Siblings’ Differential Treatment in Mexican American Families

Abstract

We investigated the patterns and correlates of parents’ differential treatment of adolescent siblings in 246 two-parent Mexican American families. In home interviews, siblings rated 7 domains of differential treatment (e.g., privileges, chores, warmth) as well as their adjustment and perceptions of parental acceptance and fairness, and both parents and adolescents reported on cultural dynamics. More gender-typed patterns of differential treatment were evident when parents were more oriented to Mexican than Anglo culture. The links between differential treatment and youth reports of adjustment, parental acceptance, and parental fairness were moderated by adolescents’ familism values, particularly for older siblings: Differential treatment was linked more strongly to adjustment and parent-youth relationship problems when youth reported lower levels of familism.

Balancing the different and sometimes competing needs and interests of siblings is a significant childrearing challenge for parents. Because children differ in many ways—such as in their personalities, interests, abilities, and maturity levels—parents may have good reasons for treating their offspring differently. An array of studies, however, has established a link between parents’ differential treatment and both child and adolescent adjustment problems and poor family relationships (e.g., Brody, Stoneman, & McCoy, 1992; Conger & Conger, 1994; Stocker, Dunn, & Plomin, 1989). As Adler explained in his theory of individual psychology (Ansbacher & Ansbacher, 1956), children experience feelings of rivalry, hostility, and low self-esteem when their parents treat a sibling more favorably, and they may act out in an effort to garner their fair share of parental attention and family resources. In contrast, children who experience more favorable treatment develop a more positive sense of self. Grounded in an analytic tradition, the phenomena of differential treatment, sibling comparison, and self-development were proposed originally as universal developmental and family dynamics (Ansbacher & Ansbacher).

A limitation of Adler’s theory and the empirical research that has been conducted more recently on parents’ differential treatment and its implications for siblings is its almost exclusive focus on majority culture (European American) families. In such families, democratic ideals set a standard for equal treatment of siblings (Parsons, 1974/1942), and an ethos of individualism promotes competition among siblings for preferred or favored status (Nuckolls, 1993). A cultural--ecological perspective, however, builds on an ecological framework to outline the ways in which cultural forces shape family beliefs, values, and practices (Spencer, 1995). For instance, the democratic ideal of equal treatment of siblings contrasts sharply with cultural anthropologists’ descriptions of family processes in non-Western societies, where siblings’ family roles are highly differentiated by gender and age (e.g., Nuckolls; Weisner, 1993). These reports suggest that equal treatment of siblings is neither the norm nor the ideal in many societies; thus, examining whether and how parents’ patterns of differential treatment are linked to their cultural orientations is an important topic for research on siblings’ family experiences.

A second component of Adlerian theory and the related neosocial comparison theories that have provided a framework for research on the implications of differential treatment for siblings (e.g., Suls & Wheeler, 2000; Tesser, 1980) is a focus on “ego enhancement associated with being better off than others” (Suls & Wheeler, p. 10). Although portrayed as normative in the context of these theoretical perspectives, a motivation toward self-enhancement at the expense of another group member may be incongruent with the values of communally oriented cultures wherein the needs and interests of the group are emphasized over those of the individual (e.g., Nuckolls, 1993; Weisner, 1993). Examining how values around group affiliation have implications for siblings’ reactions to differential treatment is an important direction for research: The typically observed negative reactions to differential treatment may not be apparent in contexts in which communal rather than individualistic values prevail.

In this study, we investigated the role of culture in differential treatment in two-parent Mexican American families. In so doing, we advance understanding of this important family dynamic and also take a step toward filling the significant gap in our knowledge of normative family and developmental processes in minority families (Hagen, Velissaris, & Nelson 2004; McLoyd, 1998). The Mexican American population is one of the fastest growing segments of contemporary U.S. society (U.S. Census Bureau, 2002), and there is a burgeoning literature on the characteristics and dynamics of these families (e.g., Coltrane & Valdez, 1993; Leaper & Valin, 1996; Phinney & Flores, 2002; Sabogal, Marin, Otero-Sabogal, Marin, & Perez-Stable, 1987). Scholars who have studied the role of culture in family functioning, however, direct attention to the diversity in values, beliefs, and socialization practices within the Mexican American population (Baca Zinn, 1980; Coltrane & Valdez; Sabogal et al.). Studying the diversity of family dynamics within ethnically homogeneous populations is an important end in its own right (e.g., Garcia Coll et al., 1996; McLoyd), and diversity in the extent of families’ cultural orientations and values provides for a natural experiment into how cultural processes operate in family dynamics such as parents’ differential treatment.

In this study, we examined two cultural phenomena, parents’ cultural orientations (i.e., involvement in Anglo and Mexican culture) and adolescents’ familism values (i.e., emphasizing loyalty, support, and interdependence among family members) given their relevance to two distinct aspects of differential treatment dynamics within families. Our study had two goals: (a) to determine whether parents’ cultural orientations were linked to their patterns of differential treatment, specifically the extent to which they treated their daughters and sons differently; and (b) to test whether adolescents’ familism values moderated the associations between differential treatment and adolescents’ adjustment, perceptions of parental acceptance, and ratings of the fairness of their parents’ treatment of themselves relative to their siblings. With respect to the first goal, we tested the hypothesis that Mexican-oriented parents would display more gender-typed differential treatment than Anglo-oriented parents. With respect to the second goal, we tested the hypothesis that stronger familism values would mitigate the negative implications of differential treatment by orienting youth to family needs and concerns and away from an interest in being “better off” than their siblings.

Parents’ Cultural Orientations and Patterns of Differential Treatment

Research into the dynamics of Mexican American families suggests that parents may subscribe to traditional ideals about the social roles of women and men. Consistent with a cultural-ecological perspective, variability in parents’ orientations regarding gender is considerable, however, and is linked to factors including immigration history, connections to Mexico, involvement in Mexican American social networks, and the like (Baca Zinn, 1980; Cauce & Domenech-Rodríguez, 2002; Coltrane & Valdez, 1993). Although parents’ orientations to Mexican culture have been linked to their own traditionality regarding family roles (e.g., Leaper & Valin, 1996; Phinney & Flores, 2002), we know very little about how these play out in parents’ gender socialization practices with offspring, specifically in the extent to which parents treat girls and boys differently.

Research on parents’ differential treatment in European American samples provides some insights about the family conditions under which differential treatment is more or less pronounced. Prior research suggests, for example, that differential treatment of daughters versus sons is more common when fathers have more traditional gender attitudes (McHale, Crouter, & Tucker, 1999). Given extant findings on gender traditionality in Mexican-oriented families, we hypothesized that parents would display more differential treatment toward daughters versus sons when parents were less acculturated to Anglo and more enculturated within Mexican culture.

We also expected that differential treatment of daughters versus sons would be most pronounced in domains relevant to gender socialization. Specifically, we studied parents’ differential allocations of housework responsibilities, privileges, and material resources. Previous work on these dimensions of differential treatment is limited, but research on European American samples suggests that parents tend to grant more privileges to older siblings as well as assign them more household responsibilities (Tucker, McHale, & Crouter, 2003). As noted, however, gender-typed assignment of household tasks (sisters having more responsibilities than brothers) has been documented in European American families with more traditional fathers (McHale et al., 1999), and some work suggests that, particularly in economically disadvantaged families, boys may be the recipients of more of a family’s material resources than are their sisters (Eccles & Harold, 1992).

We also examined differential control and warmth, the domains that have been the target of most research on parents’ differential treatment. Extant literature suggests that birth order, in combination with developmental status, may underlie patterns of differential treatment in these domains in European American samples: In middle childhood, younger siblings tend to be the targets of more discipline and may experience lower levels of warmth, but by adolescence, older siblings are the targets of more discipline and experience relatively less parental warmth (Brody et al., 1992; Kowal & Kramer, 1997; Stocker et al., 1989). To the extent that traditional orientations toward gender—which include more protected and sheltered family roles for women and girls—are pronounced in some Mexican American families (e.g., Azmitia & Brown, 2000), however, parents who are more oriented to Mexican as compared to Anglo culture may display more gender-typed patterns of control and warmth, that is, relatively more control and warmth toward daughters than sons.

Differential Treatment and Adolescent Psychosocial Adjustment

Another cultural dynamic that has been highlighted in research on Mexican American families is captured in the construct of familism or familismo (Fuligni, Tseng, & Lam, 1999; Sabogal et al., 1987). Like the construct of cultural orientation, familism is multifaceted and includes a focus on assistance and support, feelings of respect, and sense of obligation to family members (e.g., Fuligni et al.; Sabogal et al.). With some important exceptions (Fuligni et al.), however, we know little about the implications of familism values for youth’s family experiences and well-being.

In the present study, we investigated the potential role of familism values in siblings’ reactions to parents’ differential treatment. A series of studies documents that negative outcomes of differential treatment are not inevitable and provided a foundation for our hypothesis about the role of familism values in youth’s reactions to this family dynamic. McHale and Pawletko (1992), for instance, found that children whose siblings had disabilities reacted less negatively to differential treatment than children with nondisabled siblings and argued that, when children perceive differential treatment as legitimate or fair, adjustment and family relationship problems may not ensue. Along these lines, Kowal and Kramer (1997) showed that children’s attributions about the reasons for differential treatment moderated the links between this dynamic and sibling relationships quality, and both McHale, Updegraff, Jackson-Newsom, Tucker, & Crouter (2000) and Kowal, Kramer, Krull, and Crick (2002) found that children’s and adolescents’ perceptions of the fairness of differential treatment were more closely linked to their individual adjustment and relationship evaluations than was differential treatment, per se. Taken together, these studies underscore the importance of studying differential treatment in context. Consistent with a cultural-ecological perspective, the implications of differential treatment may not be universal; rather, the meanings youth attribute to differential treatment—and thereby, its psychosocial implications—are likely to vary across family ecologies. In this study, we tested the hypothesis that adolescents’ familism values would mitigate the negative implications of differential treatment in the areas of youth adjustment and parent-adolescent relationship quality by focusing youth’s concerns on the needs of the family group rather than the interests of the self.

METHOD

Participants

The data came from a short-term longitudinal study of family gender socialization and adolescent development in Mexican American families, the Juntos (“Together”): Families Raising Successful Teens Project. The 246 participating families were recruited through schools in and around a southwestern metropolitan area. Given the goal of the larger study, to examine normative family processes in Mexican American families, specifically family gender socialization processes, the criteria for participation were that (a) family membership included a seventh grader and at least one older adolescent sibling living in the home, (b) family membership included both a mother and a father figure living in the home (all nonbiological fathers had lived with the target children for at least 10 years), (c) mothers were of Mexican origin, and (d) fathers were employed for pay for at least 20 hours/week. Although it was not a criterion for participation, 93% of fathers were also of Mexican origin. Importantly, our sampling criteria and our focus on a local population mean that our sample was not chosen to be representative of Mexican American families as a whole. The overall study goals directed our attention to two-parent families so that we could examine the role of fathers, with two siblings in a circumscribed age range so that we could examine the differential experiences of siblings, and pragmatic considerations directed our focus on a local sample.

To recruit families, letters and brochures in both English and Spanish that described the study were sent to families, and follow-up telephone calls were made by bilingual staff to determine each family’s eligibility and interest in participation. Families’ names were obtained from junior high schools in five school districts and five parochial schools. Schools were selected to represent a range of socioeconomic situations, with the proportion of students receiving free/reduced lunch varying from 8% to 82% across schools. Letters were sent to 1,851 families with a Hispanic seventh grader who was not learning disabled. For 438 families (24% of the original roster), the contact information was incorrect and repeated attempts to find updated information through school personnel or public listings were unsuccessful. An additional 42 families (2.4% of the roster) moved between the initial screening and final recruitment contact, and 148 (8%) refused to be screened for eligibility. Eligible families included 421 families (23% of the initial rosters; 32% of those we were able to contact and screen for eligibility). Of those that were eligible, 284 (67% of eligibles) agreed to participate, 95 (23% of eligibles) refused, and we were unable to contact the remaining 42 families (10% of eligibles) to determine whether they would participate. Interviews were completed by 246 families. Those who agreed and were eligible but did not participate in the final sample (n = 38) were families that we were unable to locate at the time of scheduling, those that were unwilling to participate when the interview team arrived at their home, and those that were not home for repeated interview attempts. Because we had surpassed our target sample size (N = 240), we did not continue to recruit these families.

Characteristics of the sample are shown in Table 1. Families represented a range of education and income levels from poverty to upper class. The percentage of families that met federal poverty guidelines was 18.3%, similar to the 18.6% of two-parent Mexican American families who were in poverty in the county from which the sample was drawn (U.S. Census Bureau, 2000). Most parents (about 70% of both mothers and fathers) had been born outside the United States; this subset of parents had lived in the United States an average of 12.4 (SD = 8.9) and 15.2 (SD = 8.9) years for mothers and fathers, respectively. Further, in the case of both mothers and fathers, about two thirds of the interviews were conducted in Spanish. With respect to the target adolescents, the sample was approximately equally divided by the gender constellation of the sibling dyad (68 sister-sister pairs, 55 sister-brother pairs, 57 brother-sister pairs, and 66 brother-brother pairs). Older siblings were 15.7 (SD = 1.6) years of age on average, slightly under half (47%) had been born outside the United States, and 82% were interviewed in English. Their younger siblings averaged 12.8 (SD = .58) years of age, slightly over a third (38%) had been born outside the United States, and 83% were interviewed in English.

Table 1.

Sample Characteristics (N = 246)

	M	SD
Family income	$53,183.00	$45,381.00
Years of educationa
Mothers	10.34	3.74
Fathers	9.83	4.40
Hours worked per week
Mothers	35.96	11.98
Fathers	46.88	11.52
Job prestigeb
Mothers	36.85	12.14
Fathers	37.92	10.88
Age (in years)
Mothers	39.00	4.63
Fathers	41.70	5.77
Older siblings	15.70	1.54
Younger siblings	12.77	.58
No. of children in household	3.79	1.60
Age spacing between siblings	2.93	1.57

^a12 = high school graduate; 16 = college graduate.

^b35 = child-care worker; 40 = car mechanic.

Procedures

Data were collected using two procedures. First, during home interviews, lasting an average of 3 hours for parents and 2 hours for adolescents, parents and adolescents reported on their personal qualities and family relationships. Then, during the 3–4 weeks following the home interviews, families were telephoned on seven evenings and reported on their activities during the day of the call. Bilingual interviewers conducted home and phone interviews in the language of each respondent’s preference. Data for the present analyses came from the home interviews. These interviews began with an overview of the study and review of informed consent procedures. Then, parents and adolescents were interviewed individually. Interview questions were read to parents and adolescents in either English or Spanish given the variability in reading levels for family members.

Measures

All measures were forward and back translated into Spanish (for Mexican dialect in the local area) following the procedures outlined by Foster and Martinez (1995). A third individual of Mexican origin reviewed all final translations, and discrepancies were resolved.

Parents’ differential treatment was rated by adolescents using an adaptation of the Sibling Inventory of Differential Experience (Daniels & Plomin, 1985). Specifically, we used single-item indices of maternal and paternal control and warmth and added three items that tapped parents’ differential allocation of privileges, chores, and monetary resources to the target siblings. On this measure, siblings used a five-point rating scale to describe whether a certain form of treatment (e.g., “Who would you say gets more privileges such as going to a friend’s house, going to parties, staying up or staying out later?”) was experienced by 1 = me a lot more; 2 = me a little more; 3 = both of us the same; 4 = my sister/brother a little more; or 5 = my sister/brother a lot more. Given our interest in the well-being of individual siblings, we used a relative difference store, which reflects which sibling was favored, rather than an absolute difference store, which assesses the total amount of differential treatment. Preliminary analyses revealed that adolescents’ ratings of differential treatment in different domains (privileges, chores and monetary allocations, maternal control, maternal warmth, paternal control, paternal warmth) were not highly correlated (range, r = −.22, p < .01, to r = .19, p < .01). Given this pattern of correlations, we treated the domains of differential treatment separately in the analyses.

Parents’ enculturation/acculturation was measured using the Acculturation Rating Scale for Mexican Americans II (Cuellar, Arnold, & Maldonado, 1995). On this 30-item scale, respondents used a five-point rating scale to show how often (1 = not at all to 5 = extremely often or always) each experience applied to themselves during the past year (e.g., “I like to identify myself as a Mexican American”). The Acculturation Rating Scale for Mexican Americans II includes two subscales, one focused on Mexican and the other on Anglo orientation. In this sample, alpha reliability was high for each subscale, ranging from .75 to .89 for mothers’ and fathers’ reports on the English and from .77 to .85 on the Spanish version of the scales. The two scales also were highly correlated, r = −.50, p < .01, for mothers and r = −.56, p < .01, for fathers. Furthermore, the enculturation and acculturation scores for mothers and fathers were highly correlated, r = .62, p < .01, for mothers’ and fathers’ Mexican orientations, r = .68, p < .01, for mothers’ and fathers’ Anglo orientations. Given this pattern of correlations, we combined the measures of mothers’ and fathers’ enculturation and acculturation scores in the analyses.

Familism values of adolescents were measured with the 17-item familism scale of the Mexican American Enculturation/Acculturation Scale (Gonzales, Knight, & Saenz, 2000). This scale includes three conceptual domains: (a) support/closeness (e.g., “Family provides a sense of security because they will always be there for you”), (b) family obligations (e.g., “Children should be taught that it is their duty to care for their parents when their parents get old”), and (c) family as referent (e.g., “Children should always do things to make their parents happy”). Of the 17 items, 5 were adapted from Sabogal et al. (1987), and the remaining items were developed through focus group work with Mexican American parents and adolescents. Older and younger siblings rated the items using a five-point scale (1 = strongly disagree to 5 = strongly agree); items were averaged to create a scale score for each sibling. Cronbach’s alphas were .87 for English-speaking younger siblings, .79 for Spanish-speaking younger siblings, .89 for English-speaking older siblings, and .92 for Spanish-speaking older siblings. Older and younger siblings’ familism scores were not highly correlated, r = .17, p < .01.

Adolescents’ depression symptoms were assessed using the 20-item Center for Epidemiological Studies Depression Scale (CES-D; Radlolf, 1977). On this measure, adolescents used a four-point rating scale (1 = rarely or none of the time to 4 = all of the time) to describe the frequency of their experiences (e.g., “I had crying spells,” “I felt sad”). The reliability of the English and Spanish versions of the CES-D has been established for Hispanic populations (e.g., Mosicicki, Locke, Rae, & Boyd, 1989), and in our sample, alpha reliabilities ranged from .84 for Spanish-speaking younger siblings to .86 for Spanish-speaking older siblings and English-speaking younger and older siblings.

Adolescents’ risky behavior was measured using an index adapted from Eccles and Barber (1990), which was originally developed for an ethnically diverse sample and parallels measures used in other work with adolescents from a range of cultural backgrounds (e.g., Mason, Cauce, Gonzales, & Hiraga, 1994). On this measure, youth used a five-point rating scale (1 = never to 4 = more than 10 times) to describe how often during the past year they engaged in particular behaviors including, “Done something you knew was dangerous just for the thrill of it?” and “Gotten drunk or high?” Items were summed such that higher scores signified greater involvement in risky behavior. Older siblings’ scores were based on 24 items, but because there was no variability for younger siblings on one item (“How often have you stayed out all night without your parents’ permission?”), younger siblings’ scores were based on 23 items. Cronbach’s alphas were .91 for older English- and Spanish-speaking siblings, .93 for English-speaking younger siblings, and .88 for Spanish-speaking younger siblings.

Parental acceptance was assessed separately for mothers and fathers using an eight-item subscale from the Child’s Report of Parental Behavior Inventory (Schludermann & Schludermann, 1970); this scale has been cross-validated in terms of ethnic and language equivalence on a Hispanic sample (Knight, Tein, & Shell, 1992). On this measure, adolescents rated the frequency with which their mothers and their fathers behaved toward them in particular ways (e.g., “My mother/father understands my problems and worries”) using a five-point rating scale (1 = almost never to 5 = almost always). Items were summed to create separate acceptance scales for mothers and fathers, and alpha reliabilities ranged from .82 to .86 for older and younger siblings who spoke English and .79 to .89 for older and younger siblings who spoke Spanish.

Parents’ fairness was rated by each sibling for each domain of differential treatment we assessed (i.e., privileges, chores, maternal warmth) using a measure developed by McHale et al. (2000). Specifically, after rating the extent of differential treatment in a particular domain, siblings used a three-point scale (1 = not fair to 3 = very fair) to describe “how fair” their parents’ treatment was in that domain.

RESULTS

In a preliminary step, we provide descriptive data on the extent of differential treatment reported by siblings. The remainder of the results is organized around our two research goals. First, to test the hypothesis that parents with stronger orientations to Mexican culture display more gender-typed differential treatment, we compared patterns of differential treatment in families with Mexican- versus Anglo-oriented parents. Next, to test the hypothesis that youth with stronger familism values react less negatively to differential treatment, we examined the links between differential treatment and adolescent adjustment, perceptions of parental acceptance, and ratings of parents’ fairness for youth with stronger versus weaker familism values.

Preliminary Analyses: Patterns of Differential Treatment

We compared older versus younger siblings’ reports of differential treatment in seven domains using a 2 (sibling) × 7 (domain) within-groups analysis of variance (ANOVA); given the correlated error terms in this mixed-model design, we calculated adjusted ds as the measure of effect size. Results revealed an overall significant effect of domain, F (6, 240) = 17.08, p < .01. Follow-up tests showed significantly more differential treatment favoring older siblings in the area of privileges, M = 3.56, SD = .87, as compared to each of the other domains of differential treatment, M = 3.23, SD = .79, (adjusted) d = .37 for differential monetary allocations; M = 3.26, SD = .80, d = .36 for paternal strictness; M = 3.21, SD = .87, d = .32 for chores; M = 3.12, SD = .82, d = .54 for maternal strictness; M = 3.09, SD = .58, d = .62 for maternal warmth; and M = 2.91, SD = .66, d = .81 for paternal warmth. In addition, in the domains of chores and monetary resources, the extent of differential treatment was significantly different (favoring older siblings) from that in maternal and paternal warmth. Finally, differential treatment in the area of paternal warmth differed from all other domains (note that this was the only domain in which the mean differential treatment score was less than 3, i.e., favoring younger siblings). These descriptive data provide a benchmark for understanding the role of culture in parental patterns of differential treatment.

Parents’ Cultural Orientations and Their Patterns of Differential Treatment

Although some researchers have examined parents’ Mexican and Anglo orientations separately (e.g., Phinney & Flores, 2002), other researchers have created a single measure of cultural orientation by subtracting respondents’ scores on the enculturation scale from their scores on the acculturation scale (e.g., Flores & O’Brien, 2002; Parke et al., 2004). As reported above, in our sample, the acculturation and enculturation scores were highly correlated within and between mothers and fathers; thus, we created a single score of parents’ cultural orientations by subtracting the mean of mothers’ and fathers’ enculturation scores from the mean of their acculturation scores. This subtraction process resulted in a group of 180 families in which parents were relatively more oriented to Mexican culture, that is, scores less than 0, and 66 families in which parents were relatively more oriented to Anglo culture, that is, scores greater than 0 (there was no instance in which the subtraction process yielded a score of 0). As Table 2 indicates, the two groups differed significantly on both mothers’ and fathers’ enculturation and acculturation scores, F (1, 242) = 223.30, p < .01, d = 3.07, F (1, 242) = 282.30, p < .01, d = 2.75, for mothers’ enculturation and acculturation scores, and F (1, 242) = 267.52, p < .01, d = 1.65, F(1, 143) = 193.87, p < .01, d = 1.62, for fathers’ enculturation and acculturation scores, respectively. Additional analyses revealed that the two groups were different on other markers of their cultural orientations: 90% or more of mothers and fathers in the Mexican-oriented group were born in Mexico and interviewed in Spanish, whereas in the Anglo-oriented group, 10% or fewer of mothers and fathers were born in Mexico and less than 2% were interviewed in Spanish. The groups also differed significantly on family background variables, including family income; number of children in the home; and age spacing between target siblings; As we describe below, these factors were therefore controlled in the analyses.

Table 2.

Means (SDs) for cultural orientations of mexican-versus anglo-oriented parents

	Mexican Oriented (n = 180)	Anglo Oriented (n = 66)
Mexican orientation
Mothers	4.31 (.45)	3.23 (.64)
Fathers	4.25 (.46)	2.96 (.74)
Anglo orientation
Mothers	2.50 (.73)	4.07 (.38)
Fathers	2.60 (.74)	3.97 (.51)
Parents’ cultural orientationa	−1.73 (.74)	0.93 (.67)

^aNegatively signed score signifies parents’ stronger orientations to Mexican culture; positively signed scores signify parents’ stronger orientations to Anglo culture.

To study the implications of parents’ cultural orientations for differential treatment, we compared siblings’ reports of domains of treatment across these two groups of families. Although parents’ cultural orientation scores were correlated, siblings’ ratings of the seven domains of differential treatment were not; mixed-model analysis of covariance (ANCOVA) allowed us to examine multiple dependent measures in the same analysis, that is, older versus younger siblings’ perceptions of differential treatment and siblings’ perceptions of mothers’ versus fathers’ warmth and control. Further, because we were interested in the role of cultural factors in shaping differential treatment, we controlled for factors that were confounded with cultural orientation group in this sample, that is, family income, number of children, and sibling age spacing.

Specifically, we conducted a series of 2 (culture group: Mexican vs. Anglo) × 4 (sibling gender constellation: sister-sister, sister-brother, brother-sister, brother-brother) × 2 (sibling) mixed-model ANCOVAs, with sibling as a within-groups factor and siblings’ reports of differential treatment as the dependent variables; in the case of siblings’ reports of warmth and control by mothers and fathers, we used a 2 (culture group) × 4 (gender constellation) × 2 (sibling) × 2 (parent) design, with sibling and parent as within-groups factors so that we could also test for mother-father differences in differential treatment. Family income, number of children in household, and sibling age spacing were treated as covariates, and adjusted ds were used as measures of effect size.

Using this design, evidence consistent with our hypothesis of more gender-typed differential treatment of siblings in Mexican-oriented than in Anglo-oriented families is reflected in a Culture Group × Sibling Gender Constellation interaction (i.e., larger differences in the treatment of sisters vs. brothers in the Mexican-oriented group); given our directional hypothesis, we report effects at p < .10 or higher. Evidence of larger differences between mothers and fathers in Mexican-oriented families is reflected in a Culture Group × Parent interaction. Because of the complexity of the design and our interest in parents’ cultural values, we report only the highest order interaction effects involving the culture group factor.

The results of these analyses are shown in Table 3. The predicted Culture Group × Sibling Gender Constellation interaction emerged for three of the five domains of differential treatment, that is, privileges, chores, and monetary resources, and a Parent × Culture effect emerged for differential control. There were no significant effects involving culture group for parental warmth. Beginning with the allocation of privileges, Tukey follow-ups of the overall Culture Group × Sibling Gender interaction, F(3, 234) = 4.09, p < .01, revealed that, whereas there were no sibling constellation differences for the Anglo group, within the Mexican group, older sisters in girl-boy dyads received relatively fewer privileges as compared to their younger brothers, and older siblings in boy-girl, d = 1.13, boy-boy, d = .87, and girl-girl dyads, d = .80 were granted relatively more privileges than their younger siblings. Turning to the allocation of household tasks, we again found a significant overall Culture Group × Sibling Gender Constellation interaction, F(3, 234) = 10.88, p < .01. Similar to the findings for privileges, Tukey follow-ups revealed that, whereas there were no differences as a function of gender constellation in the Anglo group, in the Mexican group, older sisters in girl-boy pairs scored highest and older brothers in boy-girl pairs scored lowest in their relative involvement in household tasks. Indeed, as Table 3 indicates, the only families in which older siblings were not assigned more household tasks (i.e., scores greater than 3) were Mexican-oriented families that included an older brother–younger sister dyad. In these families, younger sisters were relatively more involved in house-hold tasks than their older brothers. The results for allocation of monetary resources revealed the same Culture Group × Sibling Gender Constellation interaction, F(3, 233) = 2.98, p < .05. Although the pattern of means in the Mexican group was similar to that in the prior two analyses (older sisters with younger brothers received the least favorable treatment), follow-up Tukey tests revealed no gender constellation differences in monetary allocations in either cultural group.

Table 3.

Means (SDs) for differential treatment as a function of sibling gender constellation and parents’ cultural orientation

	Family Context
	Mexican Oriented				Anglo Oriented
Differential Treatment	Girl-Girl	Girl-Boy	Boy-Girl	Boy-Boy	Girl-Girl	Girl-Boy	Boy-Girl	Boy-Boy
Privileges	3.64b (.85)	2.92a (.87)	3.84b (.76)	3.64b (.79)	3.65 (.78)	3.71 (.72)	3.47 (1.02)	3.80 (.92)
Chores	3.31b (.66)	3.98a (.74)	2.17c (.62)	3.16b (.64)	3.21 (.72)	3.75 (.72)	3.33 (.72)	3.03 (.85)
Money	3.37 (.73)	3.07 (.81)	3.24 (.73)	3.34 (.71)	2.86 (.77)	3.35 (1.02)	3.41 (1.00)	3.03 (.77)

Note: Means in the same row that do not share subscripts differ at p < .05 using Tukey Honestly Significant Difference comparison.

As noted, the analyses focusing on parental control revealed a different pattern of findings. Here, an overall Culture Group × Parent effect emerged, F(1, 234) = 9.55, p < .01. Follow-ups revealed that, in families with more Mexican-oriented parents, adolescents rated fathers as being more strict with older siblings than mothers, M = 3.32, SD = .75 and M = 3.10, SD = .77, for Mexican-oriented fathers versus mothers, d = .26; in contrast, in the Anglo group, fathers were rated as being less strict with older siblings than mothers, M = 3.09, SD = .90 and M = 3.18, SD = .96, for Anglo-oriented fathers versus mothers, d = .11. This analysis also revealed an overall Culture Group × Sibling Gender Constellation × Sibling effect, F(3, 234) = 3.59, p < .01. Follow-ups documented differences between siblings’ perceptions of parents’ control that varied as a function of gender constellation in the Mexican-oriented families: Sisters with younger brothers perceived relatively more control directed toward themselves than did their brothers, M = 3.51, SD = .86, for older sisters, d = .27, and M = 2.96, SD = .73, for younger brothers, d = .45. In contrast, older brothers with younger sisters reported that they experienced relatively less parental control than did their sisters, M = 2.97, SD = .93, for older brothers, and M = 3.11, SD = .79, for younger sisters, d = .16. No gender constellation differences emerged in Anglo-oriented families.

In a follow-up step, we examined mothers’ and fathers’ cultural orientations and their links to differential treatment separately to see whether either parent’s orientations were more consistently linked to differential treatment patterns. The results for mothers and fathers were similar to the overall parental pattern, validating our use in this sample of a measure that incorporated both parents’ cultural orientations.

Cultural Values, Differential Treatment, and Adolescent Well-Being

Our next step was to test a cultural-ecological perspective on the implications of differential treatment for youth well-being: Moving beyond the notion that sibling social comparison processes are universal across families, we asked whether adolescents’ familism values moderated the links between differential treatment and siblings’ adjustment, perceptions of parental acceptance, and ratings of parents’ fairness. Specifically, we tested the hypothesis that strong familism values would mitigate the links between differential treatment and youth well-being by focusing youth’s attention on family needs. As noted, the low to modest correlations between siblings’ ratings of the different domains of differential treatment meant that we had to examine the well-being correlates of differential treatment separately by domain and by sibling. Given the number of predictor and criterion variables, in an effort to reduce the possibility of chance findings, we assessed the correlations between specific domains of differential treatment and the adjustment and parent-adolescent relationship measures for adolescents with high versus low familism values, and then we tested whether the overall pattern of correlations differed across groups using structural equation modeling.

We began by creating groups of older and younger siblings who were high versus low in their familism values using a mean split on the familism measure. (A mean split was used because the mean was closer to the midpoint of the familism scale than was the median.) This procedure resulted in groups of n = 146 older siblings with high familism scores and n = 100 older siblings with low familism scores and groups of n = 142 younger siblings with high familism scores and n = 104 younger siblings with low familism scores. Preliminary analyses revealed that, for both older and younger siblings, the resulting groups differed significantly in their familism values, M = 76.56, SD = 3.68 and M = 61.69, SD = 9.40, F(1, 244) = 295.19, p < .01, d = 2.26, for older siblings’ familism group differences, and M = 76.54, SD = 3.36 and M = 63.96, SD = 7.89, F(1, 244) = 289.40, p < .01, d = 2.26, for younger siblings’ familism group differences. The groups also differed on family background characteristics including number of children in the household and maternal education, which were therefore controlled in the analyses.

Next, we conducted a series of partial correlations, examining the links between differential treatment and youth’s well-being separately by familism group and partialing out the control variables. We then tested whether the correlation matrices differed for adolescents with stronger versus weaker familism values: Significant correlations between differential treatment and well-being in the low-familism but not in the high-familism group support our hypothesis that strong familism values mitigate the expected negative implications of differential treatment. Note that, because correlations between self-report measures should be inflated because of rater bias, this analytic approach provides for a strong test of our hypothesis that the linkages between differential treatment and well-being will be low or nonsignificant for youth with strong familism values. Because of our directional hypothesis, we again report results at p < .10 and higher.

The results of the correlational analyses are shown in Table 4 (older siblings) and Table 5 (younger siblings). In the case of older siblings, 26 of 35 correlations for the low-familism group reached trend level or higher, and all were in the expected direction: Treatment that favored the self was linked to fewer depression symptoms, lower levels of involvement in risky behavior, lower levels of mother-adolescent and father-adolescent conflict, and higher ratings of parents’ fairness. In contrast, only five correlations were significant in the case of the high-familism group.

Table 4.

Partial correlations between perceptions of differential treatment and adjustment, parent-adolescent acceptance, and fairness ratings for older siblings with high (N = 146) versus low (N = 100) familism values

	Depression Symptoms		Risky Behavior		Maternal Acceptance		Paternal Acceptance		Fairness Ratings
Older Sibling More:	High Familism	Low Familism	High Familism	Low Familism	High Familism	Low Familism	High Familism	Low Familism	High Familism	Low Familism
Privileges	−.10	−.31**	.08	−.06	.01	.17†	.11	.06	.14	.37**
Chores	.07	.30**	.01	.08	−.02	−.21*	.04	−.23*	−.33**	−.46**
Monetary resources	.00	−.19*	.01	−.06	.03	.23*	.03	.03	.12	.12
Maternal control	.06	.33**	.08	.18*	−.09	−.37***	−.07	−.14	−.33**	−.46**
Paternal control	.00	.28**	−.01	.22*	.15†	−.21*	−.01	−.28**	−.30**	−.47**
Maternal warmth	.10	−.19*	.07	−.23*	.17*	.46***	−.11	.12	−.03	.39**
Paternal warmth	−.04	.03	−.01	−.30**	−.02	.34***	.24**	.54***	.12	.54**

^†p < .10.

^*p < .05.

^**p < .01.

^***p < .001.

Table 5.

Partial correlations between perceptions of differential treatment and adjustment, parent-adolescent acceptance, and fairness ratings for younger siblings with high (N = 141) versus low (N = 104) familism values

	Depression Symptoms		Risky Behavior		Maternal Acceptance		Paternal Acceptance		Fairness Ratings
Older Sibling More:	High Familism	Low Familism	High Familism	Low Familism	High Familism	Low Familism	High Familism	Low Familism	High Familism	Low Familism
Privileges	.03	.15	−.05	.17†	−.06	−.08	−.14	−.06	−.28**	−.35**
Chores	−.10	−.01	.12	−.21*	.10	.19†	.03	.40***	.32**	.32**
Monetary resources	.05	.17†	.12	−.08	−.03	−.07	.09	.11	−.42**	−.42**
Maternal control	−.23**	−.20*	−.15†	−.23*	.24**	.14	.06	.20*	.15†	.44**
Paternal control	.01	.05	.02	.01	.03	−.18†	−.03	−.10	.02	.20*
Maternal warmth	.22**	.31**	.30**	.16†	−.16†	−.21*	−.03	−.07	−.29**	−.40**
Paternal warmth	.13	−.35**	.08	−.17†	.10	.09	.03	−.18†	.12	−.08

^†p < .10.

^*p < .05.

^**p < .01.

^***p < .001.

We tested for differences between the correlation matrices of the low- versus high-familism groups using structural equation modeling with AMOS 4.0 (Arbuckle & Wothke, 1999). Specifically, we tested the relative fit of two models: In the first model, the correlation matrices were constrained to be equivalent across the high- and low-familism groups; in the second model, they were allowed to vary. χ² tests were used to evaluate whether the constrained model (matrices equal) produced a significantly different (i.e., poorer) fit than the freely estimated (matrices different) model. The constrained and unconstrained models were estimated separately for the three sets of youth outcomes: (a) adjustment (depression and risky behaviors), (b) parental acceptance, and (c) fairness ratings.

Results revealed that the unconstrained models fit the data better than the constrained models, Δχ² = 22.03, df = 14, p < .08; Δχ² = 28.91, df = 14, p < .01; Δχ² = 21.90, df = 7, p < .01, for the adjustment, acceptance, and fairness matrices, respectively. Given the sensitivity of the χ² statistic, in combination with our directional hypothesis, these results provide good support for the proposition that negative effects of differential treatment are mitigated by strong familism values.

One basis for the different pattern of correlations across familism groups is that adolescents in the low-familism group might actually be treated less favorably than those in the high-familism group. To test this possibility, we conducted a 2 (sibling gender) × 2 (familism) ANOVA with adolescents’ reports of differential treatment as the dependent measures. The results revealed only one effect for familism: Older siblings in the high-familism group reported that their siblings were treated more warmly by fathers as compared to older siblings in the low-familism group, M = 2.73, SD = .84 for the high-familism group, and M = 2.97, SD = .91 for the low-familism group. In other words, this pattern was inconsistent with the notion that significant correlations between differential treatment and well-being problems were found in the low-familism group because these youth were treated less favorably.

Findings for younger siblings were less consistent with our hypothesis about the role of familism. For the low-familism group, 20 of 35 correlations reached trend level or higher, whereas in the high-familism group, 11 of 35 correlations reached this criterion. Again a structural modeling approach was used to test for differences between the correlation matrices of the two groups. Results were consistent with our hypothesis for the adjustment and fairness rating matrices, Δχ² = 40.96, df = 14, p < .001; Δχ² = 12.92, df = 7, p < .08, but the comparison for the parental acceptance ratings was not significant, Δχ² = 16.65, df = 14, ns. These results provide some support for our hypothesis about the moderating role of familism.

We next explored whether the group differences in correlations emerged because younger siblings in the high-familism group were treated more favorably. Comparisons of the two familism groups’ scores on the differential treatment indices using ANOVA procedures revealed only one group difference: Younger siblings in the high-familism group reported that they received less favorable treatment in the area of maternal control, M = 2.84, SD = 1.02, as compared to younger siblings in the low-familism group, M = 3.09, SD = .91. As was the case with older siblings, these results suggest that the more negative reactions to differential treatment by younger siblings in the low-familism group did not emerge because they received more unfavorable treatment.

In a final test, we explored one possible basis for the stronger pattern of findings for older as compared to younger siblings. In this study, birth order and age were confounded, but there was some variability in older siblings’ ages. Accordingly, we reran the analyses for older siblings, comparing the patterns of association between differential treatment and well-being for older siblings who were above 15 years of age (the median age for this sample of older siblings) versus older siblings who were below 15 years of age. The resulting smaller sample size limits our ability to draw conclusions about the role of age, but we found no evidence that the pattern of results was stronger for older siblings of age 15 and above than for older siblings below age 15.

DISCUSSION

A body of research and theory suggests that parents’ differential treatment of siblings is a central dynamic in family life (Ansbacher & Ansbacher, 1956; Brody et al., 1992). The present study adds to this literature by documenting the family patterns and correlates of differential treatment in two-parent Mexican American families. Study of differential treatment in Mexican American families contributes to our understanding of this family dynamic in evidencing its operation in another cultural context: Consistent with a cultural-ecological perspective (Spencer, 1995), our findings suggest that differential treatment dynamics are not universal but vary as a function of cultural processes within families.

The present research also contributes to our understanding of Mexican American families in demonstrating that parents with stronger orientations to Mexican culture were more likely to display gender-typed patterns of differential treatment. Some prior research on Mexican American families has suggested that these families are more traditional in terms of gender orientations (Valenzuela, 1999), but as researchers in this area have argued, it is important to separate the “myths” from the realities in this cultural group (Cauce & Domenech-Rodríguez, 2002). This study provides a step in this direction. Further, by using an ethnic homogeneous design, we were able to document variability among Mexican American families in how family roles and activities are organized by gender and provide insights into the processes underlying gender-typed patterns of differential treatment. Mexican American families are a diverse group, and they vary along many dimensions (e.g., ties to traditional Mexican culture, parent education, family size). As such, it is important to move beyond social address variables (i.e., ethnic group status) to study the qualities of families (e.g., parents’ cultural orientations) that make a difference for youth’s family experiences and development. Given our interest in fathers’ family roles, the present study included only two-parent families, and an important direction for future research is to examine these kinds of dynamics in families that vary in structure.

Our findings showed that culture matters for differential treatment dynamics in at least two ways. First, the data were consistent with the idea that parents’ cultural orientations influenced the way they organized their children’s family roles and activities. Although chores and privileges were the prerogative of older siblings in the sample as a whole, in a pattern consistent with extant research on European American families (e.g., Tucker et al., 2003), older brothers with younger sisters performed relatively fewer chores and older sisters with younger brothers were granted relatively fewer privileges when their parents were more enculturated within Mexican culture and less acculturated to Anglo culture. Importantly, our findings showed that gender-typed treatment of offspring required both a set of values that disposed parents toward particular parenting strategies as well the opportunity to enact those values, that is, the presence of both a daughter and a son. In other words, both family structure and family dynamics must be taken into account in understanding these adolescents’ gendered family experiences. Although parents’ cultural orientations had implications for adolescents’ family roles and activities (i.e., chores, privileges), we found less evidence that they made a difference for relationship dynamics, that is, parental warmth and control. Prior research on differential treatment has focused almost entirely on parent-child relationship dynamics, but our findings suggest that other domains of differential treatment that pertain to siblings’ family roles and activities may be where culture makes its mark, and these should be explored in future work.

In addition to shaping family roles and activities, our findings suggested that culture operates by imbuing family experiences with meaning. When siblings reported stronger familism values, even though we were testing correlations between self-reports, we found few significant links between differential treatment and youth well-being. In contrast, siblings with weaker familism values showed the pattern predicted by Adler’s theory about sibling social comparison dynamics: Less favorable treatment was correlated with poorer well-being. Familism values highlight family responsibilities and the needs and interests of family members; concern for the group may make adolescents less inclined toward social comparisons with their siblings and less likely to perceive unequal treatment as a sign that they are unloved or less valued by their parents. Extant theory about the negative implications of differential treatment emerged within a cultural context that highlights the rights of the individual, the ideal of equal treatment, and the way sense of self is enhanced when individuals think they are better than others (Nuckolls, 1993; Parsons, 1974/1942; Suls & Wheeler, 2000). Much of the empirical literature on differential treatment has been conducted with the seeming assumption that this value system is universal. The findings of this study are consistent with a handful of investigations that suggest that the meanings youth attribute to their parents’ differential treatment—rather than the fact of parents’ treatment—have the most important implications for youth well-being (e.g., Kowal & Kramer, 1997; McHale & Pawletko, 1992). A limitation of the present study is that we did not collect information about phenomena such as sibling comparisons or adolescents’ perceptions of the reasons for their parents’ differential treatment. An important direction for future research on the role of culture in family dynamics such as differential treatment would be to learn how family members interpret their family experiences. At the level of application, however, our findings suggest that children’s values have implications for the ways in which they respond to parental strategies and that different strategies are effective in different family contexts. Our results also imply that promoting values that highlight family needs and interests may mitigate the negative implications of differential treatment. Another important direction for future research would be to determine whether familism values mitigate the negative implications of differential treatment in European American families and other ethnic groups.

Our findings on the links between differential treatment and adolescent well-being revealed somewhat different patterns for older and younger siblings. Although significant effects were evident for both groups of siblings, the results were more consistent for older siblings. Because age and birth order were confounded in this study, we were not able to disentangle which of these factors was responsible for these findings. Although we tested for differences in reactions to differential treatment as a function of older siblings’ age and found no differences for chronologically older versus younger siblings, small cell sizes may have precluded our detecting age effects. Some prior studies have shown that (chronologically) older siblings are more sensitive to differential treatment dynamics than younger ones (Bryant & Crockenberg, 1980; Feinberg, Neiderhiser, Simmens, Reiss, & Hetherington, 2000; Kowal & Kramer, 1997); these investigators argue that, with age, siblings may simply become more aware of and understanding of differential treatment dynamics. For siblings in our sample, familism values may play an increasingly important role in youths’ perceptions of and reactions to their experiences as they grow older. Alternatively, a birth-order interpretation of older siblings’ greater sensitivity to differential treatment dynamics would direct attention to the hierarchical structure of Mexican American families and the possibility that siblings higher in birth order are more threatened when they are not accorded favored status. Such an interpretation is consistent with Feinberg et al.’s contentions about the difference between downward versus upward social comparisons, that is, that downward social comparisons, such as an older sibling would make vis-à-vis a younger, are more psychologically threatening. A longitudinal design would provide for a test of the relative roles of age versus birth order in siblings’ reactions to differential treatment and constitutes an important direction for this research.

In conclusion, the present study adds to the knowledge base on normative family dynamics in Mexican American families. In so doing, it also expands our understanding of differential treatment by delineating the cultural processes that give rise to this dynamic and that moderate its impact on siblings. Finally, this research contributes to the literature on families, more generally: Study of differential treatment moves beyond the focus of much family research on individual family members or dyadic relationships by targeting parent-sibling triads, and our analytic approach was directed at capturing the patterning of multiple family members’ experiences. An understanding of how families work as systems requires selection of constructs that reflect processes beyond the individual and dyadic levels, data collection approaches that include information about the experiences of multiple family members, and analytic strategies that target patterns or profiles of family experience. Our research makes a beginning step in this direction. Clearly, however, much more work is needed in the development of family research methods.