Parent-Youth Differences in Familism Values from Adolescence into Young Adulthood: Developmental Course and Links with Parent-Youth Conflict

Abstract

A critical step in capturing family processes is to incorporate the perspectives and experiences of multiple family members toward characterizing how families operate as systems. Although some research has examined differences between parents' and youth's family experiences, most studies have focused on European American families, and we know little about the nature and implications of divergent parent-youth experiences in other ethnic groups. Accordingly, we focused on Mexican-origin families and assessed the links between mother-youth and father-youth differences in familism values and parent-youth conflict from early adolescence into young adulthood. Participants were mothers, fathers, and two siblings (248 female and 244 male; M_age = 14.02 years) from 246 families who were interviewed in their homes on three occasions over eight years. We operationalized parent-youth differences in familism values using difference scores, controlling for mean levels of familism. Multilevel models revealed that mothers' and fathers' familism values remained relatively stable over time, but youth's (51% female) familism values declined until age 17, stabilized, and then increased slightly in young adulthood. Lagged models tested directions of effect by examining whether parent-youth differences in familism values predicted parent-youth conflict or vice versa. The findings revealed that parent-youth conflict predicted greater differences in parent-youth familism values, but differences in familism values did not predict conflict. Our findings align with a family systems perspective in documenting the significance of differences between family members' perspectives and highlighting that such processes are dynamic. Further, by testing bidirectional associations in longitudinal models, we were able to disentangle the temporal ordering of differences in familism values and parent-youth conflict thereby advancing understanding of parent-youth discrepancies in cultural values.

Introduction

Parents and youth often have different perspectives regarding their “shared” family experiences, and these different perceptions provide unique information about family functioning (Steinberg, 2001; Telzer, 2010). Despite this empirical evidence, many studies fail to include more than one perspective on family life (Collins, 1990; Smetana, 1988). From a theoretical perspective, a critical step in capturing family processes is to incorporate the perspectives and experiences of multiple family members in an effort to characterize how families operate as systems (Minuchin, 1974). One such line of research has been directed at examining differences between parents' and youth's perceptions of their dyadic relationship experiences, primarily in European American families. This work shows that parents and youth express distinct perspectives on family life, that adolescents typically perceive the family more negatively than do their parents (Laursen & Collins, 2009; Mooney, Laursen & Adams, 2006; Smetana, Villalobos, Tasopoulos-Chan, Gettman, & Campione-Barr, 2009) and that greater discrepancies between parents' and youth's reports are often linked to more negative adjustment outcomes for youth (Gaylord, Kitzmann, & Coleman, 2003; Pelton & Forehand, 2001).

The current study built on this line of research by testing whether differences between parents' and youth's familism values were linked to the frequency of conflict in parent-youth relationships in Mexican-origin families. Familism values, which reflect a strong emphasis on family support, solidarity, and loyalty, are a critical component of Mexican American culture (Cauce & Domenech-Rodriquez, 2002; Tsai, Telzer, Gonzales, & Fuligni, 2015; Zeiders, Updegraff, Umaña-Taylor, McHale, & Padilla, 2016). Further, parents' and youth's reports of familism values are associated (using cross-sectional data) with more positive family relationship dynamics (Fuligni, Tseng, & Lam, 1999; Germán, Gonzales, & Dumka, 2009; Updegraff, McHale, Whiteman, Thayer, & Delgado, 2005; White, Roosa, & Zeiders, 2012; Zeiders, Updegraff, Umaña-Taylor, McHale, & Padilla, 2016), but we know little about discrepancies in parents' and youth's familism values and their implications for family dynamics. To advance understanding of parent-youth discrepancies in ethnic minority families, our study addressed two goals. Using longitudinal data, the first goal was to chart changes in mothers', fathers', and youth's familism values from adolescence into young adulthood in an effort to illuminate the nature of cross-generational discrepancies in familism values. The second goal was to test two models regarding the associations between intergenerational differences in familism values and parent-youth conflict: (a) mother-youth and father-youth differences in familism values as predictors of increases in mother-youth/father-youth conflict; and (b) mother-youth/father-youth conflicts as predictors of increases in intergenerational differences in familism values. We focused on a relatively neglected population in the developmental and family literatures, Mexican-origin youth and their parents. The study of family dynamics in diverse populations can expand our understanding of the complexities of family system processes. Further, Mexican Americans are the largest U.S. immigrant group, and they comprise two-thirds of Latinos, the largest and youngest ethnic minority group in the U.S. (U.S. Census Bureau, 2014), making them an important focus for study of family systems dynamics in their own right.

Development of Familism Values

Familism values are central to Latino culture (Cauce & Domenech-Rodríquez, 2002). Prior work demonstrates, for example, that Latinos endorse stronger values for family support and obligations-- such as providing economic and emotional support to kin and meeting familial expectations-- than youth from Asian and European backgrounds (Hardway & Fuligni, 2006; Telzer & Fuligni, 2009; Sabogal, Marín, Otero-Sabogal, Marín, & Perez-Stable, 1987). Although prior cross-sectional research documents that familism values are stronger for first-generation relative to third-generation immigrant youth (Fuligni, Tseng, & Lam, 1999), we know almost nothing about how familism values change across the course of individual development. To the extent that familism values change at different times and rates for individual family members, however, the development of familism values may contribute to increasing or decreasing discrepancies between family members' values, with implications for family relationship quality.

Theory and research suggest that youth experience declines in parental closeness and support across adolescence, with most of this work focused on European American families (Laursen & Collins, 2009; Mooney, Laursen & Adams, 2006). Given a mainstream cultural emphasis on autonomy and the world beyond the family during this developmental period (Steinberg & Morris, 2001), familism values of Mexican-origin youth living in the U.S. may also decline across adolescence. Consistent with this idea, research on acculturation differences between youth and their parents has shown that, while parents tend to more strongly retain their focus on Mexican culture, youth more rapidly adopt U.S. cultural values and practices (Phinney, Ong, & Madden, 2000; Telzer, 2010). Thus, research on acculturation gaps between parents and children suggests that immigrant youth generally acculturate at a quicker pace than their parents (Birman, 2006; Szapocznik & Kurtines, 1993; Telzer, 2010). Most studies of differences in parent-youth cultural orientations and values have examined this family dynamic in childhood and adolescence, however, and longer term longitudinal research is rare. We could find no longitudinal studies of parents' and youth's familism values that charted their development over time.

In addition, within the literature on the parent-youth acculturation gap as well as the literature on discrepancies between parents' and youth's perceptions of their family experiences, more generally, the focus has tended to be on mothers or on generic “parents,” and fathers have been relatively neglected (Telzer, 2010). The traditional gender roles of Mexican-origin families (Galanti, 2003), however, mean that fathers tend to be more involved in the world beyond the home as breadwinners, and thus it may be that mothers and fathers differ in how orientation toward their heritage culture, including their familism values, change over time. Accordingly, to shed new light on the familism values of Mexican-origin youth and their parents, we used an ethnic homogeneous design to illuminate within-group differences in patterns of change over time (García Coll et al., 1996), the first goal of this study was to chart changes over an 8-year period in mothers', fathers' and youth's familism values. Toward this end, we followed families from the time youth were in adolescence through early adulthood, expecting that mothers' and fathers' values would remain relatively stable, but that youth's values would decline over time, resulting in larger intergenerational differences across time.

Links between Differences in Parent-Youth Familism Values and Parent-Youth Conflict

The acculturation literature proposes that differences between parents' and youth's acculturation levels increase immigrant youth's risks for well-being and adjustment problems (Portes & Rumbaut, 1996), including family conflict (Telzer, 2010). Sluzki (1979) explained that youth's clashes with parents around cultural practices and values –such as familism values--may lead to heightened conflict. For instance, parent-youth disagreements on the centrality of the family in everyday activities, the importance of family responsibilities, and the level of support that family members owe one another (Portes & Rumbaut, 1996 Portes & Rumbaut, 2001; Sabogal et al., 1987), may give rise to parent-youth discord and conflict. Indeed, familial obligations, one dimension of familism values, have been deemed one of the greatest sources of conflict between parents and youth from Latin American families (Zhou, 1997).

With respect to parent-youth differences in familism values, prior work on Mexican-origin families focuses primarily on the mother-youth acculturation gap (Elder et al., 2005; Lau et al., 2005; Liu et al., 2009) and, as noted above, we know much less about the role of fathers in this process (Martinez, 2006; Pasch et al., 2006). In addition, as noted, in many studies, mother-youth and father-youth differences in cultural orientations are often not distinguished, with the seeming assumption that maternal and paternal acculturation processes are similar, if not identical (Telzer, 2010). Importantly, however, when mothers and fathers in Mexican-origin families have both been studied, father-youth, but not mother-youth acculturation discrepancies were linked concurrently to parent-youth conflict, particularly when overall relationship quality was poor (Schofield et al., 2008). Studying adolescents in immigrant Chinese families using a cross-sectional design, Costigan and Dokis (2006) also found that the correlates of parent-youth discrepancies varied across parents, such that father-youth differences in cultural values were linked to both parent-youth conflict and youth depression, but the links for mother-youth differences were not significant. Prior research on parent-youth conflict also highlights the role of gender, including the finding that, although conflicts with mothers are more frequent, those with fathers are more salient to youth (Fuligni, 1998; Maccoby, 1999). Most prior research has been cross-sectional, but development may play a role in these patterns. From a gender intensification perspective, for example, family gender socialization becomes more pronounced across adolescence (Galambos, Berenbaum & McHale, 2009). Traditional gender roles in Mexican-origin families that highlight the status of fathers (Galanti, 2003) may mean that conflict with fathers is increasingly influential as youth transition across adolescence into young adulthood. Accordingly, in this study, we built on the limited literature on gender in parent-youth dynamics, using a longitudinal design to examine linkages between parent-youth differences in familism values and conflict and the roles of parent and youth gender in these patterns.

We also expanded on prior research by examining the direction of effects linking differences between parents' and youth's familism values and parent-youth conflict. In previous research on both the acculturation gap and youth's familism values, the focus has been on how these cultural orientations give rise to more or less positive family relationship dynamics (Fuligni et al., 1999; Juang et al., 2007; Lau et al., 2005). From a family systems perspective, however, we would expect that these processes may be bidirectional (Minuchin, 1974). For example, given that conflict in Mexican-origin family relationships may be seen as inconsistent with familism values (Marin & Marin, 1991), such conflict may lead to decreases in youth's familism values, and thereby, larger parent-youth discrepancies. Accordingly, to test the direction of effect, we estimated lagged models in the context of a longitudinal design to assess whether intergenerational differences in familism values predicted increases in parent-youth conflict (controlling for prior conflict), and whether conflicts between parents and youth predicted differences in parent-youth familism values (controlling for prior differences in familism values). Importantly, in the parent-youth familism differences predictor models, we also controlled for the average of youth's and parents' individual familism values to ensure that parent-youth differences in values, not levels of familism values, were the operative factors. Finally, we tested youth gender as a potential moderator of links between parent-youth familism values and parent youth conflict to determine whether similar patterns emerged for mothers and fathers with their sons and daughters. A social learning perspective highlights the influence of high status models, and the more traditional gender roles that characterize Mexican-origin families, which follow a patriarchal structure in which fathers have higher status (Galanti, 2003) may mean that experiences with fathers are more significant for both sons and daughters than are those with mothers.

The Current Study

The goals of the study were: (a) to chart the longitudinal trajectories of Mexican-origin mothers', fathers', and youth's familism values from adolescence to young adulthood, and (b) to examine the associations between differences in mother-youth and father-youth familism values and parent-youth conflict. Research on intergenerational differences in acculturation (Telzer, 2010) led us to expect that differences between parents' and youth's familism values would increase over time, as parents' levels of values remained stable while youth's levels declined. Further, based on research on discrepancies between parents' and youth's relationship evaluations and literature on parent-youth differences in cultural orientations (Schofield, Parke, Kim, & Coltrane, 2008), we expected that larger differences between parents' and youth's familism values (i.e., youth reporting lower familism values than their parents) would be linked to higher levels of parent-youth conflict. As noted, expanding on prior literature, we also used our longitudinal data to test whether parent-youth conflict predicted differences in familism values, providing insights about the direction of effects. Finally, we also tested youth gender as a potential moderator of these mother-youth and father-youth relationship patterns.

Method

Participants

The data came from a longitudinal study of 246, two-parent Mexican-origin families (McHale, Updegraff, Shanahan, Crouter, & Killoren, 2005). Participating families were recruited through schools in a southwestern metropolitan area. Given the goals of the larger study, the criteria for participation at Time 1 were that: (a) family membership included a seventh grader, at least one older adolescent sibling, a biological mother and a biological or adoptive father figure (all non-biological father figures had lived with the target children for at least 10 years), all living together; (b) mothers were of Mexican origin (93% of fathers also were of Mexican origin, although this was not a study criterion); and (c) fathers were employed for pay for at least 20 hours/week. Mothers, fathers, and older and younger siblings were interviewed in their homes on three occasions over eight years (in 2002-2003-2007-2008, and 2009-2010). To recruit families, letters in English and Spanish were sent to families, and follow-up telephone calls were made by bilingual staff to determine eligibility and interest in participation. Families' names were obtained from five school districts and five parochial schools. Schools were selected to represent a range of socioeconomic situations, with the proportion of students receiving free or reduced lunch varying from 8% to 82% across schools.

At Time 1 (T1), families represented a range of education and income levels. The percentage of families that met federal poverty guidelines was 18.3%, a figure similar to the 18.6% of two-parent Mexican-origin families living in poverty in the county from which the sample was drawn (U.S. Census Bureau, 2000). The median family income was $41,000 (SD = $45,381; range = $3,000 to over $250,000). Mothers and fathers had completed an average of 10 years of education (M = 10.34, SD = 3.74; M = 9.88, SD = 4.37, respectively). Most parents had been born outside the United States (70%); this subset had lived in the U.S. an average of 12.4 (SD = 8.9) years (mothers) and 15.2 (SD = 8.9) years (fathers). About two thirds of the parents were interviewed in Spanish. With respect to youth, 248 were female and 244 were male. Older siblings were 15.48 (SD = 1.57) years old, on average, 47% were born in Mexico, and 82% were interviewed in English. Younger siblings were 12.55 (SD = .60) years of age, on average, 38% had been born in Mexico, and 83% were interviewed in English.

Time 2 (T2) interviews were completed when older siblings averaged 20.65 (SD = 1.56) years of age, and younger siblings, 17.72 (SD = .57) years. Time 3 (T3) interviews were completed when older siblings averaged 22.57 (SD = 1.57) years of age, and younger siblings, 19.6 (SD =.66) years. The majority of youth (88% of older and 89% of younger siblings) were interviewed in the same language at all three phases.

We used a range of strategies to promote retention across the 8 years of the study, including a hands-on approach to data collection via the home interviews, which gave families close contact with project staff. In addition, we mailed newsletters and postcards every 3-4 months and followed up with phone calls and visits to homes if families did not return postcards. Retention rates were 75% and 70% for T2 and T3, respectively. Those who did not participate: could not be located (n = 44 at T2; n = 45 at T3), had moved to Mexico (n = 2 at T2; n = 4 at T3), could not presently participate or were difficult to contact (n = 5 at T2; n = 12 at T3), or refused to participate (n = 10 at T2; n =12 at T3). At T2, participating families differed from non-participating families on T1 maternal education (M = 10.62, SD = 3.80 versus M = 9.48, SD = 3.45) and T1 family income (M = $59,517, SD = $48,395 versus M = $37,632, SD =$28,606, respectively). At T3, participating families differed from non-participating families on T1 maternal education (M = 10.75, SD = 3.75 versus M = 9.35, SD = 3.53), T1 paternal education (M = 10.46, SD = 4.37 versus M = 8.49, SD = 4.08), and T1 family income (M = $59,136, SD = $46,674 versus M = $41,635, SD =$39,095). Thus, family socioeconomic status (SES), a composite of parents' education and family income, was a covariate in all analyses.

Procedures

After obtaining informed consent and assent (for youth under age 18), data were collected in home interviews separately with each family member. Interviews lasted an average of 3 hours for parents and 2 hours for youth. Bilingual interviewers read questions aloud to all participants to account for variability in reading levels and entered their response into laptop computers. Families received $100 for in-home interviews at T1 and $125 at T2, and each family member received $75 at T3. The University's Institutional Review Board approved all procedures.

Measures

Youth's and parents' familism values

These were assessed at each of the three time points using the 16-item familism subscale of the Mexican American Cultural Values Scale (MACVS; Knight et al., 2010). Youth and parents used a 5-point scale (1 = strongly disagree to 5 = strongly agree) to rate their agreement on items (e.g., “It is always important to be united as a family”). Items were averaged for each family member, with higher scores indicating stronger familism values. Cronbach's alphas ranged from .82 to .95 for youth and from .77 to .85 for mothers and fathers. We used parents' and youth's individual reports to calculate the measures of differential maternal-youth and paternal-youth familism values. Specifically, youth's reports of familism values were subtracted from their parents' reports of familism values, for mothers and fathers separately. Thus, negatively signed familism difference scores reflected that youth reported higher familism values than their mother or father, whereas positively signed familism difference scores reflected that youth reported lower familism values than their mother or father. As noted, we calculated the mean of parents' and youth's familism values to use as a control variable so that we were able to determine that parent-youth differences explained unique variance in parent-youth conflict beyond the effects of the level of familism values.

The use of differences scores has had a long and sometimes contentious history (Bereiter, 1963; Edwards, 2001; Rogosa & Willett, 1983; Willett, 1987; Zimmerman & Williams, 1982). With relatively few exceptions (e.g. Laird & De Los Reyes, 2013), there is general agreement that the difference score provides an unbiased estimate of a particular kind of within-unit variation (Allison, 1990; Edwards, 2001; Rogosa, Brandt, & Zimowski, 1982; Willett, 1987), and support for its use comes from the developmental literature, which has long focused on within-unit differences. As described in the developmental literature, for two repeated, or within-unit measures of a construct (e.g., person over time; individuals within families), the original variables can be transformed into a difference score and an average variable, which is equivalent to the well-known orthogonal polynomial transformation for two occasions of data (Hertzog & Rovine, 1985; McCall & Appelbaum, 1973) and to the generally preferred polynomial approach to differences as outcome variables in regression (Allison, 1990; Edwards, 2001; Rovine, 1993). One way to see that the average and difference scores provide non-redundant information is by plotting two data points and connecting them with a straight line to indicate the difference between the two measures. Here, for an average value, the possible differences range from large positive to large negative suggesting that the two are not necessarily correlated but that their degree of association is an empirical question.

Another concern is that the difference score is correlated with the component scores, but this is not necessarily the case either because of the mathematical relation based on the inner product of the coding vectors used to create these variables (Strang, 1988). The difference score tends to have a small negative correlation with the first value used in the construction and a small positive value with the second value used in the construction, but typically, these correlations are not large. For instance, for the difference score and the average variable, the transformation equations are:

$$y_{\mathit{\text{dif}}}=-1\ast y_{t1}+1\ast y_{t2}\text{and}{y}_{\mathit{\text{ave}}}=\frac{1}{2}\ast y_{t1}+\frac{1}{2}\ast y_{t2}$$

The inner product is:

$$-1\ast \frac{1}{2}+1\ast \frac{1}{2}=0$$

which implies that the correlation will tend toward 0, meaning again, that the relation between the average and difference scores is an empirical question.

In short, both the mathematical relations and the graphical relations among the different forms of variables establish that the difference score does not necessarily provide redundant information to the average score, and thus, the difference score also tends to provide information that is not redundant with its constituent variables. Further, the difference and average scores are mathematical transformations of the original variables and provide distinctive ways of summarizing the information contained in those variables: Either or both can be more informative than the original variables depending on the research question. The one case where the difference score would provide redundant information with the component scores would be the trivial situation wherein one of two measures has essentially no variance, in which case the difference score would be highly correlated with the remaining variable. Our descriptive results indicate that this is not the case in the current study.

Parent-youth conflict

We assessed the frequency of parent-youth conflict in 11 areas (e.g., choosing activities, social life) using a measure adapted from Smetana (1988). At each time of measurement, youth rated the frequency of conflict with their mother and father on a 6-point scale (1 = not at all, 6 = several times a day) at separate points in the interview. Conflict areas were adapted to be developmentally relevant. For instance, at T1 youth were asked about disagreements regarding “whether or not [you] can have a girlfriend/boyfriend or go out on dates,” whereas in T3 youth responded about “who [you are] dating/married to.” Items were averaged, with high scores reflecting more conflict at each time point. Cronbach's alphas ranged from .76 to .86 for youth.

Family background characteristics

At T1, mothers and fathers reported on their highest level of education on a scale ranging from less than a high school degree (e.g., 10 for 10^th grade) to graduate or professional degree (e.g., 21 for PhD, JD, or MD). Parents reported on their annual income at T1, and scores were log-transformed to correct for skewness. Family SES was the standardized average of mothers' and fathers' education and log-transformed family income. Information on youth age, gender and birth order were also collected from parents at T1.

Analysis Plan

Given the clustered (time within sibling, siblings within families) and unbalanced design (i.e., siblings were assessed at different ages with different intervals between data collection time points), we used a multilevel modeling (MLM) approach in the context of an accelerated longitudinal design (Aiken & West, 1991; Duncan, Duncan, & Hops, 1996; Raudenbush & Bryk, 2002; Singer & Willett, 2003). Another advantage of MLM is that it accommodates missing data, and thus effectively reduces biases and standard errors (Schaefer & Graham, 2002).

We began by charting the development of mothers', fathers', and youth's familism values. To address this first study goal, we used an accelerated longitudinal design in which older and younger siblings were treated as two age cohorts, using younger siblings' age as the metric of time. This approach allowed us to examine different age cohorts over the same data collection period; it is advantageous because short-term longitudinal data points are combined into a single longitudinal growth pattern (Duncan, Duncan, & Hops, 1996). We centered at age 12 (the mean age across all younger siblings at T1) and estimated a saturated means model, in essence using an ANOVA model to estimate the mean pattern with the fewest parameters as possible. Deviance tests comparing the log likelihoods of nested models were used to determine which random variances (i.e., better error structure) to include for each dependent measure.

To address our second goal, we tested two competing models, separately for mothers and fathers: (a) differences in parent-youth familism values predict parent-youth conflict, and (b) parent-youth conflict predicts differences in familism values. Specifically, we tested mother-youth and father-youth differences in familism values as lagged (i.e. Time N-1 predicts Time N), time-varying predictors of mother-youth and father-youth conflict, controlling for prior conflict, as well as whether mother-youth and father-youth conflict, as lagged, time-varying predictors, explained parent-youth differences in familism values on the following occasion, controlling for prior differences in familism values. In these lagged models, the coefficients predict residualized change in each of the dependent variables, thus discounting the interdependence between variables that weakens the ability to make causal inferences as in cross-sectional models. For all models, we used the PROC MIXED procedure in SAS 9.3.

Two, three-level models were estimated for mother-youth and then for father-youth relationships. In the first, parent-youth familism differences predictor models, mother-youth relative difference scores were included as lagged, time-varying, grand-mean centered (i.e., centered at the sample mean) predictors of mother-youth conflict, and father-youth difference scores were included as lagged, time-varying, grand-mean centered predictors of father-youth conflict. In these models, parents' and youth's average levels of familism values were included as lagged, time-varying, grand-mean centered covariates. By including each dyad's average familism values, we were able to test whether differences between parents' and youth's familism values predicted conflict, after accounting for parents' and youth's average level of familism. In addition, parents' and youth's reports of conflict were included as lagged, time-varying, grand mean centered covariates to control for prior levels of conflict in these familism differences-conflict predictive models. Finally, youth age at T1, gender (0 = female; 1 = male) and birth order (0 = older siblings; 1 = younger siblings) were included at Level 2 and tested as potential moderators of the familism values-conflict linkages, and family SES was included at Level 3 as a time-invariant control.

The parent-youth conflict predictor models were similarly structured, but for these, mother-youth conflict was included as a lagged, time-varying, grand-mean centered (i.e., centered at the sample mean) predictor in the mother-youth model, and father-youth conflict was included as a lagged, time-varying, grand-mean centered predictor in the father-youth model. Both the average of and differences in mother- and father-youth familism values also were included as lagged, time-varying, grand-centered controls (i.e., average of and differences in familism values) to account for prior levels of familism and isolate the effects of differences. Further, so that conflict as predictor and familism differences as predictor models were similarly structured, mother and father reported parent-youth dyadic conflict ratings were also included as time-varying, grand-mean centered controls. In addition, gender (0 = female) and birth order (0 = older siblings) were entered at Level 2 and tested as moderators of the conflict-familism linkages, and family SES was entered at Level 3 as a control variable.

Results

Developmental Trajectories of Youth's and Parents' Familism Values

Tables 1 and 2 show the correlations, means, and standard deviations for parent-youth conflict, average familism values, and parent-youth differences in familism values by phase of study. On average, mothers and fathers reported stronger familism values than youth (i.e., difference scores were positively signed) and both conflict reports and familism values were somewhat stable across phases. Correlations between average familism values and differences in familism values tended to be significant but generally low or moderate. Figure 1 shows the trajectories of mothers', fathers', and youth's familism values as a function of youth age. Largely supporting our hypothesis, both mothers', γ = -0.001, SE = 0.00, p = .59, and fathers', γ = 0.001, SE = 0.00, p = .64 familism values remained relatively stable over time, while youth's familism values declined from age 12 until age 17, stabilized and then increased slightly to age 22, cubic effect: γ = -0.001, SE = 0.00, p < .01.

Table 1. Mothers' Correlations, Means, and Standard Deviations (SDs) for Study Variables (N = 246).

		1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.
1.	Parent-youth familism differences (T1)	-	0.37**	0.13*	-0.40**	0.03	0.06	0.23**	-0.03	0.10*	0.08**	0.14**	0.08†
2.	Parent-youth familism differences (T2)	0.11*	-	0.42**	-0.07	-0.27**	0.00	0.14**	0.29**	0.13*	0.08	0.10†	0.16**
3.	Parent-youth familism differences (T3)	0.15**	0.35**	-	-0.01	-0.03	-0.11*	-0.01	0.05	0.07	-0.02	-0.03	0.05
4.	Parent-youth mean familism values (T1)	-0.27**	0.02	0.12*	-	0.37**	0.39**	-0.14**	-0.03	-0.17**	-0.16**	-0.10*	-0.08†
5.	Parent-youth mean familism values (T2)	0.15**	-0.24**	-0.06	0.44**	-	0.49**	-0.06	-0.18**	-0.21**	-0.12*	-0.08	-0.06
6.	Parent-youth mean familism values (T3)	0.07	0.08	-0.03	0.36**	0.57**	-	0.10†	-0.01	0.00	-0.10†	0.07	0.01
7.	Dyadic youth-parent conflict (T1)	0.03	0.09†	0.02	-0.01	-0.02	0.06	-	0.42**	0.52**	0.27**	0.28**	0.31**
8.	Dyadic youth-parent conflict (T2)	-0.14*	0.20**	0.03	0.02	-0.11*	0.09	0.36**	-	0.54**	0.16**	0.24**	0.20**
9.	Dyadic youth-parent conflict (T3)	-0.23	0.07	0.06	0.07	-0.11*	0.08†	0.38**	0.62**	-	0.36**	0.45**	0.48**
10.	Dyadic parent-youth conflict (T1)	0.00	0.01	-0.15**	-0.10**	-0.02	0.07	0.30**	0.37**	0.24**	-	0.42**	0.30**
11.	Dyadic parent-youth conflict (T2)	0.01	0.22**	-0.08	-0.02	-0.14**	0.07	0.34**	0.43**	0.33**	0.49**	-	0.58**
12.	Dyadic parent-youth conflict (T3)	-0.08	0.16**	-0.10*	-0.01	-0.10†	-0.07	0.48**	0.41**	0.45**	0.44**	0.67**	-
	Younger sibling means	0.17	0.25	0.07	4.35	4.27	4.39	2.71	2.26	2.08	2.31	2.23	2.10
	SD	0.64	0.53	0.51	0.33	0.33	0.30	0.87	0.76	0.75	0.77	0.71	0.72
	Older sibling means	0.20	0.27	0.15	4.33	4.27	4.35	2.61	2.01	1.80	2.22	1.97	1.86
	SD	0.70	0.62	0.55	0.37	0.30	0.30	0.85	0.79	0.55	0.85	0.85	0.68

Note. T1 = Time 1; T2 = Time 2; T3 = Time 3; this time metric differs from the more substantively meaningful metric, youth age metric of time used in the core analyses. Mother-younger sibling correlations are reported below the diagonal; mother-older sibling correlations are reported above the diagonal. For dyadic conflict, the reporter is listed first, such that parent-youth conflict is reported by parents and youth-parent conflict is reported by youth.

^†p < .10,

^*p < .05,

^**p < .01,

^***p < .001.

Table 2. Fathers' Correlations, Means, and Standard Deviations (SDs) for Study Variables (N = 246).

		1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.
1.	Parent-youth familism differences (T1)	-	0.53**	0.35**	-0.35**	0.11*	0.22**	0.24**	0.12*	0.20**	0.10**	0.16**	0.28**
2.	Parent-youth familism differences (T2)	0.40**	-	0.44**	-0.02	-0.27**	0.06	0.09†	0.17**	0.11†	0.14**	0.15**	0.23**
3.	Parent-youth familism differences (T3)	0.20**	0.49(*	-	0.08	0.08	-0.06	-0.04	0.00	-0.06	-0.11*	0.09	-0.05
4.	Parent-youth mean familism values (T1)	-0.21**	0.11*	0.15**	-	0.32**	0.34**	-0.18**	0.13*	-0.14**	-0.20**	-0.05	-0.12*
5.	Parent-youth mean familism values (T2)	0.25**	-0.20**	0.03	0.39**	-	0.59**	-0.12*	-0.13*	-0.30**	-0.15**	-0.11*	-0.14*
6.	Parent-youth mean familism values (T3)	0.23**	-0.14*	-0.03	0.39**	0.60**	-	-0.03	0.00	-0.07	-0.04	-0.02	0.04
7.	Dyadic youth-parent conflict (T1)	0.06	0.21**	-0.04	0.01	-0.07	-0.07	-	0.41**	0.46**	0.17**	0.20**	0.32**
8.	Dyadic youth-parent conflict (T2)	-0.06	0.16**	-0.03	0.09†	-0.20**	-0.06	0.37**	-	0.49**	0.06	0.29**	0.24**
9.	Dyadic youth-parent conflict (T3)	-0.16**	-0.01	-0.03	0.15**	-0.13*	-0.02	0.37**	0.66**	-	0.13*	0.27**	0.38**
10.	Dyadic parent-youth conflict (T1)	-0.05	0.01	0.14**	-0.09*	-0.03	-0.02	0.29**	0.13**	0.09†	-	0.42**	0.42**
11.	Dyadic parent-youth conflict (T2)	-0.09†	0.08	-0.02**	-0.01	-0.06	0.06	0.22**	0.29**	0.43**	0.42**	-	0.45**
12.	Dyadic parent-youth conflict (T3)	0.12*	0.22**	0.05	-0.02	-0.01	0.00	0.40**	0.32**	0.37**	0.37**	0.50**	-
	Younger sibling means	0.20	0.33	0.10	4.36	4.31	4.41	2.69	2.10	1.88	2.25	2.11	2.05
	SD	0.65	0.57	0.50	0.34	0.31	0.30	0.99	0.74	0.71	0.83	0.69	0.69
	Older sibling means	0.23	0.33	0.16	4.34	4.28	4.37	2.57	1.98	1.69	2.19	1.91	1.85
	SD	0.74	0.62	0.54	0.36	0.32	0.32	0.94	0.75	0.58	0.89	0.75	0.70

Note. T1 = Time 1; T2 = Time 2; T3 = Time 3; this time metric differs from the more substantively meaningful metric, youth age metric of time used in the core analyses. Father-younger sibling correlations are reported below the diagonal; father-older sibling correlations are reported above the diagonal. For dyadic conflict, the reporter is listed first, such that parent-youth conflict is reported by parents and youth-parent conflict is reported by youth.

^†p < .10,

^*p < .05,

^**p < .01,

^***p < .001.

Figure 1. The trajectories of mothers', fathers', and youth's familism values.

Longitudinal Links between Differences in Parent-Youth Familism Values and Parent-Youth Conflict

Coefficients for fixed effects for the parent-youth familism values predictor models can be found in Table 3. Mother-youth and father-youth differences in familism values were not significant predictors of later parent-youth conflict, as rated by parents or youth. Among control variables, older youth reported less conflict than younger/later-born youth, and youth-reported conflict with fathers was stable over time.

Table 3. Gamma Coefficients (γ) and t-Ratios for Multilevel Models with Parent-Youth Familism Differences Predicting Parent-Youth Conflict.

	Conflict with Mother		Conflict with Father
	γ	t-ratios	γ	t-ratios
Intercept	1. 99	16. 64**	2. 14	16. 99**
Age	-0. 14	-7. 31**	-0. 14	-7. 19**
Family SES	-0. 07	-1. 50	-0. 04	-0. 74
Birth order (0 = older;1 = younger)	-0. 17	-2. 13**	-0. 29	-3. 51**
Gender (0 = boy; 1 = girl)	0. 06	0. 83	0. 08	1. 16
Prior youth conflict report	-0. 31	-7. 28	-0. 22	-5. 30**
Prior parent conflict report	0. 20	5. 30	0. 10	2. 44
Prior average parent-youth familism	-0. 02	-0. 17	0. 09	0. 84
Prior familism difference	-0. 04	-0. 73	0. 02	0. 39

^*p < .05.

^**p < .01.

In contrast, the parent-youth conflict predictor models (see Table 4) revealed that youth-reported mother-youth and father-youth conflict predicted increases in mother-youth and father-youth differences in familism values, respectively, controlling for both prior parent-youth average levels and prior parent-youth differences in familism values. In addition, father-reported conflict, but not mother-reported conflict, predicted greater parent-youth differences in familism values on the following measurement occasion (i.e., a within-person effect). Among control variables, higher family SES was related to smaller differences between parents' and youth's familism values.

Table 4. Gamma Coefficients (γ) and t-Ratios for Multilevel Models with Parent -Youth Conflict Predicting Differences between Parents' and Youth's Familism Values.

	Mother-Youth Familism Difference		Father-Youth Familism Difference
	γ	t-ratios	γ	t-ratios
Intercept	0. 03	0. 11	0. 24	2. 10*
Age	0. 00	0. 12	-0. 00	-0. 24
Family SES	-0. 13	-3. 36**	-0. 19	-4. 33**
Birth order (1 = younger)	-0. 04	-0. 65	-0. 09	-1. 35
Gender (1 = girl)	0. 03	-0. 56	-0. 07	-1. 27
Prior average parent-youth familism	-0. 10	-1. 22	-0. 06	0. 62
Prior familism difference	-0. 40	-7. 70**	-0. 40	5. 82**
Prior parent conflict report	-0. 02	-0. 78	0. 10	2. 75**
Prior youth conflict report	0. 08	2. 66**	0. 06	2. 13*

Note. Bolded are the significant conflict predictor variables.

^*p < .05.

^**p < .01.

Sensitivity Analyses

We used several statistical checks to test the robustness of our findings. First, to test direction of effects we used lagged models that controlled for the prior levels of the dependent variables to determine whether the predictors explained additional variance. We also controlled for the average of youth and parent familism values in these lagged models to isolate the effects of parent-youth differences in values in predicting conflict. In addition, we assessed the generality of these patterns by collecting data from both mothers and fathers and by testing for youth gender moderation. Finally, in alternate models we tested whether the absolute value of the differences between parents' and youth's familism resulted in a different pattern of findings, and results were the same as those reported here, which were based on relative difference scores.

Discussion

Grounded in a family systems perspective, this study built on prior research on differences between parents' and youth's perspectives on their families (Laursen & Collins, 2009) and on the literature on parent-youth acculturation gaps (Telzer, 2010) to examine familism values and parent-youth conflict in Mexican-origin families. Our study goals were to chart changes in parents' and youth's familism values across eight years, from early adolescence into young adulthood, and assess whether and how (changes in) the differences between the familism values of parents and youth were linked to (changes in) parent-youth conflict. We expanded on prior research by: (a) examining the bidirectional linkages between differences in parents' and youth's familism values and parent-youth conflict using longitudinal data in the context of a multilevel model to address direction of effect; (b) using an ethnic homogenous design to illuminate within-group variations in family processes and their implications in Mexican-origin families (García Coll, et al., 1996); and (c) studying the experiences of daughters and sons vis a vis their mothers and fathers to illuminate the potential role of gender in these family dynamics.

An important step in capturing family systems processes is to incorporate the potentially different perspectives and experiences of multiple family members (Minuchin, 1974). Given the pertinence of cultural values in Mexican-origin families, to better understand the nature and implications of parents' and youth's differential perspectives, we tested whether discrepancies in familism values predicted subsequent intergenerational conflict --and vice versa. Most prior research on acculturation gaps between Mexican-origin parents and youth has been grounded in the idea that youth's more rapid acculturation relative to that of their parents gives rise to increases in parent-youth conflict (Telzer, 2010). By testing bi-directional associations in longitudinal models, we were able to disentangle the temporal ordering of differences in familism values and conflict in multiple subsystems in the family, shedding new light on parent-youth cultural discrepancies.

Our results revealed that conflict in Mexican-origin family relationships predicted increases in parent-youth discrepancies in familism values -- but that discrepancies in values did not predict subsequent parent-youth conflict. Conflict in Mexican-origin family relationships is inconsistent with cultural values that emphasize the centrality of family (Marin & Marin, 1991; Bush & Peterson, 2013), and thus parent-youth conflicts may be a motivating force in decreasing the strength of youth's values in the acculturation process and thereby increasing intergenerational differences in familism values. An important next step is to identify the sources and subjects of parent-youth conflict that are most influential in informing parent-youth discrepancies in cultural values. It may be, for example, that conflicts emerge around youth's increasing interest and engagement in U.S. (mainstream) culture. These findings have important practical implications for parent education programs given prior research documenting the protective effects of strong familism values for adjustment in Mexican-origin youth (e.g., Germán, Gonzales, & Dumka, 2009). To the extent that parent-youth conflict drives the declines in familism values that we observed in our longitudinal analyses, programs aimed at helping parents and youth better manage their differences and disagreements may not only preserve family harmony, but be protective for youth, more generally, to the extent that youth can be supported in maintaining strong familism values.

More generally, our study contributed to an understanding of the familism values of Mexican-origin parents and youth by using longitudinal data collected from mothers, fathers, and youth to describe how familism values changed across time as youth developed across the adolescent years and into young adulthood (i.e., 12 to 22 years of age). Most prior studies of familism values have focused on youth in childhood and adolescence (Fuligni et al., 1999). To our knowledge, this is the first study to examine changes over time in youth's familism values across adolescence and into young adulthood; neither could we find prior research that explored changes over time in parents' (i.e., “midlife adults”) values. Our results revealed that mothers' and fathers' familism values remained stable across this time period. Although subject to acculturation pressures, it appears that by middle adulthood, familism values were established, and that while acculturation may continue, adults' levels of enculturation within these Mexican values regarding family may not shift. In contrast, youth's familism values declined until age 17 and then stabilized in young adulthood. Particularly given the stability of their parents' values in midlife, an important research direction will be to examine whether and how the familism values of Mexican-origin young adults change across the third decade of life, and when they begin to stabilize. More generally, our findings are consistent with prior research on acculturation differences between youth and their parents in demonstrating that parents tend to retain their focus on Mexican culture, while youth may decline in ethnic cultural values over time (Phinney, Ong, & Madden, 2000). Further, our findings provide evidence that parent-youth discrepancies will decrease or stabilize later in development. At the most general level, our findings also align with a family systems perspective in documenting differences in family members' perspectives, highlighting that such processes are dynamic and tied to the development of individual family members.

Our study also contributed to the literature on the acculturation gap in Mexican-origin families by examining discrepancies in youth's values in relation to both their mothers' and their fathers' values. Most prior studies of the acculturation gap have focused on mothers or generic “parents” and have not explicitly focused on mothers and fathers as distinct socialization agents. Including fathers is particularly important given work documenting traditional gender roles in Mexican-origin families and the unique roles of mothers and fathers (Galanti, 2003). Our findings revealed no significant effects for youth gender, but suggested that effects of conflicts with fathers on intergenerational differences in familism values were somewhat stronger than those for mothers. Although between-groups effects were significant for conflicts with both mothers and fathers, effects for fathers were also significant at the within-group level. In other words, controlling for average conflict, when conflicts with fathers were more frequent than usual, the intergenerational difference in conflict was even greater than usual on the next occasion of measurement. Gender norms in Mexican-origin families highlight fathers' status as authority figures and their connections to the world beyond the home in the breadwinner role (Galanti, 2003). Further, prior research has shown that conflicts with fathers tend to be more salient to youth than those with mothers (Maccoby, 1999), and our results were also consistent with prior research that found that father-youth acculturation differences were more closely linked to father-youth conflict than were mother-youth differences (Schofield et al., 2008). In discussions of the role of gender in Mexican-origin families, fathers have been a relatively neglected focus of study. Our findings direct attention to their significance as socialization agents for youth and, taken together, the results underscore the need for future research aimed at examining Mexican-origin youth's potentially different experiences in their relationships with mothers and fathers.

Although our study makes several important contributions to the literature, it is not without limitations. Using an ethnic homogenous design, we aimed to illuminate variability within this sample of Mexican-origin families, but our sample is not perfectly representative of the larger Mexican-origin population in the U.S. Future investigations should include families from other geographic regions and examine families that differ in structure and roles, including mothers' and fathers' work roles, to determine whether similar patterns emerge in the links between intergenerational discrepancies in familism values and youth outcomes and the role of gender in these processes. Additional research also is needed to examine parent-youth differences in other cultural values and practices that may make a difference for youth well-being and family dynamics. Finally, the developmental scope of our study was limited to adolescence and early adulthood, and given the significance of family relationships and supports in Mexican-origin families, future research should examine these dynamics further into adulthood.

Conclusion

The current study advances understanding of a key cultural value within Mexican-origin families and also contributes to the larger literatures on acculturation gaps and family systems processes in parent-youth relationships. Overall, the findings indicated that Mexican-origin mothers' and fathers' familism values were stable over time, but youth's familism values declined until age 17, stabilized, and then increased slightly in young adulthood. Furthermore, higher parent-youth conflict was associated with greater differences in parent-youth familism values, whereas differences in parent-youth familism values did not predict parent-youth conflict. Given changing U.S. demographics, it is important to understand the diversity of family members' experiences within ethnic groups and the cultural factors that impact family functioning (García Coll et al., 1996). From a family systems perspective, examining within family differences in the experiences of family members, in the context of changes over time within individual family members, is equally important for understanding how families operate as social and socializing systems.

Biographies

Jenny Padilla, M.S., is a doctoral student at The Pennsylvania State University (PSU) where she earned a Master's degree in Human Development and Family Studies. Her research focuses on sibling-related family dynamics, including siblings' direct influences on one another such as in their everyday exchanges, and their indirect influence through their effects on larger family processes such as parents' differential treatment (PDT). With respect to the larger contexts of family systems dynamics, she is most interested in the role of culture, and her work has focused on Mexican-origin families.

Susan McHale, Ph.D., is Distinguished Professor of Human Development and Family Studies and Demography, and Director of the Social Science Research Institute at Penn State. She holds a Ph.D. in Developmental Psychology from The University of North Carolina at Chapel Hill. Her research interests include children's and adolescents' family relationships, roles, and everyday activities. Highlighted in her work are sibling relationship dynamics and the family experiences that foster similarities and differences among sisters and brothers as well as the cultural contexts of family dynamics.

Michael Rovine, Ph.D., is Professor of Human Development and Family Studies. He holds a Ph.D. in Educational Psychology from The Pennsylvania State University (PSU). His research interests include the relationships (including equivalences and differences) of models based on the general linear mixed model, including a comparison of repeated measures ANOVA and multilevel growth curve modeling.

Kimberly Updegraff, Ph.D., is Cowden Distinguished Professor of Family and Human Development at Arizona State University. She holds a Ph.D. in Human Development and Family Studies from Penn State University. Her research interests include the role of family and peer relationships (i.e., mothers, fathers, siblings, and close friends) in youth development from early adolescence into young adulthood. Central to her work is understanding the role of gender and culture in adolescents' lives, including the different family experiences of girls and boys and the roles of mothers versus fathers.

Adriana Umaña-Taylor, Ph.D., is Foundation Professor of Family and Human Development at Arizona State University. She holds a Ph.D. in Human Development and Family Studies from the University of Missouri-Columbia. She uses an ecological approach to inform her research, taking into account how individuals and families influence and are influenced by their surrounding ecologies. Her research focuses on ethnic identity formation during adolescence and parent-adolescent relationships.