Analysis of Influencing Factors and Construction of Nomogram of School Bullying Suffered by Middle School Students in Beijing in 2022

Introduction

School bullying is the persistent and repeated negative behavior of an individual or group toward another individual or group in a school setting, including verbal teasing, name-calling, spreading rumors to isolate others, and physical contact, such as hitting, kicking, and pushing.¹ Bullying at school is becoming increasingly serious among children and adolescents, with around a third (32%) of students worldwide reporting having experienced bullying by their peers,² and a survey of 11 provinces in China showed that around 11% of secondary school students surveyed had experienced bullying,³ according to a Chinese study. 57.29% of middle school students have experienced at least one type of bullying in the past year.⁴ School bullying will have different degrees of negative impact on a student’s physical and mental health, including injuries, bruises, anxiety, depression, sleep difficulties, mental disorders, chronic bullying victimization can also lead to a cumulative decrease in life satisfaction, and even suicidal ideations or behaviors in severe cases.^5–9 The damage is not limited to children, a study shows that children’s bullying victimization is associated with an increased likelihood of suicidal ideation among marriage migrant mothers.¹⁰ Therefore, it is urgent to take effective measures to intervene the influencing factors of school bullying.

There is a series of social and family problems underlying the school bullying problem, and family and personal factors are important predictors of school bullying. Previous studies have found that many factors are associated with school bullying. Boys are more likely to be bullied in school than girls,^11–13 low socioeconomic status is associated with being a victim of bullying,¹⁴ depressive symptoms that mediate between intentional self-harm and bullying victimization.¹⁵ Changes in family structure, parent–child relationships, and peer relationships are also important factors influencing bullying.^16–18 Parents can also help children and adolescents recover from the adversity of bullying.¹⁹ So it is necessary to study the impact of personal and family factors on bullying. This study fills the gap in the data on school bullying among middle school students in Beijing and visualizes the risk of school bullying by using a nomogram model. This study analyzed the influencing factors of school bullying, which showed that inadequate family support, poor diet, depression, and internet use were associated with having been bullied. To identify and intervene in the high-risk group of school bullying early and then reduce the risk of school bullying, a novel nomogram can be used to identify students who need extra support from families, schools, and society.

Methods

Study Population and Design

This study was based on the “Monitoring and Intervention of Common Diseases and Health Influencing Factors among Students in Beijing in 2022” project, which was a cross-sectional survey. From September 2022 to November 2022, a multistage stratified cluster random sampling was used to select research participants among middle school students in 16 districts of Beijing (five urban and four suburban). Cluster sampling was done by class unit, in which each grade must have had at least 80 students; if the number was insufficient, it was supplemented by other nearby schools of the same type.

Data Collection

The health impact factor items of the “2022 Beijing Students’ Common Diseases and Health Impact Factors Monitoring and Intervention Work Project” included basic personal information on smoking, alcohol consumption, diet, internet addiction, family atmosphere (family type, whether they were beaten and scolded by parents in the past 30 days), physical exercise (days of medium-to-high-intensity exercise/week, number of physical education classes/week), sleep status, depression, and school bullying. Before the investigation was conducted, the national ethical review was passed, and informed consent was obtained from the participants. The questionnaire survey was conducted in the school, and the questionnaire response method was filled in by the students. The presence of school teachers was avoided during the whole survey process. In the process of investigation, retesting and verification were organized in a random way at a 5% ratio, and quality control works such as coding requirements, logic verification, data verification, and outlier verification were strengthened in the process of questionnaire investigation. A total of 17750 questionnaires were sent out, 17735 questionnaires were recovered, and 6 questionnaires with incomplete information were deleted. Finally, 17729 middle school students were selected as research objects, with an effective recovery rate of 98.2%.

Definition of Some Concepts

School bullying was defined as “frequent” if one of the following behaviors occurred in the previous 30 days: being teased maliciously, asked for property, intentionally excluded or isolated from group activities, threatened and intimidated, hit, kicked, pushed, jostled or locked in a room, and being teased because of physical deficiencies or looks.²⁰

The Internet addiction scale compiled by Young was used to evaluate students’ Internet addiction.²¹ Internet addiction was defined as internet use ≥4 hours/day in the past 7 days and ≥4 of the following conditions: frequent internet use, including thinking about internet-related things; feeling uncomfortable or unwilling to do other things if you cannot access the internet; increased time spent online for satisfaction; losing interest in other recreational activities because of the internet; inability to stop surfing the internet; unable to complete homework or skip school because of the internet; hiding internet access from parents, teachers, and classmates; continuing to use the internet despite the negative consequences (lack of sleep, being late for class, and arguing with parents); and to escape from reality, banish difficulties, depression, helplessness, and anxious feelings.

Depression was measured using the Center for Epidemiologic Studies Depression Scale, which includes 20 items on depression, positive emotions, and somatic and interpersonal symptoms. The options for each item are “no or occasionally”, “sometimes”, “often or half of the time”, and “most of the time or continuously”, scored on a scale of 0–3. A total score of ≤15 points indicates no depressive emotions, 16–19 points indicate possible depressive emotions, and ≥20 points indicate depression.²²

In the past 7 days, drinking sugary drinks ≥1 times a day, eating fried food ≥1 times a day, eating fresh fruit ≤1 times a day, eating fresh vegetables ≤1 times a day, one of the above behaviors is defined as poor diet.²²

The normal family includes the nuclear family and the extended family. The nuclear family refers to living only with the father and mother. Extended family means living with grandparents, father, and mother. Abnormal families include single-parent families, reorganized families, and intergenerational families. A single-parent family is one living with one’s mother or father; Reorganized family means living with a father and stepmother or a mother and stepfather; A generational family is one living with grandparents without a father/mother/stepparent.²⁰

Data Analysis

SPSS (version 26.0, IBM Corp) and R (version 4.3.3, www.R-project.org/) were used for statistical analysis. The chi-squared test was used for attribute data, and a binary logistic regression model was used to screen risk factors. Statistical significance was considered at P < 0.05. The nomogram prediction model was established, and the AUC was calculated using the rms package; the receiver operating characteristic (ROC) curve was drawn using the ROCR and rms packages, and bootstrapping was verified using the caret package.

Results

Single Factor Analysis of School Bullying in Middle School Students

As shown in Table 1, among the 17,729 middle school students in Beijing in 2022, 477 (2.69%) were victims of school bullying. Single-factor analysis showed that sex, boarding, family type, whether parents beat and scolded them in the past 30 days, smoking status, alcohol use, days of medium-to-high-intensity exercise/week, number of physical education classes/week, poor diet, depression, and internet addiction were all associated with school bullying (P < 0.05).

Table 1 Single-Factor Analysis of School Bullying Among Middle School Students in Beijing [n(%)]

Multivariate Logistic Regression Analysis of School Bullying in Middle School Students

As shown in Table 2, variables with statistical significance in the single-factor analysis were included in the binary logistic regression analysis, which showed that being beaten and scolded by parents in the past 30 days, family abnormalities, poor diet, depression (possibly depressed, depressed), and internet addiction were risk factors for school bullying (odds ratio [OR]=1.61, OR = 1.31, OR = 1.73, OR = 2.16, OR = 6.85, and OR = 1.47, respectively), while female sex, non-boarders, 3–4 days/week of moderate to high-intensity exercise, and 2–3 physical education classes/week were protective factors for bullying (OR = 0.47, OR = 0.76, OR = 0.67, and OR = 0.58, respectively) (P < 0.05).

Table 2 Multivariate Logistic Regression Analysis of School Bullying Among Middle School Students in Beijing

Establishment and Verification of the Nomogram Model

A nomogram was established to visualize the risk of school bullying using variables screened by multivariate logistic regression to predict the risk of having been bullied. First, the value of each risk factor for an individual, corresponding to a single score, was entered in the first line of the column diagram, and then the score of the nine risk factors was added to obtain the total score. Finally, the probability of an individual experiencing school bullying was predicted based on the total score. The higher the total score, the greater the risk of bullying. Most students scored 200–350 points (Figure 1).

Figure 1 Nomogram prediction model of school bullying risk for middle school students.

Figure 1 shows an example of a nomogram prediction model used to predict an individual’s risk of bullying. A female student without internet addiction, not depressed, with a normal family, who does not board, has not been beaten and scolded by her parents in the past 30 days, engages in moderate to high-intensity exercise 5–7 days/week, has ≥4 physical education classes/week, and a poor diet has a total score of 263. The risk of school bullying was 2.16%.

The nomogram’s AUC was 0.801 (95% confidence interval [CI]: 0.782–0.821), indicating good discrimination (Figure 2). The Hosmer–Lemeshow test results (χ²=6.65 and P = 0.57) indicated an acceptable goodness of fit. After 1000 re-samplings using bootstrap internal verification, the corrected AUC was 0.797, and the calibration curve showed good prediction (Figure 3).

Figure 2 The ROC curve of secondary school bullying risk predicted using the nomogram model.

Figure 3 Calibration curve of secondary school bullying risk predicted using the nomogram model.

Discussion

The rate of middle school students being bullied in this study was 2.69%, similar to survey results in Jiangsu Province.²⁰ This study found that males were at a higher risk of bullying than females, which is consistent with previous research, considering that Chinese parents are more protective of girls than boys, and girls spend more time at home, which may be the reason why girls are less likely to be victims of bullying.^23,24 Students engaging in moderate to high-intensity exercise 0–2 days/week are more likely to experience school bullying than students who exercise 3–4 days/week. This may be because physically strong people seem “hard to bully”, suggesting that strengthening physical exercise and physical fitness can help prevent school bullying.^25,26 Students having 0–1 physical education classes/week are more likely to experience school bullying than students who have 2–3 physical education classes/week, suggesting that school policies and measures play an important role in preventing campus bullying.²⁷ In the next step, schools should increase the length of physical education classes, encourage students to go outside, pay attention to students’ psychology, and promote students’ physical and mental health and comprehensive development.

Family factors have an important predictive role on whether one suffers from school bullying, which was higher among students living in families with abnormalities. An incomplete family structure may lead to a lack of close familial relationships, a lack of parental care and guidance, and increased vulnerability to bullying. Students who have been beaten and scolded by their parents in the past 30 days are more likely to suffer from insecurity, mood swings, and personality traits associated with low self-esteem and cowardice, and are more likely to suffer school bullying.²⁸ At the same time, parents’ rejection parenting style may lead students to adopt negative coping styles to deal with problems, resulting in more pain of non-physical bullying at school.^29–31 Students who board have a chronic lack of communication and emotional support with their parents, which may result in a higher risk of school bullying.

Finally, individual-level factors such as internet addiction, poor dietary behaviors, and depressive symptoms were also important predictors of school bullying. Students with internet addiction are usually more solitary, lack the care of parents and peers, place their hopes on the internet world, and lack social skills and social adaptability; such students who escape from reality are more likely to suffer from school bullying.^32,33 Students with depressive symptoms tend to have low emotional performance, resist communication, and lack a sense of collective identity and are more likely to suffer from school bullying.^34,35 Consistent with relevant studies, this study found that poor dietary behaviors are associated with school bullying, possibly because poor eating behaviors are positively correlated with depressive symptoms.³⁶

Previous studies have used Nomogram models to analyze the risk of bullying. This study uses individual and family variables screened by multivariate logistic regression as predictive variables to establish a Nomogram prediction model. It can visualize the risk of school bullying and provide the basis for preventing school bullying in the future. According to the results of this study, at the individual level, middle school students who are male, have less medium to high-intensity exercise and physical education class, have Internet addiction, depressive symptoms, and poor diet are more likely to suffer school bullying. At the family level, middle school students with incomplete family structure, lack of emotional support from parents, and lack of close contact with parents are more likely to suffer school bullying. Parents should pay attention to students’ mental health, provide them with emotional support, oppose violent parenting, and get along with children equally; Schools should offer physical education courses on a regular basis, strengthen publicity and education, set up anti-bullying courses, and provide psychological counseling for students; Students themselves should also establish a “zero tolerance” attitude toward campus bullying, improve their physical fitness and mental toughness, regulate eating behaviors, try more to interact with others, avoid indulging in the online world, and dare to say “no” to campus bullying. Parents, schools, and society should unite, pay attention to students’ mental health, oppose bullying in schools, and provide appropriate support for high-risk students to prevent bullying in schools.

Limitations of This Study

First, it had a cross-sectional design, which limits the inference of causality. Second, moreover, individual-level and family-level heterogeneity not observed in this study may have an impact on the study results, and more comprehensive individual and family factors should be studied to explain their association with school bullying. Finally, the model lacked external verification. Thus, large multicenter longitudinal studies with follow-up are needed for verification in order to provide a more accurate model where diverse groups can be effectively included in the prevention of school bullying.

Conclusion

In conclusion, parents should focus on students’ mental health, provide them with emotional support, oppose violent parenting, and interact with children equally. Schools should offer physical education courses regularly, strengthen awareness and education, initiate anti-bullying courses, and provide psychological counseling for students. Students should establish a “zero tolerance” attitude toward school bullying, improve their physical fitness and mental toughness, and dare to say “no” to school bullying. Parents, schools, and society should unite, focus on students’ mental health, oppose bullying in schools, and provide appropriate support to high-risk students to prevent bullying.