Becoming More Egalitarian: A Longitudinal Examination of Mexican-Origin Adolescents’ Gender Role Attitudes

Abstract

The current study examined the trajectory of gender role attitudes of 471 Mexican-origin adolescents (236 girls, 235 boys) from 5^th grade (M_age = 10.86 years) to 11^th grade (M_age = 16.75 years), investigating how situating identities (i.e., gender, nativity, SES), ethnic identity (i.e., ethnic pride), and familial context (i.e., parents’ attitudes) contributed to adolescents’ gender role attitudes across time. Participant interviews were conducted every other year, resulting in 4 waves of data. Most parents (96%) were Mexico natives, with an average immigration age of 18.16 years for fathers and 14.01 years for mothers. Results revealed linear and quadratic trends in gender attitude traditionality for all adolescents, characterized by a linear decline through age 16 years that leveled off through age 18 years. Although both girls and boys trended toward egalitarian gender role attitudes across adolescence, girls endorsed more egalitarian attitudes than did boys. Adolescents from higher-SES backgrounds endorsed more egalitarian attitudes than those from lower-SES backgrounds. Significant within-person effects of ethnic pride surfaced, such that children with higher levels of ethnic pride at any given time also reported more traditional gender role attitudes. Significant between-person effects of mothers’ and fathers’ attitudes were found, such that parents with more traditional gender role attitudes tended to have children with relatively more traditional gender role attitudes. Overall, these findings highlight the importance of studying gender development in conjunction with situating identities, cultural identities, and the broader context, particularly when children are embedded in multiple cultures with contrasting gender role expectations.

Immigrant individuals currently represent 13.4 percent of the total United States population, a figure expected to increase in the coming years (López & Radford, 2017). Mexican-origin individuals represent 27 percent of this population, making Mexican-origin families the largest immigrant group in the United States. Although it is important to include Mexican-origin families in all areas of developmental research, the immigrant context is particularly important when considering culturally-specific constructs such as gender development. Throughout their childhoods, immigrant children may need to balance multiple cultures as they receive messages from both their parents’ home culture and the United States’ culture. Gender, a construct that denotes the social and cultural qualities associated with being male or female (e.g., gender roles, Blakemore, Berenbaum, & Liben, 2009), represents a prime example of this balance between multiple cultures (Schroeder & Bámaca-Colbert, 2019), with home and outside contexts sending potentially different messages which affect children’s gender-related identity and attitude development. Thus, immigrant children are simultaneously exposed to multiple gender role expectations and may internalize parts or all of a given gender role ideology. In the current study, we examine gender development in an immigrant context across a 7-year period, specifically examining the roles of situating identities (i.e., gender, socioeconomic status, nativity), cultural identities (i.e., ethnic identity), and contextual variables (i.e., mother and father gender role attitudes) in the formation of gender attitudes among Mexican-origin youth.

Theoretical Framework

García Coll and colleagues’ (1996) model of minority children’s development brings attention to children’s multiple social position variables (e.g., social class, ethnicity, gender) and their role in situating a child’s developmental context. These social position variables not only are a reflection of the child’s own identity, but also are impacted by broader social stratification systems, affecting the way in which minority youth fit into the social hierarchy. Importantly, these social positions, or situating identities, do not exist separately, but are instead overlapping, compounding, and intersecting; it is the unique combination of various social position variables that situates a child’s development. When considering gender role attitude development, perhaps the most relevant situating identity is an individual’s gender. However, other situating identities such as nativity (Updegraff et al., 2014) and socioeconomic status (Pepin & Cotter, 2018) have also been found to influence the development of gender role attitudes.

In addition to understanding the role of situating identities, we also focus on the ways in which individuals may perceive and internalize their social and cultural identities. Knight, Safa, and White (2018) suggest that as children gain knowledge of the behaviors, expectations, attitudes, beliefs, and values relevant to various social identities (e.g., gender, ethnicity), those identities are incorporated into their self-concept. A relevant and developmentally important identity among ethnic minority individuals is ethnic identity (Rivas-Drake et al., 2014). Ethnic minority individuals undergo a process of internalization of their native culture, moving through age-related stages of exploring and understanding their ethnicity and ultimately incorporating ethnicity into their overall identity (Phinney, 1993). Eventually, ethnic identity, in conjunction with other social identities, shapes the way in which children interpret their environments, guiding their own behaviors, attitudes, and beliefs.

Contexts of development (e.g., family, peers, culture) can also shape the way in which individuals’ identities are perceived and incorporated into the self-concept (Abes, Jones, & McEwen, 2007; Knight et al., 2018). Abes and colleagues (2007) theorize that contextual factors, such as family beliefs and norms, may influence an individual’s self-perception of their identities. When thinking about gender development, family gender roles and family members’ broader gender ideologies are particularly important to consider. As children learn about gender and incorporate their gender identity into their self-concept, the way in which family members think about and enact gender roles may affect their own attitudes and beliefs related to gender (Tenenbaum & Leaper, 2002). To capture this familial influence, our research incorporates the role of parents’ gender attitudes in influencing their adolescents’ gender attitudes.

Importantly, none of these processes are static - adolescents in immigrant families are simultaneously experiencing developmental change and acculturative change, both of which affect the formation of their intersecting identities. The acculturation process is multidimensional, with immigrant adolescents balancing practices, values, and identities related to their home and host cultures (Schwartz, Unger, Zamboanga, & Szapocznik, 2010). The identity aspect of acculturation affects an individual’s endorsement of situating identities related to ethnicity and culture. The values aspect of acculturation has the potential to shape additional situating identity variables such as gender. The degree to which an individual adheres to a given culture’s values, often marked by strength of ethnic identity and acculturation status, may affect both the way in which their own gender is expressed and the way in which they view others’ gendered behaviors. Altogether, we approached the current study using a combination of these theoretical models, examining the ways in which a child’s socio-ethnocultural and gender identities may intersect and become internalized within the immigrant context across time. Gender attitudes, a belief associated with gender identity but also embedded within ethnic identity, is particularly interesting to study within these frameworks.

Gender Attitude Development

As is the case with the development of several other identity-related constructs, gender development is particularly salient during adolescence, with gender-relevant changes occurring in multiple domains (i.e., cognitive, social, and biological). During middle childhood, children undergo a sharp increase in their cognitive abilities, emerging in adolescence with a more complex and better defined sense of identity, especially when pertaining to collective identities (i.e., gender, ethnicity, race; Eccles, 1999). This cognitive growth extends through adolescence, marked by increases in identity maturity, growth of cognitive empathy, and growth of support and understanding in relationships (Meeus, 2016). These relationship- and identity-related cognitive changes hold an important role in adolescents’ gender development, allowing for better reflection on and selective endorsement of the gender roles apparent in everyday experiences.

As adolescents notice and gain knowledge of the various gendered expectations in their environments, this information not only shapes the behaviors, interests, and relationships that are internalized into children’s gender identities, but also forms the attitudes related to gender that drive the formation of children’s gender ideologies. Unlike gender identity or gender expression, which are viewed as more internal processes, gender attitudes have greater external consequences, potentially relating to individuals’ occupation aspirations, religious orientation, and political views (Barth, Guadagno, Rice, Eno, & Minney, 2015; Dolan, 2010; Fulcher & Coyle, 2011; Kulik, 2000). Additionally, the socialization of gender attitudes through parents’ attitudes have similar implications (Lawson, Crouter, & McHale, 2015).

Across childhood, children hold strict ideas about appropriate and inappropriate gendered behaviors, demonstrating a tendency toward more traditional gender attitudes (Blakemore et al.,2009). However, this clear trajectory muddles in adolescence. In some studies, early adolescence (sixth and seventh grade) marks one of the most flexible periods in gender attitude development (Liben & Bigler, 2002). In others, boys and girls demonstrate different attitude trajectories, with girls becoming more flexible and boys becoming more traditional (Galambos, Almeida, & Petersen, 1990). More consistent findings have revealed that both child gender and parents’ gender attitudes are significant influences on adolescents’ gender attitudes (Crouter, Whiteman, McHale, & Osgood, 2007; Davis, 2007; Lam, Stanik, & McHale, 2017; Updegraff et al., 2014). Specifically, girls’ attitudes are generally more egalitarian than are boys’, and parents’ more egalitarian attitudes inspire more egalitarian attitudes in their children, an effect shown to last through adulthood (Cunningham, 2001). In late adolescence and the transition to adulthood, potential influencing factors increase, with parenthood, pursuing education, and occupational attainment affecting gender role attitudes (Fan & Marini, 2000; Toomey, Updegraff, Umaña-Taylor, & Jahromi, 2015).

Gender Attitudes and Ethnic Identity

When studying racial-ethnic minority youth, it is critical to consider the ways in which concurrent identities may derive meaning from one another (Cole, 2009). As mentioned previously, for immigrant adolescents, gender attitude formation is embedded in developmental processes related to ethnic identity and acculturative processes, both of which may affect the way in which their gender attitudes are formed. Specifically, gender role attitude shifts are considered a marker of acculturation, given how profoundly gender ideologies are embedded in Latino culture (Glass & Owen, 2010; Kulis, Marsiglia, & Hurdle, 2003). For example, the social implications and interpretations of being a girl are quite different from those of being a Latina girl, or Latino boy. With each additional identity, new meaning is gained, determined by the knowledge of previously named categories. Particularly important are the mutually derived meanings of race-ethnicity and gender, as gender roles and gender power disparities manifest differently in different cultures (Hurtado, 2009; Milville & Ferguson, 2014). Thus, expectations for masculinity and femininity follow culturally-specific norms, meaning that for ethnic minority populations, it is especially important to consider how their gender identity may be situated in, and shaped by, their particular ethnic background.

In order to capture the way in which unique aspects of Mexican-origin adolescents affects gender role attitude development, previous research (Updegraff et al., 2014) included nativity as a predictor of gender attitude trajectories, finding that nativity played a significant role in the formation of gender attitudes. Whereas boys and girls born in Mexico showed gender- differentiated trajectories, with boys remaining stable and girls showing attitudes that were more egalitarian over time, boys and girls born in the U.S. did not differ in their gender attitude trajectories. Research studying the role of nativity allows for an interesting look at how adolescents’ native culture might influence their cultural values, including those related to gender roles. However, nativity represents only a demographic characteristic related to an individual and does not necessarily reflect the ways in which that individual may internalize that characteristic into their self-concept. Ethnic identity, on the other hand, is a process of internalization of one’s native culture (Phinney, 1993), with ethnic pride being the aspect of ethnic identity that reflects on individuals’ positive attitudes towards their ethnic group (Phinney, 1990). Thus, we incorporated ethnic pride as a more direct way in which a child’s native culture can contribute to gender attitude development.

Gender Development and the Latino Immigrant Context

Beyond a child’s own identities, there are contextual factors that play important roles in the development of gender attitudes. In an immigrant context, gender role attitude development involves exposure to multiple cultures, with immigrant children responsible for balancing and internalizing varying definitions of male and female gender roles. In addition, immigration itself is a gender-differentiated experience that may come with distinctive benefits and challenges for men and women (Espiritu, 2001; Itzigsohn & Giorguli-Saucedo, 2005). In many cultures, males play a dominant role in the gender hierarchy, holding authority in work, family, and societal domains (Hyde, 2014; Kane, 2000). However, in some countries, the hierarchy appears to be much steeper than in others, which, for immigrant families, may mean adjusting to a new set of gender role expectations. Mexico is one country that endorses a rather steep gender hierarchy, with adolescents’ reports demonstrating relatively less favorable attitudes toward gender equality and census ratings of the gender inequality index falling higher than other countries’ (Dotti Sani & Quaranta, 2017). The United States’ comparatively more egalitarian gender ideology thus poses a particularly challenging adjustment for individuals from countries like Mexico with steep gender hierarchies. For women, this change allows for an expansion of social roles and potentially greater opportunities outside the home, thereby inspiring a faster acculturation to the United States’ gender role attitudes (Updegraff et al., 2014). However, while their work opportunities and experiences may expand, immigrant women may still be holding traditional roles in their home and community contexts (Pessar, 1999; Zhou & Bankston, 2001). For men, the same experience may occur, but with the opposite valence attached. Men may feel more restricted or less appreciated in their work roles upon immigrating, but then may continue their authoritative roles in home and community settings to preserve some semblance of their home country’s patriarchal tradition (Zhou & Bankston, 2001).

Not only are the gender hierarchies different, but the cultural emphasis placed on gender also differs across cultures (Cole, 2009). For Mexican-origin individuals, gender is a particularly salient aspect of culture, with specific words, expectations, and traditions tied to masculinity and femininity. For instance, marianismo and machismo provide a special emphasis on and rigid definition of femininity and masculinity. Marianismo suggests that women and girls should be cherished and protected, maintaining sexual abstinence until marriage. Conversely, machismo outlines starkly different expectations for masculinity: males are supposed to demonstrate dominant and aggressive personality traits. In recent years, this narrower concept of masculinity has expanded, allowing for a bilinear conceptualization that includes caballerismo, which emphasizes men’s role as protectors with central responsibility toward family (Arciniega, Anderson, Tovar-Blank, & Tracey, 2008). The endorsement of these traditional Mexican gender ideals varies by individual characteristics (e.g., gender and generation; Karver, Sorhaindo, Wilson, & Contreras, 2016; Updegraff, Umaña-Taylor, McHale, Wheeler, & Perez-Brena, 2012) and physical location (e.g., Mexico vs. U.S. or urban vs. rural; Hirsch, 1999). Additionally, it is possible for individuals to endorse only some aspects of these ideologies. For example, Mexican- origin men have been found to identify with certain components of machismo but not others (Torres, Scott, & Carlstrom, 2002). Gender differences in endorsement of marianismo have also been found, with boys showing greater endorsement of self-silencing value of marianismo compared to girls and girls showing greater endorsement of family and spiritual roles of marianismo compared to boys (Piña-Watson, Castillo, Jung, Ojeda, & Castillo-Reyes, 2014).

Within the family context, differential adjustments and expectations for men and women also transfer to parenting, as both fathers and mothers attempt to preserve aspects of their home cultures in their children. Complicating these gendered undertones is the fact that boys and girls are socialized differently, expected to fulfill different cultural roles within the family. For example, girls in immigrant families are often expected to be the “keepers of culture,” learning and maintaining cultural traditions (Qin, 2009; Stritikus & Nguyen, 2007). This expectation of greater cultural responsibility for girls is a pattern that holds across several domains. In a retrospective study of Latino gender socialization, women reported having less social freedom, stricter dating rules, and greater household chores than did men (Raffaelli & Ontai, 2004). The social and dating restrictions are typically tied to the emphasis placed on daughters’ sexual purity for the family’s reputation (Lee, 2005; Suárez-Orozco & Qin, 2006). Interestingly, even as other gender role expectations have become more equal in Mexican culture, the role of women as the family cultural hub and the need to remain a virgin until marriage are ideals that have persisted (Karver et al., 2016). For immigrant boys, sexual activity and risky behavior does not jeopardize the family’s social status, allowing for greater freedom and independence. This “double standard” in parental monitoring, as coined by Espiritu (2001), is not unique to a single cultural background; in almost every ethnic or social group, boys are granted more independence across development (Blakemore et al., 2009; Suârez-Orozco & Qin, 2006).

Further complicating the gender dynamics in the family, in all domains of gender socialization, fathers seem to enforce the most gender pressure on their sons whereas mothers focus on their daughters’ gender socialization. Immigrant women remember their mothers encouraging, and sometimes enforcing, their stereotypically feminine behavior, as mothers discouraged “tomboy” behavior in favor of having long hair, wearing dresses, and being “ladylike” (Raffaelli & Ontai, 2004). This research is in line with other qualitative findings, which cite mothers explaining what it means to be a “good” wife, daughter, or mother (Gallegos-Castillo, 2006) and explicitly stating expectations for their daughters’ sexuality (Ayala, 2006). For boys, fathers have been found to encourage extreme masculinity, advocating for masculine activity involvement and reduced emotional expressivity (Raffaelli & Ontai, 2004).

As immigrant children receive messages from their parents about appropriate gender behaviors in their home culture, they also experience gender socialization in the host culture context. In many cases, messages from peers may stem from a more egalitarian gender ideology than the messages from parents, causing children to juggle contrasting definitions of what it means to be a boy, girl, man, or woman (Lee, 2005). This balancing act then pushes children either to make decisions about which gender roles to internalize (i.e., home vs. host culture’s gender ideology) or to learn to code-switch, demonstrating the home culture’s gender roles in the family context and the host culture’s gender roles in the peer context (Schindler, Reinhard, Knab, & Stahlberg, 2016). Parents, too, have to balance their home and host cultures’ gender role ideologies, but may acculturate at a slower rate, endorsing their home cultures’ gender roles more persistently than their children (Phinney, Ong, & Madden, 2000). With respect to cultural gender roles, this quicker acculturation is especially true for immigrant girls, who may be drawn more strongly to a gender ideology that affords a broader range of opportunities (Dasgupta, 1998).

The Current Study

The current study seeks to extend the existing gender development literature by examining the trajectory of Mexican-origin adolescents as a function of their situating identities, cultural identities, and familial contexts. In line with existing theoretical and empirical work, the following hypotheses were tested:

1. We expected that both linear and quadratic trends would emerge in adolescents’ gender role attitude trajectories. Specifically, we expected that adolescents’ gender role attitudes would show the greatest decline in early adolescence and would level off in late adolescence.

2. We hypothesized that three situating identities would affect children’s gender role development:

   1. We predicted that adolescents’ gender would affect gender role attitude development, such that girls’ endorsement of traditional Mexican gender roles would begin at a lower point and would decline faster than would boys’.
   2. We predicted that adolescents’ nativity (i.e., immigrant status) would affect gender role development, such that Mexico-born adolescents would more strongly endorse traditional gender role attitudes compared to US-born adolescents.
   3. We predicted that adolescents’ SES would affect gender role attitude development, such that adolescents growing up in lower SES families would show stronger endorsement of traditional gender role attitudes than would adolescents growing up in higher SES families.

3. We hypothesized that individuals’ internalization of their ethnic identity (measured as ethnic pride) would affect their gender role attitude development, over and above the effect of situating identity variables.

4. We hypothesized that individuals’ contextual factors, measured as parents’ gender attitudes, would affect their gender role attitude development, over and above the effects of situating and ethnic identities.

   1. We hypothesized that parents’ gender role attitudes would affect their children’s, such that parents with greater endorsement of traditional Mexican gender role attitudes would have children with relatively higher endorsement of those attitudes.
   2. We hypothesized that parents’ attitudes would differentially affect their children’s gender role attitude development, with mothers’ attitudes being more impactful than fathers’.
   3. We expected an interaction between parents’ gender role attitudes and children’s gender, such that fathers’ attitudes may have greater impact on sons than daughters and mothers’ attitudes may have greater impact on daughters than sons (though mothers’ impact may be significant for both sons and daughters, given the rationale in hypothesis 4b).

Method

Participants and Procedure

Data were collected as part of a large longitudinal research study of Mexican-origin families in Northern California (e.g., Aizpitarte, Atherton, & Robins, 2017; Martin, Conger, & Robins, in press). The project, entitled the California Families Project, was granted IRB approval through the University of California, Davis with the following IRB protocol number: 217484–23. The broad aim of the study was to identify individual, family, school, community, and cultural factors that increase or reduce risk for emotional and behavioral problems during late childhood and adolescence. The current research centered on a small portion of the project, looking at only relevant cultural and demographic variables.

The original sample consisted of 674 single and two-parent Mexican origin families. Retention rates for the full sample were 86%, 91%, and 90% at Waves 3, 5, and 7, respectively. Families were selected for recruitment if they had a fifth-grade child (M_age = 10.85, 50.0% boys) who was drawn at random from local school rosters. Parents were subsequently contacted via telephone or home visits during the 2006–2007 and 2007–2008 school years. Both single-and two-parent families were eligible to participate. Single-mother families could participate only if there was not another adult living in the household. Two-parent families were eligible to participate only if the child was living with their biological mother and father.

Considering the current project’s specific interest in the relative influence of mothers and fathers, only data from two-parent families were included. Additionally, only families with at least one immigrant parent were included in analyses. Data for the variables of interest were available for 471 (96.5%) of those families at four time points: fifth grade, seventh grade, ninth grade, and eleventh grade. Of those 471 families included in the current research, 236 were families with daughters and 235 were families with sons. Most parents were immigrants, with 94.9% of mothers and 81.3% of fathers were born in Mexico, and most children were US-born, with 71% of daughters and 63% of sons born in the United States. At the first time point (child in fifth grade), mothers reported an average annual income of $30,001–35,000 (SD = $20,000, range = less than $5,000 - more than $95,001) and fathers reported an average annual income of $35,001–40,000 (SD = $20,000, range = less than $5,000 - more than $95,001). Additionally, at the first time point, the average highest educational degree completed was ninth grade for mothers (M = 9^th grade, SD = 3.67 years of education, range = no education - Master’s degree) and fathers (M = 9^th grade, SD = 4.08 years of education, range = no education - Doctorate degree). See Table 1 for detailed breakdowns of demographic information.

Table 1.

Participant Demographic Information (N = 471)

Categorical Variables
Variable Name	Category Levels	Percent of Sample
Parent Nativity	One parent born in Mexico	6%
	Two parents born in Mexico	93%
	Missing	1%
Mother’s Education Level	Elementary/middle school education	43%
	Some high school education	23%
	GED	2%
	Completed high school	13%
	Some college, vocational, or technical school education	11%
	4-year college degree	4%
	Graduate school	0.2%
	Missing	4%
Father’s Education Level	Elementary/middle school education	34%
	Some high school education	21%
	GED	0.7%
	Completed high school	16%
	Some college, vocational, or technical school education	9%
	4-year college degree	3%
	Graduate school	0.3%
	Missing	16%
Continuous Variables
Variable Name	M (SD)
Mothers’ Time Spent in US (yrs)	14.01 (8.40)
Fathers’ Time Spent in US (yrs)	18.16 (9.76)

Measures

Traditional gender role attitudes.

Parents’ and adolescents’ gender role attitudes were measured with a subscale (5 items) of the Mexican American Cultural Values Scale (MACVS; Knight et al., 2010) specifically designed to tap traditional gender role attitudes. This measure was completed every two years, beginning at the first time point (fifth grade). Adolescents completed the scale through the seventh time point (eleventh grade), but parents were administered the scale through the fifth time point (ninth grade). Participants responded on a 4-point scale indicating the degree to which they agreed with the statement. Responses to the five items were averaged to create a single score for each participant, with higher scores indicating more traditional gender attitudes. The full MACVS scale was developed based on focus groups conducted with Mexican-origin mothers, fathers, and adolescents, with each subscale representing one value central to Mexican American or Anglo American culture (Knight et al.,2010). The specific subscale used in this research was the traditional gender roles value, representative of Mexican American culture. Items touch on various traditional gender attitudes, including breadwinner/caregiver dynamics (e.g., “Men should earn most of the money for the family so women can stay home and take care of the children and the home) and marianismo/machismo beliefs (e.g., “Families need to watch over and protect teenage girls more than teenage boys”). The scale has since been used with Mexican-origin populations, producing acceptable to strong reliability in adult and adolescent samples (Knight et al., 2010). The scale demonstrated acceptable reliability with this sample across waves (α_adolescent = .63, .73, .74, .75; α_mother = .71, .70, .69; α_father = .70, .72, .69).

Ethnic pride.

Adolescents’ ethnic pride was measured with the Mexican-American Ethnic Pride scale developed for the overall project from which this research stems. Most items in the measure were pulled from a scale developed by Phinney (1992), with two items extracted from a scale that has been used successfully in other studies with Mexican-American adolescents (Berkel et al., 2010; Thayer, Valiente, Hageman, Delgado, & Updegraff, 2002). The measure was designed to tap identification with Mexican culture, including both the degree to which adolescents participate in cultural traditions and the degree to which adolescents hold positive feelings about their ethnic group. Adolescents responded to nine items (e.g., “You like people to know that your family is Mexican.”) on a 4-point scale indicating their agreement with each statement. Adolescents completed the measure every two years, beginning at the first time point (fifth grade). The scale produced acceptable reliability across waves with the current sample (α = .77, .84, .85, .87). Individual item responses were averaged to create a single score for each participant, with higher scores indicating greater ethnic pride.

Demographic information.

At Wave 1, parents reported their own and their child’s nativity, their time spent in the United States (when applicable), their family’s annual income, and their education level. Nativity was recorded as the country of birth, either United States or Mexico. Parents’ time spent in the United States was collected as an open-ended variable, with parents reporting the number of years since immigration to the United States. Annual income was reported in increments of $5,000, with a possible range of “less than $5,000” to “more than $95,001.” Parents’ education level was reported as the highest level of education completed, with a possible range of “1^st grade” to “last year of doctorate or other advanced degree.” Mothers’ and fathers’ reports of annual income and education level were z-scored and averaged to create the composite socioeconomic variable used in analyses.

Results

Plan of Analyses

The current study used growth models to examine the trajectories of adolescents’ gender role attitudes. Models were conducted within a multilevel modeling (MLM) framework using the Linear and Non-Linear Mixed-Effects Models (nlme) package in R (Pinhiero, Bates, DebRoy, Sarkar, & Team, 2017). For the current study, missing data represented 12.8% of all data points. Logistic regression models revealed no associations between missingness and exogenous constructs; therefore, missing data was handled using maximum likelihood estimation.

A series of two-level growth models were tested, with time nested within individuals. Time was included in the model as adolescents’ age in years, with age centered at 10 years old (the average age at the first wave of data collection). We used polynomial contrasts to describe linear and quadratic trajectories of adolescents’ gender attitudes. At Level 1, we included time- varying predictors: ethnic pride, mother’s gender attitudes, and father’s gender attitudes. To assess both within-person and between-person effects, each time-varying predictor was included as two separate variables (Hoffman & Stawski, 2009). The first variable, entered at Level 1, assessed within-person effects. This variable was calculated by centering the time-varying predictor on its group mean (i.e., each time point’s score minus the cross-time average for that individual). The second variable, entered at Level 2, assessed between-person effects. This variable was calculated by centering the time-varying predictor on the grand mean (i.e., each participant’s cross time average minus the sample mean). At Level 2, we also entered time-invariant variables that differed across individuals. These included adolescent gender (contrast coded; girls = 1, boys = −1), adolescent nativity (contrast coded; US = 1, Mexico = −1), and SES (z-score composite of parents’ education and income).

First, an unconditional means model was tested to determine the best-fitting growth model for the gender attitude trajectories (i.e., linear and quadratic effects). Next, predictor variables were entered in groups, with one group added for each sequential model: (1) the situating identity variables (gender, nativity, SES), (2) the cultural identity variable (ethnic pride), and (3) the contextual variables (mothers’ gender attitudes, fathers’ gender attitudes). See Table 2 for descriptive information and bivariate correlations for variables of interest.

Table 2.

Descriptive Statistics and Bivariate Correlations (wave 1) for Variables of Interest

Wave 1 study variables	1	2	3	4	5	6
1. Adolescent Nativity	-	.30	−.17	−.03	−.09	−.16
2. Socioeconomic Status	.25	-	−.15	−.29	−.40	−.09
3. Adolescent EP	−.01	.06	-	.11	.14	.09
4. Mother’s GRA	−.03	−.20	−.05	-	.19	.14
5. Father’s GRA	−.01	−.36	−.08	.06	-	.22
6. Adolescent’s GRA	−.05	−.26	.06	.24	.13	-
Gender	.14	−.02	.05	.06	−.02	−.10
% or M (SD) for Boys	71% US-born	-	3.55 (0.40)	2.49 (0.72)	2.61 (0.68)	2.68 (0.62)
% or M (SD) for Girls	63% US-born	-	3.58 (0.40)	2.59 (0.69)	2.63 (0.67)	2.55 (0.67)

Note. EP = ethnic pride. GRA = gender role attitudes. Gender is coded −1 = boys, 1 = girls. Nativity is coded −1 = Mexico-born, 1 = US-born. Correlations for boys appear above the diagonal and correlations for girls appear below the diagonal.

Unconditional Means Model

The unconditional means model yielded an intra-class correlation coefficient (ICC_between = 0.44) that showed considerable variance to be modeled at each level. This number suggests that 44% of the variance in adolescents’ gender attitudes was between-persons and 56% of the variance was within-persons.

Initial models tested whether random intercepts, random slopes, or random intercepts and slopes would best improve the model fit. We found that a random intercepts and slopes model best served the model, such that each individual would receive a predicted value for both their intercept and the shape of their trajectory. This determination was based on comparisons of relative fit statistics (i.e., AIC, BIC, −2 Log Likelihood).

Linear and quadratic effects of age were also tested at this point, with the best-fitting model yielding significant linear and positive quadratic slopes. To determine the points at which linear and quadratic trajectories were present, follow-up tests were conducted in which time was centered at different ages. We found that adolescents’ gender attitudes became less traditional from age 10 to age 16 (γ = −0.15, SE = 0.02, p < .001, 95% CI = [−0.18, −0.11]), but leveled off through age 18 (γ = 0.01, SE < .01, p < .001, 95% CI = [0.01, 0.02]).

Situating Identities

Situating identities (adolescent gender, adolescent nativity, and SES) were next added to the model. We chose these variables primarily based on previous data suggesting associations between gender attitudes and these key variables (e.g., Updegraff et al., 2014). We also tested (1) mothers’ and fathers’ time spent living in the US and (2) mothers’ and fathers’ nativity as potential control variables but found no significant associations with the outcome variable above and beyond those variables already included in the model. The model with situating identity variables revealed that boys demonstrated more traditional attitudes across time than did girls (γ = −0.10, SE = 0.02, p < .001, 95% CI = [−0.15, −0.05]). Adolescents from lower SES families reported more traditional attitudes than those from higher SES families (γ = −0.20, SE = 0.03, p < .001, 95% CI = [−0.26, −0.13]). The gender attitudes of adolescents born in the US did not differ from those born in Mexico. As an exploratory test of intersectionality effects, we examined the interactions between various situating identities, but found no significant effects. Thus, those interaction terms were dropped from the model. Separate models were tested to determine whether any situating identities interacted with the linear and quadratic growth patterns, with Bonferroni corrections applied to each test. These models revealed no significant interactions, meaning that the situating identities did not affect the shape of individuals’ trajectories.

Ethnic Identity Variable

The cultural identity variable (ethnic pride) was added in the next model. Because it was a time-varying predictor, ethnic pride was entered into the model as both a within-person and between-person effect. We found that ethnic pride was significantly related to adolescents’ gender role attitudes at the within-person level (γ = 0.13, SE = 0.04,p < .001, 95% CI = [0.05, 0.20]), but not at the between-person level. This effect indicates that at times when adolescents reported more ethnic pride, they also reported more traditional gender role attitudes. To explore ways in which ethnic identity worked in tandem with situating identity variables, we also included interactions between ethnic pride and gender, nativity, and SES. We did not find evidence of interaction effects; therefore, interaction terms were dropped from the model.

Contextual variables

Contextual variables (mother’s gender attitudes and father’s gender attitudes) were added in the final model. As aforementioned, the effects of both parents were included in the model twice: once as within-person effects and once as between-person effects. Notably, assessments of parent attitudes were only completed at the first three time points, meaning that within-person effects may only be interpreted for the first three time points. The between-person effects, however, were calculated as aggregate scores across those three waves and used to predict adolescents’ trajectories through all four time points. We found that mothers’ and fathers’ gender role attitudes affected their children’s, but only at the between-person level. Mothers’ gender role attitudes significantly affected adolescents’, such that mothers with more traditional gender attitudes tended to have adolescents with more traditional gender attitudes (γ = 0.24, SE = 0.05, p < .001, 95% CI = [0.15, 0.33]). The same effect was found for fathers’ attitudes, where fathers with more traditional attitudes tended to have adolescents with more traditional attitudes (γ =0.13, SE = 0.05,p < .001, 95% CI = [0.03, 0.22]). To test our hypothesis that mothers and fathers may differentially affect boys’ and girls’ attitudes, we included interactions between mothers’ and fathers’ attitudes and adolescent gender. We found no significant interaction effects, suggesting that mothers’ and fathers’ attitudes do not differentially affect their sons’ and daughters’ gender attitude trajectories. To examine whether the effect of one parents’ attitudes depended on the attitudes of the other parent, we also included interactions between mothers’ and fathers’ gender role attitudes; however, we did not find evidence of interaction effects. See Table 3 for information on all model effects.

Table 3.

Gamma Coefficients (y) and Standard Errors (SE) for Growth Models of Adolescents’ Traditional Gender Role Attitudes (N = 471)

	Situating Identities Model		Cultural Identity Model		Contextual Model
	γ	SE	γ	SE	γ	SE
Intercept	2.61***	0.03	2.61***	0.03	2.62***	0.04
Linear Time	-0.15***	0.02	-0.15***	0.02	-0.15***	0.03
Quadratic Time	0.01***	< 0.01	0.02***	< 0.01	0.01†	0.01
Adolescent Gender	-0.10***	0.02	-0.10***	0.02	-0 09***	0.03
Adolescent Nativity	−0.01	0.03	−0.01	0.03	−0.02	0.03
SES	−0.20***	0.03	−0.20***	0.03	−0.01**	0.04
EP (BP)			0.04	0.07	−0.01	0.08
EP (WP)			0.12***	0.04	0.11*	0.06
Mother’s GRA (BP)					0 24***	0.05
Mother’s GRA (WP)					0.01	0.05
Father’s GRA (BP)					0.13**	0.05
Father’s GRA (WP)					0.01	0.05
Mother’s GRA (BP) * Father’s GRA (BP)					0.06	0.07
Mother’s GRA (WP) * Father’s GRA (WP)					0.15	0.19
Mother’s GRA (BP) * Gender					0.04	0.04
Mother’s GRA (WP) * Gender					0.03	0.05
Father’s GRA (BP) * Gender					−0.04	0.05
Father’s GRA (WP) * Gender					0.01	0.05

Note. WP = within-person. BP = between-person. EP = ethnic pride. GRA = gender role attitudes. Gender is coded −1 = boys, 1 = girls. Nativity is coded −1 = Mexico-born, 1 = US-born.

^*p < .05,

^**p < .01,

^***p < .001,

^†= p < .10

Discussion

The current study sought to expand the current understanding of gender role attitude development by investigating the unique contributions of situating identities, cultural identities, and context on the trajectory of Mexican-origin adolescents’ gender attitudes. Findings demonstrated that both identities and context hold importance for gender role attitude development: SES and gender emerged as significant situating identities, ethnic pride was an influential cultural identity, and both mothers’ and fathers’ attitudes showed associations with their children’s gender attitudes. First, we will elaborate on specific study findings and will subsequently discuss the complexities in interpreting the relative importance of these identities and contextual factors, ending with implications for theory and practice.

The first growth model we tested focused on child characteristics related to situating identities, or social positions as described by García Coll and colleagues (1996). We predicted that adolescent gender would contribute to gender role attitude trajectories, with girls’ attitudes starting more egalitarian and becoming increasingly egalitarian compared to boys’. Although we found differences between girls’ and boys’ intercepts, with girls’ attitudes demonstrating less traditionality across time compared to boys’, we did not find differences in attitude trajectory by gender. That is, although girls reported less traditional attitudes overall than did boys, girls and boys showed similar declines in traditionality across adolescence. Previous research with Mexican-origin youth has found that girls are quicker to adapt to their host country’s relatively more egalitarian gender role ideology (Dasgupta, 1998; Updegraff et al., 2014). However, this research suggests that boys and girls may decline in traditionality at a similar rate. Importantly, the trajectory of gender role attitudes in our sample was nonlinear, revealing a decline in traditionality through age 16 years which levels off through age 18 years. These results are similar to those of Lam and colleagues (2017) and Crouter and colleagues (2007), who demonstrated similar linear and quadratic trajectories for African American and European adolescents’ gender roles across adolescence. Previous research with Mexican American adolescents (Updegraff et al., 2014) has examined only linear change; it is possible that once the nonlinear trajectories are modeled, the gender differences in slope are not as pronounced. However, future research is required to clarify these mixed findings.

A second situating identity predictor, SES, was expected to influence gender role attitudes, such that adolescents from higher SES families would report more egalitarian attitudes than those from lower SES families. Our findings supported this hypothesis, in line with prior research indicating that greater parent education level is associated with greater egalitarianism in children (e.g., Pepin & Cotter, 2018). A child’s SES may not only impact the ideologies to which they are exposed, with parents’ gender attitudes becoming more egalitarian with higher SES (Leaper & Valin, 1996), but could also affect their exposure to gender roles in the family. For example, Pinto and Coltrane (2009) found that higher SES was associated with less gender- segregation of household labor between spouses. Thus, a child from higher SES family who witnesses more egalitarian gender roles being enacted may be more likely to endorse attitudes in line with those gender role behaviors.

We also predicted that adolescent nativity would affect the development of gender role attitudes, such that adolescents born in Mexico would endorse more traditional attitudes than adolescents born in the US. Unlike previous research (e.g., Updegraff et al., 2014) which showed significant effects of nativity, our research found no difference between adolescents born in Mexico and those born in the US. Although unexpected based on prior research, this finding does reiterate the importance of studying the internalization of identities over and above the effects of demographic proxy variables. Nativity as a demographic variable may not be enough to characterize the ways in which country-specific norms (i.e., Mexican gender roles) are integrated into adolescents’ identities.

Following this logic, our next hypothesis focused on cultural identities. We predicted that ethnic pride would influence gender role attitudes, such that greater ethnic pride would be associated with greater endorsement of traditional gender role attitudes. Results provided evidence for this hypothesis by demonstrating a time-varying association between ethnic pride and gender role attitudes. It was found that as ethnic pride fluctuated, so did adolescents’ endorsement of traditional gender roles. In comparison to adolescents’ nativity, ethnic pride allowed for a more dimensional understanding of how cultural orientation might affect the endorsement of Mexican gender roles. In other words, this finding might reflect not just the time spent in the United States, but may reflect a child’s acculturation to the United States’ cultural values. When a given value type is intertwined with cultural orientation (e.g., gender attitudes; Glass & Owen, 2010; Kulis, Marsiglia, & Hurdle, 2003), acculturation-based measurements may emerge as a stronger predictor of those values than might nativity-based measurements. The current study demonstrates that gender role attitude development depends not where children are born, but on the sense of connection that they feel toward their ethnic group. In contrast, when studies rely on nativity to account for cultural orientation, researchers may be missing the nuances of the mechanisms or processes by which nativity matters.

Finally, we tested hypotheses related to the role of contextual influences on adolescents’ gender role development. We predicted that parents’ gender role attitudes would affect their children’s attitude trajectories, such that parents who had more traditional attitudes would have children who showed more traditional attitudes over the course of adolescence. Results showed a significant association for both mothers’ and fathers’ attitudes in line with our prediction. When mothers and fathers, on average, had more traditional attitudes, their children also had more traditional attitudes. This result, too, is in line with prior research on the effect of parents’ gender attitudes. In a meta-analysis examining the role of parent socialization in children’s gender development, Tenenbaum and Leaper (2002) found that parents’ attitudes were significantly positively related to their children’s attitudes, with differences in effect size found by age - for adolescent samples, the effect size was larger than for childhood samples. Prior research with Mexican-origin adolescents has also shown an association between parent and child gender attitudes (Updegraff et al., 2014), but found this association only for mothers. Especially considering that gendered information is embedded in immigrant parents’ more intentional ethnic socialization practices (e.g., Raffaelli & Ontai, 2004), it is not surprising that children who are socialized by gender-traditional parents would tend to adopt traditional gender role ideologies.

The hypothesis that mothers and fathers would have differential effects on their sons’ and daughters’ gender role attitudes was not supported by our findings. This is inconsistent with prior research with Mexican-origin immigrant adolescents (Updegraff et al., 2014), which showed that mothers’ gender role attitudes were significantly related to their adolescents’ attitudes, regardless of adolescent gender, whereas fathers’ attitudes were only related to their daughters’. In part, this may be due to the respective age range of samples in the two studies. In the current sample, adolescents’ gender role attitudes were measured across 5^th to 11^th grade, whereas in the Updegraff and colleagues (2014) study, the sample ranged from 7^th grade through college age. It may be that mothers and fathers play different socializing roles for their sons and daughters across these developmental periods, with fathers holding generally less on their children’s gender role attitudes in early adulthood compared to child and adolescent developmental periods. Additionally, the impact of parent gender socialization may manifest differently for sons and daughters; with boys potentially moving out of the house earlier than girls, there may be less room for fathers’ attitudes to be transmitted across generations. It may also be that fathers’ own attitudes are influenced by their child’s gender, with fathers of girls showing less traditional gender attitudes over time compared to fathers of boys (Shafer & Malhotra, 2011). If fathers with daughters tend to reduce their support for traditional gender roles over time, then the association between fathers’ and daughters’ attitudes may gradually become stronger, and that the association would be more easily detectable in early adulthood. Exploratory tests with the current data did not reveal differential effects of mothers’ and fathers’ attitudes across time. However, further longitudinal research with a wider age range could better discern the unique roles that mothers and fathers play in their children’s gender development from childhood through late adolescence and early adulthood.

The lack of differential influence from mothers and fathers is also surprising given the emphasis on mothers as “keepers of culture” in immigrant families (Qin, 2009; Stritikus & Nguyen, 2007). This family role obliges mothers to socialize their children toward the cultural norms of their home country, preserving cultural traditions, language learning, and other practices - including gender roles. The emphasis on women providing cultural generational continuity would suggest that mothers may take extra care in the ethnic socialization of their daughters, knowing that someday they would grow to be the “keepers of culture” for their own families. However, results from this study suggest that, at least in the domain of gender role attitudes, mothers are socializing their sons and daughters equally. Again, this finding may be due to the age range in this study. Perhaps adolescence is a time period in which passing on cultural gender role traditions is perceived as equally important for boys and girls. In any case, further research is needed to resolve the mixed findings in gender role attitude research.

Study Limitations

This study is not without limitations. Despite the diversity in family income and education level in the sample, this sample was not geographically diverse. Families in the current sample were recruited from a region in Northern California and may not generalize to immigrant families who are residing in other areas of the U.S. Especially with the recent geographic dispersion of immigrant families to new destination cities (Hall, 2013; Tienda & Fuentes, 2014), it is important to consider how family processes such as gender socialization may be affected by the ethnic makeup of the surrounding context. While the current research models the development of gender ideology in a region with a high density of Latino families, it may not explain this development in an area with a sparser Latino population. In that case, parents may be more acculturated and children may not have as much exposure to their parents’ cultural gender roles. Thus, caution should be taken when generalizing these results to families in those areas.

An additional limitation in the current study is methodologically related. The gender attitude scale used in this research was a subscale of the Mexican American Cultural Values Scale (Knight et al., 2010) and consisted of five items related to gender roles in the family. First, these items were developed with an adolescent population, whereas our sample was recruited in childhood. Psychometric analyses revealed a relatively lower reliability coefficient (α = 0.69) for the first time point, when participants were in fifth grade, which could have resulted from using a scale intended to target older audiences. Alternatively, the lower reliability could reflect a changing cultural orientation, which may have become more stable in adolescence (Knight et al., 2018). Secondly, although the family-related aspects of gender attitudes are undoubtedly important, they do not represent the entirety of culturally based gender-related attitudes. In practice, the range of gender-related attitudes extends to activities and occupations stereotyped for males and females, gender-related political views, women’s rights, and gender-stereotyped personality traits. Using a more comprehensive gender attitude survey (e.g., Gender Attitude Inventory, Ashmore, Del Boca, & Bilder, 1995; Occupations, Activities, and Traits scale, Liben & Bigler, 2002; Attitudes toward Women Scale, Spence, Helmreich, & Stapp, 1973) could have provided a more holistic understanding of which aspects of gender attitudes (beyond those related to family) may be transmitted across generations. This is certainly an important area for future research focusing on gender development among immigrant populations.

Finally, it is important to acknowledge the methodological limitations related to unequal waves of data available for parents and children. Because parents were assessed at fewer time points than were children, the resulting MLM included predictor variables with three waves and an outcome variable with four waves. Therefore, when interpreting findings, associations for within-person effects represent the relation between parents’ and children’s attitudes from only grades 5 to 9, and associations for between-person effects represent the relation between parents’ attitudes across grades 5–9 and children’s attitude trajectory from grades 5 to 11. It could be that during late adolescence (i.e., during grade 11), associations between parents’ and children’s attitudes show a different pattern than demonstrated here. Thus, further research is needed to explore the relation between parents’ and children’s gender role attitudes across the entire adolescent period.

Implications for Gender Development Research

Despite the aforementioned limitations, findings from the current study provide greater understanding of the ways in which situating identities, cultural identities, and contextual factors may affect children’s gender role development. As described earlier, identities do not develop in isolation, but instead are incorporated into the self-concept in overlapping ways (Knight et al., 2018). Within this framework, it is important to consider how the interpretation of one identity, such as gender, may be influenced by the development of another identity, such as ethnicity. The current research demonstrates that intersecting identities may inform each other, such that the centrality of ethnic identity may shape the values and attitudes formed about gender. It is important for future research on gender development to include constructs related to ethnic identity, considering the way in which gender ideologies derive meaning from ethnic-related values. This research should not only examine the concurrent associations between ethnic identity and gender attitudes, as in the current research, but should also investigate the ways in which ethnicity and gender interact together across development using cross-lagged models. Such research may better capture the mutual influences of adolescents’ multiple identities.

Beyond identity constructs, there are also important influences on gender development that stem from the ethnocultural context. When considering an immigrant child’s development, both the home and host cultures play an important role in the gendered information that a child internalizes. The complexity of multiple cultural influences is a consideration that is currently scarce from gender development theories but could serve as an insightful addition in teasing apart the roles of parents, peers, and other social contexts in children’s gender development. More generally, current gender development theories could benefit from studies that address whether parents, schools, and out-of-home contexts (i.e., neighbors) differ in the ways (e.g., how and what) gender role socialization takes place. This is not only relevant for children of immigrant parents, but also can be applied to families with different political or religious orientations, in comparison to the broader culture, considering the associations between political, religious, and gender attitudes (Bryant, 2003; Röder, 2014).

As is true for many aspects of child development, context is highly influential in the formation of children’s gender role attitudes. Gender role ideologies may exist on a variety of levels, in a fashion similar to the bioecological model of development (Bronfenbrenner &Morris, 2006), with levels of influence ranging from personal to cultural, political, and historical domains. In other words, a child’s development is dependent not only on individual factors, such as sensitivity to gendered information and personal gender typicality, but is also situated within certain contexts and historical time periods that endorse varying gender role ideologies. Importantly, the most proximal gender contexts (e.g., church messages) may or may not be in alignment with the gender norms of the broader culture or political regime, or those popular in the given historical time. Studying gender development in immigrant populations allows for a closer look at one example of such variations in embedded contexts, examining the impact of mismatched gender ideologies in a child’s immediate and remote contexts. The current study demonstrated the contribution that a person’s situating identities (i.e., gender, nativity, SES), cultural identities (i.e., ethnic pride), and immediate contexts (i.e., parents) have on the endorsements of gender role attitudes from multiple competing remote contexts (i.e., Mexican and US), highlighting the interplay between those situating contexts in the development of gender role attitudes. Future research should recognize the importance of context in gender role development, specifically in relation to multiple competing contexts.

Implications for Practice

Finally, in addition to considering the benefits to theory, it is also important to consider how current findings might impact prevention and intervention efforts. The current study found that ethnic pride influences Mexican-origin adolescents’ endorsement of gender role attitudes, with stronger ethnic pride predicting greater endorsement of traditional Mexican gender role attitudes. Considering that previous research has demonstrated that ethnic pride serves as a protective factor for immigrant children (Castro, Stein, & Bentler, 2009; Hernandez, Robins, Widaman, & Conger, 2017), it is thought-provoking to consider what the combination of these findings might mean for immigrant girls. On one hand, ethnic pride has been shown to decrease girls’ likelihood of risky sexual behavior, tied to greater endorsement of marianismo, but increased ethnic pride’s association with more traditional gender attitudes may also restrict their perceived opportunities for gender-atypical occupations and activities (Fulcher & Coyle, 2011; Harren, Kass, Tinsley, & Moreland, 1979). Research examining the impact of ethnic pride and other cultural variables on immigrant girls’ risky behaviors in combination with their gender-atypical opportunity pursuits would shed light on these potential outcomes.

In sum, the current study contributes to existing literature on gender role development among immigrant children and underscores the importance for the field of gender development to advance our understanding on how different situating (e.g., SES), cultural identities (e.g., ethnic identity), and contexts (e.g., family, schools) shape gender developmental domains (e.g., roles, attitudes, gender-related pressure). This focus is critical for the advancement of theory and practice given the embeddedness of gender development in the day-to-day experiences of children and youth across cultures and across contexts.