Introduction

A body of research shows that parents and children disagree about parenting behaviors and family functioning in domains such as nurturance, support, monitoring, conflict, and discipline; furthermore in some domains these discrepancies are consistent over time (De Los Reyes et al. 2010; 2016; Ehrlich et al. 2011; Stuart and Jose 2012). Overall, parents tend to report more positive views of the family and parenting behaviors compared to children, and these discrepant views are uniquely related to youth’s adjustment, beyond what is accounted for by parents’ and/or youth’s individual reports (Gaylord et al. 2003; Guion et al. 2009). Importantly, the parent–adolescent relationship is a subsystem of the larger family unit. A family systems perspective directs attention to the larger family context within which dyadic parent–youth relationships are embedded (Cox and Paley 1997; Minuchin, 1985), but we know almost nothing about how discrepancies in the relationship perspectives of one parent–youth dyad are linked to other relationship experiences in the family. In addition, most research on discrepancies between youth and parent perceptions has studied individual dyads to assess the links between discrepant relationship perspectives and youth outcomes. In contrast to such a variable-oriented approach, a pattern analytic approach aims to capture a more holistic picture of individuals’ experiences (Magnusson 2003). Applied to families, such an approach places dyadic experiences in the context of larger family patterns. In this study we adopted such an approach, using latent profile analysis to identify patterns in the dyadic relationship experiences of mothers, fathers and two siblings from each family.

Numerous studies show that, in comparison to White families, African American and other minority families report more discrepant views about the parent–adolescent relationship in domains such as discipline and conflict (Guion et al. 2009; Ehrlich et al. 2011). Few studies, however, have extended this research to study within group variation in this ethnic group (Coll et al. 1996). In addition, most research to date on discrepant views about parenting behaviors and the parent–child relationship has focused on discrepancies within the mother–adolescent subsystem, limiting our understanding of parent–adolescent relationships within the context of the larger family unit. The congruence between youth’s and mothers’ relationship perspectives, however, may have different implications for their adjustment depending on the congruence in relationship perceptions of other parent–youth dyads such as siblings’ relationships with their mothers or the congruence between youth’s and their fathers’ relationship perceptions. We know almost nothing about discrepancies involving fathers’ views of family functioning and parenting behaviors, particularly in African American families. This neglect is consistent with the limited attention to African American fathers in the research literature, in general—in the face of data showing that significant numbers of African American children under the age of 18 reside in two parent households (over 30 %) and that many African American fathers, residential and non-residential, are involved in their children’s lives (Coley and Chase-Landsdale 1999).

One important dimension of parent–adolescent relationships that has implications for youth adjustment is conflict. Although adolescents and parents generally have positive views about the parent–child relationship, parent–adolescent conflict is frequent, and is often considered a normative part of the development process in U.S. families (Laursen and Collins 2009). Parent–adolescent conflict also has been discussed in the literature on discrepancies between parents’ and youth’s views of family functioning, but few studies have directly examined discrepancies about conflict in parent–adolescent relationships. Instead, scholars have theorized that conflict underlies discrepancies between parents’ and adolescents’ views, or that parent–adolescent discrepancies in perceptions lead to conflict (Holmbeck and O’Donnell 1991; Treutler and Epkins 2003). Indeed, in one study, conflict was examined as a mediator between discrepancies in parental control and youth’s depressive symptoms (Juang et al. 2007). Current research shows that parents’ and adolescents’ views about parent–adolescent conflict are only moderately correlated (Ehrlich et al. 2015), however and, if parents and adolescents do not see eye-to-eye about conflict in their relationship, discrepancies in this domain could have consequences for youth’s adjustment as frustrations and feelings of alienation, brought about by a perceived lack of understanding of one another’s perspectives, take their toll.

To advance understanding of parent–youth differences in their perceptions of their shared relationship, this study addressed two goals. First, using an ethnic homogeneous design to better understand within-group variation (Coll et al. 1996) and focusing on a sample of two-parent, African American families, our first goal was to identify groups of families that varied in their patterns of parent–youth dyadic conflict as reported by mothers, fathers, and two adolescent-aged siblings. Next, using 3 years of longitudinal data, our second goal was to test whether there were differences in youth’s well-being, specifically, depressive symptoms, risky behaviors, and perceptions of parental warmth, as a function of family type as defined by family patterns of conflict reports.

Parent–Adolescent Conflict

Among African American families, parent–adolescent conflict tends to be frequent but moderate in intensity. These conflicts tend to involve everyday issues such as school-work, chores, interpersonal issues and peer relationships (Smetana et al. 2003; Smetana and Gaines 1999; Smetana et al. 2004). Some scholars suggest that conflict may not be a major area of divergent relationship perspectives between parents and adolescents because it is a stable indicator of broader family dynamics, or because conflict is more overt in comparison to other indicators of the parent–child relationship such as cohesion or control (Gonzales et al. 1996; Stuart and Jose 2012). Discrepancies between parents and adolescents regarding conflicts in the parent–adolescent relationship, however, have been found. For example, African American mothers report more conflict with adolescents and rate conflict as less intense over time than adolescents (Gonzales et al. 1996; Smetana et al. 2003; Smetana and Gaines 1999; Smetana et al. 2004). Furthermore, parents and adolescents often view conflict differently, with parents describing some conflicts in terms of violations of social convention, but adolescents describing the same conflicts as involving issues of personal jurisdiction (Smetana et al. 2003). It is therefore not surprising that, similar to other domains of family functioning, adolescents’ and mothers’ and adolescents’ and fathers’ perceptions of conflicts or disagreements are only modestly correlated (Ehrlich et al. 2011, 2015).

Although mothers’ and fathers’ perceptions about family are often correlated (Gaylord et al. 2003), it is not clear whether conclusions drawn from studies about discrepancies between mothers and youth are generalizable to father–youth relationships. For example, an early study of majority White families showed that mothers were more discrepant from their children than fathers in their perceptions of family cohesion and power (Feldman et al. 1989). In contrast, a more recent study of White families revealed that mothers’ and adolescents’ reports of parent–adolescent conflict were generally congruent, but that fathers reported more conflict than did adolescents (Ehrlich et al. 2011). Secular increases in paternal involvement (Parke and Buriel 1998) mean that fathers may experience more conflict due to their greater engagement. Studies that include data collected from both mothers and fathers from the same families are especially rare, however, particularly in the case of minority families, which have also been neglected in research on these kinds of normative family processes. Thus, a direction for research is to determine whether parent–youth relationship perspective discrepancies are gendered. In addition, from a family systems perspective, experiences in one dyad are inter-connected with those of other dyads in the family. For example, discrepant perspectives that are evident in one dyad may spill over to affect experiences in other dyadic relationships, such that, for example father–child relationship experiences may have implications for mothers’ relationships with their children.

Although older adolescents report more family conflict than younger adolescents, the few available data suggest that discrepancies in relationship reports are similar for younger and older adolescents (Stuart and Jose 2012). We could find no studies, however, that assessed discrepancies between parents’ and adolescents’ relationship perceptions for more than one adolescent in a family. Within-family studies of parent–adolescent conflict might help us better understand age differences in conflict discrepancies because other family characteristics that may serve as “third variable” explanations are essentially controlled. In addition, as we noted, from a family systems perspective, relationship experiences can be better understood when studied within a larger family context. Accordingly, in this study we built on the existing research literature by examining parent–adolescent conflict frequency as reported by mothers, fathers and two adolescent-aged siblings from the same African American families.

Parent–Adolescent Discrepancies and Youth Adjustment

Developmental theories suggest and empirical research documents that parent–youth interactions have implications for adolescents’ adjustment. We know less about how discrepancies between parents’ and adolescents’ views about family are related to youth’s well-being. Some early studies found that parent–adolescent discrepancies were related to less negative outcomes for youth, perhaps because the discrepancies that were evident, for instance, girls reporting high but mothers reporting low communication, reflect adolescents’ developmentally appropriate individuation from the family (Holmbeck and O’Donnell 1991; Ohannessian and De Los Reyes 2014). From a family systems perspective, however, parent–adolescent discrepancies may reflect a lack of family cohesion or family disorganization (Minuchin 1985). Consistent with this perspective, greater discrepancies between adolescents’ and parents’ reports of family functioning were positively related to internalizing and externalizing problems for both children and adolescents (Gaylord et al. 2003). Importantly, mother–adolescent and father–adolescent discrepancies may have different implications for youth adjustment, highlighting the need for studies that include both fathers and mothers.

Parent–adolescent discrepancies about conflict in parent–adolescent relationships may have negative implications for adolescents given that these discrepancies might indicate a lack of understanding and acceptance of one another’s reactions to their “shared” relationship experiences. Indeed, recent research shows that open communication in parent–adolescent relationships is linked to fewer parent–adolescent discrepancies in perceptions of conflict (Ehrlich et al. 2015). Ehrlich et al. (2015) suggested that discrepant views regarding mother–adolescent and father–adolescent conflict emerge because of poor communication about the disagreements that occur. Emotions that are generated during conflicts between parents and adolescents may also spill over to affect youth’s perceptions about other dimensions of their parent–child relationships. To examine potential spillover of emotions across relationships, in this study, we expanded on prior research to examine whether family patterns of discrepancies in parent–adolescent conflict perceptions were related to parents’ and youth’s perceptions of parental acceptance as well as youth’s internalizing and externalizing problems.

The Current Study

Taken together, research shows that discrepancies in parent–adolescent relationship perceptions occur across many domains of relationship and family functioning, including parent–adolescent conflict. Building on this literature, in this study we focused on parent–adolescent discrepant views regarding the frequency of parent–adolescent conflict using a person-centered approach, latent profile analysis (LPA), to determine whether we could identify groups of families that varied in both the levels of and differences between partners’ reports of four parent–youth relationships within the same families: mother–older sibling, mother–younger sibling, father–older sibling, and father–younger sibling. To illuminate the potential implications of family patterns of parents’ and youth’s experiences of conflict in their relationship, we then tested whether youth differed over time in their reports of depressive symptoms, risky behavior, and parental acceptance as a function of family type.

Method

Participants

Participants were 187 African American families who were part of a 3-year longitudinal study on gender socialization and gender development in two-parent African American families living in the mid-Atlantic region of the United States. To be eligible for the study, families had to identify as Black or African American and include a mother and a father figure who resided together and who were raising at least two adolescent-aged children. In families with more than two children, the two consecutively born children closest to age 13 were the focal youth in the study. About half of the families were recruited by African American recruiters in the community who posted advertisements in local businesses, provided information to local churches and distributed flyers at youth activities. Interested families contacted the recruiters, who then provided families’ contact information to the project managers. The remaining families were recruited via mailings purchased from a marketing firm. A total of 202 families participated in the first year of the study. However, we did not include data from families in which the parents were not romantically involved (e.g., mother living with her father and raising her children), or were divorced during the first year of the study. Attrition rate across the 3 years of the study was 5 %.

All participating youth self-identified as Black or African American, as did 94 % of mothers and 97 % of fathers. In the first year of the study, older siblings averaged 13.92 years (SD = 1.91; range = 10.03–19.07), and younger siblings averaged 10.30 years (SD = 1.11; range = 7.84–12.92) years of age. Average ages of mothers and fathers were 40.43 (SD = 5.61; range = 27.69–60.71) and 43.11 (SD = 7.22; range 26.50–66.12) years, respectively. Most mothers reported that they were the biological parent for older (93 %) and younger (96 %) children. Similarly, most fathers were the biological fathers of older (75 %) and younger (80 %) siblings. In year 1 of the study, the average educational level was 14.63 years (SD = 1.82; range = 9–19) for mothers and 14.23 years (SD = 2.36; range = 5–19) for fathers (a score of 12 indicated high school graduate). The median family income was $60,000 (SD = $48, 387; range = $3000–$275,000).

Procedure

During each of the 3 years of the study, two African American interviewers conducted separate home interviews with fathers, mothers, and each of the two target children. In the present study, we used data from parents and youth on parent–child relationship (conflict and acceptance), youth reports of their psychological well-being, and parents’ reports of family background characteristics (family income). Informed consent/assent was obtained from all participants prior to data collection, and families received a $200 honorarium for their participation. The project was approved by the university’s IRB.

Measures

Parent–Adolescent Conflict

Parent–adolescent conflict was measured during the first year of the study using an instrument adapted from Smetana (1988) that assessed conflict frequency in nine areas (chores, appearance, homework/school work, social life, health, independence, religion, behavior and personality, and relationships with siblings). For each topic, participants were asked to rate the frequency of conflicts/disagreements. At different points in the home interviews, youth rated the frequency of conflicts with mothers and fathers and mothers and fathers rated the frequency of conflicts or disagreements with their older and younger children using a 6-point scale (1 = not at all, 6 = several times a day). Ratings were averaged, with higher scores reflecting more conflict during the past year. A separate score was computed for each parent–child dyad, resulting in eight scores reflecting parent–adolescent conflict across reporters and dyads (i.e., mother rating of conflict with older sibling, older sibling rating of conflict with mother, mother–younger sibling, younger sibling–mother, father–older sibling, older sibling–father, father–sibling child, younger sibling–father). Cronbach’s alphas ranged from .77 to .87 across reporters.

Depressive Symptoms

Older and younger siblings reported on their depressive symptoms in Years 1, 2, and 3 using the 10-item version of the Childhood Depression Inventory (CDI; Kovacs, 2001). For each item, youth indicated which of three statements best described their feelings during the past 2 weeks (e.g., 0 = I am sad once in a while; 1 = I am sad many times; 2 = I am sad all the time). Responses were summed with higher scores indicating more depressive symptoms. Cronbach’s alphas ranged from .68 to .82. A log transformation was applied to correct for positive skew.

Risky Behavior

Older siblings reported on risky behavior in Years 1, 2, and 3, and younger siblings, in Years 2 and 3 using an 18-item scale developed by Eccles and Barber (1990). Items assessed participation in behaviors such as smoked cigarettes, skipped a day of school, and disobeyed parents on an important issue in the past year (1 = never, 4 = more than ten times). Items were summed, and Cronbach’s alphas ranged from .79 to .87. A log transformation was applied to correct for positive skew.

Parent–Adolescent Acceptance

In all 3 years of the study, older and younger siblings rated their mothers’ and fathers’ acceptance at separate points in the interview using eight items from the Child’s Report of Parental Behavior Inventory (Schwarz et al. 1985). Youth rated their parents on a 1 (really unlike) to 4 (really like) scale (e.g., “My mother understands my problems and worries”). Items were averaged to create total scores, with higher scores indicating more acceptance. During all 3 years of the study mothers and fathers also rated their relationship with each of their children on a 1 (not at all) to 5 (very much) scale (e.g., “I am a person who sees my child’s good points more than his/her faults”) using the parents’ version of the Child’s Report of Parental Behavior Inventory (CRPBI). On these measures, Cronbach’s alphas ranged from .78 to .94 across reporters and times of assessment.

Overview of Analyses

Analyses were conducted in two steps. First, we conducted latent profile analysis (LPA) using Mplus version 7.4 (Muthén and Muthén 1998–2015) to identify groups of families that were similar in their patterns (i.e., levels of and within-family differences in) parent–adolescent conflict across mothers’, fathers’ and two siblings’ reports. LPA is a data driven approach, and we made no a priori assumptions about the nature or number of profiles that would emerge. We began with a one-profile model and increased the number of profiles until optimal fit was achieved (as indicated by various fit indicators, described below). In our LPA, mothers’ and fathers’ reports of conflict with older and younger siblings, and older siblings’ and younger siblings’ reports of conflict with each parent were used as the observed variables that served as indicators for the categorical latent variable, c. We used three model fit indicators—the Akaike information criterion (AIC), the Bayesian information criterion (BIC), and the bootstrap likelihood ration test (BLRT)—as well as substantive considerations in deciding upon the number of latent profiles to extract from the data. None of these indicators is an absolute measure of model fit; rather, these statistics are useful in comparing the relative fit of models with varying numbers of profile. For the AIC and BIC, the model that produces the lowest value is regarded as the best fitting model. The BLRT produces a p value that is used to compare the LPA model being estimated to the model with one less profile. A non-significant p value for the BLRT indicates that the model with one less profile is preferred.

The second step of the analyses entailed using family latent-profile membership to predict changes in depressive symptoms, risky behavior and parental acceptance. Modal profile membership was exported to SAS, which was then used to test a series of multi-level models (MLM) using the MIXED procedure to examine if latent profile membership was linked to the youth well-being outcomes. We used MLM because of the clustered nature of the data. Family-wide characteristics, specifically, latent profile membership, were entered at the between-family level (Level 3), siblings were clustered within families (Level 2), and time was clustered within individuals (Level 1). We used youth age as the metric of time, and tested for linear, quadratic and cubic polynomial terms in the adjustment and parental acceptance outcomes. Youth age was centered at age 13 (the average age across siblings and time points).

In all MLMs we included as covariates, youth gender and birth order (Level 2). In addition, two- and three-way, cross level interactions of latent profile membership (Level 3), birth order and gender (Level 2), and their interaction with time were included in each model to test for birth order and gender differences in the effects of profile membership at the intercept as well as in the developmental trajectories of each outcome. Only significant interactions were retained in the final models.

Results

Parent–Youth Conflict Profiles

Indicators of fit for models with 1–4 latent profiles are shown in Table 1. Although the indicators suggested that the four-profile solution may be preferable to the three-profile solution, the four-profile solution yielded two conceptually similar profiles wherein parents reported more conflicts than their children. Given that the change in BIC from the three profile model to the four profile model was relatively small, we retained the more parsimonious three-profile model. Results from a one-way ANOVA showed that the family profiles of conflict we identified were not significantly different in demographic characteristics including mothers’ income, F(2, 172) = .03, p = .97, fathers’ income, F(2, 164) = 2.37, p = .10, sibling gender composition, F(2, 184) = 1.87, p = .16, mothers’ age, F(2, 184) = .78, p = .46, fathers’ age, F(2, 184) = .91, p = .41, older siblings’ age, F(2, 184) = .2.65, p = .07, younger sibling’s age, F(2, 184) = .51, p = .60 or family size, F(2, 184) = .11, p = .13.

Table 1 Fit indices for LPA models with 1–4 profiles (N = 187 families)
Full size table

Means (and standard deviations) for each family member’s report of parent–youth conflict, by profile, are displayed in Table 2, and correlations between all study variables are shown in Table 3. Figure 1 illustrates the means of each family profile. The largest profile (65 % of the sample) was labeled the low conflict profile. This group was characterized by overall low levels of conflict reported by each member of each dyad as compared to the other two groups based on one-way ANOVAs. This group was also characterized by the lowest overall mean level of parent–adolescent conflict. Paired t tests revealed within-family differences within this family type, however, such that youth reported less conflict with their parents than their parents reported with them, and mothers reported the highest levels of conflict.

Table 2 Means (SDs) for family conflict profiles and overall dyadic conflict
Full size table
Table 3 Correlations between Study Variables
Full size table
Fig. 1
figure 1

Mean levels of conflict frequency over the past year as reported by each family member in the three family profiles. Note Sib O = older sibling; Sib Y = younger sibling. First partner signifies reporter, e.g., Mother Sib O = mothers’ report of conflict with older sibling; Sib O mother = older siblings’ report of conflict with mothers

Full size image

The second profile (17 % of the sample) was labeled the father high conflict profile. The results from a one-way ANOVA with Tukey post hoc comparisons showed that fathers’ ratings of conflict were higher in this group in comparison to those of fathers in each of the other two family profiles. In addition, the overall mean of parent–adolescent conflict in this profile was significantly higher than that for the low family conflict profile but not different from that of the younger sibling high conflict profile. Fathers in this profile also reported more frequent conflict with both older and younger siblings in comparison to other family members’ reports based on paired samples t tests (Table 2).

The third profile (18 % of the sample) was labeled the younger sibling high conflict profile. Results from a one-way ANOVA with Tukey post hoc comparisons showed that younger siblings reported higher levels of conflict in comparison to younger siblings in any of the other profiles. Overall parent–adolescent conflict in this profile was significantly higher than overall parent–adolescent conflict in the low family conflict profile but was not significantly different from the father high conflict profile. In addition, younger siblings in this group reported more conflict with fathers and with mothers in comparison to mothers’ and fathers’ reports of conflict with them, parents’ reports of conflict with older siblings, and older siblings’ reports of conflict with mothers or fathers (Table 2).

Family Conflict Profile, Youth Adjustment and Family Relationships

We next used MLM to test whether the family conflict profiles were associated with both levels of and changes in depressive symptoms, risky behaviors and paternal and maternal acceptance. In these models, the low family conflict profile served as the reference group against which the other two family conflict profiles were initially compared, and we then used the younger sibling high family profile as the reference group to compare this profile with the father high conflict profile. A series of dummy predictor variables was used to indicate youth’s family conflict group membership. As noted, youth birth order and gender were tested for potential moderation effects (older sibling and girl were the reference groups).

Depressive Symptoms

As shown in Table 4 (Model A), youth in the younger sibling high conflict profile reported more depressive symptoms than youth in the low family conflict profile at age 13, γ = .26, SE = .08, p < .001, but there was no significant difference between the low conflict profile and the father high conflict profile, γ = .06, SE = .08, p = .34. Youth in the younger sibling high conflict profile also reported marginally significantly more depressive symptoms than those in the father high conflict profile, γ = −.20, SE = .10, p = .05 (Model B). There were no significant moderation effects by birth or gender, however, and profile effects did not change over time.

Table 4 Coefficients (and SEs) for tests of differences in youth depressive symptoms and risky behaviors as a function of family profile (N = 187 families)
Full size table

Risky Behaviors

As shown in Table 4 (Model A), youth in the younger sibling high profile reported more risky behaviors than those in the low conflict profile at age 13, γ = .07, SE = .03, p = .03. Youth in the father high conflict profile, however, differed neither from those in the low conflict profile at age 13, γ = .05, SE = .03, p = .09 nor those in the younger sibling high conflict profile, γ = −.01, SE = .04, p = .69 (Model B). There were no significant moderation effects by birth order or gender and the effects did not change over time.

Mothers’ Acceptance

As shown in Table 5 (Model A), results revealed no difference between the low conflict profile and the younger sibling high conflict profile, γ = −.06, SE = .08, p = .40. A significant family profile x birth order interaction, γ = .34, SE = .11, p = .002, in combination with follow-ups showed that in comparison to the low conflict profile, older siblings in the father high conflict profile reported lower maternal acceptance, γ = −.27, SE = .08, p = .001, but there were no differences for younger siblings, γ = .07, SE = .06, p = .27 (Fig. 2). In addition, youth in the younger sibling high conflict profile were not significantly different from youth in the father high conflict profile at age 13, γ = −.06, SE = .10, p = .58 (Model B). No significant moderation effects for birth order or gender emerged for youth’s reports of mothers’ acceptance, and the effects did not change over time.

Table 5 Coefficients (and SEs) for tests of differences in mothers’ and fathers’ acceptance as a function of family profile (N = 187 families)
Full size table
Fig. 2
figure 2

Relations between family conflict profile and youth’s report of mothers’ acceptance by birth order

Full size image

Mothers’ reports of their own acceptance revealed lower acceptance of youth in the father high conflict profile in comparison to the low conflict profile, γ = −.19, SE = .09, p = .03, but there was no difference between mothers’ report of acceptance in the younger sibling high conflict profile versus the low conflict profile, γ = −.12, SE = .09, p = .18 (Model A). In addition, there was no difference between youth in the younger sibling high conflict profile and the father high conflict profile, γ = −.07, SE = .11, p = .52 (Model B). Finally, no significant moderation effects for birth order or gender emerged for mothers’ reports of acceptance, and the effects did not change over time.

Fathers’ Acceptance

Results for fathers’ acceptance revealed a significant interaction between family conflict profile x birth order for the low conflict profile- father high conflict profile comparison, γ = .33, SE = .12, p = .01 (Model A). Follow-up tests showed that younger siblings in the father high conflict profile reported more acceptance than younger siblings in the low conflict profile, γ = .17, SE = .08, p = .03. Differences for older siblings, however, were not significant (Fig. 3). A marginally significant interaction between birth order and profile for the younger sibling high conflict- low conflict profile comparison also emerged, γ = .22, SE = .12, p = .07. Follow-up tests showed significant differences between older siblings in the low conflict profile versus the younger sibling high conflict profile such that the former group reported more acceptance from fathers, γ = −.22, SE = .09, p = .01, but there were no such differences for younger siblings, γ = .03, SE = .08, p = .71 (Fig. 4). Finally, no significant differences emerged for the younger sibling high conflict profile versus father high conflict profile comparison, γ = 13, SE = .12, p = .25 (Model B).

Fig. 3
figure 3

Relations between family conflict profile and youth’s report of fathers’ acceptance by birth order: low conflict–father high conflict profile comparison

Full size image
Fig. 4
figure 4

Relations between family conflict profile and youth’s report of fathers’ acceptance by birth order

Full size image

Turning to fathers’ reports, in comparison to the low conflict profile, fathers reported less acceptance of youth in the younger sibling high conflict profile, γ = −25, SE = .09, p = .01, but there was no differences between the low conflict profile and father high conflict profile, γ = −.05, SE = .09, p = .59 (Model A). In addition, in comparison to the younger sibling high conflict profile, fathers reported greater acceptance of youth in the father high conflict profile, γ = .30, SE = .12, p = .01 (Model B). No significant birth order or gender moderation emerged, and there were no significant changes over time for fathers’ acceptance.

Sensitivity Analysis

As a test of reliability of the profile solution we examined whether the three profile solution provided a good fit to the data at later points of data collection. In both Year 2 (N = 173) and Year 3 (N = 157), the same three profiles emerged, and the low conflict profile always included the majority of families.

Discussion

Parent–adolescent conflict is a common family dynamic, but parents and adolescents often have different perspectives on this relationship experience (Gonzales et al. 1996; Ehrlich et al. 2015; Smetana et al. 2003). Although previous research shows that parent–adolescent conflict is negatively related to youth’s adjustment (Weymouth et al. 2016), most studies rely on either mothers’ or adolescents’ reports of conflict. Grounded in a family systems perspective (Cox and Paley 1997), which directs attention to the larger family context of dyadic relationship dynamics, in this study we used latent profile analysis to identify groups of African American families that varied in their reports of parent–youth dyadic conflict. In addition to capturing the experiences of four family members, the LPA approach allowed us to capture both levels of conflict and differences, as well as the direction of differences in perceptions of conflict in parent–youth dyads within families and between groups of families. Three family groups emerged: families in which each member of each dyad reported low frequencies of parent–adolescent conflict relative to other families (low conflict families), families in which fathers reported high frequencies of conflict relative to other family members and other fathers (high father conflict families), and families in which younger siblings reported high frequencies of conflict relative to other family members and other younger siblings (younger sibling high conflict families).

Moreover, analyses of the adjustment correlates of these family types revealed that, in comparison to low conflict families, youth in the younger sibling high conflict families reported more depressive symptoms and risky behaviors. Importantly, these group differences were not moderated by birth order, suggesting some spillover across younger and older siblings. Further, and again suggestive of a spillover process, in comparison to low conflict families, older siblings in father high conflict families reported less acceptance from mothers, and mothers also reported less acceptance of youth in this profile. Results for fathers’ acceptance, however, showed no differences for older siblings in the father high conflict profile compared to the low conflict profile, but younger siblings in the father high conflict profile reported more acceptance. Finally, fathers reported less acceptance of youth in the younger sibling high conflict profile in comparison to the low conflict profile, and this effect was not moderated by birth order.

Our study contributed to the literature on discrepancies in parent–youth relationship experiences in several key ways. First, we collected data from fathers, mothers and two siblings from each family to capture both the levels of and discrepancies between parents’ and youth’s reports of dyadic conflict within their larger family context. In addition, using an ethnic homogeneous design, we studied differences among African American families, which have been relatively neglected in the research literature on normative family processes (McLoyd 1998). Below we discuss conclusions that we can draw from this study and highlight directions for future research.

Parent–Adolescent Conflict in African American families

Overall, this sample of African American families reported relatively low levels of parent–adolescent conflict. Further, consistent with previous research (Ehrlich et al. 2015; Smetana et al. 2003), mothers and fathers, on average, reported more conflicts with adolescents than adolescents reported having with their parents, and both older and younger siblings reported more conflicts with mothers than with fathers. As noted, however, beyond these overall results we were able to identify groups of families that varied in parents’ and youth’s reports of levels of, differences in, and directions of differences in perceptions of parent–youth conflict. More than half of the families were in the low conflict group, whereas smaller numbers of youth belonged to families in which adolescents (younger siblings, specifically) or fathers reported higher conflict than other family members and their counterparts in other groups of families. Studies of normative processes in African American families are rare (Coll et al. 1996; McLoyd 1998), but one previous investigation showed that, as in more typically studied samples of White families, parent–adolescent conflict is a regular occurrence in African American families (Smetana and Gaines 1999). We added to this literature by identifying family patterns of parent–youth conflict.

Although fathers have been relatively neglected in research on parent–adolescent conflict, one prior study of White American families of boys found that mothers reported more positive views about family functioning than did fathers and sons (Feldman et al. 1989). In contrast, in this study, although conflict reported by African American mothers did not distinguish the family types we identified, in the low conflict families, mothers generally reported more conflict than other family members, especially their children. In the case of fathers, our findings were consistent with results from a recent study of majority White families with adolescent offspring, which showed that, in comparison to mothers, fathers were more discrepant from their children in their reports of parent–youth conflict (Ehrlich et al. 2015). In contrast, in our study, fathers’ reports of more frequent conflict were limited to a small proportion of families. Fathers’ discrepant experiences, however, appeared to engender substantial spillover to other family members and dyads. For example, mothers in the father high conflict group reported less acceptance of their children, and older siblings reported less acceptance from mothers. These findings suggest that much more research is needed to understand the role of fathers in two parent African American families and their impact on adolescent development.

The emergence of a profile characterized by younger siblings’ reports of more frequent conflict relative to other family members may reflect the developmental stage of these adolescents given that parent–adolescent conflict generally peaks in early adolescence and declines in later adolescence (Shanahan et al. 2007). Beyond group and family member differences, our findings also suggested that reporter effects were in operation. That is, adolescents’ reports of conflict with their mothers and their fathers were highly correlated as were parents’ reports of their conflicts with older and younger siblings. Such a pattern underscores the need to collect data on parent–child relationships from both relationship partners.

Importantly, the three family groups did not vary in family background characteristics including parents’ income, youth’s age, family size, and sibling gender constellation suggesting that such “third variable” did not explain the family group differences that we observed. Given the links between harmonious parent–adolescent relationships and positive youth outcomes (Dotterer et al. 2014), future research should examine factors such as youth and parent personality characteristics and well-being or sociocultural factors such as cultural values and stressors that might contribute to family patterns in perceptions of conflict.

Family Profiles of Conflict and Youth Well-Being

Drawing from a family systems perspective which suggests that parent–adolescent discrepancies reflect problems in the family (Minuchin 1985) and with prior research on differences between parents’ and adolescents’ reports of their relationship experiences (e.g., Ehrlich et al. 2015; Smetana et al. 2003), we expected to find differences in youth’s adjustment as a function of differences between parent and youth reports. Our findings were somewhat more nuanced. First, the direction of parent–youth discrepancies appeared to make a difference: Although youth in the younger sibling high conflict family exhibited more adjustment problems, those in the father high conflict profile did not differ from those in the low conflict profile in their reports of depression or risky behaviors. The overall level of family level of conflict also may play a role: Although youth in the low conflict profile—who had the lowest levels of depressive symptoms and risky behavior—reported significantly less conflict than their parents, the average level of conflict in these families also was lower than that of other groups. Whereas most prior research on parent–youth discrepancies in relationship reports has examined dyads in isolation, our LPA approach allowed us to capture family patterns of parent–youth conflict, and our results suggest that viewing parent–youth discrepancies within their larger family context provides new insights on the implications of discrepancies for youth adjustment.

Underscoring the significance of the larger family context, our findings also suggested that dyadic conflict experiences may spill over to have implications for other dyadic family relationships. For example, older siblings and mothers in the father high conflict profile reported lower maternal–but not paternal—acceptance than those in the low conflict group. And, along these same lines, both older and younger siblings in the younger sibling high conflict group reported more depressive symptoms and risky behaviors than youth in the low conflict group.

In contrast to these negative spillover effects, younger siblings in the father high conflict group reported more acceptance from fathers than younger siblings in the low conflict group, and their fathers reported higher levels of acceptance overall as compared to fathers in the younger sibling high conflict group. Although this group was characterized by high conflict reports by fathers, younger siblings in this group reported lower levels of conflict with their fathers than did their older siblings, and fathers in this group also reported less conflict with their younger as compared to their older child. Thus, it is possible that within-family social comparisons may have operated to engender younger siblings’ perceptions of an especially accepting relationship with their fathers. Fathers’ reports of higher levels of acceptance as compared to those of fathers in the younger sibling high conflict group also may mean that fathers in the father high conflict group were more engaged with their children overall—which also could help to explain why, in the face of discrepancies between father and youth relationship perceptions, neither older nor younger siblings in this group reported elevated rates of individual adjustment problems. Such an interpretation suggests that future pattern analytic research should incorporate multiple dimensions of parent–youth relationships to better understand the implications of discrepant perspectives.

In the face of our study’s contributions to the literature on parent–adolescent relationships, some limitations imply directions for research. First, our sample was located in one geographic region within the U.S. and not representative of the larger population of two-parent African American families in the U.S. Future research should be directed at recruiting nationally representative samples of two-parent African American families to further examine normative family dynamics in this sociocultural group, especially the roles of fathers and potential differences between the implications of mother- and father- youth relationships for youth adjustment. Second, our study was limited to examining 3 years of longitudinal data on youth adjustment as a function of the family profiles. Significant effects involved cross-time average family profile differences, and a direction for research is to collect longer term longitudinal data to learn more about how family patterns of discrepancies in parent–youth relationship perceptions have their implications over the longer term. In addition, youth and parents provided reports of conflict and also reported on the well-being and relationship outcome measures, raising the possibility that mono-reporter bias might explain some of the linkages we observed. More objective measures such as in home observations of parent–child relationships or reports from other family members are important directions for future research. Finally, families were assigned to a profile based on their modal profile membership, and modal profile membership was subsequently used in the MLM analyses. Recent research suggests that such standard classify-analyze approaches can result in attenuation of the relations between latent profile membership and distal outcomes however, methods that provide for more precise grouping estimates and are suitable for MLM analyses with continuous distal outcomes have not yet been developed (Lanza and Cooper 2016; Vermunt 2010). In the face of this concern, we were able to detect family profile differences in youth well-being, but it will be important to replicate and extend our use of a pattern analytic approach to advance understanding of the implications of discrepancies between parents’ and youth’s dyadic relationship experiences.

Conclusion

As with other dimensions of family and parent–youth relationships, parents and adolescents often disagree about the extent of conflict in their relationships. This study shed new light on the nature and adjustment implications of discrepancies between parents’ and youth’s views of conflict by studying dyadic relationship experiences within their larger family context. Supporting tenets of family systems theory and recent research suggesting that discrepant perspectives reflect family problems such as poor communication (Ehrlich et al. 2015; Minuchin, 1985), we found differences among adolescents in depression, risky behaviors and parental acceptance as a function of family patterns characterized by higher levels of parent–adolescent conflict and youth’s reports of higher levels of conflict relative to parent reports. Importantly, our study of an ethnically homogeneous sample of two parent, African American families allowed us to capture variability in normative processes among African American families. Further, we extended the research on African American fathers’ relationships with their adolescent children and its implications for adolescent’s adjustment and relationships.

In the face of substantial interest and discussion (e.g., Cox and Paley 1997), family systems constructs have been difficult to operationalize, and the empirical literature on family systems processes remains limited. In this study, incorporating the experiences of two parents and two siblings from each family provided a window into how families operate as systems to affect youth wellbeing.