Mexican-origin Mothers’ and Fathers’ Involvement in Adolescents’ Peer Relationships: A Pattern-Analytic Approach

Abstract

Using latent profile analysis, this study examined patterns of mother-father involvement in adolescents’ peer relationships along three dimensions, support, guidance, and restrictions, in 240 Mexican-origin families. Three profiles were identified: (a) High Mother Involvement (mothers higher than fathers on all three dimensions); (b) High Support/Congruent (mothers and fathers reported the highest levels of peer support and similar levels of guidance and restrictions); and (c) Differentiated (more guidance and restrictions by fathers than by mothers, similar levels of parent support). These profiles were linked to mothers’ and fathers’ familism values, traditional patriarchal gender role attitudes, and socioeconomic status, and to adolescents’ friendship intimacy and risky behaviors measured longitudinally from early to late adolescence. Adolescent gender moderated the linkages between parents’ involvement in adolescents’ peer relationships and youth adjustment.

Mexican-origin Mothers’ and Fathers’ Involvement in Adolescents’ Peer Relationships: A Pattern-Analytic Approach

Parents’ involvement in their children’s peer relationships is linked to child and adolescent friendship quality and peer social competence (e.g., Ladd & Pettit, 2002; Updegraff, McHale, Crouter, & Kupanoff, 2001), and to adolescent adjustment, more generally (e.g., Mounts, 2004). Highlighting parents’ significance in their offspring’s peer experiences, Parke and Buriel’s (1998) tripartite model points to parents’ role as “managers of their children’s social lives” (p. 468), as a conceptually distinct aspect of parent-child relationships. Extant research likewise documents the ways that parents shape their children’s peer experiences, such as providing support or advice about peer relationships or facilitating involvement with peers (Ladd & Pettit, 2002; Parke et al., 2003). In the face of research and theoretical attention, we still know very little about fathers’ involvement in their children’s peer relationships, and almost nothing about the coordination of mothers’ and fathers’ involvement, including the sociocultural characteristics that give rise to coparenting in this domain, or how coparenting around peer socialization is related to youth adjustment.

A family systems perspective directs attention to the interrelated subsystems that comprise families (Cox & Paley, 1997; Minuchin, 1974) and underscores the importance of examining the extent to which mothers and fathers coordinate their child rearing activities and engage in effective coparenting (Feinberg, 2003; J. P. McHale et al., 2002). Consistent with family systems premises are investigations of patterns in mothers’ and fathers’ parenting that reveal connections between inter-parental congruence in childrearing and youth adjustment (e.g., Fletcher, Steinberg, & Sellers, 1999; Kan, S. M. McHale, & Crouter, 2008). We extend this line of work to examine a specific domain of co-parenting, parents’ involvement in adolescents’ peer relationships, focusing on a sample of Mexican-origin families. Given that Mexican Americans comprise the majority of Latinos and are the largest and fastest growing minority group in the U.S. (U.S. Census Bureau, 2010), investigation of family dynamics in this group is timely.

Our first goal was to identify profiles of Mexican-origin mothers’ and fathers’ involvement in young adolescents’ peer relationships along three dimensions: support, restrictions, and guidance. This approach extends prior, variable-oriented research that has examined single dimensions of maternal or paternal involvement in youths’ peer relationships, by studying how multiple dimensions operate together in youths’ everyday experiences. Grounded in a cultural-ecological perspective (García Coll et al., 1996; McAdoo, 1993) and research on the ecology of coparenting (Feinberg, 2003), our second goal was to examine how profiles of mother-father involvement were linked to sociocultural factors, including familism values, traditional patriarchal gender role attitudes, and socioeconomic status (SES). Finally, informed by family systems and gender socialization perspectives, our third goal was to explore how profiles of parent involvement in early adolescence were linked to adolescent psychosocial adjustment over a five-year period, from early to late adolescence.

Profiles of Mother-Father Involvement in Adolescents’ Peer Relationships

Research on parents’ involvement in adolescents’ peer relationship builds on conceptual frameworks of parental involvement in young children’s peer relationships (Ladd & Pettit, 2002; Parke & Buriel, 1998). Ladd and colleagues conceptualized four dimensions of parent involvement in children’s peer relationships (Ladd & Pettit, 2002). First, as designers, parents select proximal settings, such as schools and neighborhoods, that can promote peer relationships, though little research has examined this dimension of parenting. Second, as mediators, parents help their children meet peers, choose friends, and form friendships. In adolescence, when negative peer influences increase, scholars have focused on parents’ guidance, restrictions, and prohibitions in examining this dimension of their peer involvement (e.g., Mounts, 2004; Updegraff, Kim, Killoren, & Thayer, 2010). Third, as supervisors parents directly intervene in peer interactions among young children, and provide more distal supervision and monitoring in later childhood and adolescence. Fourth, as advisors and consultants parents provide support and advice regarding peer relationships (e.g., Mounts, 2002; Tilton-Weaver & Galambos, 2003). In this study, our investigation of Mexican-origin parents’ support, guidance, and restrictions reflects a focus on (a) peer involvement via direct interactions with adolescents (advice, mediation) rather than more distal processes (design, supervision), and (b) dimensions highlighted in qualitative work on parents’ management of peer relationships in Latino families (Brown, Alvarez, & Quijada, 1999).

What remains unknown is how mothers and fathers coordinate their socialization in this important domain of youth development. As noted, family systems researchers emphasize the need to understand the interrelations among different family subsystems (Cox & Paley, 1997), and particularly, how parental figures function together as coparents in childrearing (J. P. McHale et al., 2002; Minuchin, 1974). Empirical work substantiates the theoretical significance of coparenting as a distinct feature of families that is linked to parent and youth adjustment (Feinberg, 2003; J. P. McHale et al., 2002). In this study, we used a pattern-analytic approach (Magnusson, 1988), which allows for the identification of distinct profiles across multiple dimensions of mothers’ and fathers’ involvement with peers. Based on prior research (e.g., Fletcher et al., 1999; S.M. McHale, Crouter, McGuire, & Updegraff, 1995), we anticipated that a congruent pattern, with mothers and fathers displaying relatively similar levels of support, guidance and restrictions, would emerge.

Research also highlights differences in mothers’ versus fathers’ parenting roles, suggesting that some parents may display complementary patterns of involvement (J. P. McHale et al., 2002). Consistent with evidence that Latino mothers are more involved in caregiving than fathers (Azmitia & Brown, 2002), one complementary pattern may involve mothers assuming the primary role in managing adolescents’ peer relationships and fathers displaying lower levels of peer socialization. Another possibility is that parents’ traditional patriarchal gender role orientations will be reflected in gender differentiated roles, with mothers as caregivers and fathers as authority figures. In this scenario, parents’ peer involvement might take the form of mothers providing support and fathers taking on a controlling role (restricting, guiding). Research in this area is limited, however, and other patterns also may be evident. Thus, our goal was to identify distinct family-level profiles of Mexican-origin mothers’ and fathers’ peer involvement.

Sociocultural Correlates of Mother-Father Involvement with Peers

Our second goal was to explore the sociocultural correlates of patterns of mother-father involvement with peers and to test whether these correlates differed for families of girls versus families of boys. The rationale for this goal is grounded in a cultural-ecological perspective (García Coll et al., 1996; McAdoo, 1993), which emphasizes the substantial variability among Mexican American families in characteristics ranging from socioeconomic resources to cultural orientations and values (Baca Zinn & Wells, 2000), and in Feinberg’s (2003) ecological model of coparenting, which highlights parents’ characteristics as predictors of coparenting. This study advanced research on the sociocultural correlates of parenting in two ways. First, we moved beyond the examination of proxy measures of culture, such as nativity, to consider how specific cultural values, namely familism and traditional patriarchal gender role attitudes (referred to as traditional gender attitudes hereafter), were linked to parents’ profiles of peer involvement. Such an approach provides for a direct test of the hypothesized mechanisms through which ethnic group status has implications for parenting. Second, our study design enabled us to investigate both between-family differences, as reflected in mother-father averages or family-level cultural values as well as within-family (mother versus father) differences in parents’ cultural values as correlates of their patterns of peer involvement. Consideration of cultural values as a family-level dynamic is rare, but has the potential to provide insights about families as systems compromised of members with potentially different values/beliefs (e.g., Ferree, 1990).

Although there is substantial variability among Mexican American families (Baca Zinn & Wells, 2000), strong ties to Mexican culture are associated with more traditional gender role attitudes that specify distinct roles for mothers versus fathers (e.g., Leaper & Valin, 1996). Thus, parents’ traditional gender role attitudes may give rise to complementarity in mothers’ and fathers’ socialization in terms of their adolescents’ peer relationships. Complementarity may be reflected in higher overall levels of involvement by mothers as compared to fathers or by differentiated parental roles wherein mothers provide support and fathers take on guidance and limit setting activities. We anticipated that profiles characterized by gender differentiated parental involvement, as compared to congruent inter-parental involvement, would be associated with more traditional attitudes in mothers and/or fathers that are consistent with their division of parenting.

Familism values, another key aspect of Mexican American culture, refer to beliefs about interdependence, support, and obligations among family members (Cauce & Domenech-Rodríguez, 2002; Sabogal, Marín, Otero-Sabogal, Marín, & Perez-Stable, 1987). One possibility is that parents who display congruent patterns of involvement with peers, compared to those in complementary profiles, also may endorse stronger familism values because they are on the same page in terms of their family-oriented values and their strategies for managing adolescents’ involvement with peers. A second possibility is that parents who display more traditional gender role attitudes, and thus complementary profiles of peer involvement, also may describe higher levels of familism values because of their strong ties to traditional Mexican cultural values. Thus, in the case of familism values, we anticipated between-family (mother-father average) differences across peer involvement profiles, and considered these analyses as exploratory given these two competing predictions.

We also considered the role of family SES. First, family SES was examined as a sociocultural correlate of mother-father profiles of peer involvement. Extant research suggests that SES is associated with parents’ support for and facilitation of activities outside the home (Lareau, 2003), and thus, parents in profiles characterized by high levels of mother-father support may have more socioeconomic resources than parents in other profiles. Second, scholars note that culture and SES are often confounded in studies of ethnic minority families (Knight, Roosa, & Umaña-Taylor, 2009; McLoyd, 1998), underscoring the importance of controlling for SES when testing the links between parents’ cultural values and peer involvement profiles.

We examined the moderating role of adolescent gender to assess whether the links between profiles of mother-father involvement with peers and sociocultural correlates differed for families with daughters versus sons. Theory and prior research suggest that mothers and fathers take a special interest in the socialization of their same-sex offspring in early adolescence (Hill & Lynch, 1983; Updegraff et al., 2001), and that such preferences are pronounced in traditional family contexts (Crouter, Manke, & McHale, 1995). As such, we anticipated that gender differentiated profiles of parents’ peer involvement would be associated with more traditional gender attitudes for mothers of girls as compared to mothers of boys and for fathers of boys as compared to fathers of girls. For family SES and parents’ familism values, we viewed these moderation analyses as exploratory and did not advance specific hypotheses.

Profiles of Mother-Father Involvement and Trajectories of Adolescent Adjustment

The third goal was to examine how profiles of mother-father involvement in adolescents’ peer relationships were linked to youth’s psychosocial adjustment, as measured by friendship quality, depressive symptoms, and risky behaviors, from early to late adolescence.

In adolescence, peer relationships can provide benefits to youth, including emotional support and companionship (Way & Chen, 2000), and they also can place youth at risk, such as through exposing them to deviant activities (Barrera et al., 2001). Extant research shows that parents’ efforts to manage adolescents’ peer relationships are associated with youth’s friendship quality (e.g., Updegraff et al., 2001) and problem behavior (e.g., Mounts, 2004). Given that early adolescence marks the onset of substantial changes in youth psychosocial functioning, including increases in internalizing and externalizing symptoms (Zahn-Waxler, Shirtcliff, & Marceau, 2008) and in reliance on peers as sources of support (Berndt, 2004), we examined whether parents’ involvement in adolescents’ peer relationships was associated with (a) average levels of and (b) changes over time from early to late adolescence in youth adjustment.

Two conceptual perspectives informed our ideas about how profiles of parent involvement would be linked to youth adjustment. First, drawing from a family systems perspective, congruence in mothers’ and fathers’ parenting is thought to reflect cooperation within the marital subsystem and appropriate generational boundaries within the family. Consistent with these ideas, a growing body of research shows that coordination and support between adults who share parenting roles is linked to more positive family dynamics and youth well-being (J. P. McHale et al., 2002), including in Hispanic families (Lindahl & Malik,1999). Thus, one possibility is that adolescents whose parents exhibit congruent patterns of peer involvement will report closer relationships with friends and fewer adjustment problems over time than adolescents whose parents differ in their peer involvement.

A second possibility emerges from a gender socialization perspective, which underscores the importance of same-gender parents in family socialization processes in early adolescence (Galambos, Berenbaum, & McHale, 2009; Hill & Lynch, 1983). Among European American families, one investigation showed that mothers’ involvement in adolescents’ peer relationships was more strongly linked than fathers’ to the qualities of girls’ peer relationships, and that only fathers’ involvement was associated with boys’ peer relationships (Updegraff et al., 2001). Based on this work, we would expect girls to report more positive trajectories of friendship quality and adjustment when mothers are highly involved, and that boys would benefit from paternal involvement. To examine this possibility, we tested the moderating role of adolescent gender in in the links between profiles of parent peer involvement and youth adjustment.

Method

Participants

The data came from a study of family dynamics and adolescent development in Mexican American families (Updegraff, McHale, Whiteman, Thayer, & Delgado, 2005) including 246 families recruited in 2002 and 2003 from a southwestern metropolitan area. Interest in how gender, culture, and family socialization was linked to adolescent adjustment led to the focus on 7^th graders as the target youth, as early adolescence marks the onset of significant change in youth adjustment (Zahn-Waxler et al., 2008). Criteria for participation were as follows: (a) 7^th graders and at least one older adolescent sibling lived at home; (b) biological mothers and biological or long-term adoptive fathers lived at home (for a minimum of 10 years); (c) mothers were of Mexican origin; and (d) fathers worked at least 20 hours/week. Although not a criterion, 93% of fathers also were of Mexican descent. The present analyses focus on parents’ roles in the 7^th graders’ peer relationships as data were not collected on parents’ involvement in older siblings’ peer relationships.

Families of 7^th graders were recruited from public junior high schools in five school districts and from five parochial schools. Schools were selected to represent a range of socioeconomic situations as indicated by the proportion of students eligible for free/reduced lunch (i.e., 8% to 82% across the schools). To recruit families, letters and brochures describing the study (in both English and Spanish) were sent to families, and follow-up telephone calls were made by bilingual staff to determine eligibility and interest in participation. Letters were sent to 1,856 families with a Hispanic 7th grader who was not learning disabled. For 396 families (21%), the contact information was incorrect and repeated attempts to find updated information were unsuccessful, and 146 (10%) refused to be screened for eligibility. Eligible families included 421 families (32% of those we were able to contact and screen for eligibility). Of those who were eligible, 284 (or 67%) agreed to participate, 95 (23%) refused, and we were unable to re-contact the remaining 42 families (10%). Interviews were completed by 246 families.

At Phase 1, families represented a range of socioeconomic levels, with annual household income ranging from $5,000 to over $250,000 per year and including a range of $5,000 to $32,000 for the lower third of the sample and $58,000 to over $250,000 for the upper third. The median income was $41,000, and the average income was $54,000 (SD = $45,222), with 18.3% of the sample meeting federal poverty guidelines. Parents’ educations ranged from less than 6^th grade to advanced graduate degrees (e.g., JD, MD, Ph.D.), with an average of about 10 years of education (M = 10.34; SD = 3.74 for mothers, and M = 9.88; SD = 4.37 for fathers). Parents born outside the U.S. (70% of mothers and fathers) resided in the U.S. an average of 12.4 (SD = 8.9) and 15.2 (SD = 8.9) years for mothers and fathers, respectively. Adolescents (51% female) were more likely to be born in the U.S. (62%) and averaged 12.8 years of age (SD = .58). Two thirds of parents and 17% of adolescents completed the survey in Spanish.

Comparisons of the sample relative to Mexican-headed two-parent families in the county from which the sample was drawn (U.S. Census Bureau, 2000) revealed that similar percentages of families met federal poverty guidelines (18.3% versus 18.6%, respectively) and that similar percentages of mothers (19.1%) and fathers (22.4%) completed high school relative to Mexican female (22.1%) and male adults (20.3%) in the county. Median household income was slightly higher in our two-parent sample ($41,000) relative to the county ($32,000). In terms of immigrant status, 59% of family members in the sample (70% of parents and 38% of adolescents), compared to 42% of Mexican Americans in the county, were foreign-born. Finally, 65% of the households in the sample, compared to 76.6% of households with Mexican-origin individuals over age 5 in the county, spoke Spanish.

At Phase 2, 91% of the adolescents participated. There were no significant differences in the background characteristics of adolescents who participated in Phase 2 versus those who did not. Phase 3 interviews were conducted with all four family members three years after Phase 2 when target adolescents were 18 years of age; over 75% of the families participated (n = 184). Those who did not participate: could not be located (n = 43), had moved to Mexico (n = 2), could not presently participate or were difficult to contact (n = 8), or refused to participate (n = 8). Non-participating families at Phase 3 (n = 62) compared to participating families (n = 184) reported significantly lower income at Phase 1 (M = $37,632; SD =$28,606 for non-participating families and M = $59,517; SD = $48,395 for participating families) and lower maternal education (M = 9.48; SD = 3.45 for non-participating families and M = 10.62; SD = 3.80 for participating families). Thus, income and education were accounted for in longitudinal analyses. For the present analyses, two families were excluded due to missing data, and four families were excluded because fathers were not born in the U.S. or Mexico.

Procedures

Data were collected during in-home interviews that lasted approximately three hours for mothers and fathers and two hours for adolescents at Phase 1. Individual interviews were conducted using laptop computers by bilingual interviewers (who were primarily Latino) with questions read aloud due to variability in family members’ reading levels. At Phase 2 (two years after the initial interview), target adolescents were re-contacted and invited to participate in a one-hour interview over the phone using the same procedures as in the home interviews at Phase 1, with interviewers reading items and entering adolescents’ responses into the computer. At Phase 3 (five years after Phase 1), family members participated in home interviews again using the same procedures as Phase 1. Informed consent was obtained prior to the interviews at all phases. Families received $100 and $125 at Phases 1 and 3, respectively, and adolescents received $40 at Phase 2. The University Institutional Review Board approved all procedures.

Measures

All measures were forward translated into Spanish and back translated into English by two separate individuals. Discrepancies were reviewed and resolved by the research team. For all subscales below, higher scores indicate higher levels of the construct of interest.

Cultural background (Phase 1)

Parents reported on their nativity (1 = U.S.-born; 2 = Mexico-born) and number of years living in the U.S. (for Mexico-born parents). To account for parent nativity at the family-level, we coded U.S.-born = 1 (i.e., both parents were U.S.-born; 25%) versus immigrant = 2 (at least one parent was born in Mexico; 75%).

Parents’ involvement in adolescents’ peer relationships (Phase 1)

Mothers and fathers rated their support, restrictions, and guidance in their adolescent’s peer relationships using a measure adapted for Mexican-origin parents (see Updegraff et al., 2010). The support subscale (8 items) represented parents’ interest in and support for adolescents’ peer relationships (e.g., “I encourage my child to invite friends over to the house”). The restrictions subscale (4 items) tapped parents’ efforts to minimize involvement with peers (e.g., “I arrange activities with family members so that my child can’t spend as much time with friends”). The guidance scale (4 items) measured parents’ efforts to provide direction and guidance (e.g., “I talk to my child about the pros and cons of hanging around with certain people”). All items were rated on a 5-point scale ranging from 1 (Not at all) to 5 (Very often). Items were averaged to form the three subscale scores separately for mothers and fathers. Cronbach’s αs were .83 and .84 on the support subscale, .78 and .73 on the restrictions subscale, and .68 and .62 on the guidance subscale, for mothers and fathers, respectively.

Family SES (Phase 1)

Mothers’ and fathers’ reports of their education levels in years completed (e.g., 14 = 2 years beyond high school) and their annual household income were used to create a family SES score, as income and parents’ education levels were positively correlated (r = .43, p < .01 for mothers, and r = .45, p < .01 for fathers). After transforming income to correct for skewness, we created the family SES score by standardizing and averaging mothers’ and fathers’ education levels and household income (α = .78).

Cultural values (Phase 1)

Mothers’ and fathers’ familism values and traditional patriarchal gender role attitudes were assessed using subscales of the Mexican American Cultural Values Scale (Knight et al., 2010). The familism subscale includes 16 items (e.g., “It is always important to be united as a family”), and the traditional patriarchal gender role attitudes scale includes 5 items (e.g., “Men should earn most of the money”). For both subscales, items were rated on a 5-point scale from 1 (Strongly disagree) to 5 (Strongly agree) and averaged to create subscale scores for each parent. Cronbach’s αs were .80 and .85 for mothers’ and fathers’ familism, and .71 and .65 for mothers’ and fathers’ traditional gender role attitudes, respectively.

Adolescent psychosocial adjustment (Phases 1, 2, and 3)

We assessed three indicators of psychosocial adjustment: friendship intimacy, depressive symptoms, and risky behaviors. First, adolescents described the degree of intimacy they experienced with their closest same-sex friend (e.g., “How much do you go to ‘friend name’ for advice or support?”) using an 8-item index developed by Blyth and Foster-Clark (1987); adequate reliability and validity has been demonstrated with Mexican American youth (Updegraff, Madden-Derdich, Estrada, Sales, & Leonard, 2002). Response choices were on a 5-point scale, ranging from Not at all to Very much. Cronbach’s alpha was .83, .88, and .84, for Phases 1 – 3, respectively. Next, we used the Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977) to measure depressive symptoms in the past month. On this 20-item measure, adolescents rated the frequency of cognitive, behavioral, and affective symptoms on a 4-point scale, ranging from 1= Rarely or none of the time to 4 = Most of the time (αs = .85, .87, and .89, for Phases 1 through 3, respectively). Finally, adolescents rated the frequency with which they engaged in each of 23 risky behaviors in the past year (e.g., skip a day of school, got drunk or high) on a 4-point scale (Never to More than 10 times) using a measure developed for ethnically diverse youth (Eccles & Barber, 1990). Cronbach’s αs were .92 for Phase 1, .89 for Phase 2, and .88 for Phase 3.

Results

The results are organized around our three research goals: (a) to identify family-level profiles of mothers’ and fathers’ support for, guidance of, and restrictions on adolescents’ peer relationships; (b) to examine sociocultural correlates of mother-father profiles in terms of family SES and parents’ cultural values; and (c) to investigate how profiles of parents’ peer involvement were linked to adolescent psychosocial adjustment from early to late adolescence. For the second and third goals, we also tested the moderating role of adolescent gender.

Goal 1: Identifying Mother-Father Profiles of Peer Involvement

We used latent profile analysis, a type of mixture modeling, to identify profiles of parent involvement in adolescents’ peer relationships. An advantage of mixture modeling is that it uses a model-based procedure to determine the optimal profile structure in the data and provides measures of statistical fit, both of which are useful in determining the optimal number of profiles (Whiteman & Loken, 2006). In addition, mixture modeling assigns probabilities of group membership to each case. The sensitive nature of these probabilities is another benefit of mixture models over other person-oriented methodologies (Whiteman & Loken, 2006).

Latent profile analyses were conducted using data from 240 families at Phase 1 using Mplus version 6.1. Mothers’ and fathers’ ratings of peer support, restrictions, and guidance were entered into the latent profile analysis, and parent nativity status was included as a control variable to account for within-sample heterogeneity.

We retained three profiles based on the following criteria: (a) the fit criteria, which in combination, supported the three-profile solution (i.e., lowest BIC and significant LMRT; see Table 1); (b) each latent profile had an adequate sample size; (c) the solution made intuitive sense; and (d) the solution was determined to be sufficiently stable via replication with different start values. As noted, LPA assigns each family a probability of membership in each profile. Ideally, each family has a high probability of being in one profile and a low probability of being in the other profiles. In this analysis, the average probabilities for the most likely latent profile memberships were high for the three-profile solution (i.e., .85, .88, and .92, respectively), indicating that each family in our sample fit clearly within one of the profiles. Because the average probabilities were high, the model had an entropy score of .86, and thus, it was appropriate to use a classify-analyze approach (Clark & Muthén, 2010). Specifically, we used the resulting LPA probabilities to assign families to their highest probability profile and computed the remaining analyses in an ANCOVA framework. The ANCOVA approach allowed us maintain the family as the unit of analysis and examine both between-family and within-family differences in our sociocultural correlates.

Table 1.

Fit Criteria and Average Latent Profile Probabilities for One-, Two- Three- and Four-Profile Solutions (N = 240)

Fit Criteria	1 Profile	2 Profiles	3 Profiles	4 Profiles
AIC	3780	3671	3638	3615
BIC	3822	3737	3728	3731
ABIC	3784	3677	3646	3626
LMRT		122.75***	47.40**	36.30
Df		7	7	7

Note. The BIC is considered the best information criteria (IC) for selecting the number of profiles as the AIC can often overestimate the necessary number of profiles (Nylund et al., 2007). The LMRT (Lo-Mendell-Rubin Likelihood Ratio Test) is a nested model test using bootstrapped methodology to replicate a normal distribution in order to compare the model fit of a model solution with k number of profiles against a model solution with k-1 profiles (e.g. comparing a 3-profile solution to a 2-profile solution). A significant p -value indicates the k profile solution improves fit as compared to the k-1 solution.

^*p < .05;

^**p < .01;

^***p < .001.

Means and standard deviations for mothers’ and fathers’ ratings of support, restrictions, and guidance, the observed variables used to create latent profiles, as a function of assigned profile membership can be found in Table 2. To further describe these profiles, we conducted 3 (Profile) × 2 (Parent: Mother versus Father) mixed model ANOVAs with Profile as the between-groups factor, Parent as the within-groups factor, and parents’ ratings of peer involvement as the dependent variables. Significant Profile × Parent interactions indicated within-family (i.e., mother versus father) differences in peer involvement across profiles for support, F (2, 237) = 16.25, p < .01, restrictiveness, F (2, 237) = 29.21, p < .01, and guidance, F (2, 241) = 22.84, p < .01. Our follow-up analyses of these interactions revealed significant differences across profiles in the difference scores (mother minus father) for peer support, guidance and restrictions (see Table 2).

Table 2.

Means (and Standard Deviations) by Latent Profile Membership for Mothers’ and Fathers’ Involvement in Adolescents’ Peer Relationships (N = 240)

	High Mother Involvement (n = 75)		High Support/ Congruent (n = 51)		Differentiated (n = 114)
	Mean	(SD)	Mean	(SD)	Mean	(SD)
Support
Mothers (M)	2.89b	(.83)	3.71a	(.58)	2.96b	(.76)
Fathers (F)	2.16c	(.50)	3.35a	(.52)	2.99b	(.66)
M-F Difference	.73a	(.90)	.36a	(.83)	−.03b	(.94)
Restrictiveness
Mothers (M)	2.76a	(1.09)	2.07b	(.62)	3.00a	(1.12)
Fathers (F)	1.95c	(.57)	2.23b	(.64)	3.46a	(.63)
M-F Difference	.82a	(1.20)	−.16b	(.86)	−.46b	(1.20)
Guidance
Mothers (M)	3.41a	(.85)	3.44a	(.84)	3.55a	(.91)
Fathers (F)	2.83c	(.69)	3.52b	(.78)	4.00a	(.56)
M–F Difference	.57a	(1.03)	−.08b	(1.01)	−.45c	(1.03)

Note. Profile means with different subscripts within a row are significantly different at p < .05.

We labeled the first profile High Mother Involvement (31.25% of the sample), as mothers reported significantly higher levels of peer support, restrictiveness, and guidance compared to fathers. In the second profile, labeled High Support/Congruent (21.25%), both mothers and fathers reported significantly higher levels of peer support than parents in the other two profiles, and did not differ from one another in their peer guidance and restrictions, which were low to moderate in range. In the third profile, Differentiated (47.50% of the sample), fathers reported higher levels of guidance and restrictions than mothers, and parents did not differ in support. A Pearson chi-squared test revealed no adolescent gender differences across profiles, χ² (2, N = 240) = 3.30, ns, but significant differences in parent immigrant status, χ² (2, N = 240) = 143.54, p < .01, such that 34% and 63% of immigrant parents were in the High Mother and Differentiated profiles, respectively, with the remaining 3% in the High Support/Congruent profile.

Goal 2: Mother-Father Peer Involvement Profiles and Sociocultural Correlates

Our second goal was to examine the sociocultural correlates (i.e., SES, cultural values) of these profiles and test the moderating role of adolescent gender in these linkages. Given our study goals, we focused on significant main effects and interactions involving Profile. For family SES, we conducted a 3 (Profile) × 2 (Adolescent Gender) ANOVA and found a significant Profile effect, F (2,237) = 44.58, p < .01. Tukey follow-up tests showed that parents reported higher SES in the High Support/Congruent profile versus the other profiles (see Table 3).

Table 3.

Means (and Standard Deviations) for Family SES and Parents’ Cultural Values as a Function of Profile Membership (N = 240)

	High Mother Involvement (n = 75)		High Support/ Congruent (n = 51)		Differentiated (n = 114)
SESa	−.22a	(.79)	.83b	(.46)	−.24a	(.75)
Familism Values
Mother-Father Average	4.40b	(.33)	4.27a	(.30)	4.55c	(.28)
Gender Role Attitudes
Mother-Father Average	2.97b	(.74)	2.55a	(.57)	3.26c	(.74)
Families of Girls
Mothers (M)	3.03	(1.01)	2.46	(.79)	3.22	(1.06)
Fathers (F)	2.94	(.73)	2.79	(.66)	3.43	(.84)
M – F Difference	.09d	(1.06)	−.33	(1.08)	−.22	(1.19)
Families of Boys
Mothers (M)	2.63	(1.06)	2.56	(.87)	3.15	(.91)
Fathers (F)	3.27	(.86)	2.38	(.84)	3.25	(.97)
M – F Difference	−.64e	(1.06)	.17	(1.12)	−.10	(1.16)

Note: Subscripts a, b, and c indicate significant between-profile differences within a row at p < 05. Superscripts d and e indicate significant mother-father differences for families of boys versus families of girls within the High Mother Involvement profile at p < .05.

^aFamily SES is a standardized score with a mean of zero and a standard deviation of 1.

Next, to test profile differences in parents’ cultural values, we conducted a series of 3 (Profile) × 2 (Adolescent Gender) × 2 (Parent: Mother versus Father) ANCOVAs, with Profile and Adolescent Gender as between-group factors, and Parent as a within-group factor. SES was included as a covariate to account for group differences in SES noted above and to test whether cultural values were associated with profile membership after accounting for SES. Notably, the findings were the same whether or not SES was included as a covariate.

For both parents’ familism values and traditional gender role attitudes, we found significant Profile effects, F (2,232) = 6.79, p < .01, and F (2,232) = 5.10, p < .01, respectively. Follow-up tests showed that all three groups differed significantly from one another, with parents in the Differentiated profile reporting the highest and parents in the High Support/Congruent profile reporting the lowest familism values and traditional gender role attitudes (see Table 3). In addition, for parents’ traditional gender role attitudes, there was a Profile × Adolescent Gender × Parent interaction, F (2,232) = 5.14, p < .01. Follow-up analyses examined mother-father differences in traditional patriarchal gender role attitudes for families of girls versus boys in each of the three profiles. Significant differences emerged only in the High Mother Involvement Profile, such that families of girls were characterized by more traditional mothers than fathers, but families of boys were characterized by more traditional fathers than mothers (see Table 3).

Goal 3: Mother-Father Peer Involvement Profiles and Adolescent Adjustment Over Time

To address our third goal, we investigated the connections between profiles of mother-father involvement in peer relationships and adolescents’ reports of friendship intimacy, depressive symptoms, and risky behaviors, and we tested the moderating role of adolescent gender. Specifically, we conducted 3 (Profile) × 2 (Adolescent Gender) × 3 (Time) mixed model ANCOVAs with SES as a covariate using Proc MIXED in SAS 9.2 with full maximum likelihood estimation to manage missing data in the dependent measures without inflating the models’ standard errors (Wolfinger & Chang, 1998). Proc MIXED also allowed us to examine the overall effects of the mother-father involvement profiles and the moderator on adolescent adjustment, as well as patterns of change in adjustment over time. First, each model was estimated with time as a fixed effect. If the model showed a significant effect for time, then a second model was estimated in which time was added to the CLASS statement to create contrast codes for time, allowing us to test mean differences in the outcome variable at each time point. Dependent variables were measured at Phase 1 (7th grade), Phase 2 (9^th grade), and Phase 3 (12^th grade). Below, we describe the highest order effects involving the Profile factor.

Beginning with friendship intimacy, there was a Profile × Adolescent Gender × Time interaction, F (2, 233) = 2.49, p < .05. Follow-up analyses revealed significant increases in friendship intimacy from middle to late adolescence for boys in the High Support/Congruent, F (2, 21) = 7.37, p < .01, and Differentiated groups, F (2, 61) = 19.21, p < .001, and for girls in the High Support/Congruent group, F (2, 26) = 7.07, p < .01 (see Table 4). In contrast, no significant effects were found for adolescents’ depressive symptoms. For adolescents’ risky behaviors, there was a significant Profile × Adolescent Gender interaction, F (1, 235) = 4.85, p < .05. Follow-up analyses examining differences in risky behavior (averaged across the three points in time) revealed that boys in the High Support/Congruent profile (M = 1.74, SE = .07) reported significantly more risky behaviors, on average, than boys in the High Mother Involvement (M = 1.53, SE = .06), and Differentiated groups (M = 1.41, SE = .04), F (2, 114) = 7.51, p < .01. No significant profile differences emerged for girls; their risky behaviors were M = 1.30 (SE = .06) in the High Support/Congruent profile, M = 1.35 (SE = .04) in the High Mother Involvement profile, and M = 1.43 (SE = .04) in the Differentiated profile.

Table 4.

Comparison of Adjusted Means (and Standard Errors) by Latent Profile, Adolescent Gender, and Grade for Adolescents’ Friendship Intimacy

	7th Grade		9th Grade		12th Grade
	M	(SE)	M	(SE)	M	(SE)
Boys
High Mother Involvement	3.36	(.11)	3.38	(.12)	3.62	(.15)
High Support/Congruent	3.67ab	(.13)	3.41a	(.12)	3.78b	(.08)
Differentiated	3.24a	(.08)	3.17a	(.08)	3.67b	(.09)
Girls
High Mother Involvement	4.02	(.07)	4.08	(.12)	4.22	(.10)
High Support/Congruent	4.21a	(.10)	4.16a	(.10)	4.52b	(.07)
Differentiated	4.22	(.07)	4.22	(.07)	4.18	(.08)

Note. Means within a row with different subscripts are significantly different at p < .05.

Discussion

This study drew on family systems, cultural-ecological, and gender socialization perspectives to explore patterns of mothers’ and fathers’ involvement in adolescents’ peer relationships in two-parent Mexican-origin families. Our findings contribute to existing scholarship in several ways. First, the identification of distinct profiles of mother-father involvement in adolescents’ peer relationships extends research on coparenting and on parents’ involvement in peer relationships in new directions. To our knowledge, this study is the first to document within-family similarities and differences in mothers’ and fathers’ involvement in adolescents’ peer relationships across multiple dimensions (i.e., support, restrictions, guidance). Second, our findings revealed that both SES and parents’ cultural values were associated with mother-father patterns of peer involvement; these findings underscore the role of the larger sociocultural context in parents’ involvement with peers and provide a novel look at inter-parental coordination of parenting within a cultural group. Third, mother-father involvement with peers in early adolescence predicted changes in friendship quality across adolescence and average levels of involvement in risky behaviors during adolescence differentially for girls versus boys, highlighting the significance of adolescents’ gender in parenting processes.

Profiles of Mother-Father Involvement and Parents’ Sociocultural Characteristics

Evidence of distinct profiles of mothers’ and fathers’ peer involvement underscored the value of systemic approaches to identifying family dynamics that extend beyond the level of individual dyads (Cox & Paley, 1997; Minuchin, 1974). Further, our findings revealed that different patterns of parents’ peer involvement were linked to distinct sociocultural characteristics, consistent with a cultural-ecological framework (García Coll et al., 1996; McAdoo, 1993) and research documenting substantial variability among Mexican American families (Baca-Zinn & Wells, 2000). Contrary to research with European American families where congruent patterns of coparenting are most common (Fletcher et al., 1999; S.M. McHale et al., 1995), most families in this sample were characterized by complementary patterns of mother-father involvement with peers. Given that complementary patterns exemplified gender differentiated parenting roles, and more traditional patriarchal roles are linked to stronger ties to Mexican culture (Leaper & Valin, 1996), it is not surprising that complementary patterns were predominant (79%) in this largely immigrant sample of Mexican-origin parents.

The first complementary pattern, the Differentiated profile, was characterized by a gendered division of parenting with fathers as authority figures. Specifically, fathers in this profile reported higher levels of restrictions and guidance than mothers. Consistent with their strong ties to Mexico (63% were Mexico-born), these parents endorsed the most traditional patriarchal gender role attitudes and highest familism values (Leaper & Valin, 1996; Sabogal et al., 1987), after accounting for SES. Together, these findings illuminate the sociocultural context in which this differentiated pattern of parents’ peer involvement emerged. The more limited socioeconomic resources that characterized this group may mean that parents have fewer resources to support adolescents’ peer involvement and that families may be living in more dangerous neighborhoods. As such, fathers’ control over adolescents’ peer relationships may be attributed to the risks and limitations imposed by socioeconomic circumstances and to the expression of culturally embedded values.

The second complementary pattern, the High Mother Involvement Profile, was characterized by gender differentiated roles emphasizing mothers as primary caretakers (Azmitia & Brown, 2002), as mothers displayed significantly higher levels of support, guidance, and restrictions relative to fathers. Like parents in the Differentiated profile, these parents had more limited socioeconomic resources but relatively strong familism values; one third of these parents were born in Mexico. Unique to this profile was within-family (mother versus father) differences in traditional patriarchal gender role attitudes. In particular, mothers of daughters reported more traditional patriarchal attitudes than did fathers of daughters and fathers of sons reported more traditional patriarchal attitudes than did mothers of sons in this group. Thus, a pattern of mothers being primarily responsible for socialization around peers was most common when adolescents’ same-gender parents held relatively more traditional gender attitudes. Many studies of family gender dynamics have relied on between-family comparisons, but our findings point to the value of incorporating a within-family component. In this way, our multi-informant (mother, father, adolescent) design is an important strength of this study.

The third profile, High Support/Congruent, included mothers and fathers who displayed the highest levels of support for peer relationships and who were relatively similar in their guidance and restrictions. This was the smallest of the three profiles and included families with the most socioeconomic resources and the fewest immigrant-born parents (3%). Further, parents in this profile reported significantly lower familism values and less traditional gender role attitudes, on average, than parents in the other two groups, after accounting for SES. The high support for peer relationships exhibited by this group of parents may be a result of their greater socioeconomic resources, to the extent that these resources are associated with safer living environments and financial support for adolescents’ involvement with peers, as well as their lower emphasis on family-oriented values and weaker ties to Mexico.

Our findings take a first step in investigating how different dimensions of parents’ involvement are interrelated using a pattern-analytic approach (Magnusson, 1988), with the measures of peer involvement falling within parents’ roles as mediators and advisors/consultants (Ladd & Pettit, 2002). It will be important to extend this pattern-analytic approach to incorporate other dimensions of peer involvement, including parents’ roles as designers and supervisors (Ladd & Pettit, 2002). Given the associations between mother-father patterns of peer involvement examined here and parents’ sociocultural characteristics, it will be particularly important to examine parents’ roles as designers, or parents’ efforts to shape peer involvement via their selection of proximal settings (e.g., neighborhoods, extracurricular activities) that influence their adolescents’ opportunities for peer relationships. Consistent with a cultural-ecological perspective, our findings also revealed that both socioeconomic resources and parents’ cultural values differentiated the profiles of peer involvement, such that two profiles were characterized by families with fewer socioeconomic resources in combination with stronger familism values and more traditional patriarchal gender role attitudes relative to parents in the third profile, who had more economic resources but less traditional cultural values. As other scholars have argued (García Coll et al., 1996; McLoyd, 1998), SES and cultural factors are intertwined in the lives of U.S. ethnic minority families and must be considered simultaneously to understand family dynamics and youth development.

Mother-Father Patterns of Involvement and Adolescent Psychosocial Adjustment

Patterns of mother-father involvement were associated with adolescent adjustment, but in different ways for girls versus boys. We posited, from a family systems perspective (Cox & Paley, 2003), that congruence in mother-father involvement with peers may be associated with more positive friendship quality and adjustment. Findings linking patterns of mother-father involvement with peers to trajectories of girls’ friendship intimacy were consistent with this idea. Specifically, girls in the High Support/Congruent profile reported significantly higher intimacy with their same-sex best friend in late adolescence, as compared to early and middle adolescence, but no significant changes in friendship intimacy were found for girls in the other two profiles. The combination of high levels of mother and father support, congruence in parents’ control over peer relationships that was in the moderate to low range, and higher socioeconomic resources, may provide an ideal context to support the development of girls’ close friendships in adolescence in this family and cultural context. Although profiles of parent involvement were not directly linked to girls’ (or boys’) depressive symptoms, the possibility that friendship quality serves as a mediating process linking parent involvement to more positive future adjustment deserves consideration as a next step.

The findings for boys were more complex. Similarly to the results for girls, boys in the High Support/Congruent profile reported increases in friendship intimacy during middle to late adolescence, but these same boys also reported the highest levels of risky behaviors. Together, these findings suggest the importance of exploring the characteristics of boys’ close friends in future work on parents’ involvement with peers, as boys in these families may be developing close ties with peers who are engaging in deviant activities, and as a result, engaging in more risky behaviors themselves. In addition, high levels of support, greater socioeconomic resources, and moderately low restrictions may mean that boys have more freedom and less supervision when spending time with their friends. The combination may allow boys to develop close friendships, but also place them at risk for misconduct.

From a gender socialization perspective, we anticipated associations between the same-sex parents’ peer involvement and adolescent adjustment (Galambos et al., 2009; Hill & Lynch, 1983). Consistent with research on European American families (Updegraff et al., 2001), we found some evidence that fathers’ involvement was important for boys. Specifically, in the Differentiated profile, wherein fathers displayed the highest levels of guidance and restrictions and moderately high support, boys reported increases in friendship intimacy from middle to late adolescence. To the extent that fathers’ multifaceted involvement (i.e., combination of support, restrictions, and guidance) can be interpreted as interest in boys’ peer relationships, boys may be better able to develop friendships. An important next step will be to begin to identify mediating processes linking parents’ peer involvement to youth’s psychosocial adjustment.

Limitations, Future Directions, and Conclusions

This is one of the first studies to examine coparenting and its sociocultural correlates among ethnic minority families, but it is not without limitations. First, we measured parents’ involvement in adolescents’ peer relationships at a single point in time. Thus, illuminating processes of change in parents’ involvement in peer relationships and how such changes are linked to changes in adjustment is an important future direction. Second, we focused on three dimensions of parents’ peer involvement and it will be important to extend pattern-analytic approaches to include other dimensions of parental involvement in their children’s peer relationships. In addition, our measures of guidance and restrictions included only four items each, which may have contributed to lower reliabilities, and also suggest the need to expand these measures in future work. Third, the generalizability of our findings is constrained by our sample’s characteristics (i.e., predominantly immigrant, two-parent families), and it will be important to explore these mother-father dynamics in families in different sociocultural contexts.

In closing, our findings of distinct inter-parental profiles and correlates suggest the importance of moving beyond universal characterizations of Mexican-origin families to identify how variability within this cultural group is related to variability in parenting and youth adjustment. Further, by examining multiple dimensions of culture from the perspectives of both mothers and fathers, we were able to provide a deeper understanding of the role of culture in family dynamics. Investigating such within-group variability in culture and parenting dynamics provides a window into the processes underlying development and adjustment of the growing U.S. population of ethnic minority youth and their families.