Family dysfunction and Adolescents' anxiety and depression: A multiple mediation model

Highlights

• Longitudinal sample: 921 Chinese early adolescents with three waves.

• T1 Family dysfunction (FD) positively predicted T3 anxiety (AN) and depression (DE).

• T2 loneliness and self-esteem mediated the relations between T1 FD and T3 AN and DE.

• Findings have implications for prevention and intervention on anxiety and depression.

Abstract

Based on a social-cognitive model, we examined the multiple mediating roles of self-esteem and loneliness in linking family dysfunction to anxiety and depression in adolescents. Participants at baseline included 921 junior high school students (M_age = 12.98 years; 51.7% girls) from a midsized city located in Northern China. The students completed a multi-measure questionnaire at three time points, six months apart, starting from the initial grade of junior high school (Grade 7). Using structural equation modeling, we found that: High levels of family dysfunction at Time 1 were significantly related to increases in anxiety and depression at Time 3; both self-esteem and loneliness at Time 2 mediated the relations between family dysfunction at Time 1 and anxiety and depression at Time 3. These results suggest the importance of multisystemic efforts (i.e., addressing social context and intrapersonal cognitive factors simultaneously) to prevent or reduce adolescents' anxiety and depression.

Introduction

Anxiety and depressive disorders are among the most common forms of psychopathology in adolescence (Inchley, Kirby, & Currie, 2011). These disorders represent typical responses to the internal and external pressures corresponding to the unique physical and psychological changes associated with adolescence, including the development of greater self-awareness (Rasing, Creemers, Janssens, & Scholte, 2017). The specific prevalence rates of anxiety and depressive disorders are 38% and 18% respectively, among people of ages 13–17 in America (Kessler, Petukhova, Sampson, Zaslavsky, & Wittchen, 2012). Among adolescents in non-Western countries, such as China, the prevalence of anxiety disorders ranges approximately from 7% to 38%, and the prevalence of depressive disorders ranges approximately from 8% to 20% (Chen et al., 2014). Adolescents who suffer from anxiety or depression will likely experience multiple adverse outcomes, such as academic difficulties (e.g., poor academic performance), school dropout, maladaptive social relations, and higher risk for substance abuse and suicide (Ingul & Nordahl, 2013; Seipp, 2007). In addition, entering junior high school is an important step for many young adolescents in their academic journey (Guo, Tian, & Huebner, 2018). In Mainland China, this transition occurs in the seventh grade, following six years of the more familiar and secure setting of elementary school. When students enter junior high school, they perceive more pressure to succeed and must adapt to new academic workloads, school expectations, and social relationships (Duchesne, Ratelle, & Roy, 2012). Thus, junior high school students are especially prone to experiencing anxiety and depression. Therefore, it is warranted to pay attention to adolescents' anxiety and depression during this transitional period.

However, Chinese adolescents' internalizing problems have largely been neglected by researchers because Chinese people who are deeply influenced by traditional Chinese culture usually consider individual well-being less important than the welfare of the collective or long-term personal success (Chen & Li, 2000). Therefore, to date, there have been few systematic, empirically-informed prevention or intervention efforts related to internalizing problems among students in Chinese schools. Moreover, the prevalence rates and levels of Chinese adolescents' anxiety and depression have been increasing in the past decades, likely as the result of facing the many rapid and dramatic social changes in China, such as economic and education reforms as well as changes in family structure (Zheng, Rijsdijk, Pingault, McMahon, & Unger, 2016). Therefore, it is important to understand the determinants and the specific psychological mechanisms accounting for the development of adolescents' anxiety and depression in China, in order to promote healthy outcomes as early as possible.

Some important determinants likely exist in the family context, since the family is one of the most important contexts for the developmental of adolescents' mental health outcomes (Tian & Li, 2005). In early adolescence, individuals develop two conflicting tendencies of independence and attachment with parents. On the one hand, adolescents seek increasing autonomy from their parents. On the other hand, adolescents still benefit from a secure attachment to their parents as they are not yet sufficiently mature to deal with the challenges in their life without some support (Xu, Fang, Zhang, Lin, & Zhang, 2008). These two conflicting tendencies result in unique challenges in the interactions between parents and children (Xu et al., 2008), and thus children's psychological problems are often related to the quality of their family functioning. Family dysfunction (e.g., poor problem-solving strategies, ineffective communications among family members) means that a family system does not facilitate appropriate functioning (Mousavi, 2004). According to previous research, higher levels of family dysfunction are closely related to higher levels of anxiety and depression in adolescents (Xie, Xie, Zhang, & Zhang, 2008). However, previous research has mainly focused on the direct effects of family dysfunction, with few attempts to elucidate the psychological mechanisms that account for the association between family dysfunction and adolescents' anxiety and depression.

The overarching purpose of our study was thus to reveal the psychological mechanisms linking family dysfunction and adolescents' anxiety and depression. Social-cognitive models highlight the cognitive pathways between social contexts and individuals' developmental outcomes (Wood, Maltby, Stewart, Linley, & Joseph, 2008). Previous literature has suggested that the family system not only relates to the formulation of individuals' self-concepts, but it may also relate to interpersonal interactions (Ferro & Boyle, 2015; Wu & Chow, 2013). Therefore, family dysfunction could be linked with individual differences in psychopathology (e.g., anxiety and depression) via adolescents' feelings about themselves (i.e., self-esteem) and their interpersonal relationships (i.e., loneliness). Given that it is possible to ignore the correlations among mediating variables by testing each mediating variable separately (Kong et al., 2012), we examined the mediation effects of both self-esteem and loneliness simultaneously in the relations between family dysfunction and anxiety and depression in adolescents. Overall, we evaluated a multiple mediating model following a social-cognitive theoretical framework: social context (i.e., family dysfunction) → perceptions of self (i.e., self-esteem) and perceptions of interpersonal relationship with parents and peers (i.e., loneliness) → psychopathology outcomes (i.e., anxiety and depression) among Chinese adolescents. Furthermore, we employed a longitudinal design to provide a relatively stringent test of the model.

The McMaster Model of Family Functioning (MMFF) postulates that a positive family environment is necessary for the healthy development of the physical, psychological, and social functioning of its family members (Epstein, Bishop, & Levin, 1978). The MMFF proposes that the family system is comprised of six key dimensions of functioning (Epstein, Baldwin, & Bishop, 1983): Problem Solving (the family's ability to resolve problems), Communication (how family members exchange verbal information), Roles (the clarity and appropriateness of the allocation of family roles), Affective Responsiveness (the extent to which family members experience appropriate affect over a range of stimuli), Affective Involvement (the extent to which family members are interested in and value each other's activities and concerns), and Behavior Control (how family members express and maintain standards of behavior). Family dysfunction, the evaluation of the functioning of the total family system, refers to a family system that does not facilitate appropriate functioning in one or more of the aforementioned six dimensions (Mousavi, 2004). Given that family dysfunction might be an important determinant of anxiety and depression based on MMFF, we chose it as the construct of interest for this study.

Focusing on family dysfunction is of special importance for designing interventions for adolescents' anxiety and depression. First, the family is a crucial setting for individual development (Thariq, 2018), especially for junior high school students who are experiencing developmental changes and higher academic expectations and loads. Second, family dysfunction reflects a systemic perspective that considers the family as a whole, which is greater than the sum of its parts (Xu et al., 2008). Third, family dysfunction is amenable to intervention, and can be detected early and influenced by parents (Sandler, Schoenfelder, Wolchik, & MacKinnon, 2011). In addition, most parents have an inherent motivation to take preventive actions concerning the development of their children (Yap, Pilkington, Ryan, & Jorm, 2014). Therefore, we focused on family dysfunction as a predictor of young adolescents' anxiety and depression based on its likely critical importance for intervention.

Previous research has demonstrated that family dysfunction is related to adolescent psychopathology (Chapman & Woodruff-Borden, 2009; Ferro & Boyle, 2015). According to family function theory (Miller, Ryan, Keitner, Bishop, & Epstein, 2000), the overall family environment is associated with the formation of anxiety and depression. Some studies have revealed that higher levels of family dysfunction represent a risk factor for the development of higher levels of anxiety (Katz & Low, 2004) and depression (Martin, Rozanes, Pearce, & Allison, 2010). For instance, Weeks et al. (2014) found that family dysfunction at age 8–9 increased the possibility of severe internalizing symptoms (i.e., anxiety and depression) at age 12–13 with a sample of 4405 Canadian children. Pagani, Japel, Vaillancourt, Côté, and Tremblay (2008) indicated that a prolonged duration of high levels of family dysfunction was associated with the most extreme developmental trajectories of anxious behavior (i.e., a higher base level at age 2 and steady increments over time to age 11) in another large Canadian sample. Therefore, it seems plausible that family dysfunction would also operate as an important determinant of Chinese adolescents' anxiety and depression.

Although family dysfunction has been linked with adolescents' anxiety and depression, few studies have explored the possible mediation effects for these associations. Mediation refers to the “transmission” of the effect of an explanatory variable on an outcome variable through an intermediate variable termed the mediator (Mackinnon, Fairchild, & Fritz, 2007). The current mediation literature discusses two different types of mediation (Nitzl, Roldan, & Cepeda, 2016): full mediation (i.e., only the indirect path is statistically significant) and partial mediation (both the direct and indirect paths are statistically significant). The mediation model may potentially explicate the specific psychological mechanism or mechanisms that underlie the association between an antecedent variable and outcomes via the inclusion of mediating factors (Mackinnon et al., 2007). Exploring such mechanisms could answer the question of how family dysfunction is translated into adolescents' anxiety and depression. Based on a social-cognitive model, which integrates social situations, mediating cognitive mechanisms, and individual differences (Wood et al., 2008), our study focused on adolescents' perceptions about themselves (i.e., self-esteem) and interpersonal relationships with parents and peers (i.e., loneliness) as key cognitive mediators linking family dysfunction and mental health outcomes among Chinese adolescents.

Self-esteem, defined as a global positive or negative view of the self (Rosenberg, 1965), is one aspect of an individual's self-concept. It emerges from individuals' interactions with society, family members, and other people, such as teachers and peers (Baumeister & Bushman, 2010; Rosenberg, 1965). The family environment represents the earliest and most proximal environmental context for children (Harter, 2012), and family functioning is highly related to children's self-esteem. Individuals develop perceptions of self-worth through participating in problem solving, communicating with parents, and incorporating family roles, all of which are important aspects of family functioning (Ferro & Boyle, 2015). Additionally, adolescents' self-esteem increases when they experience their parents' acceptance and emotional support, and such acceptance and emotional support reflect the aspects of affective responsiveness and affective involvement in family functioning (Yen, Yang, Wu, & Cheng, 2013). Therefore, family dysfunction was expected to be an important determinant of children's self-esteem. Empirical research has also supported this relation. For example, using a sample of 816 children aged between 7 and 16, Wu et al. (2015) revealed that family dysfunction was associated with low self-esteem.

In addition, low self-esteem is an important predictor of subsequent psychopathology (Arslan, 2016). According to the vulnerability model, low self-esteem contributes to anxiety and depression (Sowislo & Orth, 2013). The underlying assumption of the vulnerability model is that self-esteem, like other personality traits, is a diathesis exerting causal influences on the onset and maintenance of psychopathology (Orth, Robins, & Roberts, 2008). For depression, low self-esteem contributes to the processes of reassurance seeking, negative feedback seeking, and rumination, which are theoretically linked to depression (Sowislo & Orth, 2013). For anxiety, as suggested by terror management theory, high self-esteem alleviates anxiety elicited by awareness of human mortality (Pyszczynski, Greenberg, Solomon, Arndt, & Schimel, 2004). In addition, terror management theory postulates that all anxiety is ultimately derived from, and related to, the fear of death (Pyszczynski et al., 2004). Therefore, high self-esteem might be associated with the mitigation of all anxiety. Many empirical studies using longitudinal designs have shown a substantial negative relation between self-esteem and individuals' depression (see Sowislo & Orth, 2013, for a review). For instance, using a large longitudinal data set with four repeated assessments, Orth et al. (2008) found that low self-esteem operated as a risk factor for depression. Furthermore, a meta-analysis of longitudinal studies demonstrated that low self-esteem also predicted anxiety (Sowislo & Orth, 2013).

In addition to the empirical support for the separate parts of the mediating paths (i.e., from family dysfunction to self-esteem and from self-esteem to anxiety and depression), a few studies have evaluated the mediating roles of self-esteem between similar social contexts and psychological outcomes. For instance, Yen et al. (2013) found that poor family functioning was related to low self-esteem among Chinese adolescents, and that lower levels of self-esteem were related to higher levels of social anxiety. Also, Hu and Ai (2016) observed that self-esteem partially mediated the link between parent-adolescent relationships and depression in Chinese junior high school students. Finally, using latent growth curve modeling, Ferro and Boyle (2015) found that self-esteem partially mediated the path from family functioning to children's anxiety and depression in Canadian children of ages 10 to 15 years. Overall, children who live in higher levels of dysfunctional families are more likely to report lower self-esteem, and in turn are inclined to experience higher levels of depression and anxiety.

Loneliness is defined as feelings and thoughts of being isolated and disconnected from others (Russell, Salazar, & Negrete, 2000). Loneliness can occur in all periods of development, but it is experienced most intensely during adolescence (Antognoli-Toland, 2001). Family function theory posits that the quality of the family environment is associated with the formation of loneliness (Miller et al., 2000). Specifically, loneliness is associated with dysfunctional family experiences because adolescents' needs for intimacy or close relationships cannot be met in such a family environment (Wu & Chow, 2013). Moreover, when adolescents are over-controlled and misunderstood by their parents, adolescents often report feeling lonely (Frisén, 2007). In addition, in dysfunctional families, adolescents likely experience fewer positive social interactions to observe and model, so they are less able to develop effective and lasting relationships with peers (NICHD Early Child Care Research Network, 2009). Given that peer relationships serve as social capital and are highly influential during this developmental stage (Farrell, Thompson, & Mehari, 2017), a perceived lack of intimate peer relationships would especially increase adolescents' feeling of loneliness.

Adolescents with higher levels of loneliness are more likely to experience anxiety and depression. Despite the overlap in common features of loneliness and depression, they appear to be distinct phenomena (Heinrich & Gullone, 2006). Loneliness is more prevalent than depression (Erzen & Çikrikci, 2018). It is a universal experience, which can occur in differing degrees between individuals and within individuals at differing stages in their lives. Moreover, loneliness involves a specific appraisal of the social domain in one's life, whereas depression involves appraisals that are more global and heterogeneous, covering multiple life domains (Heinrich & Gullone, 2006). Lonely individuals report beliefs, such as, “I have no one to talk with.” and “I feel that others do not pay attention to me.”. These beliefs, which often reflect cognitive errors, may yield a sense of depression and even suicidal thoughts (Pössel, 2017). Lonely adolescents display a fear of negative evaluations because they believe that nobody likes them, which may increase their anxiety. Empirical investigations have linked loneliness to anxiety and depression (see Michalska da Rocha, Rhodes, Vasilopoulou, & Hutton, 2017 for a review). For example, research on left-behind children in China (i.e., rural children left at home when one or both of their parents migrate to urban areas for work), has also shown that the frequent experience of loneliness is significantly related to their levels of anxiety and depression (Yuan, Jin, & Yang, 2014). In a longitudinal study with a sample of 296 British children, Qualter, Brown, Munn, and Rotenberg (2010) established prospective links between loneliness and depression over an 8-year period from early/middle childhood to adolescence.

A few studies have evaluated the mediating roles of loneliness between social contexts and psychological outcomes. For instance, using a cross-sectional design, He, Zhou, Li, Cao, and Guan (2014) found that loneliness played a partial mediating role between social support and depression in a sample of students displaying internet addictions. Overall, individuals from dysfunctional families appear less likely to feel understood or supported in their interpersonal interactions with family members and peers, and are in turn more inclined to report higher levels of anxiety and depression.

Self-esteem and loneliness are the two focal mediators in this study. They are distinguishable, but interrelated for several reasons. First, adolescents with low self-esteem are more likely to experience feelings of alienation and detachment because of their beliefs that they are not worthy of love (Nayak & Kochar, 2016). Several studies have verified the path from self-esteem to loneliness (e.g., Kong & You, 2013; Liu, Shen, Xu, & Gao, 2013). Second, adolescents with feeling of loneliness may attribute the cause of their loneliness to “not being good enough” (Nayak & Kochar, 2016). Some studies have verified the path from loneliness to self-esteem (e.g., Yildiz, 2017). Third, both self-esteem and loneliness might be associated with another predictor (Nayak & Kochar, 2016). For instance, Nayak and Kochar (2016) found in a sample of undergraduate students that both self-esteem and loneliness were predicted by parenting style. Therefore, given that loneliness and self-esteem are both possible mediators and they are interrelated, it seems necessary to consider the mediating roles of loneliness and self-esteem simultaneously, thus controlling for the correlation between loneliness and self-esteem.

Gender differences have been documented in the levels of anxiety and depression among adolescents (Nivard et al., 2015). Compared to boys, girls are more likely to experience higher levels of anxiety and depression (Nivard et al., 2015), and lower levels of self-esteem in adolescence (Bleidorn et al., 2016). In addition, gender differences might also moderate the links between variables. For instance, the association between family functioning and loneliness was significant only for girls (Shi, Wang, & Zou, 2017); the associations between loneliness and anxiety/depression appeared stronger among women than men (Chang, 2018); and the association between self-esteem and depression was stronger for girls than for boys (Moksnes & Espnes, 2012). A possible explanation for such findings might be that girls display a stronger interpersonal orientation and are more sensitive to relationships (Shi et al., 2017). However, the association between some variables did not differ as a function of gender. Specifically, family functioning was a very strong predictor of self-esteem for both genders (Mandara & Murray, 2000), and the associations between self-esteem and anxiety also did not differ for genders (Moksnes & Espnes, 2012). Considering the possible gender differences in some associations, we tested for gender differences in the multiple mediation model to see whether the overall pattern of effects differed for boys and girls.

Beyond gender, there were other demographic variables that potentially related to our study variables. Specifically, age and SES (i.e., father's and mother's work status and education level) have been revealed to be associated with family dysfunction, loneliness, self-esteem, anxiety, and depression (e.g., Martin et al., 2010; Nayak & Kochar, 2016; Sowislo & Orth, 2013). Therefore, these demographic variables (i.e., gender, age, and SES) were treated as covariates in our study.

Overall, based on the MMFF, higher levels of family dysfunction would be related to higher levels of anxiety and depression among adolescents. Furthermore, based on a social-cognitive model, students' perceptions of high loneliness and low self-esteem might account for these associations. However, few studies have evaluated the direct effects of family dysfunction on adolescents' anxiety/depression or the multiple mediating roles of self-esteem and loneliness using longitudinal designs, limiting the confidence in the existing results about the associations.

Therefore, we aimed to examine a multiple mediation model in which loneliness and self-esteem were hypothesized to mediate the relations between family dysfunction and Chinese adolescents' anxiety and depression using a three-wave longitudinal design. In addition, given that gender might moderate some associations in such a multiple mediation model, we also aimed to examine gender differences in the mediation model as a supplementary analysis. The multiple mediation model is a conservative statistical approach for testing the full model, because it allows evaluation of the indirect effect of a specific mediator in the presence of other mediators, and can reduce the likelihood of parameter bias due to omitted variables (Preacher & Hayes, 2008). In addition, compared with a simple mediation model, the multiple mediation model allows for simultaneous examinations of multiple psychological mechanisms, illuminating complementary indirect paths, and clarifying the complex interplay among the variables that promote the development of anxiety and depression.

Specifically, we formulated the following specific hypotheses (see Fig. 1): (1) Higher levels of family dysfunction at Time 1 will predict Chinese adolescents' anxiety and depression at Time 3. (2) Self-esteem and loneliness at Time 2 will simultaneously mediate the relations between family dysfunction at Time 1 and Chinese adolescents' anxiety and depression at Time 3. (3) As for the supplementary analysis, based on the previous literature, we hypothesized that the various associations would be stronger for girls than for boys.