Introduction
Adolescence is a stage of rapid psychological and physical development, during which individuals learn to develop their own unique outlooks and behaviors. However, 10–20% of children and adolescents worldwide are affected by mental health problems, which should be attended to and prioritized [Reference Kieling, Baker-Henningham, Belfer, Conti, Ertem and Omigbodun1, 2].
Parenting style composed of parents’ attitudes and behaviors in daily interactions with offspring has been reported to influence children’s mental health [Reference Yap, Pilkington, Ryan and Jorm3]. Different researchers have identified different dimensions of parenting, such as two dimensions of care and overprotection, or three dimensions of emotional warmth, rejection, and protection [Reference Parker, Tupling and Brown4, Reference Eun, Paksarian, He, Merikangas and Daeseleire5]. Eun et al. found that children who perceived high levels of maternal care were less likely to exhibit depressive symptoms, eating and behavioral disorders, while higher perceived paternal control was associated with greater odds of alcohol abuse. Ma et al. [Reference Ma, Gao, Chiu, Ding, Wang and Wang6] reported that better daily interactions between parents and adolescents are a protective factor against depression in adolescents. Parenting style also influences children’s subjective well-being (SWB), defined as self-evaluation on how well life is according to the individuals’ own standards [Reference Eun, Paksarian, He, Merikangas and Daeseleire5, Reference Xu, Sun, Zhu, Bai, Yu and Duan7, Reference Chen, Haines, Charlton and VanderWeele8]. Few studies have explored the influence of parenting style on subjective interpersonal popularity (SIP) or the level of individuals’ satisfaction with interpersonal relationships, which correlates with negative emotions such as depression as well [Reference Hames, Hagan and Joiner9].
Another variable that may be related to adolescents’ mental health is the adolescents’ achievement attribution tendency. Achievement attribution refers to how individuals explain their own or others’ success. It influences the expectations of achievement and the adaption of future behaviors [Reference Brun, Pansu and Dompnier10] as well as psychological outcomes such as depression, SWB, and SIP [Reference Ge and Hou11, Reference Määttä, Nurmi and Stattin12]. One study has reported the mediation role of achievement attribution pattern in the influence of parenting style on adolescents’ academic outcomes. Adolescents perceiving nonauthoritative parenting style tended to ascribe achievements to external factors or low individual capacity, which led to lower school attendance [Reference Glasgow, Dornbusch, Troyer, Steinberg and Ritter13]. However few studies have explored the role of achievement attribution pattern in the relationship between parenting style and children’s psychological outcomes.
In the current study, we aimed to first investigate the associations of one aspect of parenting, specifically, parent–child communication, with adolescents’ psychological outcomes including achievement attribution, depression symptoms, SWB, and SIP. Next, we aimed to explore the role of achievement attribution tendency in the relationships between parent–child communication and the remaining psychological outcomes in order to better understand the pathway between parent–child interactions and adolescents’ mental health and thereby inform interventions designed to promote adolescent mental health. We hypothesized that better parent–child communication will predict better psychological outcomes and that internal attribution (IA) would mediate the relationships between parent–child communication and the other psychological outcomes.
Methods
Data source and study sample
Data used for this study were obtained from the China Family Panel Studies (CFPS), a national longitudinal survey, carried out by the Institute of Social Science Survey (ISSS) at Peking University. CFPS covers 25 provinces (or municipalities) in China, representing 95% of the total population. To date CFPS has completed six rounds of surveys, beginning in the year 2010, and every 2 years since then. CFPS tracks and collects data at the individual, family, and community levels to reflect society, economy, demography, education, and health changes in China. All participants sign informed consent forms. The 2020 survey included a series of psychological scales including the Center for Epidemiologic Studies Depression scale (CES-D), an adapted version of the Parental Bonding Instrument (PBI), an achievement attribution scale, a well-being measure, and an interpersonal popularity measure. Of the 28,590 individuals who participated in this survey, there were 3,101 adolescents aged between 9 and 18 years. The average age was 13.71, with 1,390 female adolescents accounting for 47.5% of the total. Of these, 2,725 completed all the required measures and were included in our study. A detailed flowchart is shown in Figure 1 and the detailed demographic characteristics are shown in Table 1.
Flowchart of sample selection.
Sample characteristics.
Abbreviations: CES-D8, Center for Epidemiologic Studies Depression scale; PBI, parenting bonding instrument.
a There are 406 missing data points, 433 missing data points, 438 missing data points, 468 missing data points, 694 missing data points, 808 missing data points, 750 missing data points, and 41 missing data points for maternal age, paternal age, maternal education, paternal education, maternal employment, paternal employment, parental marriage state, and living areas, respectively.
Exposure measure
The Parental Bonding Instrument (PBI) developed by Parker [Reference Parker, Tupling and Brown4] was adapted by ISSS for the 2020 survey. Seven items assessing the adolescent’s perception of the communications pattern between parents and the adolescents were included. Participants were asked how often in the past 12 months when they did something wrong, their parents/guardians asked for reasons and guided them; encouraged them to do things with great effort; were gentile while talking with them; encouraged them to think independently; gave them reasons when asking them to do something; liked to talk with them; and praised them. Participants responded to these items on a scale ranging from one (never) to five (always), with two signifying rarely, three sometimes and four often. Scores also range from 1 to 5, and total score range from 7 to 35. Cronbach α of this seven-item scale was 0.82.
Mediator measure
The achievement attribution measure adopted in this survey was a seven-item scale designed for the Chinese population [Reference Ge and Hou11]. Participants were asked about the importance of family’s social status, family’s economic condition, education level, gifted talent, effort, luck, and family social connections to future achievements. Answers to every item were scored from 0 to 10. The scale has two factors: external attribution (EA; family’s social status, family’s economic condition, gifted talent, luck, and family social connections) and IA (education level and effort). The Cronbach α of the EA and IA were 0.76 and 0.62, respectively.
Outcome measure
The eight-item short version of the CES-D developed by Radloff [Reference Radloff14] was adopted to evaluate the depressive symptoms. The CES-D has been widely used in large international investigations [Reference Beydoun, Obhi, Weiss, Canas, Beydoun and Evans15, Reference Harshfield, Pennells, Schwartz, Willeit, Kaptoge and Bell16], and the CES-D8 is a brief version that takes less time and is more acceptable to respondents [Reference Karim, Weisz, Bibi and Rehman17, Reference Alvarez-Galvez and Rojas-Garcia18] who are asked to indicate how often in the past week they have felt depressed, felt everything that they did was an effort, slept restlessly, were happy, felt lonely, felt sad, enjoyed life, and could not get going. There are four response options: Hardly (less than a day); Sometimes (1–2 days); Often (3–4 days); and Most of the time (5–7 days). Responses are scored 0, 1, 2, or 3, respectively, with the two items indicating positive feelings reverse scored. The total CES-D8 score ranges from 0 to 24. Cronbach α for the eight-item scale was 0.77.
SWB was assessed by a single question “How happy are you”. Participants responded using an 11-point scale ranging from 0 (lowest score) to 10 (highest score). Likewise, participants’ SIP was assessed by a single question: “How popular are you with people?” with response options ranging from 0 (lowest score) to 10 (highest score).
Covariates
Covariates at individual, parent, and household levels were recorded. Potential confounding factors at the individual level included age, gender, the degree of education, and frequency of exercise. Potential confounding factors at the parental level included age, education, marital status, and employment status. The potential confounding factor at the household level was whether the family lived in an urban or rural area.
Statistical analysis
Firstly, we conducted descriptive analyses of the psychological and demographic characteristics of participants. Continuous variables and categorical variables are displayed as mean (standard deviation) and percentages, respectively, in Table 1. A series of linear regression analyses were used to explore the associations of parent–child communication with offspring’s achievement attribution, depression, SWB, and SIP. To test the robustness of this finding, we performed multiple linear regressions controlling for the demographic factors.
Next, we examined the mediation effect of achievement attribution tendency in the relationships between parent–child communication and offspring’s depression, SWB, and SIP. We conducted the Harman’s single-factor test first to examine the common method bias by SPSS 26.0 and obtained a variance of 24.49% for the first common factor, indicating that there was no common method bias. A structural equation model was evaluated using the R package “lavaan.” We packaged the items of CES-D8 into two parcels according to “internal-consistency approach” (one parcel of six items and one parcel of two items) and the items of PBI into three parcels according to “item-to-construct balance approach” (one parcel of three items and two parcels of two items) (Supplementary Material) [Reference Little, Cunningham, Shahar and Widaman19]. The structure of the packaged CES-D8 and PBI items was tested by confirmatory factor analysis, and the packaged parcels were included in the measurement model of structural equation model. As for structure model, achievement attribution was set as the mediator, parent–child communication was the independent factor, and depression symptoms, SWB, and SIP were dependent variables. Further, we included demographic variables as control factors to test the robustness of the model. For all paths, standardized indirect and total effects were estimated. Bootstraping (1,000 replications) to test the significance of indirect and total effects was conducted.
Goodness of overall fit was evaluated as follows: χ 2/df < 3 [Reference Kakemam, Rouzbahani, Rajabi and Roh20], comparative fit index (CFI) > 0.95, Tucker-Lewis index (TLI) > 0.95 [Reference Browne and Cudeck21], adjusted goodness-of-fit index (AGFI) > 0.9 [Reference Bentler and Bonett22], and root mean square error of approximation (RMSEA) < 0.08 [Reference Lt and Bentler23]. The value of χ 2 was influenced by sample size and the complexity of the model, so normed χ 2 (χ 2/df) was taken into consideration. The closer the values of CFI, TLI, and AGFI approach 1, the better the model fits. The closer the value of RMSEA approaches 0, the better the model fits.
Results
Sample characteristics
The study included 2,725 participants aged between 9 and 18 years. All participants had valid psychological measurements. Specific psychological measurements and demographic characteristics were all shown in Table 1.
Associations between parent–child communication and adolescents’ psychological well-being
We hypothesized that parent–child communication would predict adolescent offspring’s psychological outcomes. The linear regression analyses showed parent–child communication was significantly associated with adolescent offspring’s IA (β = 0.152, p < 0.001), depression symptoms (β = −0.296, p < 0.001), SWB (β = 0.356, p < 0.001) and SIP (β = 0.141, p < 0.001), while it was not significantly associated with EA. To control for the influence of demographic variables, we included potential cofounding factors at adolescent level and at three levels of adolescent, parent, and household, respectively, the associations of parent–child communication still remained significant with offspring’s psychological outcomes except EA (see Table 2).
Linear regression analyses of parenting-child communication on adolescents’ psychological outcomes.
Note: Standardized regression coefficients are displayed, with standard errors in parentheses. Model 1, raw model; model 2, including potential cofounding factors at adolescent level (gender, education level, exercise frequency); model 3, including potential cofounding factors at parental and household levels in addition to model 2 (parental age, parental education, and living areas).
Abbreviations: CES-D8, Center for Epidemiologic Studies Depression scale; EA, external attribution; IA, internal attribution; PBI, parenting bonding instrument; SIP, subjective interpersonal popularity; SWB, subjective well-being.
* p < 0.001.
The mediating effect of achievement attribution tendency
We hypothesized that IA, one of the factors of achievement attribution would mediate the relationships between parent–child communication and the other psychological outcomes. Figure 2 shows the detailed overall model. Indicators of model fit are as follows: χ 2 = 21.937, p = 0.287, χ 2/df = 1.155; CFI = 0.999; TLI = 0.999; AGFI = 0.996; RMSEA = 0.008, (90% CI, 0.000, 0.019), indicating a good model fit. As shown in Figure 2, white nodes and connected edges represent the measurement model and numerals on connecting lines represent standardized factor loads. The factor loadings of CES-D8, PBI, and IA range from 0.42 to 0.79 all with significance. Blue nodes and connected edges compose the structural model. One-way arrows represent probable causality, and bipolar arrows represent correlation. Red and blue arrows represent positive effects and negative effects, respectively. Numerals on connecting edges represent standardized coefficients ranging from −0.57 to 0.35 all with statistical significance.
Mediation analysis of achievement attribution pattern in the associations between parent–child communication and adolescents’ depression symptom, subjective well-being, and subjective interpersonal popularity. The overall structural equation model has two parts, the measurement model (white nodes and connected edges) and the structural model (blue nodes and connected edges). In the measurement model, factor loadings of nodes connected by dotted lines are fixed at one and are presented as standardized form. In the structure model, probable causation relationship is shown with a single arrow, correlation with a double arrow. Red arrows represent positive effects and blue negative effects. Numerals stand for standardized coefficients. CES-D8, Center for Epidemiologic Studies Depression scale; IA, internal attribution; PBI, Parenting Bonding Instrument; SIP, subjective interpersonal popularity; SWB, subjective well-being.
The indirect effects through IA tendency of achievement and total effects of parent–child communication on adolescents’ psychological outcomes are shown in Table 3. As can be seen, IA tendency partly mediates the relationships between parent–child communication and adolescents’ psychological outcomes. For depression symptoms, IA mediates 5.3% of the effects indirectly, for SWB and SIP, IA mediates 12.3 and 31.8%, respectively. The overall model including demographic variables as control factors was tested and also showed a good fit as well (Supplementary Table S1).
The total effects of parent–child communication and the indirect effects through achievement attribution on adolescents’ psychological well-being.
Note: Beta, the standardized estimates; “LLCI”, the lower limit of 95% confidence interval of coefficients; “ULCI”, the upper limit of 95% confidence interval of coefficients.
* p < 0.01;
† p < 0.001.
Discussion
As hypothesized, we found moderate associations between the quality of day-to-day parent–child communication and adolescents’ psychological outcomes. Parent–child communication accounted for 8.5% of the variance in adolescent depression, consistent with previous studies [Reference McLeod, Weisz and Wood24]. The proportion of variance explained by parent–child communication in adolescents’ SWB was a little higher, 14.2%, but the proportion for SIP was lower, 4.1%. Better parent–child communication, as perceived by adolescents, helps to reduce their depressive symptoms and to promote SWB and SIP.
Mediation analysis revealed that IA of achievement partially mediates the associations between parent–child communication and psychological well-being. Better parent–child communication helps adolescents ascribe their achievements to their own effort or growth through education. Consequently, a tendency to attribute achievement to effort or education engagement helps to reduce individuals’ depression symptoms, and improve SWB and SIP. These findings are consistent with previous attribution theories. Children’s achievement attribution pattern is usually affected by the attitudes and behaviors of significant others like parents or guardians [Reference Reid, Ramey and Burchinal25]. McFarland and Ross [Reference McFarland and Ross26] and Weiner [Reference Weiner27] have suggested attribution styles were the determinants of mood reactions to success or failure, ascribing achievements to internal factors could produce more positive emotion and less negative emotion. Studies have linked parental factors, children’ attribution pattern, and academic outcomes, which suggested that parental attribution style interacted with children’s attribution style, and affected children’s academic achievements [Reference Khodayarifard, Brinthaupt and Anshel28–Reference Aunola, Stattin and Nurmi30]. In the background of Chinese culture, previous researches revealed parents tended to ascribe children’s achievement to working hard and this pattern of attribution could predict children’s academic achievement [Reference Phillipson31], and children inclined to ascribe success or failure to internal factors like ability and effort than external factors [Reference Ho, Salili, Biggs and Kit‐Tai32]. The interactions between parents and adolescents and adolescents’ achievement attribution in our study showed this similar pattern. Further, we linked parent–child communication, adolescents’ achievement attribution pattern, and mental health, revealing the importance of parent–child communication like encouraging autonomy and effort in adolescents’ positive attribution and mental health. Better communication perceived by adolescents from their parents can bring encouragement, confidence, and warmth, and helps them to ascribe achievement to their internal factors like efforts in some degree, which facilitates a positive attribution pattern formation and improved psychological well-being. Notably, the outcome variables in the mediation analysis we conducted incorporated depression, SWB, and SIP, showing the mutual correlations of these mental health indicators, and reflecting mental health as a whole.
Till now few studies have connected parent–child communication, achievement attribution pattern, and psychological well-being together to explore their relationships. Our study provided insights into the associations of parent–child communication and achievement attribution pattern with adolescents’ mental health, which can inform future interventions for adolescents’ mental health promotion. While our sample size is relatively large which potentially makes our results robust, our study has some limitations. Firstly, our study utilized a cross-sectional design, and future longitudinal studies are needed to inform causality in the explored relationships. Secondly, the proportions of variance explained by parent–child communication quality in adolescents’ psychological outcomes were relatively small, which may indicate there were potential confounders or other important variables that were not measured. Although it is difficult to identify and include all possible confounders, the proportions of variance we explained are consistent with those reported in past studies. Thirdly, we experienced some missing data which may impacted the findings. Fourthly, the scales used were based on self-reports which may bring reporting bias, though our statistical test showing there existed no common method bias. Fifthly, the measurements of SWB and SIP were both based on a single item, which may bring potential bias. Sixthly, the measurement of parent–child communication did not distinguish between fathers and mothers, an issue which needs further exploration.
In conclusion, our research revealed that parent–child communication associates with adolescents’ psychological outcomes including depression symptoms, SWB, and SIP. These relationships are partly mediated by adolescents’ IA of achievement. Interventions targeting improved parent–child communication and positive achievement attribution formation may promote adolescents’ mental health.
Abbreviations
- CES-D
-
Center for Epidemiologic Studies Depression
- EA
-
external attribution
- IA
-
internal attribution
- PBI
-
parenting bonding instrument
- SIP
-
subjective interpersonal popularity
- SWB
-
subjective well-being
Supplementary Materials
To view supplementary material for this article, please visit http://doi.org/10.1192/j.eurpsy.2022.2314.
Data Availability Statement
Data were available from http://www.isss.pku.edu.cn/cfps/sjzx/gksj/index.htm.
Author Contributions
Conceptualization: N.L., Y.L.; Methodology: Y.L., X.H., S.X., Q.H., C.L.; Supervision: Q.H., C.L., P.J., Y.X., H.F., J.C.; Validation: X.H., S.X.; Writing—original draft: N.L., Writing—review and editing: D.M., Y.X., H.F., J.C.
Financial Support
This study was supported by Peak Discipline Project of Shanghai Jiao Tong University School of Medicine-Clinical Medicine “Research Physician” Project, Program of Shanghai Academic/Technology Research Leader (21XD1423300), Shanghai Pujiang Program (21PJD063).
Conflict of Interest
The authors declare none.
Open access
Comments
No Comments have been published for this article.