Stepfather–Adolescent Relationship Quality During the First Year of Transitioning to a Stepfamily

Abstract

This study contributes to the growing literature on factors associated with the formation of close relationships between stepfathers and stepchildren. The authors extend prior research by using nationally representative data from Add Health (N = 179) to examine how factors existing prior to stepfamily formation are associated with the quality of stepfather–adolescent ties within the first year after married stepfathers join the household. Results from structural equation models revealed that both the quality of the mother–adolescent relationship and adolescent adjustment prior to stepfamily formation were significantly associated with the perceived quality of adolescents’ relationships with their stepfathers.

The general instability of child-bearing unions in the United States, combined with the tendency for parents to form new partnerships following relationship disruption, means that stepfamilies have become a central feature of the American family system. Kreider and Ellis (2011) reported that almost 8% of all U.S. children lived with a stepparent in 2009, and about 30% of U.S. children will live with a stepparent at some time before reaching adulthood (Bumpass, Raley, & Sweet, 1995). The majority of stepfamilies involve residential stepfathers (Stewart, 2007), who were the focus of the current study.

This study builds on recent research on factors that are associated with the formation of close relationships between stepfathers and stepchildren (e.g., Ganong, Coleman, & Jamison, 2011; Jensen & Shafer, 2013; King, 2009; King, Thorsen, & Amato, 2014). In contrast to earlier studies that focused on differences between stepfamilies and two-parent biological families, more recent studies have focused on variation within stepfamilies. Stepfamily-focused research, like the current study, investigates processes unique to stepfamilies (e.g., the presence of stepsiblings and the number of years in a stepfamily) that may produce positive or negative stepfamily outcomes. In this study we drew on Waves I and II of the National Longitudinal Study of Adolescent to Adult Health (Add Health; http://www.cpc.unc.edu/projects/addhealth) and examined adolescents who transitioned from single-mother households to married mother–stepfather households between waves. Our goal was to use variables measured prior to the entry of stepfathers to predict the quality of stepchild–stepfather relationships during the first year after stepfamily formation.

Background and Conceptual Perspective

Understanding adolescents’ relationships with stepfathers in newly formed stepfamilies is a topic of particular importance and the current study captured families when they are particularly vulnerable. Many families exhibit declines in parental attention and increases in parent–child conflict as children enter adolescence (Hetherington & Stanley-Hagan, 2000), placing children at greater risk for poor outcomes (Hetherington & Kelly, 2002). This developmental period can be a particularly difficult time for stepfamily formation and integrating stepfathers into the family, given that adolescents (and preadolescents) are more likely than young children to reject mothers’ new partners (Bray & Easling, 2005; Hetherington & Jodl, 1994). The early phase of stepfamily formation is a critical time involving major changes to the family system because new roles and relationships have to be negotiated (Pryor, 2014). Early tension between adolescents and stepfathers can spill over and disrupt stepfamily functioning more generally (Bray, 1999), yet relatively little is known about the factors that affect whether adolescents accept or reject stepfathers during the first critical year of stepfamily life.

When mothers form new unions, children often benefit from the economic resources provided by stepfathers (Sweeney, 2010). Although remarriage improves the standard of living of most single mothers and their children, children with stepfathers have the same risk of behavioral and emotional problems as children with single mothers (Coleman, Ganong, & Fine, 2000). The absence of positive stepfather effects has led researchers to focus less on the presence of stepfathers and more on the quality of relationships between stepfathers and stepchildren. Children’s relations with stepfathers are not always close, and there is little reason to assume that children benefit when relationships with stepfathers are distant or hostile (Hetherington & Kelly, 2002). For this reason, recent studies have attempted to document the conditions under which children and stepfathers form close relationships—relationships that may benefit children and help compensate for any disadvantages associated with single-parent households. Many of these studies, like the current one, also have focused on stepfamilies formed through marriage because relationships between stepfathers and stepchildren under these conditions are likely to be more enduring and consequential.

Ecological–contextual theory (Doherty, Kouneski, & Erickson, 1998) holds that fatherhood (unlike motherhood) is largely socially constructed. As a consequence, the father–child relationship is more variable in quality than the mother–child relationship and thus more sensitive to a variety of contextual influences. Given the normative ambiguity that surrounds the stepfather role, relationships with stepfathers are even more sensitive to family and interpersonal factors than are relationships with biological fathers (Sweeney, 2010). Consistent with King et al. (2014), we assume that other family relationships are particularly important in understanding children’s relationships with stepfathers. This assumption follows from family systems theory (Minuchin, 1974), which holds that all parts of a family system are interconnected, with one subsystem (e.g., the mother–child dyad) affecting all other subsystems (e.g., the stepfather–child dyad). Like King et al., we view adolescents’ relationships with mothers and nonresident biological fathers prior to stepfamily formation as potential predictors of subsequent relationships with stepfathers.

We assume that child characteristics that predate the entry of the stepfather into the household also affect the quality of subsequent adolescent–stepfather relationships. We focus here on two dimensions of adolescent adjustment that developmental researchers generally consider to be central dimensions of child adjustment (Achenbach & McConaughy, 1997): (a) externalizing problems (delinquency) and (b) internalizing problems (depressive symptoms). Adolescents who engage in delinquent and antisocial activities (e.g., shoplifting, damaging property, and getting into physical fights) may reject the authority of new stepfathers. Similarly, adolescents who experience depressive symptoms (who are focused on their own distress) may find it difficult to form positive emotional attachments to new household members. Moreover, adolescents with behavioral or emotional problems are unlikely to behave in ways that endear themselves to new stepfathers. Previous research has shown that adolescent delinquency and depressive symptoms make it difficult for children and adolescents to establish close relationships with other people, peers as well as adults (Oland & Shaw, 2005), and the same processes are likely to occur with stepfathers.

The conceptual model that guides the current study is shown in Figure 1. In this model, the mother–child relationship, the nonresident father–child relationship, and child adjustment are proximal factors that affect the adolescent–stepfather relationship. The model treats other background factors (e.g., adolescent gender, adolescent age, and the number of previous father figures) as distal variables. Distal variables can affect the adolescent–stepfather relationship both directly and indirectly. Adolescent age, for example, may have a direct effect on the adolescent–stepfather relationship. But if age also affects the adolescent–mother relationship, and if the adolescent–mother relationship affects the adolescent–stepfather relationship, then age also will have an indirect effect. Distinguishing between proximal and distal factors makes it possible to estimate direct and indirect effects in stepfamilies—an approach that reflects the complexity of stepfamily processes.

Figure 1. Conceptual Model.

Note. Correlations between the residuals of all latent variables at Wave 1 are included in the statistical model but are not shown in the figure to improve clarity.

Previous research has supported the importance of some of the variables in the conceptual model. The quality of the child–mother relationship, for example, is positively correlated with the quality of the child–stepfather relationship (King, 2009; King et al., 2014). Research is less clear about whether children’s relationships with nonresident biological fathers have implications for stepfathers, although the number of studies that have examined this topic is small (King, 2006). Some studies have shown associations between children’s relationships with stepfathers and children’s emotional and behavioral problems (e.g., Yuan & Hamilton, 2006). Because these studies tend to be cross-sectional, however, whether problems in children’s adjustment precede or follow the introduction of stepfathers into children’s households remains unclear.

With respect to other variables in the model, several studies have shown that boys tend to report better relationships with stepfathers than do girls (Jensen & Shafer, 2013; King et al., 2014; Pasley & Moorefield, 2004). One study found that older adolescents report weaker ties with stepfathers than do younger adolescents (King et al., 2014). Another study found that stepfather education is positively associated with stepfather involvement (Cooksey & Fondell, 1996). Relatively little is known about the importance of other variables in the model, such as the number of siblings, whether children were born within marriage, and the number of prior father figures in children’s lives. Moreover, studies that have compared the stepfamily relationships of children with different racial backgrounds have yielded mixed results (e.g., Hofferth & Anderson, 2003). Overall, more research is necessary to clarify the factors associated with positive relationships between stepchildren and stepfathers.

In summary, in the present study we focused on adolescents’ descriptions of their relationships with stepfathers (closeness, caring, warmth, positive communication, and general relationship quality). We concentrated on adolescents who lived with single mothers in Wave I of Add Health and transitioned to residential, married stepfather families by Wave II (including stepfamilies in which the parents briefly cohabited before marrying or in which the marriage coincided with coresidence). Our goal was to understand the factors (measured prior to stepfamily formation) that predict close (or weak) ties between adolescents and stepfathers during their first year of living together. The fact that our study was longitudinal, with the independent variables measured prior to stepfamily formation, increases our confidence that we had the correct causal ordering of variables.

Our study builds directly on a previous study conducted by King (2009), which also used the first two waves of Add Health. The main focus of King’s study was on how the entrance of a stepfather was associated with changes in adolescents’ ties to mothers and nonresident fathers, but it also included one analysis that explored how closeness to each biological parent prior to stepfather entry was associated with adolescent reports of feeling close to the stepfather at Wave II (with a single item used to assess closeness). King found that closeness to mothers was associated with adolescent reports of closeness to the stepfather but that closeness to the nonresident father was unrelated to stepfather–adolescent closeness. The current study went beyond King’s by (a) including measures of adolescent adjustment as predictors of closeness to stepfathers, (b) relying on structural equation modeling to represent relationships as latent variables measured without error, and (c) using a larger sample of stepfamilies.

Method

We used data from the first two waves of Add Health. The Wave I in-home interview included 20,745 adolescents in Grades 7–12 during the 1994–1995 school year and is nationally representative when appropriate sample weights are used. Parent data (n = 17,670) were collected from one parent, usually the biological mother (see Harris et al., 2009, for a detailed description of the data). In 1996, 14,738 of the adolescents were reinterviewed. (The Wave II sample design purposefully excluded adolescents who were in 12th grade at Wave I.) The analytic sample for this study was restricted to adolescents with valid sample weights who reported living with a single biological mother (no partner in the household) at Wave I and who transitioned to a married-stepfather family by Wave II (n = 179).

We relied on structural equation modeling techniques, a particularly appropriate approach for this study given the multiple pathways proposed and the underlying latent constructs outlined in the conceptual model. Analyses were conducted in Mplus Version 6 (Muthén & Muthén, 2010). Mplus uses full-information maximum-likelihood to deal with missing data. Results are based on weighted data, with standard errors adjusted for clustering and stratification in the Add Health sample design. We report descriptive statistics based on nonmissing, weighted data.

The dependent variable, stepfather–child relationship quality, was treated as a latent construct with five observed indicators taken from the Wave II adolescent interview: (a) how close adolescents feel to their stepfather, (b) how much they feel he cares about them, (c) how much they feel he is warm and loving, (d) how satisfied they are with their communication, and (e) how satisfied they are with their overall relationship (1 = “not at all/strongly disagree,” 2 = “very little/disagree,” 3 = “somewhat/neither agree nor disagree,” 4 = “quite a bit/agree,” 5 = “very much/strongly agree”). A scale based on these observed indicators produced an alpha reliability coefficient of .90 (M = 3.75, SE = .08 ) Although a majority of adolescents perceived having good relationships with stepfathers (as indicated by responses in the top two agreement categories), a substantial minority gave low to moderate ratings on closeness (43%), feeling the stepfather cares (23%), stepfather warmth (31%), satisfaction with communication (30%), and the relationship in general (28%).

The indicators for the mother–child relationship quality latent construct were identical to those for stepfather–child relationship quality but were measured at Wave 1 (α = .87, M = 4.46, SE = 0.06). The latent construct nonresident father–child relationship quality was measured with three observed indicators from the Wave 1 adolescent interview (α = .89): (a) how often they talked with their father or received a letter from him (M = 1.94, SE = 0.21), (b) how often they stayed overnight with their father (0 = not at all to 5= more than once a week; M = 1.11, SE = 0.16), and (c) how close they felt to their father (range: 1 = not at all close to 5 = extremely close; M = 2.69, SE = 0.16).

The latent construct child delinquency was measured with 13 observed indicators from the Wave 1 adolescent interview. These items tapped the frequency of behaviors such as stealing, lying to parents, damaging property, and fighting during the past 12 months (see Table 1). Response categories originally ranged from 0 (never) to 3 (five or more times) but were dichotomized (0 = never, 1 = once or more) to reduce extreme skewness and then summed (α = .81, M = 2.67, SE = 0.28).

Table 1.

Measurement Model With Standardized Factor Loadings and Correlations

Variable	Coefficient	SE
Stepfather–child relationship
Closeness	.878***	.071
Caring	.688***	.107
Warmth	.833***	.070
Communication	.632***	.097
Overall relationship	.536***	.104
Mother–child relationship
Closeness	.787***	.059
Caring	.370***	.060
Warmth	.702***	.064
Communication	.825***	.046
Overall relationship	.860***	.043
NRF–child relationship
Closeness	.899***	.080
Number of nights	.851***	.097
Amount of contact	.736***	.061
Child delinquency
Damaged property	.844***	.081
Stole more than $50	.965***	.062
Sold drugs	.622***	.107
Stole less than $50	.607***	.103
Seriously hurt someone	.696***	.100
Used/threatened with weapon	.808***	.127
Was in a group fight	.820***	.102
Lied to parents about location	.554***	.101
Shoplifted	.701***	.072
Took car without permission	.611***	.148
Got into serious fight	.765***	.080
Child depression
Couldn’t shake blues in last week	.764***	.059
Felt depressed in last week	.717***	.046
Felt lonely in last week	.766***	.070
Felt sad in last week	.558***	.072
Felt life not worth living in last week	.511***	.062
Correlations
SF–child with mother–child	.392***	.082
SF–child with NRF–child	.194†	.108
SF–child with delinquency	−.352***	.097
SF–child with depression	−.109	.114
Mother–child with NRF–child	.140	.095
Mother–child with delinquency	−.153†	.090
Mother–child with depression	−.396***	.067
NRF–child with delinquency	.118	.129
NRF–child with depression	−.272*	.116
Delinquency with depression	.293*	.125

Note. Correlations between residuals of observed indicators that were included to improve model fit are not shown in the table. χ² (347) = 359.01, root-mean-square error of approximation = .01, comparative fit index = .98. SF = stepfather; NRF = nonresident father.

^†p < .10.

^*p < .05.

^***p < .001. (two-tailed)

The latent construct child depression was measured with five observed indicators from the Wave 1 adolescent interview. Adolescents were asked how many times within the past week (range: 0 = never, 3 = five or more times) they (a) could not shake the blues, (b) felt depressed, (c) felt lonely, (d) felt sad, and (e) felt that life was not worth living (α = .85, M = 0.43, SE = 0.05). Although the interview included a larger number of depressive symptoms, a preliminary analysis revealed that the items did not represent a single underlying dimension. For this reason, we relied on five items with high loadings on the first general factor to emerge from an exploratory factor analysis.

The adolescent’s age was measured in years (M = 15.01, SE = 0.22). Gender was coded as a binary variable for which 0 = male and 1 = female (45%). Race was a binary variable with Black (24%) coded 1 and all others (mostly White) coded 0. The stepfather’s level of education was measured as a continuous variable ranging from 1 (less than a high school education) to 4 (college degree or beyond; M = 2.22, SE = 0.09). The number of siblings with whom the adolescent reported to be living in the household was measured as a continuous variable (M = 1.57, SE = 0.14). We created the above background variables using Wave I adolescent reports, with the exception of stepfather’s education, which came from the Wave II adolescent interview because it was not available in Wave 1. Because it is unlikely that the stepfather’s level of education changed during the year, we treated this measure as a Wave 1 background characteristic along with the other Wave 1 background measures. (Study findings were similar when we substituted the Wave I mother’s education for the stepfather’s education; we retained the stepfather’s education in our model because it is likely more relevant for predicting the quality of the stepfather–adolescent relationship.) The number of prior father figures experienced by adolescents (M = 1.58, SE = 0.10) drew on a series of questions from the Wave I mother interview regarding her relationship history and was calculated as the number of coresidential relationships (marriages and cohabitations) to which the child had been exposed since birth. A binary variable was also created to indicate that the adolescent was born in marriage (72%).

Results

Measurement Model

We used confirmatory factor analysis to assess how well the latent variable model fit the data and the root-mean-square error of approximation and comparative fit index to indicate the overall goodness of fit. The measurement model is presented in Table 1. This model included correlations between all latent variables, along with several correlations between the residuals of observed indicators (not shown) to improve model fit. The overall fit of the measurement model was good, χ²(347) = 359.01, p > .05; root-mean-square error of approximation = .01; comparative fit index = .98).

The standardized factor loadings on all five latent variables were acceptable, with factor loadings ranging from .37 to .97. Correlations between the latent variables revealed that stepfather–stepchild relationship quality was positively and significantly associated with mother–child relationship quality (p < .001) and approached significance (p < .10) with nonresident father–child relationship quality. These findings suggest that adolescents’ relationships with stepfathers are more positive when relationships with both biological parents were positive before the stepfather entered the household. The correlation between stepfather–child relationship quality and child delinquency was negative and significant (p < .001), which suggests that adolescents’ relationships with their stepfather are less positive when the child exhibited more delinquent behaviors before the stepfather entered the household. The correlation between stepfather–child relationship quality and child depression was not significant, although mother–child and nonresident father–child relationship quality were negatively associated with child depression (p < .001 and p < .05, respectively). Child delinquency and depression were positively correlated (p < .05), which is consistent with previous research showing that externalizing and internalizing problems tend to co-occur in children and adolescents (e.g., Marmorstein & Iacono, 2004).

Structural Model

Following the conceptual model (see Figure 1), the structural model assumes that the mother–child relationship, the nonresident father–child relationship, child delinquency, and child depression directly affect the extent to which adolescents develop positive relationships with stepfathers within approximately the first year of his entry into the household. The model also assumes that background variables have direct effects on the quality of the stepfather–child relationship as well as indirect effects through other family relationships, child delinquency, and child depression.

As shown in Table 2, regression coefficients for the structural model indicated that both the quality of the mother–child relationship (b = .34) and the level of child delinquency (b = −.58) prior to stepfather entry were significantly associated (both ps < .01) with adolescent reports of positive stepfather–child relationships at Wave II, controlling for child and family characteristics. These findings are consistent with our assumption that the quality of the mother–child relationship and the child’s level of delinquency prior to stepfather entry are good predictors of the quality of children’s relationships with stepfathers after he enters the household. The regression coefficient for the quality of the nonresident father–child relationship (b = .24), however, was smaller in magnitude and not statistically significant. The level of child depression (b = .36) prior to stepfather entry was significantly (p < .01) and positively associated with adolescent reports of positive stepfather–child relationships at Wave II, controlling for child and family characteristics. Contrary to expectations, this finding suggests that adolescents who experienced depressive symptoms were more, not less, likely to report positive relationships with stepfathers, a finding to which we return later in this article.

Table 2.

Structural Model Predicting Stepfather-Child Relationship at Wave II (WII, N = 179)

Predictor	Stepfather–child relationship (WII)	Mother–child relationship (WI)	Nonresident father –child relationship (WI)	Child delinquency (WI)	Child depression (WI)
Family relationships and child adjustment
Mother–child relationship  (WI)	.34**
Nonresident father–child  relationship (WI)	.24
Child delinquency (WI)	−.58**
Child depression (WI)	.36**
Child characteristics
Female	−.67*	−.07	−.68**	−.57***	.34***
Age	−.03	−.06	−.08	.18**	.14***
Black	−.83*	−.07	.54	−.77**	−.14
Number of siblings	−.10	−.11*	−.24*	−.14†	.05
Family characteristics
Stepfather’s education	.02	.02	−.003	−.24*	.002
Number of prior father  figures	.01	.04	.43†	−.08	−.03
Child was born in  marriage	−.35	.18	1.01**	−.48†	−.18

Note. Correlations between residuals of observed indicators that were included to improve model fit are not shown in the table. Coefficients for family relationship and child adjustment variables are fully standardized. Coefficients for child and family characteristics are partially standardized on the dependent variable. χ²(515) = 544.22, root-mean-square error of approximation = .02, comparative fit index = .94. WI = Wave I.

^†p < .10.

^*p < .05.

^**p < .01.

^***p < .001. (two-tailed)

With respect to other covariates, two background variables had direct effects on stepfather–child relationship quality: Girls reported lower quality relationships with their stepfathers than boys (b = −.67, p < .05), and Black adolescents reported lower quality relationships with their stepfathers than other adolescents (b = −.83, p < .05).

We also assessed the statistical significance of the indirect effects implied by our model. Some observers have recommended the use of bootstrapping methods to generate standard errors for indirect effects when sample sizes are small, as in the present study (MacKinnon, 2008). Bootstrapping is not available in Mplus, however, for complex sample designs such as Add Health’s. Therefore, we relied on traditional Sobel tests to assess the indirect effects. Because these results are approximate, readers should treat them cautiously.

As noted earlier, daughters reported less positive relationships with stepfathers than did sons. The data in Table 2 also show that daughters reported less delinquency than did sons (b = −.57), and delinquency, in turn, was negatively associated with the quality of the stepfather relationship (b = −.58). This corresponds to an indirect estimated effect of .33 (p < .05). Daughters also reported more depressive symptoms than did sons (b = .34), and depressive symptoms, in turn, were positively associated with the quality of the stepfather relationship (b = .36). This corresponds to an indirect estimated effect of .12 (p < .05). These findings suggest that the tendency for daughters to engage in less delinquency and experience more symptoms of depression (compared with sons) helped them become closer to stepfathers. With delinquency and depressive symptoms controlled, however, daughters were considerably less close to stepfathers than were sons.

The direct association between adolescent age and the quality of the stepfather relationship in Table 2 was close to 0 (b = −.03) and not statistically significant. Nevertheless, age was positively associated with delinquency (b = .18), and delinquency, in turn, was negatively associated with the stepfather relationship. This corresponds to an indirect estimated effect of −.10 (p < .05). At the same time, age was positively associated with depressive symptoms (b = .14), and depressive symptoms, in turn, were positively associated with the stepfather relationship. This corresponds to an indirect estimated effect of .05 (p < .05). In other words, older adolescents were no more or less close to stepfathers than were younger adolescents because the negative effect of being more delinquent offset the positive effect of having more depressive symptoms.

Two other indirect effects emerged from Table 2. First, Black adolescents were less close to stepfathers than were adolescents of other races (b = −.83). At the same time, however, Black adolescents reported less delinquency than did other adolescents (b = −.77). Their lower level of delinquency produced an indirect estimated effect of .45 (p < .05) on the stepfather–child relationship. In other words, Black adolescents’ lower level of delinquency helped to offset their tendency to feel less close to stepfathers. Finally, stepfather education was not directly related to the quality of the stepfather–child relationship (b = .02). Stepfather education was negatively related to delinquency (b = −.26), however, which produced an indirect estimated effect of .14 (p < .05) on the stepfather–child relationship.

Earlier, we reported that reports of depressive symptoms were positively (rather than negatively) associated with the quality of stepfather relationships. In a supplementary analysis, we considered whether stepfather relationships were associated with changes in symptoms of depression. To conduct this analysis, we created a latent variable for adolescent depression at Wave II based on the same five indicators of depression at Wave I. In a structural model (not shown), we assumed that Wave II depression was affected by Wave I depression as well as the quality of the stepfather–adolescent relationship. (With Wave I scores controlled, the Wave II depression variable was conceptually similar to a change score; that is, the variable reflected the extent to which adolescents’ reports of symptoms at Wave II differed from the values predicted by the Wave I scores.) The standardized path from Wave I to Wave II depression was .55 (p < .001), which indicates that reports of depressive symptoms were relatively stable over time. More important, the standardized path from the stepfather–adolescent relationship to Wave II depression was −.22 (p < .05). This finding indicates that adolescents who were closer to their stepfathers reported fewer depressive symptoms than expected, given the number of symptoms they reported prior to the stepfathers’ arrival. This finding helps us understand our earlier, unexpected finding; that is, adolescents who reported a relatively high number of depressive symptoms in Wave I may have formed especially close bonds with their new stepfathers because these relationships helped them feel better.

Discussion

By focusing on children who gain a stepfather during adolescence and examining factors that are associated with the formation of positive relationships between stepfathers and stepchildren within the first year, this study captured families when they were particularly vulnerable. During the early phase of stepfamily formation, family members must (re)negotiate roles and relationships, reorganize many aspects of their lives, and adapt to major changes in the household (Pryor, 2014). For these reasons, the first year is when stepfathers experience the most difficulty integrating themselves into the family (Bray & Easling, 2005) and children are at the greatest risk for developing problems (Hetherington & Kelly, 2002). Despite these challenges, many stepfamilies adapt successfully and, as this study confirms, a majority of married stepfathers appear to establish close and satisfying relationships with their adolescent stepchildren.

Results from the present study point to a number of factors that are associated with positive relationships between stepfathers and adolescent stepchildren during the first year of stepfather entry to the household. Consistent with our conceptual model and a growing body of research (King, 2009; King et al., 2014), the quality of the mother–child relationship was a strong predictor of the quality of adolescent’s relationships with stepfathers after he entered the household. Other scholars have noted the pivotal role mothers play in enhancing, or detracting from, family functioning and child well-being in stepfamilies (Pryor, 2014; Smith, 2008), and it appears that a positive mother–child relationship may also help facilitate the development of positive ties between adolescents and their new stepfathers (see also Weaver & Coleman, 2010).

After controlling for child and family characteristics, the quality of the nonresident father–child relationship was not significantly associated with adolescent ties to stepfathers. This finding is consistent with a few other studies suggesting that the quality of adolescents’ relationships with stepfathers and nonresident fathers are largely independent and that having close ties to a nonresident father does not preclude developing close ties to a stepfather (King, 2006, 2009).

A unique aspect of the current study was the consideration of adolescent adjustment prior to stepfather entry as a predictor of later stepfather–child relationship quality. Adolescent delinquency was a strong predictor (indeed, the strongest predictor in our model) of later stepfather–child ties. Some mothers may (re)marry hoping that the presence of a father figure in the household will reduce the delinquent activities of children who are getting into trouble. These mothers may find that their children and new husband have difficulty getting along with one another, which could exacerbate (rather than improve) tension in the family system.

Contrary to expectations, adolescents who reported high levels of depressive symptoms prior to stepfather entry were significantly more likely than other adolescents to report having positive relationships with their stepfathers. This finding is noteworthy, given that depressive symptoms were negatively associated with the quality of adolescents’ relationships with mothers and biological fathers—a finding consistent with prior research (Hawkins, Amato, & King, 2007; Marmorstein & Iacono, 2004). Our supplementary analysis found that adolescents who reported higher levels of depressive symptoms prior to stepfather entry at Wave I, but positive relationships with stepfathers at Wave II, were more likely to report a decrease in depressive symptoms at Wave II.

Perhaps adolescents with depressive symptoms living with single mothers are especially responsive to the benefits that often accompany the entrance of stepfathers (e.g., increased economic resources, the availability of a father figure in the household, and greater emotional support for mothers). These improvements in quality of life may lead adolescents with depressive symptoms to appreciate their stepfathers and develop close relationships with them (Ganong et al., 2011) that may in turn lessen depressive symptoms. Similarly, adolescents exhibiting depressive symptoms may be emotionally needy and attracted to a new source of social support in the household. Correspondingly, some stepfathers may make special efforts to reach out to such adolescents. Further research is certainly needed to shed light on the ways in which adolescent adjustment affects the development of stepfather–stepchild ties (and how these ties may be associated with later adolescent adjustment), but these results suggest the fruitfulness of pursing additional work on this topic.

Our findings have implications for practitioners who work with recently formed stepfather families or with single mothers planning to remarry. Practitioners should recognize that the development of close ties between adolescents and stepfathers is likely to be challenging when adolescents exhibit high levels of delinquent behavior. These families may need more time and attention to reach satisfactory outcomes. Depressive affect among adolescents, however, does not appear to be a barrier to the formation of close ties between adolescents and newly acquired stepfathers. Indeed, adolescents experiencing symptoms of depression may benefit from his presence in the household.

On the basis of earlier research (e.g., Hetherington & Jodl, 1994), we expected girls to report less positive relationships with stepfathers than boys. This expectation was supported in the analysis. Our analysis also revealed, however, that daughters reported less delinquency and more depressive symptoms than did sons—trends that indirectly increased their closeness to stepfathers. Were it not for the tendency for daughters to report more internalizing than externalizing problems, they would have reported even weaker ties with stepfathers.

Contrary to some previous research (King et al., 2014), we did not find that older adolescents reported especially weak ties with stepfathers. The analysis of indirect effects provided an explanation for the absence of an age effect. Older adolescents reported more delinquency (which weakened ties with stepfathers) as well as more depressive symptoms (which strengthened ties with stepfathers). These two trends offset one another, resulting in no net effect of age on stepfather relationships.

Our finding that Black adolescents reported less positive relationships with stepfathers deserves future attention. Prior studies have reported mixed findings regarding how stepfamily relationships vary by race and ethnicity. For example, King’ study (2009) of families who recently transitioned to a stepfamily also found that Black adolescents reported less positive relationships with stepfathers. Another study that examined adolescents’ relationships with married stepfathers at Wave I of Add Health, however, found no racial differences (King et al., 2014), although this latter cross-sectional study included stepfamilies of all durations. It may be that establishing positive relationships with stepfathers is more difficult for Black youth only when this occurs during adolescence, or only during the initial transition, or to some other constellation of factors.

Although this study identified a number of important factors associated with positive stepfather–stepchild ties, data limitations precluded an examination of other potentially important factors that should be examined in future research. For example, data were not available on the stepfather–mother relationship or stepfather–child relationship prior to his entry into the household. The current study focused on only one time point after stepfather entry and did not address how the stepfather–child relationship unfolds over time. There is no measure that captures exactly when the transition occurred between the two waves, so the number of months the stepfather had been in the household (or whether the marriage was preceded by a brief period of cohabitation) could not be taken into account, yet this might also have implications for the quality of stepfather–stepchild ties.

Our findings are also limited to married-stepfather families; although this is a common stepfamily form, it does not include the growing diversity of stepfamilies, including cohabiting stepfamilies, stepmother families, and gay and lesbian stepfamilies. In particular, our results speak to stepfamilies formed through marriage (or through a very short period of cohabitation that transitioned to marriage) and may not apply to stepfamilies in which parents are unmarried. Unions in which mothers are cohabiting with new partners tend to be unstable and not last very long. The dynamics and factors associated with the development of stepparent–child bonds (and their consequences) are likely different in cohabiting stepfamilies.

We should also note that our findings are based on a relatively small sample that precluded a full examination of other important issues, such as the possibility of bidirectional associations or long-term child outcomes. For example, stepfather–adolescent relationship quality may be associated with later adolescent adjustment, too. Finally, all of our observations came from adolescents’ reports, which is a limitation of the Add Health data. Having information on the perspectives of stepfathers as well as adolescents would have broadened the scope of the study.

The transition to a stepfamily is a critical life course event for adolescents and other family members, although few studies have captured stepfamilies during this juncture. The current study contributes to the growing literature on factors associated with the formation of close relationships between stepfathers and stepchildren. We have extended prior research in this area by using nationally representative data to look prospectively at stepfather–adolescent relationships, providing a better understanding of how factors existing prior to stepfamily formation are associated with the quality of stepfather–adolescent ties within the first year after married stepfathers join the household. Our findings suggest that both the quality of the mother–adolescent relationship and the adolescent’s adjustment prior to stepfamily formation are significantly associated with the development of adolescent–stepfather ties.