Parent-Child Discrepancies in Perceived Parent-Child Communication and Depressive Symptoms in Early Adolescents in China

Although recent studies demonstrated that parent-child discrepancies in the perceived family processes were associated with children’s developmental outcomes, few studies have addressed this issue in different types of families in mainland China. The present study investigated that how discrepancies in parents’ and adolescents’ perceptions of parent-adolescent communication were associated with early adolescent depressive symptoms in a nationally representative sample (N = 15,377) with 7010 father-adolescent dyads (adolescents: Mage = 14.24 years, SD = 1.25 years; 5960 adolescents from two-parent families, 443 adolescents from single-father families) and 8367 mother-adolescent dyads (adolescents: Mage = 14.02 years, SD = 1.18 years; 6670 adolescents from two-parent families, 1362 adolescents from single-mother families) in China. Adolescent respondents completed a measure of depressive symptoms and all informants reported on the perceived levels of parent-adolescent communication. Results indicated that adolescents reported parent-child communication more negatively than did their parents. Father-adolescent discrepancies were also greater in intact families than non-intact families. Polynomial regression analyses indicated that while there was a significant interactive effect of father-reported and adolescent-reported father-adolescent communication in Chinese two-parent families, no significant interaction was found for mother-adolescent dyad. Besides, adolescent-reported mother-child communication interacted with mother-reported communication in Chinese single-mother families only. The findings clarify parent-adolescent discrepancies in parent-child communication in different types of families in China and they have theoretical and practical implications on the role of discrepancies in parents and adolescent children on perceived parent-adolescent communication in early adolescent depressive symptoms.


Introduction
Compared to children, early adolescents face greater challenges and stresses arising from significant biological, cognitive, and social-emotional developmental changes [1][2][3], resulting in mental health issues such as internalizing and externalizing problems [4,5]. For example, high prevalence rates of adolescent depressive symptoms in the United States (13.3%) and China (24.6%) have been reported [6,7]. As many studies have demonstrated that depressive symptoms have detrimental impacts on adolescents' concurrent and later psychosocial functioning [8,9], it is very important to identify factors influencing the risk of depressive symptoms during early adolescence.
Parent-adolescent communication as a key dimension of attachment plays an important protective role in shaping adolescent depressive symptoms [10][11][12]. Generally speaking, while positive parental communication was negatively associated with adolescent depressive symptoms [10,11,13], low quality parental communication such as rejection municate their emotions openly [46]. As a result, they are more likely to communicate openly and discuss problems with their parents [37,47], resulting in a low discrepancy in perceived parent-child communication. Moreover, securely attached adolescents may get more timely support from their parents through open communication that would decrease the risk of developing depressive symptoms [48,49]. In contrast, adolescents with an insecure attachment are less willing to communicate with their parents, which would lead to a discrepancy in perceived parent-adolescent communication [40]. Besides, parents' overestimation of parent-adolescent communication will lead to a higher risk of depressive symptoms [25], because parents are unable to provide effective and timely support based on limited information.

Chinese Family Processes with Reference to Parent-Child Discrepancies in Parent-Adolescent Communication
In traditional Chinese culture, parental control is very strict. Shek [49] argued that traditional Chinese parental control included psychological control (expectation of total obedience of the child) and behavior control (high expectation and strict discipline). Relative to Western parents, Chinese parents are less likely to show their love through open communication, and Chinese adolescents are less willing to discuss disagreement with their parents openly [50]. As a result, Chinese parents and their children may perceive more discordance in parent-adolescent communication.
Moreover, in traditional Chinese families, fathers and mothers play different roles in parenting. This is reflected in the Chinese saying "men take care of things outside the family whereas women take care of things inside the family" (nan zhu wai, nv zhu nei). A father is seen as the head of a family and a mother is expected to take care of the basic needs of the children. Fathers were expected to play the role of teachers supervising the children to ensure that they behaved well, as there is a saying, "it is the fault of the father if he only raises the child without teaching him" (yang bu jiao, fu zhi guo) [51]. Thus, in traditional Chinese culture, fathers are expected to give more behavioral control and moral guidance to their children, and children are required to be more obedient. Finally, father and their adolescent children rarely negotiate disagreement and share decisionmaking, leading to greater discrepancies in father-adolescent communication. Furthermore, a recent study found that mothers showed higher levels of parental control and parentchild relational quality than did fathers during their high school years in Hong Kong [52]. Another study found that Chinese adolescents perceived maternal parenting attributes were more positive than perceived paternal parenting attributes, and this effect was found only among adolescent girls [53]. These findings suggest that father-child discrepancies in perceived father-adolescent communication are greater than mother-child discrepancies in perceived mother-adolescent communication in China. Unfortunately, despite recent work on Chinese families and adolescent development [54,55], studies on parent-adolescent discrepancies in perceived family processes are few.
Furthermore, due to industrialization, urbanization, and population migration, marriage and family have undergone profound changes in China, one of which is characterized by the diversification of family structure or living arrangement pattern. First, China's crude divorce rate rose from 0.9 per 1000 people in 2002 to 2.79 in 2015 [56,57]. Second, intergenerational parenting is very common in China. A considerable number of couples send their children to grandparents' homes for long periods of time, and the proportion of such "generational separation families" is increasing year by year [58]. Third, restricted by the household registration system and their own economic conditions, most rural migrant workers leave their children in their rural hometowns, resulting in a large number of "left-behind children" [59].
According to the living arrangement of children and parents, the Chinese family structure can be divided into four categories: two-parent families, single-mother families, single-father families, and both parents absent families (such as parents working in places remote from their home). Different types of families may have different characteristics regarding family processes. Shek [49] found that parental behavioral control and parent-child relational quality were poorer, while parental psychological control was higher, in

Participants and Procedure
We used the nationally representative data from the China Education Panel Survey (CEPS) collected during September 2013 and March 2014. A total of 19,487 adolescents and their mothers (or fathers) from 438 classes in 112 middle schools in 28 counties (districts) in China completed the child-reported and parent-reported questionnaires, respectively.
In the present study, we deleted 175 participants whose depressive symptom scores were the maximal value to ensure normality of distribution. Additionally, we deleted 3176 participants with missing data. Finally, there were 15,377 parent-adolescent dyads (adolescents: Mage = 14.12 years, SD = 1.22 years; Nboys = 7677, Ngirls = 7700; Ngrade 7 = 8060, Ngrade 9 = 7317) participating in the study, including 7010 father-adolescent dyads and 8367 mother-adolescent dyads. Only one parent (father or mother) in a family participated. Please see Table 1. Following previous studies [10,12], we measured paternal and maternal communication as perceived by adolescents, fathers, and mothers with five items used to measure father-and mother-adolescent communication as perceived by adolescents, fathers, and mothers. Adolescents were asked, "How often did your father/mother have a discussion with you on the following issues?" The five response items are "What happened at school," "Your relationship with your friends," "Your relationship with your teacher," "Your feelings," and "Your mind or troubles." Similarly, we assessed father-and mother-adolescent communication by fathers or mothers using the same five items (except replacing "your" by "child's"), and fathers or mothers were asked, "How often did you have a discussion with your child on the following issues?" All items were rated on a three-point response scale (1 = never, 2 = once in a while, and 3 = often). In this study, internal consistency for parent-adolescent communication was found to be good across informants. Specifically, Cronbach's α was 0.83 for adolescent reports on the mother-child communication, 0.83 for adolescent reports on the father-child communication, 0.87 for mother-reports on the mother-child communication, and 0.85 for father-reports on the father-child communication.

Depressive Symptoms
According to previous measurement [66,67], depressive symptoms was measured by five items, including "depressed", "upset", "unhappy", "sad", and "life has no meaning". Following the question "Have you had any of the following feelings in the past seven days?", all items were rated on a five-point Likert-type scale ranging from 1 (almost never) to 5 (almost always). The previous study has demonstrated this measurement was valid in Chinese adolescents [19]. Cronbach's α in this study was 0.85.

Data Analyses
Regarding Research Question 1, paired t test was used to investigate father/motheradolescent differences in perceived parent-child communication. Cohen's d was used to determine the effect size [68]. Regarding Research Question 2, an independent-samples t test was used to examine whether parent-adolescent discrepancy in parent-child communication was different between intact families and non-intact families.
Regarding Research Question 3, polynomial multiple regression analyses [69] were used to test the interaction effect of father-reported (or mother-reported) and adolescentreported parent-adolescent communication on adolescent depressive symptoms. The polynomial multiple regression model, used in testing parent and adolescent discrepancies of perceived parent-adolescent communication on the prediction of adolescent developmental outcomes, is shown in the equation: while P and A represent mothers'/fathers' and adolescents' reports of parent-adolescent communication, b 2 and b 5 represent the linear and quadratic effects of mother/father reports at the mean level of adolescent reports. Besides, b 1 and b 3 represent the linear and quadratic effects of adolescent reports at mean levels of mother/father reports, and b 4 represents the interaction term between mother/father and adolescent reports. To avoid the misrepresentation of the quadratic effect of mother/father or adolescent reports by the interaction term, A 2 and P 2 were included in the equation [70,71]. We standardized the scores of parent-child communication. Simple slope analysis and plotting were used when interaction was significant [72]. Moreover, analyses were conducted separately for the mother-child and father-child communication. Additionally, analyses were conducted separately in four types of family structure, including two-parent families, single-father families, single-mother families, parent-absent families.
Regarding Research Question 2, results showed that parent-child discrepancy in perceived mother-child communication was not different between intact families (M = 0.76) and non-intact families (M = 0.76), t = −0.008, p = 0.994, whereas parent-child discrepancy in perceived father-child communication was larger in intact families (M = 1.51) than non-intact families (M = 1.21), t = 2.98, p = 0.003. The findings can be seen in Table 3. Hypothesis 2 was not supported. Table 3. Comparisons of parent-child discrepancies in perceived parent-adolescent communication between intact families and non-intact families. Regarding Research Question 3, polynomial regression analyses were used to test the interaction effect of father-reported (or mother-reported) and adolescent-reported father-adolescent/mother-adolescent communication on adolescent depressive symptoms.
Regarding single-mother families, the interaction between mother-reported and adolescent-reported mother-adolescent communication on adolescent depressive symptoms was significant, b = −0.287, SE = 0.127, t = −2.256, p = 0.024, 95% CI = [−0.537, −0.037]. The findings are presented in Table 5. Simple slope analysis showed that when adolescent perceived low mother-adolescent communication, adolescent depressive symptoms were higher when mother-reported mother-adolescent communication was In other words, high mother-perceived mother-child communication but low adolescent-perceived mother-child communication was linked with the highest adolescent depression, whereas high mother-perceived and high adolescent-perceived mother-child communication was linked with the lowest adolescent depression (see Figure 2a). However, no significant interaction effect between father-reported and adolescent-reported father-adolescent communication was found in single-father families (see Figure 2b). In addition, for parent-absent families, the interactions between motherreported (or father-reported) and adolescent-reported mother-adolescent/father-adolescent communication on adolescent depressive symptoms were not significant. Hypothesis 4 was supported.

Discussion
The present study makes a pioneer contribution to understanding how congruency or discrepancy between parent-reported and adolescent-reported parental communication influence adolescent depressive symptoms in Chinese intact and non-intact families. First, we investigated the discrepancies in mother-adolescent and father-adolescent communication, and their linkages with adolescent depressive symptoms in intact and non-intact families, by using the nationally representative sample in China. Second, we used polynomial regression analyses to examine the interactive effects of how informant discrepancies would influence early adolescent depressive symptoms in China. This methodology is superior because recent studies showed that polynomial regression analyses could yield more accurate findings than difference scores [63,70]. Third, our findings provide support for the view that parent-child discrepancy in the perceived family process is pathological where parent-child discrepancies in perceived family processes have maladaptive impacts on children's developmental outcomes.

Discrepancies between Parents and Adolescents on Perceived Parent-Child Communication
Results indicated that parents and adolescents had a different perception of parentchild communication, with adolescents having more negative perceptions than parents in China. The results are consistent with previous studies showing adolescents perceive more negative parent-adolescent communication than did their parents in Western societies [23,29,63]. In the Chinese context, discrepancies between perceptions of parentadolescent communication may be intensified due to hierarchical decision-making and the lack of emotional expression within the Chinese families [73]. For Chinese parents, they are responsible for educating and nurturing their children, so they are more likely to overestimate their efforts in parent-adolescent communication, whereas Chinese adolescents often demonstrate their autonomy and independence by showing their differences from their parents. In Confucian culture, Chinese adolescents are required to be obedient, so they are less likely to communicate openly on disagreements and conflicts with their parents [74]. Besides, Chinese parents expressed less positive affection to their children compared to American parents [75]. The less affectional communication circumstance may lead to adolescents' lower evaluations of parent-adolescent communication relative to their parents [55,76]. Additionally, as Chinese fathers showed less responsiveness, warmth, and concern, adolescents reported more negative feelings when communicating with their fathers than did fathers report [77].
Moreover, we found that father-adolescent discrepancies in perceived paternal communication were greater in intact families than non-intact families, whereas mother-adolescent discrepancies in perceived maternal communication were the same in intact families as well as non-intact families. This finding did not fully support Hypothesis 2 which argues that parent-adolescent discrepancies were greater in non-intact families than intact families. The reason may be that fathers perceived lower levels of father-child communication in non-intact families (such as single-father families, parent-absent families) than intact families (two-parent families). This can also be explained by the fact that as mothers are the main socialization agent, their communication with their children is more intense as compared to that of the fathers.

Parent-Adolescent Discrepancies in Perception of the Parent-Child Communication and Adolescent Depressive Symptoms in Two-Parent Families
Results indicated that, for adolescents from two-parent families, the interaction between father-reported and adolescent-reported communication predicted adolescent depressive symptoms. Specifically, adolescent-perceived father-child positive communication was more strongly linked with less adolescent depressive symptoms when father-perceived father-child communication was high than when it was low. In other words, congruence in high adolescent-perceived and father-perceived father-child positive communication was linked with the lowest levels of adolescent depressive symptoms. Moreover, although the interaction was not significant for mother-adolescent communication discrepancies in two-parent families, a similar pattern was found. Specifically, adolescents have fewer depressive symptoms when both the mother and the adolescent perceived high mother-child communication. These results are consistent with previous Western findings that congruence in low adolescent-perceived and parent-perceived parent-child negative interaction was linked with the lowest adolescent depression [64].
In summary, in two-parent families, high parent-reported and high adolescent-reported parent-adolescent communication were associated with the lowest levels of adolescent depressive symptoms. Theoretically, attachment theory and family ecological models can explain this result. Attachment theory argues that adolescents with secure attachment to their parents are less likely to suffer depressive symptoms relative to adolescents with insecure attachment [78]. Adolescents with secure attachment to their parents usually have more positive parent-child communication compared with insecurely attached adolescents [79], thus both securely attached adolescents and their parents may perceive high levels of parent-child communication. Moreover, family ecological models suggest that the father-child relationship, mother-child relationship, and father-mother relationship in two-parent families together influence children's developmental outcomes. If both parents and adolescents reported high levels of parent-child communication in two-parent families, which suggests family functioning is well, children are less likely to have detrimental developmental outcomes such as depressive symptoms.

Discrepancies in Perception of Parent-Adolescent Communication and Early Adolescent Depressive Symptoms in Single-Parent Families
For adolescents in single-mother families, results indicated that the interaction between adolescent-reported mother-adolescent communication and mother-reported communication predicted adolescent depressive symptoms. Specifically, low adolescentperceived but high mother-perceived mother-child communication was linked with the highest adolescent depressive symptoms, while congruence of high adolescent-perceived and mother-perceived mother-child communication was linked with the lowest adolescent depressive symptoms. This observation is consistent with previous studies showing that greater informant discrepancies were linked with more internalizing problems [22,41]. This observation is also consistent with the previous finding that congruence between mothers' and adolescents' perception of low conflict was linked with fewer adolescent depressive symptoms [70]. The explanation for this finding is that as the mother-child dyad is the cornerstone of single-mother families, mother-child discrepancies in perceived parenting would create conflict and be a stressor for the adolescent child. In Chinese single-mother families, mothers are responsible for raising their children and they usually have higher expectations and involvement for their children. Chinese mothers are also more concerned about closeness and intimacy, which cannot satisfy early adolescents' needs for more autonomy and individuality [80,81]. Besides, Chinese mothers pay more attention to adolescents' academic performance than mothers in western culture [82]. Therefore, in Chinese singlemother families, mothers are likely to perceive high mother-adolescent communication, while adolescents might perceive low mother-adolescent communication, so that mothers would ignore adolescent internalizing behaviors and adolescents cannot seek help from others in families. As a result, adolescents would suffer more depressive symptoms.
However, for adolescents in single-father families, results indicated that interaction between adolescent-reported father-adolescent communication and father-reported communication on adolescent depressive symptoms was not significant. This result suggested that congruence of adolescent-perceived and father-perceived low levels of father-child communication was linked with highest adolescent depression, while congruence of adolescentperceived and father-perceived high levels of father-child communication was linked with lowest adolescent depression. This result may be explained by low attachment and low bonding between children and their fathers in single-father families. In traditional Chinese culture, children usually have low attachment and bonding with their fathers due to less emotional expression in father-child communication [51]. Moreover, in Chinese single-father families, as fathers need to spend more time on earning money to support the family, they have less time to communicate with their children. Previous studies showed that a high level of father-child communication reduced adolescents' depressive symptoms [48,83]. As a result, both children and their fathers in single-father families would perceive low levels of father-child communication, in turn, they develop low attachment and low bonding, which could contribute to depressive symptoms.

Theoretical Implications
Our findings have important theoretical implications. Our findings suggest that parent-child discrepancy in the perceived family process is not normative but pathological, and it is not a measurement error, but it has theoretical significance. Specifically, parent-child discrepancies in perceived family processes are associated with maladaptive children's developmental outcomes, which provide support for the pathological perspective on the nature and impact of parent-adolescent discrepancy in the perceived family process. Furthermore, our findings in the Chinese context suggest the maladaptive impacts of parent-child discrepancy in the perceived family process are universal in different cultures.

Implications for Clinical Practice
Our findings provide information regarding maladaptive impacts of discrepancies between parents and adolescents in the perceived parent-adolescent communication on adolescent developmental outcomes in Chinese two-parent families and single-mother families. First, for adolescents from two-parent families, congruence of adolescent-reported and father-or mother-reported high communication was linked with low levels of depressive symptoms. Thus, for early adolescents in two-parent families, practitioners should pay more attention to adolescents with the congruence of adolescent-reported and parent-reported low parent-adolescent communication. Importantly, intervention programs should be developed to improve the quality of parent-adolescent communication, in particular for those families where there is low father-adolescent communication and larger discrepancies between fathers and adolescents.
Second, among early adolescents from single-mother families, practitioners should be sensitive to adolescents with the discrepancies of adolescent-reported low motheradolescent communication and mother-reported high communication, because these adolescents were at the highest risk for depressive symptoms. Intervention programs should be designed to improve communication skills and strategies for these adolescents and their mothers in single-mother families. Intervention should focus on raising mothers' awareness of their perception of mother-child communication as well as discrepancies between mothers and adolescent children.

Limitations of the Study
The present study has several limitations. First, as the present study was a crosssectional survey, causality between variables cannot be effectively explained. Thus, future studies should employ a longitudinal survey. Second, we only measure parent-driven communication which might have different impacts on adolescent depression compared with child-driven communication [10]. Both parent-driven and child-driven communication should be considered in future studies. Third, we used an established database with fixed measures that may limit the comparisons of other related studies and the clinical application. Fourth, some important factors, such as the size of the family nucleus, the order in which the adolescent occupies the children depending on the number of siblings, have not been controlled. Further studies should discuss these factors. Nevertheless, as families in China are mostly single-child families, this problem should not be great. Fifth, as only one parent (father or mother) in a family participated our study, we could not compare the effects of discrepancies in the mother-adolescent and father-adolescent communication on adolescent depressive symptoms. This should be further studied because family systems theory proposes that the mother-child and father-child relationships are related but different family subsystems, which may be differentially linked with adolescent