The effect of bullying victimization on internet addiction: mediated by cyberbullying perpetration and moderated by social support

Abstract

Background

Given the importance and high prevalence of Internet addiction (IA), this study aimed to investigate the relationship between bullying victimization and Chinese adolescent students’ IA and its mechanism—mediated by cyberbullying perpetration and moderated by social support.

Methods

Participants completed the Internet Addiction Test (IAT) and provided information on their demographics, experiences with bullying, and social support. General linear modeling (GLM), structural equation modeling (SEM), and interaction item analysis were conducted.

Results

The study included 60,268 middle and high school students with a mean age of 15.9 ± 1.651 years, comprising 45.6% males, 88.7% Han ethnicity, and 82.7% rural residents. Among the participants, 35.1% were classified as having IA, and 14.3% reported experiencing any bullying victimization. We found: (1) Bullying victimization was positively related to adolescent IA, except for physical victims; (2) reduced cyberbullying perpetration positively mediated the relation between social bullying victimization and adolescent IA while negatively mediating the link between physical and cyber victims and IA; (3) the positive moderation roles were observed in the effect of subjective support on the relationship between experiences of being cyberbullied and IA, objective support on the correlation between physical victimization and IA, and utilization of support on the link between social victimization and IA, while objective support negatively moderated the correlation between social victimization and IA.

Conclusions

The study suggests that cyberbullying perpetration and social support are important factors in understanding the impact of bullying victimization on adolescent IA and underscores the need for improving the quality of bullying interventions and social support for at-risk adolescents.

Background

Prevalence and importance of Internet addiction among adolescents

As of mid-2020, 1.079 billion people in China used the Internet, accounting for 76.4% of the country’s entire population [1]. An appropriate level of Internet use may be helpful for business or school-related needs, as well as for keeping in touch with relatives and close friends. However, excessive and unrestrained Internet use can develop into Internet addiction (IA), which is defined as a persistent desire to engage in online activities, which can cause dysfunction or distress [2]. IA has become a major public health issue globally [3], reporting a 20.0–40.3% prevalence rate [4–6], particularly in East Asia [3], with China reporting a rate of 36.7% [7]. Individuals suffering from IA often face negative outcomes in terms of physical health (vision, sleep, obesity, sedentary lifestyle, and musculoskeletal disorders), psychological well-being (depression, anxiety, and loneliness), academic performance, cognitive ability, social interactions, and many other aspects [8–10].

The effect of bullying victimization on IA

Bullying has become a growing public health problem in recent years, defined as repetitive and deliberate harm in an unbalanced relationship of power [11]. Traditionally, direct bullying in schools has included physical aggression (e.g., pushing, hitting, and kicking) and verbal harassment (e.g., threats and verbal abuse), as well as indirect forms of social aggression (e.g., spreading rumors and socially excluding behaviors) [12]. Cyberbullying is defined as deliberate and repeated victimization by individuals or groups using electronic devices such as computers and mobile phones [13]. Global studies have shown a widespread occurrence of bullying victimization, with an average rate of 39.4% in 29 countries [14], and profound negative effects on mental health, such as increased emotional symptoms and reduced well-being [15, 16].

Previous studies suggested bullying victimization might be a key factor contributing to IA among adolescents. A meta-analysis including Chinese adolescents and college students revealed a positive association between adverse life experiences and IA [17]. Additionally, a review that included 14 prospective studies confirmed a robust positive link between cyberbullying victimization and IA [18]. Chinese adolescents who had experienced more peer victimization were shown to be at a higher risk of IA [19]. Being a bullying victim turned out to be another risk factor for IA tendency in Korean adolescents [20]. Research revealed that the unique contribution of these three traditional victimization forms to poorer well-being, with relational and physical victimization, was more harmful than verbal victimization [21]. Therefore, it is necessary to keep abreast of the effects of bullying victimization types on IA among Chinese adolescents.

Behaviorally, IA can result from both positive reinforcement, like the pleasure from Internet use, and negative reinforcement, such as negative emotions without using the Internet and then using it to escape negative emotions [22]. The problem-behavior theory and the self-control model provide the theoretical foundation for the association [23, 24]. According to the problem-behavior theory, bullying victimization disrupts the balance between personality and environment systems, increasing the likelihood of engaging in problematic behaviors such as IA [23]. Besides, based on Baumeister’s psychological resource model, self-control is a limited psychological resource and stressful experiences such as bullying victimization can deplete it, increasing vulnerability to maladaptive behaviors like IA [24]. However, previous studies often focused on only traditional or cyber types and failed to provide a comprehensive understanding of the impacts of bullying victimization on IA, as well as always considering variables as categorical ones, which might lead to information loss and reduced statistical power.

Based on these previous findings, we propose Hypothesis 1: Bullying victimization (relational, physical, verbal, and cyber) could lead to IA among Chinese adolescent students.

Cyberbullying perpetration may lead bullying victims to develop into IA

Current research reported that a higher level of cyberbullying perpetration is associated with a higher level of IA [25]. Subsequently, a review of 14 prospective studies suggested that cyberbullying perpetrators were more prone to IA than normal ones [18]. Engagement in cyberbullying perpetration was typically linked to deficits in personal resources, social skills, and peer relationships [26], factors that were highly associated with IA [27]. According to the problem-behavior theory [23], problematic behaviors often emerge from common underlying risk factors and tend to co-occur.

Some studies found that bullying victims were more likely to commit cyberbullying [28–31]. Based on the frustration-attack hypothesis, frustration induced by physical and cyber attacks may increase an individual’s propensity for aggressive behavior, particularly in the online environment where retaliation is more accessible [32, 33]. In addition, the general aggressive model (GAM) further explains how individual and contextual factors contribute to this behavioral transition, that input enters the assessment and decision-making process through the cognition, emotion, and wakefulness paths, and determines the resulting aggressive behaviors [34].

Thus, combining the above arguments provides reasons for exploring cyberbullying perpetration as a mediator in the link between bullying victimization and IA. However, to the best of our knowledge, there is currently no direct evidence to support the mediating role of cyberbullying perpetration, especially among Chinese adolescents and across different types of victimization.

Therefore, the present study proposes Hypothesis 2: Adolescents who have experienced being bullied are positively correlated with cyberbullying perpetration, which in turn is positively correlated with IA. In other words, cyberbullying perpetration mediates the relationship between bullying victimization and IA among Chinese adolescents.

Social support may decrease the risk of IA among bullying victims

Social support is widely recognized as a protective factor against both IA and bullying victimization among adolescents [35]. Studies have shown that adolescents with IA felt a lack of social support [20, 36], and bullying victims tended to receive less social support than non-victims [37, 38]. While social support is negatively correlated with the co-occurrence of IA and cyberbullying victimization [18, 39, 40], some research suggests that social support might intensify the positive correlation [21, 41]. The different findings of social support can be explained from two perspectives. On one hand, it alleviates stress and enhances the psychological resilience of bullying victims [42]. On the other hand, based on the goal-striving stress theory, discrepancies between expected and actual support may lead to disappointment and increased psychological stress, ultimately contributing to poorer behavior like IA [43–45].

However, existing research primarily focused on different sources of social support, with limited attention to its subjective, objective, and utilizing dimensions. Furthermore, cultural differences may influence its effectiveness. Since there are inconsistent and unclear conclusions on the effect of social support, this study proposes Hypothesis 3: Social support acts as a protective factor against the impact of adolescent experiences of being bullied on Chinese IA. In other words, social support negatively moderates the relationship between bullying victimization and IA among Chinese adolescents.

Given the mentioned research gaps, the study aimed to (1) investigate the impact of bullying victimization (social, verbal, physical, and cyber) on IA; (2) assess the mediating roles of cyberbullying perpetration on the association between bullying victimization and IA; and (3) identify the moderating effects of social support (subjective, objective, and utilizing support) in the correlation between bullying victimization and IA. Multiple models (Fig. 1) were established for this study to test all possible mechanisms of the mediator and moderator.

Fig. 1.

The total pathways of the association between different types of bullying victimization and Internet addiction, mediating pathway by cyberbullying perpetration the moderating role of social support

Methods

Study design and participants

In this cross-sectional study, participants were recruited online using a convenience sampling method from December 14, 2022, to February 28, 2023. A total of 122 middle schools and high schools participated in the survey, ensuring a broad and diverse sample. The following were the criteria for inclusion: (1) Middle and high school students in Sichuan Province, China; (2) the ability to read and understand the questionnaire; and (3) voluntary and willing participation. The following were the exclusion criteria: (1) not being enrolled in schools in Sichuan Province; (2) having difficulty filling out the questionnaire; and (3) unwillingness to take part [46–48].

A total of 65,509 students accessed the survey, and 60,268 (92.0%) completed the questionnaires, while 5,241 (8.0%) declined to participate or did not complete the survey. Data quality was monitored by checking for inconsistencies and incomplete responses to identify and exclude inconsistent or unreliable responses.

Measurements

IA

The Internet Addiction Test (IAT) by Kimberly S. Young, consisting of 20 items, was developed based on the DSM-IV-TR criteria for pathological gambling [49]. Scored on a 5-point Likert scale, from “very rarely” to “very frequently,” the IAT yields a total score between 20 and 100 [49], with a higher score indicating a higher level of IA. The Internet addicts were differentiated from non-Internet addicts, with a score above 50 indicative of IA and one below it indicative of no evidence, which demonstrated excellent reliability (r =.86) and Cronbach’s alpha (α = 0.90) in previous studies [50]. Initial investigations into the validity of the original IAT have demonstrated strong internal consistency (α = 0.90-0.93) and good test-retest reliability (r =.85) [51]. The IAT’s validity and reliability were also confirmed among Chinese adolescents, with Cronbach’s alpha reaching 0.93 [52]. Besides, due to the simplicity of its administration, interpretation, and application, we selected the IAT for assessing IA.

Traditional bullying victimization

We used a 6-item checklist from the Programme for International Student Assessment (PISA) to assess traditional bullying victimization [53], which covers social, verbal, and physical victimization. This scale uses Likert 4-point scoring such that 0 means “never or rarely,” 1 means “a few times a year,” 2 means “a few times a month,” and 3 means “once a week or more.” Respondents who score at least 2 (a few times a month) in any category indicate that they have experienced that type of bullying victimization [53, 54]. In PISA, conducted across 34 countries, the checklist demonstrated strong reliability with a Cronbach’s alpha of 0.88 [55]. In PISA across four regions of China (Beijing, Shanghai, Jiangsu, and Guangdong), the checklist showed a Cronbach’s alpha of 0.84 [21].

Cyberbullying victimization and perpetration

The Chinese Cyberbullying Intervention Project Questionnaire (C-CIPQ) is modified for cultural relevance and application [56]. This 14-item test uses a 5-point Likert scale to evaluate cyberbullying, with 7 items measuring victimization and the other 7 measuring perpetration. Cyber victims and cyber-aggressors have been identified with scores equal to or higher than 2 (a few times a month) in any of the items of the sub-scale [57]. In the original European version, Cronbach’s alpha values for bullying victimization and perpetration were 0.77 and 0.71, respectively [58]. In the Chinese context, the C-CIPQ has been validated among Chinese teenage samples, with Cronbach’s alpha values of 0.77 for victimization and 0.78 for perpetration [59].

Social support

The Social Support Rating Scale (SSRS) consists of 10 questions divided into three domains [60]. Objective support measures the real assistance someone may receive from his or her social network. Subjective support relates to a person’s opinion of whatever aid is available to them, and support utilization reflects the actions that a person might take to obtain help. On the basis of an overall score range of 50%, social support scores are classified as low or high [61]. This methodology has established the scale’s substantial validity and reliability within the Chinese population, achieving a Cronbach’s alpha of 0.78 among Chinese adolescents [62].

Other variables

Besides, the study used self-made structured questions to collect demographic information (gender, age, ethnicity, and residence) (Table 1).

Table 1.

IAT scores in all participants by demographic characteristics

	N (%)	Mean ± SD	P value
Overall	60,268 (10.0)	44.8 ± 17.065
Age (year)
12–15	22,129 (36.7)	41.6 ± 16.627	< 0.001 ***
16–19	38,139 (63.3)	46.6 ± 17.058
Gender
Male	27,459 (45.6)	45.9 ± 28.228	< 0.001 ***
Female	32,809 (54.4)	43.8 ± 15.964
Ethnicity
Han	53,459 (88.7)	44.7 ± 17.135	0.160
Others	6,809 (11.3)	45.0 ± 16.509
Registered residence
Urban	10,449 (17.3)	44.1 ± 17.010	0.001 **
Rural	49,819 (82.7)	44.9 ± 17.074
Education Level
Middle school	24,157 (4.1)	41.8 ± 16.705	< 0.001 ***
High school	36,111 (59.9)	46.7 ± 17.024
Social bullying victimization
No	55,027 (91.3)	43.7 ± 16.397	< 0.001 ***
Yes	5,241 (8.7)	55.8 ± 19.780
Verbal bullying victimization
No	55,207 (91.6)	43.7 ± 16.416	< 0.001 ***
Yes	5,061 (8.4)	56.1 ± 19.674
Physical bullying victimization
No	56,914 (94.4)	44.0 ± 16.486	< 0.001 ***
Yes	3,354 (5.6)	58.4 ± 2.619
Cyberbullying victimization
No	55,319 (91.8)	43.6 ± 16.373	< 0.001 ***
Yes	4,949 (8.2)	57.5 ± 19.290
Cyberbullying perpetration
No	57,580 (95.5)	44.0 ± 16.504	< 0.001 ***
Yes	2,688 (4.5)	61.1 ± 2.378
Subjective support
Low	12,147 (2.2)	48.7 ± 19.145	< 0.001 ***
High	48,121 (79.8)	43.7 ± 16.346
Objective support
Low	53,259 (88.4)	45.1 ± 17.132	< 0.001 ***
High	7,009 (11.6)	41.9 ± 16.284
Utilization of support
Low	31,601 (52.4)	48.1 ± 17.408	< 0.001 ***
High	28,667 (47.6)	41.0 ± 15.869

SD Standard deviations. *: P value < 0.05; **: P value < 0.01; ***: P value < 0.001

Procedure

The study was submitted to and approved by the Ethics Committee of West China Hospital of Sichuan University (Ethical Approval Number: 2022–1790). According to the Helsinki Declaration, the study followed all the necessary guidelines. The survey was conducted using the Wenjuanxing platform, a widely used online survey tool in China, providing functions equivalent to Amazon Mechanical Turk [63]. To ensure data quality, the survey was initially sent to teachers in middle and high schools across Sichuan Province. The teachers then were asked to disseminate the survey to all students in their classes, minimizing selection bias. Teachers were available to assist with explaining the questionnaire. Participants were informed that their responses would be anonymous, confidential, and used solely for research purposes. No monetary or material incentives were provided to avoid bias or coercion. Students’ voluntary involvement is ensured by parents as their legal guardians and teachers as their guardians in the school. Informed consent was obtained from all participants themselves, their parents or legal guardians, and their teachers as guardians in school.

Statistical analysis

To describe the demographics, distribution, and prevalence, these characteristics were specified as categorical variables using frequencies and percentages, and the IAT scores by these characteristics were using means and standard deviations (SD). Independent samples t-tests and analysis of variance (ANOVA) were conducted to examine whether IAT score differences existed in these groups. Besides, in the next analysis, bullying victimization, IAT, and the mediating and moderating variables were considered continuously to avoid losing information and exaggerating statistical power, adjusting for age, gender, ethnicity, and residence. To address missing values in our dataset, we specifically used the Multiple Imputation by Chained Equations (MICE) technique. All statistical analyses were performed using R 4.3.3, and all two-tailed tests had a predefined significance threshold of 0.05.

To evaluate the mediating effect of cyberbullying perpetration and the moderating effect of social support, we first utilized general linear modeling (GLM) to determine the association between each score of four types of bullying victimization, cyberbullying perpetration, three dimensions of social support, and IA, adjusting for demographics (Fig. 1, the total and mediating pathways). By analyzing regressions, we calculated the beta (β) values of the bullying victimization scores per 1 standard deviation increase in each of the IAT scores and their 95% confidence intervals (95% CIs).

Then, cyberbullying perpetration and all dimensions of social support were entered into the initial model as mediators and moderators. However, variance inflation factors (VIFs), which were calculated to identify multicollinearity, exceeded acceptable limits (VIF > 5) [64], with unacceptable VIF values ranging from 5 to 9. To address this, and because our main purpose did not include the mutual or combined effect of cyberbullying perpetration and social support, the mediating and moderating roles were explored separately.

Structural equation modeling (SEM) and the Sobel Z test were utilized to examine the multiple mediators between each bullying victimization score and IAT score, adjusting for demographic variables. The output of mediation analysis was presented in terms of total effect, direct effect, and indirect effect. The direct effect denotes the influence of bullying victimization on IA while maintaining the mediator as a constant. The indirect effects, on the other hand, represent the influences of bullying victimization on IA transmission through the mediators [65]. We utilized the “lavaan” package in R, the most common package to conduct SEM [66]. We also calculated the percentage of each mediation as the ratio of the indirect association to the total association.

In examining the moderating effect of social support on the relationship between bullying victimization and IA, the study adopted the original variables and variables mean-centered, which can reduce multicollinearity and improve the interpretability of the main effects [67]. Mean-centering makes the main effect coefficients easier to interpret and more intuitive in that they represent the effect of the moderator on the outcome when exposure is at its mean and reduces confounding between interaction and main effects (by decreasing multicollinearity, leading to more stable estimates of interaction effects).

Results

The study included 60,268 Chinese middle and high school students (63.3% aged 16–19 years, 45.6% males, 88.7% Han ethnicity, and 82.7% residents in rural areas). The total sample had an average IAT score of 44.8, and with the cutoff that individuals who scored 50 points or more were classified as having IA [50], 21,135 (35.1%) of our total sample experienced IA.

Bullying victimization was also not rare, with any victimization (14.3%), social bullying (8.7%), verbal bullying (8.4%), and cyberbullying (8.2%) sharing almost near prevalence, and physical bullying (5.6%) being relatively the lowest. Furthermore, 4.5% of students admitted to committing cyberbullying. Regarding social support, a majority of students (79.8%) felt they had high subjective support, but objective support was notably lower, with up to 88.4% experiencing low levels. Support utilization was also relatively lower, with only 47.6% reporting having high levels. The demographic characteristics and IAT scores of the groups of participants are shown in Table 1.

Effects of different types of bullying victimization on IA

Table 2 details all of the outcomes of multiple linear regression analyses, including the relationship between four types of bullying victimization—social, verbal, physical, and cyber—and IA, as measured by the IAT score. The results showed that higher levels of bullying victimization were associated with higher IAT scores, indicating increased IA problems across participants, while physical victims showed none of significance. As for different victims, verbal bullying was identified as the most significant factor (β = 1.04), followed by social (β = 0.82) and cyberbullies (β = 0.76).

Table 2.

The significance of each exposure-outcome, exposure-mediator/moderator, and mediator/moderator-outcome path by using general linear regression in all total

	IAT score	Cyberbullying perpetration	Subjective support	Objective support	Utilization of support
	β (95% CI) a	β (95% CI) a	β (95% CI) a	β (95% CI) a	β (95% CI) a
Social bullying victimization	0.82 (0.62, 1.01) ***	− 0.35 (−0.36, − 0.33) ***	− 0.51 (−0.56, − 0.45) ***	− 0.06 (−0.1, − 0.02) **	− 0.23 (−0.26, − 0.21) ***
Verbal bullying victimization	1.04 (0.82, 1.27) ***	− 0.04 (−0.06, − 0.02) ***	− 0.33 (−0.39, − 0.26) ***	− 0.07 (−0.12, − 0.02) **	− 0.14 (−0.17, − 0.10) ***
Physical bullying victimization	0.19 (−0.05, 0.42)	0.41 (0.39, 0.43) ***	0.11 (0.04, 0.18) **	− 0.14 (−0.19, − 0.10) ***	0.05 (0.02, 0.09) **
Cyberbullying victimization	0.76 (0.67, 0.86) ***	0.85 (0.85, 0.86) ***	− 0.04 (−0.05, − 0.02) ***	− 0.08 (−0.09, − 0.07) ***	− 0.01 (−0.01, 0.00)
Cyberbullying perpetration	− 0.14 (−0.23, − 0.05) **	/	/	/	/
Subjective support	− 0.19 (−0.22, − 0.16) ***	/	/	/	/
Objective support	− 0.03 (−0.07, 0.01)	/	/	/	/
Utilization of support	−1.22 (−1.29, −1.16) ***	/	/	/	/

^a Adjusted for age, residence, ethnicity, and gender (only in the total sample)

*: P value < 0.05; **: P value < 0.01; ***: P value < 0.001

Association among cyberbullying perpetration, bullying victimization, and IA

Of the total sample, there was a significant negative association between cyberbullying and IA (β = − 0.14). On the other hand, there was a significant negative association between social (β = − 0.35) and verbal bullying victimization (β = − 0.04), as well as cyberbullying perpetration, while physical (β = 0.41) and cyber victimization (β = 0.85) were significantly positively correlated with cyberbullying perpetration.

Correlation among social support, bullying victimization, and IA

As for the association between social support and IA, there was a significantly negative link between IA and subjective (β = − 0.19) and utilizing support (β = −1.22), while the effect of objective support was not significant. As for social support and bullying victims, except for the insignificant link between cyberbullying victimization and utilization of support and the positive correlation between physical bullying victimization and subjective and utilizing support, the other types of victimization and social support were independently negatively associated.

Mediation of cyberbullying perpetration in the association between bullying victimization and IA

Table 3 illustrates the total, direct, and indirect associations between bullying victimization and the IAT score (Fig. 1), along with the mediated proportions after controlling for covariates. The total and direct association between each bullying victimization and the IAT score was statistically significant, except for the direction association of physical victims and IA.

Table 3.

The mediating effect of cyberbullying perpetration in the correlation between types of bullying victimization and internet addiction

	Total association	Direct association	Indirect association		Sobel Z
	β (95% CI) a	β (95% CI) a	β (95% CI) a	% b
Social bullying victimization	1.25 (1.05, 1.45) ***	1.18 (0.98, 1.38) ***	0.07 (0.03, 0.11) ***	5.5	−3.55 ***
Verbal bullying victimization	1.28 (1.05, 1.51) ***	1.27 (1.04, 1.51) ***	0.01 (0.00, 0.02) *	0.6	− 0.70
Physical bullying victimization	0.05 (−0.22, 0.31)	0.13 (−0.14, 0.39)	− 0.08 (−0.13, − 0.03) ***	NA	−2.74 **
Cyberbullying victimization	0.66 (0.59, 0.73) ***	0.83 (0.72, 0.94) ***	− 0.17 (−0.26, − 0.08) ***	−25.7	−3.40 ***

^a Adjusted for age, gender, ethnicity, and residence

^b The mediated association was computed as the ratio of the indirect association to the total association

NA Not applicable

*: P value < 0.05; **: P value < 0.01; ***: P value < 0.001

The analysis revealed that the most significant indirect association with cyberbullying perpetration was between cyber victims and IAT scores, accounting for − 25.7%, followed by social victimization, accounting for 5.5%. In verbal victims, the mediation analysis showed no significance (Sobel Z = − 0.70, p >.05). Cyberbullying perpetration had a mediation effect of more than 100% on the impact of physical victimization on IA, which means an inhibition effect on the original item.

Moderation of social support in the association between bullying victimization and IA

As shown in Table 4, subjective support only positively moderated the relationship between experiences of being cyberbullied and IA, while objective support moderated the correlation between social victimization and IA negatively and the effect of physical victimization positively. Besides, the utilization of support only positively moderated the link between social victimization and IA.

Table 4.

The moderating effect of each dimension of social support in the correlation between bullying victimization and internet addiction

	Raw	Mean-centering a
	β (95% CI) b	β (95% CI) b
Social bullying victimization	1.27 (0.40, 2.14) **	0.85 (0.64, 1.06) ***
Verbal bullying victimization	1.36 (0.31, 2.41) *	1.01 (0.77, 1.25) ***
Physical bullying victimization	− 0.86 (−1.87, 0.15)	0.23 (−0.02, 0.49)
Cyberbullying victimization	0.23 (0.02, 0.44) *	0.72 (0.66, 0.78) ***
Subjective support	− 0.19 (−0.22, − 0.15) ***	− 0.19 (−0.22, − 0.16) ***
Social bullying victimization × Subjective support	− 0.03 (−0.08, 0.02)	− 0.03 (−0.08, 0.02)
Verbal bullying victimization × Subjective support	− 0.01 (−0.06, 0.05)	− 0.01 (−0.06, 0.05)
Physical bullying victimization × Subjective support	− 0.03 (−0.08, 0.03)	− 0.03 (−0.08, 0.03)
Cyberbullying victimization × Subjective support	0.02 (0.01, 0.03) ***	0.02 (0.01, 0.03) ***
Objective support	− 0.04 (−0.08, 0.01)	− 0.04 (−0.08, 0.01)
Social bullying victimization × Objective support	− 0.08 (−0.15, − 0.01) *	− 0.08 (−0.15, − 0.01) *
Verbal bullying victimization × Objective support	0.00 (−0.08, 0.08)	0.00 (−0.08, 0.08)
Physical bullying victimization × Objective support	0.15 (0.07, 0.22) ***	0.15 (0.07, 0.22) ***
Cyberbullying victimization × Objective support	0.00 (−0.02, 0.01)	0.00 (−0.02, 0.01)
Utilization of support	−1.31 (−1.38, −1.24) ***	−1.22 (−1.29, −1.16) ***
Social bullying victimization × Utilization of support	0.11 (0.01, 0.21) *	0.11 (0.01, 0.21) *
Verbal bullying victimization × Utilization of support	− 0.03 (−0.15, 0.09)	− 0.03 (−0.15, 0.09)
Physical bullying victimization × Utilization of support	0.05 (−0.07, 0.16)	0.05 (−0.07, 0.16)
Cyberbullying victimization × Utilization of support	0.02 (0.00, 0.05)	0.02 (0.00, 0.05)

^a Calculated on the basis of the mean centering of the exposure and mediator and are interpreted as a standard deviation change in the outcome variable as the predictor variable changes from 0 to 1

^b Adjusted for age, gender, ethnicity, and residence

*: P value < 0.05; **: P value < 0.01; ***: P value < 0.001

Discussion

The current study showed that the effect of bullying victimization on IA was partly mediated by cyber perpetration and moderated by social support. The most significant mediating role of cyber perpetration was observed in cyberbullying victims and IA, accounting for − 25.7%. Besides, the most positively moderating role of social support was observed in the relationship between physical victims, objective support, and IA, while the only significantly negative relationship was in social bullying, objective support, and IA.

The present study found that 35.1% of adolescent students in Sichuan Province, China, were at risk of IA, aligning with other studies in Chinese (36.7%) [7] and international adolescents (40.3%) [5]. This indicates that the high frequency of IA may be a similar issue across racial and geographic groups of adolescents. The prevalence of any victimization was 14.3%, consistent with 9.4% in Korea, 13.3% in Chinese Taipei [53], and 16.1% in Japan [68]. Our study revealed that 8.7% of students were classified as victims of social bullying, 8.4% of verbal bullying, 8.2% of cyberbullying, and 5.6% of physical bullying, which had an emphasis on a more severe situation of indirect bullying [69]. We also showed that 4.5% of students reported being perpetrators of cyberbullying, which is consistent with the Iran study (7.0%). These findings have emphasized the global challenge of IA and bullying and called for international action to reduce these issues.

Regarding social support, a majority of students (79.8%) felt they had high subjective support, while a majority of students experienced low levels of objective support (88.4%) and support utilization (52.4%), which may be because students lacked actual social support but could feel sufficient support through psychological adjusting and did not know how to use support resources effectively.

Effects of different types of bullying victimization on IA

Partly aligning with our Hypothesis 1, the study showed that various forms of indirect bullying victimization had significant impacts on the IA of adolescents, as verbal bullying was the most significant factor in IA, whereas physical bullying showed no significance. A possible explanation is that physical bullying may be more easily detected and prompts timely intervention by teachers and parents. In contrast, indirect bullying—characterized by behaviors that are subtle, persistent, and difficult to escape—may cause prolonged psychological distress, making victims more vulnerable to maladaptive coping mechanisms such as IA.

From a cultural perspective, Chinese society emphasizes collectivism, social harmony, and conformity, which may shape adolescents’ coping strategies when facing bullying victimization. Rather than seeking confrontation or external support, victims may resort to avoidance, cognitive restructuring, or withdrawal [70, 71]. Excessive Internet use serves as a way of escaping from negative emotions and social distress [25]. Consequently, in the context of China, instead of addressing the bullying directly, victims may increasingly turn to online environments, reinforcing IA tendencies.

Additionally, the self-control model highlights that an adolescent’s reservoir of self-control is a pivotal psychological resource [24]. Under normal conditions, individuals can replenish self-control over time. However, exposure to severe environmental stressors—such as persistent indirect bullying victimization—may lead to chronic depletion, making it difficult to recover. This weakened self-regulatory capacity increases susceptibility to compulsive behaviors, including IA.

Furthermore, according to the problem-behavior theory, bullying victimization disrupts the dynamic interplay between the personality and environment systems [23], leading to heightened social isolation [72] and shame [73]. Thus, victims of indirect bullying may feel an intensified sense of exclusion, further driving them toward Internet use as a refuge. Unfortunately, this withdrawal reinforces the risk of IA. Overall, our findings suggest that indirect bullying victimization, particularly verbal bullying, plays a crucial role in the onset and persistence of IA among adolescents. Future IA interventions should prioritize early detection of indirect bullying and promote adaptive coping strategies that mitigate its long-term consequences.

The mediating role of cyberbullying perpetration on the relationship between bullying victimization and IA

Contrary to our Hypothesis 2 and previous research indicating a positive correlation [74], the results of this study showed that there was a significant negative association between cyberbullying perpetration and IA. The possible reason may be that engaging in cyberbullying may occupy adolescents’ online time, potentially reducing engagement in other online activities that contribute to IA. Moreover, in the virtual world, adolescents can enhance their sense of power and self-efficacy through engagement in cyberbullying perpetration, which could increase their perceived control and reduce susceptibility to IA [75]. It indicated that the internal mechanism of cyberbullying perpetration, such as moral disengagement and Internet literacy [74], reduced the network behavior model of addiction. These factors may inhibit the development of IA.

On one hand, being physically or cyberbullied was positively correlated with cyberbullying perpetration, supported by the continuous and multi-dimensional nature of bullying. According to the frustration-attack hypothesis, an individual’s frustration caused by others’ physical and cyber attacks can lead to readiness for attack behavior [32]. In addition, the behavioral pattern of individuals who experienced physical or cyberbullying may be more prone to expressing aggressive behaviors and can be explained by the general aggressive model (GAM): the cognitive and emotional pathways allow both individual and environmental inputs to enter the appraisal and decision-making process, which in turn determines the aggressive actions [34]. At the same time, the Internet platform allowed for a more effortless shift from victim to attacker, requiring only online revenge in the virtual world instead of physical force in the real world [33]. This suggests a potential chain reaction, where physical and cyber victimization leads to increased cyberbullying perpetration, ultimately reducing IA. Thus, the results revealed that the most significant indirect association by cyberbullying perpetration was between cyber victims and IAT score, accounting for − 25.7%, while cyberbullying perpetration might have a mediation effect of more than 100% on the impact of physical bullying victimization on IA, which meant an inhibition effect of cyberbullying perpetration on the positive impact of being physical and cyber victims on IA.

On the other hand, social victims were significantly negatively correlated with cyberbullying perpetration. Victims of social bullying were more likely to internalize their negative emotions rather than display aggressive behaviors [76]. Victims of social bullying often experience significant emotional distress, such as sadness, anxiety, and low self-esteem [77]. As a result, they may cope with bullying through avoidance or passive responses rather than engaging in cyberbullying. The result of the study showed that decreased cyberbullying perpetration played a partly positive mediating role in the transition from social victimization to IA. Future research should further explore the protective role of reduced cyberbullying perpetration in the relationship between social and verbal victimization and IA.

The moderating role of social support on the correlation between bullying victimization and IA

Our study further clarified the moderating role of different dimensions of social support in the effect of bullying victimization on IA. While previous studies have generally identified social support as a protective factor against IA [18, 39, 40], our findings revealed a more nuanced pattern of different dimensions. Specifically, subjective support positively moderated the relationship between cyberbullying victimization and IA, objective support positively moderated the link between physical victims and IA, and utilization of support positively moderated the effect of social victims on IA. Though partially aligning with prior research [21, 41], these findings were reluctant with Hypothesis 3 that all dimensions of social support uniformly mitigate the risk of IA among bullying victims.

The observed positive moderation effects can be interpreted through the goal-striving stress theory [43]. Adolescents who perceive strong support may develop high expectations for assistance, and when these expectations are unmet, they may experience disappointment and mistrust, leading to increased psychological stress and subsequently a higher likelihood of IA [44, 45]. In some cases, their social networks may lack the necessary resources or skills to effectively intervene in bullying situations, exacerbating feelings of helplessness and reinforcing maladaptive coping strategies such as IA.

Conversely, objective support negatively moderated the relationship between social victims and IA, supporting the notion that tangible support—such as material resources and structured guidance—helps adolescents manage stress or emotional problems more effectively, enhancing psychological resilience and offering emotional and physical resources to enable them to arrange their behavior when facing the enticements of the Internet [42]. Adolescents with sufficient objective support are better equipped to cope with these stresses and problems posed by bullying, reducing their reliance on the Internet as a coping mechanism [78].

The results of this study suggest that the role of social support is complex and culturally influenced. In the context of China, subjective and utilization-based support may sometimes function as risk factors rather than protective factors, whereas objective support remains beneficial. Future studies should further explore the cultural and contextual factors that shape the effectiveness of different support dimensions in mitigating IA risk among adolescent victims.

Strengths, limitations, and future directions

This study had some strengths. First, this study comprised a diverse adolescent sample across a wide array of ages and schools, enhancing the generalizability of its findings. Second, to minimize the confounding effects, potential confounders such as demographics and family background were progressively controlled in the analyses. Third, this study sheds new light on the mediating and moderating roles between various types of bullying victimization and IA, which might help to identify high-risk populations of IA and develop psychosocial interventions for improving the mental health of these students.

This study also had several limitations. First, this study was conducted only in Sichuan Province, China, which may constrain the generalizability of our findings to adolescents in other regions of China or globally. Thus, we have taken care to present our findings within the specific cultural and socioeconomic context of Sichuan. As a populous and economically vital province, Sichuan can provide a distinct perspective through which to examine the interplay of IA and bullying behaviors, which may not be fully captured in national or international studies, thus enriching the global body of knowledge. Second, as this study was cross-sectional, no causal inference can be drawn from it. Further longitudinal follow-up research should be carried out. Third, this convenience sampling survey was conducted online, and all items were self-reported, which might cause bias. To alleviate these possible biases, the study sought out a diverse participant composition and adopted tried-and-true tools. Fourth, cyberbullying was not further categorized as cyberattacks, cyberstalking, or sex-based cyberbullying [33]. However, this study nevertheless contributed substantially to our understanding of the complex relationships between bullying behaviors and IA. Lastly, while the IAT is widely used and validated [79], its diagnostic accuracy and cross-cultural applicability have been questioned [80]; caution should be exercised with respect to using the IAT for diagnosis.

Implications for practice

The findings of this study provided valuable insights for both research and practical interventions aimed at mitigating adolescent IA. Since bullying victimization effects vary in IA, anti-bullying programs and IA prevention in schools should focus on different types of bullying. The mediating role of cyberbullying perpetration suggested that physical and cyberbullying may reduce engagement in cyberbullying themselves, thereby increasing their risk of IA. Schools and policymakers should focus on digital literacy and online behavioral interventions to break this chain and mitigate IA risk. Social support played a crucial moderating role, either positively or negatively affecting the relationship between bullying victimization and IA. It suggests that intervention strategies should carefully assess how external support is provided to ensure its effectiveness and encourage adolescents to actively seek and use social support resources, which can be beneficial.

Conclusions

The study suggests that cyberbullying perpetration and social support are important factors in understanding the impact of bullying victimization on adolescent IA. Schools should implement comprehensive anti-bullying policies that address both traditional and cyberbullying. Parental guidance and community-based interventions should be strengthened to enhance adolescents’ access to and utilization of effective social support systems. Future studies should explore longitudinal designs to establish causal relationships and investigate additional protective factors that may buffer the impact of bullying victimization on adolescent IA.

Abbreviations

IA

Internet addiction

IAT

Internet Addiction Test

GLM

General Linear Modeling

SEM

Structural Equation Modeling

GAM

General Aggressive Model

PISA

Programme for International Student Assessment

C-CIPQ

Chinese Cyberbullying Intervention Project Questionnaire

SSRS

Social Support Rating Scale

SD

Standard Deviation

ANOVA

Analysis of Variance

MICE

Multiple Imputation by Chained Equation

CI

Confidence interval

VIF

Variance Inflation Factors

Authors' contributions

MSR designed this study. MSR, MLL, CW, JC, YFM, ZYD, YW, APD, HJS, LZ, JZ, YH, LY, TTJ, and WWS conducted this study. YX, MMZ, and MLL conducted data analysis. YX, MMZ, and MLL wrote the first draft of the paper. All authors participated in the data collection and made contributions to the critical revision of the manuscript.

Funding

This study was supported by the Initial Research Fund, West China Hospital (WCH, No. 136220012, PI: Prof. Ran), and the National Science and Technology Innovation 2030 “Brain Science and Brain Research” major project of China (2021ZD0202100, 2021ZD0202102). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Data availability

The de-identified data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author, MSR, via email at msrancd@outlook.com upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was submitted to and approved by the Ethics Committee of West China Hospital of Sichuan University (Ethical Approval Number: 2022–1790). All procedures followed were by the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2013. Informed consent was obtained from all participants, their parents or legal guardians, and teachers as their guardians in the school.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.