Effect of perceived school environment on adolescent depression: the moderating role of intergenerational interactions

Abstract

Objective

To examine the impact of perceived school environment on adolescent depression, and the moderating effects of intergenerational interactions.

Methods

Based on the 2018 adolescent questionnaire data from the China Family Panel Studies (CFPS), regression analysis was conducted using Stata 17.0 on a sample of 1058 individuals, focusing on depression status, perceived school environment, and intergenerational interactions.

Results

(1) Adolescent depression was significantly associated with lower parent-child communication (β= -0.06, p < 0.001), lower school satisfaction (β=-0.10, p < 0 0.001), higher academic pressure (β = 0.35, p < 0.001), and more frequent externalizing problem behaviors (β = 1.29, p < 0.001). (2) Intergenerational interaction played an important buffering role: behavioral evaluation negatively moderated the effect of academic pressure on depression (β = -0.05, p < 0.05), while parent-child communication negatively moderated the effect of externalizing problem behaviors on depression (β = -0.06, p < 0.01).

Conclusion

Perceived school environment is an important factor affecting adolescent depression, and intergenerational interaction (especially behavioral evaluation and parent-child communication) can effectively buffer the negative effects of some risk factors (academic pressure, externalizing problem behaviors).

Introduction

During the period of social transformation, rapid economic development, cultural diversity, and environmental uncertainties can profoundly impact the mental health of adolescents [1]. The prevalence of depressive tendencies among adolescents is showing a gradual increase [2]. According to data released by the World Health Organization, approximately one in seven 10–19-year-olds experiences a mental disorder, accounting for 15% of the global burden of disease in this age group. Depression, anxiety and behavioral disorders are among the leading causes of illness and disability among adolescents [3]. The 2024 Report on the Mental Health of Rural Students in Underdeveloped Regions of China further indicates that 21.5% of rural students are at risk of mild depression, while 8.1% are at high risk of depression [4]. These data underscore the urgency of studying adolescent depression.

Depression encompasses depressive tendencies and major depressive disorder. The former refers to depressive symptoms exhibited by an individual at certain moments in daily life, while the latter, as a typical representative of depressive disorders, is also termed major depressive disorder. Its core symptoms include persistent low mood and anhedonia [5]. Since depressive disorders with clinical manifestations are difficult to improve through enhanced social resources, this study focuses on depressive tendencies as an emotional state and a dimension of mental health [6], rather than depression meeting clinical diagnostic criteria.

Research demonstrates that stress is a significant risk factor for depression [7]. According to the Transactional Model of Stress proposed by Lazarus and Folkman, stress does not arise solely from external events themselves, but rather from the dynamic and complex interaction between the individual and their environment. When facing stressful situations, individuals first engage in primary assessment to internally evaluate the nature and threat level of the stressor. If perceived as a threat, they proceed to a secondary assessment phase, evaluating available coping resources. This series of assessments directly influences the coping strategies ultimately chosen by individuals, thereby playing a critical role in their mental health [8–10].

Therefore, individuals’ subjective interpretation of stressors and their coping strategies play a central role in the process of stress transforming into psychological distress. This also implies that the causes of depression involve not only the abundance of external social resources but are also closely related to the individual’s internal psychological evaluation mechanisms.Especially for adolescents in the critical stage of physical and mental development, their emotional and cognitive functions are still evolving, and their ability to perceive and cope with stress is still forming. Thus, the occurrence of adolescent depression cannot be simply attributed to a single stressful event or lack of social resources, but should be understood within the broader ecosystem in which they are situated.

Bronfenbrenner’s ecological system theory provides a framework for conducting research on adolescent depression. This theory posits that individual development is embedded within a series of interconnected environments. These systems, from the innermost to the outermost, include the microsystem, mesosystem, exosystem, and macrosystem [11, 12]. Among these, the most closely related to the individual is the micro system, which primarily includes the family, school, and peer groups [13]. Existing research has identified numerous risk or protective factors within families and schools that influence adolescent depression [14–16].

As the core unit of adolescent socialization, the family subtly shapes young people’s personality and behavioral patterns through daily parenting and parent-child interactions, playing an irreplaceable role in preventing and intervening adolescent depression. A review of intergenerational family interaction factors influencing adolescent depression reveals three key elements: parental educational involvement, parental behavioral evaluation of children, and parent-child communication. Among these, parental educational involvement enhances students’ emotional functioning, with parental engagement showing positive correlations with life satisfaction and self-esteem levels, while exhibiting negative correlations with depression and anxiety levels [17]. Negative parental evaluations are established risk factors for various mental health issues, including depression [18]. Maternal evaluations directly influence adolescent daughters’ emotional states: praise significantly boosts positive emotions, whereas criticism intensifies negative emotions [19]. Parent-child communication significantly predicts depression in left-behind adolescents. High communication frequency between parents and children can mitigate the impact of primary caregivers’ parenting styles on adolescent depression, thereby protecting their mental health [20]. These findings collectively confirm a significant association between intergenerational interaction and adolescent depression. Based on this evidence, this study proposes Hypothesis 1: Intergenerational interaction is associated with adolescent depression levels, specifically:

• Hypothesis 1a: More positive parental educational involvement is associated with lower levels of adolescent depression.

• Hypothesis 1b: Higher parental behavioral evaluation is associated with lower depression levels in adolescents.

• Hypothesis 1c: More positive parent-child communication is associated with lower depression levels in adolescents.

As a crucial domain in adolescents’ daily lives, the school system exerts its influence primarily through dimensions such as academic pressure, teacher-student relationships, peer relationships, and overall satisfaction with the school. School satisfaction is regarded as a significant indicator of adolescent well-being [21], and has been shown to negatively predict adolescent depression [22]. Students with positive teacher-student relationships tend to exhibit relatively fewer depression, and stronger perceived teacher support is generally associated with milder depression [23]. Beyond interpersonal networks, academic pressure is a risk factor that exacerbates student anxiety [24]. Furthermore, poor peer relationships among adolescents in school can trigger externalizing problem behaviors [25], which in turn severely impact their interpersonal relationships and academic development, ultimately leading to a series of issues such as depression and anxiety [26]. Based on this, grounded in students’ subjective experiences, this study proposes Hypothesis 2: adolescents’ perceived school environment is associated with their depression levels, specifically encompassing the following three aspects:

• Hypothesis 2a: Higher school satisfaction is associated with lower depression levels in adolescents.

• Hypothesis 2b: Higher academic pressure is associated with an increase in adolescent depression levels.

• Hypothesis 2c: More severe externalizing problem behaviors are associated with higher levels of depression in adolescents.

Building upon ecological system theory, the theory of overlapping spheres of influence further posits that family, school, and community are the three primary environments influencing student development.These spheres independently and collectively shape children’s learning and development through shared experiences, values, and behaviors [27]. This suggests that the family and school will form an overlapping sphere where home-school collaboration is crucial. Based on longitudinal survey data, Rothon et al. examined the relationship between school-family interaction and adolescent mental health in the UK. The results showed that higher parental involvement in school activities was associated with better adolescent mental health outcomes and fewer depressive symptoms [28]. Furthermore, Stromshak et al. found that when the entire family system participated in school-based intervention programs, adolescents exhibited fewer antisocial behaviors and substance abuse issues, which indirectly reflects the potential role of family-school collaboration in mitigating adolescent depression risks [29]. In summary, family-school interaction not only contributes to improving adolescents’ overall mental health but may also exert a positive impact on depression alleviation. Accordingly, this study proposes the following hypotheses:

• Hypothesis 3: Intergenerational interaction moderates the effect of perceived school environment on adolescent depression.

In conclusion, existing research has mostly focused on the relationship between single factors, such as intergenerational interaction or school environment with adolescent depression. Although the overlapping influence theory highlights the importance of collaborative effects between family and school, relevant empirical studies remain relatively scarce. Currently, few studies have concretely developed an integrated model incorporating both “intergenerational interaction” (a core family element) and “perceived school environment” (a core school element), systematically examining their interactive effects (i.e., moderating roles) on adolescent depression. Therefore, this study intends to take adolescent depression as the dependent variable, intergenerational interaction and perceived school environment as the main independent variables, and intergenerational interaction as the moderating variable, to deeply explore the relationship among them.

Methods

Participants

The data used in this study comes from the China Family Panel Studies (CFPS) conducted by the Institute of Social Science Survey of Peking University. The survey longitudinally tracks and collects data at three levels: individual, household, and community, covering a wide range of topics such as village/neighborhood profiles, family relationships, and population mobility [30]. The CFPS questionnaire consists of four modules: community, household, adult, and children. The children’s questionnaire targets individuals under the age of 16. For children under 10 years old, the questionnaire is completed by parents (or guardians), while for children aged 10 and above, it includes proxy responses from parents (or guardians) as well as self-reported responses from the children. This study retained self-reported data from children aged 10–16 in the CFPS 2018 child database. After data processing and the exclusion of missing values, the final sample size included in the analysis was 1,058.

Measures and procedure

The survey covered basic information (gender, age, urban/rural residence status, parental age, and academic performance reflected by the composite scores of Chinese and mathematics tests in the CFPS questionnaire [31]), depression, intergenerational interactions (parental educational involvement, parental behavioral evaluation, parent-child communication), and perceived school environment (school satisfaction, academic stress, and externalized problem behaviors).

Center for epidemiologic studies depression scale (CES-D)

This study employs the Chinese version of the 8- item Center for Epidemiologic Studies Depression Scale (CES-D 8) to assess depressive symptoms [32, 33]. Respondents are asked “How often you have felt this way during the last week” with “I am in a low spirit,” “everything felt effortful,” and “poor sleep quality.” A 4-point Likert scale was used (1 = rarely, 2 = sometimes, 3 = occasionally, 4 = mostly), with a higher score indicating more severe depressive symptoms [34]. The Cronbach’s α coefficient for this scale in the study was 0.71.

Educational involvement

Measured the frequency of parental behaviors over the past 12 months, such as “sacrificing TV time for the child’s studies,” “discussing school matters with child,” “ensuring homework completion,” “checking child’s homework,” “restricting TV watching,” and “restricting types of TV programs” [35]. A 5-point Likert scale was used (from 1 “Never” to 5 “very often”, with higher scores indicating more positive educational involvement). The Cronbach’s α coefficient for this scale was 0.68.

Behavioral evaluation

Assessed evaluations based on daily observations of adolescents, including 7 items such as “studies diligently,” “checks homework after completion,” and “finishes homework before playing” [36]. A 4-point Likert scale was used, with higher scores indicating more positive behavioral evaluations. The Cronbach’s α coefficient for this variable was 0.85.

Parent-child communication

Covered two aspects: “the frequency of monthly heart-to-heart conversations with parents” and “the frequency of monthly arguments with parents“ [37]. The scores for argument frequency are reverse-coded and added to the conversation frequency. A higher total score indicates more positive parent-child communication.

School satisfaction

Included respondents’ satisfaction with their school, homeroom teacher, Chinese teacher, math teacher, and English teacher [38]. A 5-point Likert scale was used (from 1 “very dissatisfied” to 5 “very satisfied”). The reliability coefficient for this variable was 0.80.

Academic pressure

Measured using the item “How much academic pressure do you feel“ [39]. A 5-point Likert scale was used, with higher scores indicating greater academic pressure.

Externalizing problem behaviors

The following six items were used for measurement, such as “frequently arguing with peers,” “causing trouble by interrupting others due to excessive talking,” and “causing trouble by fighting with peers,” etc. A 4-point Likert scale was employed, with higher scores indicating more severe externalizing problem behaviors [40]. The reliability coefficient of this variable was 0.79.

Data analysis

Data processing was performed using Stata 17.0 statistical software.Firstly, descriptive statistic was employed to characterize the dependent variable independent variable, and control variables.

Secondly, we conducted linear regression analysis using the Ordinary Least Squares (OLS) method. The specific analytical strategy was as follows: In the first place, based on the basic control variables, we separately added the intergenerational interaction variables for educational involvement, behavioral evaluation, and parent-child communication, corresponding to Model 1a, Model 1b, and Model 1c, respectively. On this basis, a combined model for intergenerational interaction was constructed by including all three variables into one model, referred to as Model 1. This model aims to examine the correlation between the three independent variables and adolescent depression. Next, based on the basic control variables, we separately added the three variables for perceived school environment: school satisfaction, academic pressure, and externalizing problem behavior, corresponding to Model 2a, Model 2b, and Model 2c, respectively. Simultaneously, Model 2 was constructed to comprehensively examine the relationship between the three perceived school environment variables and depression. Lastly, we integrated six variables related to intergenerational interaction and perceived school environment into the comprehensive Model 3. After regression, the Variance Inflation Factor (VIF) test was conducted to ensure that the models did not suffer from severe multicollinearity issues.

Finally, we performed mean-centering on the raw scores of three variables measuring perceived school environment and three variables measuring intergenerational interaction to reduce multicollinearity and facilitate result interpretation. Using the centered variables, we constructed nine corresponding interaction terms (3 perceived school environment variables × 3 intergenerational interaction variables). After controlling for main effects, we sequentially incorporated these interaction terms into the regression model to establish a moderation effect model for testing.

Results

Descriptive statistics

To examine the structural validity of “intergenerational interaction” in this study and extract its latent factor structure for constructing measurement indicators in subsequent analyses, we first conducted exploratory factor analysis on this variable (see Table 1). Feasibility tests revealed that the KMO value for the exploratory factor analysis sample met the acceptable criterion (KMO = 0.82), and Bartlett’s test of sphericity was statistically significant (χ²= 3822.14, P < 0.001), suggesting that the data were suitable for factor analysis. Subsequently, the principal component analysis and Promax oblique rotation were used to analyze the data.Given that the factor number of intergenerational interaction is 3, this study set the factor number to 3 for exploratory factor analysis, and all items must meet the following conditions:①Item commonality > 0.30; ②The absolute difference of item cross-loadings between two factors > 0.05; ③Item loadings in the structural matrix > 0.50. The results showed that after rotation, the item communality ranged from 0.30 to 0.55, the item loadings ranged from 0.51 to 0.72, 2, and the difference of cross load and absolute value between two factors in each item was exceeded 0.05. Thus, the exploratory factor analysis revealed that the education involvement dimension comprised six items, with factor loadings ranging from 0.51 to 0.70, accounting for 15.31% of the variance; the behavioral evaluation dimension included 7 items, with factor loadings between 0.52 and 0.71, accounting for 21.26% of the variance; and the parent-child communication dimension consisted of two items, with factor loadings ranging from 0.71 to 0.72, accounting for 7.31% of the variance.

Table 1.

Item loadings and commonality of exploratory factor analysis for the intergenerational interaction scale

Education involvement dimension			Behavioral evaluation dimension			Parent-child communication dimension
Items	Loadings	Commonality	Items	Loadings	Commonality	Items	Loadings	Commonality
Checking child’s homework	0.70	0.49	Maintains focus during tasks	0.71	0.51	The frequency of monthly arguments with parents	0.72	0.55
Restricting TV watching	0.66	0.38	Completes tasks once started	0.70	0.48	The frequency of monthly heart-to-heart conversations with parents	0.71	0.52
Ensuring homework completion	0.66	0.44	Checks homework after completion	0.70	0.50
Restricting types of TV programs	0.61	0.46	Studies diligently	0.70	0.48
Discussing school matters with child	0.57	0.34	Keeps belongings organized	0.66	0.45	-	-	-
Sacrificing TV time for the child’s studies	0.51	0.30	Finishes homework before playing	0.65	0.42	-	-	-
			Adheres to rules and discipline	0.52	0.30	-	-	-
Variance explained 15.31%			21.26%			7.31%

Table 2 presents the descriptive statistics of the study variables. Among all samples, the mean age was 12 years (SD = 1.66). Male participants accounted for 52.46% (n = 555), while females constituted 47.54% (n = 503). Urban residents represented 43.1% (n = 456), and rural residents comprised 56.9% (n = 602). The mean score for academic performance was 35.73 (SD = 9.18). The mean score on the CES-D scale was 11.61 (SD = 2.98).

Table 2.

Descriptive statistical

Variable	Mean	SD	Minimum	Maximum	%	Sample Size
Individual Level
Age	12	1.66	10.00	15.00		1058
Gender						1058
Female					47.54	503
Male					52.46	555
Urban-Rural Distribution						1058
Rural					56.9	602
Urban					43.1	456
Academic performance	35.73	9.18	0.00	53.00		1058
Depression Level	11.61	2.98	8.00	26.00		1058
Intergenerational Interactions
Paternal age	41.61	5.37	29.00	64.00		1058
Maternal age	39.48	5.33	29.00	64.00		1058
Educational Involvement	19.92	4.54	6.00	30.00		1058
Behavioral evaluation	20.01	3.34	9.00	28.00		1058
Parent-Child Communication	50.63	5.26	10.00	80.00		1058
Perceived School Environment
School Satisfaction	21.19	3.81	5.00	25.00		1058
Academic Pressure	2.91	1.13	1.00	5.00		1058
Externalizing Problem Behaviors	2.08	0.66	1.00	5.00		1058

In terms of intergenerational interaction, the average age of fathers was 41.61 years (SD = 5.37), and that of mothers was 39.48 years (SD = 5.33). The average score for parental educational involvement was 19.92 (SD = 4.54), and the average score for behavioral evaluation was 20.01 (SD = 3.34). The average frequency of monthly conversations in parent-child communication was 50.63 (SD = 5.26).

Regarding perceived school environment, the average score of school satisfaction was 21.19 (SD = 3.81), the average score of academic pressure was 2.91 (SD = 1.13), and the average score of externalizing problem behavior was 2.08 (SD = 0.66).

Multiple regression results

After controlling for gender, age, urban-rural residence status, and academic performance, we progressively analyzed the relationships between intergenerational interaction, perceived school environment, and adolescent depression. Following the regression analysis, a multicollinearity test was conducted on the data.The results showed that the variance inflation factor (VIF) was significantly below 10 (maximum VIF = 3.18, average VIF = 1.54), indicating no collinearity issues with the independent variables. The detailed analytical results are presented in Table 3.

Table 3.

OLS regression model results

Variable	Model 1a	Model 1b	Model 1c	Model 1	Model 2a	Model 2b		Model 2c	Model 2	Model 3
	Intergenerational Interactions				Perceived School Environment					Full Model
Intergenerational Interactions
Educational Involvement	−0.03			−0.02						−0.02
Behavioral evaluation		−0.07*		−0.05						0.00
Parent-Child Communication			−0.08***	−0.08***						−0.06***
Perceived School Environment
School Satisfaction					−0.14***				−0.10***	−0.10***
Academic Pressure							0.50***		0.36***	0.35***
Externalizing Problem Behaviors								1.50***	1.31***	1.29***
F-statistic	11.06***	11.62***	14.78***	15.88***	14.06***		9.37***	7.71***	9.92***	13.67***
Sample Size (N)	1058	1058	1058	1058	1058		1058	1058	1058	1058
R2	0.04	0.05	0.06	0.07	0.07		0.08	0.14	0.18	0.19
adj_R2	0.04	0.04	0.06	0.06	0.07		0.07	0.14	0.17	0.18
F-statistic	6.61	7.30	9.84	8.12	11.79		12.32	24.63	24.94	20.37

*p<0.05, **p<0.01, ***p<0.001. The following is the same

Model 1 in Table 3 examines the relationship between intergenerational interaction and adolescent depression. After controlling for other variables, Model 1a indicates no significant correlation between parental educational involvement and adolescent depression levels (β= −0.03). Model 1b shows a negative correlation between behavioral evaluation and adolescent depression levels (β= −0.07, p < 0.05), with each one-point increase in behavioral evaluation resulting in a 0.07-point decrease in depression scores. Model 1c reveals a negative correlation between parent-child communication and depression scores (β= −0.08, p < 0.01), with a one-point increase in parent-child communication resulting in a 0.08-point decrease in depression scores. When three intergenerational interaction variables were simultaneously included in Model 1, the correlations between educational involvement and behavioral evaluation became non-significant, whereas the negative correlation of parent-child communication remained significant (β = −0.08, p < 0.001). According to the coefficient of determination (R²), the model 1 can explain approximately 7.00% of the total variance in adolescent depression scores.

Model 2 reveals the correlation between perceived school environment and adolescent depression. The estimation results of Model 2a indicate a significant negative correlation between school satisfaction and depression (β= −0.14, p < 0.001), suggesting that a 1-point increase in satisfaction leads to a 0.14-point decrease in depression scores. Model 2b demonstrates a significant positive correlation between academic stress and depression (β = 0.50, p < 0.001), with each 1-point increase in academic stress resulting in a corresponding rise of approximately 0.50 points in depression scores. Model 2c demonstrates a significant positive correlation between externalizing problem behaviors and depression (β = 1.50, p < 0.001), with higher levels of problem behaviors correlating with higher depression levels. When Model 2 incorporates all three perceived school environment predictors, the estimated effect remains statistically significant despite a slight reduction in magnitude. This model exhibits superior fit, accounting for approximately 18.00% of total variance (R²=0.18).

Model 3 presents the results of the multiple regression analysis incorporating all variables. In the perceived school environment, all three independent variables (school satisfaction, academic stress, and externalizing problem behaviors) show significant correlations with adolescent depression levels (p < 0.001). Regarding intergenerational interactions, after controlling for school environment factors, only parent-child communication maintains a significant negative correlation (β = −0.06, p < 0.001), while the effects of educational involvement and behavioral evaluation become non-significant.In terms of model explanatory power (R²), Model 3 (R² = 0.19) showed an improvement compared to Model 1 (R² = 0.07) and Model 2 (R² = 0.18), with an overall explanatory power of approximately 19.00%. Overall, the full model (Model 3), which includes all predictive variables, demonstrated the best predictive performance.

To address the potential heteroscedasticity in the model, we further employed robust regression (robust test) for robustness testing (see Table 4). The robustness results are largely consistent with the empirical findings, indicating that the model exhibits a certain degree of stability.

Table 4.

Robustness test results based on Robust regression

Variable	Model 1a	Model 1b	Model 1c	Model 1	Model 2a	Model 2b	Model 2c	Model 2	Model 3
	Intergenerational Interactions				Perceived School Environment				Full Model
Intergenerational Interactions
Educational Involvement	−0.03			−0.02					−0.02
Behavioral evaluation		−0.07*		−0.05					0.00
Parent-Child Communication			−0.08***	−0.08***					−0.06***
Perceived School Environment
School Satisfaction					−0.14***			−0.10***	−0.10***
Academic Pressure						0.50***		0.36***	0.35***
Externalizing Problem Behaviors							1.50***	1.31***	1.29***
F-statistic	11.06***	11.62***	14.78***	15.88***	14.06***	9.37***	7.71***	9.92***	13.67***
Sample Size (N)	1058	1058	1058	1058	1058	1058	1,058	1,058	1,058
R2	0.04	0.05	0.06	0.07	0.07	0.08	0.14	0.18	0.19
adj_R2	0.04	0.04	0.06	0.06	0.07	0.07	0.14	0.17	0.18
F-statistic	6.02	6.35	9.43	8.00	10.44	11.09	21.36	23.11	20.38

Moderation effect analysis

The results of the moderation effect analysis (see Table 5) indicate that two interaction terms had a statistically significant impact on adolescent depression: academic pressure × behavioral evaluation (β=−0.05,p < 0.05) and externalizing problem behaviors × parent-child communication (β= −0.06, p < 0.01).

Table 5.

Moderating effect results

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
School Satisfaction×Educational Involvement	−0.01
Academic Pressure × Educational Involvement		0.01
externalizing problem behaviors ×Educational Involvement			−0.04
School Satisfaction×behavioral evaluation				−0.01
Academic Pressure × behavioral evaluation					−0.05*
externalizing problem behaviors ××behavioral evaluation						−0.01
School Satisfaction × Parent-Child Communication							0.00
Academic Pressure × Parent-Child Communication								−0.03
externalizing problem behaviors ×Parent-Child Communication									−0.06**
Intercept	11.82***	11.83***	12.73***	11.89***	11.71***	11.83***	11.85***	11.71***	11.89***
Sample Size (N)	1058	1058	1058	1058	1058	1058	1058	1058	1058
R2	0.19	0.19	0.19	0.19	0.19	0.19	0.19	0.19	0.19
adj_R2	0.18	0.18	0.18	0.18	0.18	0.18	0.18	0.18	0.18
F-statistic	18.73	18.90	20.16	18.78	19.36	18.78	18.99	19.47	19.28

All significant interaction coefficients were negative, indicating that behavioral assessment has a buffering effect on the relationship between academic stress and depression. Similarly, parent-child communication also has a buffering effect on the relationship between externalizing problem behaviors and depression.

The following two figures also support the above findings. As shown in Fig. 1, adolescent depression levels increase with rising academic pressure. Under the same academic pressure, the adolescents with lower behavioral evaluation had higher depression level, which indicated that behavioral evaluation moderates the relationship between academic pressure and depression. Specifically, when the behavioral evaluation level is high, the positive predictive effect of academic stress on depression is weak; whereas when the behavioral evaluation level is low, the driving effect of academic stress on depression is stronger.

Fig. 1.

The moderating effect of behavioral evaluation on the relationship between academic pressure and depression

Figure 2 demonstrates that parent-child communication significantly moderates the relationship between externalizing problem behaviors and depression. When the level of parent-child communication is high, the impact of externalizing problem behaviors on depression is relatively weak; when communication is at an average level, its influence gradually becomes apparent; whereas when communication level is low, the negative effect of externalizing problem behaviors on depression is most pronounced.Therefore, positive parent-child communication can effectively mitigate the adverse effects of externalizing problem behaviors on adolescent depression.

Fig. 2.

The moderating effect of parent-child communication on the relationship between externalizing problem behaviors and depression

Discussion

Using a framework combining ecological systems theory and the theory of overlapping spheres of influence, this study leverages national longitudinal household survey data to explore the effect of intergenerational interactions and perceived school environment on adolescent depression.The analysis results demonstrated a significant negative correlation between parent-child communication in intergenerational interactions and adolescent depression, supporting Hypothesis 1c. However, no significant correlation was observed between educational involvement and behavioral evaluation with depression levels, thus Hypotheses 1a and 1b were not supported. Consequently, Hypothesis 1 was partially confirmed.

This finding is notably inconsistent with previous research [41]. Previous studies have found that a moderate negative correlation between parental educational involvement and adolescent depression [42], but this study did not observe such a significant association. This discrepancy may be attributed to the following reasons:

First, the connotation of educational participation is different. Existing research typically defines parental educational involvement as comprehensive parental engagement in their children’s education, encompassing both home-based involvement (such as assisting with homework and reading) and school-based involvement (such as attending parent-teacher meetings and communicating with teachers) [43]. In contrast, this study focuses primarily on educational involvement within the family, such as homework assistance, academic discussions, and media restriction, without including dimensions like parent-school interactions. This may have resulted in a measurement scope that is insufficiently comprehensive.

Second, educational involvement exhibits heterogeneous characteristics. The scale employed in this study encompasses both supportive involvement (e.g., discussing school affairs) and restrictive involvement (e.g., limiting TV watching). Restrictive involvement may reflect parental control and pressure over the academic process, with existing research indicating a positive correlation between this form of participation and adolescent depression, particularly when manifested as excessive control [44]. The effects of supportive and restrictive involvement on depression were in opposite directions, which may partially offset the overall effect. Additionally, the Cronbach’s α coefficient of this scale was 0.68, suggesting that measurement errors might have influenced the significance of the results.

Third, the prevalence of “shadow education” may dilute the impact of parental involvement. In the context of China, the participation rate of “shadow education” (private tutoring) among primary and secondary school students in China is as high as 62.33% [45]. Many parents partially transfer the responsibility of academic supervision to tutoring institutions, thereby weakening the actual role of direct parental educational involvement and potentially diminishing its correlation with adolescent depression.

Furthermore, while previous studies have demonstrated that parental behavioral evaluations significantly influence adolescent depression [18], this study did not observe such an effect.This may be attributed to the fact that adolescents may enter an adaptation phase in their relationship with parents, and dolescents in adolescence may pursue independence and freedom, which may make the relationship between parents and children tend to be tense and antagonistic [46]. This may lead to a weakening of parental influence as adolescents grow older [47]. At this stage, parental evaluation of the child is no longer the sole influencing factor. The role of peer groups becomes increasingly prominent. Due to the “comparative pressure” effect, adolescents who perceive their peers as more capable often experience greater academic pressure and exhibit poorer mental health [48]. Given that adolescents in this study were primarily engaged in school activities and frequent peer interactions, the reduced correlation between behavioral evaluations and depression levels after controlling for the school environment may be attributed to the significant influence of peer effects.

The results confirm Hypothesis 2, that adolescents’ perceived school environment directly predicts depression level. Specifically, school satisfaction was significantly negatively correlated with adolescent depression, confirming Hypothesis 2a; academic pressure and externalizing problem behaviors were significantly positively correlated with adolescent depression, confirming Hypotheses 2b and 2c.This is consistent with the conclusion of previous studies [41]. With the widespread implementation and expansion of compulsory education, schools have evolved into the second most important microsystem after the family, significantly influencing adolescent depression. Through daily activities and learning experiences, adolescents develop diverse social networks. The nature of interactions within these networks plays a crucial role in shaping their mental health and behavioral patterns. A safe, harmonious, and balanced campus environment fosters healthy growth in adolescents [49]. In this study, the perceived school environment variables demonstrated significant relationships with adolescent depression in both sub-models and the full model. This demonstrates that adolescent depression is not merely a family-level issue but requires systemic attention to the school environment. This involves fostering a positive school environment, including healthy teacher-student and peer relationships, appropriately reducing academic pressure, and enhancing adolescents’ sense of school belonging within a harmonious and supportive campus atmosphere.

Further analysis of moderating effects shows that behavioral evaluation in intergenerational interaction can effectively moderate the relationship between academic stress and depression, and parent-child communication can also effectively moderate the relationship between externalized problem behaviors and depression. Hypothesis 3 is partially confirmed.This finding further highlights the importance of family-school collaboration. According to the theory of overlapping spheres of influence, if the three key stakeholders—family, school, and community—can engage in high-quality positive interactions, ensuring children receive consistent information across different environments, it will significantly promote the healthy development of children and adolescents [50]. Based on our research findings, schools, as a key setting influencing adolescent mental health, should play a leading role in family-school collaboration. Schools can establish collaborative mechanisms with families through parent-teacher conferences, thematic lectures, and online interactive groups, maximizing the overlap of influence spheres to prevent the occurrence of adolescent depression.

Limitations

This study has several limitations. First, the cross-sectional nature of the study design restricts causal inference.The cross-sectional design of this study means that only the statistical correlation between parent-child communication, school satisfaction, academic pressure, externalizing problem behaviors, and adolescent depression can be determined, but not the causal relationship. Particular attention should be paid to the fact that depression tendencies themselves may significantly influence an individual’s perception of stressful events, evaluation of the school environment, and interaction patterns with parents, potentially establishing a clear reverse causal relationship. Therefore, in the analysis and discussion of this study, causal statements have been avoided as much as possible, with emphasis placed on the associative relationships among variables. Furthermore, variables may interact with each other—for instance, school satisfaction could be influenced by academic pressure and externalized problem behaviors. Statistical methods are inadequate for analyzing these complex interrelationships, as they fail to adequately account for dependencies, reciprocity, and indirect effects among variables, which constitutes one of the limitations of this study.

Second, the measurement scales have certain limitations, which may affect the interpretation of the results.The school satisfaction scale is primarily designed for evaluating the teaching staff, and fails to encompass multiple dimensions of satisfaction such as teaching quality, curriculum content, school facilities, peer relationships, and management fairness.The Cronbach’s α coefficient of the Educational Participation Scale was 0.68, which is within the acceptable critical level, and its reliability may have a certain impact on the statistical results. Academic pressure is measured using a single item, a common but limited approach that cannot capture the multifaceted sources of stress or potential threshold (non-linear) effects. The externalizing problem behavior scale, while useful as an indirect indicator of students’ difficulties in adapting to school, is more oriented toward individual behaviors and lacks direct measures of perceptions of the school environment, such as safety, sense of belonging, and peer support. The academic performance indicator was derived from combined Chinese and mathematics test scores. To account for age and grade differences, this indicator was standardized to enhance comparability across samples.

Third, uncontrolled potential additional family socioeconomic indicators may affect the accuracy of the estimates. Although this study has controlled for variables such as age, gender, and urban/rural residence status, some important covariates were not included due to data availability limitations, including detailed indicators of family socioeconomic status (income, wealth), parental mental health status, school type, and regional characteristics.The absence of these factors may lead to residual confounding, preventing the complete separation of the independent effects of family and school on depression.Therefore, the interpretation of the results should be approached with caution.

Conclusion

This study aims to examine the correlations between perceived school environment, intergenerational interaction, and adolescent depression. The results indicate that in the dimension of intergenerational interaction, parent-child communication is negatively correlated with adolescent depression. In the dimension of perceived school environment, school satisfaction is negatively correlated with depression, while academic stress and externalizing problem behaviors are positively correlated with depression. Further analysis of moderating effects revealed that behavioral evaluation and parent-child communication could effectively buffer the negative impact of academic stress and externalizing problem behaviors on depression, respectively.The findings of this study contribute to a better understanding of the relationships among intergenerational interactions, perceived school environment, and adolescent depression.

Abbreviations

CFPS

the China Family Panel Studies

CES-D

Center for Epidemiologic Studies Depression Scale

Authors’ contributions

CJN contributed to the analysis of the data, and drafted the manuscript.LX contributed to the drafted the manuscript.YSM helped to perform the revision of the manuscript, and provided final approval for the manuscript. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by the Shanxi Provincial Basic Research Program of China (No. 202403021222212), the Shanxi Province Education Science 14th Five-Year Plan Research Project (No. GH-240007).

Data availability

The raw data supporting the conclusions of this article will be made available for download on the official CFPS website at http://www.isss.pku.edu.cn/cfps/.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the 1964 Helsinki declaration and its later amendments. The analysis is based on publicly available data from the China Family Panel Studies (CFPS). The CFPS project strictly complies with relevant regulations and periodically submits applications for ethical review or continuing review to the Peking University Biomedical Ethics Committee. Data collection is carried out only after obtaining ethical review approval. The ethical review approval number for the CFPS project remains consistent across all survey waves and is uniformly designated as IRB00001052-14010. Written informed consent was obtained from all participants prior to each wave of CFPS data collection.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.