Changes in Ethnic Identity in Middle Childhood: Family and Neighborhood Determinants

Abstract

The present study examined changes in ethnic and racial identity (ERI) over one year among 353 Black and Latino early adolescents in relation to ethnic-racial socialization (ERS) and neighborhood economic and demographic characteristics. Parent and child reports of ERS were collected, and child ERI was assessed via self-report. Neighborhood measures included economic disadvantage and ethnic diversity. Results indicated parent report of preparation for bias was associated with greater increases in exploration but not among boys living in non-impoverished neighborhoods. In contrast, child report of preparation for bias was associated with decreases in belonging independent of child gender or neighborhood context. These findings highlight the complexities of neighborhood economic disadvantage and ERI development. Implications for future research examining gender differences and children’s interpretation of ERS messages are discussed.

Introduction

Ethnic and racial identity (ERI) has been defined as a clear understanding and commitment to one’s ethnic/racial group and a secure, confident sense of group membership (Roberts et al., 1999). The achievement of ERI is a critical developmental task for Black and Latino youth in the United States (U.S.). For Black adolescents, ERI attenuates the negative effects of racial discrimination and is associated with higher levels of academic self-efficacy and social competency (Rivas-Drake et al., 2014). Similarly, among Latino adolescents and emerging adults, higher levels of ERI are associated with lower rates of depression and higher social competence and school efficacy (Meca et al., 2022).

ERI is a multidimensional construct used by developmental scientists seeking to examine this nuanced process among minoritized individuals (Phinney, 1992; Sellers et al., 1998; Umaña-Taylor et al., 2014). In developing the Multigroup Ethnic Identity Measure (MEIM) and its revised version (Phinney & Ong, 2007), Phinney (1992) identified two conceptual dimensions of ERI: exploration and commitment. Exploration is the process of seeking experiences and information about one’s ethnic/racial group. Meanwhile, commitment (formerly affirmation and belonging) represents an individual’s sense of belonging to their ethnic/racial group and the positive feelings they have about being a member of this group. While Phinney’s conceptualization of commitment represents both an affective component of ERI and a developmental process of ERI, other approaches distinguish between the developmental processes of exploring and clarifying what one’s ethnic/racial group membership means to them (i.e., exploration and resolution; Umaña-Taylor, 2004) from the positive feelings felt toward one’s ethnic/racial group (i.e., private regard, Sellers et al., 1998). To align with how ERI was measured in this study, we use the term exploration to reference youth’s efforts to learn about their ethnic/racial background and its meaningfulness in their lives, and we use the term belonging to reference the concepts of affirmation, belonging, and commitment. Other ERI terms used are specific to the cited literature.

Informed by theories of ego identity development, Phinney identified four stages of ERI development: diffuse, foreclosed, moratorium, and achieved (Phinney, 1989). Achieved ERI is defined as a clear and secure acceptance and understanding of one’s own ethnicity marked by high levels of both exploration and belonging (Phinney, 1989). Achievement of ERI usually occurs by late adolescence or early adulthood (Phinney, 1989), although some scholars have identified ERI development as continuing into and throughout young adulthood (Williams et al., 2020).

Ethnic and Racial Identity Development through Ethnic-Racial Socialization

The ethnic-racial socialization (ERS) practices of minoritized parents significantly shape children’s ERI development (Hughes et al., 2006). ERS has been described as the process of learning the significance of race and ethnicity, membership and identity, social stratification, and inter- and intra-group interactions (Hughes et al., 2006). Hughes and Chen (1997) originally identified four dimensions of ERS across ethnic/racial groups, but two have become particularly common in cited literature: cultural socialization and preparation for bias (Umaña-Taylor & Hill, 2020). Cultural socialization encourages pride in one’s ethnic heritage by providing information about the history and traditions of their ethnic group. Preparation for bias is focused on increasing awareness of and strategies for coping with discrimination (Hughes et al., 2006).

Scholars have also conceptualized ERS constructs specific to children of certain ethnic/racial groups. For example, Stevenson (1995) identified spirituality and religion as key ERS constructs that both Black parents and adolescents use when coping with discrimination and racism. Meanwhile, parents of Latino adolescents and young adults discuss acculturation and maintenance of their native culture more frequently compared to Black families as part of the ERS process (Umaña-Taylor & Guimond, 2010). However, consistent across the ERS practices of all minoritized groups are messages socializing youth and alerting them to the existence of discrimination to prepare children for these experiences (Umaña-Taylor & Hill, 2020).

Ethnic-racial socialization (ERS) practices support ERI development and help prepare children to cope with experiences of racism, discrimination, and prejudice (Hughes & Chen, 1997; Neblett et al., 2002). For example, cultural socialization is associated with greater ethnic affirmation, self-esteem, and cultural connectedness for Black children and youth (Hughes et al., 2009; Peck et al., 2014). Cultural socialization in Latino families is also associated with positive outcomes for adolescents including positive ERI development (Umaña-Taylor et al., 2014). However, research on the effects of ERS messages that prepare youth for bias and how children and youth interpret these messages is mixed (Umaña-Taylor & Hill, 2020). Some report these messages are associated with more maladjustment in Black youth (Peck et al., 2014; Williams & Smalls-Glover, 2014). For Latino children, preparation for bias promotes positive adjustment while providing coping strategies against discrimination (Umaña-Taylor & Hill, 2020).

Ethnic and Racial Identity Development in Middle Childhood

Scholars agree that the process of ERI development begins prior to adolescence (Umaña-Taylor et al., 2014), although much of existing research focuses on adolescence (Phinney, 1992; Sellers et al., 1998; Umaña-Taylor et al., 2014). During pre-adolescence, many of the same dimensions of ERI are present, yet nascent, and begin to develop rapidly (Byrd, 2012). Examining ERI during this period can help elucidate the role internal and external influences on ERI development. Quintana (1998) posited that normative changes in cognitive functioning undergird substantive shifts in ERI across childhood and adolescence, an assertion supported by others (Byrd, 2012; Williams et al., 2020). For example, the second stage of ERI development (age 6–10) is heralded by the achievement of concrete operational thought (Piaget, 1974), which allows the child to reduce reliance on external, superficial indicators of ethnicity/race and supports the development of ethnic/racial constancy.

The age of youth in this study corresponds with Quintana’s third stage (age 10–14), which represents the age at which children are moving toward achievement of formal operational thought. In terms of ERI, children in this stage are increasingly aware of the social implications of ethnicity/race, such as socioeconomic differences between ethnic/racial groups and the likelihood of experiencing discrimination based on one’s ethnic/racial group. Selman (1980), as cited by Quintana (1998) describes this stage as self-reflective because “children can see themselves through the eyes of others” (p. 38). Children begin to consider what their ethnicity/race means for themselves and their place in the world. However, this developmental period is relatively understudied, and the data that do exist are mixed (Williams et al., 2020).

Neighborhood Context and the Development of Ethnically/Racially Minoritized Children

In their lifespan model of ERI, Williams et al. (2020) identified neighborhoods as a context that shapes the “identity-relevant experiences” of ERI development. Neighborhoods have been defined as “segregated localities that share similar sentiments, traditions, and history,” creating a shared culture between residents (Guo & Bhat, 2007, p.32). Neighborhood ethnic concentration and neighborhood-level economic disadvantage, influence ERI development in Black and Latino youth (Stevenson & Arrington, 2009; White et al., 2018a). Living in predominantly Black neighborhoods gives Black youth a safe and supportive place to develop racial attitudes and their own ERI (Byrd & Chavous, 2009; McBride Murry et al., 2011; Stevenson & Arrington, 2009; White et al., 2018b). For Latino children, living in ethnically matched neighborhoods allow youth to maintain their native culture resulting in higher levels of ERI exploration and belonging (Supple et al., 2006; White et al., 2018b; Witherspoon et al., 2021). Furthermore, cross-sectional research indicates that markers of neighborhood structural disadvantage (i.e., economic segregation) negatively impact ERI for both Black and Latino youth (Supple et al., 2006).

There is evidence that the socialization priorities of Black and Latino parents and their impacts differ depending on characteristics of neighborhoods. Higher levels of ERS have been observed in neighborhoods characterized by poverty and social/physical disorder (Caughy et al., 2006; Witherspoon et al., 2019). Furthermore, White and colleagues (2018a) found that parent ERS practices were unrelated to Mexican American adolescents’ ERI exploration when families lived in neighborhoods with more ethnically-matched families but with higher levels of exploration when families lived in neighborhoods with fewer ethnically matched families. For Black first graders, Caughy et al. (2006) reported that preparation for bias messages were associated with more behavior problems but only in neighborhoods marked by high social capital.

Gender Differences in Ethnic/Racial Socialization and Neighborhood Contexts

Differences related to child gender are evident in parents’ ERS practices, in children’s ERI development, and in the influence of neighborhood contexts. Parents of minoritized children engage in more cultural socialization with daughters and more preparation for bias with sons (Byrd & Chavous, 2009; Hughes et al., 2009), although not all studies report gender differences in ERS (Byrd & Ahn, 2020; Witherspoon et al., 2016). Regarding differences by child gender in the influence of ERS on ERI and of neighborhood contexts on parents’ ERS use, Umaña-Taylor and Guimond (2010) found ERS was more strongly associated with ERI development among Latina adolescent girls relative to Latino adolescent boys. Meanwhile, Black girls living in neighborhoods with less cultural diversity and reporting prior racism experiences reported receiving more cultural socialization from parents, whereas boys with matching neighborhood contexts and past experiences reported significantly less cultural socialization (Stevenson et al., 2005).

The Present Study

ERI development is a critical task for ethnically/racially minoritized youth, and individual differences in ERI development have significant implications for youth well-being. The purpose of this longitudinal study was to examine how family and neighborhood factors influence changes in ERI exploration and belonging in a sample of Black and Latino 10- to 11-year-old children. Children at this age are developing the social-cognitive skills necessary for self-reflection and identity development, but ERI development remains relatively unstudied prior to adolescence (Quintana, 1998; Williams et al., 2020). Neighborhoods are important contexts that shape identity-relevant experiences for minoritized youth. We focus on neighborhood ethnic/racial composition (White et al., 2018b) and neighborhood economic conditions (White et al., 2018a) to examine how these factors influence change in ERI during a period when children’s growing social-cognitive skills spur ERI development.

The current study was guided by three research questions. First, we examined whether parents’ ERS practices were associated with changes in ERI development in this sample of Black and Latino 10- to 11-year-old children. Aligning with prior research (Else-Quest & Morse, 2015), we predicted higher levels of cultural socialization would be associated with greater increases in both components of ERI. In contrast, we predicted preparation for bias would be associated with increases in exploration only (Hughes et al., 2009). Second, we examined how ERS practices and whether relations between ERS practices and ERI development differed depending on neighborhood context; specifically, by neighborhood poverty and ethnic/racial diversity. We predicted parents would report greater use of preparation for bias and cultural socialization messages when living in neighborhoods characterized by higher levels of poverty (Witherspoon et al., 2021) and greater cultural socialization in neighborhoods characterized by greater ethnic/racial diversity (Stevenson et al., 2005). We also examined whether the relation between ERS and ERI development differed by neighborhood context but did not establish a priori hypotheses regarding the direction of effects. Some researchers have reported that the relation of ERS to child outcomes is magnified when parents report more neighborhood social problems (e.g., less safety, vandalism; Witherspoon et al., 2021) while others report the impact of ERS is attenuated in such neighborhoods (Caughy et al., 2006). Finally, girls and boys are often differentially receptive to ERS messages (Caughy et al., 2006; Davis & Stevenson, 2006; Hughes et al., 2009), which can lead to distinct developmental outcomes. Although there is empirical evidence that neighborhoods moderate the impact of ERS differently for boys versus girls (Caughy et al., 2006; Stevenson et al., 2005), the available data are too limited to provide a basis for making predictions in the present study (Byrd & Chavous, 2009; Roosa et al., 2003).

Method

Participants

Participants were drawn from a larger longitudinal study of academic achievement among low-income Black and Latino children living in a metropolitan area in the southwestern United States. Eligibility criteria included (1) target child between ages 28 and 31 months of age; (2) at least one parent self-reported as either Black or Latino; (3) family income below 200% of the federal poverty level; (4) child hospitalized for less than 7 days at birth; and (4) family intended to remain in the area for at least one year. Recruitment was conducted by two middle-aged women well-known in the local community, one Black and one Latina. Fliers in both English and Spanish were distributed to agencies serving low-income families such as WIC clinics, recreational centers, and schools. In addition, families were recruited via word of mouth.

A total of 407 families were enrolled between November 2009 and February 2011. Families were enrolled when target children were 2.5 years of age (Time 1), and follow-up data were collected when children were 3.5 years old (Time 2), and in kindergarten (Time 3), first grade (Time 4), fourth grade (Time 5), fifth grade (Time 6), sixth grade (Time 7), and seventh grade (Time 8). The present study focuses on data collected at Time 5 and Time 6. Families excluded from follow-up beginning in Time 5 included children diagnosed with a significant developmental disability (n = 8, 3 Black; 5 Latino), families that voluntarily withdrew (n = 15, 6 Black; 9 Latino), and families not seen since Time 1 (n = 29, 20 Black; 9 Latino). Additionally, two Black children passed away before Time 5. Excluded families did not differ from those who were included in terms of parent education or family income. However, girls (14%) more likely to be withdrawn/lost-to-follow-up compared to boys (8%), χ² (1) = 4.04, p < .05.

Demographic characteristics for the remaining 353 families are displayed in Table 1. Black families were more likely to be living below the U.S. federal poverty level, while Latino primary caregivers were less likely to have completed high school. Among Latino caregivers, a little less than two-thirds were Spanish-dominant, and another third were Spanish-English bilingual. The majority of Latino caregivers were foreign-born and from Mexico.

Table 1.

Characteristic of Study Sample (N = 353).

	Child Ethnicity
	Black (N = 152)	Latino (N = 201)	χ2
	N (%)	N (%)
Child gender			0.39
Boy	86 (56.6)	107 (53.2)
Girl	66 (43.4)	94 (46.8)
Child race/ethnicity
Black	140 (92.1)	0 (0.0)	316.82***
Latino	0 (0.0)	167 (83.1)
Multiracial/multiethnic	12 (7.9)	34 (16.9)
Primary caregiver race/ethnicity			299.55***
African American	143 (94.1)	8 (4.0)
Latino	0 (0.0)	177 (88.1)
Multiracial/multiethnic	5 (3.3)	12 (6.0)
Other	4 (2.6)	4 (2.0)
Primary caregiver relationship to child			23.01***
Mother	127 (83.6)	197 (98.0)
Grandmother	13 (8.6)	2 (1.0)
Father	9 (5.9)	2 (1.0)
Aunt/Uncle	2 (1.3)	0 (0.0)
Primary caregiver’s level of education			37.31***
Less than high school	22 (14.5)	84 (41.8)
High school/GED	68 (40.8)	70 (34.8)
More than high school	62 (40.8)	47 (23.4)
Average family poverty level (Times 1–4)			66.32***
Less than 100%	70 (46.8)	20 (10)
100–149% federal poverty level	15 (9.9)	54 (26.9)
150%+ federal poverty level	65 (42.8)	127 (63.2)
Latino families only
Primary caregiver language
Spanish-dominant	—	113 (56.2)
English-dominant	—	11 (5.5)
Bilingual	—	63 (31.3)
Missing	—	14 (7.0)
Primary caregiver nativity
U.S. born	—	54 (26.9)
Foreign born	—	144 (71.7)
Country of origin (foreign-born caregivers)
Mexico	—	140 (69.7)
Central America	—	6 (3.0)
South America		1 (0.5)

^*p < .05;

^***p < .001.

Procedure

The current analysis utilizes data collected at Time 5 (the first time point for which ERI data were collected), when target children averaged 10.3 years old (SD = .26, range 9.6 – 11.1), and Time 6, when they averaged 11.1 years old (SD = .30, range 10.3 – 11.8). Starting at Time 5, the study adopted a planned missing design. Participants were randomly selected to complete specific data collection time points as well as randomly assigned to complete data collection during a home visit or over the phone. Home visits were completed by two research assistants and lasted approximately 1.5 hours. Telephone interviews lasted approximately one hour. For both home visits and telephone interviews, youth completed questionnaires on a tablet, and a research assistant provided help by reading the questions if needed. For Spanish-speaking parents and children, data collectors for both home visits and phone interviews were fully bilingual. At Time 5, 255 families were selected for participation, and data collection was completed with 168 parents and 167 children. There were no differences between families that did not complete data collection at Time 5 and the families that did in terms of family income or child gender. However, families who completed data collection were more likely to have a high school degree or higher, 70.2% versus 29.8%, χ² (3) = 8.05, p < .05. At Time 6, 266 families were randomly selected for participation, and data collection was completed with 156 parents and 150 children. There were no differences between the families that did not complete data collection at Time 6 and the families that did in terms of parent education, family income, or child gender.

Measures

Ethnic and Racial Identity.

Child report was used to assess ERI at both Time 5 and Time 6 using the Multigroup Ethnic Identity Measure (Phinney, 1992; Roberts et al., 1999). The MEIM consists of 12 items in two subscales: affirmation, belonging, and commitment (referred to as belonging, seven items) and ethnic identity search (referred to as exploration, five items). All items used a four-point Likert scale ranging from 1 (Strongly Disagree) to 4 (Strongly Agree).

Confirmatory factor analyses of the MEIM at Time 5 and Time 6 indicated the two-factor model fit well. At Time 5, model fit indices were: χ² (42) = 89.15, p = .06, CFI = .91, TLI = .88, RMSEA = .085, and SRMR = .060. At Time 6, model fit indices were: χ² (51) = 67.00, p = .06, CFI = .94, TLI = .93, RMSEA = .046, and SRMR = .051. Longitudinal invariance tests supported strong factorial invariance across the two time points (see supplemental material).

Ethnic/Racial Socialization.

Both parent-report and child-report measures of ERS were collected at Time 5 using survey measures developed by (Hughes & Chen, 1997). Parents responded to five items about cultural socialization practices and six items about preparation for bias practices. If a parent responded “yes” to an item, the follow-up question asks the parent to report the frequency of the behavior in the past year on a six-point Likert scale ranging from 1 (Never) to 6 (Eight Times or More). The two parts of each item were combined to form a seven-point Likert scale ranging from 0 (Never) to 6 (Eight Times or More in the Last Year). We utilized two subscales from the child reported measure of ERS at Time 5: cultural socialization (nine items) and preparation for bias (six items; Hughes & Johnson, 2004; Hughes et al., 2006). Children responded to questions on a six-point Likert scale ranging from 1 (Never) to 6 (Eight Times or More).

Confirmatory factor analysis of ERS for both the parent- and child-report measures indicated the two-factor model fit well. For parent report, model fit indices were: χ² (5) = 18.13, p = .002, CFI = .99, TLI = .97, RMSEA = .09, and SRMR = .02. For child report, the model fit indices were: χ² (6) = 17.35, p = .008, CFI = .99, TLI = .98, RMSEA = .07, and SRMR = .03.

Neighborhood Concentrated Economic Disadvantage.

Family addresses at Time 5 were geocoded and linked to census tract data that included the unemployment rate, proportion of female-headed households with children, percentage of households below the poverty level, and the proportion of families receiving supplemental nutrition assistance program (SNAP) benefits. Census indicators were standardized against the state and averaged to create an index of neighborhood economic impoverishment (Korbin & Coulton, 1997; Sampson et al., 1997). The internal reliability for the index was .76, and higher values indicated high levels of concentrated economic impoverishment. Based on previous experience that the moderating impact of neighborhood economic impoverishment is not linear, rather a threshold with family level processes operating differently in high poverty neighborhoods, neighborhood economic disadvantage is included in multivariate models as a binary variable.

Neighborhood Ethnic Diversity.

Neighborhood ethnic/racial diversity was calculated as the sum of the pairwise disparities between groups, each weighted in proportion to that group’s individual contribution, using a formula proposed by Stirling (2007). Higher values indicate greater neighborhood ethnic/racial diversity. Neighborhood diversity is retained as a continuous variable because we have no a priori expectations of threshold effects.

Average Family Income-to Needs Ratio During Early Childhood.

From Time 1 to Time 4, household composition and income data were used to calculate family income-to-needs ratios and averaged to create an estimate of family economic resources during early childhood. Household income data were not collected at Time 5 or Time 6.

Analysis Methods

Missing Data Imputation.

Before conducting analyses to address the study aims, missing data were imputed using auxiliary variables derived from a block of study variables using principal components analysis (PCA; Howard et al., 2015; Stevenson et al., 2005). Specifically, approximately 100 auxiliary variables representing all domains of data across all time points of the larger study as well as interactions between auxiliary variables were used to estimate a reduced set of components. These PCA-derived variables were used to impute 100 data sets of the study variables. The quality of the imputations was verified by confirming the distributions of imputed data were consistent with the observed data. A final grand mean dataset was created after imputations were deemed to be satisfactory.

Analyses to Address Research Questions

Descriptive statistics were generated using the SPSS Missing Values module. Next, Mplus (Muthén & Muthén, 1998–2017) was used to fit multivariate models. Specifically, to address whether ERS was related to changes in ERI development, we regressed ERI at Time 6 on ERS at Time 5 while adjusting for ERI at Time 5 as well as family income-to-needs ratio. Next, we examine whether the relation between ERS and changes in ERI differed by neighborhood and/or child gender. First, we fit a series of models to examine the main and interactive effects of neighborhood ethnic/racial diversity and each ERS construct. To examine the impact of neighborhood economic impoverishment, we ran a multiple group model stratified by high neighborhood economic impoverishment versus less neighborhood economic impoverishment. High poverty neighborhoods were defined as those at or above 1.5 standard deviations above the mean of neighborhood economic impoverishment for the state overall. Approximately 29% of the children in the present study were living in high poverty neighborhoods. Differences in structural parameters by neighborhood type were tested systematically using a nested model approach and selectively fixing specific parameters to be equivalent across groups.

Results

Bivariate correlations between the ERI subscales (exploration and belonging), neighborhood characteristics, and parent and child report of parents’ ERS strategies are reported in Table 2. Because these analyses are based on a multiply imputed data set, the magnitude of correlations are best evaluated using guidelines provided by Cohen (1988): |.10-.30| as weak, |.30-.50| as moderate, and |>.50| as strong. Neither neighborhood economic impoverishment nor ethnic/racial diversity were significantly correlated with exploration and belonging concurrently or one year later. Parent report of cultural socialization was moderately positively correlated with concurrent measures of child exploration and belonging and weakly positively correlated with child exploration and belonging one year later. In contrast, child report of cultural socialization was strongly correlated with concurrent measures of child belonging and exploration. Furthermore, child report of cultural socialization was unrelated to exploration one year later but positively correlated with child belonging one year later.

Table 2.

Intercorrelations and Descriptives of Study Variables.

		1	2	3	4	5	6	7	8	9	10	11
Ethnic Identity
1	Exploration (T5)	1.00
2	Belonging (T5)	.984	1.00
3	Exploration (T6)	.072	.083	1.00
4	Belonging (T6)	.189	.205	.936	1.00
Neighborhood characteristics
5	Concentrated economic disadvantage (T5)	.065	.024	−.036	−.086	1.00
6	Ethnic diversity	.076	.070	−.021	.030	−.103	1.00
Ethnic racial socialization (parent)
7	Cultural socialization (T5)	.328	.244	.158	.181	−.151	.035	1.00
8	Preparation for bias (T5)	.057	−.006	.261	.174	−.064	−.041	.634	1.00
Ethnic racial socialization (child)
9	Cultural socialization (T5)	.849	.847	.096	.206	.002	.079	.244	.059	1.00
10	Preparation for bias (T5)	.440	.370	.025	.015	.037	−.013	.345	.249	.483	1.00
Covariates
11	Family income-to-needs ratio	.053	.063	−.132	−.061	−.072	.143	.004	−.072	.002	−.028	1.00
	Mean	2.05	2.40	2.04	2.43	.79	.07	2.59	1.45	3.07	2.01	0.86
	SD	0.69	0.91	0.70	0.82	1.36	.03	1.30	1.13	1.87	1.01	0.56

Note. T5 = the first time point for which ERI data were collected; M age = 10.3 years old and Time 6 = second time point for which ERI data were collected; M age = 11.1 years old.

Parent report of preparation for bias was unrelated to concurrent child exploration and belonging but moderately positively associated with exploration and weakly positively associated with belonging one year later. Child report of preparation for bias showed the opposite association, being positively correlated with concurrent measures of exploration and belonging but unrelated to these same measures one year later.

We recognized the correlation between the two components of ERI within time points exceed .90 at both Time 5 and Time 6. To examine whether these constructs were separate dimensions, we constrained these estimates to a perfect correlation (1.00). This approach fundamentally compares a 1-factor and 2-factor solution. Chi-square difference tests between these competing models indicated the original model with correlations above .90 fit the data better than correlations constrained to 1.00, supporting a two-factor model of ERI.

Neighborhood Diversity and Ethnic and Racial Identity Development

To examine whether neighborhood ethnic/racial diversity moderated the relation between ERS and ERI development, we regressed Time 6 exploration and belonging on neighborhood diversity, Time 5 cultural socialization and preparation for bias, and the interaction variable between neighborhood diversity while controlling for Time 5 exploration and belonging. Results are displayed in Table 3. In Model 1, we test the direct effect of neighborhood diversity as well as the effects of both child-reported and parent-reported cultural socialization and preparation for bias. In Model 2, we test the interaction between cultural socialization, and neighborhood diversity, and in Model 3, we test the interaction between preparation for bias and neighborhood diversity. The model fit indices for Model 1 were: χ² (248) = 786.44, p < .001, CFI = .93, TLI = .92, RMSEA = .08, SRMR = .07. As shown in Table 3, neighborhood diversity was not related to change in exploration and belonging nor did it moderate the relation between cultural socialization and preparation for bias and change in exploration and belonging. In addition, we fit a multi-group model grouped by child gender to examine whether these relations differed between boys and girls, which they did not. (Results available upon request).

Table 3.

Structural Equation Model Results Regressing Time 6 Ethnic Identity on Neighborhood Diversity and Time 5 Parent-Report and Child-Report Ethnic Racial Socialization.

	Model 1		Model 2		Model 3
	b (se)	t	b (se)	t	b (se)	t
Exploration (T6)
Cultural socialization (Parent report)	−0.02 (0.10)	−0.23	−0.11 (0.18)	−0.62	−0.02 (0.10)	−0.15
Preparation for bias (Parent report)	0.31 (0.09)	3.26**	0.32 (0.10)	3.07**	0.33 (0.20)	1.68+
Cultural socialization (Child report)	0.16 (0.15)	1.02	−0.08 (0.24)	−0.31	0.14 (0.17)	0.83
Preparation for bias (Child report)	−0.11 (0.08)	−1.40	−0.10 (0.09)	−1.07	−0.26 (0.17)	−1.54
Exploration (T5)	−0.01 (0.16)	−0.07	0.01 (0.18)	0.03	0.00 (0.18)	0.01
Neighborhood ethnic diversity	−0.69 (1.88)	−0.37	−0.58 (1.98)	−0.29	−0.92 (1.87)	−0.50
Diversity × cultural socialization (parent)	—	—	1.20 (1.98)	0.61	—	—
Diversity × cultural socialization (child)	—	—	2.88 (2.30)	1.26	—	—
Diversity × preparation for bias (parent)	—	—	—	—	−0.22 (2.16)	−0.10
Diversity × preparation for bias (child)	—	—	—	—	2.04 (1.96)	1.04
Belonging (T6)
Cultural socialization (Parent report)	0.07 (0.09)	0.76	−0.06 (0.17)	−0.34	0.07 (0.10)	0.72
Preparation for bias (Parent report)	0.17 (0.09)	2.00*	0.18 (0.10)	1.91+	0.06 (0.17)	0.36
Cultural socialization (Child report)	0.18 (0.13)	1.34	−0.08 (0.22)	−0.37	0.18 (0.13)	1.35
Preparation for bias (Child report)	−0.18 (0.08)	−2.32*	−0.16 (0.08)	−1.93+	−0.24 (0.15)	−1.61
Belonging (T5)	0.12 (0.12)	0.94	0.14 (0.12)	1.16	0.12 (0.12)	0.96
Neighborhood ethnic diversity	0.45 (1.78)	0.25	0.54 (1.87)	0.29	0.14 (1.79)	0.08
Diversity × cultural socialization (parent)	—	—	1.85 (1.84)	1.00	—	—
Diversity × cultural socialization (child)	—	—	3.16 (2.08)	1.52	—	—
Diversity × preparation for bias (parent)	—	—	—	—	1.58 (1.83)	0.86
Diversity × preparation for bias (child)	—	—	—	—	0.93 (1.74)	0.54

⁺p < .10;

^*p < .05,

^***p < .001.

Neighborhood Economic Impoverishment and Ethnic and Racial Identity Development

Before examining whether neighborhood poverty moderated the relation between cultural socialization and preparation for bias and exploration and belonging, we conducted invariance testing to confirm the measurement model for cultural socialization and preparation for bias and child exploration and belonging across four groups: 1) boys in non-impoverished neighborhoods, 2) boys in impoverished neighborhoods, 3) girls in non-impoverished neighborhoods, and 4) girls in impoverished neighborhoods. Configural, weak factorial, and strong factorial invariance assumptions were marginally met (see supplemental material).

Using a latent variables framework in Mplus (Muthén & Muthén, 1998–2017), we examined how cultural socialization and preparation for bias at Time 5 predicted changes in child exploration and belonging from Time 5 to Time 6 and whether associations differed by neighborhood poverty and/or child gender. Results are shown in Table 4. The fit of the final model was acceptable, χ² (1031) = 2085.811, p < .001, CFI = .88, TLI = .87, RMSEA = .108, SRMR = .099. There was no indication of modifications to appreciably improve model fit.

Table 4.

Multigroup Structural Equation Model Results Regressing Time 6 Ethnic Identity on Time 5 Parent-Report and Child-Report Ethnic Racial Socialization.

	Non-lmpoverished Neighborhoods				Impoverished Neighborhoods
	Boys		Girls		Boys		Girls
	b (se)	t	b (se)	t	b (se)	t	b (se)	t
Exploration (T6)
Cultural socialization (Parent report)	ns	—	ns	—	ns	—	ns	—
Preparation for bias (Parent report)	ns	—	0.32 (.08)	3.84***	0.32 (.08)	3.84***	0.60 (0.12)	4.77***
Cultural socialization (Child report)	ns	—	ns	—	ns	—	ns	—
Preparation for bias (Child report)	ns	—	ns	—	ns	—	ns	—
Exploration (T5)	ns	—	ns	—	ns	—	ns	—
Belonging (T6)
Cultural socialization (Parent report)	ns	—	ns	—	ns	—	ns	—
Preparation for bias (Parent report)	ns	—	0.31 (0.08)	3.77***	0.23 (0.09)	2.46*	0.31 (0.08)	3.77***
Cultural socialization (Child report)	ns	—	ns	—	ns	—	ns	—
Preparation for bias (Child report)	−0.08 (0.04)	−2.14*	−0.08 (0.04)	−2.14*	−0.08 (0.04)	−2.14*	−0.08 (0.04)	−2.14*
Belonging (T5)	0.09 (0.02)	3.83***	0.10 (0.03)	3.56***	0.09 (0.02)	3.83***	0.09 (0.02)	3.83***

Note. ‘ns’ indicates that the coefficient was not significant and was set to zero for the sake of model parsimony. Model fit indices: χ² (1031) = 2085.81, p =.000, CFI = .88, RMSEA = .11.

^*p < .05,

^***p < .001.

Coefficients predicting change in exploration are reported at the top of the table, while coefficients for belonging are reported on the bottom. Only parent report of preparation for bias at Time 5 predicted significant change in exploration between Time 5 and Time 6. More parent-reported preparation for bias at Time 5 was associated with a greater increase in child-reported exploration from Time 5 to Time 6, but only among boys and girls living in impoverished neighborhoods and girls living in non-impoverished neighborhoods. There were no significant predictors of change in exploration for boys living in non-impoverished neighborhoods.

Preparation for bias was also associated with greater increases in belonging, but associations differed by child gender, neighborhood poverty, and reporter (parent versus child). Parent-reported preparation for bias was associated with an increase in child-reported belonging but only among boys and girls in impoverished neighborhoods and girls in non-impoverished neighborhoods. Furthermore, this relation was stronger for girls in impoverished neighborhoods than it was for boys in these neighborhoods. In contrast, child report of preparation for bias was associated with declines in belonging. This association did not differ by neighborhood economic disadvantage or child gender. The magnitude of this association was smaller than the association of parent report of preparation for bias with change in ERI development.

Discussion

In this study, we examined the moderating effects of child gender and neighborhood contextual factors on the relation between ERS practices and ERI development in Black and Latino early adolescents over the span of one year. Based on prior research with Black and Latino adolescents (Byrd & Chavous, 2009; Roosa et al., 2003), we predicted cultural socialization would be associated with increases in ERI exploration and belonging (Else-Quest & Morse, 2015) and preparation for bias would be associated with increases in ERI exploration only (Hughes et al., 2009). While findings of the current study did not support our hypothesis about cultural socialization, preparation for bias was associated with changes in ERI exploration and belonging, albeit differentially depending on who was reporting on these messages. Specifically, while parent-reported preparation for bias was associated with increases in exploration and belonging for girls, overall, and for boys living in economically impoverished neighborhoods, child-reported preparation for bias was associated with declines in ERI belonging for all children in the sample regardless of their gender or their neighborhood’s economic conditions. We discuss these conflicting findings and the contributions they make next in light of the young age of this sample’s youth, the inclusion of two ethnic/racial groups, and with an eye for the future of ERS and ERI research in neighborhood contexts.

Findings Related to Cultural Socialization and Preparation for Bias Regardless of Child Gender or Neighborhood Context

Very little research has examined longitudinal effects of ERS messages on ERI development of Black and Latino youth prior to and during early adolescence. Based on the high correlation between belonging and exploration, it is plausible to assume that at this early developmental stage, children’s’ ERI development appears to be less differentiated that what studies with older youth have shown. Additional research is needed to understand these high correlations more explicitly. Specifically, cultural socialization messages may not prompt the kinds of self-reflective behavior theorized to promote ERI exploration and belonging for youth of this age (Quintana, 1998). In contrast, preparation for bias was associated with changes in both subscales of ERI, but the significance and direction of associations depended on the reporter. To our knowledge, no research has examined whether caregivers and youth interpret the preparation for bias messages that researchers list on ERS measures in the same way (Hughes et al., 2009). It may be that early adolescent children perceive preparation for bias messages differently from the intent of parents.

An example to help highlight the importance of youth perceptions of preparation for bias messages is a conversation common to Black families known as “The Talk,” based in preparation for bias socialization (Whitaker & Snell, 2016). In “The Talk,” Black parents discuss with their children topics like racial profiling, effective ways to interact with police officers, and how to diffuse negative stereotypes of Black Americans (Whitaker & Snell, 2016). Both Black and Latinx parents engage in this and other preparation for bias socialization with the intent of making their children aware of and prepared for racial inequalities in the world (Whitaker & Snell, 2016). Despite this, “The Talk” is associated with anger toward other ethnic/racial groups and feelings of not belonging in this world among Black adolescents (Whitaker & Snell, 2016). This suggests youth-report and not parent-report of preparation for bias may be associated with declines in feelings of ERI belonging. The version of the MEIM used in this study conceptualizes ERI belonging to include both feelings of belonging to one’s ethnic/racial group and positive feelings about being a member of this group. The sociopolitical context at the time of data collection (2017–2019) was marked by police violence toward Black individuals and violent deportation raids in Latino communities. Particularly for youth of this age who may have yet to achieve formal operational thinking, preparation for bias messages amidst a context of racial violence toward Black and Latino individuals may result in more discontent about belonging to their ethnic/racial group, not in terms of disliking Black or Latino culture but because of how these groups are treated in the U.S.

Parents can promote feelings of belonging and self-esteem in their children by emphasizing that experiences of discrimination are less about their personal attributes and more about other’s individual beliefs regarding specific ethnic/racial groups (Neblett et al., 2002). Our findings further emphasize the need for continued exploration of the impact of parent versus youth perceptions of preparation for bias messages with pre- and early adolescent youth to ensure that parents’ intentions with messages are aligned with their impact.

Preparation for Bias in the Context of Child Gender and Neighborhood Economic Impoverishment

Our results also indicated gender differences in the influence of preparation for bias messages, as parents’ reports of preparation for bias were associated with greater increases in exploration one year later for boys in impoverished neighborhoods and for girls in neighborhoods regardless of economic conditions. It is important to consider how experiences of living in poverty might incite curiosity in youth related to the economic disparities that exist in racially stratified societies (Chetty et al., 2020). While processing parents’ preparation for bias messages, children are simultaneously learning more about their ethnic-racial group membership and what it means to be a minoritized individual in the U.S. (Umaña-Taylor et al., 2014). Moreover, they may see confirmation of the biases and prejudices their parents warned them about in the circumstances of their neighborhoods and ‘turn toward’ their ERI, resulting in higher scores of ERI belonging.

It is important to note that the association of preparation for bias with increases in exploration was not found among boys in non-impoverished neighborhoods. Compared to girls, boys are more likely to move beyond what feels familiar culturally and engage with peer groups that are ethnically diverse (Way & Chen, 2000). Moreover, compared to living in contexts of neighborhood poverty, boys living in non-impoverished neighborhoods have access to greater resources and, therefore, have more opportunities to engage in racially integrated activities (i.e., community/school sports teams, youth organizations, etc.). As these youth engage with interracial groups more, they may also have an increased awareness of racialized and gender stereotypes. Being aware of societal perceptions attached to their race and gender identity may serve as a barrier against ERI exploration. Way & Rogers (2015) asserted that youth from all racial backgrounds engage with stereotypes in varied fashions as they form their identities, which can include challenging them or even accepting them. It is important to note that this finding is inconsistent with prior research linking greater neighborhood poverty and lower familial SES with less ERI exploration and belonging among Latino (Supple et al., 2006) and Black youth (Bennett, 2006). Determining how specific aspects of impoverished versus non-impoverished neighborhoods contribute to ERI exploration and belonging at this age awaits further explication.

Extant research has examined Black and Latinx children’s experiences of ERS and ERI development both together as well as separately while acknowledging distinct differences between these groups (Pasco et al., 2021; Witherspoon et al., 2021). In the present study, we examined Black and Latinx children together without the intent of comparing them due to differences in the historical experiences of these two groups within the U.S. These differences result in varying experiences in the context of ERI development for Black and Latino children as well as how parents engage in ERS practices with their children. Likewise, in the absence of a theoretical rationale for race/ethnic differences in the relation between neighborhood context, ERS and ERI, we believed it was not justified to examine differences between groups.

Limitations

The findings of the current study should be considered in the context of several important limitations. First, our measures of neighborhood were limited to concentrated economic disadvantage and ethnic/racial diversity and did not include measures of neighborhood physical or social characteristics such as physical incivilities and/or collective efficacy. Economic and ethnic/racial composition measures of neighborhoods do not capture important variation in neighborhood physical and social characteristics (Caughy et al., 1999). Therefore, although our findings did not indicate direct relations between neighborhood conditions and ERI development, future research should utilize more nuanced measures of neighborhood context. Additionally, due to the lack of findings regarding neighborhood ethnic/racial diversity in our sample, future research can benefit from further examination of the effects of neighborhood ethnic/racial diversity, particularly youth’s perceptions of their neighborhood, collective efficacy, and experiences of discrimination, as research has shown that neighborhood ethnic/racial diversity is important for youth’s overall development (White et al., 2018a). Since our measure of neighborhood diversity did not examine these factors, it may be limited in explaining the effects of neighborhood diversity for children of this age. Another possible explanation for our null findings could be the age of the children in our sample who were about four years younger than those in the White et al. (2018a) article. Perhaps the influence of neighborhood ethnic composition is not detectable until children are further along in their ERI development and/or are more able to navigate the neighborhood on their own. Second, our study was unable to examine cross-level interactions due to minimal clustering of families across census tracts as two-thirds of the children live in census tracts with no other study children, and 93% live in census tracts with two or fewer other participants. Future research should collect data across a number of diverse census tracts to increase variability within samples and examine cross-level interactions. Further, our study is based on parent and child self-reported ERS measures which could produce a social desirability bias due to the sensitive topic of race and discrimination. Additionally, although both parent-and child-report measures of ERS were utilized, the scales differed in substantive ways that did not allow for a direct comparison of the two scales, or for measurement invariance to be conducted to determine, in particular, whether items of the preparation for bias subscales were being interpreted differently by children and parents. Lastly, the analyses employed are fundamentally correlational, thus limiting our ability to draw conclusions regarding causation.

Future Directions

Despite the limitations identified, this study makes important contributions regarding the influence of neighborhood conditions, child gender, and parents’ ERS messages on the ERI development of Black and Latino early adolescents. Our research highlights the need to explore how youth make sense of parents ERS messages, in addition to the role that the environment plays in how children interpret these messages. Future examinations like ours may be enhanced by the inclusion of qualitative and other specific research methodologies. Because the research surrounding preparation for bias messaging is relatively mixed (Umaña-Taylor & Hill, 2020) in-depth qualitative data could be useful in broadening our perspectives around the effectiveness of ERS messages, particularly preparation for bias, from the perspective of minoritized children.

Additionally, conducting a related study of ERI development that includes both the Multigroup Ethnic Identity Measure (MEIM; Phinney, 1989) along with the Multidimensional Measure of Racial Identity could be useful in further elucidating the experiences of minoritized children within neighborhoods (Sellers et al., 1998). While there are similarities across the two measures, there are also subtle differences. The MEIM specifically focuses on the four stages of ERI development, while the MMRI examines racial centrality and regard. Sellers et al. (1998)’s measure of racial centrality and regard can extend our understanding of the identity process for minoritized youth, as it examines both positive and negative public regard (i.e., how Black Americans believe other people view their racial group) and private regard (i.e., how Black Americans view their own racial group). In addition, the MMRI takes ERI development a step further by examining additional racial ideologies that are not explored by Phinney (1989) in the MEIM. Including both the MEIM and MMRI in a single study could provide more nuanced results that could advance our understanding of these complex identity development process.

Finally, our study highlights the important contribution of the environment for early ERI development while also indicating the complexities of neighborhood economic disadvantage for children. Future studies should further examine gender differences in neighborhood contexts and ERI development as well as the ways that minoritized youth make meaning about their experiences in communities in relation to their understanding of race and ethnicity in society. Continuing to explore the complicated relation between ERS, ERI development, and the neighborhood environment for minoritized children would further understandings of how these environments promote or hinder minoritized children’s overall development.

Biographies

LaRen Morton, M.S. is a fifth-year doctoral student and graduate research assistant at the University of Georgia in the Department of Human Development and Family Studies. As a graduate research assistant, LaRen currently works on a longitudinal research project examining cultural, familial, and contextual factors that impact Black and Latinx children during their transition to middle school. Her research interests encompass gender differences in educational contexts and the effects of the school environment on minoritized children’s social, emotional, and academic development.

Leslie A. Anderson, PhD, LMFT is an Assistant Professor of Child and Family Sciences and Marriage and Family Therapy at The University of Southern Mississippi. Her research broadly focuses on Black American familial processes and specifically, their processes of racial socialization. Her scholarship is rooted in a commitment to social justice and undergirded by Critical Race Theory.

Margaret O’Brien Caughy, Sc.D. is the Georgia Athletic Association Professor in Family Health Disparities in the Department of Human Development and Family Science at the University of Georgia. Dr. Caughy’s research combines the unique perspectives of developmental science, epidemiology, and public health in studying the contexts of risk and resilience affecting young children with a specific focus on race/ethnic disparities in health and development and how these disparities can be understood within the unique ecological niches of ethnic minority families. Dr. Caughy has been the principal investigator of a number of studies focused on how inequities in family and community processes affect the cognitive development, socioemotional functioning, and early academic achievement of young children in diverse race/ethnic groups.

Omolola Odejimi, PhD is a senior postdoctoral research fellow at Children’s Hospital Los Angeles. Her main research interests are developmental trends in youth and measuring positive youth development in relation to academic performance and behavior in primary and secondary school settings.

Kimberly Osborne, PhD is a Postdoctoral Research Associate with the Psychology Department’s REACH Institute at Arizona State University. Broadly speaking, her research focuses on the ethnic-racial socialization of Black and Latinx children from low socioeconomic backgrounds as they navigate the sociopolitical demands placed upon them.

Katharine Suma is a doctoral student and research specialist at the University of Georgia. Her research interests focus on socioculturally sensitive observation of parenting behavior and exploring how demographic information informs implications. She has also worked on the development of observational rating measures for dyadic interactions.

Todd D. Little, PhD is a professor in the research, evaluation, measurement and statistics program in the College of Education at Texas Tech University. He is a fellow in AAAS, APS, and APA. He founded, directs, and teaches in the internationally recognized Stats Camps (statscamp.org), operated by Stats Camp Foundation a 501c3 not-for-profit dedicated best-practice training seminars in advanced statistical methods.