Regarding direct effects, authoritative parenting provides clear behavioral standards alongside emotional support, creating secure environments where children develop adaptive coping strategies (Williams et al., 2009; Karaer and Akdemir, 2019), and may predict IA negatively (H1a). Authoritarian parenting creates emotional deficits that may be consistent with compensatory Internet use according to the compensatory Internet use theory (Kardefelt-Winther, 2014; Lei and Wu, 2007; Xu et al., 2014), while excessive control may paradoxically increase prohibited behaviors through psychological reactance based on psychological reactance theory (Brehm, 1966; Van Petegem et al., 2015), may predict IA positively (H1b). Permissive parenting may leave children with limited guidance to regulate Internet use (Eastin et al., 2006; van den Eijnden et al., 2010), which in turn may also predict IA positively (H1c).

Parenting styles may influence Internet addiction (IA) indirectly through psychological mechanisms, particularly self-control—the capacity to regulate behavior, thoughts, and emotions (Tangney et al., 2004). Low self-control increases vulnerability to impulsive behaviors, including excessive Internet use (Baumeister et al., 2007; Hofmann et al., 2012), with robust evidence linking self-control deficits to IA among East Asian students (Li et al., 2021).

Importantly, parenting styles differentially shape self-control development. Authoritative parents help children internalize behavioral standards and develop impulse control through reasoning rather than punishment (Baumrind, 1991; Steinberg, 2001), strengthening self-regulatory capacities. Conversely, authoritarian parenting may undermine self-control by fostering external regulatory orientation wherein children rely on parental punishment and surveillance rather than internal self-regulation (Grolnick and Ryan, 1989); when supervision is absent, they may lack internal resources to resist impulsive behaviors (Larose et al., 2003). Similarly, permissive parenting associates with higher impulsivity and poorer self-regulation (Pinquart and Gerke, 2019; Piotrowski et al., 2013) as children rarely receive explicit instruction in impulse regulation or adequate self-regulatory models. Therefore, we hypothesize that self-control mediates the relationships between all three parenting styles and IA (H2a–c).

Rejection sensitivity—anxious anticipation and overreaction to rejection cues (Downey and Feldman, 1996)—is identified as a significant IA risk factor (Tao et al., 2022; Fan et al., 2022; Xu et al., 2025), as rejection-sensitive individuals seek relief through excessive Internet use and prefer virtual environments where rejection cues are less salient (Kardefelt-Winther, 2014).

Parenting styles differentially influence rejection sensitivity development. Authoritative parents provide substantial emotional support and rarely reject children, weakening formation of “rejection–anxiety” associations (Downey and Feldman, 1996). The emphasis on reasoning and guided autonomy helps children develop effective emotion regulation strategies when facing interpersonal challenges (Bernier et al., 2010), reducing their tendency to catastrophize potential rejection scenarios. In contrast, authoritarian environments with harsh discipline, rejection, and coldness contribute to rejection sensitivity development (Downey et al., 2004); longitudinal research demonstrates that parental psychological control and coercion positively predict rejection sensitivity in early adolescents (Rowe et al., 2015). Children experience frequent negative encounters involving criticism and conditional acceptance, fostering anticipatory anxiety and defensive overreaction to rejection cues. Permissive parenting, however, involves low conflict and infrequent explicit rejection, lacking the critical rejection-related experiences that heighten rejection sensitivity. Given that permissive parenting’s influence likely operates through alternative mechanisms such as poor self-regulation (Steinberg, 2001), we hypothesize that rejection sensitivity mediates the relationships between authoritative parenting (H3a) and authoritarian parenting (H3b) and IA, but not the relationship between permissive parenting and IA (H3c).

Building on developmental cascade models (Masten and Cicchetti, 2010), we propose that self-control and rejection sensitivity operate sequentially in linking parenting to IA. Low self-control reduces children’s ability to regulate emotions and impulses, increasing susceptibility to anxiety-driven interpretations of social situations and heightened rejection sensitivity (Blair and Raver, 2015; Fan et al., 2022). Heightened rejection sensitivity, in turn, motivates excessive Internet use as children seek to avoid interpersonal risks (Kardefelt-Winther, 2014; Tao et al., 2022). Accordingly, we hypothesize that all three parenting styles influence IA through this sequential pathway: authoritative parenting reduces IA via higher self-control and lower rejection sensitivity, whereas authoritarian and permissive parenting increase IA via lower self-control and higher rejection sensitivity (H4a–c).

While parenting and mediating mechanisms explain how IA risk develops, individual differences may determine whether these risks materialize. Achievement motivation—an individual’s drive to pursue success and master challenging tasks (Martin, 2012)—may moderate the effects of self-control and rejection sensitivity on IA. Drawing on Achievement Goal Theory (Dweck and Leggett, 1988; Elliot, 1999), we propose that achievement motivation serves as a boundary condition, particularly in Chinese primary school contexts where academic achievement is highly emphasized (Chao, 1994) and motivational patterns are forming (Wigfield et al., 2015; Gottfried et al., 2001). High achievement motivation may promote goal-directed regulation and buffer against self-regulatory deficits and emotional distress (Duckworth et al., 2007; Hofer and Fries, 2016).

For the self-control pathway, children with high achievement motivation may rely on future-oriented goals to regulate Internet use, reducing dependence on trait self-control. Thus, even children with lower self-control may restrain excessive use, weakening the protective effect of self-control at high motivation levels. Conversely, when achievement motivation is low, self-control is more critical for regulating Internet use (Fujita, 2011; Inzlicht and Schmeichel, 2012; Milyavskaya et al., 2015). Therefore, achievement motivation is hypothesized to moderate the self-control–IA link (H5).

For the rejection sensitivity pathway, high achievement motivation provides adaptive coping and a future-oriented sense of competence, reducing reliance on Internet use to manage social stress and rejection-related anxiety (Dweck, 2006; Burnette et al., 2013; Duckworth et al., 2007). In contrast, children with high rejection sensitivity and low achievement motivation face a “double vulnerability,” heightening IA risk. Thus, achievement motivation is expected to moderate the rejection sensitivity–IA link, weakening its effect at higher levels (H6).

The present study investigates the associations between Baumrind’s three parenting styles and IA among Chinese primary school students using an integrated moderated sequential mediation framework (see Figure 1). Specifically, this framework advances existing literature by: (1) providing culturally grounded evidence on the direct effects of parenting styles on IA (H1a–c); (2) identifying self-control (H2a–c) and rejection sensitivity (H3a–c) as distinct yet complementary mediating mechanisms; (3) testing a novel sequential mediation pathway that specifies the temporal ordering of these psychological processes (H4a–c); and (4) examining achievement motivation as a moderator that delineates critical boundary conditions (H5, H6). By clarifying the ordered psychological mechanisms linking parenting to IA within the Chinese cultural context, this integrative framework advances theory and informs multi-level interventions, including family-based programs targeting parenting practices, individual-focused strategies to strengthen self-regulation, and school-based initiatives that cultivate adaptive achievement motivation.

A total of 721 grade 4–6 students were recruited from selected classes in two public primary schools in Changsha, Hunan Province, using a convenience sampling method, targeting classes that were readily accessible. Within the selected classes, all students whose parents provided consent were invited to participate. The researchers reviewed all completed questionnaires for data quality. Responses were excluded if more than 10 items were missing across all questionnaires, if more than two consecutive scenarios (six items) were missing in the rejection sensitivity scale, or if identical response options were selected across all items. After applying these criteria, 661 valid responses remained, yielding a valid response rate of 91.68%. The mean age of the participants was 11.23 years (SD = 0.77). The gender distribution was nearly balanced, with 49.5% boys and 50.5% girls. Of the 661 participants, 4.4% were in grade 4, 50.7% in grade 5, and 44.9% in grade 6. Additionally, 65.2% of the participants reported having siblings, and 75.6% resided in urban areas. Regarding parental education, 31.8% of fathers held a bachelor’s degree, which was the same as that (31.8%) observed for mothers. A one-way ANOVA indicated that IA levels significantly differed across groups based on sibling status (F = 5.70, p < 0.01) and family residence (F = 3.70, p < 0.05), suggesting that these demographic factors may influence IA. Therefore, these two sociodemographic variables were included as covariates in the subsequent analyses to control for potential confounding effects. Additional demographic information is presented in Table 1.

To determine the appropriate sample size, we conducted a Monte Carlo power analysis for indirect effects (Schoemann et al., 2017). Results indicated that a minimum of 205 participants would be required to achieve a statistical power of 90% in a serial mediation model. Therefore, the current sample size was sufficient for robust statistical analysis.

This study received ethical approval from the Ethics Committee of Science and Technology at our university and adhered to the principles outlined in the Declaration of Helsinki. Data were collected between February and May 2023. Prior to participation, written informed consent was obtained from both the children and their parents. Participants were asked to complete a structured questionnaire that assessed demographic characteristics, parenting styles, self-control, rejection sensitivity, achievement motivation, and IA. They were instructed to finish the questionnaire within 40 min and were allowed to seek clarification from a trained assistant teacher if needed. Participants were informed of their right to withdraw at any point prior to submitting the completed questionnaire. The researchers reviewed the completed questionnaires for data quality, and responses that did not meet the predefined exclusion criteria were excluded from the analysis.

All statistical analyses were performed using IBM SPSS Statistics version 22.0 and Mplus version 8.3. Descriptive and inferential analyses were conducted to examine participants’ gender, grade level, only-child status, place of residence, paternal education level, and maternal education level. Independent-samples t-tests were used to assess gender differences in IA. One-way analyses of variance (ANOVAs) were applied to examine differences in IA across grade level, only-child status, place of residence, paternal education level, and maternal education level. Pearson’s correlation analysis was performed to examine the relationships among authoritarian, authoritative, and permissive parenting styles, rejection sensitivity, self-control, achievement motivation, and IA. Confirmatory factor analysis (CFA) was performed in Mplus 8.3 to evaluate the construct validity of the scales. Harman’s single-factor test was employed to assess potential common method bias. Missing data were handled using listwise deletion, consistent with the default procedure in PROCESS. Model 88 of the PROCESS was employed to test the moderated sequential mediation model. The model examined authoritative parenting style, authoritarian parenting style, permissive parenting style, as the predictor, with self-control and rejection sensitivity as serial mediators, achievement motivation as the moderator, and IA as the outcome. All continuous variables were mean-centered prior to analysis to facilitate interpretation of interaction terms and reduce multicollinearity (Aiken and West, 1991). Indirect effects were estimated using 5,000 bootstrap resamples, and statistical significance was determined based on 95% bias-corrected confidence intervals that did not include zero.

Harman’s (1967) single-factor test was conducted to assess common method bias. Results indicated that 28 factors had eigenvalues greater than 1, with the largest factor accounting for 11.66% of the variance, well below the conventional 40% threshold. While these results suggest that no single factor dominated, all variables were collected via child self-report in a single session, and other sources of method bias cannot be entirely ruled out. To reduce potential common method bias, several procedural safeguards were implemented, including assurances of confidentiality, standardized instructions, and exclusion of questionnaires exhibiting inattentive or uniform response patterns.

Pearson correlation analyses revealed that authoritative parenting was negatively correlated with IA (r = −0.13, p < 0.001) and rejection sensitivity (r = −0.14, p < 0.001), and positively correlated with self-control (r = 0.17, p < 0.001) and achievement motivation (r = 0.13, p < 0.001). Authoritarian parenting was positively correlated with IA (r = 0.14, p < 0.001) and rejection sensitivity (r = 0.13, p < 0.001), negatively correlated with self-control (r = −0.26, p < 0.001), and not significantly correlated with achievement motivation (r = −0.02, p > 0.05). Permissive parenting was not significantly correlated with IA. Additionally, self-control was negatively correlated with rejection sensitivity (r = −0.16, p < 0.001), and achievement motivation was negatively correlated with IA (r = −0.27, p < 0.001) (see Table 2). Overall, these are small associations.

The total effects of the three parenting styles indicated that authoritative parenting negatively predicted IA (β = −0.17, p < 0.05), whereas authoritarian parenting positively predicted IA (β = 0.17, p < 0.05); permissive parenting, however, showed a non-significant effect (β = 0.07, p > 0.05). After controlling for mediators and the moderator, the direct effects of all three parenting styles—authoritative (β = −0.01, p > 0.05), authoritarian (β = 0.09, p > 0.05), and permissive (β = −0.02, p > 0.05)—became non-significant (see Table 3). These results suggest that the associations between parenting styles and IA are primarily transmitted indirectly through self-control and rejection sensitivity.

As the direct effects of authoritative parenting style, authoritarian parenting style, and permissive parenting style, on self-control and rejection sensitivity differed in the model, the subsequent mediation chain from self-control through rejection sensitivity to IA remained structurally identical with equivalent coefficients. We established three separate models based on the three parenting styles, as shown in Figures 2–4. The varying coefficients are presented in the first three columns of Table 3, while the invariant coefficients are presented in the fourth column.

Authoritative parenting’s negative association with IA, supporting Hypothesis 1a, reflects not merely high expectations but the optimal combination of structure and emotional support. This finding is particularly significant in the Chinese context, challenging traditional emphasis on strict discipline (guan) (Chao, 1994) while validating the importance of parental warmth. When parents set clear boundaries with warmth and open communication, children develop secure attachment and internalized self-regulation, reducing reliance on Internet for emotional fulfillment or escape (Monacis et al., 2017). This suggests contemporary Chinese families benefit from integrating traditional high expectations with modern emphases on emotional connection and autonomy support, aligning with Self-Determination Theory (Ryan and Deci, 2017) that environments satisfying autonomy, competence, and relatedness promote adaptive behaviors.

Conversely, authoritarian parenting’s positive association with IA, supporting Hypothesis 1b, indicates that excessive control, harsh punishment, and emotional neglect—rather than high demands per se—are associated with increased risk of problematic Internet use in children. This carries particular significance given authoritarian practices’ roots in Confucian traditions emphasizing filial piety and parental authority (Chao, 1994). However, when parents prioritize obedience over emotional connection, children experience stress, reduced autonomy, and family disconnection, making the Internet a refuge for autonomy and social connection (Kunz and Grych, 2013). This compensatory function suggests that as Chinese society modernizes, traditional authoritarian approaches may be diminishing in effectiveness, particularly where children face both traditional control and unprecedented digital freedoms (Ren et al., 2024). The psychological reactance triggered by excessive control may further lead children to engage in Internet use as rebellion (Van Petegem et al., 2015), aligning with Self-Determination Theory’s premise that environments thwarting basic psychological needs are associated with maladaptive compensatory behaviors (Ryan and Deci, 2017).

The non-significant direct effect of permissive parenting on IA contrasts with some Western research (Valcke et al., 2010) but aligns with studies reporting inconsistent or null effects (Anandari, 2016; Dogan et al., 2015), and thus failed to support Hypothesis 1c. This pattern may reflect the structurally heterogeneous nature of permissive parenting in contemporary China, where the construct encompasses qualitatively distinct subtypes with opposing influences. Specifically, “permissive parenting” in China includes both “left-behind children” experiencing involuntary parental absence due to rural-to-urban migration (Meng and Yamauchi, 2017) and children of Western-influenced urban parents who intentionally grant autonomy to foster independence (Ren and Edwards, 2015). These opposing pathways—neglect versus intentional autonomy-granting—likely cancel out at the population level, producing null aggregate direct effects. Additionally, central Chinese urban families may exhibit distinct dynamics compared to coastal metropolitan areas (Liu S.-W. et al., 2025), further contributing to regional heterogeneity. The borderline internal consistency of the permissive parenting subscale may have slightly attenuated the observed association with IA, and the near-null correlation should be interpreted with caution. Critically, permissive parenting negatively correlated with self-control, which predicted IA, indicating indirect influence. This underscores the importance of examining mediation mechanisms even when direct effects are non-significant.

Self-control emerged as a significant mediator in the relationships between all three parenting styles and children’s IA, supporting hypotheses 2a, 2b, and 2c. These findings confirm recent meta-analytic evidence establishing self-regulation as a primary protective factor against addictive behaviors (Li et al., 2021). Moreover, these results extend the strength model of self-control (Baumeister et al., 2007) to the digital domain, demonstrating that children lacking self-regulatory resources struggle to resist impulse-driven behaviors offering immediate gratification, such as gaming, social media, and video consumption (Brand et al., 2019).

Authoritative parenting predicted higher self-control, which negatively predicted IA. This finding aligns with extensive research demonstrating that supportive, structured parenting environments foster self-regulatory capabilities in children (Marcone et al., 2020). The combination of clear expectations, behavioral guidance, and positive role modeling inherent in authoritative parenting provides children with both the motivation and the skills necessary for effective self-regulation. In Chinese cultural contexts where self-regulation and behavioral discipline are highly valued (Chao, 1994), this pathway represents optimal integration of traditional “guan” (discipline) with autonomy support—children develop value-based rather than fear-based regulation, enabling resistance to Internet temptations even without parental surveillance.

Authoritarian parenting predicted lower self-control, which increased IA risk. This pathway suggests that excessive parental control may undermine children’s development of autonomous self-regulation. When children rely primarily on external punishment to control their behavior rather than developing internal regulatory mechanisms, they may struggle to resist the immediate gratification offered by Internet activities in the absence of parental surveillance. This finding extends previous research on parenting and adolescent self-control (Lei et al., 2021) to younger children, emphasizing the early importance of fostering internal regulation rather than relying solely on external control.

Notably, although permissive parenting showed no total effect on IA, it significantly predicted lower self-control, which in turn increased IA risk. This indirect-only mediation pattern reveals an important nuance: permissive parenting may undermine self-control development through insufficient guidance and structure, yet other unmeasured protective factors in permissive families (e.g., parental warmth, child autonomy) may counterbalance this risk at the total effect level. This finding demonstrates the value of examining indirect pathways even when total effects are non-significant (Preacher and Hayes, 2008) and suggests that interventions targeting self-control may be particularly relevant for children from permissive households.

Contrary to hypotheses 3a and 3b but consistent with 3c, rejection sensitivity did not function as an independent mediator directly following parenting practices. This unexpected finding provides crucial theoretical insights into the developmental ordering of psychological mechanisms. Rather than operating as a parallel pathway alongside self-control, rejection sensitivity emerged as a subsequent consequence—a discovery with important implications for understanding IA etiology.

The pattern of results suggests that parenting practices influence rejection sensitivity primarily through their impact on self-regulatory capacities. Authoritative parenting showed no direct effect on rejection sensitivity, indicating its protective influence operates entirely through enhancing self-control. Even authoritarian parenting, which showed a significant direct path to rejection sensitivity, exerted its ultimate influence on IA through the self-control → rejection sensitivity cascade rather than through an independent direct pathway. These findings align with emerging research emphasizing executive function’s primacy in emotional development (Blair and Raver, 2015).

Recent developmental neuroscience research provides mechanistic insights into this temporal ordering. Self-regulatory capacities, grounded in prefrontal cortex maturation, develop earlier and more continuously across childhood than complex socio-emotional processing patterns like rejection sensitivity (Zelazo, 2020). Children with well-developed self-control can effectively modulate emotional responses to ambiguous social cues, reducing tendency toward anxiety-driven interpretations characteristic of high rejection sensitivity (Fan et al., 2022). Conversely, children lacking self-regulatory capacities become vulnerable to emotional dysregulation in social contexts, manifesting as heightened vigilance for rejection cues and amplified responses to perceived rejection (Shafiq and Batool, 2021).

Permissive parenting showed no association with rejection sensitivity (Hypothesis 3c). Low conflict, minimal criticism, and infrequent rejection provide fewer rejection cues than authoritarian households, not significantly impacting anticipatory anxiety. Rejection sensitivity relates more to critical or rejecting behaviors (Downey et al., 2004) than lack of structure, thus permissive parenting does not appear to be linked to this pathway to IA.

The most significant contribution lies in demonstrating that parenting styles influence IA through an ordered developmental cascade: self-control shapes rejection sensitivity, which subsequently affects Internet use. This sequential pattern emerged consistently across all three parenting styles, supporting a developmental cascade model (Masten and Cicchetti, 2010) wherein self-control serves as the proximal mechanism directly shaped by parenting, and rejection sensitivity emerges as a distal consequence. Although this study is cross-sectional in design, the findings align with longitudinal evidence showing that early self-regulation predicts subsequent emotional reactivity, but not vice versa (Dollar et al., 2023; Spinrad et al., 2012), and extends the strength model of self-control (Baumeister et al., 2007) to developmental processes. This integrates with compensatory Internet use theory (Kardefelt-Winther, 2014): children lacking self-regulatory capacities struggle to manage emotional challenges, developing heightened rejection sensitivity that motivates escape into virtual environments. Children with both poor self-regulation and high rejection sensitivity face “double jeopardy,” explaining individual differences in IA vulnerability.

However, the directional patterns and underlying mechanisms differed meaningfully across parenting types. Authoritative parenting initiated a protective cascade: enhanced self-control reduced rejection sensitivity, ultimately decreasing IA risk. The observed mediation pattern indicates that the protective influence is mainly transmitted via psychological mechanisms. Children from authoritative homes possess both the self-regulatory capacity to resist impulsive online engagement and sufficient emotional resilience to manage offline challenges—a “double protection” eliminating the psychological need for compensatory Internet use.

In contrast, authoritarian parenting initiated a risk cascade but uniquely demonstrated both a direct effect on rejection sensitivity and an indirect effect through self-control—a “dual-pathway” mechanism. The direct pathway reflects children’s emotional responses to harsh, psychologically controlling practices, developing heightened vigilance to disapproval. The indirect pathway reflects the developmental cascade: children who fail to develop self-regulation under external control subsequently struggle to manage interpersonal challenges. Despite its emphasis on discipline, authoritarian parenting undermines internalized self-regulation by relying on external controls rather than fostering autonomous motivation (Baumrind, 1991; Steinberg, 2001). This creates “double jeopardy”: children lack both the self-regulatory capacity to resist impulsive online engagement and experience strong rejection sensitivity motivating escape into virtual environments.

Permissive parenting also initiated a risk cascade through impaired self-control and heightened rejection sensitivity but demonstrated no direct effect on rejection sensitivity—an “indirect-only” pattern indicating its influence on emotional functioning operates entirely through failing to develop self-regulatory capacities. Despite surface differences from authoritarian parenting (passive neglect versus active control), both share the critical commonality of failing to foster internalized self-regulation (Baumrind, 1991). While authoritarian parenting may actively create emotional distress (reflected in the direct path to rejection sensitivity), permissive parenting’s influence appears more passive—failing to build protective capacities rather than actively causing harm. In the Chinese context, “permissive parenting” encompasses heterogeneous subtypes—from involuntary neglect due to parental migration to intentional autonomy-granting—yet the common consequence is impaired self-regulatory development cascading into emotional vulnerabilities.

The findings offer several practical implications for preventing children’s IA. First, parent education programs focusing on authoritative parenting may be promising targets for intervention (Theopilus et al., 2024; Wang et al., 2024), which integrates clear behavioral expectations with emotional warmth and responsiveness. A recent systematic review of preventive interventions found that parenting strategy approaches showed promising efficacy in reducing IA risks with small-to-medium effect sizes (Theopilus et al., 2024). Specifically, parents are encouraged to establish consistent rules for Internet use, explain the rationale behind these rules, encourage open dialogue about online activities, and provide emotional support when children encounter difficulties (Theopilus et al., 2024).

For authoritarian families, interventions could emphasize the integration of discipline and warmth, consistent with the Chinese concept of “guan” (governance with affection) (Wu, 2012). Simply imposing rules without emotional responsiveness may hinder children’s ability to internalize behavioral norms (Grusec and Goodnow, 1994; Li et al., 2023). In contrast, providing warmth and support alongside clear expectations helps children develop self-control (Wang et al., 2007; Li et al., 2023), which in turn may reduce rejection sensitivity and lower the risk of IA (Rohner, 2017; Rothenberg et al., 2022). For permissive families, guidance should consider the cultural and structural realities of Chinese society. Evidence from community-based interventions suggests that in left-behind or migrant families, community-based supervision may compensate for limited parental monitoring (Zhao et al., 2017; Hesketh et al., 2020; Zhang et al., 2020). For instance, a study of community-based children’s centers in rural China demonstrated improvements in psychosocial well-being among left-behind children, particularly in areas where homework supervision and structured activities were provided (Hesketh et al., 2020). In urban families influenced by liberal parenting beliefs, education programs could highlight the importance of structure and routine in fostering children’s self-control within a supportive family environment (Maccoby and Martin, 1983; Li et al., 2023; Li et al., 2025).

Second, interventions could target self-control as a key mechanism. Evidence from mindfulness-based and self-regulation training programs have been shown to enhance children’s attention, impulse control, and emotion-regulation capacities (Bockmann and Yu, 2023; Lawler et al., 2019; Porter et al., 2024). Such programs may also help children reinterpret rejection experiences and develop healthier emotional responses. Moreover, given the sequential mediation pathway identified in this study, early intervention focusing on parenting and self-control may prevent the cascade toward heightened rejection sensitivity and IA. Consistent with the developmental cascade framework (Masten and Cicchetti, 2010), “breaking the chain” during early childhood—when self-regulatory capacities are still forming and family influences are most salient—could yield more enduring protective effects.

Third, fostering achievement motivation may serve as a protective buffer. Educational programs helping children set meaningful goals and develop mastery experiences can enhance resilience against both low self-control and rejection sensitivity. However, healthy achievement motivation should be distinguished from excessive performance pressure, which may be associated with increased Internet use as an escape in children.

Finally, tailored, multi-level intervention strategies are needed. School-based universal prevention programs that integrate parent training, children’s self-regulation skill development, and institutional support may be most effective in reducing IA risk across diverse family contexts, complementing family-focused interventions.

Several limitations should be acknowledged. First, the cross-sectional design substantially limits causal inference and prevents examination of temporal or developmental relationships among parenting styles, self-control, rejection sensitivity, and IA, as these constructs are likely to change across childhood development. Longitudinal or experimental studies are needed to clarify causal pathways and test developmental trajectories. Second, all variables were assessed via child self-report in a single session, which raises concerns regarding common method bias and social desirability effects. Although Harman’s single-factor test did not indicate a single dominant factor, other sources of method bias cannot be fully ruled out. Procedural safeguards—including confidentiality, standardized instructions, and exclusion of inattentive responses—were implemented to reduce bias, but future research should incorporate multi-informant assessments (e.g., parent reports of parenting and children’s IA, teacher reports of self-control), behavioral observations, and objective measures of Internet use to enhance measurement validity.

Third, listwise deletion was used to handle missing data, which may introduce bias if data are not missing completely at random (MCAR). Although the proportion of missing data was small, future studies should consider alternative approaches such as multiple imputation or full information maximum likelihood (FIML) to address potential biases associated with missing data. Fourth, the permissive parenting subscale demonstrated borderline internal consistency, which may have attenuated correlations with IA and contributed to the near-null association observed. Accordingly, findings related to permissive parenting should be interpreted with caution. In addition, although confirmatory factor analyses were conducted for all measures, some fit indices (CFI/TLI) fell below the commonly recommended threshold of 0.90. Specifically, the Western-developed parenting measure showed suboptimal fit, which is consistent with prior findings that such instruments may not fully capture the nuances of parenting in Chinese cultural contexts. Similarly, the Grasmick Self-Control Scale demonstrated high internal consistency but suboptimal factor structure, possibly because it was originally developed for adults rather than children. While these measurement limitations did not affect our primary analyses—which were based on composite scores with acceptable internal consistency rather than latent variable modeling—future studies should develop and validate culturally appropriate parenting measures and age-appropriate self-control instruments for Chinese primary school populations.

Fifth, the study sampled children from two urban public schools in central China, which may limit generalizability to regions with substantially higher or lower socioeconomic levels. Given China’s substantial regional variation in economic development and educational philosophies, the identified patterns may differ in highly developed metropolitan areas or less developed regions. Future research should examine whether these findings replicate across regions with diverse economic and cultural contexts. Sixth, the present study did not distinguish between specific types of Internet use or examine potential gender differences in the proposed pathways. Not all Internet activities are equally problematic, as educational use may differ substantially from gaming or social media engagement. Moreover, boys and girls may respond differently to parenting practices and exhibit distinct Internet use patterns. Although preliminary analyses suggested no significant gender differences in the key study variables, future studies with larger samples are warranted to investigate type-specific and gender-specific pathways. Finally, the study did not assess how children’s characteristics may influence parenting behaviors. Transactional models suggest that parent–child influences are reciprocal, with children’s IA, self-control, or rejection sensitivity potentially eliciting different parenting responses. Future research employing cross-lagged panel designs could examine these bidirectional relationships.