Trajectories of Loneliness and Aggression From Childhood to Early Adolescence in Taiwan: The Roles of Parenting and Demographic Predictors

Abstract

This study examines three parenting behaviors (support, involvement, and harsh control) in predicting children’s loneliness and aggression across developmental transitions to adolescence in a Taiwanese sample. Two cohorts (n = 4,990) were followed for 5 years: a younger cohort (first-graders, 51.5%) and an older cohort (fourth-graders). Multilevel modeling for change across time was used to examine the baseline and rate of change of child outcomes. Overall, loneliness decreased and aggression increased across time. Maternal involvement predicted less loneliness at the baseline and a faster decreasing rate of loneliness. Harsh control predicted more aggression and support predicted less aggression at the baseline. Girls showed a faster increasing rate of aggression and slower declining rate of loneliness than boys. The findings highlight the need to identify effective parenting behaviors for loneliness and aggression with acknowledgment of cultural beliefs, and address girls’ aggression. Implications for prevention/intervention in children’s social functioning problems were discussed.

Keywords

parenting loneliness aggression developmental trajectory cultural values

Introduction

Loneliness and aggression are internal and external forms of impaired social functioning. Internally, children feel lonely when they perceive a discrepancy between their desired and actual levels of social relationships (Peplau & Perlman, 1982). Research has found that loneliness is associated with several adjustment problems, such as low self-esteem, victimization, depression, and suicidal ideation (Jones, Schinka, & van Dulmen, 2011; McWhirter, Besett-Alesch, Horibata, & Gat, 2002; Storch & Masia-Warner, 2004). Externally, aggression is a visible sign of deficits in social competence and relationships and is linked to lower levels of prosocial behavior, social acceptability and popularity, and maladjustment (Card, Stucky, Sawalani, & Little, 2008; Schoffstall & Cohen, 2011). Therefore, the first aim of the present study is to investigate the developmental trajectories of loneliness and aggression across time to help us better understand how both internal and external forms of social functioning problems change across developmental transitions.

A child first establishes relationships with other human beings in his/her family, and the mother is generally the child’s initial point of contact. That is, children’s social functioning first develops within families through children’s interactions with their mothers. Bronfenbrenner’s (2001) ecological theory of developmental processes and process-person-context-time model (Bronfenbrenner & Morris, 1998) suggests that interactions between the individual and the immediate environment (e.g., family) constitute the “primary engines of development.” Because family interactions are central to a child’s development, this article focuses on three fundamental patterns of family interaction: support, involvement, and harsh control. We developed an index of parenting behaviors to approximate the responsiveness and demandingness dimensions suggested by Baumrind (1971) and Maccoby and Martin (1983). “Support” includes encouragement, praise, and comfort. “Involvement” includes shared activities, such as eating meals together or working on homework together. “Harsh control” is characterized by parents’ strict and punitive control techniques—practices that are geared toward establishing children’s unquestioning obedience. The second aim of the present study is to examine ways that maternal parenting behaviors affect children’s loneliness and aggression at the baseline and across elementary and middle school years.

Moreover, parenting and children’s social functioning may reflect cultural context. In collectivist cultures, such as the East Asian culture of China and Taiwan, individuals emphasize their connections to other members of society (Triandis, 1995). Collectivist cultures consider parental control and an authoritarian parenting style to be less destructive (Chao, 1994) compared with individualist cultures, such as those in the West. Most studies on parenting and social functioning have been conducted in Western cultures so less is known about these relationships in the East, and therefore, the present study aim to study these associations across developmental transitions in a Taiwanese (East Asian) sample.

Children’s Loneliness Across Time

Loneliness is one of the major internalizing behaviors in childhood and adolescence. Loneliness is a distress and aversive feeling in one’s social relationship that most people experience at some point in their lives. For some people, loneliness is a prolonged experience, but for most people, loneliness is a transient experience. Individuals in different developmental stages face different challenges in their social relationships, and thus, sources of loneliness can differ at different developmental stages because of changes in belonging needs (Parkhurst & Hopmeyer, 1999). Individuals’ belonging needs and expectations of friendships focus on enjoying activities together in early childhood, being liked by a close friend in middle childhood, being accepted by a peer group in early adolescence, and having a romantic relationship in late adolescence (Qualter et al., 2015). The levels of loneliness may change over time and across developmental transitions. The present study investigates ways that loneliness may change over time during the developmental transitions for two cohorts from early childhood to early adolescence and from middle childhood to adolescence, respectively. Using time as a factor, how does children’s loneliness change over time? When looking at the average trajectory of loneliness, most of the studies conducted in Western societies found decreasing patterns of change. A study in the United States found that children’s loneliness decreased from Grade 3 through Grade 5, and adolescents’ loneliness decreased from Grade 5 through age 15 (Jones et al., 2011). Another longitudinal study in the United Kingdom reported decreasing loneliness over time for children aged 8 through 11 (Harris, Qualter, & Robinson, 2013). Adolescents in the Netherlands also exhibited a slightly decreasing trajectory of loneliness from ages 15 through 20 (Vanhalst, Goossens, Luyckx, Scholte, & Engels, 2013). Ladd and Ettekal (2013) reported a similar declining trajectory from early adolescence through late adolescence, with 12th graders (age 18) perceiving less loneliness than sixth graders (age 12), on average. The aforementioned longitudinal studies conducted in Western societies suggest a deceasing trajectory of loneliness over time.

Children’s Aggression Across Time

Aggression is one of the major externalizing behaviors in childhood and adolescence. Overt aggression is characterized by physical and verbal attacks and norm-breaking behaviors, such as hitting, fighting, cursing, and destroying things. The present study investigates ways that aggression may change over time, on average, for two cohorts during their developmental transitions across elementary and middle school years. Using time as a factor, how does children’s aggression change over time? A longitudinal study in Canada categorized low socioeconomic status boys into three different levels of physical aggression (low, medium, and high) and examined the patterns of change from age 6 to age 13 (Brame, Nagin, & Tremblay, 2001). During the transition from childhood to adolescence, they found that half of the boys reported a relatively low level of aggression with a decreasing trajectory, 30% reported a declining trajectory with medium aggression, and 20% reported a declining trajectory of high aggression. For younger children, one study identified either stable or decreasing trajectories of physical aggression from toddlerhood to age 9 (Campbell, Spieker, Burchinal, & Poe, 2006). Other studies indicate that physical aggression decreases slightly over time as children grow older and reach adolescence (Lacourse et al., 2002; Nagin & Tremblay, 1999). In short, the aforementioned longitudinal studies conducted in Western societies suggest a deceasing trajectory of aggression over time.

Parenting and Children’s Loneliness

Children’s loneliness can be understood via social needs approaches, and the connections between parenting and child loneliness can be explained using attachment theory. According to the social needs approach (Weiss, 1973), loneliness occurs when individuals lack an interpersonal relationship, have deficits in social integration or nurturance, and/or cannot meet their own social needs (Weiss, 1987). How do parenting and parent-child relationships promote or deter children’s social functioning? Attachment theory (Bowlby, 1969) argues that when individuals have an insecure attachment with their caregiver, they view themselves and others as unacceptable and unworthy (Cassidy, 2000). This internal working model diminishes children’s self-esteem and trust regarding self and others and hinders the ability to approach others and form desired relationships with others to meet their social needs; in turn, this internal model can lead to loneliness in childhood and adolescence (Chipuer, 2001).

Previous research has found that lack of maternal support has a direct effect on adolescents’ loneliness at a baseline, but not on a slope (which represents change; van Roekel, Goossens, Scholte, Engels, & Verhagen, 2011). In contrast, restrictive harsh parenting and pathological harsh control are linked to higher levels of children’s loneliness (Annear & Yates, 2010; Minzi, 2006). Children living under harsh control have been shown to have fewer prosocial skills and competencies than other children when interacting with peers (Kinard, 1999), to be less liked by their peers (Dodge, Pettit, & Bates, 1994), and to experience more social withdrawal and isolation (Nelson, Hart, Wu, et al., 2006), which may, in turn, lead to loneliness.

Parenting and Children’s Aggression

Both too much and too little control can be problematic. Harsh control predicts later externalizing behaviors, such as aggression among school-age children and adolescents (Kuppens, Grietens, Onghena, & Michiels, 2009; Nelson, Hart, Yang, Olsen, & Jin, 2006). The connections between harsh control and child aggression can be explained using the spillover hypothesis, social learning theory, and social-information processing theory. According to the spillover hypothesis, behavior such as aggression that children learn from one setting (e.g., family) is likely to transfer to other settings (e.g., school). In addition, social learning theory (Bandura, 1977) argues that negative parent-child interactions and parenting behaviors (such as harsh control) are associated with children’s aggressive behaviors via observational learning, modeling, and coercive family processes (Parke, MacDonald, Beitel, & Bhavnagri, 1988; Patterson, 1982). When parents use harsh control at home, children become more likely to use coercive and aggressive behaviors when interacting with others (Patterson, 1995). Moreover, drawing on social-information processing, Dodge, Pettit, Bates, and Valente (1995) proposed that children who experience harsh control have deficits in encoding relevant social cues. These deficits may lead to hostile attributions for their peers’ behaviors and intentions, and then the children respond with aggressive behavior. A longitudinal study in a sample of 1,993 Canadian children aged 6 to 12 revealed that mothers’ harsh and controlling parenting increased the odds of following a trajectory of high levels of physical aggression, even above and beyond other risk factors, such as a reactive temperament and parental separation (Joussemet et al., 2008). In contrast, higher perceived parental support predicted lower levels of aggression at age 15 in the Netherlands (Van der Graaff, Branje, De Wied, & Meeus, 2012). Similarly, maternal involvement is negatively associated with aggression for boys and girls in the Midwest of the United States (Simons, Johnson, & Conger, 1994).

Child Gender and Outcomes

In addition to family context, the process-person-context-time ecological model (Bronfenbrenner & Morris, 1998) emphasizes the importance of individual characteristics in understanding development. Individual characteristics such as gender have the “capacity to invite or discourage reactions from the social environment.” For example, studies have found that boys are more likely than girls to exhibit overt aggression in childhood and adolescence (Blitstein, Murray, Lytle, Birnbaum, & Perry, 2005; Hyde, 1984; Maccoby & Jacklin, 1980). Furthermore, boys’ physical characteristics and displays of masculinity may influence their aggressive behaviors, and their desire to appear strong and capable promotes physical aggression (Pepler & Craig, 2005). In contrast, girls’ sensitivity to social control and social expectations of gender role may deter physical aggression. Overt aggression may be perceived as a violation of the feminine stereotype (Crick, Bigbee, & Howes, 1996). Thus, in the present study we expected that girls would display less aggression than boys, and we controlled for gender in the aggression model. With regard to the rate of change, although boys show higher initial levels of physical aggression than girls at age 11, the trajectories of aggression between boys and girls are parallel (Karriker-Jaffe, Foshee, Ennett, & Suchindran, 2008).

In terms of loneliness, the findings for gender differences are mixed. Among adolescents, some studies show that males reported higher levels of loneliness than females (Rich, Kirkpatrick-Smith, Bonner, & Jans, 1992), whereas other studies revealed that females reported higher levels of loneliness than males (Brennan, 1982) or found no gender differences in loneliness (Koenig & Abrams, 1999). In the present study, we controlled for gender in the loneliness model and expected that girls would experience less loneliness than boys, because girls have closer and more intense group friendships than boys (Pepler & Craig, 2005). Furthermore, girls’ levels of social skills are higher than those of boys (Deniz, Hamarta, & Ari, 2005), with girls being generally more socially responsive, empathic, and intimate (Wheeler, Reis, & Nezlek, 1983). In addition, previous research has found gender differences in social support, with girls receiving more social support than boys (Call & Mortimer, 2001).

The Role of Culture in Parenting

Relationships between parenting and child social functioning should be considered within cultural contexts. We may first acknowledge some unique cultural beliefs with regard to the purpose and interpretation of being harsh and critical, and link those beliefs to Chinese and Taiwanese mothers’ use of harsh control when disciplining their children or pushing their children toward success. Chinese parents have little tolerance for “losing face” (Chao, 1994, 2001), and tend to compare their children with other children in various circumstances. To help their children improve and achieve goals and win competitions, Chinese parents often regard harsh control and harsh criticism as an acceptable and efficient way to discipline and educate children. Because the parents’ ultimate goal is to help their children succeed, harsh control is considered acceptable and widely used in traditional Chinese culture. Moreover, harsh control is sometimes even viewed by Chinese parents as a demonstration of involvement and concern and an expression of care or love (Chao, 1994). Chao (1994) described the unique parenting concepts of harsh control, chiao Shun (training) and guan (to govern or to care for), which are viewed as being very negative by the European-American mothers (were associated with “strict,” “militaristic,” or “regimented” adjectives), and are viewed as being very positive by the Chinese mothers (were associated with “teaching,” “educating,” or “inculcating” adjectives). Therefore, harsh control in childrearing tend to be considered normative among Chinese parents and thus less likely to lead to problem behaviors than among Western parents (Chao, 1994; Chiu, 1987). Some studies have even found the associations between parental harsh control and positive outcomes such as academic achievement and less depression among Chinese adolescents (Li, Costanzo, & Putallaz, 2010). On the other hand, living within these cultural norms, Chinese children appear to have higher level of conformity toward authority figures (parents) and are more obedient, compared to American children (Leung & Fan, 1996).

In sum, the goals of the present study are to (a) investigate how loneliness and aggression change over time during developmental transitions across elementary and middle school years, and (b) examine how parenting behaviors predict both the baselines and rate of changes (slopes) of children’s loneliness and aggression across 5 years. In terms of developmental trajectories, we hypothesize the following: first, a slightly decreasing trajectory of loneliness as children grow older and become more competent to form and maintain friendships, especially in the cultural context that educates and emphasizes mutual dependence and collectivity, and second, a slightly decreasing trajectory of aggression as children age and are able to regulate their emotions and behaviors better, especially under years of Chinese parenting in educating and promoting harmony, rule obedience, and aggression suppression. In terms of a mesosystem ecology of loneliness and aggression between individuals and families, we hypothesize the following: third, higher levels of maternal support and involvement to predict less loneliness and less aggression in children and adolescents. Fourth, unlike previously published findings from Western societies, we hypothesize that harsh control does not predict loneliness and aggression among Taiwanese children, given the unique cultural conceptions of harsh control (chiao Shun and guan). Finally, we expect girls to display less loneliness and less aggression than boys.

Method

Participants

The data used in this study were drawn from the Children and Adolescent Behaviors in Long-Term Evolution (CABLE) study, which was conducted by National Taiwan University and Taiwan’s National Health Research Institutes (Yen, Chen, Lee, Hsiao, & Pan, 2002). The CABLE study was approved by the Institutional Review Board of the National Health Research Institutes. The present study focuses on the first 5 years of the CABLE study (2001 through 2005) and follows two cohorts of children over time. The reason for having two cohorts is that the first graders (the younger cohort) were the first group to experience a reform of the educational system in Taiwan. In order to gain an understanding of the influences on student outcomes under the old and new educational systems, an additional group of fourth graders (the older cohort) was chosen for the project. The data were collected annually for the younger cohort from Grades 1 to 5 (ages 6 through 10) and in the older cohort from Grades 4 to 8 (ages 9 through 13).

In the first year of the CABLE study, 18 public elementary schools were randomly selected from an urban area in Taipei (nine schools) and a rural area in Hsinchu County (nine schools). The random sample reflected the following distribution of school size to ensure that the numbers of children chosen from each type of school was about equal: six small schools (50-199 children), two medium-sized schools (200-399 children), and one large school (more than 400 children). With the assistance of each school’s contact person, all teachers of the first- and fourth-grade children distributed informed consent forms for parents to complete. Only those children whose parents and children agreed to participate in the study were included in the sample. In Year 1, 2,218 out of 3,584 possible first graders (61.9 %) and 2,075 out of 3,612 possible fourth graders (57.4 %) had obtained parents’ consent and participated in this longitudinal study (52% boys). In Year 2, 1,264 additional children from the original 18 schools joined the study, and 327 children who participated in Year 1 dropped out. Therefore, a total of 5,218 children participated in Year 2. In Year 3, 5,021 children participated; in Year 4, this figure was 4,633; and in Year 5, a total of 4,607 children remained. The attrition rate was 12%. To test whether children included in the models tested differed from children excluded because of missing data, the two groups were compared for each of the study variables. The two groups significantly differed in the following ways. Children who were missing data and were not included in the analyses had parents with fewer years of education ( $\bar{X}$ = 11.41, SD = 2.95 vs. $\bar{X}$ = 12.87, SD = 2.39; p < .001), lower levels of maternal support ( $\bar{X}$ = 3.26, SD = .58 vs. $\bar{X}$ = 3.32, SD = .48; p < .05), and lower levels of maternal involvement ( $\bar{X}$ = 2.8, SD = .57 vs. $\bar{X}$ = 2.9, SD = .49; p < .01), compared to children who were included in the analyses. Children’s mothers also were invited to participate in the study. The number of mothers participating in the study was 3,856 in Year 1, 4,744 in Year 2, 4,612 in Year 3, and 4,075 in Year 4. Only baseline scores of mothers’ reports were used for analyses in this study. There are no differences in children’s outcomes between children with and without mothers’ reports.

Procedures and Measurements

Every fall for five successive years, children in the CABLE study completed a pen-and-paper survey in their classrooms to assess their loneliness and aggression. Trained project staff read the questions and responses slowly to the children who were younger than 9 years old (fourth grade); older children completed the questionnaire by themselves and received help from project staff if necessary. In Year 4, the older cohort graduated from elementary school and went to various middle schools. Thus, some of the children received the survey by mail and individually completed the questionnaire for the follow-up. After completing the questionnaire at school each year, children brought home parenting questionnaires for their mothers. The mothers completed the questionnaires to assess their parenting behaviors and reported their education and family income. Mothers returned the questionnaires to the school in provided envelopes.

Loneliness and aggression

Children completed the CABLE Social Loneliness Scale and Aggression Scale (Yen et al., 2002) questionnaire. This questionnaire was adapted from the Child Loneliness Scale (Asher, Hymel, & Renshaw, 1984) and the Youth Risk Behavior Survey (Centers for Disease Control and Prevention, 2001), with consideration of social-cultural context and age-appropriate rewording. Pilot studies also had been conducted to ensure the reliability and validity of the questionnaire, and then the questionnaire was carefully refined (Yen et al., 2002; Yen, Chiu, Wu, & Pan, 2006).

The CABLE Loneliness Scale consists of five items on a 3-point rating scale (1 = I do not feel this way, 2 = once or twice, and 3 = many times). Children reported how frequently in the past 2 weeks they experienced the followings: Did you feel that you did not know who to talk to? Did you feel that you were alone and had nobody to be with? Did you feel that other peers would not play with you when they were playing? Did you feel that you do not have friends? Did you feel lonely when you were by yourself? Scale scores were computed as the mean of the items. Cronbach’s alphas for each of the 5 years were .63, .71, .75, .76, and .79, respectively.

The CABLE Aggression Scale consists of four items on a 4-point rating scale (1 = no, 2 = one or two days, 3 = many days, and 4 = every day). Children reported how often in the last month they had engaged in the following activities: Did you curse or say dirty words? Did you quarrel with others? Did you hit someone? Did you destroy things when you got angry? Scale scores were computed as the mean of the items. Cronbach’s alphas for each of the 5 years ranged from .64 to .69. To test the construct validity, the factor structure was evaluated using principal-axis-factoring analysis with promax rotation. The results support the hypothesized two factors, that is, loneliness (five items) and aggression (four items), for each of the 5 years. Using the Year 5 data as an example, the items loaded from .65 to .82 for loneliness and .66 to .75 for aggression. The two factors explained 53% of the variance.

Three types of parenting behaviors

Mothers completed the CABLE Family Interactions Scale questionnaire, which is a 17-item questionnaire with three subscales, to assess maternal support, involvement, and harsh control (Yen et al., 2002). The questionnaire was developed by a panel of Taiwanese experts by adapting a measure created for Western samples (Lamborn, Mounts, Steinberg, & Dornbusch, 1991), with consideration of social-cultural context. To validate the measures for Taiwanese samples, a pilot study of 84 fourth graders was conducted, and the validity and reliability information from the pilot study was used to modify the questionnaire to develop the final version; further validation information was reported (Wu, Kao, Yen, & Lee, 2007; Yen et al., 2002).

The parenting questionnaire contains five support items, five involvement items, and seven harsh control items. Mothers reported how frequently they practiced each type of parenting behavior on a 4-point Likert scale (1 = no, 2 = once or twice, 3 = many times/days, 4 = every time/day). The five maternal support items are as follows: In the past month, did you encourage your child when he or she was having trouble doing something? Did you praise your child when he or she was good or performed well? Did you comfort your child when he or she was in a bad mood? Did you look after your child when he or she was sick? Have you cared about what happens to your child at school? The five involvement items relate to how often the respondents had engaged in the following activities during the past week: Did you have a chat with your child? Did you eat with your child (any meal during the day)? Did you do housework with your child? Did you help your child with his or her homework? Did you take your child out for a fun activity? The seven harsh-control items relate to how often the respondents used specific punishments for their children’s misbehavior, as follows: In the past month, did you lock your child in the bathroom or bedroom when he or she misbehaved? Did you prohibit your child from doing something he or she likes? Did you take away your child’s pocket money or toys? Did you make your child do housework or write words over and over? Did you yell at your child? Did you make your child stand or kneel? Did you hit your child? Cronbach’s alphas for maternal support, involvement, and harsh control were .82, .77, and .74, respectively.

To test the construct validity, we evaluated the factor structure of the parenting measures by using principal-axis-factoring analysis with promax rotation. Using the Year 5 data as an example, the analysis supported the hypothesized three-factor solution with five support items loading on the first factor (range from .68 to .85), seven harsh-control items loading on the second factor (range from .58 to .68), and five involvement items loading on the third factor (range from .63 to .78). The three factors explained 50% of the variance.

Maternal education and family income

Mothers provided information about their education level and monthly family income. Education level ranged from 1 (elementary school) to 7 (graduate school), and it was recoded to be years of education for analysis. Monthly family income range from 1 (less than $624, the poverty line) to 9 (more than $5,000). Education level and family income were asked each year, but only the baseline score was used in the analyses.

Analytic Plan

For the current study, we used two-level multilevel modeling (also called hierarchical liner modeling [HLM]; Raudenbush & Bryk, 2002) to model changes across time (Singer & Willett, 2003) for loneliness and aggression, with equal spacing between measurement occasions (five waves, 5 years), and to model the associations between independent variables and the two child outcomes (loneliness and aggression) over time. Separate models were estimated for child outcomes. This study operated on the assumption that repeated measures of loneliness and aggression were nested within individuals. Level 1 of each model describes a child’s changes over time (within-person), including the initial status (intercept) and the rate of change (slope). Time was rescaled from 1-5 to 0-4. The first-year assessment thus represents the intercept. Level 2 describes the different patterns of change across children (between-person) based on the predictors.

The first step in the analyses was to estimate unconditional growth models to examine if there was significant variability in trajectories of loneliness and aggression. Repeated assessments of loneliness and aggression were modeled as a function of time. The variable Time represented the time in years at each assessment and it was within-child. In terms of variance, the Level 1 residual variance ( $σ_{ε}^{2}$ ) summarizes the average scatter of a child’s observed outcome values around his or her own true change trajectory. The Level 2 variance components quantify the amount of unpredicted variation in the individual growth parameters: $σ_{0}^{2}$ assesses the unpredicted variability in true initial status; $σ_{1}^{2}$ assesses the unpredicted variability in true rates of change (Singer & Willett, 2003). The next step in the analyses was to estimate conditional models to examine between-child associations between parenting behaviors and child outcomes across elementary and middle school years, when controlling for child and mother and household characteristics. To test whether between-child differences in loneliness and aggression over time were associated with parenting behaviors, measures of three baseline parenting behaviors (involvement, support, and harsh control) were added as predictors in both models. In addition, measures of child characteristics (gender and cohort) and baseline mother (education) and household (family income) characteristics were included in the models as well to estimate between-child effects. All the variables at Level 2 were centered on the grand mean for the sample so that the Level 2 intercepts represent adjusted means for the average participants in the sample. Using grand mean centering gives the intercept and slope terms a meaningful interpretation. These seven predictors were tested simultaneously in both loneliness and aggression models and had their own unique contributions to the models.

Results

The preliminary analysis results indicate that the variables were not skewed (values between +2 and −2); however, some variables, such as loneliness in Year 5 and aggression in Years 1, 2, and 4, were affected slightly by kurtosis (values slightly over 2; Gravetter & Wallnau, 2014). The distributions for these variables have heavier tails, but were not considered problematic (values below 8; Kline, 2016). Table 1 includes the means, standard deviations, and correlations for all variables. While maternal harsh control was not correlated to children’s loneliness, maternal support and involvement and education and family income were negatively correlated to loneliness. On the other hand, while harsh control was positively correlated to children’s aggression, maternal support and involvement and education and family income were negatively correlated to children’s aggression. In addition, children’s loneliness was correlated to aggression.

Table 1.

Bivariate Correlations, Means, and Standard Deviations for Variables in the Models of Loneliness and Aggression.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
1. Child gender	—
2. Child cohort	.01	—
3. Support	.02	−.18**	—
4. Involvement	−.03*	−.19**	.57**	—
5. Harsh Control	.17**	−.05**	.02	.08**	—
6. Loneliness (Y1)	.01	.16**	−.05**	−.07**	.01	—
7. Loneliness (Y2)	−.06**	.12**	−.08**	−.10**	−.01	.29**	—
8. Loneliness (Y3)	−.02	.08**	−.06**	−.06**	.01	.22**	.38**	—
9. Loneliness (Y4)	−.04**	−.13**	−.04*	−.02	.01	.13**	.25**	.36**	—
10. Loneliness (Y5)	−.11**	.05**	−.05**	−.03	.01	.18**	.27**	.33**	.39**	—
11. Aggression (Y1)	.15**	.17*	−.07**	−.07**	.05**	.20**	.10**	.08**	.05**	.07**	—
12. Aggression (Y2)	.16**	.14**	−.09**	−.08**	.11**	.14**	.20**	.15**	.09**	.11**	.34**	—
13. Aggression (Y3)	.11**	.22**	−.10**	−.10**	.07**	.13**	.15**	.23**	.13**	.15**	.26**	.46**	—
14. Aggression (Y4)	.10**	.15**	−.11**	−.08**	.09**	.11**	.11**	.16**	.20**	.17**	.23**	.36**	.49**	—
15. Aggression (Y5)	.09**	.19**	−.11**	−.08**	.09**	.12**	.10**	.14**	.13**	.20**	.21**	.33**	.41**	.51**	—
16. Education	.02	−.09**	.27**	.18**	−.01	−.06**	−.05**	−.07**	−.02	−.03	−.01	−.04*	−.04**	−.03*	−.07**	—
17. Family income	.04**	−.02	.23**	.11**	−.03	−.03*	−.04**	−.07**	−.05**	−.06**	−.02	−.02	−.05**	−.04**	−.06**	.57**	—
$\bar{X}$	—	—	3.31	2.89	1.90	1.53	1.41	1.36	1.35	1.30	1.50	1.51	1.59	1.60	1.67	12.78	4.54
SD	—	—	0.49	0.50	0.37	0.46	0.44	0.42	0.43	0.42	0.58	0.50	0.51	0.50	0.52	2.4	2.0
n	5,540	5,540	5,202	5,209	5,163	4,266	5,214	5,021	4,623	4,606	4,274	5,214	5,020	4,632	4,606	5,307	5,216

Note. SD is standard deviation. Y1 is Year 1.

p < .05. **p < .01.

Loneliness and Aggression Trajectories Across Elementary and Middle School Years

The analyses began by estimating unconditional growth models to determine whether there was significant variability in initial status and growth of loneliness and aggression over time. The average loneliness score at the baseline for the average child in the sample was 1.49 (SE = .01, p < .001) on a 3-point rating scale, and the average aggression score for the average child in the sample was 1.48 (SE = .01, p < .001) on a 4-point rating scale. The average rate of change in loneliness across time points was −.05 (SE = .00, p < .001). The negative and significant coefficient on the slope indicates that loneliness trajectories decreased across elementary and middle school years. The average rate of change in aggression across time points was .05 (SE = .00, p < .001). The positive and significant coefficient on the slope indicates that aggression trajectories increased across elementary and middle school years. In terms of variance, the estimated within-child variance for loneliness, $σ_{ε}^{2}$ , is .12 (p < .001); the estimated between-child variance for loneliness, $σ_{0}^{2}$ , is .08 (p < .001). The estimated within-child variance for aggression is .15 (p < .001); the estimated between-child residual variance for aggression is .14 (p < .001). All of these variances were significant, indicating that loneliness and aggression in the first year and the changes of loneliness and aggression across 5 years varied significantly across the sample.

Parenting Behaviors and Child Development Across Elementary and Middle School Years

Conditional models were estimated to examine between-child associations between parenting behaviors and child functioning outcomes. The results of the conditional models for loneliness and aggression are displayed in Table 2.

Table 2.

Multilevel Models for Change Across Time for Loneliness and Aggression (Each n = 4,990).

	Loneliness		Aggression
	Coefficient	SE	Coefficient	SE
Fixed effects
Initial status, π_0i
Intercept	1.49***	.01	1.48***	.01
Maternal support	−.01	.02	−.05**	.02
Maternal involvement	−.06***	.01	−.03	.02
Maternal harsh control	.02	.02	.09***	.02
Child cohort	.06***	.01	.08***	.01
Child gender	−.00	.01	.08***	.01
Education	−.01	.00	.01**	.00
Family Income	−.00	.00	−.01	.00
Rate of change, π_1i
Intercept	−.05***	.00	.05***	.00
Maternal support	−.01	.01	−.01	.01
Maternal involvement	.01**	.01	.00	.01
Maternal harsh control	.00	.01	.01	.01
Child cohort	−.02***	.00	.00	.00
Child gender	−.01***	.00	−.01***	.00
Education	.00	.00	−.00**	.00
Family Income	−.00*	.00	.00	.00
Variance components
Level 1
Level 2
Within-person, $σ_{ε}^{2}$	.12***		.15***
In initial status, $σ_{0}^{2}$	.07***		.12***
In rate of change, $σ_{1}^{2}$	.01***		.01***
Correlation between ζ_0i and ζ_1i, τ	−.52		−.53

Note. Codes for cohorts are 1 = older cohort, −1 = younger cohort; codes for child gender are 1 = boys, −1 = girls.

p < .05. **p < .01. ***p < .001.

Loneliness Model

After controlling for child gender and cohort as well as mother education and family income, children with involved mothers reported less loneliness at the baseline (β = −.06, p < .001) and had slower rate of decreasing trajectories of loneliness (β = .01, p < .01) than other children. However, maternal support did not predict the initial status (intercept) and the rate of change (slope) of loneliness. Lastly, consistent with our hypothesis, harsh control did not predict the intercept and slope of loneliness. No gender differences in loneliness were found initially; however, loneliness decreased more rapidly for boys (β = −.01, p < .001) than for girls across years. The older cohort reported more loneliness (β = .06, p < .001) than the younger cohort at the baseline, the older cohort showed a greater declining slope (β = −.02, p < .001) than the younger cohort.

Aggression Model

After controlling for child gender and cohort as well as mother education and family income, children with supportive mothers reported less aggression at the baseline (β = −.05, p < .01) than other children, but no difference in the rate of change was found. However, maternal involvement did not predict the initial status (intercept) and the rate of change (slope) of aggression. Lastly, children whose mothers used more harsh control reported more aggression than other children at the baseline (β = .09, p < .001), but no difference in the rate of change was found. Boys reported more aggression than girls initially (β = .08, p < .001); however, aggression increased less rapidly for boys than for girls across time (β = −.01, p < .001). The older cohort reported more aggression than the younger cohort at the baseline (β = .08, p < .001), but no difference in the rate of change was found.

Discussion

Developmental Trajectories of Loneliness and Aggression

This study first examined the trajectories of children’s loneliness and aggression across their developmental transition years and then investigated the roles of three parenting behaviors as well as child cohort and gender in predicting the initial status and rate of change in loneliness and aggression in a Taiwanese sample. Consistent with most of the longitudinal studies conducted in Western societies (Harris et al., 2013; Ladd & Ettekal, 2013; Vanhalst et al., 2013), our results suggest a declining trajectory of loneliness over time. That is, children’s loneliness gradually decreases over time as they grow older. As for aggression, children’s aggressive behavior greatly conflicts with expectations in the traditional Chinese culture in Taiwan that emphasizes collectivism, filial piety, family-oriented values, harmony with others, and interdependence between self and others (Hsu, 1971). Aggression toward peers or others may threaten such harmony. Therefore, aggression is viewed as highly problematic in Taiwan and is actively suppressed (Sung, 1985). Although this result indicated the average aggression score for the average child in the Taiwanese sample was relatively low (1.48 on a 4-point rating scale), it unexpectedly suggested an increasing trajectory of aggression over time as children transition into early adolescence. Although aggression is actively suppressed in Taiwan, one possible reason for the increasing trajectory of aggression across developmental transitions may be the adolescent storm and stress (Hall, 1904) and puberty-related changes in mood and behavior (Buchanan, Eccles, & Becker, 1992).

Parenting in Relation to Children’s Loneliness and Aggression

The descriptive statistics show that maternal harsh control is normally distributed in our sample, whereas it is usually skewed in Western samples. Given the unique concept of guan or chiao shun in Chinese culture (Chao, 1994) that is used to discipline and educate children and promote success, harsh control is probably acceptable and widely used in Chinese culture.

The significant parenting predictor of child loneliness in this study was maternal involvement, but not support and harsh control. The more the Taiwanese mothers were involved with their children, the less lonely their children were. Mothers are often the facilitators of their children’s social interactions with peers (Ladd, Profilet, & Hart, 1992). When mothers spend more time with their children, they may have more opportunities to know about their children’s thoughts, feelings, and relationships with peers. In turn, mothers may be able to reduce their children’s loneliness by providing companionship or giving advice for social and communication skills. Unexpectedly, maternal support did not predict loneliness. To reduce the likelihood of child loneliness, being involved and being there for children seems to be a more effective parenting behavior than being supportive among Taiwanese families. Finally, consistent with the hypothesis, harsh control did not predict loneliness. When harsh control is practiced by Taiwanese mothers with the unique parenting concepts of guan or chiao shun (Chao, 1994), it tends to be considered as normative and thus less likely to lead to child maladjustment and problem behaviors (Chiu, 1987).

On the other hand, significant parenting predictors of child aggression in this study were maternal support and harsh control, but not involvement. The more the Taiwanese mothers were supportive, the less aggressive their children were. Previous studies that used Western samples have also found parental support to be linked to low aggression levels (Van der Graaff et al., 2012). Unexpectedly, maternal involvement did not predict child aggression. To reduce the likelihood of child aggression, being supportive seems to be a more effective parenting behavior than being involved among Taiwanese families. Finally, because parental harsh control was viewed as normative and acceptable in traditional Chinese culture and less likely to lead to child maladjustment (Chao, 1994; Chiu, 1987), we hypothesized that harsh control did not predict aggression among Taiwanese families. However, in contradiction of the hypothesis, the results indicated that harsh control predicted aggression at the baseline. One possible explanation is that when children encounter conflict between their desire for greater autonomy and the high levels of control/obedience that Taiwanese parents expect from their children, the children are more likely to act out against the strict norms set by their parents and then exhibit aggressive behavior. Another possible explanation for the association can be considered using social learning theory (Bandura, 1977). Among Taiwanese families, when parents use harsh control to gain children’s immediate compliance and obedience, children may learn the behaviors via observation and modeling in the family and become more likely to use coercive and aggressive behaviors as the most immediate and effective strategies when interacting with their peers in school settings to solve conflicts or problems. Moreover, Chinese mothers tend to respond to children’s aggressive behavior with anger, embarrassment, and disappointment (Cheah & Rubin, 2004). When mothers respond to children’s aggression with aggressive discipline strategies, such as harsh control, this response may actually lead to more child aggression. Thus, a negative cycle between harsh control and child aggression can occur.

Gender Differences in Aggression

Consistent with the hypothesis, the current study found that boys in Taiwan reported more overt aggression than girls. Boys tend to be more physically aggressive and exhibit more problem behaviors than girls (Card et al., 2008). Girls are more prosocial (Zahn-Waxler, 1993) and view interpersonal relationships as more important than is the case with boys. Therefore, in comparison with boys, girls tend to have lower aggression levels to protect group harmony and preserve good interpersonal relationships with their peers during childhood. However, as children grow older, our results revealed that girls’ aggression levels increased at a faster rate than boys’ aggression levels across the transition years to early adolescence. The possible explanation for this outcome is that girls usually begin puberty and enter adolescence earlier than boys, and may exhibit puberty-related changes in mood and behavior earlier than boys, and then lead to increased use of aggression (Susman et al., 2007). Another explanation is that in having lower initial aggression levels, the girls’ aggression levels tend to increase faster than those of the boys during the transition years to early adolescence.

Strengths and Limitations

The present study is one of the first to explore the relative roles of three parenting behaviors and child gender in predicting the baselines and trajectories of loneliness and aggression across developmental transitions in a Taiwanese sample. A strength of the study is that we examined the unique contributions of parenting behaviors and child characteristics using multilevel modeling with regard to both the initial status and the rate of change that characterize two problematic social functioning outcomes (loneliness and aggression) across elementary and middle school years. By using multilevel modeling to model change across time, we explored the intercept and slope in loneliness and aggression as well as both levels of between-child differences and within-child differences as they changed across time. Furthermore, having five assessments in five successive years for both the younger and the older cohort made it possible to model changes in child outcomes across two major developmental transitions, that is, from early childhood to early adolescence and from middle childhood to adolescence. Another strength of the study is that, because mothers reported their maternal parenting behaviors and children reported their child social functioning problems, the associations found among parenting behaviors and children’s social functioning problems were not due to shared method variance. Other strengths are the large sample size and the stratified random sampling, which provided sufficient power for the analyses and enhanced sample representation. Finally, studying Taiwanese families in East Asia provided us with an opportunity to explore whether Western theories and research that focus on Western parenting behaviors and child social functioning remain valid in a notably different cultural context, especially with the unique concepts of harsh control in Chinese culture.

This study also has several limitations. First, the developmental trajectories of overt aggression found in this study cannot be generalized to other types of aggression, such as relational aggression or covert aggression. Future studies could explore whether boys and girls follow different trajectories for different types of aggression. Second, this study focused on the role of maternal parenting behaviors in child functioning. Thus, the results should be generalized only for mothers, because the effects of paternal parenting behaviors on child adjustment may differ from the effects of maternal parenting behaviors (Chen, Liu, & Li, 2000). Future studies could include paternal parenting behaviors in the models in order to clarify the roles that fathers play in children’s loneliness and aggression in Chinese cultural contexts. Third, the current study relies exclusively on self-reported measures. Future studies could include multiple data sources, such as peer nominations or teacher ratings for loneliness and aggression. Fourth, low alphas (ranged from .64 to .69) were questionable for the measure of aggression. The possible explanation for low alphas might be due to there only being four items and different forms of aggression were included in the scale. George and Mallery (2003) suggested that alpha greater than .70 is acceptable, and alpha between .60 and .70 is questionable, but not poor or unacceptable. Fifth, the measures used in the current study were developed by the CABLE research team and have not been well replicated and standardized, although reliability and validity were tested by the CABLE team and by a few studies (Wu et al., 2007; Yen et al., 2002; Yen et al., 2006). Finally, the correlation table shows that loneliness positively correlates with aggression (from .05 to .20). Chen, Wang, Li, and Liu (2014) also found that aggression is positively associated with loneliness. Future research can further extend the current study by examining the interplay between loneliness and aggression under hierarchical linear modeling.

Conclusion and Implications

The results from the current study begin to fill gaps in our understanding of the associations between Chinese parenting and children’s internal and external forms of social functioning problems (loneliness and aggression, respectively) across time from early childhood to adolescence. First, the results revealed a decreasing trajectory of loneliness and an increasing trajectory of aggression across developmental transitions. Second, no matter how normative and positive harsh control was viewed in traditional Chinese culture, maternal harsh control still predicted child aggression, but did not predict child loneliness. Third, maternal involvement seems to be more effective in reducing the levels of loneliness, while maternal support seems to be more effective in reducing the levels of aggression among Taiwanese families. Fourth, while girls were less aggressive than boys at the baseline, girls displayed a faster rate of increase in aggression over time than boys.

The findings of the present study offer some implications for prevention and intervention efforts by educators, counselors, social workers, and other professionals who work with Chinese families. First, although Chinese parents are widely believed to be more restrict and controlling than American parents (Kelley & Tseng, 1992; Lin & Fu, 1990) and harsh control is viewed as normative and even positive in Chinese culture (Chiu, 1987), we found that harsh control predicts high levels of child aggression in a Taiwanese sample, which is similar to the findings in Western samples. Chinese mothers often regard harsh criticism and punishment as acceptable and efficient ways to discipline children and to help children improve and achieve goals, and these strategies are even viewed as expressions of care. Although these mothers’ intentions are probably good, their children’s outcomes appear to point in another direction. Children may not share the same values with their mothers toward the concept of harsh control, guan. Children may not perceive that harsh control is “for your own good” and “out of love,” and the outcomes of this harsh control often turn out to be negative for the children (Chang, Schwartz, Dodge, & McBride-Chang, 2003). In addition, our results suggested that maternal involvement seems to be more effective in reducing the likelihood of loneliness, while maternal support seems to be more effective in reducing the likelihood of aggression among Taiwanese families. Each specific parenting behavior may work differently in terms of the effectiveness for different child outcomes. Finally, the increasing trajectories of aggression, especially for girls, caught our attention and require prevention and intervention. Our results show a higher rate of increase in aggression for girls than for boys. Thus, the traditional assumption that girls are often less aggressive than boys may be erroneous and, in turn, suggests that more, not less, attention should be paid to girls than boys when trying to mitigate aggressive behavior. Finally, the present study help professionals and practitioners acknowledge unique cultural beliefs and use evidence-based practice models to integrate the research evidence and inform practice decisions may help to prepare and begin prevention and intervention efforts.

Footnotes

Acknowledgements

We would like to thank the Education Departments of Taipei City and Hsinchu County for administrative support, the 18 participating schools for providing venues and time for the surveys, the children and their parents who were involved in the survey for their support, and the interviewers and supervisors that helped carry out the survey.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study carried out an analysis of part of the 2001-2005 data from the National Health Research Institutes supported (2001: HP-090-SG-03; 2002: HP-091-SG-02; 2003: HP 092-PP-07; 2004: HP-093-PP-03) Children and Adolescent Behaviors in Long-Term Evolution (CABLE) Project.

Author Biographies

Yi-Ping Hsieh is an assistant professor and an interdisciplinary scholar at the Department of Social Work at University of North Dakota. She received her PhD in educational psychology (minor in counseling) from Indiana University, Bloomington. Her research interests include child maltreatment, parenting, victimization, problematic Internet use, and behavioral and mental health problems.

Lee-Lan Yen is an emeritus professor at Institute of Health Policy and Management at National Taiwan University. Her research interests include child and adolescent development, substance use, and health promotion.

References

Annear

K. D.

Yates

G. C. R.

(2010). Restrictive and supportive parenting: Effects on children’s school affect and emotional responses. The Australian Educational Researcher, 37, 63-82.

Asher

S. R.

Hymel

Renshaw

P. D.

(1984). Loneliness in children. Child Development, 55, 1456-1464.

Bandura

(1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall.

Baumrind

(1971). Current patterns of parental authority. Developmental Psychology, 4(1, Pt. 2), 1-103.

Blitstein

J. L.

Murray

D. M.

Lytle

L. A.

Birnbaum

A. S.

Perry

C. L.

(2005). Predictors of violent behavior in an early adolescent cohort: Similarities and differences across genders. Health Education & Behavior, 32, 175-194.

Bowlby

(1969). Attachment and loss: Attachment (Vol. 1). New York, NY: Basic Books.

Brame

Nagin

D. S.

Tremblay

R. E.

(2001). Developmental trajectories of physical aggression from school entry to late adolescence. The Journal of Child Psychology and Psychiatry, and Allied Disciplines, 42, 503-512.

Brennan

(1982). Loneliness at adolescence. In Peplau

L. A.

Perlman

(Eds.), Loneliness: A sourcebook of current theory, research and therapy (pp. 269-290). Hoboken, NJ: John Wiley.

Bronfenbrenner

(2001). The theory of human development. In Smelser

N. J.

Baltes

P. B.

(Eds.), International encyclopedia of the social and behavioral sciences (Vol. 10, pp. 6963-6970). New York, NY: Elsevier.

10.

Bronfenbrenner

Morris

P. A.

(1998). The ecology of developmental processes. In Denton

(Series Ed.) Lerner

R. M.

(Vol. Ed.), Handbook of child psychology: Theory (Vol. 1, pp. 993-1028). New York, NY: John Wiley.

11.

Buchanan

C. M.

Eccles

J. S.

Becker

J. B.

(1992). Are adolescents the victims of raging hormones? Evidence for activational effects of hormones on moods and behavior at adolescence. Psychological Bulletin, 111, 62-107.

12.

Call

K. T.

Mortimer

J. T.

(2001). Arenas of comfort in adolescence: A study of adjustment in context. Mahwah, NJ: Lawrence Erlbaum.

13.

Campbell

S. B.

Spieker

Burchinal

Poe

M. D.

(2006). Trajectories of aggression from toddlerhood to age 9 predict academic and social functioning through age 12. The Journal of Child Psychology and Psychiatry, 47, 791-800.

14.

Card

N. A.

Stucky

B. D.

Sawalani

G. M.

Little

T. D.

(2008). Direct and indirect aggression during childhood and adolescence: A meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Development, 79, 1185-1229.

15.

Cassidy

(2000). Adult romantic attachments: A developmental perspective on individual differences. Review of General Psychology, 4, 111-131.

16.

Centers for Disease Control and Prevention. (2001). Youth risk behavior survey. Available at: www.cdc.gov/YRBSS.

17.

Chang

Schwartz

Dodge

K. A.

McBride-Chang

(2003). Harsh parenting in relation to child emotion regulation and aggression. Journal of Family Psychology, 17, 598-606.

18.

Chao

R. K.

(1994). Beyond parental control and authoritarian parenting style: Understanding Chinese parenting through the cultural notion of training. Child Development, 65, 1111-1120.

19.

Chao

R. K.

(2001). Extending research on the consequences of parenting style for Chinese Americans and European Americans. Child Development, 72, 1832-1843.

20.

Cheah

C. S. L.

Rubin

K. H.

(2004). Comparison of European American and Mainland Chinese mothers’ responses to aggression and social withdrawal in preschoolers. International Journal of Behavioral Development, 28, 83-94.

21.

Chen

Liu

M. W.

(2000). Parental warmth, control, and indulgence and their relations to adjustment in Chinese children: A longitudinal study. Journal of Family Psychology, 14, 401-419.

22.

Chen

Wang

Liu

(2014). Loneliness in Chinese children across contexts. Developmental Psychology, 50, 2324-2333.

23.

Chipuer

H. M.

(2001). Dyadic attachments and community connectedness: Links with youths’ loneliness experiences. Journal of Community Psychology, 29, 429-446.

24.

Chiu

L. H.

(1987). Child-rearing attitudes of Chinese, Chinese-American, and Anglo-American mothers. International Journal of Psychology, 22, 409-419.

25.

Crick

N. R.

Bigbee

M. A.

Howes

(1996). Gender differences in children’s normative beliefs about aggression: How do I hurt thee? Let me count the ways. Child Development, 67, 1003-1014.

26.

Deniz

Hamarta

Ari

(2005). An investigation of social skills and loneliness levels of university students with respect to their attachment styles in a sample of Turkish students. Social Behavior and Personality: An International Journal, 33(1), 19-32.

27.

Dodge

K. A.

Pettit

G. S.

Bates

J. E.

(1994). Socialization mediators of the relation between socioeconomic-status and child conduct problems. Child Development, 65, 649-665.

28.

Dodge

K. A.

Pettit

G. S.

Bates

J. E.

Valente

(1995). Social information-processing patterns partially mediate the effect of early physical abuse on later conduct problems. Journal of Abnormal Psychology, 104, 632-643.

29.

George

Mallery

(2003). SPSS for Windows step by step: A simple guide and reference. 11.0 update (4th ed.). Boston, MA: Allyn & Bacon.

30.

Gravetter

F. J.

Wallnau

L. B.

(2014). Essentials of statistics for the behavioral sciences. Belmont, CA: Cengage Learning.

31.

Hall

G. S.

(1904). Adolescence: Its psychology and its relation to physiology, anthropology, sociology, sex, crime, religion, and education. Englewood Cliffs, NJ: Prentice Hall.

32.

Harris

R. A.

Qualter

Robinson

S. J.

(2013). Loneliness trajectories from middle childhood to pre-adolescence: Impact on perceived health and sleep disturbance. Journal of Adolescence, 36, 1295-1304.

33.

Hsu

F. L. K.

(1971). Psychosocial homeostasis and Jen: Conceptual tools for advancing psychological anthropology. American Anthropologist, 73(1), 23-44.

34.

Hyde

J. S.

(1984). How large are gender differences in aggression? A developmental meta-analysis. Developmental Psychology, 20, 722-736.

35.

Jones

A. C.

Schinka

K. C.

van Dulmen

M. H. M.

(2011). Changes in loneliness during middle childhood predict risk for adolescent suicidality indirectly through mental health problems. Journal of Clinical Child & Adolescent Psychology, 40, 818-824.

36.

Joussemet

Vitaro

Barker

E. D.

Côté

Nagin

D. S.

Zoccolillo

. . . Tremblay

R. E.

(2008). Controlling parenting and physical aggression during elementary school. Child Development, 79, 411-425.

37.

Karriker-Jaffe

K. J.

Foshee

V. A.

Ennett

S. T.

Suchindran

(2008). The development of aggression during adolescence: Sex differences in trajectories of physical and social aggression among youth in rural areas. Journal of Abnormal Child Psychology, 26, 1227-1236.

38.

Kelley

M. L.

Tseng

H.-M.

(1992). Cultural differences in child rearing a comparison of immigrant Chinese and Caucasian American mothers. Journal of Cross-Cultural Psychology, 23, 444-455.

39.

Kinard

E. M.

(1999). Perceived social skills and social competence in maltreated children. American Journal of Orthopsychiatry, 69, 465-481.

40.

Kline

R. B.

(2016). Principles and practice of structural equation modeling. New York, NY: Guilford Press.

41.

Koenig

L. J.

Abrams

R. R.

(1999). Adolescent loneliness and adjustment: A focus on gender differences. In Rotenberg

K. J.

Hymel

(Eds.), Loneliness in childhood and adolescence (pp. 296-322). Cambridge, UK: Cambridge University Press.

42.

Kuppens

Grietens

Onghena

Michiels

(2009). Associations between parental control and children’s overt and relational aggression. British Journal of Developmental Psychology, 27, 607-623.

43.

Lacourse

Cote

Nagin

D. S.

Vitaro

Brendgen

Tremblay

R. E.

(2002). A longitudinal-experimental approach to testing theories of antisocial behavior development. Development and Psychopathology, 14, 909-924.

44.

Ladd

Ettekal

(2013). Peer-related loneliness across early to late adolescence: Normative trends, intra-individual trajectories, and links with depressive symptoms. Journal of Adolescence, 36, 1269-1282.

45.

Ladd

G. W.

Profilet

S. M.

Hart

C. H.

(1992). Parents’ management of children’s peer relations: Facilitating and supervising children’s activities in the peer culture. In Parke

R. S.

Ladd

G. W.

(Eds.), Family-peer relationships: Model of linkage (pp. 215-253). Hillsdale, NJ: Lawrence Erlbaum.

46.

Lamborn

S. D.

Mounts

N. S.

Steinberg

Dornbusch

S. M.

(1991). Patterns of competence and adjustment among adolescents from authoritative, authoritarian, indulgent, and neglectful families. Child Development, 62, 1049-1065.

47.

Leung

Fan

R. M.

(1996). Adolescent delinquent behavior in Chinese societies. In Lau

(Ed.), Growing up the Chinese way: Chinese child and adolescent development (pp. 237-264). Hong Kong: Chinese University Press.

48.

Costanzo

P. R.

Putallaz

(2010). Maternal socialization goals, parenting styles, and social-emotional adjustment among Chinese and European American young adults: Testing a mediation model. The Journal of Genetic Psychology, 171, 330-362.

49.

Lin

C. Y. C.

V. R.

(1990). A comparison of child-rearing practices among Chinese, immigrant Chinese, and Caucasian-American parents. Child Development, 61, 429-433.

50.

Maccoby

E. E.

Jacklin

C. N.

(1980). Sex differences in aggression: A rejoinder and reprise. Child Development, 51, 964-980.

51.

Maccoby

E. E.

Martin

J. A.

(1983). Socialization in the context of the family: Parent–child interaction. In Mussen

P. H.

Hetherington

E. M.

(Eds.), Handbook of child psychology: Socialization, personality, and social development (4th ed., Vol. 4, pp. 1-101). New York, NY: John Wiley.

52.

McWhirter

B. T.

Besett-Alesch

T. M.

Horibata

Gat

(2002). Loneliness in high risk adolescents: The role of coping, self-esteem, and empathy. Journal of Youth Studies, 5, 69-84.

53.

Minzi

M. C. R.

(2006). Loneliness and depression in middle and late childhood: The relationship to attachment and parental styles. The Journal of Genetic Psychology, 167, 189-210.

54.

Nagin

Tremblay

R. E.

(1999). Trajectories of boys’ physical aggression, opposition, and hyperactivity on the path to physically violent and nonviolent juvenile delinquency. Child Development, 70, 1181-1196.

55.

Nelson

D. A.

Hart

C. H.

Yang

C. M.

Olsen

J. A.

Jin

S. H.

(2006). Aversive parenting in China: Associations with child physical and relational aggression. Child Development, 77, 554-572.

56.

Nelson

L. J.

Hart

C. H.

Yang

Roper

S. O.

Jin

(2006). Relations between Chinese mothers’ parenting practices and social withdrawal in early childhood. International Journal of Behavioral Development, 30, 261-271.

57.

Parke

R. D.

MacDonald

K. B.

Beitel

Bhavnagri

(1988). The role of the family in the development of peer relationships. In Peters

R. D.

McMahon

R. J.

(Eds.), Social learning and systems approaches to marriage and the family (pp. 17-44). Philadelphia, PA: Brunner.

58.

Parkhurst

J. T.

Hopmeyer

(1999). Developmental change in the sources of loneliness: Contributions of attachment theory. In Rotenberg

K. J.

Hymel

(Eds.), Loneliness in childhood and adolescence (pp. 56-79). New York, NY: Cambridge University Press.

59.

Patterson

G. R.

(1982). Coercive family process: A social learning approach (Vol. 3). Eugene, OR: Castalia.

60.

Patterson

G. R.

(1995). Coercion as a basis for early age of onset for arrest. In McCord

(Ed.), Coercion and punishment in long-term perspectives (pp. 81-105). New York, NY: Cambridge University Press.

61.

Peplau

L. A.

Perlman

(1982). Perspectives on loneliness. In Peplau

L. A.

Perlman

(Eds.), Loneliness: A sourcebook of current theory, research and therapy (pp. 1-20). Hoboken, NJ: John Wiley.

62.

Pepler

D. J.

Craig

W. M.

(2005). Aggressive girls on troubled trajectories: A developmental perspective. In Pepler

D. J.

Madsen

K. C.

Webster

Levene

K. S.

(Eds.), The development and treatment of girlhood aggression (pp. 3-28). Mahwah, NJ: Lawrence Erlbaum.

63.

Qualter

Vanhalst

Harris

Van Roekel

Lodder

Bangee

. . .Verhagen

(2015). Loneliness across the life span. Perspectives on Psychological Science, 10, 250-264.

64.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models: Applications and data analysis methods. In DeLeeuw

Berk

R. A.

(Eds.), Advanced quantitative techniques in the social sciences (Vol. 1). Thousand Oaks, CA: Sage.

65.

Rich

A. R.

Kirkpatrick-Smith

Bonner

R. L.

Jans

(1992). Gender differences in the psychosocial correlates of suicidal ideation among adolescents. Suicide and Life-Threatening Behavior, 22, 364-373.

66.

Schoffstall

C. L.

Cohen

(2011). Cyber aggression: The relation between online offenders and offline social competence. Social Development, 20, 587-604.

67.

Simons

R. L.

Johnson

Conger

R. D.

(1994). Harsh corporal punishment versus quality of parental involvement as an explanation of adolescent maladjustment. Journal of Marriage and Family, 56, 591-607.

68.

Singer

J. D.

Willett

J. B.

(2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York, NY: Oxford University Press.

69.

Storch

E. A.

Masia-Warner

(2004). The Relationship of peer victimization to social anxiety and loneliness in adolescent females. Journal of Adolescence, 27, 351-362.

70.

Sung

B. L.

(1985). Bicultural conflicts in Chinese immigrant children. Journal of Comparative Family Studies, 16, 255-269.

71.

Susman

E. J.

Dockray

Schiefelbein

V. L.

Herwehe

Heaton

J. A.

Dorn

L. D.

(2007). Morningness/eveningness, morning-to-afternoon cortisol ratio, and antisocial behavior problems during puberty. Developmental Psychology, 43, 811-822.

72.

Triandis

H. C.

(1995). New directions in social psychology: Individualism and collectivism. Boulder, CO: Westview Press.

73.

Van der Graaff

Branje

De Wied

Meeus

. (2012). The moderating role of empathy in the association between parental support and adolescent aggressive and delinquent behavior. Aggressive Behavior, 38, 368-377.

74.

van Roekel

Goossens

Scholte

R. H.

Engels

R. C.

Verhagen

. (2011). The dopamine D2 receptor gene, perceived parental support, and adolescent loneliness: Longitudinal evidence for gene–environment interactions. The Journal of Child Psychology and Psychiatry, 52, 1044-1051.

75.

Vanhalst

Goossens

Luyckx

Scholte

R. H. J.

Engels

R. C. M. E

. (2013). The development of loneliness from mid- to late adolescence: Trajectory classes, personality traits, and psychosocial functioning. Journal of Adolescence, 36, 1305-1312.

76.

Weiss

R. S.

(1973). Loneliness: The experience of emotional and social isolation. Cambridge, MA: MIT Press.

77.

Weiss

R. S.

(1987). Reflections on the present state of loneliness research. Journal of Social Behavior and Personality, 2(2), 1-16.

78.

Wheeler

Reis

Nezlek

J. B.

(1983). Loneliness, social interaction, and sex roles. Journal of Personality and Social Psychology, 45, 943-953.

79.

W. C.

Kao

C. H.

Yen

L. L.

Lee

T. S. H.

(2007). Comparison of children’s self-reports of depressive symptoms among different family interaction types in northern Taiwan. BMC Public Health, 7, Article 116.

80.

Yen

L. L.

Chen

Lee

S. H.

Hsiao

Pan

L. Y.

(2002). Child and adolescent behaviour in long-term evolution (CABLE): A school-based health lifestyle study. Promotion & Education, 1(Suppl. 1), 33-40.

81.

Yen

L. L.

Chiu

C. J.

W. C.

Pan

L. Y.

(2006). Aggregation of health behaviors among fourth graders in northern Taiwan. Journal of Adolescent Health, 39, 435-442.

82.

Zahn-Waxler

(1993). Warriors and worriers: Gender and psychopathology. Development and Psychopathology, 5, 79-89.