816*** 0.208 0.101 0.397 Threshold 1.207*** 0.386 *P<0.05, **P<0.01, ***P<0.001 Vol. 1, No. 1, 2019 24 Table 4. Mediation Analysis Using Bias-Corrected Bootstrap Method. Influence Path Estimate 95% Bootstrap CI Coefficient SE Lower Upper Family ESCS → parental support → students’ exposure to school bullying -0.012** 0.004 -0.015 -0.008 Family ESCS → teacher support → students’ exposure to school bullying -0.010*** 0.003 -0.012 -0.008 Total indirect effects -0.022*** 0.005 -0.027 -0.017 *P<0.05, **P<0.01, **P<0.001 and economic welfare in adulthood (Brown et al., 2008). As a result, students from dis- advantaged families in the four provinces of China may not enjoy equal opportunities to flourish due to an increased exposure to school bullying. We identify the mechanism by which family ESCS influences students’ exposure to school bullying: family ESCS indirectly impacted students’ exposure to school bully- ing through its effect on parental and teacher support available to students. Specifically, family ESCS inequality leads to inequality in levels of parental support and teacher support, which further created inequality of students’ being bullied. In terms of the me- diating role of parental support, parents from families with higher ESCS tend to have higher educational levels. This may mean they make more conscious efforts to get in- volved in their children’s education and development and are more willing to com- municate with their children to help solve the learning and living problems their chil- dren confront (Lereya et al., 2013). Caring and supportive ties between parents and children are conducive to the development of students’ academic adaptability and social skills and thus can protect students from being bullied (Rivara et al., 2016). In contrast, parents from families with lower ESCS have limited resources, capabilities, and time and thus may be less able to provide their children with timely and effective help when their children face difficulties, making these students more likely to be bullied. In terms of the mediating role of parental support, parents from families with higher ESCS can mobilize more social and cultural resources to become better involved in their child’s school educational activities and can actively build cooperative relations with teachers, which will help their children get more support from teachers (McNeal et al., 1999). Support from teachers helps students gain peer acceptance and enhance their social competency, which can reduce students’ exposure to school bullying (Troop-Gordon et al., 2011). In contrast, parents from families with lower ESCS lack the relevant re- sources and capabilities to create a favorable external developmental environment for their children, making these students more likely to be bullied. Implications Vol. 1, No. 1, 2019 25 In light of the discussion above, the following recommendations are offered. First, China should continue to improve its system of school bullying prevention and treatment to reduce bullying victimization among students from disadvantaged backgrounds. School bullying prevention and treatment has become a major concern for the country’s policy-makers. To promote the implementation of effective school bully- ing prevention and treatment initiatives, China should continue to improve its system of school bullying prevention and treatment to comprehensively tackle all forms of school bullying and to guarantee that all students have equal and sufficient opportunities for healthy physical and mental development. For students in schools within Beijing, Shanghai, Jiangsu, and Guangdong provinces, family ESCS inequality resulted in stu- dents from disadvantaged families experiencing disproportionately more bullying. Therefore, while formulating the school bullying prevention and treatment policy, the country should give more policy support to students from disadvantaged ESCS families. Likewise, educational authorities should invest more resources and professional support in rural schools, disadvantaged schools, and migrant schools that have higher propor- tions of students from disadvantaged family backgrounds and should help these schools build up their bullying prevention and treatment system to reduce the risk of school bul- lying in disadvantaged students. Second, parents in disadvantaged families should give more support and care to their children to enhance their abilities to resist bullying victimization. The role of par- ents in helping students counteract school bullying is critically important (Liao et al., 2017). However, parents from families with lower ESCS often cannot offer such help, care, and support to their children, leaving their children exposed to bullying victimiza- tion. In this study, we propose that even though parents of disadvantaged families may lack the corresponding resources and capabilities to provide their children with a better developmental environment, they can still be actively involved in their children’s edu- cation, pay more attention to their children’s mental and behavioral development, and improve the parent-child relationship through increased close communication and inter- action to help their offspring cope with difficulties in learning, living and peer interac- tion and to protect them from the harmful effects of being bullied. We suggest that par- ents make a conscious effort to learn how to prevent and treat school bullying and to help their children recognize school bullying and make appropriate responses. In addi- tion, parents who find that their children have been exposed to school bullying should seek the support of communities and schools to help these students overcome bullying victimization. Third, teachers should treat students fairly and guide students to socialize with their peers appropriately. Teachers are a valuable resource that students can rely on for help with school bullying. The help and care teachers give to students is conducive to broadening student-teacher communication channels and building positive student- teacher relations. This enables students to actively seek their teachers’ support when in need and helps student get the acceptance and recognition of peers. In this study, we found that the teacher support that students receive varied for students from different Vol. 1, No. 1, 2019 26 family backgrounds, such that students with higher family ESCS have a lower risk of being bullied and students with lower family ESCS have a higher risk of being bullied. To reduce exposure to school bullying for all students, we suggest that teachers treat students fairly, analyze the behavioral and mental progress of their students, and pro- vide timely help and support to students facing learning, living and socializing difficul- ties. Additionally, teachers can help students cultivate good conduct, such as being friendly and helpful, and respecting others, and can guide students to socialize with their peers appropriately so that students from disadvantaged family backgrounds better integrate themselves into their peer group. Notes 1. In the first stage of sampling, PISA employed probability proportional to size (PPS) to select sam- ple schools. In the second stage of sampling, PISA randomly selected students from sample schools. 2. For a path analysis model with continuous and categorical variables, model fit is acceptable if the weighted root mean-square residual (WRMR) is less than 0.9. For more detail, please refer to Schreiber, J. B., Nora, A., Stage, F. K., Barlow, E. A., & King, J. (2006). Reporting structural equation modeling and confirmatory factor analysis results: a review. The Journal of Educational Research, 99(6), 323-338. 3. Results of the path analysis model estimate the correlation coefficient between parental support and teacher support as 0.150 (P < 0.001). This result was not included in Table 3. References Analitis, F., Velderman, M. K., Ravenssieberer, U., Detmar, S., Erhart, M., & Herdman, M., et al. (2009) Being bullied: associated fac- tors in children and adolescents 8 to 18 years old in 11 european countries. Pediatrics; 123(2):569-577. Brown, S., & Taylor, K. (2008) Bullying, educa- tion and earnings: evidence from the nation- al child development study. Econ Edu Rev; 27(4):387-401. Delprato, M., Akyeampong, K., & Dunne, M. (2017) The impact of bullying on students’ learning in Latin America: a matching ap- proach for 15 countries. Int J Edu Dev; 52:37-57. Due, P., Merlo, J., Harelfisch, Y., Damsgaard, M. T., Soc, M.S. , & Holstein B.E., et al. (2009) Socioeconomic inequality in exposure to bullying during adolescence: a comparative, cross-sectional, multilevel study in 35 coun- tries. Am J Pub Health; 99(5):907-914. Hayes, & Andrew, F.. (2009) Beyond baron and kenny: statistical mediation analysis in the new millennium. Commun Monograph; 76(4):408-420. Jansen, D.E., Veenstra R., Ormel, J., Verhulst, F.C., & Reijneveld, S.A.. (2011) Early risk factors for being a bully, victim, or bul- ly/victim in late elementary and early sec- ondary education. the longitudinal trails study. BMC Pub Health; 11(1):440. Kaltiala-Heino, R., Rimpel, M., Rantanen, P., & Rimpel, A. (2000) Bullying at school—an indicator of adolescents at risk for mental disorders. J Adolesc; 23(6):661-674. Lereya, S.T., Samara, M., & Wolke, D. (2013) Parenting behavior and the risk of becoming a victim and a bully/victim: a meta-analysis study. Child Abus Negl; 37(12):1091-1108. Liao, J.X., Jin, Y.L. (2017) The prevention of the American school bullying: history, expe- rience and enlightenment. Edu Sci; 33(5):89- 96. Maynard, B.R., Vaughn, M.G., Salaswright, C.P., & Vaughn, S.R. (2016) Bullying victimiza- tion among school-aged immigrant youth in Vol. 1, No. 1, 2019 27 the united states. J Adol Health Offic Pub Soc Adol Med; 58(3):337-344. McNeal Jr, R.B., & Ralph, B. (1999) Parental involvement as social capital: differential ef- fectiveness on science achievement, truancy, and dropping out. Soc Forc; 78(1):117-144. OECD (2016) PISA 2015 Assessment and Ana- lytical Framework: Science, Reading, Math- ematics and Financial Literacy, PISA, OECD Publishing, Paris. OECD (2017) PISA 2015 Results (Volume III): Students’ Well-Being, PISA, OECD Pub- lishing, Paris. OECD (2009) PISA data analysis manual: SPSS, second edition. PISA, OECD Publishing: 72- 74. Olweus, D. (1994) Bullying at school: basic facts and effects of a school based interven- tion program. J Child Psychol Psych, 35(7):1171-1190. Rivara, F., & Le Menestrel, S. (2016) Preventing Bullying Through Science, Policy, and Prac- tice. Washington, DC: The National Acade- mies Press. Smith, P. & Morita, Y., Junger-Tas, J., Olweus, D., Catalano R.F., & Slee, P. (1999) The Na- ture of School Bullying: A Cross National Perspective. London, United Kingdom: Routledge. Troop-Gordon, W., & Kopp, J. (2011) Teacher– child relationship quality and children’s peer victimization and aggressive behavior in late childhood. Soc Dev; 20(3):536-561. Wang, M.C. (2014) Latent variable modeling using Mplus. Chongqing, P.R.China: Chongqing University Press, 77. Wang, W.D. (2016) How much school bullying that should not have occurred? Guangming Daily. Retrieved September 10, 2017 from http://news.gmw.cn/2016- 12/15/content_23265749.htm. Woods, S., & Wolke, D. (2004). Direct and rela- tional bullying among primary school chil- dren and academic achievement. J Sch Psychol; 42(2):135-155. Received: 27 October 2018 Revised: 11 November 2018 Accepted: 25 November 2018 The Chinese version of this article has been published in Education Science 2018; 34(1):7-13. The English version has been authorized for being publication in BECE by the author(s) and the Chi- nese journal. 黄亮, 赵德成. 家庭社会经济文化地位与学生遭受校园欺凌关系的实证研究-家长支持和教师 支持的中介作用. 教育科学 2018; 34(1):7-13. Article-Liang Huang_Title_20Jan2019 Article-Liang Huang_Main text_20Jan2019