Parental Psychological Control and Bullying Victimization in Chinese Adolescents: Roles of Deviant Peer Affiliation and Self-Compassion

1 Introduction

Bullying victimization (BV) is characterized by repeated, intentional peer aggression within a power-imbalanced relationship, causing psychological or physical harm to victims unable to defend themselves [1]. Recent meta-analytic evidence shows that one in four adolescents worldwide are affected by BV [2]. BV not only leads to psychopathological symptoms (e.g., depression, anxiety) [3,4,5], but also may alter victims’ social cognition, leading to maladaptive coping strategies such as aggressive cognitive styles (increased hostility) [6]. Specifically, this can perpetuate a cycle of victim-to-perpetrator transition, in which a victim may later bully others [7]. Research investigating modifiable risk factors and fundamental processes of BV is therefore essential for informing the design of tailored intervention approaches.

Bullying is widely conceptualized as an ecological phenomenon embedded within adolescents’ everyday social contexts, rather than an isolated individual experience [8]. Ecological systems theory [9] emphasizes that development unfolds through dynamic interactions between individuals and their nested environments, particularly within proximal microsystems such as the family, peer group, and school. Extending this perspective, the Social–Ecological Diathesis–Stress model proposes that environmental stressors interact with individual vulnerabilities to shape adolescents’ adjustment within peer contexts, including experiences of bullying victimization [8]. Together, these frameworks suggest that family processes may influence bullying risk indirectly by shaping adolescents’ peer environments, and that individual characteristics may condition adolescents’ susceptibility to these contextual influences. Within this framework, negative parenting practices such as parental psychological control (PPC) represent a distal microsystemic factor that shapes adolescents’ relational schemas and emotion regulation capacities, which in turn influence peer interactions [10,11]. Deviant peer affiliation (DPA), which is defined as selective association with peers who engage in problem or delinquent behaviours [12], has been identified as one such peer-context mechanism.

Empirical studies consistently indicate that adolescents exposed to dysfunctional or coercive parenting environments are more likely to affiliate with deviant peers [13,14]. From a social–ecological diathesis–stress perspective, psychologically controlling parenting may heighten adolescents’ cognitive and emotional vulnerabilities, including diminished autonomy, poor emotion regulation, and maladaptive interpersonal schemas, which in turn shape their peer selection processes [15,16]. Hence, DPA constitutes a critical peer-level mechanism through which family risk may be translated into victimization experiences. Although prior research suggests that the association between DPA and bullying dynamics may be bidirectional—serving both as a precursor to and a consequence of victimization [11,17,18]. The present study conceptualizes DPA primarily as a peer-context mechanism through which family-level risk is associated with bullying victimization. This conceptualization is consistent with ecological systems theory, which emphasizes the cascading influence of family processes on adolescents’ peer environments.

Importantly, not all adolescents exposed to controlling parenting engage with deviant peers, suggesting the presence of intrapersonal resilience factors. Self-compassion (SC) may serve as such a factor by modulating adolescents’ cognitive and affective responses to coercive parenting [12,19]. Specifically, SC promotes adaptive self-regulation, mindful awareness of personal distress, and balanced interpersonal engagement [20,21]. Prior research has largely examined SC as a buffer against psychological distress, demonstrating its protective role in reducing depression, anxiety, and stress and in moderating associations between stressors and maladjustment [22,23]. However, less attention has been paid to the role of SC in adolescents’ social-contextual processes, particularly in shaping peer selection under conditions of family adversity.

Drawing on the Social–Ecological Diathesis–Stress model, the present study examines SC as a moderator of peer-context formation rather than solely as a buffer of adjustment outcomes. Adolescents with higher SC may interpret coercive parental behaviour less self-critically and maintain more autonomous coping strategies, which reduces the likelihood of affiliating with deviant peers despite experiencing high PPC. In contrast, adolescents lower in SC may be more vulnerable to internalizing parental rejection and may gravitate toward deviant peers as an alternative source of acceptance and identity [11,24]. Thus, rather than exerting a direct effect, SC is theorized to buffer the transmission of family risk through peer contexts.

Taken together, the present study advances a process-oriented ecological model that integrates family-level risk, peer-context mechanisms, and individual resilience within a single conditional pathway. Specifically, we test whether (a) DPA mediates the association between PPC and BV, and (b) SC moderates the pathway linking PPC to DPA. By examining these associations within an ecological framework, this study aims to clarify the conditional processes through which family risk may be associated with adolescent bullying victimization.

1.1 Parental Psychological Control and Bullying Victimization

Ecological system theory positions the family as a foundational microsystem that actively scaffolds children’s development and configures their future social-relational capacities [9]. Early interactions with parents are pivotal to adolescents’ socialization, as they rely on their parents as critical sources of self-identity and self-evaluation [25]. Therefore, adolescents’ adverse social interactions in school settings, such as victimization, are often rooted in the negative relationship dynamics they encounter in their early family life [26]. Previous studies have identified parenting as a predictor influencing the occurrence of bullying victimization through the ecological system theory framework [27,28]. Emerging as a culturally entrenched disciplinary practice in China [29], parental psychological control (PPC) manifests as a parenting approach that employs coercion and manipulation of rewards and punishment to regulate offspring conduct, cognitive schemas, and affective regulation patterns. This approach often employs guilt induction, shaming, and love withdrawal to enforce compliance, ultimately undermining autonomy and self-regulatory development [30]. Self-determination theory posits that PPC hinders adolescents’ autonomy, diminishes their self-esteem, and impedes their ability to establish a sense of control in peer relationships. These factors, in turn, make them susceptible to being dominated and manipulated [31]. When interacting with peers, children tend to mimic the interpersonal style they see in their families. Children who are exposed to family environments that are characterized by negative and excessively controlling parenting may develop acquiescent behavioural schemas. This is because they may internalize maladaptive relational schemas and emotion regulation patterns from familial interactions and then exhibit similar behaviours in peer interactions [32]. Within the developmental psychopathology framework, this trajectory significantly increases vulnerability to peer victimization among PPC-exposed individuals [26,33]. We therefore hypothesize:

Hypothesis 1: PPC is positively associated with adolescents’ BV.

1.2 Mediating Role of Deviant Peer Affiliation (DPA)

Empirical evidence confirms that dysfunctional family contexts predict heightened deviant peer affiliation (DPA) [34,35]. Furthermore, the Social-Ecological Diathesis-Stress model posits that psychosocial adversity amplifies neurocognitive vulnerabilities, resulting in more severe adverse outcomes, such as bullying victimization [8]. According to the theoretical framework of this model, PPC may trigger cognitive vulnerabilities in adolescents, which may diminish their psychological autonomy and lead to lower levels of volitional functioning and independent decision-making. Consequently, adolescents may develop a more defiant and maladaptive approach to peer relationships [15,16]. Specifically, they might be drawn to associations with peers who exhibit delinquent behaviours. This preference stems from a desire to cope with parental and peer rejection and to seek a sense of belonging that they may not find in their family or among their peers [36,37,38].

DPA refers to the behaviour of being selectively affiliated with peers who exhibit more serious behavioural problems (e.g., smoking, truancy, cheating). This affiliation pattern significantly amplifies adolescents’ vulnerability to both bullying perpetration and victimization. Previous research indicated that interaction with deviant peers may lead to increased peer victimization [11,17], some research even suggests that the association between DPA and bullying dynamics may be bidirectional—serving both as a precursor to and consequence of victimization [18]. Consistent with the process-oriented framework outlined above, the present study treats DPA as a peer-context mechanism linking family-level risk to bullying victimization. This analytic focus is consistent with ecological systems theory, which posits that family dynamics shape adolescents’ peer environments, and is supported by empirical evidence indicating that affiliation with deviant peers increases exposure to peer conflict and social marginalization [10]. Accordingly, DPA is treated as a mediator in the current model to test a theoretically specified family-to-peer-to-victimization pathway.

Hypothesis 2: DPA plays a mediating role between PPC and BV among Chinese adolescents.

1.3 Moderating Role of Self-Compassion

Although negative parenting often elevates risk of affiliation with delinquent peers, notably, a proportion of adolescents reared in such environments do not exhibit such an association with delinquent peers [12]. Individual differences among adolescents may serve as moderating variables, influencing the extent to which negative parenting predicts affiliation with deviant peers, leading to diverse outcomes. This suggests that while negative parenting can be a contributing factor, the interplay of personal and environmental factors ultimately shapes adolescents’ social affiliations and behavioural trajectories. One potential factor in this association is self-compassion (SC) [19,39,40].

SC constitutes a resilient mindset that encompasses kindness towards oneself, recognition of our shared human experiences, and maintaining mindfulness when facing adversity [20]. As a robust protective mechanism, SC buffers adolescent mental health across multiple domains [22,23]. Empirical evidence demonstrates its moderating capacity—higher SC weakens the perfectionism-depression linkage [22,41]. While enhancing stress regulation [23] and reducing depression, anxiety, and stress symptomatology [42]. Within the Social–Ecological Diathesis–Stress framework, SC may function as a protective factor that attenuates the impact of environmental stressors by enhancing adaptive coping and stress regulation [8]. Importantly, SC may be particularly relevant in buffering the peer-selection processes linking PPC to DPA. Adolescents with higher SC tend to maintain more balanced self-evaluations, tolerate interpersonal distress, and preserve autonomy in the face of relational challenges [21]. As a result, they may be less inclined to seek sense of belonging from deviant peer groups when familial relationships are psychologically controlling [12,43]. In contrast, adolescents low in SC may be more vulnerable to internalizing parental rejection and may compensate for unmet emotional needs by affiliating with deviant peers who provide a sense of acceptance [11,24]. Thus, SC is expected to moderate the association between PPC and DPA, rather than exerting a direct influence on bullying victimization. On the basis of the reasoning and evidence in the literature, we propose the hypothesis below:

Hypothesis 3: SC moderates the association between PPC and DPA, such that higher levels of SC weaken the positive association between PPC and DPA, thereby indirectly attenuating adolescents’ BV.

1.4 Present Study

Guided by the above framework, the present study tests a moderated mediation model in which PPC is linked to BV through DPA, with SC moderating the PPC–DPA pathway (Fig. 1). This model allows us to examine whether and under what conditions family-level risk is translated into peer victimization during adolescence.

Figure 1: Structural equation model of overall associations of parental psychological control and bullying victimization. Note: PPC, parental psychological control; DPA, deviant peer affiliation; BV, bullying victimization; SC, self-compassion. Hypothesis 1: PPC is positively associated with adolescents’ BV. Hypothesis 2: DPA plays a mediating role between PPC and BV among Chinese adolescents. Hypothesis 3: SC moderates the association between PPC and DPA, such that higher levels of’ SC weaken the positive association between PPC and DPA, thereby indirectly attenuating adolescents’ BV.

2 Methods

2.1 Participants

Participants were recruited via convenience sampling from two middle schools in Shanxi Province, China. A total of 962 participants (Mean = 14.08, SD = 1.07; 49.83% male) was recruited, and 101 participants were excluded due to inattentive responses or having more than 20% missing data. The efficiency rate of the data was 89.5%.

2.2 Procedure

The study received formal authorization from the principals of the school and the parents of the participants. We collected data from the students at two randomly selected middle schools in Lvliang city, Shanxi Province, between 10 and 20 April 2025. After the principals approved the study, students and parents received information packets with consent forms. With the consent of the parents, the survey was conducted during each class meeting. Students who fulfilled the specified criteria and provided informed consent were given paper questionnaires; then, the completed questionnaires were collected directly by the researcher. PPC was assessed using the Chinese Psychological Control Scale (measuring perceived maternal/paternal control), DPA with the Deviant Peer Affiliation Scale, BV via the Bullying Victimization Scale, and SC with the Self-Compassion Scale. This study was approved by the Ethics Committee of the Faculty of Psychology, Beijing Normal University (No. BNU202503110083).

2.3 Measures

2.3.1 Parental Psychological Control

The Chinese Psychological Control Scale was included to measure participants’ perceived PPC in this study [44]. The scale is composed of two subscales and 20 items. Each subscale consists of 10 questions divided into five dimensions: feelings of neglect (2 items), expression restriction (1 item), personal attack (2 items), withdrawal of love (3 items), and excessive interference (2 items). The scale adopted a 4-point Likert scale, with 1 being “completely disagree” and 4 being “strongly agree”. Consequently, elevated sum scores on the scale were associated with elevated levels of perceived PPC (Cronbach’s α = 0.924).

2.3.2 Self-Compassion

The measurement of self-compassion was conducted using the Self-Compassion Scale [20], which comprises 26 items divided into six dimensions: self-kindness (5 items), mindfulness (4 items), common humanity (4 items), self-judgement (5 items), over-identification (4 items), and isolation (4 items). The items were evaluated using a 5-point Likert scale: from 1 (“almost never”) to 5 (“almost always”). Among the dimensions mentioned above, the latter three were reverse-scored. Higher total scores suggested elevated levels of SC (Cronbach’s α = 0.878).

2.3.3 Deviant Peer Affiliation

DPA was measured via the Deviant Peer Affiliation Scale [45]. Eight items were included in this scale, and each of which was evaluated utilizing a 5-point Likert scale ranging from 1 (“no one”) to 5 (“everyone”). Higher sum scores suggested a heightened probability of deviant peer affiliation (Cronbach’s α = 0.727). The reliability of this scale is relatively modest, which is acceptable [46].

2.3.4 Bullying Victimization

BV was assessed utilizing the culturally adapted Chinese Delaware Bullying Victimization Scale–Student [47]. Given China’s educational situation and the limited accessibility of electronic devices among middle school students, we omitted cyberbullying items (4 items) from the original scale. The 13-item revised scale consists of four items each for verbal bullying, physical bullying, and relational bullying, as well as one nondimensional scored item, Item 13, “I’m bullied at this school”, capturing overall victimization perception. Responses used a 6-point Likert scale (1 = never, 6 = every day), with higher total scores indicating greater BV frequency (Cronbach’s α = 0.865). In line with established bullying research emphasizing repetition as a defining feature [48], participants reporting a frequency of 3 (“once or twice a month”) or higher were classified as victims. This criterion reflects repeated victimization rather than occasional exposure.

2.4 Data Analysis

Prior to data collection, a Monte Carlo power analysis was conducted in Mplus 8.3 (Muthén & Muthén, Los Angeles, CA, USA) to estimate the required sample size for the hypothesized moderated mediation model. Based on parameter values drawn from prior literature, the simulation indicated that a sample size of approximately 200 participants was sufficient to achieve 80% power for detecting the indirect and interaction effects [49]. The final sample size (N = 861) exceeded this requirement, indicating adequate statistical power for the analyses. Missing data were imputed using the expectation-maximization (EM) algorithm.

Descriptive statistics, demographic correlations (sex, age), and bivariate associations were computed in SPSS 27.0 (IBM Corp., Armonk, NY, USA). Structural equation modelling (SEM) in Mplus 8.3 tested direct and mediated pathways linking PPC and DPA to BV. In mediated SEM, the measurement model (Model 1) had one latent variable (parental psychological control, PPC), two observed variables (deviant peer affiliation, DPA, and self-compassion, SC), and two covariates (age and sex). The main effects of PPC and DPA were subsequently introduced into the model (Model 2). Finally, latent moderated structural equations (LMS) tested SC’s moderating role in the DPA pathway (Model 3) (LMSs, Model 3). Model adequacy was evaluated using a three-step comparative procedure [50].

SEM Model fit was assessed by the comparative fit index (CFI), Tucker-Lewis index (TLI), root mean squared error of approximation (RMSEA), and standardized root mean square residual (SRMR). Consistent with prevailing guidelines, acceptable fit required: CFI/TLI > 0.90, RMSEA < 0.10, and SRMR ≤ 0.08. The significance of the indirect effect of DPA was determined through the bootstrapping method, employing a sample size of 5000 and establishing a 95% confidence interval (CI), which was significant if the 95% CI did not cover 0. The relative fitness of the LMS was evaluated based on the Akaike information criterion (AIC) and log-likelihood ratio (LR) test. Comparing the above information of Model 2 and Model 3 revealed that the decrease in the AIC value and the significance of the l LR test could demonstrate the superiority of Model 3 over Model 2 [51,52].

Furthermore, the bootstrapping method [53] was employed to examine the significance of the moderated mediation effect with a sample size of 5000 and a 95% CI, with effects significant when 95% CIs excluded zero.

3 Results

3.1 Descriptive Statistics

Based on the repetition-based criterion (≥3), 59 participants (6.85%) reported repeated verbal victimization (Mean = 1.57, SD = 0.78), 13 participants (1.51%) reported repeated physical victimization (Mean = 1.23, SD = 0.46), and 22 participants (2.55%) reported repeated relational victimization (Mean = 1.29, SD = 0.54). Using the global item (item 13), 1.4% of adolescents reported being bullied at least once or twice per month. Table 1 reports descriptive statistics (means, standard deviations) and Spearman’s rank-order correlations for all study variables. The findings indicated a positive association between PPC and DPA, as well as between PPC and BV, whereas a negative association was observed between PPC and SC.

3.2 SEM Results

Table 2 presents the standardized path estimates for both the mediation model (SEM) and the fully moderated mediation model (LMS). Model fit indices for the measurement and structural models are reported below. Model 1 included only the measurement model, which exhibited an optimal fit, as indicated by the fit index: χ2(19) = 87.87, CFI = 0.98, TLI = 0.97, RMSEA = 0.07 (95% CI [0.05, 0.08]). The direct effect of PPC on BV was incorporated into Model 2, along with the mediating effect of DPA. The fit of Model 2 also met the criterion: χ2(37) = 189.71, CFI = 0.95, TLI = 0.93, RMSEA = 0.07 (95% CI [0.06, 0.08]). As shown in the SEM results in Table 2, PPC is positively associated with BV both directly (β = 0.268, p < 0.001) and indirectly by relating to DPA (β = 0.043, p < 0.01, 95% CI [0.014, 0.072]), with the indirect effect accounting for 13.83% of the total effect.

3.3 LMS Model Results

To examine whether self-compassion (SC) moderates the association between parental psychological control (PPC) and deviant peer affiliation (DPA), we estimated a latent moderated structural model (Model 3) by adding the PPC × SC interaction term to Model 2 (the mediation model). Model comparison using the log-likelihood ratio (LR) test indicated that Model 3 provided a significantly better fit than Model 2, LR (df = 2) = 8.57, p < 0.05, supporting the inclusion of the interaction term.

As shown in Table 2, the latent interaction between PPC and SC was significantly associated with DPA (β = −0.077, SE = 0.038, 95% CI [−0.159, −0.011]), indicating that higher levels of SC attenuated the positive association between PPC and DPA.

To reveal the mechanisms by which SC moderates the effect of PPC on DPA, a simple slope analysis was performed (see Fig. 2). Results confirmed that at low levels of SC (−1 SD), PPC is significantly associated with DPA (b = 0.079, t = 2.244, p = 0.04 < 0.05), with a 95% CI of [0.014, 0.152]; Conversely, at high SC (+1 SD), PPC’s predictive effect became non-significant (b = −0.017, t = −0.620, p = 0.535), with a 95% CI of [−0.073, 0.034]. These findings indicated the mitigation of the adverse impacts of PPC on DPA by SC.

We further examined whether the indirect effect of PPC on BV via DPA was conditional on levels of SC by testing a moderated mediation model using bootstrapping procedures with 5000 resamples. Results indicated that the conditional indirect effect was significant at low levels of SC (−1 SD; β = 0.044, 95% CI [0.009, 0.093]), but became non-significant at mean and high levels of SC. Moreover, the index of moderated mediation was significant (β = −0.043, 95% CI [−0.097, −0.007]), providing formal evidence that SC moderated the indirect effect of PPC on BV through DPA.

Figure 2: Interactive effect of parental psychological control (PPC) and self-compassion (SC) on deviant peer affiliation.

4 Discussion

Rather than solely referring to the statistical findings, the present discussion focuses on interpreting the observed associations within an integrated theoretical framework. Drawing on ecological systems theory [9] and the Social–Ecological Diathesis–Stress model [8], the findings are interpreted as reflecting how family-level risks, peer-context processes, and individual psychological resources jointly shape adolescents’ vulnerability to bullying victimization. Consistent with prior research, PPC was positively associated with BV [26,33,54,55]. Importantly, from an ecological systems perspective, the findings highlight how family-level processes may become embedded within adolescents’ broader social ecologies. Ecological systems theory emphasizes that more distal contexts (e.g., family dynamics) influence development primarily through their impact on more proximal interactional settings [9]. In this sense, PPC may shape adolescents’ emotional and interpersonal functioning in ways (e.g., by invalidating children’s emotional experiences) [30] that become salient (e.g., being perceived as vulnerable by peers) within peer environments, where bullying victimization actually occurs [56,57]. The present findings therefore align with ecological assumptions by suggesting that parenting practices are relevant to victimization, not because they serve as a direct trigger for bullying, but because they shape adolescents’ cognitive processes when selecting peers.

Importantly, although the standardized effects observed in the present study were in the small-to-moderate range (e.g., β = 0.27 for the direct PPC–BV association; β = 0.29 for PPC–DPA; β = 0.56 for DPA–BV), these magnitudes are broadly consistent with prior meta-analytic findings indicating that parenting and peer-related predictors of bullying involvement are typically modest in size [58,59]. From a developmental and ecological perspective, however, even small-to-moderate effects may carry meaningful practical significance, particularly when risk factors accumulate over time or operate across multiple contextual systems [60,61]. Given the high prevalence and long-term consequences of bullying victimization, incremental increases in risk associated with family and peer processes may therefore translate into meaningful population-level impact.

The mediating role of deviant peer affiliation (DPA) further clarifies this ecological process. Although the association between PPC, DPA, and BV has been examined in prior studies, the current findings contribute by situating DPA explicitly as a proximal peer-context mechanism linking family risk to victimization experiences. From a social–ecological diathesis–stress perspective, PPC may be conceptualized as a distal vulnerability factor that interacts with proximal influencing factors (peer contexts) to heighten adolescents’ exposure to interpersonal stressors [8]. Our findings provide evidence that maladaptive parent‒child relationships predispose adolescents to delinquent peer affiliations [34,35]. While this pathway is largely confirmatory, replication within a Chinese adolescent sample strengthens confidence in the cross-cultural relevance of peer-mediated pathways linking family processes to bullying experiences.

One possible explanation is that an absence of love and care in familial environments may render adolescents more prone to seeking intimacy from peer relationships [30,62]. In accordance with social network theory, there are two possible pathways through which PPC shapes adolescent peer network dynamics [63]. The first mechanism, “homophily selection”, involves proactive affiliation with peers exhibiting congruent psychosocial profiles. Adolescents exposed to elevated PPC demonstrate heightened psychosocial maladjustment patterns along the internalizing–externalizing spectrum [32,64,65]. Consequently, they may actively seek peers with shared behavioural profiles through similarity-based selection [63]. Moreover, when relational competence deficits from coercive family dynamics restrict access to prosocial peer networks, adolescents may gravitate toward delinquent peers—a process termed “default selection” due to limited alternatives. Grounded in social learning theory, chronic exposure to PPC impedes the acquisition of prosocial behavioural repertoires essential for maintaining adaptive peer relationships; consequently, adolescents exposed to PPC may experience marginalization within normative peer groups. Such relational exclusion systematically predisposes adolescents to the spontaneous formation of nonconforming affiliations [66]. In summary, they exhibit a greater propensity to affiliate with deviant and delinquent cohorts to gain a sense of belonging and self-worth [36,37].

Third, the moderation by SC represents one of the most theoretically informative aspects of the present study. Rather than functioning as a general protective factor across all adolescents, SC emerged as a conditional resilience resource that weakened the association between PPC and DPA. This pattern refines diathesis–stress assumptions by demonstrating that vulnerability processes are not uniform, but depend on individual-level regulatory capacities. Within the diathesis–stress framework [8], PPC and DPA can be understood as risk-related conditions, whereas SC operates as an intrapersonal resource that alters how adolescents respond to interpersonal stress and maintain healthy interpersonal connections. First, adolescents with higher levels of SC may be better able to maintain balanced self-evaluations [67], regulate negative affect [68], and avoid excessive self-blame [20,69] in response to controlling parenting experiences. These regulatory capacities may, in turn, reduce the tendency to seek affirmation or belonging through maladaptive peer affiliations. Second, SC is a crucial element in fostering positive and healthy interpersonal connections [21]. It facilitates the harmonious integration of autonomy and intimacy, allowing more self-compassionate adolescents to possess the fundamental social competencies necessary to sustain typical peer relationships, which fosters a sense of belonging. Consequently, they are less prone to forming deviant peer connections.

At the same time, the findings suggest that SC is not universally protective. Its buffering role appears most relevant under conditions of elevated familial risk (e.g., controlling parenting), highlighting the importance of considering when and for whom individual strengths are most consequential. This nuanced pattern extends prior work by positioning SC not merely as a correlate of adjustment, but as a contextual moderator within family–peer risk processes. However, it is also important to acknowledge that, from a social–ecological and diathesis–stress perspective [8], under conditions of chronic or severe psychological control, or when adolescents are embedded in highly entrenched deviant peer environments, individual self-regulatory resources such as SC may be insufficient on their own to fully offset accumulated risk.

4.1 Limitations

Several limitations warrant acknowledgment. First, the cross-sectional approach prevents definitive causal claims about observed relationships, while unmeasured variables may drive effects. Additionally, there may be a range of alternative models that could account for the observed outcomes. Future work should prioritize longitudinal or experimental designs to disambiguate the directionality and mechanisms underlying these phenomena. Second, the self-report method used for all the scales in this study may bias the results, although a common method bias test was conducted. According to Barber and Xia [70], subjective reports of PPC tend to mirror adolescents’ interpretations of their relationship with their parents. However, such interpretations may be susceptible to personal biases, emotional states, and developmental stages, which may result in discrepancies between the perceived and actual levels of PPC. Moreover, self-report surveys are susceptible to social desirability bias. Podsakoff et al. discovered that participants frequently adjust their responses in alignment with social norms, particularly in studies on sensitive topics such as peer relationships and parental behaviours [71]. To ensure the replicability of our results, future research should adopt multi-informant approaches (e.g., teacher/parent/peer reports) to mitigate single-reporter bias. Third, the present study focused exclusively on traditional forms of bullying victimization and did not include measures of cyberbullying victimization. Although cyberbullying frequently co-occurs with traditional bullying and may share overlapping family-related and peer-related risk mechanisms [72,73], its exclusion was informed by both measurement considerations and contextual factors. Specifically, the sample consisted of early and middle adolescents from public middle and high schools in mainland China, where school policies and educational regulations typically restrict students’ access to smartphones during school days. As a result, opportunities for cyberbullying exposure in daily school contexts may be relatively limited for this population. Nevertheless, by omitting cyberbullying items, the present study captures only a partial representation of adolescents’ victimization experiences. Future research should incorporate both traditional and cyber forms of bullying to more comprehensively examine the interplay between parental psychological control, peer processes, and adolescent victimization across contexts. Fourth, the reliance on the Chinese early adolescents from two Shanxi middle schools necessitates caution regarding generalizability across cultural contexts, regions, or developmental stages due to the convenience sampling method. Replication in more diverse samples is warranted.

4.2 Implications

The findings offer several tentative implications that should be interpreted cautiously, given the cross-sectional design. First, the observed associations between PPC, DPA, and BV highlight the potential relevance of parenting practices for adolescents’ peer experiences. Parenting education efforts that emphasize autonomy-supportive communication and emotional validation may be promising avenues for future intervention-oriented research, particularly for families characterized by higher levels of psychological control.

Second, the identification of DPA as a proximal correlate of BV underscores how affiliations with high-risk peers are associated with BV. Programs aimed at strengthening adolescents’ social competencies and facilitating engagement with prosocial peer networks may help reduce exposure to high-risk peer contexts. However, longitudinal and experimental studies are necessary to determine whether modifying peer affiliation patterns leads to reductions in BV.

Third, the conditional buffering role of SC suggests that individual psychological resources may shape how adolescents navigate family- and peer-related stressors. Importantly, self-compassion appeared to be most relevant under conditions of elevated familial risk, rather than functioning as a universal protective factor. School- or community-based programs designed to cultivate self-compassion (e.g., mindfulness-based approaches) [74], may therefore warrant further investigation as complementary strategies within broader prevention frameworks, rather than as standalone interventions.

5 Conclusions

This study identified significant associations between PPC and adolescent BV, with DPA statistically accounting for part of this association. SC was found to moderate the PPC–DPA link, such that the association was weaker among adolescents reporting higher levels of self-compassion. These findings highlight the interplay between familial factors, peer affiliations, and individual psychological resources within adolescents’ social ecologies. Accordingly, cultivating self-compassion may represent a promising area for future longitudinal and intervention-based research examining protective processes within adolescent peer ecologies. This insight can help parents, educators, and counsellors develop better strategies to support teens facing family stress and peer challenges, ultimately creating safer and more supportive social environments.