HHS Public Access Author manuscript Author Manuscript

J Marriage Fam. Author manuscript; available in PMC 2017 August 01. Published in final edited form as: J Marriage Fam. 2016 August ; 78(4): 1114–1130. doi:10.1111/jomf.12322.

Factors Associated With Perceptions of Family Belonging Among Adolescents Valarie King and Lisa M. Boyd Department of Sociology, Pennsylvania State University, 211 Oswald Tower, University Park, PA 16802 ([email protected])

Abstract Author Manuscript

Adolescents’ perceptions of family belonging are associated with several well-being indicators, yet we know little about which factors influence these perceptions or how they differ by family structure. The current study uses nationally representative data from Add Health to examine predictors of adolescents’ perceptions of family belonging in two-biological-parent families (n = 9686). The results are compared to a recent study using Add Health that examined family belonging in married mother–stepfather families. Findings suggest both similarities and differences across family structure in the factors associated with family belonging.

Keywords Adolescence; biological parents; family belonging; family relations; parent–adolescent relations; stepfamilies

Author Manuscript

Adolescence is a critical point in the life course for accomplishing key developmental tasks. As children grow into adolescence, they desire greater autonomy and spend increasing amounts of time with peers (Furstenberg, 2000). Parent–child conflict tends to increase and engagement in shared activities declines (Smetana, Campione-Barr, & Metzger, 2006). When these relationships remain emotionally close, however, parents continue to be valuable resources for their children. In addition, a positive home environment and feelings of family belonging can promote positive adolescent development (Cavanagh, 2008; King, Boyd, & Pragg, 2016). Adolescents still need their families as foundations from which to move out into the world and gain independence and autonomy (Chubb & Fertman, 1992). Indeed, adolescents’ emotional separation from the family has been found to be associated with negative outcomes, including deviance and substance use (Baer & Schmitz, 2007).

Author Manuscript

Family belonging encompasses feelings of inclusion within one’s family, including feelings of being understood, of having fun together, and of being paid attention to (Goodenow, 1992; Leake, 2007). Feelings of belonging are conceptually distinct from the quality of an individual’s relationships with each family member (King, Boyd, & Thorsen, 2015), although the quality of these relationships likely influences perceptions of family belonging (e.g., Broderick, 1993). Empirical evidence supports this distinction, indicating that parent– child relationships and the extent to which children and adolescents feel they belong to their families are statistically independent predictors of well-being (e.g., Cavanagh, 2008). Despite evidence that perceptions of family belonging contribute to child well-being above

King and Boyd

Page 2

Author Manuscript

and beyond the quality of parent–child relationships, and although a large literature aims to identify predictors of positive relationships between parents and children, few studies have examined which factors contribute to children’s perceptions of belonging in families. To address this gap in the literature, the current study uses nationally representative data from the National Longitudinal Study of Adolescent to Adult Health (Add Health) to examine the correlates of adolescents’ perceptions of family belonging in two-biologicalparent families. Guided by family systems theory, particular attention is paid here to how perceptions of family belonging are related to the quality of ties between family members. We compare our results to a recent study that examined the correlates of family belonging in married mother–stepfather families and also employed data from the Add Health study (King et al., 2015), with an eye toward identifying similarities and differences in the correlates of family belonging across family structures.

Author Manuscript

Background Individuals have a basic psychological need to feel they belong to a social group (Baumeister & Leary, 1995; Maslow, 1954/1970). Family members, and parents in particular, can help meet this need in children by providing love and affection. A positive home environment in which children feel understood and supported, and where family members share enjoyable experiences together, can also help children feel that they belong to the larger family group. Several studies suggest that family belonging is a protective factor against a range of negative adolescent outcomes, including emotional distress, delinquency, violence, suicidal thoughts and behaviors, substance use, early sexual debut, and negative academic behaviors (Cavanagh, 2008; Crosnoe & Elder, 2004; King et al., 2016; Resnick et al., 1997).

Author Manuscript Author Manuscript

Measurement of family belonging in the extant literature varies somewhat across studies and is sometimes referred to as family connectedness (e.g., Brown & Manning, 2009; Cavanagh, 2008) or positive family environment (e.g., Amato & Kane, 2011). Although often sharing similar types of items, two different approaches to measurement of this construct are evident. One approach (e.g., Cavanagh, 2008, King et al., 2015; Leake, 2007) is to measure family belonging as a construct that is distinct from parent–child relationship quality (the latter often indicated by items capturing parent–child closeness and/or parental involvement with children). The second approach is to combine these dimensions of family relationships into a single scale (e.g., parent–family connectedness; Borowsky, Ireland, & Resnick, 2001; Jacobson & Rowe, 1999; Resnick et al., 1997). Both approaches yield measures that demonstrate adequate internal consistency reliability and have been employed by researchers using Add Health data. We take the first approach in this study, however, because it is consistent with our theoretical framework, which distinguishes family relationships from characteristics of the family unit as a whole. Empirical support for this approach is provided by studies indicating that measuring family relationships and perceptions of family belonging as distinct constructs is preferable to combining them (King et al., 2015; Leake, 2005). Evidence also suggests that feelings of family belonging and parent–child relationship quality are independently associated with child well-being (Cavanagh, 2008; Leake, 2005).

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 3

Author Manuscript

Despite this, little research has sought to identify factors that predict adolescents’ perceptions of family belonging, and to our knowledge, no existing studies have sought to identify predictors of perceived family belonging among children or adolescents in twobiological-parent families using a nationally representative sample. This group is important to examine as it comprises almost 60% of children in the United States (U.S. Census Bureau, 2013). Because a strong sense of belonging to one’s family appears to benefit children, it is important to understand where a sense of belonging comes from and how its development can be encouraged.

Author Manuscript

Two studies have examined factors related to adolescents’ feelings of family belonging in stepfamilies, one using a sample of 60 students in a small Midwestern city (Leake, 2007), and the other using data from Add Health (King et al., 2015). Both studies found that the perceived quality of the stepparent–adolescent relationship and, especially, the perceived quality of the mother–adolescent relationship were the factors most strongly associated with adolescents’ feelings of family belonging. The quality of the mother and stepfather’s marital relationship was not directly associated with adolescent perceptions of family belong (King et al., 2015).

Author Manuscript

Our approach to identifying predictors of family belonging is informed by family systems theory (FST), a framework that views families as comprising multiple interdependent subsystems that mutually influence one another and in turn determine family functioning (e.g., Becvar & Becvar, 1999). FST treats the family as a nonsummative system with properties beyond those of its constituent interpersonal relationships—that is, family-level characteristics are not the simple sum of within-family relationship characteristics (Broderick, 1993; Hill & Hansen, 1960). The interconnected nature of family system subunits means that changes or problems in one family subsystem—with subsystems referring to either individuals or relationships between family members—affect all other subsystems (e.g., Broderick, 1993).

Author Manuscript

Family systems theory suggests that an adolescent’s perception of family belonging will be influenced by the quality of the relationships that exist between family members, and the conceptual model we test reflects this supposition (see Figure 1). Although all family relationships have the potential to contribute to a child’s sense of belonging, relationships with—and between—parents are likely to be key. In two-biological-parent families in particular, a close relationship between parents serves as a foundation for positive relationships among other family members, especially between parents and their children (Hetherington & Clingempeel, 1992). Mother–child and father–child relationship quality are defined in this study as adolescents’ perceptions of the affective qualities of the parent–child relationship, which include closeness, warmth, and satisfaction with each relationship. We test two main hypotheses. First, we hypothesize that the quality of the mother–child relationship and the father–child relationship are the two primary predictors of family belonging among adolescents in two-biological-parent families. Less clear is whether one of the parent–child relationships will be more strongly associated with perceptions of family belonging than the other. Research on stepfamilies has found that the quality of the mother– child relationship has a stronger association with adolescents’ perceptions of family

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 4

Author Manuscript

belonging than does the quality of the stepfather–child relationship (King et al., 2015; Leake, 2007). It is unknown, however, whether the stronger influence of the mother–child relationship in stepfather families is a result of the fact that this relationship is with (a) a biological parent, (b) the mother rather than the father, or (c) the primary caretaker or attachment figure, usually the mother in mother–stepfather families. Biological fathers who reside with their children are both more likely to be involved with their children than stepfathers are with stepchildren (Cooksey & Fondell, 1996; Hofferth & Anderson, 2003) and have, on average, participated in the child’s life longer than a stepfather has. Children are also more likely to include a biological father in their definition of family than stepchildren are to include their stepfather (Furstenberg & Nord, 1985). Thus, stepfathers may exert less influence on adolescents’ perceptions of family belonging than biological fathers because the adolescent does not consider them part of the family.

Author Manuscript

The second hypothesis we test is that a positive marital relationship between the mother and father will be associated with feelings of family belonging both directly and indirectly through close parent–child relationships—this is known as the spillover hypothesis (Easterbrooks & Emde, 1988; Engfer, 1988) and is consistent with FST. A poor-quality or conflictual relationship between parents may “spill over” into other areas of family life, creating a tense family environment that detracts from family cohesiveness and negatively affects interactions with children. Alternatively, parents in happy and supportive marriages may be more available to respond to children’s needs and be more attuned to supporting each other’s relationships with children. Several studies have revealed that high-quality marital relationships are positively associated with close and caring parent–child relationships (see Erel & Burman, 1995, for a meta-analysis).

Author Manuscript

Several studies also suggest that the quality of the marital relationship has greater consequences for the quality of the father–child relationship than for the quality of the mother–child relationship (e.g., Amato, 1986; Werneck, Eder, Yanagida, & Rollett, 2014). This finding is consistent with the father vulnerability hypothesis, which postulates that the quality of a father’s parenting is more likely to suffer as a result of family-related stressors than is the quality of a mother’s parenting (Cummings, Goeke-Morey & Raymond, 2004).

Author Manuscript

We also consider the influence of a number of other family characteristics ( “background variables” in Figure 1) that are more distal determinants of family belonging. These factors may be associated with feelings of family belonging either directly or indirectly through the proximate family relationship variables. In general, prior research suggests that higher income, more parental education, the presence of fewer siblings, and the absence of nontraditional siblings (e.g., half-siblings) in the household predict more positive family relationships (Amato, 1998; Blake, 1985, 1989; Halpern-Meekin & Tach, 2008). The quality of sibling relationships may also contribute to feelings of family belonging. The extensive contact and close interaction typical of these relationships provides opportunities for siblings to influence one another both positively and negatively, and in ways that can influence larger family dynamics (McHale, Updegraff, & Whiteman, 2012). Unfortunately, it is not possible to take into account the quality of adolescents’ relationships with their siblings in the current study, given limitations of the Add Health data.

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 5

Author Manuscript

As for child characteristics, adolescent boys, younger adolescents (compared to older adolescents), and religious adolescents often report feeling closer to their parents (and for fathers, especially boys) (King, 2010; King, Thorsen, & Amato, 2014). Race/ethnicity and immigrant status are often associated with family relationships and family processes, sometimes in complex ways (Hardway & Fuligni, 2006). Perhaps most relevant to the current study, some research suggests that individual feelings of family belonging (King et al., 2015), as well as support between family members (including family cohesion and familism), may be higher for immigrant families and/or Hispanics than for non-Hispanic Whites (Baer & Schmitz, 2007; Miranda, Estrada, & Firpo-Jimenez, 2000).

Author Manuscript

Prior research has found that children living in two-biological-parent families report higher levels of family belonging than children living in stepfamilies (Brown & Manning, 2009; Cavanagh, 2008), but this research has little, if anything, to say about whether the factors that promote family belonging are similar across family structures. Some factors, such as the quality of an adolescent’s relationship with his or her mother, are likely similarly important in contributing to feelings of family belonging in two-biological-parent families and stepfamilies, but the role of other factors is less clear. For example, is the quality of an adolescent’s relationship with his or her biological father in two-biological-parent families more strongly related to feelings of family belonging than the relationship between adolescents and stepfathers in stepfamilies? Is a close marital relationship less beneficial for adolescents’ perceptions of family belonging in stepfamilies because the stepfather may be competing with the child for the mother’s time and attention? In addition to the possibility that some factors may differentially contribute to feelings of family belonging across family structures, other factors may be relevant for only certain types of families. For example, in stepfamilies, the length of time the stepfather has been in the household, the age of the child when he entered, the presence of stepsiblings in the household, the number of prior father figures the child has been exposed to, and the relationship between the child and the nonresident biological father may be relevant factors to consider in assessing feelings of family belonging, but they play little or no role in understanding family belonging for children living with both biological parents.

Author Manuscript

The current study parallels an earlier study of Add Health adolescents (n = 2,085) who were living in married mother–stepfather families (King et al., 2015). Where relevant, we note similarities and differences in key findings between these two studies. We also conduct a supplementary analysis that compares adolescents living in married mother–father stepfamilies (the same 2,085 adolescents studied by King et al., 2015) to those living with two biological parents in order to further our understanding of the correlates of family belonging across family structure.

Author Manuscript

Method We used data from the National Longitudinal Study of Adolescent to Adult Health (Add Health). The full Wave 1 sample includes 20,745 adolescents in Grades 7–12 during the 1994–1995 school year and is nationally representative with the use of appropriate sample weights. Parent data (n = 17,670) were collected from one parent of each respondent, usually the resident mother (see Harris et al., 2009, for a detailed description of the data). The

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 6

Author Manuscript

analytic sample was restricted to adolescents with valid sample weights who reported residing with their biological mother and father in the Wave 1 interview (n = 9,686).

Author Manuscript

We also conducted two supplementary analyses. The first of these was a lagged dependent variable analysis using the subset of adolescent respondents in two-biological-parent families who were interviewed on a second occasion, approximately one year later in 1996. This allowed us to test a model using Wave 1 measures to predict levels of family belonging at Wave 2, thereby providing a robustness check on the validity of the cross-sectional findings. This approach better captures the correct temporal ordering between the independent and dependent variables, which is assumed only in the cross-sectional analysis. A drawback of the lagged dependent variable analysis is that it is based on a smaller, less representative sample. The Wave 2 sample design purposively excluded adolescents who were in 12th grade at Wave 1. This exclusion reduced the number of adolescents in our sample to 8,271. An additional 1,297 adolescents living with both biological parents at Wave 1 did not provide interviews in 1996. Finally, we excluded 238 adolescents who completed the Wave 1 and Wave 2 interviews but were no longer living at home with both biological parents at Wave 2. These restrictions reduced the final longitudinal sample to n = 6,736. The second supplementary analysis compared two-biological-parent families and stepfamilies in terms of associations between family relationships and belonging. The stepfamily sample for this analysis comprised adolescents with valid sample weights residing with their biological mothers and married stepfathers in Wave 1 (n = 2,085).

Author Manuscript

We relied on structural equation modeling (SEM) techniques to conduct the analysis. SEM is a particularly appropriate approach for our study given the multiple structural pathways proposed and the modeling of key constructs as latent variables. Analyses were conducted in Mplus version 6 (Muthén & Muthén, 2010). Mplus uses full-information maximum likelihood (FIML) to deal with missing data. Results are based on weighted data, and all analyses take into account the Add Health sample design (i.e., clustering and stratification). Measures

Author Manuscript

Table 1 provides descriptive information for the main study measures. The dependent variable family belonging was measured with four observed indicators from the adolescent interview (each with five response options, 1 = very little to 5 = very much): “How much do you feel your family understands you?,” “How much do you feel you want to leave home?” (reverse coded), “How much do you feel you and your family have fun together?,” and “To what extent do you feel your family pays attention to you?” The items had good internal consistency reliability as a scale, with Cronbach’s alpha (α) = 0.75 (unless otherwise noted, all variables were created using reports from the Wave 1 adolescent in-home interview). Although adolescents on average reported fairly high scores on the four indicators of family belonging, moderate variation existed in this measure. For example, approximately 10% of adolescents disagreed either quite a bit or very much with the statements that their families understand them and that they have fun together, whereas about 13% reported that they would like quite a bit or very much to leave home. Five percent of respondents felt “very little or not at all” that their families pay attention to them, and another 20% agreed only

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 7

Author Manuscript

“somewhat” that their families pay attention to them. The longitudinal analysis used identical indicators of family belonging taken from the Wave 2 adolescent interview (α = 0.76).

Author Manuscript

The latent construct mother–child relationship was measured with five items reflecting several dimensions of mother–child closeness. Items were coded so that higher values corresponded to closer relationships; response options were 1 (not at all/strongly disagree) to 5 (very much/strongly agree). The five items asked respondents how close they feel to their mother, how much they feel she cares about them, how warm and loving they feel she is toward them, how satisfied they are with their communication with her, and how satisfied they are with their overall relationship with her. The items had good internal consistency when taken as a scale (α = 0.83). The items for the father–child relationship latent construct were identical to those for the mother–child relationship. The overall scale again had good internal consistency (α = 0.88). The latent construct mother–father relationship was measured with three observed indicators from the parent interview, using responses that were reported by biological mothers: current relationship happiness (measured on a 1–10 scale), whether the mother and father had talked about separation in the previous year (1 = no, have not talked about separating, 0 = yes, have talked about separating) and how infrequently the mother and father fight (1 = fight a lot, 4 = not at all). Each variable was coded such that a higher score indicated better relationship quality.

Author Manuscript

The adolescent’s gender was a dichotomous variable (1 = female, 0 = male). The adolescent’s age was measured in years. Race/ethnicity was measured as a set of dummy variables for the mutually exclusive categories “non-Hispanic White” (reference group), “non-Hispanic Black,” “Hispanic,” “non-Hispanic Asian,” and “other race.” Immigrant status was a set of three dummy variables representing the following statuses: (a) the adolescent respondent and both parents were born in the United States (or the respondent was born outside the United States but was a U.S. citizen); (b) the adolescent was born in the United States but at least one parent was not born in the United States; and (c) the respondent was not born in the United States. The adolescent’s religiosity was measured with a three-item scale that combined measures of church attendance, perceived importance of religion, and participation in church activities (on a 4-point scale).

Author Manuscript

Income was reported in the parent interview and was transformed using the equation ln(income + 1). The mother’s educational attainment was reported by mothers (or taken from the adolescent interview when mother reports were missing) and was dichotomized so that 1 indicated that the mother had a bachelor’s degree or higher. A continuous variable captured the total number of siblings present in the household. The presence of half-siblings in the household was reflected in a dummy variable coded 1 if the respondent resided with one or more half-siblings. Correspondence With a Prior Study of Stepfather Families Where possible, we compare our results to those from a recent study of family belonging in a different family type, stepfather stepfamilies (King et al., 2015). Although both studies

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 8

Author Manuscript Author Manuscript

employ Wave 1 data from Add Health and include many of the same measures, they are not identical, and conclusions based on this comparison are necessarily only suggestive. Two key differences in the structural model between the two studies are important to note. First, the structural model predicting family belonging in stepfamilies included several additional predictors that were not relevant (or much less relevant) to the current study of twobiological-parent families (i.e., years in a stepfamily, number of stepsiblings). The stepfamily model also did not look at indirect effects of the mother–stepfather relationship through the parent–child relationships. Second, the latent constructs for the mother–child and the father–child relationships were measured differently in the two studies. In the stepfamily study, the mother–child and stepfather–child relationships were measured with three scales: closeness (same five items as in the current study), shared activities (five items), and communication (three items). Confirmatory factor analysis indicated that the measurement model provided a good fit for the latent relationship variables in the stepfamily study. In preliminary analyses for the current study, however, confirmatory factor analysis indicated that including activities and communication along with closeness did not provide a good fit for the latent parent–child relationship variables, which suggests that the dimensions of parent–child relationships that influence family belonging are different depending on family type. As a result, the latent variables for the mother–child and father–child relationship in the current study only reflect parent–child closeness or affective qualities of these relationships. This latter difference underlines the importance of considering that apparently “similar” measures may not be the same across different family structures despite being composed of identical questions, and it points to fundamental differences in family dynamics between the family structures.

Author Manuscript Author Manuscript

We also conducted a new, simplified analysis that applies identical structural equation models to the two-biological-parent and stepfather-family samples. Although, as already discussed, the same constructs do not necessarily measure the same aspect of family functioning in different family types, a parallel analysis is the best way to draw tentative comparisons. The two-biological-parent and stepfamily models include the three family relationship constructs introduced earlier (mother–child relationship, father–child relationship, and mother–(step)father relationship) measured with the same items in both samples. The mother–child and father–child relationship latent variables are each measured with five items reflecting closeness and do not include the additional items on shared activities and communication that King et al. (2015) included as indicators of the stepfather– child relationship. In addition, this parallel analysis includes fewer control variables than our main model predicting family belonging in two-biological-parent families. This model is therefore pared down from both the main analysis in this article and the main analysis in King et al. (2015) and represents a compromise adopted for the sake of model fit and the ability to compare between samples. All variables, latent and manifest, are measured identically. The model fits the data adequately for both samples (two-biological-parent sample: root mean square error of approximation, RMSEA =.03, and comparative fit index, CFI = .95; stepfamily sample: RMSEA = .03, CFI = .95).

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 9

Author Manuscript

Results Two-Biological-Parent Families

Author Manuscript

Measurement model—A confirmatory factor analysis indicated that the measurement model provided a good fit for the latent relationship variables (see Figure 2). Examination of modification indices revealed that the fit of the final measurement model could be improved by including correlations between the residuals of several observed variables. All the suggested correlations were between indicators of the same latent construct or indicators of the mother–child and father–child relationship, for example, between closeness to mother and closeness to father. After adding these correlations to the model, the root mean square error of approximation (.02) and comparative fit index (.99) indicated good fit. Standardized loadings for the observed indicators of the relationship and belonging variables and correlations between latent variables were high (.438–.835). The mother–child and father– child relationships were highly correlated with family belonging (.708 and .674, respectively) and were also fairly highly correlated with each other (.540). Structural model—Consistent with the conceptual model (Figure 1), the structural model posits that interpersonal within-family relationships—namely mother–child, father–child, and mother–father relationships—directly affect the extent to which adolescents feel they belong to their families. We hypothesized that the mother–father relationship would be associated with family belonging both directly and indirectly by way of the mother–child and father–child relationships. Additional background variables were also hypothesized to have direct effects on family belonging, as well as indirect effects through the family relationships. According to well-accepted fit indices, the model as proposed fits the data well: RMSEA = 0.03, CFI = 0.95.

Author Manuscript Author Manuscript

Standardized regression coefficients for the structural model (see Table 2) indicate that two of the three family relationships do directly predict adolescents’ perceptions of family belonging when selected characteristics of the child and family are controlled for. The mother–child relationship is most strongly associated with family belonging, with a standardized coefficient of 0.475 (p < .001). This coefficient tells us that a one-standarddeviation increase in mother–child relationship quality predicts an increase in belonging of almost half of a standard deviation, a fairly large effect size. The father–child relationship is also a significant predictor of perceived family belonging (b = 0.401, p < .001). An increase of one standard deviation in father–child relationship quality is associated with an increase in belonging of two-fifths of a standard deviation. The quality of the mother–father relationship is indirectly, rather than directly, associated with family belonging (discussed later). Girls reported higher belonging (b = 0.079, p < .01) than boys, while older adolescents reported lower perceived belonging than younger adolescents (b = –0.072, p < .001). These effects were small, with girls reporting on average a increase in standard deviation of .08 in belonging over boys, and every year of age corresponding to a decrease of .07 of a standard deviation in belonging. Asian adolescents reported lower perceived belonging than White adolescents (b = –0.150, p < .05). Adolescents who were not born in the United States reported significantly stronger feelings of belonging than those who, along with their J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 10

Author Manuscript

parents, were born in the United States (b = 0.257, p < .01). Religious adolescents reported higher perceived family belonging (b = 0.04, p < .01) than their less religious peers.

Author Manuscript

Indirect effects—Several variables were associated with family belonging indirectly by way of the mother–child or father–child relationship. Indirect effects were calculated using the Sobel (1982) test for mediation of significant pathways (results not shown; all indirect effects were calculated using fully or partially standardized coefficients, consistent with Table 2). As our conceptual model illustrates (Figure 1), we hypothesized that the mother– father relationship would be indirectly associated with family belonging via the mother– child and father–child relationships. We found this to be the case. The indirect effect of the mother–father relationship on family belonging by way of the mother–child relationship (indirect effect = 0.06, p < .001) reflects the fact that stronger marital relationships tend to predict closer relationships between mothers and their adolescent children, which in turn are associated with higher levels of perceived family belonging. A similar indirect effect was found for the mother–father relationship by way of the father–child relationship (indirect effect = 0.11, p < .001).

Author Manuscript

Gender, age, race, religiosity, and number of siblings were also indirectly associated with reports of family belonging. Results from the Sobel test revealed an indirect negative effect of gender such that females reported less close relationships with both mothers (indirect effect = –0.12, p < .001) and fathers (indirect effect = –0.09, p < .001), which in turn were associated with lower levels of belonging. As noted earlier, however, girls reported higher levels of family belonging than boys. Together these findings suggest that the tendency for girls to feel less close to mothers and fathers decreased feelings of family belonging. With the quality of parent–child relationships controlled, however, girls reported greater feelings of family belonging. Respondent’s age was shown to indirectly predict family belonging through both the mother–child relationship (indirect effect = –0.08, p < .001) and the father–child relationship (indirect effect = –0.07, p < .001) as well, reflecting the fact that older adolescents tend to have less close relationships with parents, which in turn are associated with lower levels of belonging. Relative to White adolescents, Black adolescents reported less close relationships with fathers only (indirect effect = –0.01, p < .05), which in turn are associated with lower levels of belonging.

Author Manuscript

We found a significant, indirect effect of religiosity on family belonging via both of the parent–child relationships (indirect effect through the mother–child relationship = 0.06, p < . 001; indirect effect through the father–child relationship = 0.05, p < .001), as religious adolescents reported closer relationships with parents. The number of siblings in the household also had an indirect effect on belonging through the mother–child (indirect effect = –0.02, p < .001) and father–child (indirect effect = –0.01, p < .05) relationships. The more siblings there are in the household, the less close the adolescents’ relationships with both the mother and father tend to be, which in turn is associated with lower perceived belonging. All of these indirect effect sizes, however, were

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 11

Author Manuscript

very small in magnitude, with none even approaching one-tenth of a standard deviation for a one-unit increase in the independent variable. Lagged dependent variable analysis—We replicated the structural model in Table 2, but substituted Wave 2 indicators of family belonging for Wave 1 family belonging. Results revealed relationships between the Wave 1 mother–father, mother–child, and father–child latent variables and Wave 2 belonging very similar to those in the main analysis that used Wave 1 measures. The parent–child relationships were directly and significantly associated with Wave 2 belonging (mother–child standardized b = 0.309, p < .001; father–child standardized b = 0.353, p < .001), while the mother–father relationship had smaller but still significant indirect effects on Wave 2 belonging through the mother–child and father–child relationships (results available on request).

Author Manuscript

Comparing Two-Biological-Parent and Stepfather Families

Author Manuscript

To directly compare two-biological-parent families and stepfather families, we combined the two samples and performed a test of factorial invariance to determine whether the measurement of the three latent family relationship variables was comparable across family structures. We compared results from unconstrained and constrained models using several model fit indices. Two of these indices, the Satorra-Bentler chi-square test and the loglikelihood ratio test, indicated that the father–child relationship construct should be measured differently in the two family types. This suggests that factor loadings for the father–child relationship construct vary between samples, or that the indicators that best capture father–child closeness may differ between two-biological-parent families and stepfather stepfamilies. Although chi-square tests are sample-size dependent, the factor structure variance for the father–child relationship construct was sufficient to produce significant chi-square results when we tested measurement invariance for the model as a whole, even though the other two latent constructs appeared to be identical across samples. This finding supports the observation from our comparison of the current study to a previous study of stepfamily belonging (King et al., 2015) that family relationships cannot be assumed to work the same way in different family types.

Author Manuscript

We proceeded to run separate but identical structural equation models for the two-biologicalparent and stepfather family samples. Results from these models (reported in the Appendix) reveal that the most important predictors of belonging in both family types were the mother– child and father–child relationships (all coefficients are fully standardized, as in Table 2). In the two-biological-parent sample, the mother–child relationship coefficient predicting belonging was .476 (p < .001) and the father–child relationship coefficient was .401 (p < . 001). In the stepfamily sample, the corresponding coefficients were .475 (p < .001) and .370 (p < .001). The mother–father relationship was associated with belonging indirectly through both the mother–child and father–child relationships in stepfather families, as in twobiological-parent families, although the mother–father relationship was more weakly associated with the mother–child relationship in particular when compared to the twobiological-parent sample (two-biological-parent sample: mother–child on mother–father relationship, b = 0.129, p < .001, and father–child on mother–father relationship, b = 0.262, p < .001; stepfamily sample: mother–child on mother–father relationship, b = 0.099, p < .01,

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 12

Author Manuscript

and father–child on mother–father relationship, b = 0.330, p < .001). The relatively larger statistical effect of the mother–father relationship on the father–child relationship in the stepfamily sample supports the idea that the father–child relationship fills a different role in stepfamilies than in two-biological-parent families. This conclusion is also supported by the weaker effect of the stepfather–child relationship on family belonging (relative to the mother–child relationship) in stepfather families compared to the more similar mother–child and father–child relationship coefficients in two-biological-parent families.

Discussion

Author Manuscript

Despite a growing body of research documenting a positive association between perceptions of family belonging and child well-being, surprisingly little research exists on which factors influence these perceptions of family belonging, nor do we know how these factors might differ for children living in different family structures. The current study helps address this gap in the literature by using nationally representative data to examine the correlates of adolescents’ perceptions of family belonging in two-biological-parent families. Consistent with a family systems perspective and our first hypothesis, the quality of the mother–child relationship and the father–child relationship were the two primary predictors of family belonging among adolescents in two-biological-parent families. Notably, adolescent relationships with each parent strongly and independently contributed to feelings of family belonging. Thus, mothers and fathers appear to play similarly important roles in helping their adolescents feel they belong.

Author Manuscript Author Manuscript

A previous study of family belonging in stepfamilies (King et al., 2015) serves as a useful comparison on this point. In this previous study, the mother–child and stepfather–child relationship were the strongest predictors of family belonging as well. However, in the present study, the mother–child and father–child relationships appeared to contribute roughly equally to family belonging, whereas in the other study, the mother–child relationship was a better predictor than was the stepfather–child relationship. This same pattern existed in our own supplementary analysis comparing two-biological-parent families and married stepfather families. This difference could result from several factors associated with two-biological-parent families that differ in stepfather families, such as reflecting adolescent ties to a biological father, the biological father’s greater involvement with the adolescent during the child’s entire life, or the biological father’s greater likelihood of being included in the adolescent’s definition of a family member. In these respects, residential biological fathers are similar to residential biological mothers, which perhaps results in both parents having a similar effect on adolescents’ perceptions of family belonging. The larger role of mothers in predicting family belonging in stepfamilies is also consistent with the notion that mothers play a pivotal role in successful stepfamily functioning (Smith, 2008). Our second hypothesis was partially supported. Although the quality of the mother–father relationship was not directly associated with adolescent perceptions of family belonging, it was indirectly associated with family belonging through the mother–child and father–child relationships. That is, a positive marital relationship was significantly associated with adolescent reports of higher-quality relationships with their mothers and, even more

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 13

Author Manuscript

strongly, with their fathers, which in turn was associated with higher levels of family belonging. This finding supported the spillover hypothesis and, more expressly, the vulnerability hypothesis, according to which father–child relationships are more susceptible than mother–child relationships to outside influence—in this case that of the marital relationship.

Author Manuscript

The stepfamily study (King et al., 2015) also found no direct association between marital quality and adolescent perceptions of family belonging, but it did not examine indirect effects through the mother–child or stepfather–child relationships. A reanalysis by those authors that specifically examined these indirect effects (personal communication; results available upon request) revealed a significant indirect effect of marital quality on family belonging via the stepfather–child relationship, again of a somewhat larger magnitude than that through the mother–child relationship. This same pattern was echoed in our own supplementary analysis of stepfather families. These findings suggest the need to consider the indirect effects of marital quality when examining parent–child relationships or family dynamics more broadly, and they provide additional evidence for the notion that marital quality has a stronger association with the quality of the father–child relationship than with the mother–child relationship.

Author Manuscript

The significant associations between the quality of family relationships and adolescents’ sense of family belonging found in both this study and the stepfamily study (King et al., 2015) provide support for a basic premise of FST that families comprise interdependent subsystems that influence one another. Families have the potential to nurture (or detract from) a child’s sense of belonging and provide a secure foundation from which to navigate the transition through adolescence and into young adulthood. Although feelings of family belonging tend to be higher among adolescents in two-biological-parent families than in stepfamilies (King et al., 2015), parents in both types of families appear to play a role in helping foster a sense of belonging with the potential to contribute to the well-being of their children (King et al., 2016).

Author Manuscript

As mentioned previously, our tests of the measurement model in the current study indicated that mother–child and father–child relationship latent variables incorporating measures of parent–child closeness, activities, and communication did not provide the best fit for the two-biological-parents sample. However, this tripartite latent variable—rather than a variable that measured only closeness—fit the data well in the stepfather study (King et al., 2015). This discrepancy suggests that, although parent–child relationships are important predictors of family belonging for adolescents in both two-biological-parent families and stepfather families, the structure and nature of these relationships vary by family type. For example, supplementary analyses we conducted (not presented) revealed that communication and activities with stepfathers were associated with closeness, whereas these aspects of the father–child relationship were relatively unimportant for closeness to resident biological fathers. In other words, for this latter group closeness is much more “built in” and less dependent on discrete shared activities. This finding sheds light on an important difference in family dynamics between these two family forms and emphasizes the need for further research on how to best measure and promote positive parent–child relationships for adolescents in different types of families. This conclusion is supported by our new

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 14

Author Manuscript Author Manuscript

supplementary analysis of stepfather stepfamilies, which provided additional evidence that fathers occupy a different role in stepfather stepfamilies than in two-biological-parent families. Not only did our analyses indicate that the father–child relationship is more influenced by the mother–father relationship than the mother–child relationship in stepfamilies than in two-biological-parent families; tests of measurement invariance also indicated that the pathway from the father–child relationship to family belonging is not analogous in the two samples. However, if we were to measure the father–child relationship differently in the two-biological-parent and stepfather family samples, we would introduce new variables to the model and results would not be comparable. The parallel analysis we conducted was a compromise that allowed us to draw tentative conclusions about similarities and differences between two-biological-parent families and stepfather families. The ability to test measurement invariance is a strength of SEM, given that it allows for the identification of incorrectly specified models. It can also point us toward new avenues for research, as in this case, where our results raise the question of how to best model family relationships in different family types, given that they cannot be assumed to be the same for all families.

Author Manuscript

Several child characteristics were directly associated with perceptions of family belonging, although effect sizes were modest. Feelings of belonging were higher among girls, younger adolescents, immigrants, and religious adolescents, and lower among Asian youth compared with Whites. Very modest indirect effects were also found for gender, age, race, religiosity, and number of siblings. One of these findings in particular deserves more attention in future research: Asian adolescents reported lower levels of family belonging than Whites, although they did not significantly differ in their reports of ties to mothers or fathers. We are not aware of prior research examining levels of family belonging among Asian adolescents. A few studies have examined other aspects of family relationships such as levels of conflict and cohesion between adolescents and their parents (Fuligni, 1998) and levels of family identification (Hardway & Fuligni, 2006), but findings have tended to suggest more similarities between White and Asian adolescents on these dimensions than differences. Future research would benefit from more attention to the mechanisms and family processes underlying Asian adolescents’ lower levels of perceived family belonging.

Author Manuscript

As with any study, our study has some limitations. Our findings are based on data collected from adolescents in the mid-1990s. Although it is not clear why the predictors of adolescents’ perceptions of family belonging (especially the importance of positive family relationships) would necessarily be any different today, the rapid pace of family change opens the possibility to cohort differences in many aspects of family life. Although the Add Health data contained detailed information on the parent–child relationship and information from mothers on the marital relationship, detailed information on relationship quality with siblings or other potentially important family members (e.g., grandparents) was not available. These relationships may also play a role in feelings of family belonging. Findings regarding differences between two-biological-parent families and stepfamilies are only suggestive, as the different models and measures employed in this study and the King et al. (2015) study precluded a direct empirical test of significant differences. The supplementary analysis of two-biological-parent families and stepfamilies presented in this article relied on

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 15

Author Manuscript

a reduced comparative model, and even those results indicated that the two models should not be considered equivalent. Overall, findings from the current study suggest both a number of similarities and a few differences between adolescents living with two biological parents and those living with biological mothers and married stepfathers regarding the factors associated with adolescent perceptions of family belonging. Future research should explore whether and how predictors of family belonging are similar or different in other types of families, and it should consider other potentially important factors that could be associated with feelings of family belonging. For adolescents living with both biological parents, it appears that positive relationships with mothers and with fathers are both important in helping children feel they belong.

Author Manuscript

Acknowledgments This research was supported by funding from the National Science Foundation (NSF) to Valarie King, principal investigator (SES-1153189), and by funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) to the Population Research Institute at the Pennsylvania State University for Population Research Infrastructure (R24 HD41025) and Family Demography Training (T-32HD007514). This research uses data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by grant P01-HD31921 from the Eunice Kennedy Shriver NICHD, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due to Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from grant P01HD31921 for this analysis.

References Author Manuscript Author Manuscript

Amato PR. Marital conflict, the parent-child relationship and child self-esteem. Family Relations. 1986; 35:403–410. Amato, PR. More than money? Men’s contributions to their children’s lives. In: Booth, A.; Crouter, AC., editors. Men in families: When do they get involved? What differences does it make?. Mahwah, NJ: Erlbaum; 1998. p. 241-278. Amato PR, Kane JB. Parents’ marital distress, divorce, and remarriage: Links with daughters’ early family formation transitions. Journal of Family Issues. 2011; 32:1073–1103. [PubMed: 21785523] Baer JC, Schmitz MF. Ethnic differences in trajectories of family cohesion for Mexican American and non-Hispanic White adolescents. Journal of Youth and Adolescence. 2007; 36:583–592. Baumeister RF, Leary MR. The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin. 1995; 117:497–529. [PubMed: 7777651] Becvar, DS.; Becvar, RJ. Systems theory and family therapy: A primer. Lanham, MD: University Press of America; 1999. Blake J. Number of siblings and educational mobility. American Sociological Review. 1985; 50:84–94. Blake J. Number of siblings and educational attainment. Science. 1989; 245:32–36. [PubMed: 2740913] Borowsky IW, Ireland M, Resnick MD. Adolescent suicide attempts: Risks and protectors. Pediatrics. 2001; 107:485–493. [PubMed: 11230587] Broderick, CB. Understanding family process: Basics of family systems theory. Thousand Oaks, CA: Sage; 1993. Brown SL, Manning WD. Family boundary ambiguity and the measurement of family structure: The significance of cohabitation. Demography. 2009; 46:85–101. [PubMed: 19348110] Cavanagh SE. Family structure history and adolescent adjustment. Journal of Family Issues. 2008; 29:944–980.

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 16

Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Chubb NH, Fertman CI. Adolescents’ perceptions of belonging in their families. Families in Society— The Journal of Contemporary Human Services. 1992; 73(7):387–394. Cooksey EC, Fondell MM. Spending time with his kids: Effects of family structure on fathers’ and children’s lives. Journal of Marriage and the Family. 1996; 58:693–707. Crosnoe R, Elder GH. Family dynamics, supportive relationships, and educational resilience during adolescence. Journal of Family Issues. 2004; 25:571–602. Cummings, EM.; Goeke-Morey, MC.; Raymond, J. Fathers in family context: Effects of marital quality and marital conflict. In: Lamb, ME., editor. The role of the father in child development. 4th. Hoboken, NJ: Wiley; 2004. p. 196-221. Easterbrooks, MA.; Emde, RN. Marital and parent–child relationships: The role of affect in the family system. In: Hinde, RA.; Stevenson-Hinde, J., editors. Relationships within families: Mutual influence. New York: Oxford University Press; 1988. p. 83-103. Engfer, A. The interrelatedness of marriage and the mother–child relationship. In: Hinde, RA.; Stevenson-Hinde, J., editors. Relationships within families: Mutual influences. Oxford, UK: Oxford University Press; 1988. p. 104-118. Erel O, Burman B. Interrelatedness of marital relations and parent-child relations: A meta-analytic review. Psychological Bulletin. 1995; 118:108–132. [PubMed: 7644602] Fuligni AJ. Authority, autonomy, and parent-adolescent conflict and cohesion: A study of adolescents from Mexican, Chinese, Filipino, and European backgrounds. Developmental Psychology. 1998; 34:782–792. [PubMed: 9681270] Furstenberg FF Jr. Sociology of adolescence and youth in the 1990s: A critical commentary. Journal of Marriage and the Family. 2000; 62:896–910. Furstenberg FF Jr, Nord CW. Parenting apart: Patterns of childrearing after marital disruption. Journal of Marriage and the Family. 1985; 44:893–900. Goodenow C. Strengthening the links between educational psychology and the study of social contexts. Educational Psychologist. 1992; 27:177–196. Halpern-Meekin S, Tach L. Heterogeneity in two-parent families and adolescent well-being. Journal of Marriage and Family. 2008; 70:435–451. Hardway C, Fuligni AJ. Dimensions of family connectedness among adolescents with Mexican, Chinese, and European backgrounds. Developmental Psychology. 2006; 42:1246–1258. [PubMed: 17087556] Harris, KM.; Halpern, CT.; Whitsel, E.; Hussey, J.; Tabor, J.; Entzel, P.; Udry, JR. The National Longitudinal Study of Adolescent to Adult Health: Research design. 2009. Retrieved from http:// www.cpc.unc.edu/projects/addhealth/design Hetherington EM, Clingempeel WG. Coping with marital transitions. Monographs of the Society for Research in Child Development. 1992; 57:2–3. Hill R, Hansen DA. The identification of conceptual frameworks utilized in family study. Marriage and Family Living. 1960; 22:299–311. Hofferth SL, Anderson KG. Are all dads equal? Biology versus marriage as a basis for paternal investment. Journal of Marriage and Family. 2003; 65:213–232. Jacobson KC, Rowe DC. Genetic and environmental influences on relationships between family connectedness, school connectedness, and adolescent depressed mood: Sex differences. Developmental Psychology. 1999; 4:926–939. King, V. The influence of religion on ties between the generations. In: Ellison, CG.; Hummer, RA., editors. Religion, families, and health: Population-based research in the United States. New Brunswick, NJ: Rutgers University Press; 2010. p. 86-105. King, V.; Boyd, LM.; Pragg, B. Parent-adolescent closeness, family belonging, and adolescent wellbeing across family structures; Poster presented at the annual meeting of the Population Association of America; 2016. King V, Boyd LM, Thorsen ML. Adolescents’ perceptions of family belonging in stepfamilies. Journal of Marriage and Family. 2015; 77:761–774. [PubMed: 26166845] King V, Thorsen ML, Amato PR. Factors associated with positive relationships between stepfathers and adolescent stepchildren. Social Science Research. 2014; 47:16–29. [PubMed: 24913942]

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 17

Author Manuscript Author Manuscript

Leake, VS. ProQuest Dissertations and Theses database. (UMI No. 3231196). 2005. Steps between, steps within: Adolescents and stepfamily belonging. Leake VS. Personal, familial, and systemic factors associated with family belonging for stepfamily adolescents. Journal of Divorce & Remarriage. 2007; 47:135–155. Maslow, AH. Motivation and personality. New York, NY: Harper & Row; 1954/1970. McHale SM, Updegraff KA, Whiteman SD. Sibling relationships and influences in childhood and adolescence. Journal of Marriage and Family. 2012; 74:913–930. [PubMed: 24653527] Miranda AO, Estrada D, Firpo-Jimenez M. Differences in family cohesion, adaptability, and environment among Latino families in dissimilar stages of acculturation. Family Journal. 2000; 8:341–350. Muthén, LK.; Muthén, BO. Mplus user’s guide. 6th. Los Angeles, CA: Muthén & Muthén; 2010. Resnick MD, Bearman PS, Blum RW, Bauman KE, Harris KM, Jones J, Udry JR. Protecting adolescents from harm. Journal of the American Medical Association. 1997; 278:823–832. [PubMed: 9293990] Smetana JG, Campione-Barr N, Metzger A. Adolescent development in interpersonal and societal contexts. Annual Review of Psychology. 2006; 57:255–284. Smith, M. Resident mothers in stepfamilies. In: Pryor, J., editor. The international handbook of stepfamilies: Policy and practice in legal, research, and clinical environments. Hoboken, NJ: Wiley; 2008. p. 151-174. Sobel ME. Asymptotic confidence intervals for indirect effects in structural equation models. Sociological Methodology. 1982; 13:290–312. U.S. Census Bureau. Table C9: Children by presence and type of parent(s), race, and Hispanic origin. 2013. Retrieved from https://www.census.gov/hhes/families/data/cps2013C.html Werneck H, Eder MO, Yanagida T, Rollett B. Predicting adolescents’ parent-child relationship quality from parental personality, marital conflict and adolescents’ personality. European Journal of Developmental Psychology. 2014; 11:159–176.

Appendix. Structural Models Predicting Family Belonging: Two-BiologicalAuthor Manuscript

Parent and Stepfamily Samples Two-Biological-Parent Sample (N= 9,686)

Mother–Child Relationship

Father–Child Relationship

−0.001

0.129***

0.262***

0.078**

0.251***

0.221***

−0.045

0.071***

0.171***

0.183***

0.005

Black

0.011

0.009

−0.028*

−0.028

Hispanic

0.132

0.021

−0.033

−0.132

−0.137**

−0.007

−0.015

−0.032

0.252**

−0.090

−0.139

0.297**

0.031

−0.059

−0.088

0.043

0.041**

0.114***

0.119***

0.124***

Family Belonging

Mother–Father Relationship

Family Relationships Mother–child relationship

0.476***

Father–child relationship

0.401***

Mother–father relationship

Child Characteristics Female Age

Author Manuscript

Racea

Other race Immigration statusb R not born in U.S. 1+ parent not born in U.S. Religiosity

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 18

Author Manuscript

Two-Biological-Parent Sample (N= 9,686)

Family Belonging

Mother–Child Relationship

Father–Child Relationship

Mother–Father Relationship

Income (natural log of)

0.000

0.003

0.008

0.056*

Mother college educated

0.008

−0.029 *

0.005

0.004

−0.010

−0.047 ***

−0.037*

−0.023

DF = 271

RMSEA = .025

CFI = .952

Family Belonging

Mother–Child Relationship

Father–Child Relationship

0.029

0.099**

0.330***

−0.012

0.270***

−0.077

0.014

−0.053*

0.147***

0.156***

−0.014

−0.013

0.037

0.050

−0.082*

Family Characteristics

Number of siblings Chi-square = 1,957.991*** Stepfamily Sample (N= 2,085)

Mother–Father Relationship

Family Relationships

Author Manuscript

Mother–child relationship

0.475***

Father–child relationship

0.370***

Mother–father relationship

Child Characteristics Female Age Racea Black Hispanic

0.164

−0.072

−0.162

0.092

−0.131

0.237*

0.007

−0.230

0.231

−0.175

0.080

0.212

−0.003

0.118

−0.042

−0.197

0.061

0.034

0.081*

0.045

Income (natural log of)

−0.044

0.036

−0.002

0.046

Mother college educated

0.027

−0.031

−0.019

−0.092*

−0.051

0.026

0.056

0.002

DF = 280

RMSEA = .026

CFI = .949

Other race Immigration statusb R not born in U.S.

Author Manuscript

1+ parent not born in U.S. Religiosity

Family Characteristics

Number of siblings Chi-square = 684.169***

Author Manuscript

Note. Coefficients for the family relationship variables are fully standardized. Coefficients for all other variables are standardized on the dependent variable only. For all four latent variables, a high score indicates a positive relationship. a Reference group for race is White. b Reference group for immigration status is “The respondent and both parents were born in the U.S.” * p < .05. ** p < .01. *** p < .001.

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 19

Author Manuscript Author Manuscript Author Manuscript

Figure 1.

Conceptual Model.

Author Manuscript J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 20

Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Figure 2.

Measurement Model Note. Chi-square = 461.29***, degrees of freedom = 98, RMSEA = .02, CFI = .987. All coefficients are fully standardized. Correlations between residuals of observed indicators included to improve model fit not shown in figure. Data from Wave 1 of Add Health. *p < .05. **p < .01. ***p < .001.

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 21

Table 1

Author Manuscript

Descriptive Information on Items and Scales Measuring Family Belonging, Family Relationships, and Child and Family Characteristics Cronbach’s Alphaa

M or %

(SD)

Family understanding

3.68

(0.97)

Desire to leave homeb

3.94

(1.19)

Family has fun

3.83

(0.96)

4.00

(0.88)

Family understanding

3.67

(1.00)

Desire to leave homeb

3.88

(1.22)

Family has fun

3.78

(0.99)

4.02

(0.87)

Closeness

4.55

(0.76)

Caring

4.89

(0.40)

Warmth

4.45

(0.73)

Satisfaction with communication

4.10

(0.97)

4.35

(0.82)

Closeness

4.36

(0.89)

Caring

4.81

(0.54)

Warmth

4.21

(0.87)

Satisfaction with communication

4.00

(1.01)

4.18

(0.93)

Relationship happinessc

8.46

(1.60)

Did not talk about separatingc

90.2

Low conflictc

2.81

Variable

Family relationships (latent) Family belonging (Wave 1)

.75

Family attentiveness Family belonging (Wave 2)

.76

Author Manuscript

Family attentiveness Mother–child relationship (Wave 1)

.83

Satisfaction with relationship Father–child relationship (Wave 1)

.88

Author Manuscript

Satisfaction with relationship Mother–father relationship (Wave 1)c

.62

(0.68)

Child characteristics (Wave 1) Female

51.0

Age

15.36

White

73.4

Author Manuscript

Black

8.4

Hispanic

11.7

Asian

4.8

Other race

1.4

Respondent and parents born in U.S.

82.6

At least one parent not born in U.S.

11.6

(1.81)

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 22

Cronbach’s Alphaa

Variable

M or %

Author Manuscript

Respondent not born in U.S. Religiosity

(SD)

5.8 .79

2.69

(0.96)

Income (natural log of)c

3.79

(0.73)

Mother college educatedc

25.4

Number of siblings

1.52

Presence of half-siblings

3.0

Family characteristics (Wave 1)

(1.16)

Note. Table values are based on weighted data and nonmissing cases. a

Cronbach’s alphas are standardized and are intended as approximations—these “scales” are modeled as latent variables.

b

Denotes a reverse-coded item.

c

Denotes a measure from the parent questionnaire (all others are based on adolescent reports).

Author Manuscript Author Manuscript Author Manuscript J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

King and Boyd

Page 23

Table 2

Author Manuscript

Structural Model Predicting Family Belonging in Two-Biological-Parent Families (N = 9,686) Mother–Child Relationship

Father–Child Relationship

−0.001

0.129***

0.262***

0.079**

−0.251***

−0.220***

−0.041

−0.072***

−0.170***

−0.184***

0.001

Black

0.012

0.009

−0.028*

−0.027

Hispanic

0.131

0.011

−0.036

−0.134

−0.150*

−0.108

−0.048

−0.011

−0.102

0.257*

0.068

−0.068

0.257**

−0.051

−0.128

0.293**

0.031

−0.048

−0.085

0.038

0.041**

0.116***

0.119***

0.121***

Income (natural log of)

0.000

0.005

0.008

0.055*

Mother college educated

0.008

−0.027*

0.005

0.002

Number of siblings

−0.010

−0.048***

−0.036*

−0.018

Presence of half-siblings

−0.035

0.019

−0.049

−0.282*

Family Belonging

Mother–Father Relationship

Family Relationships Mother–child relationship

0.475***

Father–child relationship

0.401***

Mother–father relationship

Child Characteristics Female Age Racea

Author Manuscript

Asian Other race Immigration statusb R not born in U.S. 1+ parent not born in U.S. Religiosity

Family Characteristics

Author Manuscript

Chi-square = 2,117.493***

RMSEA = .025

DF = 309

CFI = .949

Note. Coefficients for the family relationship variables are fully standardized. Coefficients for all other variables are standardized on the dependent variable only. For all four latent variables, a high score indicates a positive relationship. a

Reference group for race is White.

b

Reference group for immigration status is “The respondent and both parents were born in the U.S.”

*

p < .05.

Author Manuscript

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

J Marriage Fam. Author manuscript; available in PMC 2017 August 01.

Factors Associated With Perceptions of Family Belonging Among Adolescents.

Adolescents' perceptions of family belonging are associated with several well-being indicators, yet we know little about which factors influence these...
881KB Sizes 0 Downloads 9 Views