The Effect of Family Cohesion and Life Satisfaction During Adolescence on Later Adolescent Outcomes

Abstract

Family factors and adolescent life satisfaction have profound effects on later behavior. The present study examined the effects of family cohesion and life satisfaction on internalizing (depression and self-esteem) and externalizing (deviance) outcomes among adolescents. Moreover, this study also viewed internalizing outcomes as conduits that channeled the effects of family cohesion and life satisfaction on later adolescent deviance. Using the Taiwan Youth Project (TYP), results indicated that the family cohesion and life satisfaction of Taiwanese adolescents declined with time, and this, in turn, led to negative life outcomes, such as higher levels of depression and deviance, and lower levels of self-esteem. In addition to the direct relationship between family cohesion and adolescent outcomes, family cohesion also had indirect effects on self-esteem and depression through change in life satisfaction. Also, initial life satisfaction is indirectly related to later deviance through change in family cohesion. Limitations and implications are also discussed.

Keywords

families deviance mental health

Introduction

Adolescence is an important juncture for each individual, because a well-adjusted adolescent (e.g., one with high self-esteem and low depression) has positive life outcomes (e.g., better educational attainment and career) later in life. However, such adjustment is not easy to achieve, as research has documented that adolescence is replete with physical (e.g., puberty), psychological (e.g., unstable emotions), and behavioral changes (e.g., increased risk taking) (Cicognani, 2011; Levitt et al., 2005; Moffitt, 1993; Newman, Newman, Griffen, O’Connor, & Spas, 2007). Therefore, it is important to investigate the factors that can protect adolescents from the possible negative impacts of these changes.

Although previous studies have demonstrated that youths who reside in a more organized neighborhood are less likely to commit delinquent acts (Sampson, 2006) and feel safer about the community (Nebbitt, Williams, Lombe, McCoy, & Stephens, 2014), contextual factors that are more proximal to the individual often exert greater effects on adolescent development than factors at a distance. Consequently, two relatively important factors are relevant: family and individual perception. With regard to family, previous studies have shown that a supportive and integrated family environment (e.g., a cohesive family) influences adolescent well-being and behavior, such that a cohesive family (e.g., support and connectedness) leads to positive outcomes (e.g., low depression) (Ekas, Lickenbrock, & Whitman, 2010; Rayle & Chung, 2007). Besides influences from teens’ immediate environment (e.g., family), their self-perception of life also exerts effects on their own behavior and well-being. Studies have demonstrated that youths who show satisfaction with their lives (e.g., general satisfaction with life or satisfaction with school) have better functioning (e.g., low levels of problem behavior or school engagement) than youths who do not (Lewis, Huebner, Malone, & Valois, 2011; Proctor, Linley, & Maltby, 2009; Sun & Shek, 2012). Hence, exploration of the etiology of negative outcomes in adolescence requires consideration of both individuals’ own life experience (e.g., life satisfaction) and influences from other important institutions in the current life stage (e.g., family).

The present study used a parallel latent growth curve model to see how family environment (e.g., family cohesion) and self-perception (e.g., life satisfaction) change and interact over time during early adolescence. It then examined the immediate effects of family cohesion and youths’ own life satisfaction during early adolescence on both externalizing (e.g., deviance) and internalizing (e.g., depression and self-esteem) outcomes in a Taiwanese sample.

Family Cohesion and Adolescent Functioning

Adolescence is a stressful period of development because it is not only a transitional period that links childhood and adulthood but also a period involving various changes (Goldbeck, Schmitz, Besier, Herschbach, & Henrich, 2007; Jaggers et al., 2015; Moksnes & Espnes, 2013). One particular factor that can help adolescents to adjust well is family. This notion has received support both theoretically and empirically. Theoretically, ecological system theory indicates that one of the most influential and immediate ecological contexts during adolescence is family (Bronfenbrenner, 1992). Empirically, various studies have demonstrated that family is related to both externalizing (Nash, McQueen, & Bray, 2005; Reeb et al., 2015) and internalizing problems (Newman et al., 2007), as well as positive developments (e.g., prosocial behavior) (Fosco, Caruthers, & Dishion, 2012). One particular family factor that may be important to adolescent development is family cohesion, which can be defined as the level of support, acceptance, and connectedness in the family (Olson, Sprenkle, & Russell, 1979). Indeed, family cohesiveness has been shown to be related to both negative emotions (Shamai & Kimhi, 2007; Sturge-Apple, Davies, & Cummings, 2010; White, Shelton, & Elgar, 2014) and deviance (Baumann, Spitz, Predine, Choquet, & Chau, 2007; Hamama & Arazi, 2012; Reeb et al., 2015). For example, adolescents from a noncohesive family are more likely to have earlier sexual debut (Kapungu, Holmbeck, & Paikoff, 2006) and more likely to be involved in alcohol use and related problems (Reeb et al., 2015).

Beside the evident relationship between cohesiveness and adolescent adjustments in the West, similar results have also been found in the East. For example, one study in Hong Kong (Lai & McBride-Chang, 2001) found that in addition to parenting style, a family climate that is high in conflicts and low in warmth is related to suicide ideation. Some other studies also found that a noncohesive family is detrimental to adolescents in Taiwan (Hou, 2001; Jou, 2015; Kung & Lee, 2015; M.-Y. Wu & Jou, 2009). Specifically, M.-Y. Wu and Jou (2009) found that family cohesion is positively related to adolescents’ moral beliefs, such that adolescents from more cohesive families have higher levels of both public and private moral beliefs. Jou (2015), in addition, found that Taiwanese high school students who are from more cohesive families at the last year of junior high school (mean age = 15 years) experience lower levels of depression and higher levels of self-esteem than others. Other found that family cohesion or support is related to low delinquency (Hou, 2001), better school performance (Kung & Lee, 2015), and better psychological well-being, such as low depression (Chang, Yi, & Lin, 2008).

While studies from both Western and Eastern societies have shown that a cohesive family is beneficial to adolescent development, further study may be warranted, because most previous studies focused on how family cohesion at one point in time influences adolescents concurrently and/or longitudinally (Jou, 2015; Kung & Lee, 2015; Shamai & Kimhi, 2007; Sturge-Apple et al., 2010). Hence, these studies implicitly assumed that family cohesion is static and would not change over time. This implicit assumption may be troublesome, as there are two reasons why the family relationship might decline during adolescence. First, youths in early adolescence experience a transition to middle school, which brings various challenges (e.g., social changes, and particularly academic demands) (Doge & Sherrill, 2006; Lee & Larson, 2000; Yi & Wu, 2004). This is especially true in Taiwan as well as in other East Asian countries (e.g., Korea and China) because of various national examinations. For example, Lin (2011) described adolescents as typically having long school hours that include regular school hours (10 hours) and cram school hours (3 hours), and others described students as highly stressed and strained (Yi & Wu, 2004). Second, early adolescence also brings about status changes, that is, transition from childhood to adolescence (Shoshani & Slone, 2013). As Erikson (1994) pointed out, adolescence is a time to build identity and seek autonomy; hence, conflicts between youths and their parents might be high. All these may compromise the cohesiveness of the family.

Consistent with this notion, one study (Baer, 2002) examined the family changes during adolescence by using a cohort sequential growth curve model, and revealed that family cohesion indeed generally declined over this time period. However, it only focused on modeling change in family cohesion over time. In contrast, Jaggers et al. (2015), using parental warmth as a proxy, found no significant changes in parental warmth during early to middle adolescence (age 13-16 years) among low-income blacks in the American South. Overall, the above studies showed mixed results and reveal no further information on the relationship between the trajectory of family cohesion and adolescent outcomes. Furthermore, given that Taiwan and many Asian countries are influenced by Chinese culture, which focuses on family and interpersonal relationships (Yang & Lu, 2005), any change or decline in family cohesion may impinge differently upon adolescents than is the case in Western countries.

Life Satisfaction and Adolescent Functioning

Day and Jankey (1996) argued that subjective appraisal and evaluation of one’s own experience are most crucial and reported that subjective well-being (SWB) reflects an individual’s life experience and is meaningful (Abbott & Wallace, 2012). Hence, studying SWB is important in understanding one’s psychological well-being and overall mental health (Moksnes & Espnes, 2013), and it has been identified as both a cause of and a buffer against developing problem behavior and negative health consequences (Coker et al., 2000; MacDonald, Piquero, Valois, & Zullig, 2005; Suldo & Huebner, 2004). Whereas SWB can be conceptualized as comprising affectional (e.g., positive and negative emotion) and cognitive (e.g., life satisfaction) components, many previous studies used life satisfaction as a measure of SWB (Chan & Lee, 2006; Gray, Chamratrithirong, Pattaravanich, & Prasartkul, 2013).

Life satisfaction, which can be defined as an individual’s cognitive and self-appraisal of his or her life quality or life satisfaction based on subjective standards (Pavot & Diener, 1993), has been the subject of some research on adolescent populations. Research has demonstrated that life satisfaction is related to adolescent functioning, such that a low level of life satisfaction is related to various adjustment problems, such as substance use and other problem behaviors (MacDonald et al., 2005; Sun & Shek, 2012), depression (Gilman & Huebner, 2006), and low self-efficacy (Bradley & Corwyn, 2004). In addition, life satisfaction has been found to be related to positive outcomes, such as supportive behavior (Martin, Huebner, & Valois, 2008). One review showed life satisfaction to be “a determinant of many life outcomes” (Proctor et al., 2009, p. 605), such as positive self (e.g., self-esteem) and low levels of negative emotions (e.g., depression).

Only limited research on these topics has been conducted in Taiwan. For example, three studies that focused on adolescents and children showed that a higher level of well-being or life satisfaction is related to lower depression and better school performance (Hsien, 2003; Su, 2003; Wang, 2006, ). While these studies are limited in number, the general results echo those in the West.

Although a few studies have established the relationship between life satisfaction and its effects on other outcomes, change in life satisfaction during adolescence has not been studied much. Many studies have only considered age as a control (Leung, McBride-Chang, & Lai, 2004; Videon, 2002), and some reviews have shown that demographic variables, including age, are only weakly related to life satisfaction in adolescence (Huebner, 2004; Proctor et al., 2009). In contrast, some studies have examined change in life satisfaction during adolescence directly (Goldbeck et al., 2007; Park, 2005; Salmela-Aro & Tuominen-Soini, 2010). Goldbeck et al. (2007) and Park (2005) found that life satisfaction decreased during adolescence, while others found it to be increasing (Salmela-Aro & Tuominen-Soini, 2010). This mixed result may be due to different participants. The youths of Goldbeck et al. (2007) were in early to middle adolescence (age 11-16 years), while youths in the latter study (Salmela-Aro & Tuominen-Soini, 2010) were in middle to late adolescence (age 15-18 years). Furthermore, Goldbeck et al. (2007) used ANOVA to see the changes, whereas Salmela-Aro and Tuominen-Soini (2010) used growth curve modeling. In contrast, the Park (2005) study only compared the mean life satisfaction across children (elementary school students) and adolescents (middle and high school students) in South Korea.

Most previous studies used statistical models that examined inter-individual changes (e.g., Time 1 predicting Time 2), which ignores the within-individual change and a more dynamic approach. Those that used a dynamic approach provided mixed results (Goldbeck et al., 2007; Salmela-Aro & Tuominen-Soini, 2010). Given the inconsistent results and paucity of studies from the East, further examination of the change on a within-individual basis and with a non-Western sample is warranted. We suspect that life satisfaction will decline during adolescence in Taiwan and many other Asian countries because of the school stress mentioned earlier. Furthermore, one previous study pointed out the importance of understanding the relationship between family and life satisfaction (Suldo & Huebner, 2004). Hence, examining how changes in life satisfaction interact with family cohesion over time is important as well.

The Present Study

The above review has pointed out that family cohesion and adolescents’ life satisfaction both play important roles in influencing youths’ adjustment. The present study improves upon three limitations of previous studies and provides further understanding on adolescent development. First, only a handful of studies investigated changes in family and/or life satisfaction during adolescence and their relationship with adolescent development (Goldbeck et al., 2007; Jaggers et al., 2015; Park, 2005; Salmela-Aro & Tuominen-Soini, 2010). However, these studies investigated only how one of these two attributes changed over time and its relationship with particular adolescent outcomes, and results regarding life satisfaction are mixed. Hence, this study intends to examine how these two attributes changed across early adolescence in a non-Western sample, and the interrelationship between family cohesion and life satisfaction during early adolescence.

Second, although previous studies used both cross-sectional (Hamama & Arazi, 2012; Sun & Shek, 2012) and longitudinal data (Reeb et al., 2015; White et al., 2014) to understand the relationship between life satisfaction, family, and adolescent functioning, these studies provided only a static approach to the relationships; in other words, they looked at how life satisfaction and family cohesion measured at some time point exerted effects on adolescent functioning at some other time. To further our understanding, this study examines how changes in life satisfaction and family cohesion are related to adolescent functioning, particularly internalizing (e.g., self-esteem and depression) and externalizing problems (e.g., deviance). Various studies have provided evidence on the relationship between dysfunction during adolescence (e.g., low self-esteem and high depression) and future developmental outcomes (e.g., substance use and low SWB) (Boden, Fergusson, & Horwood, 2008; ; Orth, Robins, & Roberts, 2008; Siennick, 2007; Swislo & Orth, 2013). For example, studies have shown that low self-esteem during adolescence is related to negative future outcomes (e.g., low job success and educational attainment; Kammeyer-Mueller, Judge & Piccolo, 2008; Whitesell, Mitchell, & Spicer, 2009), that depression is related to delinquency and other negative outcomes (e.g., low income, low social support, and poorer self-rated health; Fletcher, 2013; Naicker, Galambos, Zeng, Senthilselvan, & Colman, 2013; Ritakallio et al., 2008), and that there is a high correlation between juvenile delinquency and adult negative outcomes (e.g., unemployment, an unstable career, and a criminal record; Lober & Farrington, 2012; Wiesner, Kim, & Capaldi, 2010).

Finally, prior research has often relied on Western samples, although some studies have investigated the issue in non-Western samples (Hamama & Arazi, 2012; Őnder & Yilmaz, 2012; Sun & Shek, 2012). However, these studies provided only a glimpse at the influence of family and life satisfaction on adolescent development. More research is clearly needed. This study used a panel data set from Taiwan to examine the interrelationship between family cohesion and life satisfaction during early adolescence, and their influence on adolescent adjustment in middle adolescence.

Method

Data and Sample

Data for the present study were drawn from the Taiwan Youth Project (TYP), conducted by the Institute of Sociology, Academia Sinica, Taiwan. The TYP was an 8-year longitudinal research project that began in 2000 and included two student cohorts, 2,690 first-year junior high school students (J1) (mean age = 13.3 years) and 2,890 third-year junior high school students (J3). It focused on how the social environment, especially, family and school, affects the life trajectory of adolescents.

The TYP used a two-stage stratified and clustered sampling design. In the first stage, the TYP team used two steps. The first step was to select the first-level strata based on geographic proximity to the research institute and to each other. Furthermore, the research team also considered various differences, such as occupational composition (e.g., blue- and white-collar composition) and economic development (e.g., high, middle to high, and low). The team chose three counties (Taipei and Yi-Lan Counties) and one city (Taipei) to mimic the population composition of Taiwan. The second step was to determine the first strata for further sampling. Consequently, the research team divided the city of Taipei and Taipei County into three strata each, and divided Yi-Lan County into two strata based on different levels of urbanization. After the strata were determined, the cluster sampling method was used. To select representative student samples, the project followed three principles. First, the number of students supplied by each second-level stratum was based on the proportion of students registered in each stratum relative to the entire student body in that county or city. Second, the research team divided the number of students by the average size of the stratum to determine the number of classes to be selected. Finally, on the basis of two classes for each school, the team determined the number of schools to be chosen. As a result of these calculations and determinations, 40 schools were selected: 16 from Taipei City, 15 from Taipei County, and 9 from Yi-Lan County. After the random selection of these schools and two classes for each selected school, all students in each selected class were recruited for the project.

The present study used Wave 1 (age range from 13 to 17 years) to Wave 4 of the J1 cohort, because items on the survey instrument during these time periods included ones suitable for the present study. In every wave, students received an in-class self-administered survey, and then these students were followed annually. All information about the variables used in the subsequent analyses emerged from the students’ self-report surveys based on virtually all students who were in the class on the day of the survey (n = 2,690) across all the research sites (the refusal rate was less than 1%).

Variables

Family cohesion (Wave 1 to Wave 3)

Family cohesion was measured by six items. These items were adapted from the cohesion subscale of the Family Adaptation and Cohesion Evaluation Scale III (FACES III; Olson, Portner, & Lavee, 1985) and captured the central aspects of cohesion, such as emotional bonding, support, spending time together, and sharing recreational activities. These items were very similar to those in previous studies (Reeb et al., 2015) and also captured important components of the Family Environment Scale (FES; Moos & Moos, 1994), such as coherence, expressiveness, and support. Hence, while we did not use the full scale from FACES III because of data limitations, we believe these items not only captured the essential meaning of family cohesion but also served well for the present study’s purpose. Moreover, the TYP research team modified these items to fit the Chinese culture in Taiwan, and some previous studies in Taiwan found these modified items to be useful when studying the effects of family function on adolescent adjustment (Jou, 2015; Kung & Lee, 2015; M.-Y. Wu & Jou, 2009). These six items include, for instance, “When making decisions, family members will discuss them with each other,” “When I need advice or help, I can count on my family,” and “Every family member will participate in activities together.” Each of the items was anchored on a 4-point scale (1 = it describes my family to 4 = it does not describe my family).

We first used explanatory factor analysis (EFA; principle axis factoring) to check whether these items measure this concept. The results reveal a one-factor solution for each of the first three waves, and factor loadings were all greater than .56. In addition, the reliability (α) for each measure was high (α = .82 for Wave 1; α = .84 for Wave 2; α = .83 for Wave 3). High reliability was also found in some studies that used this same scale in Taiwan (Jou, 2015; Kung & Lee, 2015; M.-Y. Wu & Jou, 2009). In the present study, all items were reverse coded, such that a higher score indicates a cohesive family. The means of these six items were used for the subsequent analysis.

Life satisfaction (Wave 1 to Wave 3)

Life satisfaction is an essential component of SWB (Proctor et al., 2009). The present measure captures youngsters’ satisfaction with their relationship with three important categories of others in their life: parents (mother and father), friends, and teachers. These individuals are representative of three major life domains for an adolescent: family, peers, and school. Besides relationships with important others, two other items asked students whether they were satisfied with their family economic situation and school performance. While the TYP did not use a published life satisfaction scale, it did have measures of adolescents’ current satisfaction with various important domains (e.g., family and school) in different perspectives (e.g., relationship and socioeconomic status). Furthermore, these items were developed by the research team to fit the local culture. Hence, the measure, though not perfect, provides a measure of life satisfaction for students in Taiwan (e.g., face validity). Furthermore, some validity and reliability were also demonstrated. For example, the confirmatory factor analyses of this measure provided acceptable fit (i.e., comparative fit index [CFI] is between .97 and .99; Tucker–Lewis index [TLI] is between .95 and .98; and root mean square error of approximation [RMSEA] is between .041 and .051), and the correlations between this measure and other known correlated are all significant (i.e., the correlation between satisfaction and happiness is from .24 to .42; the correlation between satisfaction and depressed mood is from .19 to .31). With this additional information, some validity (e.g., construct and criterion validity) is demonstrated. The reliability of this measure is also shown (α = .76 for Wave 1; α = .74 for Wave 2; α = .68 for Wave 3). Consequently, while further examination of the validity and reliability of this measure may be needed, at present, the measure should serve as a useful tool to capture the proposed attribute. At Wave 1 to Wave 3, students reported their satisfaction level regarding these six items. The response categories ranged from very unsatisfied (1) to very satisfied (4). Similar to family cohesion, these six items were reverse coded so that a high score indicated a higher level of life satisfaction, and the mean across these six items was used.

Depression (Wave 3)

For depression, 16 items from the Symptom Checklist-90-Revised (SCL-90-R; Derogatis, 1983) version were administered at Wave 3. This scale consists of 16 items asking students to report whether they have experienced depressive symptoms, and how serious the experience was in the past week (no; yes, but not serious; yes, a little bit serious; yes, serious; yes, very serious). The symptoms include both physical symptoms (e.g., headaches and numbness in some parts of the body) and emotional symptoms (e.g., depressed mood and loneliness). Summation of these items created a depressive scale, with a higher score indicating a higher level of depression (α = .88). This same scale or similar ones were used in several previous studies that examined adolescent development in Taiwan and found the scale or its equivalent (e.g., short or long form) were useful and possessed reliability (C.-I. Wu & Huang, 2010; Yi, Fan, & Chang, 2013; Yu, Chen, & Chang, 2008). Hence, cross-cultural use created no serious problems.

Self-esteem (Wave 3)

This scale adopts four items from Rosenberg’s (1965) original self-esteem scale. This scale includes the following items: (a) I take an optimistic attitude toward myself; (b) I am satisfied with myself; (c) I certainly feel useless at times; and (d) Occasionally, I think I am no good at all. Youths’ responses are measured by a 4-point Likert-type scale (1 = strongly agree, 2 = agree, 3 = disagree, and 4 = strongly disagree) for the first two items and are reversed coded for the last two items. While these four items belong to two concepts of self (Shahani, Dipboye, & Phillips, 2011), positive (i.e., self-enhancement) and negative (i.e., self-derogation), we summed them to create a self-esteem index because our purpose is to capture overall self-esteem, without intention to distinguish between these two concepts. A higher score on this scale indicates higher self-esteem (α = .72). Again, these items were used in previous studies in Taiwan and showed their usefulness in predicting related adolescent developments (Tzeng & Yi, 2009; C.-I. Wu & Huang, 2010).

Deviance (Wave 4)

At Wave 4, each student was asked to report whether they had had engaged in several common deviant behaviors in the past year (1 = yes). These items are “smoke cigarettes,” “drink alcohol,” “get into trouble in school,” “run away from home,” “skip classes,” “get into a fight,” and “use illegal drugs (e.g., MDMA).” Summation of these seven items created a deviancy scale. Hence, our measure essentially captured the variety of deviant acts a student was involved in during the previous year, and this kind of measure is also advocated by others (Hindelang, Hirschi, & Weis, 1981). A higher score indicates great involvement in various deviant acts. As can be expected, the scale is highly skewed (skewness = 3.33; kurtosis = 12.98); hence, we transformed this scale by using the logarithm. Although the transformed scale is not normally distributed (skewness = 2.20; kurtosis = 4.02), it does not depart from normal too much and is acceptable for structural equation modeling (SEM)-related analyses (Kline, 1998).

Demographic variables

We control for two demographic variables: students’ gender and family income. For gender, studies have shown that adolescent gender has different effects on the outcomes we used here, such as deviance (Liu & Kaplan, 1999), depression (Hyde, Mezulis, & Abramson, 2008), and self-esteem (Moksnes & Espnes, 2013). With regard to family income, several studies have revealed that income may influence adolescent delinquency (McNulty, Bellair, & Watts, 2013), self-esteem (Walker & Greene, 1986; Yeung & Conley, 2008), and depression (Wickrama, Noh, & Elder, 2009). Students’ gender was coded as male (1) and female (0). Family income was based on parental report and supplemented by students’ self-report. In doing this, we reduced the problem of high missingness on reporting income to the minimum. This variable was coded in 13 categories by monthly income in the local currency (New Taiwan Dollars). These 13 categories ranged from less than NT$30,000 (1), to over NT$150,000 (13) (see the descriptive statistics of all variables in Table 1).

Table 1.

The Descriptive Statistics of All Variables.

	Wave 1 (n = 2,690)	Wave 2 (n = 2,582)	Wave 3 (n = 2,520)	Wave 4 (22,44)
Family cohesion	M = 3.00; SD = .64	M = 2.88; SD = .67	M = 2.80; SD = .45
Life satisfaction	M = 3.13; SD = .50	M = 2.94; SD = .50	M = 2.84; SD = .52
Depression			M = .59; SD = .52
Self-esteem			M = 2.67; SD = .58
Deviance				M = .17; SD = .39
Gender
Male	1,378 (51.2%)
Female	1,312 (48.8%)
Family income
≤30,000	412 (15.3%)
30,000-49,999	727 (27.0%)
50,000-59,999	444 (16.5%)
60,000-69,999	276 (10.3%)
70,000-79,000	156 (5.8%)
80,000-89,999	206 (7.7%)
90,000-99,999	74 (2.8%)
100,000-109,999	212 (7.9%)
110,000-119,999	22 (0.8%)
120000-129,999	32 (1.2%)
130,000-139,999	14 (0.5%)
140,000-149,999	9 (0.3%)
≥150,000	106 (3.9%)
	M = 3.97; SD = 2.97

Note. For variables that were dichotomized in nature, we presented frequency distribution. Other variables were displayed with mean (M) and standard deviation (SD).

Analytic Plane

Change in family cohesion during early adolescence has not been examined much, especially in non-Western samples. To capture the “change,” we adopt the latent growth modeling (LGM) approach. This is a class of structural equation models that use repeated measures of an attribute (e.g., family cohesion) as indicators of latent variables (e.g., linear growth slope) which represent individual change over time. In a regular linear LGM, two latent variables are estimated: initial status or intercept, and the growth rate or slope. These two latent variables are measured by a repeated measure (e.g., family cohesion measured at Wave 1 to Wave 3). The extension of LGM is to combine two univariate LGM together, creating a parallel LGM model. This method allows us to see the interrelationship between two trajectories.

The LGM analyses were completed using Mplus Version 7.0 (Muthén & Muthén, 2010). Mplus calculates a χ² test to gauge the fit of a model to its data. Lack of significance indicates an acceptable model fit. However, because the χ² test is sensitive to sample size, scholars have suggested that investigators report multiple fit indices and specify the “critical value” for each (Brown, 2006; Hoyle & Panter, 1995). Goodness-of-fit was based on three alternative indices because of large sample size. First, we included both CFI and TLI, and a value of .95 is suggested (Hu & Bentler, 1999) for both. Second, RMSEA with values of .05 or less suggests a good fit (Brown & Cudeck, 1993). In addition to point estimates for RMSEA, some scholars have argued that reporting a 90% confidence interval (CI) of RMSEA provides further information on model fit (MacCallum, Browne, & Sugawara, 1996). Hence, we also included a 90% CI for RMSEA.

Missingness is inevitable in survey research, especially longitudinal research. Missing data were dealt with by using the full information maximum likelihood method for subsequent analyses. This method provides a correct standard error (Allison, 2002).¹

Results

From Table 1 we can see that the students are about evenly divided, with males occupying about 51% of the total sample (n = 1,378). With regard to income, we see that about three quarters of all the youths are from families with monthly income equal to or lower than $79,000. The average monthly income in this sample coincides with Income Group 4 ($60,000-$69,999). This is somewhat lower than the average family income ($90,957) of Taiwan in 2000 (Directorate-General of Budget, Accounting and Statistics, 2016). With regard to family cohesion and life satisfaction, mean score, shown in the upper part of Table 1, does show a decline over the 3 years. Also, because of the general decline in these two attributes, we expect to see these two attributes covary across these three time points. The mean of depression for the students in the northern part of Taiwan is low (M = .59), which is between no symptoms and experiencing symptoms but not serious. Finally, the average self-esteem for these adolescents is moderate to high (2.67 on a 4-point scale).

We then fit the interlocking LGM model with correlations among the observable attributes at each time point and control variables (gender and family income). The model shows acceptable fit, χ² = 84.65 (10); CFI = .986; TLI = .963; RMSEA = .053; 90% RMSEA CI: .043, .063. The results (Table 2, Model 1) indicate that both the mean of the intercept and slope for life satisfaction (intercept = 3.12; slope = −.14) and family cohesion (intercept = 3.00; slope = −.11) are significant at the conventional level. Hence, not only adolescents’ life satisfaction but also their family cohesion deteriorated gradually during early adolescence. In addition, we allowed the initial level of both life satisfaction and family cohesion to predict the change (slope) of family cohesion and life satisfaction. The results show that adolescents who reported a higher level of life satisfaction at the beginning of junior high school (age 13 years) had a slower rate of decline in family cohesion (β = −.06) during the 3 years of junior high school. Similarly, if a juvenile had a high level of family cohesion in the first year of junior high school, this particular youth tended to have a slower decline in life satisfaction during junior high school (β = −.06). Furthermore, the correlations between intercepts and slopes of both attributes are significant (r =. 65 for intercept; r = .23 for slope). This indicates that youths from more cohesive families often had a higher level of life satisfaction when entering junior high school. Moreover, adolescents whose level of life satisfaction declined over the years also experienced a decline in family cohesion. Besides the intraindividual results, some significant observations were also found in the correlation at the same time point for the three waves (r = .16 for Wave 1; r = .17 for Wave 2; r = .28 for Wave 3). This shows that individuals who had a high level of life satisfaction were also more likely to have a high level of family cohesion at the same time point. The results indicate that the within-individual change in life satisfaction and family cohesion captures most of the interrelationship between these two variables.

Table 2.

LGM of Family Cohesion and Life Satisfaction and Life Outcomes.

	Model 1	Model 2
	Coefficient (SE)	Coefficient (SE)
LGM means
Intercept(Family cohesion)	3.00 (.01)**	3.00 (.01)**
Slop(Family cohesion)	−0.11 (.01)**	−0.11 (.01)**
Intercept(Life satisfaction)	3.12 (.01)**	3.12 (.01)**
Slop(Life satisfaction)	−0.14 (.01)**	−0.14 (.01)**
Regression
IF → SL	−0.06 (.01)**	−0.06 (.01)**
IL → SF	−0.06 (.02)**	−0.06 (.02)**
Male → IF	11 (.02)**	0.11 (.02)**
Male → SF	−0.05 (.01)**	−0.05 (.01)**
Male → IL	0.09 (.02)**	0.09 (.02)**
Income → IF	0.02 (.004)**	0.02 (.004)**
Income → SL	0.01 (.002)**	0.01 (.002)**
IL → W3 Depression		−0.57 (.05)**
SL → W3 Depression		−0.83 (.15)**
IL → W3 Self-esteem		0.79 (.06)**
SL → W3 Self-esteem		1.50 (.22)**
SF → W3 Self-esteem		0.33 (.12)**
IF → W4 Deviance		−0.06 (.03)*
SF → W4 Deviance		−0.22 (.08)**
Male → W3 Depression		−0.11 (.02)**
Income → W3 Depression		0.01 (.003)*
Male → W3 Self-esteem		0.10 (.02)**
W3 Depression → W4 Deviance		0.09 (.02)**
Male → W4 Deviance		0.13 (.02)**
Indirect path	Confidence interval with bootstrapping
IF → SL → W3 Depression	.05 [.026, .094]
IF → SL → W3 Self-esteem	−.09 [−.167, −.056]
IL → SF → W4 Deviance	.01 [.001, .036]
Correlation
IF with IL	.65**	.65**
SF with SL	.23**	.23**
IF with SF	−.18**	−.18**
IL with SL	−.32**	−.33**
W1 FA with W1 LS	.22**	.22**
W2 FA with W2 LS	.19**	.19**
W3 FA with W3 LS	.35**	.36**
Depression with self-esteem		−.20**

Note. n = 2,690; LGM = latent growth modeling; IF = intercept of family cohesion; SL = slope of life satisfaction; IL = intercept of life satisfaction; SF = slope of family cohesion; FA = family cohesion; LS = life satisfaction.

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

Next, we turn to explore the relationship between the trajectories of life satisfaction and family cohesion, on the one hand, and individual internal functioning (e.g., depression and self-esteem) and external functioning (e.g., deviance), on the other. The model fit the data well based on various standards, χ² = 103.26 (16); CFI = .987; TLI = .955; RMSEA = .045; 90% RMSEA CI: .037, .054. The results from this model (Table 2, Model 2) show different pictures for these two important attributes in early adolescence. With regard to individual psychological characteristics, the “growth” and initial level of life satisfaction were significantly related to both depression (β = −.57 for intercept; β = −.83 for slope) and self-esteem (β = .79 for intercept; β = 1.50 for slope). Specifically, students whose life satisfaction was low at about age 13 years were more likely than their counterparts to have more depressive symptoms and a lower level of self-esteem 2 years later. Furthermore, a decline in life satisfaction was also related to greater depressive symptoms and a lower level of self-esteem for adolescents during these 3 years. In contrast, only the change in family cohesion was related to a lower level of self-esteem at the end of junior high school (β = .33). Finally, and importantly, as to the relationships between these variables and students’ deviant acts in the first year of high school (Wave 4), the results show that only family cohesion had significant effects on juvenile deviance (β = = −.06—intercept; β = −.22—slope). Consequently, students who started with a low level of family cohesion and whose family cohesion declined faster than other students were more likely to have a higher level of deviance when entering middle adolescence (e.g., 10th grade). In addition to family cohesion, depression was also related to deviance 1 year later (β = .09). In contrast, life satisfaction and self-esteem were not related to deviance.

The above model also shows some possible indirect effects based on the Baron and Kenny (1986) rules. Central to the present study is whether life satisfaction may be indirectly related to deviance through change in family cohesion, and whether family cohesion may be related to both self-esteem and depression through change in life satisfaction. We then check this possibility with bootstrapping CIs. The results (central part of Table 2) show that all three indirect effects are all statistically significant (i.e., CI does not include zero). Hence, while the initial level of life satisfaction was not related to deviance, for example, it did have significant indirect effects on deviance through influencing family cohesion. In other words, students who started their adolescence with a low level of life satisfaction were more likely to experience a more rapid decline in family cohesion, which in turn put them at higher risk of involvement in various deviant acts in middle adolescence (age about 15) than other youths. Similarly, although early family cohesiveness may not be directly related to adolescent internal functioning (e.g., depression and self-esteem), it may be related to it through affecting change in life satisfaction during early adolescence.

In all these models, gender and family income show some significant relationships with different variables. For example, boys had a lower level of depression (β = −.11) but higher self-esteem (β = .10) and deviant involvement than girls (β = .13). Moreover, male adolescents had higher levels of life satisfaction (β = .09) and reported higher levels of family cohesion (β = .11) than female adolescents in the first year of junior high school, but they had a steeper rate of “growth,” which indicates that boys’ subjective feeling of family cohesion declined more quickly than girls’ during this time period. Similarly, family income was also related to students’ initial life family cohesion (β = .02) and the decline in life satisfaction (β = .01), such that youths from higher income families perceived a higher level of family cohesion at age 13 years, but their level of life satisfaction declined more quickly than their counterparts’ during the junior high school years.

Conclusion and Discussion

This study set out to examine changes in family cohesion and life satisfaction during early adolescence, and their interrelationship. In addition, it investigated the influences of these two important attributes on later adolescent outcomes, including depression, self-esteem, parental attachment, and deviance. Using a Taiwanese adolescent sample and LGM, we have several interesting results.

First, consistent with previous studies (Baer, 2002; Park, 2005), we found that family cohesion and life satisfaction declined during early adolescence (13-15). The decline in both attributes during adolescence can be understood from two perspectives. First, psychologists point out that adolescence is a period full of various changes, including physical (e.g., puberty), emotional (e.g., volatile emotions), and relational ones (e.g., seeking autonomy), as well as school transition (Shoshani & Slone, 2013). As such, youngsters at this age often face many challenges, which in turn may contribute to a low level of life satisfaction. Furthermore, one of the tasks in adolescence is to build self-identity and seek autonomy; hence, adolescents may encounter increased conflicts with parents and participate in fewer family activities (Erikson, 1994), which leads to low family cohesion.

Second, the educational competition in Taiwan starts surging when individuals become junior high school students. The subjects and content of their studies become heavier and harder than they were in elementary school. In addition, awaiting them is a national entrance examination, which is used to place students in different high schools and/or educational tracking. Hence, students in Taiwan during this time period spend a great amount of time studying (Yi, Wu, Chang, & Chang, 2009), and parents use various methods to control or regulate their children. Consequently, the stress from schoolwork and tighter parental control contribute to a decline in both life satisfaction and family cohesion.

Besides the decline in family cohesion and life satisfaction, this study also modeled the interrelationship between these two attributes. The results indicate that students who had a higher level of family cohesion (e.g., emotionally bonded) had a slower rate of decline in their life satisfaction during early adolescence. Similarly, greater life satisfaction was likely to decelerate the deterioration of family cohesion in the same period of time. This is consistent with SWB studies that show that family processes (e.g., family cohesion) have important impacts on adolescents’ life satisfaction (Phillips, 2012; Shek, 2005) and are positively related to parental emotional involvement and support (Nickerson & Nagle, 2004). This dynamic interrelation between life satisfaction and family cohesion reveals a more vivid picture than previous studies have afforded.

Second, we discovered that these two attributes exerted different influences on adolescents. The initial level of life satisfaction was directly related to internal functioning (e.g., depression) but indirectly related to external functioning (e.g., deviance), such that a higher initial level of life satisfaction at age 13 years was related to higher self-esteem and lower depression 2 years later. Also, these same students may have experienced a slower decline in family cohesion, which in turn led to a lower level of deviance 3 years later. In addition, the change in life satisfaction was also significantly related to youngsters’ depression and self-esteem. That is, the slower deterioration of life satisfaction during junior high school contributed to better internal functioning (e.g., lower depression and higher self-esteem) in the last year of junior high school. These results undergird previous findings that life satisfaction is related to various life domains in adolescence (Proctor et al., 2009). However, unexpectedly, the trajectory of life satisfaction of a youth was not directly related to later deviance, which is inconsistent with previous studies (Coker et al., 2000; MacDonald et al., 2005). This inconsistency may be understood from a general strain theory (GST) perspective (Agnew, 2006). According to Agnew (2006), some strains may not be criminogenic because they are from conventional society (e.g., taking care of a sick spouse). Hence, a low level of satisfaction, as we measured here, may be one of this kind of negative life experience or strain. Furthermore, GST also argued that strain is related to deviance indirectly through low social control, a deviant learning environment, and negative emotions. Hence, there is no direct relationship between “growth” in life satisfaction and deviance, because the relationship is indirect. For the present study, this indirect link occurred through depression, which is consistent with the GST perspective (95% CI: −.15, −.03).

Third, our results indicate that both a lower initial level of and a decline in family cohesion were related to a higher level of deviance 3 years later. We also found that lower family cohesion was related to a faster decline in life satisfaction, which in turn was related to a higher level of depression and lower self-esteem. The relationship between a low level of family cohesion and later negative adolescent functioning is consistent with previous findings that family plays a central role in influencing both adolescents’ externalizing and internalizing problems (Lian & Yusooff, 2009; Marsiglia, Parsai, & Kulis, 2009; Reeb et al., 2015; Rivera, Guarnaccia, Mulvaney-Day, Torres, & Alegria, 2008). But more importantly, we show that changes in the family environment are also pernicious to adolescents; that is, a family that gradually becomes less cohesive is conducive to adolescent deviance. This result echoes previous studies that used a more static approach; hence, a low level of family cohesion at one time point and a decline in family cohesion over time are both detrimental to adolescent functioning.

Although this study offers valuable insights into adolescent development, some limitations need to be mentioned. First, this estimation of the trajectory is based on only three time points. Without more time points, the true trajectory of or change in family cohesion and life satisfaction (e.g., a curvilinear one) might not be observable. Second, while this study uses a Taiwanese sample, it is not representative. However, the variation of this sample on social economic status and urbanization is high, which may increase our confidence in generalizing the results to the whole island. Finally, adolescent development is influenced by various factors, but we only incorporate family (e.g., family cohesion) and individual factors (e.g., life satisfaction). Other important factors are not in the model, such as neighborhood. Previous studies have shown that neighborhood has significant effects on adolescent development (Leventhal & Brooks-Gunn, 2000; Nebbitt, Lombe, Yu, Vaughn, & Stokes, 2012; Nebbitt et al., 2014; Sampson, 2006).

In conclusion, adolescents, as shown, develop emotional distance from their family, so we discover a downward trajectory of positive family cohesion; however, a high level of family cohesion has both concurrent influences on adolescents’ SWB (e.g., life satisfaction) and also long-term influences on adolescents’ lives (e.g., a slower decline in life satisfaction and deviance). Hence, to prevent such emotional distancing, parents can spend time listening to their children’s worries and doing things together with them, and provide emotional and informational support. Schoolteachers may wish to promote intrafamily interaction when conducting parent–teacher meetings. Promoting family cohesiveness may not only benefit adolescent life satisfaction concurrently but also improve their later development beyond adolescence. Doing this can minimize the various risks of developing low self-esteem, deviance, and low well-being (e.g., dissatisfaction with relationships with others). As many other studies have demonstrated, a well-managed adolescence (e.g., high life satisfaction and self-esteem, and low levels of depression and deviance) can lead to fewer adjustment problems in the future and positive life outcomes in later life (e.g., a better job and educational attainment).

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Wen-Hsu Lin, PhD, is a postdoctoral research fellow at the Institute of Sociology, Academia Sinica, Taipei, Taiwan. His research focuses on adolescent development, delinquency, health-related behavior and quantitative research methods.

Chin-Chun Yi, PhD, is a research fellow at the Institute of Sociology, Academia Sinica, Taiwan. Her research interests include marital power and marital relations, intergenerational relations, and youth development.

References

Abbott

Wallace

(2012). Social quality: A way to measure the quality of society. Social Indicators Research, 108, 153-167. doi:10.1007/s11205-011-9871-0

Agnew

(2006). Pressured into crime: An overview of general strain theory. Los Angeles, CA: Roxbury Publishing.

Allison

P. D.

(2002). Missing data (SAGE University Paper Series on Quantitative Applications in the Social Sciences, 07-136). Thousand Oaks, CA: Sage.

Baer

(2002). Is family cohesion a risk or protective factor during adolescent development? Journal of Marriage and Family, 64, 668-675.

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.

Baumann

Spitz

Predine

Choquet

Chau

(2007). Do male and female adolescents differ in the effect of individual and family characteristics on their use of psychotropic drugs? European Journal of Pediatrics, 166, 29-35. doi:10.1007/s00431-006-0203-3

Boden

J. M.

Fergusson

D. M.

Horwood

L. J.

(2008). Does adolescent self-esteem predict later life outcomes? A test of the causal role of self-esteem. Development and Psychopathology, 20, 319-339.

Bradley

R. H.

Corwyn

R. F.

(2004). Life satisfaction among European American, African American, Chinese American, Mexican American, and Dominican American adolescents. International Journal of Behavioral Development, 28, 385-400.

Bronfenbrenner

(1992). Ecological systems theory. In Ross

(Ed.), Six theories of child development: Revised formulations and current issues (pp. 187-249). London, England: Jessica Kingsley.

10.

Brown

T. A.

(2006). Confirmatory factor analysis: For applied research. New York, NY: Guilford Press.

11.

Brown

M. W.

Cudeck

(1993). Alternative ways of assessing model fit. In Bollen

K. A.

Long

J. S.

(Eds.), Testing structural equation models (pp. 136-162). New Bury Park, CA: Sage.

12.

Chan

Y.-K.

Lee

R. P. L.

(2006). Network size, social support and happiness in later life: A comparative study of Beijing and Hong Kong. Journal of Happiness Studies, 7, 87-112. doi:10.1007/s10902-005-1915-1

13.

Chang

Y.-H.

C.-C.

Lin

K.-H.

(2008). Kin network and its effect on the psychological well-being of the youth: The case of Taiwan. Journal of Comparative Family Studies, 39, 19-37.

14.

Cicognani

(2011). Coping strategies with minor stressors in adolescence: Relationships with social support, self-efficacy, and psychological well-being. Journal of Applied Social Psychology, 41, 559-578. doi:10.1111/j.1559-1816.2011.00726.x

15.

Coker

A. L.

McKeown

R. E.

Sanderson

Davis

K. E.

Valois

R. F.

Huebner

E. S.

(2000). Severe dating violence and quality of life among South Carolina high school students. American Journal of Preventive Medicine, 19, 220-227.

16.

Day

Jankey

S. G.

(1996). Lessons from the literature: Toward a holistic model of quality of life. In Renwick

Brown

Nagler

(Eds.), Quality of life in health promotion and rehabilitation: Conceptual approaches, issues and applications (pp. 39-50). Thousand Oaks, CA: Sage.

17.

Derogatis

L. R.

(1983). SCL-90-R administration, scoring, and procedure manual-II. Towson, MD: Clinical Psychometric Research.

18.

Directorate-General of Budget, Accounting and Statistics. (2016). Average family income and expenditure. Retrieved from http://win.dgbas.gov.tw/fies/214.asp

19.

Doge

K. A.

Sherrill

M. R.

(2006). Deviant peer group effects in youth mental health interventions. In Landsford

J. E.

Doge

Dishion

T. J.

(Eds.), Deviant peer influences in programs for youth: Problems and solutions (pp. 97-121). New York, NY: Guilford Press.

20.

Ekas

N. V.

Lickenbrock

D. M.

Whitman

T. L.

(2010). Optimism, social support, and well-being in mothers of children with autism spectrum disorder. Journal of Autism and Developmental Disorder, 40, 1274-1284. doi:10.1007/s10803-010-0986-y

21.

Erikson

E. H.

(1994). Identity: Youth and crisis (No. 7). New York, NY: W.W. Norton.

22.

Fletcher

(2013). Adolescent depression and adult labor market outcomes. Southern Economic Journal, 80, 26-49. doi:10.4284/0038-4038-2011.193

23.

Fosco

G. M.

Caruthers

A. S.

Dishion

T. J.

(2012). A six-year predictive test of adolescent family relationship quality and effortful control pathways to emerging adult social and emotional health. Journal of Family Psychology, 26, 567-575.

24.

Gilman

Huebner

E. S.

(2006). Characteristics of adolescents who report very high life satisfaction. Journal of Youth and Adolescence, 35, 311-319.

25.

Goldbeck

Schmitz

T. G.

Besier

Herschbach

Henrich

(2007). Life satisfaction decreases during adolescence. Quality of Life Research, 16, 969-979. doi:10.1007/s11136-007-9205-5

26.

Gray

R. S.

Chamratrithirong

Pattaravanich

Prasartkul

(2013). Happiness among adolescent students in Thailand: Family and non-family factors. Social Indicators Research, 110, 706-719. doi:10.1007/s11205-011-9954-y

27.

Hamama

Arazi

(2012). Aggressive behaviour in at-risk children: Contribution of subjective well-being and family cohesion. Child & Family Social Work, 17, 284-295.

28.

Hindelang

M. J.

Hirschi

Weis

J. G.

(1981). Measuring delinquency. Thousand Oaks, CA: Sage.

29.

Hou

(2001). Family structure, family relationships, and delinquency. Research in Applied Psychology, 11, 25-43.

30.

Hoyle

R. H.

Panter

A. T.

(1995). Writing about structural equation models. In Hoyle

R. H.

(Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 159-176). Thousand Oaks, CA: Sage.

31.

Hsien

M.-H.

(2003). 國小學童之父親參與、幸福感及學業成就之相關研究 [The relationships among paternal involvement, well-being and academic achievement of primary school students] (Unpublished master’s thesis). National Changhua University of Education, Changhua, Taiwan.

32.

L.-T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

33.

Huebner

E. S.

(2004). Research on assessment of life satisfaction of children and adolescents. Social Indicators Research, 66, 3-33.

34.

Hyde

J. S.

Mezulis

A. H.

Abramson

L. Y.

(2008). The ABCs of depression: Integrating affective, biological, and cognitive models to explain the emergence of the gender difference in depression. Psychological Review, 115, 291-313. doi:10.1037/0033-295X.115.2.291

35.

Jaggers

J. W.

Church

W. T.

II Tomek

Hooper

L. M.

Bolland

K. A.

Bolland

J. M.

(2015). Adolescent development as a determinant of family cohesion: A longitudinal analysis of adolescents in the Mobile Youth Survey. Journal of Child and Family Studies, 24, 1625-1637.

36.

Jou

Y.-H.

(2015). 青少年至成年初期親子關係的變化及其影響 [Longitudinal effect of parent-child interactions on psychological well-being during the transition from adolescence to young adulthood]. Chinese Journal of Psychology, 57, 67-89.

37.

Kammeyer-Mueller

J. D.

Judge

T. A.

Piccolo

R. F.

(2008). Self-esteem and extrinsic career success: Test of a dynamic model. Applied Psychology, 57, 204-224. doi:10.1111/j.1464-0597.2007.00300.x

38.

Kapungu

C. T.

Holmbeck

G. N.

Paikoff

R. L.

(2006). Longitudinal association between parenting practices and early sexual risk behaviors among urban African American adolescents: The moderating role of gender. Journal of Youth and Adolescence, 35, 787-798.

39.

Kline

R. B.

(1998). Principles and practice of structural equation modeling. New York: Guilford Press.

40.

Kung

H.-Y.

Lee

C.-Y.

(2015). 影響國中經濟弱勢學生之學業表現與中輟傾向之因素：「脈絡－自我－行動－結果」之動機發展自我系統模式為取向 [Influence of academic achievement and dropout on economically disadvantaged students: An investigation of the “context–self–action–result” self-system model of motivational development]. Journal of Research in Education Sciences, 60, 55-92. doi:10.6209/JORIES.2015.60(4).03

41.

Lai

K. W.

McBride-Chang

(2001). Suicidal ideation, parenting style, and family climate among Hong Kong adolescents. International Journal of Psychology, 36, 81-87.

42.

Lee

Larson

(2000). The Korean “examination hell”: Long hours of studying, distress, and depression. Journal of Youth and Adolescence, 29, 249-271.

43.

Leung

C. Y.-W.

McBride-Chang

Lai

B. P.-Y.

(2004). Relations among maternal parenting style, academic competence, and life satisfaction in Chinese early adolescents. Journal of Early Adolescence, 24, 113-143. doi:10.1177/0272431603262678

44.

Leventhal

Brooks-Gunn

(2000). The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126, 309-337.

45.

Levitt

M. J.

Levitt

Bustos

Crooks

N. A.

Santos

J. D.

Telan

. . . Milevsky

(2005). Patterns of social support in the middle childhood to early adolescent transition: Implications for adjustment. Social Development, 14, 398-420. doi:10.1111/j.1467-9507.2005.00308.x

46.

Lewis

A. D.

Huebner

E. S.

Malone

P. S.

Valois

R. F.

(2011). Life satisfaction and student engagement in adolescents. Journal of Youth and Adolescence, 40, 249-262. doi:10.1007/s10964-010-9517-6

47.

Lian

T. C.

Yusooff

(2009). The effects of family functioning on self-esteem of children. European Journal of Social Sciences, 9, 643-650.

48.

Lin

W.-H.

(2011). General strain theory and juvenile delinquency: A cross-cultural study (Unpublished doctoral dissertation). University of South Florida, Tampa.

49.

Liu

Kaplan

H. B.

(1999). A longitudinal test of mediating mechanisms. Criminology, 37, 195-215.

50.

Lober

Farrington

(2012). From juvenile delinquency to adult crime: Criminal careers, justice policy, and prevention. Oxford, UK: Oxford University Press.

51.

MacCallum

R. C.

Browne

M. W.

Sugawara

H. M.

(1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130-149.

52.

MacDonald

J. M.

Piquero

A. R.

Valois

R. F.

Zullig

K. J.

(2005). The relationship between life satisfaction, risk-taking behaviors, and youth violence. Journal of Interpersonal Violence, 20, 1495-1518. doi:10.1177/0886260505278718

53.

Marsiglia

F. F.

Parsai

Kulis

(2009). Effects of familism and family cohesion on problem behaviors among adolescents in Mexican immigrant families in the Southwest United Sates. Journal of Ethnic & Cultural Diversity in Social Work, 18, 203-220.

54.

Martin

Huebner

E. S.

Valois

R. F.

(2008). Does life satisfaction predict victimization experiences in adolescence? Psychology in the Schools, 45, 705-714. doi:10.1002/pits.20336

55.

McNulty

T. L.

Bellair

P. E.

Watts

S. J.

(2013). Neighborhood disadvantage and verbal ability as explanations of the Black-White difference in adolescent violence. Crimine & Delinquency, 59, 140-160.

56.

Moffitt

(1993). “Life-course-persistent” and “adolescence-limited” antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701.

57.

Moksnes

U. K.

Espnes

G. A.

(2013). Self-esteem and life satisfaction in adolescents—Gender and age as potential moderators. Quality of Life Research, 22, 2921-2928. doi:10.1007/s11136-013-0427-4

58.

Moos

R. H.

Moos

B. S.

(1994). Family environment scale manual. Palo Alto, CA: Consulting Psychologist Press.

59.

Muthén

L. K.

Muthén

B. O.

(2010). Mplus: Statistical analysis with latent variables: User’s guide. Los Angeles, CA: Author.

60.

Naicker

Galambos

Zeng

Senthilselvan

Colman

(2013). Social, demographic, and health outcomes in the 10 years following adolescent depression. Journal of Adolescent Health, 52, 533-538.

61.

Nash

S. G.

McQueen

Bray

J. H.

(2005). Pathway to adolescent alcohol use: Family environment, peer influence, and parental expectations. Journal of Adolescent Health, 37, 19-28.

62.

Nebbitt

V. E.

Lombe

M. L.

Vaughn

M. G.

Stokes

(2012). Ecological correlates of substance use in African American adolescents living in public housing communities: Assessing the moderating effects of social cohesion. Children and Youth Service Review, 34, 338-347.

63.

Nebbitt

V. E.

Williams

J. H.

Lombe

M. L.

McCoy

Stephens

(2014). Descriptive analysis of individual and community factors among African American youths in urban public housing. Social Work, 59, 231-241.

64.

Newman

B. R.

Newman

P. R.

Griffen

O’Connor

Spas

(2007). The relationship of social support to depressive symptoms during the transition to high school. Adolescence, 42, 441-459.

65.

Nickerson

A. B.

Nagle

(2004). The influence of parent and peer attachments on life satisfaction in middle childhood and early adolescence. Social Indicators Research, 66, 35-60.

66.

Olson

D. H.

Portner

Lavee

(1985). FACES III. St. Paul: Department of Family Science, University of Minnesota.

67.

Olson

D. H.

Sprenkle

D. H.

Russell

C. S.

(1979). Circumplex model of marital and family systems I: Cohesion and adaptability dimensions, family types, and clinical applications. Family Process, 18, 3-28. doi:10.1111/j.1545-5300.1979.00003.x

68.

Őnder

F. C.

Yilmaz

(2012). The role of life satisfaction and parenting styles in predicting delinquent behaviors among high school students. Educational Science: Theory & Practice, 12, 1744-1748.

69.

Orth

Robins

R. W.

Roberts

B. W.

(2008). Low self-esteem prospectively predicts depression in adolescence and young adulthood. Journal of Personality and Social Psychology, 95, 695-708.

70.

Park

(2005). Life satisfaction among Korean children and youth: A developmental perspective. School Psychology International, 26, 209-223.

71.

Pavot

Diener

E. D.

(1993). Review of the Satisfaction With Life Scale. Psychological Assessment, 5, 164-172.

72.

Phillips

T. M.

(2012). The influence of family structure vs. family climate on adolescent well-being. Child & Adolescent Social Work Journal, 29, 103-110. doi:10.1007/s10560-012-0254-4

73.

Proctor

C. L.

Linley

P. A.

Maltby

(2009). Youth life satisfaction: A review of the literature. Journal of Happiness Study, 10, 583-630.

74.

Rayle

A. D.

Chung

K.-Y.

(2007). Revisiting first-year college students’ mattering: Social support, academic stress, and the mattering experience. Journal of College Student Retention, 9, 21-37. doi:10.2190/X126-5606-4G36-8132

75.

Reeb

B. T.

Chan

S. Y. S.

Conger

K. J.

Martin

M. J.

Hollis

N. D.

Serido

Russell

S. T.

(2015). Prospective effects of family cohesion on alcohol-related problems in adolescence: Similarities and differences by race/ethnicity. Journal of Youth and Adolescence, 44, 1941-1953.

76.

Ritakallio

Koivisto

A. M.

von der Pahlen

Pelkonen

Marttunen

Kaltiala-Heino

(2008). Continuity, comorbidity and longitudinal associations between depression and antisocial behaviour in middle adolescence: A 2-year prospective follow-up study. Journal of Adolescence, 31, 355-370.

77.

Rivera

F. I.

Guarnaccia

P. J.

Mulvaney-Day

Torres

Alegria

(2008). Family cohesion and its relationship to psychological distress among Latino groups. Hispanic Journal of Behavioral Science, 30, 357-378.

78.

Rosenberg

(1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

79.

Salmela-Aro

Tuominen-Soini

(2010). Adolescents’ life satisfaction during the transition to post-comprehensive education: Antecedents and consequences. Journal of Happiness Studies, 11, 683-701. doi:10.1007/s10902-009-9156-3

80.

Sampson

R. J.

(2006). Collective efficacy theory: Lessons learned and directions for future inquiry. In Cullen

F. T.

Wright

J. P.

Blevins

K. R.

(Eds.), Taking stock: The status of criminological theory (pp. 149-168). New Brunswick, NJ: Transaction.

81.

Seaman

S. R.

White

I. R.

(2011). Review of inverse probability weighting for dealing with missing data. Statistical Methods in Medical Research, 22, 278-295.

82.

Shahani

Dipboye

R. L.

Phillips

A. P.

(2011).Global self-esteem as a correlate of work-related attitudes: A question of dimensionality. Journal of Personality and Assessment, 54, 276-288.

83.

Shamai

Kimhi

(2007). Teenagers’ response to threat of war and terror: Gender and the role of social systems. Community Mental Health Journal, 43, 359-374. doi:10.1007/s10597-006-9080-5

84.

Shek

D. T. L.

(2005). A longitudinal study of Chinese cultural beliefs about adversity, psychological well-being, delinquency and substance abuse in Chinese adolescents with economic disadvantage. Social Indicator Research, 71, 385-409.

85.

Shoshani

Slone

(2013). Middle school transition from the strengths perspective: Young adolescents’ character strengths, subjective well-being, and school adjustment. Journal of Happiness Study, 14, 1163-1181. doi:10.1007/s10902-012-9374-y

86.

Siennick

S. E.

(2007). The timing and mechanisms of the offending-depression link. Criminology, 45, 583-615. doi:10.1111/j.1745-9125.2007.00091.x

87.

Swislo

J. F.

Orth

(2013). Does low self-esteem predict depression and anxiety? A meta-analysis of longitudinal studies. Psychological Bulletin, 139, 213-240.

88.

Sturge-Apple

M. L.

Davies

P. T.

Cummings

E. M.

(2010). Typologies of family functioning and children’s adjustment during the early school years. Child Development, 81, 1320-1335. doi:10.1111/j.1467-8624.2010.01471.x

89.

J.-H.

(2003). 青少年憂鬱與社會畏懼、雙親教養態度之相關研究 [The relationship among adolescent depression, social phobia and parenting bonding] (Unpublished master’s thesis). National Changhua University of Education, Changhua, Taiwan.

90.

Suldo

S. M.

Huebner

E. S.

(2004). The role of life satisfaction in the relationship between authoritative parenting dimensions and adolescent problem behavior. Social Indicators Research, 66, 165-195.

91.

Sun

R. C. F.

Shek

D. T. L.

(2012). Positive youth development, life satisfaction and problem behavior among Chinese adolescents in Hong Kong: A replication. Social Indicator Research, 105, 541-559. doi:10.1007/s11205-011-9786-9

92.

Tzeng

S.-P.

C.-C.

(2009, December). The nonlinear effects of self-esteem on adolescent delinquency. Paper presented at the International Conference on Youth Studies, Taipei, Taiwan.

93.

Videon

T. M.

(2002). The effects of parent-adolescent relationships and parental separation on adolescent well-being. Journal of Marriage and Family, 64, 489-503.

94.

Walker

Greene

(1986). The social context of adolescent self-esteem. Journal of Youth and Adolescence, 15, 315-322.

95.

Wang

C.-C.

(2006). 高中職生情緒智慧、憂鬱傾向與情緒調整之研究 [A study of the emotional intelligence, depression and emotional regulation of senior and vocational high school students]. Chia-Nan Annual Bulletin, 32, 484-507.

96.

White

Shelton

K. H.

Elgar

F. J.

(2014). Prospective associations between the family environment, family cohesion, and psychiatric symptoms among adolescent girls. Child Psychiatry & Human Development, 45, 544-554.

97.

Whitesell

N. R.

Mitchell

C. M.

Spicer

(2009). A longitudinal study of self-esteem, cultural identity, and academic success among American Indian adolescents. Cultural Diversity & Ethnic Minority Psychology, 15, 38-50.

98.

Wickrama

K. A. S.

Noh

Elder

G. H.

(2009). An investigation of family SES-based inequalities in depressive symptoms from early adolescence to emerging adulthood. Advance in Life Course Research, 14, 147-161.

99.

Wiesner

Kim

H. K.

Capaldi

D. M.

(2010). History of juvenile arrests and vocational career outcomes for at-risk young men. Journal of Research in Crime & Delinquency, 47, 91-117.

100.

C.-I.

Huang

Y.-T.

(2010). 青少年初期身心健康變化及其動態影響變因之討論 [The impact of dynamic factors on trajectories of early adolescent depressive symptoms]. Formosa Journal of Mental Health, 23, 535-562.

101.

M.-Y.

Jou

Y.-H.

(2009). 台灣青少年的道德信念：文社會依附的影響 [Moral beliefs of adolescents in Taiwan: Impacts of social attachment]. Taiwanese Sociology, 17, 61-100.

102.

Yang

K.-S.

(2005). Social- and individual-oriented self-actualizers: Conceptual analysis and empirical assessment of their psychological characteristics. Indigenous Psychological Research in Chinese Societies, 23, 71-143.

103.

Yeung

W. J.

Conley

(2008). Black-White achievement gap and family wealth. Child Development, 79, 303-324. doi:10.1111/j.1467-8624.2007.01127.x

104.

C.-C.

Fan

G.-H.

Chang

M.-Y.

(2013). The developmental outcome of Taiwanese youth: Effects of educational tracking during adolescence. In Yi

C. C.

(Ed.), The psychological well-being of East Asian youth (pp. 157-184). Dordrecht, The Netherlands: Springer.

105.

C.-C.

C.-L.

(2004). Taiwan. In Slater

J. J.

(Ed.), Teen life in Asia (pp. 223-241). Westport, CT: Greenwood Press.

106.

C.-C.

C.-I.

Chang

Y.-H.

Chang

M.-Y.

(2009). The psychological well-being of Taiwanese youth: School versus family context from early to late adolescence. International Sociology, 24, 397-429.

107.

S.-H.

Chen

S.-H.

Chang

K.-J.

(2008). 乳癌患者之反芻反映風格對齊術後憂鬱與創傷後壓力症狀之預測 [Rumination predicting depression and PTSD symptoms in postoperative breast cancer patients]. Chinese Journal of Psychology, 50, 289-302.