Parental psychological control and problematic smartphone use in Chinese early adolescents: shyness as a mediator and teacher–student relationship as a moderator

Abstract

Problematic smartphone use (PSU) has become an increasingly prominent concern among early adolescents. Although prior research has consistently linked parental psychological control (PPC) to adolescents’ problematic smartphone use, less is known about how family, school, and individual factors jointly relate to PSU during early adolescence. Guided by the Interactional Theory of Childhood Problematic Media Use (IT-CPU), this study examined the associations among PPC, teacher–student relationship (TSR), shyness, and PSU in a sample of 1,176 early adolescents from Shandong Province, China. Participants completed standardized questionnaires administered in classroom settings. Results indicated that PPC positively predicted PSU, and shyness partially mediated this association. In addition, TSR moderated the association between PPC and shyness. Specifically, the link between PPC and shyness was not significant under conditions of low TSR, whereas a stronger positive association emerged under high TSR, revealing an anti-buffering effect. These findings highlight the joint role of family, school, and individual factors in early adolescents’ problematic digital behavior and suggest the value of coordinated family–school support in prevention efforts within the Chinese educational context.

Introduction

With the widespread adoption of mobile Internet technologies, smartphones have become the primary device through which children and adolescents access the Internet. Recent national survey data from China indicate that the Internet penetration rate among minors has exceeded 97%, approaching saturation, and that more than 90% of minor Internet users primarily go online via smartphones [14]. These findings suggest that smartphones have become deeply embedded in minors’ daily lives and constitute a central component of their digital ecology.

This phenomenon is particularly noteworthy for early adolescents. Early adolescence, typically defined as ages 10–14, represents a critical developmental period characterized by rapid changes in peer relationships, self-identity, and social roles, as well as an important transition from late primary school to junior secondary school. During this stage, individuals are required to renegotiate the balance between attachment and autonomy within family and school contexts, while becoming increasingly sensitive to peer evaluation. In the context of widespread access to mobile devices, early adolescents may be especially susceptible to the influence of digital media use patterns. Therefore, problematic smartphone use during early adolescence warrants further attention.

Rather than focusing solely on usage duration, problematic smartphone use refers to a pattern of dysregulated use accompanied by functional impairment, such as academic difficulties, interpersonal problems, sleep disturbances, and challenges in emotion regulation. Previous research has emphasized that high-frequency use should not be equated with clinically meaningful “addiction” [2]; instead, problematic use should be understood from the perspective of interactions between individual needs and environmental factors [24].

Among the various influencing factors, parental psychological control has been identified as an important family-related risk factor associated with adolescents’ problematic smartphone use. Parental psychological control involves intrusive parenting practices—such as guilt induction, love withdrawal, and invalidation—that manipulate children’s psychological and emotional experiences, and it is conceptually distinct from behavioral control, which focuses on rule-setting and supervision [1]. Longitudinal research has suggested that, during early adolescence, parental psychological control and problematic smartphone use are bidirectionally associated [30]. In addition, meta-analytic evidence has confirmed a stable positive association between negative parenting practices (e.g., psychological control) and problematic smartphone use [33].

Although existing research has consistently demonstrated an association between parental psychological control and adolescents’ problematic smartphone use, less is known about how this association operates through more specific individual socio-emotional vulnerabilities during early adolescence, and whether relational resources within the school context are related to this process. The Interactional Theory of Childhood Problematic Media Use (IT-CPU) provides a valuable integrative framework for addressing this question. This theory emphasizes that problematic media use is jointly associated with distal ecological risk factors and proximal individual vulnerabilities, and that its development is embedded within multiple ecological contexts, including family and school [15]. Guided by this framework, the present study examines how parental psychological control (a distal family risk), shyness (a proximal individual socio-emotional vulnerability), and teacher–student relationship (a key relational resource in the school context) jointly relate to early adolescents’ problematic smartphone use.

The mediating role of shyness

Shyness is characterized by heightened sensitivity to negative evaluation in social situations due to a strong desire for social acceptance, which often leads to tension, withdrawal, and behavioral inhibition [3]. With the onset of adolescence, increases in peer competition and public social interaction, together with the continuing influence of “face” culture and parental expectations, render evaluation sensitivity particularly salient for Chinese adolescents [10]. Compared with transient state emotions, shyness represents a relatively stable dispositional trait, making it a suitable candidate for examining its role as a potential mediator linking family risk factors to problematic smartphone use [34].

On the one hand, parental psychological control is closely associated with shyness. Such control often involves emotional manipulation—such as guilt induction, love withdrawal, and devaluation—that links children’s sense of self-worth to external evaluation. During adolescence, when personal boundaries and self-concept are still developing, this parenting style may heighten self-monitoring and expectations of negative evaluation from others, which in turn may be associated with greater anxiety and behavioral inhibition. For example, didactic discipline and strong parental expectations commonly observed in Chinese families (e.g., aspirations for children’s success) are often conveyed through guilt-inducing practices, which may be related to a greater tendency toward shyness, reflected in increased evaluation sensitivity and behavioral restraint [27]. Empirical evidence supports this association: Zhang et al. [34] surveyed 1,857 Chinese middle school students and found that multiple dimensions of parental psychological control (e.g., authoritarian insistence and emotional control) were significantly associated with adolescents’ levels of shyness. Additional research has shown that, after controlling for grade level and parental education, the association between psychological control and adolescent shyness remains stable [36].

On the other hand, shyness is closely related to problematic smartphone use. In face-to-face interactions, shy individuals often worry about self-disclosure and fear immediate evaluation, which may increase the perceived burden of interpersonal communication. In contrast, smartphone-mediated and online communication allows individuals to regulate the pace of interaction, edit responses, and manage self-presentation, thereby reducing the pressure of immediate evaluation. Previous research has indicated that online communication provides shy or socially anxious adolescents with an important avenue for developing close relationships and fulfilling social needs that may be difficult to achieve through face-to-face interaction [29]. Accordingly, compared with real-life interactions that involve higher social costs and stronger evaluation pressure, shy adolescents may be more likely to turn to smartphones and other relatively low-cost communication channels to alleviate emotional discomfort and maintain social participation. Meta-analytic evidence has demonstrated a stable positive association between shyness and maladaptive digital media use [32], and this association may be strengthened by psychological factors such as social anxiety [17].

Accordingly, this study hypothesizes that shyness will partially mediate the association between parental psychological control and problematic smartphone use.

The moderating role of the teacher–student relationship

Although prior research has consistently documented an association between parental psychological control and adolescent shyness, the strength of this association may not be uniform across all early adolescents. From an ecological perspective, individual development is embedded within multiple interacting microsystems, among which family and school represent two central contexts shaping early adolescents’ socio-emotional adjustment. Therefore, it is important to identify key relational resources within the school context to better understand individual differences in the association between parental psychological control and shyness [4].

Within the school microsystem, teacher–student relationship is a core dimension of school adjustment and has been widely regarded as a potential protective resource. Evidence from Chinese samples indicates that high-quality teacher–student relationships are associated with more positive developmental outcomes. For example, multilevel analyses have shown that supportive teacher–student relationships are linked to weaker negative associations between family socioeconomic disadvantage and students’ self-efficacy and academic performance [22]. A three-level meta-analysis of Chinese studies further reported a positive overall association between teacher–student relationship and academic achievement [18]. Together, these findings suggest that teacher–student relationship is often associated with more favorable indicators of academic and school adjustment.

However, ecological theory also suggests that the function of school-based relational resources may depend on the relational climate and risk characteristics within the family microsystem. Compared with severe family adversity, parental psychological control is typically more subtle and routine, characterized by practices such as guilt induction and love withdrawal that intrude into children’s psychological experiences [1]. In the Chinese sociocultural context, parental control is often intertwined with parental involvement, responsibility, and guidance; thus, the presence of psychological control does not necessarily indicate the absence of family care and support [6, 31]. Under such conditions, the potential compensatory role of school-based relational resources may not be linear. When parental psychological control is relatively low, emotional support and autonomy experiences within the family may be more stable, and relational experiences across family and school contexts may be more consistent. In this case, teacher support may be more readily internalized as a resource that facilitates social adjustment, thereby weakening the association between parental psychological control and shyness. In contrast, when parental psychological control is relatively high, adolescents may show heightened sensitivity to evaluation and self-monitoring in family interactions. In Chinese classroom settings, high-quality teacher–student relationships are often accompanied by more frequent interaction and clearer signals of teacher expectations; under such circumstances, the salience of being noticed or evaluated may increase, potentially reducing the traditional buffering role of teacher support or even producing an enhancing pattern. Consistent with this possibility, research with Chinese adolescents has shown that teacher support does not always function as a simple buffer. For example, in the association between parental rejection and Internet addiction, teacher support demonstrated an enhancing pattern, with the association being significant only among students reporting high levels of teacher support [35].

Based on this reasoning, Hypothesis 2 proposes that teacher–student relationship moderates the association between parental psychological control and early adolescents’ shyness. Specifically, when parental psychological control is low, high-quality teacher–student relationship is expected to exhibit a buffering pattern; when parental psychological control is high, this buffering effect may weaken or even reverse, showing an enhancing pattern.

The hypothesized model of the present study is shown in Fig. 1.

Fig. 1.

Conceptual model. Note. PPC = parental psychological control; TSR = teacher–student relationship; SHY = shyness; PSU = problematic smartphone use

Methods

Participants and procedure

Following approval from the local research ethics committee, questionnaires were administered to 1,190 upper primary school students from three primary schools in Shandong Province, East China. A total of 1,176 valid responses were obtained, yielding an effective response rate of 98.82%. Among the participants, 55.4% were in Grade 5 and 44.6% were in Grade 6. The mean age was 11.45 years (SD = 1.70), with an equal gender distribution. In China’s compulsory education system, primary school typically lasts six years, and children usually enter school at around age 6. Consequently, students in Grades 5 and 6 are generally between 10 and 13 years old, largely overlapping with the developmental period commonly defined as early adolescence (10–14 years). Therefore, the present sample can be considered a relevant subsample of early adolescents.

Regarding parental education, students reported the highest educational attainment of either parent: 7.8% held a master’s degree, 16.4% a bachelor’s degree, 18.5% an associate degree, 28.8% a senior high school diploma, 26.2% a junior high school diploma, 1.6% an elementary school education, and 0.7% did not report this information. All participants were physically and mentally healthy and had normal or corrected-to-normal vision. Informed consent was obtained from students, their parents, and teachers prior to data collection.

Data collection was conducted at the class level and administered by trained psychology graduate students. Before administration, participants were informed of the confidentiality and anonymity of their responses and advised of their right to decline participation or withdraw at any time. All questionnaires were completed and collected on site, with the procedure lasting approximately 25 min.

Measures

Parental psychological control (PPC)

Students reported on their PPC using an 18-item measure rated on a 5-point scale (1 = not at all true, 5 = very true). This scale was developed by Wang et al. [31] based on previously assessment tools [1, 26]. It comprises three dimensions: guilt induction (10 items), love withdrawal (5 items), and authority assertion (3 items). A sample item is: “My parents told me that I should feel guilty when I do not meet their expectations.” The scale has demonstrated good reliability and validity among Chinese early adolescents [30]. Confirmatory factor analysis indicated a good model fit: RMSEA = 0.046, CFI = 0.953, TLI = 0.940, and SRMR = 0.045, supporting satisfactory structural validity of the questionnaire. Mean scores were computed across all items, with higher scores indicating greater parental psychological control. In the present study, Cronbach’s α was 0.90. Across subgroups: Cronbach’s α was 0.90 for both Grade 5 and Grade 6 students, and 0.88 for boys and 0.92 for girls.

Teacher–student relationship (TSR)

TSR was measured using a 5-item subscale (rated on a 5-point scale; 1 = not at all true, 5 = very true) from the School Adjustment Scale, developed by a Chinese scholar [13]. This scale has demonstrated high reliability and validity in studies involving Chinese early Chinese adolescence [20, 28]. Sample items include: “I try to avoid contact with the teacher as much as possible.” Confirmatory factor analysis indicated a good model fit: RMSEA = 0, CFI = 1.00, TLI = 1.005, and SRMR = 0.008, supporting satisfactory structural validity of the questionnaire. To facilitate analysis and interpretation, negative item scores were reversed (e.g., a score of 1 became 5) and the mean scores of the 5 items were calculated. Higher scores indicated better TSR. Cronbach’s α was 0.80 in this sample. Across subgroups, Cronbach’s α was 0.78 for Grade 5 and 0.80 for Grade 6, and 0.80 for both boys and girls.

Shyness

The Shyness Scale developed by Cheek and Buss [7] was used to assess children’s shyness. This scale has demonstrated reliable psychometric properties in Chinese middle and upper-grade primary students [11]. It is a unidimensional scale consisting of 13 items (e.g., “I feel restricted in social situations”) and uses a 5-point scale (1 = strongly disagree, 5 = strongly agree). Confirmatory factor analysis indicated a good model fit: RMSEA = 0.046, CFI = 0.948, TLI = 0.934, and SRMR = 0.038, supporting satisfactory structural validity of the questionnaire. Mean scores were computed across all items, with higher scores reflecting greater levels of shyness. Cronbach’s α was 0.85 in this sample. Across subgroups, Cronbach’s α was 0.84 for Grade 5 and 0.87 for Grade 6, and 0.83 for boys and 0.87 for girls.

Problematic smartphone use (PSU)

The Smartphone Application-Based Addiction Scale (SABAS) [12] was used to assess children’s PSU. The scale consists of six items (e.g., “My smartphone is the most important thing in my life”) rated on a 6-point scale, ranging from 1 (Strongly disagree) to 6 (Strongly agree). The Chinese version of this scale has demonstrated good reliability among Chinese upper primary school students (Grades 4–6) [8]. Confirmatory factor analysis indicated a good model fit: RMSEA = 0.063, CFI = 0.967, TLI = 0.945, and SRMR = 0.029, supporting satisfactory structural validity of the questionnaire. Mean scores were computed across all items, with higher scores indicating greater PSU. In the present study, Cronbach’s α was 0.82. Across subgroups, Cronbach’s α was 0.79 for Grade 5 and 0.84 for Grade 6, and 0.81 for boys and 0.84 for girls.

Analytical strategy

All data analyses were conducted using Mplus 8.3.

First, descriptive statistics and bivariate correlations among all study variables were computed.

Second, item parceling was applied to the shyness construct following Little et al. [21]. Given that the shyness scale is unidimensional (confirmed by CFA: RMSEA = 0.046, CFI = 0.948, TLI = 0.934, SRMR = 0.038), consists of 13 items, and the study focuses on structural relations among latent constructs rather than item-level analysis, parceling was deemed appropriate. This approach reduces item-level random error and improves model stability without altering the substantive interpretation of the findings. For teacher–student relationship (TSR; 5 items) and problematic smartphone use (PSU; 6 items), the original items were used without parceling, as the number of indicators was sufficient for model identification and stable parameter estimation.

Third, a structural equation model was constructed with parental psychological control and teacher–student relationship as exogenous latent variables, and shyness and problematic smartphone use as endogenous latent variables. Based on prior research indicating that problematic smartphone use may vary by gender and developmental stage and may also be associated with family socioeconomic background (e.g., parental education) [5, 19], gender, grade, and parental education were included as covariates in the SEM analyses. Given the nested structure of the data (students nested within classes), intraclass correlation coefficients (ICCs) were first calculated to assess the independence of observations and to determine whether a single-level or multilevel model was appropriate. Model fit for the baseline model was evaluated using conventional indices, including CFI, TLI, RMSEA, and SRMR.

Fourth, the Latent Moderated Structural Equations approach was employed to examine the moderating effect of parental psychological control on the mediation pathway. Model evaluation followed a two-step procedure. First, adequate fit of the baseline model without the interaction term was established. Second, a log-likelihood ratio test was conducted to compare the baseline model with the model including the latent interaction term (parental psychological control × teacher–student relationship). A significant likelihood ratio test combined with a reduction in AIC indicated that the latent moderation model provided superior fit relative to the baseline model. For significance testing of effects, the Delta method was used to estimate the Index of Moderated Mediation and the conditional indirect effects at high and low levels of the moderator. Effects were considered statistically significant when p < 0.05.

Results

Preliminary analyses

Harman’s single-factor test was conducted to assess potential common method bias. The results revealed that the first factor explained 22.05% of the total variance, which is below the commonly used threshold of 40%. In addition, a single-factor confirmatory factor analysis (CFA) model was tested in which all measurement items were specified to load on one latent factor. The single-factor model showed poor fit to the data (CFI = 0.509, TLI = 0.484, RMSEA = 0.085, SRMR = 0.109), suggesting that common method bias was unlikely to substantially account for the observed covariance among the study variables.

Table 1 presents the descriptive statistics and correlations of the variables. The correlation analysis revealed that parental psychological control, teacher–student relationship, shyness, and problematic smartphone use were all significantly correlated with one another.

Table 1.

Descriptive statistics and correlations (N = 1,176)

Variables	M ± SD	1	2	3	4	5	6
1.Gender	1.500 ± 0.500	1
2.Grade	1.446 ± 0.497	−0.005	1
3.PEL	3.556 ± 1.299	−0.037	0.030	1
4.PPC	2.695 ± 0.859	−0.095**	0.004	0.022	1
5.TSR	4.067 ± 0.848	0.067*	−0.093**	−0.088**	−0.302**	1
6.Shyness	2.431 ± 0.773	0.049*	−0.042	0.112**	0.229**	−0.483**	1
7.PSU	2.309 ± 1.059	−0.094**	0.145**	0.096**	0.311**	−0.456**	0.291**

Gender and Grade are dummy variables (boys = 1; girls = 2; grade 5 = 1; grade 6 = 2). PEL = parental educational level (1 = master’s degree; 2 = bachelor’s degree; 3 = associate degree; 4 = senior high school; 5 = junior high school; 6 = elementary school)

PPC Parental psychological control, TSR Teacher–student relationship, PSU Problematic smartphone use

^* p < 0.05(two-tailed). ^** p < 0.01(two-tailed)

Given the nested structure of the data (students nested within classes), intraclass correlation coefficients (ICCs) were calculated using Mplus 8.3 prior to conducting structural equation modeling (SEM). The ICC values for all variables ranged from 0.003 to 0.043, well below the commonly used threshold of 0.05. Notably, the between-class variance for the key variable parental psychological control was close to zero. These findings suggest minimal clustering effects at the class level and adequate independence of observations. Therefore, a single-level SEM was deemed appropriate for subsequent hypothesis testing.

Testing for the mediation model

The present study used Mplus 8.3 to examine the direct effect of parental psychological control on problematic smartphone use, controlling for parental education, gender, and grade. The results indicated that parental psychological control significantly and positively predicted problematic smartphone use (β = 0.335, SE = 0.038, p < 0.001).

Subsequently, after controlling for parental education, gender, and grade, the mediating role of shyness in the association between parental psychological control and problematic smartphone use was tested. The mediation model demonstrated acceptable fit indices: CFI = 0.947, TLI = 0.933, RMSEA = 0.053, and SRMR = 0.042. As shown in Table 2, parental psychological control positively predicted shyness (β = 0.304, SE = 0.035, p < 0.001), and shyness positively predicted problematic smartphone use (β = 0.291, SE = 0.039, p < 0.001). Parental psychological control also remained a significant positive predictor of problematic smartphone use (β = 0.249, SE = 0.041, p < 0.001). The indirect effect accounted for 26.21% of the total effect.

Table 2.

Mediation analysis (N = 1,176)

Variable	Equation 1 (Outcome: shyness)			Equation 2 (Outcome: PSU)
	β	z	95% CI	β	z	95% CI
Gender	0.057	1.867	[−0.005, 0.118]	−0.082**	−2.691	[−0.142, −0.021]
Grade	−0.047	−1.526	[−0.108, 0.013]	0.172**	5.937	[0.117, 0.229]
PEL	0.104**	3.392	[0.045, 0.165]	0.062*	2.209	[0.005, 0.106]
PPC	0.304**	8.214	[0.231, 0.378]	0.249**	6.206	[0.171, 0.327]
shyness				0.291**	7.539	[0.216, 0.368]
R2	0.108			0.228

PPC parental psychological control, PSU problematic smartphone use, PEL parental educational level; Bootstrap sample size = 5000. CI Confidence interval

^*p < 0.05(two-tailed). ^**p < 0.01(two-tailed). All βs are standardized coefficients

Testing for moderated mediation

To examine the moderating role of teacher–student relationship in the association between parental psychological control and problematic smartphone use, two latent variable models were estimated. First, the main effect of teacher–student relationship was added to the mediation model as a baseline model. Because latent interaction models involve nonlinear effects, parameters were estimated using maximum likelihood with robust standard errors (MLR) based on numerical integration, and model comparisons were primarily conducted using the Akaike Information Criterion (AIC). The baseline model yielded an AIC of 56,681.396. After including the latent interaction term between parental psychological control and teacher–student relationship, Model 2 (the moderated mediation model) produced an AIC of 56,676.616. Compared with the baseline model, Model 2 showed a reduction of 4.78 in AIC, suggesting better relative model fit and supporting the inclusion of the latent interaction term.

Controlling for gender, grade, and parental education, path analysis results (see Fig. 2 and Table 3) showed that parental psychological control significantly and positively predicted shyness (β = 0.091, SE = 0.038, p = 0.015), and shyness significantly and positively predicted problematic smartphone use (β = 0.327, SE = 0.041, p < 0.001). Moreover, the latent interaction between parental psychological control and teacher–student relationship significantly predicted shyness (β = 0.072, SE = 0.032, p = 0.026). The index of moderated mediation was significant (Index = 0.027, SE = 0.013, p = 0.038), indicating that the indirect effect of parental psychological control on problematic smartphone use through shyness significantly varied across levels of teacher–student relationship.

Fig. 2.

Structural model of the moderated mediation analysis. Note. PPC = parental psychological control; TSR = teacher–student relationship; SHY = shyness; PSU = problematic smartphone use. The model controls for gender, grade, parental educational level. Standardized coefficients are shown. ^*p < 0.05(two-tailed). ^**p < 0.01(two-tailed)

Table 3.

Latent moderated mediation analysis (N = 1,176)

	β	z	SE	95% CI
Mediator variable model predicting Shyness
Gender	0.082**	3.088	0.027	[0.029, 0.135]
Grade	−0.095**	−3.588	0.026	[−0.146, −0.044]
PEL	0.066*	2.411	0.027	[0.013, 0.119]
PPC	0.091*	2.427	0.038	[0.017, 0.165]
TSR	−0.604**	−17.992	0.034	[−0.671, −0.537]
PPC × TSR	0.072*	2.231	0.032	[0.009, 0.135]
Dependent variable model predicting PSU
Gender	−0.081*	−2.675	0.030	[−0.140, −0.022]
Grade	0.173**	5.888	0.029	[0.116, 0.230]
PEL	0.057*	2.009	0.028	[0.002, 0.112]
PPC	0.261**	6.060	0.043	[0.177, 0.345]
shyness	0.327**	7.900	0.041	[0.247, 0.407]

Conditional effect	Effect	SE	LLCI	ULCI
M − 1SD	0.007	0.020	−0.032	0.046
M	0.034	0.015	0.005	0.063
M + 1SD	0.062	0.019	0.023	0.099
Index of moderated mediation	0.027	0.013	0.002	0.052

Conditional indirect effects were tested via the Delta method, as numerical integration in latent moderated mediation precludes the use of bootstrapping

PPC parental psychological control, TSR teacher–student relationship, PSU problematic smartphone use, PEL parental educational level

CI = confidence interval. ^*p < 0.05(two-tailed). ^**p < 0.01(two-tailed). All βs are standardized coefficients

Conditional indirect effect analyses (see Fig. 3) further clarified the conditional indirect effects at different levels of teacher–student relationship. When teacher–student relationship was low (− 1 SD), the indirect effect of parental psychological control on problematic smartphone use through shyness was not significant (Effect = 0.007, SE = 0.020, p = 0.721). In contrast, when teacher–student relationship was high (+ 1 SD), the indirect effect was significant (Effect = 0.062, SE = 0.019, p = 0.001). These findings suggest that teacher–student relationship moderate the first stage of the mediation process: among students with higher-quality teacher–student relationship, parental psychological control was more strongly associated with problematic smartphone use through increased shyness.

Fig. 3.

Teacher–student relationship (TSR) moderated the relationship between parental psychological control (PPC) and shyness (SHY)

Discussion

Within the IT-CPU framework, the present study simultaneously examined family (parental psychological control), school (teacher–student relationship), and individual (shyness) factors. The findings suggest that problematic smartphone use in early adolescence may be associated with the interplay of multiple developmental systems. Specifically, shyness statistically mediated the association between parental psychological control and problematic smartphone use, while teacher–student relationship moderated the association between parental psychological control and shyness. These results highlight the potential role of shyness as a proximal dispositional factor associated with problematic smartphone use and underscore the joint associations of family and school contextual factors with adolescents’ digital behavior. By integrating factors across family, school, and individual domains, the present study contributes to a more comprehensive ecological understanding of problematic smartphone use in early adolescence and provides insights that may inform intervention perspectives targeting mobile media use.

It is also noteworthy that the sample in this study was drawn from Shandong, China. Within the Chinese cultural context, parents often assume strong responsibilities for guiding and regulating children’s development; consequently, a certain degree of controlling parenting behavior is not uncommon in family socialization practices. Previous research has indicated that in Chinese culture, parental control is often linked to the parenting concepts of guan (discipline) or training and is typically accompanied by high levels of parental involvement in children’s development [6, 31]. Meanwhile, schools in eastern China generally operate within a relatively strong evaluation-oriented and academically competitive environment. As a result, teacher–student relationship may provide emotional and instrumental support while also involving elements of evaluative interaction. Therefore, the interpretation of the present findings should be situated within the broader context of Chinese parenting norms and school culture.

The mediating role of shyness

The present study found that, after controlling for gender, grade level, and parental education, parental psychological control was positively associated with early adolescents’ problematic smartphone use, consistent with prior findings from Chinese adolescent samples [34]. Within the family–school socialization context examined in this study, psychological control is often conveyed through relational interactions that communicate norms and expectations. This experience of being both cared for and expected to meet standards may co-occur with adolescents’ more frequent reliance on smartphones for daily regulation or emotional distraction.

Furthermore, shyness was found to partially mediate the association between parental psychological control and problematic smartphone use, suggesting an important internal pathway linking family-related pressures to problematic smartphone use. The negative evaluation sensitivity and social withdrawal tendencies characteristic of shyness may be more likely to co-occur with the performance and interaction demands of everyday school and peer contexts. At the same time, with ongoing sociocultural changes in Chinese society, the adaptive meaning of shyness among children and adolescents has been shown to vary across cohorts, and in more recent social contexts it has been increasingly associated with peer rejection, school difficulties, and depressive symptoms [9, 23]. Within this context, shyness may serve as a link through which experiences of parental psychological control and school-based evaluation pressure become reflected in adolescents’ digital behavior. On the one hand, socialization practices associated with psychological control—such as shame-oriented feedback, protective or intrusive intervention, and an emphasis on modesty and restraint—may be related to greater attention to others’ evaluations and more constrained self-expression, which in turn may be associated with heightened sensitivity to negative evaluation in social interactions. On the other hand, in school environments characterized by strong evaluation norms, peer comparison, and performance pressure, shy adolescents may be more likely to prefer interaction formats that are perceived as lower in social risk and higher in controllability. Accordingly, when uncertainty and perceived evaluation pressure in offline social situations are more salient, shy individuals may be more inclined to seek emotional support or a sense of belonging through virtual contexts such as social media and online gaming, which may co-occur with higher levels of problematic smartphone use [17, 32]. This pattern is also consistent with prior research using Chinese samples, which has reported positive associations between parental psychological control and smartphone or Internet addiction, with shyness functioning as a partial mediator [34]. In addition, in family environments characterized by higher levels of control or close monitoring, adolescents with higher levels of shyness may exhibit stronger avoidance tendencies that are associated with higher levels of problematic smartphone use [16].

From a practical perspective, these findings suggest several implications for intervention. In contexts characterized by high family involvement and strong school-based evaluation, reducing parenting practices marked by emotional withdrawal and guilt induction and promoting more autonomy-supportive communication may help mitigate the clustering of risks associated with shyness. At the school level, providing interaction opportunities that foster psychological safety, along with social skills training and emotion regulation education, may support shy adolescents in strengthening offline social coping and media self-regulation capacities.

The moderating role of teacher–student relationship

The present study found that teacher–student relationship significantly moderated the association between parental psychological control and shyness. Specifically, when teacher–student relationship was relatively low, the association between parental psychological control and shyness was not significant; when teacher–student relationship was relatively high, the association became stronger, indicating a reverse buffering pattern. In other words, high-quality teacher–student relationships appeared to be primarily associated with lower levels of shyness in contexts of low parental psychological control, but did not show the expected weakening of the association when parental psychological control was high. This interaction pattern suggests that the role of school-based relational resources may be context-dependent, rather than uniformly functioning as protective buffers across all family contexts.

This finding may be better understood in light of everyday interaction patterns within Chinese schools. From students’ experiential perspective, a “good” teacher–student relationship often implies not only emotional support but also more frequent academic guidance, classroom interaction, and more salient signals of teacher expectations. Thus, high levels of teacher–student relationship may simultaneously involve increased experiences of being supported and being expected or evaluated. In the Chinese family socialization context, parental control practices often coexist with responsibility, involvement, and guidance [6, 31]. For adolescents exposed to higher levels of parental psychological control, evaluative demands and self-monitoring tendencies in family interactions may combine with more salient teacher attention and expectations at school, forming an accumulation of evaluative signals in daily experience. Such cumulative evaluation experiences may be associated with higher levels of shyness in social and expressive situations. Overall, the present findings suggest that the observed reverse buffering pattern may be related to the evaluative signals embedded in the coexistence of support and expectations within Chinese school interactions.

From a practical perspective, while emphasizing the value of teacher support, it is also important to consider how support is conveyed and interpreted. For students from backgrounds characterized by higher parental psychological control, teacher care that is primarily expressed through performance-oriented expectations or public evaluation may be more likely to co-occur with students’ evaluation sensitivity and be associated with higher levels of shyness. Accordingly, school practices may benefit from maintaining supportive relationships while reducing high-threat evaluative signals in daily interactions—for example, by minimizing public comparison, increasing opportunities for low-pressure expression, and strengthening home–school communication to reduce the experiential overlap of high parental psychological control and strong school-based expectations. Given that prior research has linked shyness to risk pathways associated with problematic smartphone use [16, 17, 32, 34], optimizing home–school interaction patterns may also provide an indirect entry point for integrated interventions targeting digital behavior risks.

At the same time, it is important to clarify that the above interpretation represents only one plausible pathway consistent with the observed pattern. Existing evidence indicates that teacher–student relationship is generally positively associated with academic outcomes [18] and has demonstrated buffering roles in some family risk processes [22]. Therefore, the reverse buffering pattern observed in this study is more appropriately characterized as an exploratory finding, and its stability and boundary conditions require further replication and examination. In addition, at least three alternative explanations should be considered. First, construct overlap in measurement: within the context of Chinese classroom interactions, teacher–student relationship scales may simultaneously capture both support/closeness and academic attention or demands (i.e., evaluative interaction), such that high scores do not necessarily represent low-pressure support. Second, contextual specificity: in classroom environments characterized by stronger evaluation norms, more public interaction, or more salient peer comparison, positive interaction may simultaneously increase the salience of evaluation, thereby altering its association with shyness. Third, statistical and methodological factors: given the cross-sectional design and reliance on self-report data, unmeasured confounding variables (e.g., academic performance, classroom climate, intensity of teacher expectations, or reasons for receiving teacher attention) may remain substantial, and reverse explanations—such as the possibility that more shy students are more likely to receive or report teacher attention—cannot be ruled out.

Limitations and future directions

Although the present study integrated family (parental psychological control), school (teacher–student relationship), and individual (shyness) factors within a single model, several limitations should be acknowledged, along with directions for future research.

First, the study employed a cross-sectional design; therefore, the findings reflect statistical associations among variables (including moderation and indirect effects) and are insufficient for establishing temporal order or causal inference. Future research could adopt multi-wave longitudinal designs (e.g., across semesters or academic years) or experimental and intervention-based approaches to examine the stability of associations among parental psychological control, teacher–student relationship, and shyness over time, as well as the potential interaction processes among these factors. Such designs would also allow researchers to test whether the reverse buffering pattern observed in this study can be replicated and to identify the developmental periods during which this pattern may be more pronounced.

Second, the study relied primarily on adolescent self-report data. Although Harman’s single-factor test and single-factor confirmatory factor analysis were conducted as preliminary assessments of common method bias, and the results generally suggested that a single method factor did not account for the majority of covariance—thereby reducing concerns that common method bias substantially distorted the findings—single-source data may still be influenced by common method variance and response styles [25]. Accordingly, it remains difficult to fully rule out the potential influence of common method variance on the estimation of associations and interaction effects, particularly for the anti-buffering pattern identified in this study. Future studies are encouraged to employ multi-informant and multi-method approaches, such as incorporating teacher and parent reports, classroom or teacher–student interaction observations, and peer nominations, as well as integrating objective smartphone usage indicators (e.g., system-recorded screen time, application usage logs, or passive sensing data), in order to enhance construct validity and strengthen the robustness of tests involving key interaction effects.

Third, the sample was drawn from Shandong Province, China, and the observed patterns of association may have been influenced, to some extent, by regional parenting norms and local school evaluation practices. Thus, the findings are primarily applicable to sociocultural contexts characterized by high family involvement and strong school-based evaluation, and caution is warranted when generalizing the findings to other regions or cultural contexts. Future research could replicate the proposed model across different regions, school types, and evaluation environments, and examine the boundary conditions of the findings in cross-cultural samples to enhance external validity.

Finally, future studies may further expand the scope and precision of variable selection and measurement. First, additional factors within the school microsystem could be incorporated, including peer relationships and classroom-level variables (e.g., peer pressure or peer norms, classroom comparison climate, intensity of teacher expectations, and the extent of public evaluation in classroom settings), and multilevel modeling could be used to examine whether these factors jointly constitute boundary conditions for the interaction involving teacher–student relationship. Second, measurement of teacher–student relationship could be further differentiated into specific dimensions (e.g., emotional support versus academic expectations or evaluative interaction components) to test the possibility of construct overlap and to more clearly identify which components of teacher–student interaction are more likely to be associated with shyness under different family conditions. Third, technological and environmental variables (e.g., device ownership, usage contexts, and objectively measured online duration) may also be incorporated to provide a more comprehensive ecological understanding of early adolescents’ digital behavior.

Authors’ contributions

Hong Sun: Conceptualization, Methodology, Writing – review & editing. Yang Yu: Formal analysis, Investigation, Conceptualization, Supervision, Funding acquisition. Mingxiao Zhu: Investigation Xue Chen: Investigation Yanwen Si: Investigation.

Funding

This study is supported by the 2024 National Social Science Fund of China (24BSH117).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Research Ethics Committee of the College of Teacher Education, Taishan University and was conducted in accordance with the principles of the Declaration of Helsinki.

Prior to data collection, written informed consent was obtained from the students, their parents, and teachers.

Participation was voluntary, and all responses were kept confidential and used solely for research purposes.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.