ANNOTATED BIBLIOGRAPHY

Bidirectional Influence Between African American Mothers’ and Children’s Racial Centrality From Elementary Through High School

Adam J. Hoffman and Beth Kurtz-Costes University of North Carolina at Chapel Hill

Stephanie J. Rowley University of Michigan

Elizabeth A. Adams University of North Carolina at Chapel Hill and Elon University

For many African Americans, racial identity is an important aspect of their individual identity. We explored developmental change and stability of individual differences in the racial centrality of African American youths and their mothers as well as the relation between maternal and child racial centrality across time. African American youths (N � 380) and mothers completed surveys when youths were in grades 5, 7, 10, and 12. Mean levels of youths’ racial centrality did not increase or decrease across the 7 years of the study. The stability of individual differences in youths’ racial centrality increased across time, reaching adult levels by high school. Cross-lagged panel analyses showed relations between mothers’ and children’s racial centrality in middle adolescence but not in early and late adolescence. Results document that the importance of race to the personal identities of African American youths does not show normative developmental change across adolescence, and only minimal change is linked to maternal influence.

Keywords: racial centrality, racial identity, development, African American

The exploration of the significance of race within their individ- ual identities is an important task of adolescence for African American youths (Sellers, Copeland-Linder, Martin, & Lewis, 2006). Whereas some normative aspects of cognitive development guide these processes, the formation of racial identity is also influenced by the social context (Rowley, Burchinal, Roberts, & Zeisel, 2008; Umaña-Taylor & Guimond, 2012; Umaña-Taylor, Quintana, et al., 2014; Yip, Douglass, & Sellers, 2014). Youths’

everyday social contexts are shaped by their parents, and among African American youths, the context for youths’ racial identity development includes parents’ own racial identity (Rowley, Var- ner, Ross, Williams, & Banerjee, 2012). In the current study, we examine the extent to which mothers’ reports that race is important to their self-concepts (i.e., racial centrality) influences the devel- opment of racial centrality in their children, as well as the possi- bility that children’s identities influence the racial centrality of their mothers. This longitudinal study includes data from both youths and their mothers as youths age from late childhood (Grade 5) through late adolescence (Grade 12), allowing us to examine both levels of racial centrality change across time, and bidirec- tional relations between mothers and children. Before discussing possible change in youths’ racial centrality during adolescence, we provide a brief background about racial identity in African Amer- ican youths and its importance for positive developmental out- comes.

Importance of Racial Identity for Healthy Psychological Functioning

Given growing evidence demonstrating the importance of racial identity in African Americans, Sellers and colleagues (Sellers, Smith, Shelton, Rowley, & Chavous, 1998) developed the Multi- dimensional Model of Racial Identity (MMRI), including four identity dimensions: (a) racial salience, (b) racial centrality, (c) racial regard (public and private), and (4) racial ideology. Of particular interest in the current study is centrality, the degree to which individuals define themselves with regard to their racial group membership (Sellers et al., 1998). We chose to focus on racial centrality rather than other aspects of ethnic/racial identity

This article was published Online First April 6, 2017. Adam J. Hoffman, and Beth Kurtz-Costes, Department of Psychology

and Neuroscience, University of North Carolina at Chapel Hill; Stephanie J. Rowley, Department of Psychology, University of Michigan; Elizabeth A. Adams, Department of Psychology and Neuroscience, University of North Carolina at Chapel Hill, and Department of Psychology, Elon University.

A version of this paper was submitted to fulfill partial requirements for the M.A. degree awarded to Adam J. Hoffman working with Beth Kurtz- Costes as mentor at the University of North Carolina at Chapel Hill. The authors thank Dan Bauer and Patrick Curran for helpful advice regarding the statistical analyses. Great appreciation is expressed to the youth and parents who participated in the Youth Identity Project. Data collection was supported by the National Science Foundation under Grants Nos. 0819079, 0820309, 1251862, and 1251736 awarded to Beth Kurtz-Costes and Steph- anie J. Rowley. Any opinions, findings, and conclusions or recommenda- tions expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Correspondence concerning this article should be addressed to Adam J. Hoffman, Department of Psychology and Neuroscience, The University of North Carolina at Chapel Hill, CB 3270, Chapel Hill, NC, USA, 27599- 3270. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Developmental Psychology © 2017 American Psychological Association 2017, Vol. 53, No. 6, 1130 –1141 0012-1649/17/$12.00 http://dx.doi.org/10.1037/dev0000307

1130

because it is the most basic aspect of racial identity, designating the importance of race to the individual without regard to content. Centrality is also a logical choice because of its relation with psychological well-being.

Studies that have included racial centrality or similar constructs demonstrate positive relations to psychological functioning and educational achievement (Byrd & Chavous, 2009; Neblett, Rivas- Drake, & Umaña-Taylor, 2012; Rowley, Sellers, Chavous, & Smith, 1998). For example, numerous studies show that racial centrality or feelings of connection to the African American com- munity are linked to positive outcomes such as school efficacy, academic success, and school connectedness (Altschul, Oyserman, & Bybee, 2006; Byrd & Chavous, 2009; Chavous, Rivas-Drake, Smalls, Griffin, & Cogburn, 2008; Ellis, Rowley, Nellum, & Smith, 2015; Neblett, Smalls, Ford, Nguyen, & Sellers, 2009; Sellers et al., 2006). Having a strong, positive connection to one’s racial group (e.g., feeling close to friends because of race; belief that members of one’s race have a rich shared heritage) also has indirect effects on well-being by acting as a psychological buffer against the negative impacts of racial discrimination (Wong, Eccles, & Sameroff, 2003). Although a rich body of research has documented the importance of racial centrality for the healthy development of African American youths, relatively little research has addressed relations between this aspect of youths’ identity and that of their mothers, as well as possible developmental changes in centrality. Those questions are addressed in the current study.

Normative Developmental Change: Does Racial Centrality Increase With Age?

Ruble and colleagues (2004) speculate that achieving racial constancy allows children as young as in middle childhood to begin to define themselves in terms of race. However, correctly recognizing the ethnic/racial group membership of self and others, and awareness that race does not change does not mean that ethnic/racial identity is a central part of one’s identity. Each of us has multiple social identities that might include nationality, region of country, religion, or favorite athletic team, and individuals vary in the importance of any one of these social identities to personal identity. As youths pass through late childhood, both normative cognitive development and social experiences are likely to pro- mote increased awareness of social identities such as race and gender, and richer knowledge regarding the meaning of member- ship in a specific social group (Ruble et al., 2004; Umaña-Taylor et al., 2014).

Theorists have suggested that racial identity is dynamic and continuously changing, especially during adolescence (e.g., Cross, 1991; Phinney, 1990; Quintana, 2007; Umaña-Taylor & Fine, 2004). In classic identity theory, adolescence is viewed as a time of critical social– cognitive development, when youths develop a more nuanced understanding of the self and others on the basis of social groups such as racial or ethnic groups (Erikson, 1968; Hart & Damon, 1985). Building on these ideas as well as Selman’s (1980) theory of social perspective-taking ability, Quintana (1994) developed a model of ethnic perspective-taking ability. According to this model, as youths’ racial and ethnic cognition develop from childhood to adolescence, youths move from a sense of self- identity to an achieved sense of ethnic identity, aware of them- selves as members of an ethnic/racial group.

In their theoretical synthesis of the development of ethnic/racial identity (ERI), Umaña-Taylor and colleagues (2014) elucidated the cognitive and social forces that foster ERI development during adolescence. Given both the developmental changes in social cognition as well as the wealth of cultural resources that celebrate African American heritage in the United States, it is reasonable to expect that among African American youths, race would increase in importance as an aspect of personal identity from late childhood to late adolescence. Thus, in the current study we anticipated that for our sample as a whole, the average reported level of racial centrality would increase across the seven years of the study.

The few longitudinal studies that have examined change in constructs related to ethnic/racial centrality support that hypothe- sis, although some of this research was conducted with other ethnic/racial groups. For example, the cultural identity of Ameri- can Indian youths (e.g., use of tribal language, current and in- tended future participation in tribal customs) increased across a 3-year period in late adolescence (Whitesell, Mitchell, & Spicer, 2009). Similarly, Altschul et al., (2006) found small but significant increases in African American youths’ ethnic/racial identity con- nectedness scores (e.g., “I feel part of the Black community”), measured four times between the fall of eighth grade and the spring of ninth grade. Finally, in a longitudinal study of African American boys, racial centrality increased over a 2-year span (Rogers, Scott, & Way, 2015). These findings all support the hypothesis that increases in the importance of race to personal identity would be expected during adolescence.

Stability of Individual Differences: Once High in Racial Centrality, Always High?

Normative developmental change should be distinguished from the stability of individual differences. Whereas normative devel- opmental change indicates change in the average level of racial centrality for the entire sample or population across time, the stability of individual differences assesses whether individual youths change across time in their deviation from the sample average. Thus, it is possible to have normative developmental change but no stability in individual differences or vice versa.

The degree to which race is a central part of a youths’ self- concept is likely to be shaped by factors such as the importance of race to the child’s parents (and therefore, amount of exposure to cultural traditions, artwork, literature, etc.; Hughes et al., 2006) as well as other contextual factors such as racial composition of the child’s school, neighborhood, and other everyday contexts (Umaña-Taylor & Fine, 2004). Thus, youths who are in stable home and school environments would be expected to show relative stability in individual differences across time. Alternatively, the social, cognitive, and contextual changes that occur during late childhood and adolescence might lead to differential change, low- ering the stability of individual differences. For example, a high school class on African American literature might awaken strong racial pride and awareness of race as an aspect of an adolescent’s identity, leading to an increase in racial centrality. Another youth, in contrast, might form new friendships with members of other ethnic/racial groups that lead to a decrease in racial centrality.

Although those examples explain possible change, because of the stability in most individuals’ environments and social circles, we hypothesized that individual differences in racial centrality

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1131RACIAL CENTRALITY IN AFRICAN AMERICANS

would be stable across all waves of the study for both youths and their mothers, with those who were relatively low in racial cen- trality remaining low compared to peers, and those for whom race was more central to their identity continuing to report high racial centrality compared to peers. However, because of the many cognitive and social changes that occur in early adolescence, we expected greater stability of individual differences in youths’ reports between Grades 10 and 12 than during earlier years of the study. It is likely that by late adolescence youths have a more sophisticated understanding of race than in early adolescence, contributing to greater stability in individual differences. Indeed, Pahl and Way (2006) found that Black and Latino youths spent less time exploring their ethnic/racial identity as they moved into late adolescence.

Parents and the Development of Youths’ Racial Centrality

Umaña-Taylor and Fine (2004) describe ecological factors (e.g., school and community characteristics) and socializing agents (e.g., parents, peers, and teachers) as non-normative factors that influ- ence ethnic/racial development in adolescents. Arguably, parents may be one of the most important and dominant socializing agents in youths’ racial identity development from childhood to late adolescence. Youths who receive more racial socialization from their parents tend to report higher racial centrality than youths who receive less racial socialization (O’Connor, Brooks-Gunn, & Gra- ber, 2000; Umaña-Taylor & Fine, 2004). In a review of the parent ethnic/racial socialization literature, a rich profusion of studies demonstrated that parents are socializing agents in their children’s ethnic/racial identity formation from early childhood to adulthood (Hughes et al., 2006).

Two competing theories have emerged from the literature ex- amining parents’ and children’s values and beliefs and the extent to which these either converge or diverge as the child ages to adulthood. According to Erikson’s (1993) developmental perspec- tive, adolescent youths are tasked with the job of establishing an individual sense of self. Erikson posited that as part of this process, youths may seek to differentiate themselves from their parents and their parents’ values and beliefs, and that this process sometimes is resolved with congruence between parents’ and youths’ values and beliefs, and sometimes with differentiation (Erikson, 1993, p. 269). In contrast, socialization theory posits that the relation between the values and beliefs of parents and those of their children diminishes over time, with greatest similarity occurring at early developmen- tal periods when parents have the most interaction with their children (Miller & Glass, 1989).

Research in values, attitudes, and beliefs that are unrelated to race has shown that parents and youths are likely to share similar values in adolescence and emerging adulthood (Flor & Knapp, 2001; Jennings, Stoker, & Bowers, 2009; Jodl, Michael, Malanchuk, Eccles, & Sameroff, 2001; Leonard, Cook, Boyatizis, Kimball, & Flana- gan, 2013). For example, parents and adolescents report sharing similar religious beliefs and behaviors (Flor & Knapp, 2001; Leonard et al., 2013). Similarly, youths are likely to adopt their parents’ political orientations, particularly if the family is highly politicized (Jennings et al., 2009). Finally, parents’ and adoles- cents’ reports of academic values and beliefs tend to be similar, especially if the adolescent has a positive identification with the

parent (Jodl et al., 2001). These findings provide indirect support for the hypothesis that parents and youths have similar racial centrality, especially in adolescence.

Transactional Influences Between Parents’ and Youth’s Racial Centrality

Whereas classic socialization theory emphasized the importance of parents as socializing agents of children, implying a unidirec- tional model of causality, developmental theorists now recognize the active role of children in these processes, positing bidirectional or transactional models of causality (Sameroff, 2009). In bidirec- tional models, the parent and child are viewed as dynamic agents who act on each other in a process of reciprocal influence. Bidi- rectional and transactional models not only predict concurrent similarity in parent and child values, but also posit that prior beliefs, values, and behaviors of one family member will influence the other’s future beliefs, values, and behaviors. Testing a trans- actional model, we examine whether parents and youths influence changes in each other’s racial centrality across time.

We propose that directions of influence between parents and youths would not necessarily be expected to be bidirectional throughout the child’s development from late childhood to late adolescence. Because children’s social cognition related to race and ethnicity is still developing in late childhood (Rowley et al., 2008), it is probable that parents’ and children’s racial centrality are not yet correlated or are only weakly related in late childhood. In middle- and late adolescence, bidirectional paths of influence are more likely because of youths’ more sophisticated understand- ing of race-related events in their lives (Umaña-Taylor et al., 2014). Youth may be influenced by their parents’ racial centrality during this period because of more adult-like relationships and discussions with parents about race-related experiences. Moreover, conversations about youths’ out-of-home race-related experiences might lead parents to reassess their own racial centrality as a function of the racial centrality of their adolescent children.

The Present Study

The present study is an examination of changes in racial cen- trality in a sample of African American mothers and their children as youths aged from fifth to twelfth grade. We aimed to (a) test whether racial centrality increases, on average, for the sample as a whole during adolescence; (b) examine the stability of individual differences in mothers’ and youths’ racial centrality as youths age from late childhood (Grade 5) to late adolescence (Grade 12); (c) explore if/when a concurrent relation between mothers’ and youths’ racial centrality emerges; and (d) assess if mothers and youths have a bidirectional influence on the other’s racial central- ity change across time.

Because of increased social cognition and opportunities to wit- ness the importance of race, we expected that youths’ racial centrality would increase across the seven years of the study (Hypothesis 1). We also hypothesized (Hypothesis 2) that individ- ual differences in youths’ racial centrality scores would be more stable in later years of the study than in early adolescence, but that mothers’ racial centrality would be stable across the 7 years of the study. The third hypothesis addressed concurrent relations between mothers’ and youths’ racial centrality. We expected that such

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1132 HOFFMAN, KURTZ-COSTES, ROWLEY, AND ADAMS

relations would appear only in late adolescence (Grades 10 and 12) because of early instability in youths’ racial centrality. Finally, across the 7-year span from Grade 5 to Grade 12, we hypothesized (Hypothesis 4) that a bidirectional transactional pattern of relations between mothers’ and youths’ racial centrality would emerge. We expected significant maternal influence on the child’s racial centrality change at each measurement point (i.e., Grade 5 to 7, Grade 7 to 10, Grade 10 to 12). We expected that youths would significantly influence change in their mothers’ racial centrality only in middle and late adolescence (i.e., Grade 7 to 10 and Grade 10 to 12).

Method

Participants

Data for the project were drawn from the Youth Identity Project, a longitudinal study that focused on academic achievement and racial identity development in African American youths. A sample of 380 African American youths (216 girls, 164 boys) who had participated in at least one of the four waves provided data for the current study. Youth were recruited in Grade 5 (M age � 11.14 years old) and were asked to participate again in Grades 7, 10 and 12.

Data were collected in an urban school district in the south- eastern region of the United States. Youth were originally recruited as fifth graders from seven elementary schools in which 61% to 97% of students were African American. Youth were recruited in three cohorts during the 2002–2003, 2003– 2004, and 2004 –2005 school years. Participants later attended 17 middle schools (ranging from 27% to 98% African Ameri- can; median � 68.8%) and 17 high schools (ranging from 37% to 93% African American when youths were in Grade 10, and 29% to 90% African American when youths were in Grade 12; medians � 77.8% and 76.9%, respectively). The percentages of students within participating schools who received free or re- duced price lunch ranged from 40% to 95% in the elementary schools, 21% to 91% in the middle schools, and 10% to 82% in the high schools. Retention rates of child participants across the four time points were 80%, 84%, and 82%, respectively. Only one child participated from each family.

Data are reported for 246 female caregivers, 89.8% of whom were mothers, 6.1% were grandmothers, and 4.1% were other female caregivers (e.g., aunt, step-mother). Grandmothers and other female caregivers were included in the sample if they were the target child’s primary parent. Although African Amer- ican fathers are also important in the socialization of their children (e.g., Cooper, Smalls-Glover, Metzger, & Griffin, 2015), we chose to focus on mothers and other female caregiv- ers. The importance of mothers in the socialization of children has been well documented, particularly in African Americans families (McHale et al., 2006). Although parent gender was not specified in our invitation letter for study participation, a ma- jority of parents who participated were mothers or other female caregivers. If the parent reporter changed across the four waves (e.g., mother participated in Grades 5 and 10, and grandmother participated in Grade 7), data were used from the female care- giver (henceforth referred to as “mothers”) who participated most often, and data from the other reporter(s) were excluded in

order to reliably assess the stability of parents’ racial centrality. Only mothers who identified as African American were in- cluded in the sample.

The median mother-reported education level was “some tech- nical school,” and 34% of mothers had attended college or were college graduates with a bachelor’s degree or higher. According to mothers’ reports when youths were in Grade 5, 61.3% of the families had an annual income of $29,999 or less; 24.5% were between $30,000 and $59,999 per year; and 14.2% had annual incomes of $60,000 or more (median � $20,000 to $29,999). According to United States Census data, the upper limit of income of the second fifth of African American households (i.e., the 40th percentile) in 2002 was $16,340, and the upper limit of the third fifth (i.e., at the 60th percentile) for African American households was $28,959 (U.S. Census Bureau, 2013). Thus, household income of this sample was representative of African American households in the United States (U.S.) at the time of the study. Mothers’ reports of their marital status revealed that 42.5% of the sample were single and never mar- ried, 28.7% were married and living with their partners, 10.5% were married but separated from partners, 15.3% were di- vorced, and 2.9% were widowed. Retention rates of mothers across the four time points were 92%, 84%, and 69%, respec- tively.

Procedure

In initial sample recruitment, letters explaining the goals of the study were distributed to all Grade 5 children at each participating school. The letter provided the rationale and lo- gistical information about the study and invited the child and one parent from the household to participate. Parent permission and child assent were obtained for each wave of data collection. All youth questionnaires were completed at school in small groups at a time and location (e.g., library or resource room) that was deemed suitable by school staff. For younger partici- pants (Grades 5 and 7), the surveys were read aloud by a trained research assistant. Surveys were completed in one session last- ing approximately 30 min. Upon completion of the session, the research assistant thanked the youths and distributed incentives (small toys such as a yoyo or flashlight in Grade 5; a $5 gift card in Grade 7; $10 gift cards in Grades 10 and 12).

Packets including the parent questionnaire and a stamped, return-addressed envelope were mailed to homes and were completed by participating parents. Once the laboratory re- ceived a completed questionnaire, an incentive (a gift card to a local grocery store) and thank-you note were mailed to the parent.

These data collection methods contributed to discrepancies between the numbers of students and mothers who participated. Youth completed surveys at school during school hours, whereas mothers completed surveys at a time and place of their choosing. Among the youths sample, 159 youths (42%) com- pleted all four waves, 103 (27.2%) completed three waves, 64 youths (16.9%) completed two waves, and 52 (13.7%) com- pleted a single wave of data collection. Eighty-four mothers (22.2%) provided data for all four study waves, 74 (19.5%) participated three times, 79 (20.8%) completed two waves, 78

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1133RACIAL CENTRALITY IN AFRICAN AMERICANS

(20.6%) participated once, and parents of 64 youths (16.9%) did not participate in any wave.1

Measures

Data for the current study include youth reports of racial cen- trality, mothers’ reports of racial centrality, and family demo- graphic information. Other constructs were assessed in the child and parent questionnaires that are not included in this report.

Racial centrality. Racial centrality was measured in both youths and mothers using six items from the racial centrality scale of the Multidimensional Inventory of Black Identity (Sellers et al., 1998). These items assessed the extent to which being African American is central to the respondent’s self-concept. The items were measured on a 5-point scale (1 � strongly disagree; 5 � strongly agree) where youths and parents rated how much they agreed with each item (e.g., “Being Black is an important part of my self-image”). The wording of some items was edited slightly so that the measure was appropriate for fifth graders. Items were averaged to yield a single score, with higher scores indicating greater racial centrality. The alpha reliabilities for youths at Grades 5, 7, 10, and 12 were .70, .73, .80, and .81, respectively. Alpha reliabilities for racial centrality of mothers were .76, .72, .77, and .68, respectively. These indices, considered acceptable to good, are consistent with prior research measuring racial centrality in Afri- can American adolescent and adult samples (Neblett, Banks, Coo- per, & Smalls-Glover, 2013; Rogers, Scott, & Way, 2015; Seaton, Yip, & Sellers, 2009).

Household income and mothers’ education. Mothers re- ported household income on either an annual or weekly scale. The annual scale used increments of $10,000 ranging from “less than $10,000 per year” to “more than $100,000 per year,” and the weekly scale was approximately parallel to the annual scale (“less than $200 per week” to “more than $2,000 per week”). Mothers reported their educational attainment on a 10-point scale ranging from “less than high school” to “doctoral or professional degree.” For this paper, we use household income and mothers’ education from the first wave (i.e., when youths were in Grade 5).

Mothers’ relationship to child and partnered status. Mothers indicated their relationship to the participating child by selecting the appropriate label from a list (mother, father, grand- parent, etc.). Mothers indicated their partner status by choosing from a checklist that included five choices: married and living together; married but separated; divorced; widowed; single and never married.

Results

Plan of Analyses

To better understand mothers’ and children’s racial centrality, means, standard deviations, and bivariate correlations between pairs of scores for each wave were computed (see Table 1). We used latent growth curve modeling to test normative patterns of change in youths’ racial centrality (Hypothesis 1). Repeated mea- sures autoregressive cross-lagged (ARCL) panel analysis was em- ployed to test the stability of individual differences in racial centrality (Hypothesis 2). We used both the ARCL model results and analysis of bivariate correlations to test the relations between

and bidirectional influences of children and parents on each other’s racial centrality (Hypotheses 3 and 4). All analyses were computed with Mplus Version 7 with maximum likelihood estimation to account for missing data (Muthén & Muthén, 1998 –2014).

Attrition analyses spanning the entirety of the study were con- ducted to determine if data were missing at random. Results showed that Time 1 reports of children’s and mothers’ racial centrality did not differ when comparing youths and mothers who did or did not provide data at Time 4 (children’s Ms � 3.65 and 3.61, respectively, t (370) � �.43, p � .67; mothers’ Ms � 3.42 and 3.40, respectively, t(244) � �.12, p � .91). Thus, from these results, we concluded that the data were missing at random. Data from all 380 youths and 246 mothers were included in all analyses, including data of youths whose mothers did not complete any waves of data collection. We included these youths in the sample because there was evidence that data were missing at random and because of the use of maximum likelihood estimation for missing data. Retention of these data added information to the analyses and thereby strengthened the reliability of the results (P. Curran, per- sonal communication, May 31, 2016). Cohort effects and relation to child effects (i.e., mother, grandmother, aunt) were assessed for each model and were all nonsignificant.

Preliminary Descriptive Analyses

As shown in Table 1, both parents’ and youths’ reports of racial centrality were slightly above the midpoint on the 5-point scale, indicating moderate rates of racial centrality, on average. Racial centrality scores ranged from 1.00 to 5.00 for Grade 5 youths and their mothers, reflecting considerable variability in the importance of race to these individuals. Bivariate correlations (many of which are discussed in greater detail below) of racial centrality scores across time were significant and tended to be of low magnitude for youths (ranging from .18 to .57) and of moderate magnitude among mothers (ranging from .46 to .63). Maternal education and household income were unrelated to youth reports of racial cen- trality, and were weakly related to mothers’ reports (ranging from .11 to .25).

Hypothesis 1: Increase Across Time in Average Rates of Racial Centrality

Latent growth curve modeling was used to examine the possi- bility of change in youths’ racial centrality over time. First, a conditional linear growth curve model was estimated for youths’ racial centrality across the four time points. Gender, household income, and mothers’ education were included in the model as control variables. The model was an adequate fit for the data, �2

(11) � 20.96, p � .04 CFI � .89; TLI � .83, RMSEA � .06; SRMR � .08. To ensure that a linear function was optimal in characterizing the data, a quadratic function was estimated for the

1 These numbers indicate the amount of data included in this report. Actual parent participation rates were higher. Parent data from 15 families were not used because only the father participated in that family. An additional 44 reports were not used in order to limit the data to only one parent responder from each family (e.g., if the mother participated at Times 1 and 2 and a grandmother at Time 4, the grandmother’s data were not used).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1134 HOFFMAN, KURTZ-COSTES, ROWLEY, AND ADAMS

growth curve model. However, model fit indices suggested that this model was overfitting the data or was unnecessarily complex (Preacher, 2006). Results from the linear model revealed a signif- icant intercept factor of 3.77. The slope factor of �0.03 was not significant, suggesting no change across time. Intercept and slope factors did not differ by gender, household income, or mothers’ education. Thus, Hypothesis 1 was not supported: We found no evidence of normative change in average levels of youths’ racial centrality from Grade 5 to Grade 12.

Hypothesis 2: Stability in Individual Differences Across Time

A repeated measures autoregressive cross-lagged (ARCL) panel analysis (Dwyer, 1983) was used to examine stability of individual differences in racial centrality, the relation between youths’ and mothers’ reports at each wave, and the influence of mothers’ and youths’ racial centrality in predicting each other’s subsequent racial centrality, while controlling for household income and par- ent education. ARCL panel analysis is a transactional model of causality used to examine paths of possible influence over time (Kuczynski, 2003).

A baseline model predicted children’s and mothers’ racial cen- trality at Time 1, estimated with household income and mother education as control variables. We initially included gender in the model as a control variable. However, with gender controlled, the model did not converge, and therefore gender was not retained in the model. Time-adjusted stability paths for racial centrality were added for both child and parent (i.e., Time 1 predicted Time 2; Time 2 predicted Time 3; Time 3 predicted Time 4), and cross- lagged paths of child to mother and mother to child racial central- ity scores were added across all waves (i.e., Time 1 Mother predicted Time 2 Child; Time 1 Child predicted Time 2 Mother; Time 2 Mother predicted Time 3 Child; etc.). Finally, because it was expected that children’s and mothers’ racial centrality would be correlated at each time point, within-time residual correlations were added to the model (e.g., correlating Time 1 Child and Time 1 Mother racial centrality scores, Time 2 Child and Time 2 Mother, etc.). Not surprisingly, the fit of the initial model was poor, �2(24, N � 379) � 71.91, p � .001.

Racial identity theory (Sellers et al., 1998) suggests that racial centrality is a relatively stable trait in adults. Thus, two direct prospective paths were added to the model: Time 1 Mother’s

Racial Centrality predicting Time 3 Mother’s Racial Centrality, and Time 2 Mother’s Racial Centrality predicting Time 4 Mother’s Racial Centrality. These additional paths resulted in a significant improvement in the model as confirmed by a Likelihood Ratio Test, �2 (22, N � 379) � 35.24, p � .05. Finally, to test for possible model misspecifications with the control variables, La- Grange multipliers were estimated, and based on those results a path from Household Income to Time 2 Mother’s Racial Central- ity was added. A Likelihood Ratio Test indicated that the addition of this path resulted in a significant improvement in model fit. The final model fit the data well, �2 (21, N � 379) � 31.48, p � .07; CFI � .97; TLI � .94; RMSEA � .04; SRMR � .04 (see Figure 1).

According to Hypothesis 2, individual differences in mothers’ racial centrality scores would be stable across time, whereas indi- vidual differences in youths’ racial centrality would be more stable in later years of the study than in earlier years. As shown in Figure 1, model results supported this hypothesis: Individual differences in both mothers’ and youths’ racial centrality were stable across the four time points. Moreover, Fisher’s r to z transformation for dependent correlations revealed that youths’ racial centrality scores showed higher stability from Grade 10 to Grade 12 than from Grade 5 to Grade 7 (z � 2.74, p � .003) and from Grade 7 to Grade 10 (z � 2.07, p � .019). Bivariate correlations between raw scores also reflected this pattern: The Grade 10 to 12 corre- lation of r (202) � .57 for youths was significantly greater than the correlations between youths’ Grade 5 to 7 scores (z � 3.32, p � .0005) and the correlation between Grade 7 and 10 scores (z � 2.74, p � .003) (see Table 1 for all bivariate correlations). As expected, parent racial centrality scores showed greater stability than youths’ racial centrality from Grade 5 to 7 (z � 3.48, p � .001) and Grade 7 to 10 (z � 3.75, p � .001). Individual differ- ences in youths’ racial centrality reached adult levels of stability between Grades 10 and 12 (z � 0.83, p � .406).

Whereas our primary study hypotheses addressed stability of individual differences for the sample as a whole, it is possible that stability varies as a function of the importance of race to individ- uals. For example, individuals who are exceptionally high in racial centrality might have stronger beliefs, and therefore show greater stability across time than individuals for whom race is less impor- tant. Therefore, to offer a closer examination of individual differ-

Table 1 Means, Standard Deviations, Sample Size and Bivariate Correlations for Child and Mother Racial Centrality and Control Variables

M (SD) n 1 2 3 4 5 6 7 8 9 10

1. Child racial centrality T1 3.63 (.79) 372 — 2. Child racial centrality T2 3.57 (.74) 299 .31�� — 3. Child racial centrality T3 3.52 (.78) 250 .27�� .35�� — 4. Child racial centrality T4 3.59 (.73) 204 .27�� .18� .57�� — 5. Mother racial centrality T1 3.41 (.74) 246 .08 .03 .10 .11 — 6. Mother racial centrality T2 3.54 (.67) 227 .02 .11 .23�� .13 .49�� — 7. Mother racial centrality T3 3.61 (.76) 190 .01 .05 .26�� .28�� .58�� .55�� — 8. Mother racial centrality T4 3.55 (.63) 131 .20� .10 .25�� .29�� .46�� .51�� .65�� — 9. Household income 3.65 (2.52) 270 �.11 �.03 .01 .01 .23�� .25�� .25�� .06 —

10. Mother education 5.06 (2.13) 275 �.07 �.11 �.05 �.07 .19�� .14 .16� .21� .51�� —

� p � .05. �� p � .01.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1135RACIAL CENTRALITY IN AFRICAN AMERICANS

ences in the stability of racial centrality, sensitivity analyses were conducted based on participants’ initial levels of racial centrality.

Each sample (i.e., youths and mothers) was divided at the 33rd and 66th percentiles using Grade 5 racial centrality scores to produce three groups: those who were low, average, and high in racial centrality compared to peers. Multiple groups path analyses were conducted separately for youths and mothers to allow time- adjusted stability coefficients of racial centrality to be predicted. Household income and mother’s education were included as con- trol variables. Model fit for youths was adequate, �2(27) � 33.89, p � .17; CFI � .91; TLI � .86; RMSEA � .05; SRMR � .09. Model fit for mothers was poor, �2(27) � 48.50, p � .008; CFI � .75; TLI � .62; RMSEA � .10; SRMR � .14. The poor fitting model among mothers is likely the result of low statistical power, and thus results for mothers are not presented. Means, standard deviations, and time adjusted stability coefficients of the path analysis for the youth sample appear in Table 2. Among youths with average or high levels of racial centrality, results revealed

significant stability paths between each pair of time points from Grade 5 to Grade 12. However, among youths with low levels of racial centrality, the stability path from Grade 5 to Grade 7 was not significant. Stability paths from Grade 7 to Grade 10, and from Grade 10 to Grade 12 were significant.

In summary, individual differences in mothers’ racial centrality were stable across the seven years of the study, and exceeded the stability of youths’ scores before Grade 10. Individual differences in youths’ racial centrality were also stable across time, and were greater during high school than earlier in development. Sensitivity analyses revealed that youths with lower levels of racial centrality in Grade 5 did not achieve stability until Grade 10, whereas youths with average or high levels of racial centrality showed stability in individual differences between all pairs of time points.

Hypotheses 3 and 4: Concurrent Similarity and Transactional Influence in Mothers’ and Children’s Racial Centrality

Hypothesis 3 predicted that mothers’ and children’s racial cen- trality would be concurrently correlated in Grades 10 and 12 but not in Grades 5 and 7. As shown in Table 1, bivariate correlations supported this hypothesis. Youth’s racial centrality was correlated with their mothers’ racial centrality during the high school years (r’s � .26 and .29, p’s � .05) but not earlier, r’s � .08 and .11, p’s � .10. We also tested this hypothesis using the ARCL results. Specifically, the within time residual correlations from the model were observed to examine the relation between mothers’ and children’s racial centrality scores at each wave. As anticipated, child and mother racial centrality scores were unrelated in Grade 5, r � .10, p � .10 and Grade 7, r � .12, p � .09. When youths were in Grade 10, there was a small but significant positive relation

Figure 1. Autoregressive cross-lagged model of youths’ and mothers’ racial centrality from Grade 5 to Grade 12. Regression weights for unidirectional pathways are unstandardized. Bidirectional pathways are standardized and can be interpreted as correlations. Solid lines represent paths that were significant (p � .05) and dashed lines represent paths that were not significant (p � .05).

Table 2 Means, Standard Deviations, Sample Sizes and Stability Coefficients for Youths Who Were Low, Average, and High in Racial Centrality at Grade 5

Low racial centrality

Average racial centrality

High racial centrality

M (SD) 2.76 (.54) 3.72 (.21) 4.47 (.29) n 126 129 117 Grade 5 to Grade 7 .13 .72� .70�

Grade 7 to Grade 10 .31� .42�� .32�

Grade 10 to Grade 12 .51�� .45�� .76��

� p � .05. �� p � .01.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1136 HOFFMAN, KURTZ-COSTES, ROWLEY, AND ADAMS

between mothers’ and children’s racial centrality scores (r � .19, p � .05, see Figure 1). The residual correlation was nonsignificant in Grade 12, r � .12, p � .26.

The discrepancy between results with bivariate correlations and the ARCL analyses is probably due to the more conservative nature of the ARCL analyses, in which each path shows the residual shared variance between pairs of variables above and beyond shared relations from earlier time points (Ahmed, Minn- aert, Kuyper, & van der Werf, 2012; Arnett et al., 2012). Thus, in our model, the within time residual correlation between Time 4 child racial centrality and Time 4 mother racial centrality shows the strength of that relation above and beyond the association at Time 3. With the Time 3 relation controlled, children’s and moth- ers’ racial centrality were unrelated in Grade 12.

Finally, in Hypotheses 4 we posited that mothers and youths would influence each other’s racial centrality across time. We anticipated that mothers would have significant influence on youths’ racial centrality development across all lags (i.e., from Grade 5 to Grade 7, Grade 7 to Grade 10, and Grade 10 to Grade 12) and that youths would have significant influence on mothers’ racial centrality scores in later adolescence—from Grade 10 to Grade 12. The bivariate correlations provided support for Hypoth- esis 4: Youth’s racial centrality was unrelated to mothers’ subse- quent racial centrality from Grade 5 to 7 and Grade 7 to 10, but the correlation was significant from Grade 10 to 12, r(129) � .25. Whereas mothers’ Grade 5 racial centrality was unrelated to their children’s Grade 7 reports, mothers’ racial centrality was corre- lated with youths’ later racial centrality across the last two time lags, r’s � .23 and .28, p’s � .01. In the ARCL results, only one of the cross-lagged paths was significant: Mothers’ racial central- ity when their children were in Grade 7 predicted youths’ Grade 10 racial centrality scores (see Figure 1).

Thus, both bivariate correlations and the cross-lagged panel analyses showed that, as hypothesized, relations between mothers’ and youths’ racial centrality increased from early to middle ado- lescence. However, results of the cross-lagged panel analysis in- dicated that youths did not significantly influence mothers’ later racial centrality across any of the time lags, and mothers influ- enced youths’ later centrality only in the lag from Grade 7 to 10, indicating a decreasing influence of mothers on youths’ racial centrality during the high school years.

Discussion

This study provided one of the first longitudinal examina- tions of racial centrality development in African American youths and their mothers from late childhood to late adoles- cence. Contrary to a study hypothesis, average levels of youths’ racial centrality did not change over the seven years of the study, indicating no normative change in the importance of race to African American youths’ identity as they passed through adolescence. Individual differences in reports of racial central- ity in youths and their mothers were stable across time, with the stability of youths’ scores increasing in middle adolescence and reaching adult levels in high school. Although bivariate corre- lations showed that mothers’ and youths racial centrality were related concurrently and across time beginning in middle ado- lescence, the only significant cross-lagged path in the ARCL model was from mothers’ racial centrality in early adolescence

(i.e., Grade 7) to youths’ racial centrality in middle adolescence (i.e., Grade 10). The cross-lagged panel analyses also showed a weak but significant concurrent relation between mothers’ and youths’ racial centrality when youths were in Grade 10.

Developmental Change in the Racial Centrality of African American Youth

Because adolescents have a more sophisticated understanding than children of race and of ways in which racial group member- ship affects social experiences, we expected that average levels of youths’ racial centrality would increase during adolescence. Few longitudinal studies have tested this hypothesis in African Amer- icans. Contrary to our expectations, youths’ average centrality levels did not change across the seven years of the study from late childhood to late adolescence.

Data collection for this project occurred during the historic events of Hurricane Katrina, the election of President Barack Obama, and spotlighting in the national media of police shoot- ings of young Black men. Youth in our study were teenagers when Obama was elected President of the U.S., and our last cohort of youths participated as 12th graders the year that Trayvon Martin was killed and the Black Lives Matter move- ment began. These historic events may have influenced what African Americans parents communicate to their children about race (cf. Neblett, Terzian, & Harriott, 2010). Nonetheless, youths’ racial centrality did not increase, on average, across the seven years of the study. We believe that these results indicate that the development of racial centrality on a normative level is relatively impervious to distal race-related events for African American youths. Instead, such events might have a greater impact on other aspects of racial identity such as public and private regard, or racial ideologies. Indeed, in a study with a sample from a historically Black college, 72% of students reported that the Obama Presidency had increased their racial pride (Wade & Younge, 2015). Thus, it appears that racial centrality may be less susceptible to distal historic events than other aspects of racial identity.

In contrast to our results, two recent studies demonstrated small increases in racial centrality or racial connectedness in African American youths across shorter time periods (Altschul et al., 2006; Rogers et al., 2015). As mentioned above, Altschul et al. (2006) found a small increase (two tenths of a point on a five-point scale) in racial connectedness from the beginning of eighth grade to the end of ninth grade. Using a sample of African American boys attending an all-Black, all-male school, Rogers et al. (2015) re- ported increases in racial centrality from the beginning of ninth grade to the end of tenth grade. However, raw means differed by less than two-tenths of a point across the three waves of that study, and racial centrality increased in the model only after other vari- ables were added (i.e., racial centrality increased only among boys who were high in gender centrality). Taken together, our results and those of these prior studies indicate that substantial increases in the importance of race to Black youths’ identity are not norma- tive during adolescence. Nonetheless, additional research is needed to better understand what factors might drive changes in youths’ racial centrality during adolescence.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1137RACIAL CENTRALITY IN AFRICAN AMERICANS

Developmental Change in Individual Differences in Racial Centrality

Racial centrality scores of both youths and their mothers showed considerable variability within each time point, with some partic- ipants reporting strong agreement and others disagreeing or strongly disagreeing with statements such as “Being Black is an important part of who I am.” These individual differences in mothers’ and youths’ racial centrality scores were stable over the seven years of the study, with the stability of youths’ racial centrality increasing during adolescence. The increasing stability in youths’ racial centrality supports developmental theory that posits early and middle adolescence as an important developmen- tal period for racial identity formation (e.g., French, Seidman, Allen, & Aber, 2006; Pahl & Way, 2006). In exploratory analyses, we found that youths with higher racial centrality in late childhood showed stability earlier than youths with lower racial centrality, perhaps indicating that youths who live in contexts where racial identity is especially important begin to explore the implications of their racial identity at an earlier age.

As we anticipated, individual differences in mothers’ reports of racial centrality showed high levels of stability across the seven years of the study. Whereas research regarding racial identity in youths has grown significantly in recent years, less research has examined racial identity in adults. Our results provide empirical support for the contentions of Sellers et al. (1998) that African American adults differ in the importance of race to their personal identities, and that these differences are stable across time.

Developmental Change in the Relation Between Mothers’ and Youth’s Racial Centrality

We expected to find positive concurrent relations between mothers’ and youths’ racial centrality in middle and late adoles- cence (i.e., Grades 10 and 12) but not earlier. As anticipated, concurrent relations between racial centrality of mothers and their children were not significant at the first two time points, but the relation was significant when youths were in Grade 10. Contrary to our expectations, the concurrent relation was not significant in late adolescence (i.e., Grade 12) in the cross-lagged panel analyses.

Because of the importance of parents in shaping characteristics of the home environment, in choosing children’s extracurricular activities, engaging in discourse with their children about identity- related topics, and selecting children’s neighborhoods and schools (e.g., Byrd & Chavous, 2009; French et al., 2006), one might anticipate that mothers’ and youths’ scores would be positively correlated at each measurement point. Early congruence between mothers’ and children’s identity characteristics would also be consistent with the arguments of Cross, Strauss, & Fhagen-Smith (1999), who suggested that the sense of connection to one’s racial group often starts early in life and is highly dependent on parent identity.

However, the concurrent relation between mothers’ and youths’ racial centrality was significant only in middle adolescence, con- sistent with developmental theorists’ suggestions that adolescence is a key period for identity development (Erikson, 1993; Quintana, 2007). Whereas Erikson proposed that parents’ influence is im- portant for youths’ developing values and goals, our results extend this theoretical perspective to the domain of racial identity. As

youths acquire more sophisticated social cognition during adoles- cence and individual differences in their racial centrality become more stable, shared experiences with their parents—including par- ents’ racial socialization and sharing of experiences such as racial discrimination—likely cause youths’ identity to more closely re- semble that of their parents. Although it is true that many parents engage in racial socialization already in early childhood, research has shown increased complexity during adolescence in the content of messages (e.g., discrimination or wariness of other racial groups) (Hughes et al., 2006; McHale et al., 2006). Further, with more advanced social cognition, adolescents are more likely to understand and internalize the messages they receive from their parents (Hughes et al., 2006). Thus, developmental and contextual changes during adolescence are probable factors that lead to in- creased and more complex conversations with parents about race, increasing the congruity between parents’ and children’s racial identity.

However, our analyses showed that mothers’ primary influence is likely to be in middle adolescence, and decreases by late adolescence: The concurrent relation between mothers’ and youths’ racial centrality was not significant when youths were in Grade 12. In addition to parents’ own racial identity and their racial socialization practices, youths’ experiences outside the home are likely candidates that might explain the waning relation be- tween mothers’ and youths’ racial centrality in late adolescence. As youths spend more time with peers in later adolescence and establish educational and career goals that will take them outside of their home of origin, the influence of mothers might fade while friends and romantic partners might gain prominence in shaping aspects of youths’ racial identity. Indeed, it is considered appro- priate for the healthy development of young adults to establish goals, values, and beliefs that are distinct from those of their parents (Koepke, & Denissen, 2012; Kroger, 2007).

Alternatively, it is possible that whereas some youths arrive, after identity exploration, at a point of congruence with parents, others emerge from this period choosing to be different from their parents. If that is the case, high similarity between some mother- youth pairs and high divergence between other pairs would lead to a lack of relation for the sample as a whole. Moderating factors such as quality of the parent– child relationship might influence the extent to which youths’ racial identity resembles that of their parents as they approach adulthood.

Transactional Influence Between Mothers’ and Youth’s Racial Centrality

In the examination of the transactional influence between moth- ers and youths over the course of the seven years, we anticipated that mothers would influence youths across each wave and that youths would begin to influence mothers beginning in early ado- lescence. In contrast to our hypotheses, results from the ARCL model revealed only one significant cross-lagged path: Mothers’ racial centrality in early adolescence (i.e., Grade 7) influenced youths’ racial centrality at middle adolescence (i.e., Grade 10). These results show that mothers are influential in the formation of racial centrality of their children from early to middle adolescence, a developmental period when youths’ racial centrality is becoming increasingly stable, and before peers and anticipation of adoption

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1138 HOFFMAN, KURTZ-COSTES, ROWLEY, AND ADAMS

of adult roles begin to distance many youths from their parents (Kroger, 2007).

Although we expected mothers to influence youths’ racial cen- trality from late childhood to early adolescence, it is important to consider that racial identity is likely in a nascent form for youths in middle to late childhood (Umaña-Taylor et al., 2014). Given children’s less advanced social cognition (as compared to that of adolescent youths) and smaller knowledge base about the impli- cations of their racial identity (e.g., less knowledge about the history and culture of African Americans; less knowledge about racial discrimination and systemic inequities), it is not surprising that children’s reports were unrelated to those of their mothers.

The lack of relation between mothers’ racial centrality at middle adolescence (i.e., Grade 10) and youths’ racial centrality in late adolescence (i.e., Grade 12) was also unexpected. As discussed above, it is likely that during late adolescence other influential social agents as well as youths’ growing independence from par- ents lead to a diminishing influence of parents on youths’ racial centrality. Our results also showed, contrary to expectations, that youths had little influence on changes in their mothers’ racial centrality. Although the bivariate correlation from Grade 10 youth centrality to Grade 12 mother centrality was significant, this path was nonsignificant once other variance was controlled in the cross-lagged panel analyses.

Limitations and Directions for Future Research

Notwithstanding the novelty and contributions that the current study adds to the racial identity development literature, some limitations should be addressed. The results of the current study were based solely on urban African American participants. Partic- ipants from this urban setting lived in a community with a long history of prominent and wealthy African Americans in local businesses and politics. Further, participants of the study came primarily from relatively low-income homes where the median household income was between $20,000 and $29,000. Although this income level was comparable to that of African American households across the U.S. at the time of data collection, our results might not be generalizable to upper-income families. In addition, these data were drawn from the Southeastern region of the U.S., where racial tensions and negative attitudes have histor- ically been stronger than in other regions of the U.S. Less than one third of participating mothers reported that they were living with a partner. Thus, results from the current study may represent a specific sample and might not be generalizable to other samples of African Americans.

Given the different methods of data collection between youths and mothers (i.e., collecting data in person with youths vs. col- lecting data via mailings with mothers), missing data from mothers was consistently greater than youths at each wave. All models were estimated with maximum likelihood to maximize the sample size, allowing for the inclusion of mothers who had missing data. However, a larger sample of mothers and less sample attrition should be goals of future research.

Future research might also examine mediating and moderat- ing mechanisms that shape relation between parents’ and youths’ racial centrality. Literature cited above on similarities in the values and beliefs of parents and their offspring suggests that youths who have higher quality relationships or positive

identification with their parents are more likely to share similar beliefs (Flor & Knapp, 2001; Jennings et al., 2009; Leonard et al., 2013). Thus, parent– child relationship quality might mod- erate the relation between parent and youth racial centrality. Many of our assumptions about developmental change were based on what is known about normative cognitive and social development, but the current study did not include measures of those constructs. Due to limitations of the data, we were also unable to test possible gender differences in the relation be- tween racial centrality of mothers and their children. Future research exploring if and how relations between racial identity in parent– child dyads could vary on the basis of both child and parent gender is warranted, as well as examination of specific ways in which aspects of cognitive and social development interface with youths’ racial centrality.

Another avenue for future research is to study nonparental social influences on racial centrality development. As youths establish greater autonomy from their parents during adolescence, peer relations are likely influences on racial identity (Altschul et al., 2006). Contextual factors such as the racial composition of class- rooms and schools and racial climate within the child’s neighbor- hood and community are probably also important, as are child characteristics such as skin tone (Adams, Kurtz-Costes, & Hoff- man., 2016; Kurtz-Costes & Rowley, 2012; Rowley, Kurtz-Costes, & Cooper, 2010).

Finally, a richer understanding of the development of racial identity in African American youths must also take into account other dimensions of racial identity, how they interact with one another, and how normative maturation interacts with contextual forces to shape their development. Although average levels of racial centrality did not change in our sample across the seven years of the study, levels of private and public regard as well as racial ideologies may have changed in response to national or regional events. Understanding of developmental changes in other dimensions of the MMRI, in particular racial ideology and racial regard, would contribute to theory and could have important im- plications for targeting interventions with African American youths.

This study is a first attempt to examine the longitudinal devel- opment of racial centrality in African American youths and their mothers from late childhood to late adolescence. To date, racial identity development research with African American youths has focused primarily on relations between racial identity and various indices of well-being. Little research has examined either norma- tive developmental patterns or relations between parents’ and youths’ racial identity. Thus, the current study extends the field by showing that the developmental period from late childhood to late adolescence, which is marked by increased cognitive and social- cognitive sophistication, was not accompanied by increased im- portance of race as a social identity among the African American youths in our sample. Youth in our sample showed substantial individual differences in racial centrality, with the stability of these individual differences increasing across time, reaching adult levels by high school. Contrary to our expectations, our results show that the importance of race to identity is not highly congruent across mother-child pairs, and to the extent that mothers influence their children’s racial centrality, the greatest influence occurs in middle adolescence. Given observed protective and promotive relations between racial centrality and various academic and psychosocial

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1139RACIAL CENTRALITY IN AFRICAN AMERICANS

outcomes among African American youths (Rivas-Drake et al., 2014), greater understanding of the development of youths’ racial centrality is warranted. Research on this topic will enrich devel- opmental theory regarding social identities and increase our un- derstanding of family processes within African American families.

References

Adams, E. A., Kurtz-Costes, B. E., & Hoffman, A. J. (2016). Skin tone bias among in African Americans: Antecedents and consequences across the life span. Developmental Review, 40, 93–116. http://dx.doi.org/10.1016/ j.dr.2016.03.002

Ahmed, W., Minnaert, A., Kuyper, H., & van der Werf, G. (2012). Reciprocal relationships between math self-concept and math anxiety. Learning and Individual Differences, 22, 385–389. http://dx.doi.org/10 .1016/j.lindif.2011.12.004

Altschul, I., Oyserman, D., & Bybee, D. (2006). Racial-ethnic identity in mid-adolescence: Content and change as predictors of academic achievement. Child Development, 77, 1155–1169. http://dx.doi.org/10 .1111/j.1467-8624.2006.00926.x

Arnett, A. B., Pennington, B. F., Willcutt, E., Dmitrieva, J., Byrne, B., Samuelsson, S., & Olson, R. K. (2012). A cross-lagged model of the development of ADHD inattention symptoms and rapid naming speed. Journal of Abnormal Child Psychology, 40, 1313–1326. http://dx.doi .org/10.1007/s10802-012-9644-5

Byrd, C. M., & Chavous, T. M. (2009). Racial identity and academic achievement in the neighborhood context: A multilevel analysis. Journal of Youth and Adolescence, 38, 544 –559. http://dx.doi.org/10.1007/ s10964-008-9381-9

Chavous, T. M., Rivas-Drake, D., Smalls, C., Griffin, T., & Cogburn, C. (2008). Gender matters, too: The influences of school racial discrimi- nation and racial identity on academic engagement outcomes among African American adolescents. Developmental Psychology, 44, 637– 654. http://dx.doi.org/10.1037/0012-1649.44.3.637

Cooper, S. M., Smalls-Glover, C., Metzger, I., & Griffin, C. (2015). African American fathers’ racial socialization patterns: Associations with racial identity beliefs and discrimination experiences. Family Re- lations, 64, 278 –290. http://dx.doi.org/10.1111/fare.12115

Cross, W. E. (1991). Shades of black: Diversity in African-American identity. Philadelphia, PA: Temple University Press.

Cross, W. E., Strauss, L., & Fhagen-Smith, P. (1999). African American identity development across the life span: Educational implications. In R. Sheets (Ed.), Racial and ethnic identity in school practices: Aspects of human development (pp. 29 – 47). Mahwah, NJ: Erlbaum.

Dwyer, J. H. (1983). Statistical models for the social and behavioral sciences. New York, NY: Oxford University Press.

Ellis, J. M., Rowley, L. L., Nellum, C. J., & Smith, C. D. (2016). From alienation to efficacy: An examination of racial identity and racial academic stereotypes among Black male adolescents. Urban Education, Advance online publication. http://dx.doi.org/10.1177/0042085915 602538

Erikson, E. H. (1968). Identity: Youth and crisis. New York, NY: Norton. Erikson, E. H. (1993). Childhood and society. New York, NY: Norton &

Company. Flor, D. L., & Knapp, N. F. (2001). Transmission and transaction: Predict-

ing adolescents’ internalization of parental religious values. Journal of Family Psychology, 15, 627– 645. http://dx.doi.org/10.1037/0893-3200 .15.4.627

French, S. E., Seidman, E., Allen, L., & Aber, J. L. (2006). The develop- ment of ethnic identity during adolescence. Developmental Psychology, 42, 1–10. http://dx.doi.org/10.1037/0012-1649.42.1.1

Hart, D., & Damon, W. (1985). Contrasts between understanding self and understanding others. In R. L. Leahy (Ed.), The development of the self. Orlando, FL: Academic Press.

Hughes, D., Rodriguez, J., Smith, E. P., Johnson, D. J., Stevenson, H. C., & Spicer, P. (2006). Parents’ ethnic-racial socialization practices: A review of research and directions for future study. Developmental Psy- chology, 42, 747–770. http://dx.doi.org/10.1037/0012-1649.42.5.747

Jennings, M. K., Stoker, L., & Bowers, J. (2009). Politics across genera- tions: Family transmission reexamined. The Journal of Politics, 71, 782–799. http://dx.doi.org/10.1017/S0022381609090719

Jodl, K. M., Michael, A., Malanchuk, O., Eccles, J. S., & Sameroff, A. (2001). Parents’ roles in shaping early adolescents’ occupational aspi- rations. Child Development, 72, 1247–1266. http://dx.doi.org/10.1111/ 1467-8624.00345

Koepke, S., & Denissen, J. J. (2012). Dynamics of identity development and separation-individuation in parent-child relationships during adoles- cence and emerging adulthood: A conceptual integration. Developmental Review, 32, 67– 88. http://dx.doi.org/10.1016/j.dr.2012.01.001

Kroger, J. (2007). Identity development: Adolescence through adulthood (2nd ed.). Thousand Oaks, CA: Sage.

Kuczynski, L. (2003). Beyond bidirectionality: Bilateral conceptual frame- works for understanding dynamics in parent-child relations. In L. Kuc- zynski (Ed.), Handbook of dynamics in parent-child relations (pp. 3–24). Thousand Oaks, CA: Sage Publications. http://dx.doi.org/10 .4135/9781452229645.n1

Kurtz-Costes, B., & Rowley, S. J. (2012). School transitions and African American youth. In S. Karabenick & T. Urdan (Eds.), Advances in motivation and achievement: Transitions across schools and cultures (Vol. 17, pp. 27–54). http://dx.doi.org/10.1108/S0749-7423(2012) 0000017005

Leonard, K. C., Cook, K. V., Boyatzis, C. J., Kimball, C. N., & Flanagan, K. S. (2013). Parent-child dynamics and emerging adult religiosity: Attachment, parental beliefs, and faith support. Psychology of Religion and Spirituality, 5, 5–14. http://dx.doi.org/10.1037/a0029404

McHale, S. M., Crouter, A. C., Kim, J. Y., Burton, L. M., Davis, K. D., Dotterer, A. M., & Swanson, D. P. (2006). Mothers’ and fathers’ racial socialization in African American families: Implications for youth. Child Development, 77, 1387–1402. http://dx.doi.org/10.1111/j.1467-8624 .2006.00942.x

Miller, R. B., & Glass, J. (1989). Parent-child attitude similarity across the life course. Journal of Marriage and the Family, 51, 991–997. http://dx .doi.org/10.2307/353211

Muthén, L. K., & Muthén, B. O. (1998 –2014). Mplus user’s guide (7th ed.). Los Angeles, CA: Author.

Neblett, E. W., Jr., Banks, K. H., Cooper, S. M., & Smalls-Glover, C. (2013). Racial identity mediates the association between ethnic-racial socialization and depressive symptoms. Cultural Diversity and Ethnic Minority Psychology, 19, 200 –207. http://dx.doi.org/10.1037/a0032205

Neblett, E. W., Jr., Rivas-Drake, D., & Umaña-Taylor, A. J. (2012). The promise of racial and ethnic protective factors in promoting ethnic minority youth development. Child Development Perspectives, 6, 295– 303. http://dx.doi.org/10.1111/j.1750-8606.2012.00239.x

Neblett, E. W., Jr., Smalls, C. P., Ford, K. R., Nguyên, H. X., & Sellers, R. M. (2009). Racial socialization and racial identity: African American parents’ messages about race as precursors to identity. Journal of Youth and Adolescence, 38, 189 –203. http://dx.doi.org/10.1007/s10964-008- 9359-7

Neblett, E. W., Jr., Terzian, M., & Harriott, V. (2010). From racial discrimination to substance use: The buffering effects of racial social- ization. Child Development Perspectives, 4, 131–137. http://dx.doi.org/ 10.1111/j.1750-8606.2010.00131.x

O’Connor, L. A., Brooks-Gunn, J., & Graber, J. (2000). Black and white girls’ racial preferences in media and peer choices and the role of socialization for black girls. Journal of Family Psychology, 14, 510 – 521. http://dx.doi.org/10.1037/0893-3200.14.3.510

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1140 HOFFMAN, KURTZ-COSTES, ROWLEY, AND ADAMS

Pahl, K., & Way, N. (2006). Longitudinal trajectories of ethnic identity among urban Black and Latino adolescents. Child Development, 77, 1403–1415. http://dx.doi.org/10.1111/j.1467-8624.2006.00943.x

Phinney, J. S. (1990). Ethnic identity in adolescents and adults: Review of research. Psychological Bulletin, 108, 499 –514. http://dx.doi.org/10 .1037/0033-2909.108.3.499

Preacher, K. J. (2006). Quantifying parsimony in structural equation mod- eling. Multivariate Behavioral Research, 41, 227–259. http://dx.doi.org/ 10.1207/s15327906mbr4103_1

Quintana, S. (1994). A model of ethnic perspective-taking ability applied to Mexican-American children and youth. International Journal of In- tercultural Relations, 18, 419 – 448. http://dx.doi.org/10.1016/0147- 1767(94)90016-7

Quintana, S. M. (2007). Racial and ethnic identity: Developmental per- spectives and research. Journal of Counseling Psychology, 54, 259 –270. http://dx.doi.org/10.1037/0022-0167.54.3.259

Rivas-Drake, D., Seaton, E. K., Markstrom, C., Quintana, S., Syed, M., Lee, R. M., . . . the Ethnic and Racial Identity in the 21st Century Study Group. (2014). Ethnic and racial identity in adolescence: Implications for psychosocial, academic, and health outcomes. Child Development, 85, 40 –57. http://dx.doi.org/10.1111/cdev.12200

Rogers, L. O., Scott, M. A., & Way, N. (2015). Racial and gender identity among Black adolescent males: An intersectionality perspective. Child Development, 86, 407– 424. http://dx.doi.org/10.1111/cdev.12303

Rowley, S. J., Burchinal, M. R., Roberts, J. E., & Zeisel, S. A. (2008). Racial identity, social context, and race-related social cognition in Af- rican Americans during middle childhood. Developmental Psychology, 44, 1537–1546. http://dx.doi.org/10.1037/a0013349

Rowley, S. J., Kurtz-Costes, B., & Cooper, S. (2010). The schooling of African American children. In J. Meece & J. Eccles (Eds.), Handbook of research on schools, schooling, and human development (pp. 275–292). Hillsdale, NJ: Erlbaum.

Rowley, S. J., Sellers, R. M., Chavous, T. M., & Smith, M. A. (1998). The relationship between racial identity and self-esteem in African American college and high school students. Journal of Personality and Social Psychology, 74, 715–724. http://dx.doi.org/10.1037/0022-3514.74.3.715

Rowley, S. J., Varner, F., Ross, L. L., Williams, A. D., & Banerjee, M. (2012). Toward a model of racial identity and parenting in African Americans. In J. M. Sullivan & A. M. Esmail (Eds.), African American identity: Racial and cultural dimensions of the Black experience (pp. 273–288). Lanham, MD: Rowman & Littlefield Publishing.

Ruble, D., Alvarez, J., Bachman, M., Cameron, J., Fuligni, A., Coll, C., & Rhee, E. (2004). The development of a sense of “we”: The emergence and implications of children’s collective identity. In M. Bennett & F. Sani (Eds.), The development of the social self (pp. 29 –76). East Sussex, England: Psychology Press. http://dx.doi.org/10.4324/9780203391099_ chapter_2

Sameroff, A. (2009). (Ed.) The transactional model of development: How children and contexts shape each other. Washington, DC: American Psychological Association. http://dx.doi.org/10.1037/11877-000

Seaton, E. K., Yip, T., & Sellers, R. M. (2009). A longitudinal examination of racial identity and racial discrimination among African American

adolescents. Child Development, 80, 406 – 417. http://dx.doi.org/10 .1111/j.1467-8624.2009.01268.x

Sellers, R. M., Copeland-Linder, N., Martin, P. P., & Lewis, R. H. (2006). Racial identity matters: The relationship between racial discrimination and psychological functioning in African American adolescents. Journal of Research on Adolescence, 16, 187–216. http://dx.doi.org/10.1111/j .1532-7795.2006.00128.x

Sellers, R. M., Smith, M. A., Shelton, J. N., Rowley, S. A., & Chavous, T. M. (1998). Multidimensional model of racial identity: A reconceptu- alization of African American racial identity. Personality and Social Psychology Review, 2, 18 –39. http://dx.doi.org/10.1207/s15327957pspr 0201_2

Selman, R. (1980). The growth of interpersonal understanding: Develop- mental and clinical analysis. New York, NY: Academic Press.

Umaña-Taylor, A. J., & Fine, M. A. (2004). Examining ethnic identity among Mexican-origin adolescents living in the United States. Hispanic Journal of Behavioral Sciences, 26, 36 –59. http://dx.doi.org/10.1177/ 0739986303262143

Umaña-Taylor, A. J., & Guimond, A. B. (2012). A longitudinal examina- tion of parenting behaviors and perceived discrimination predicting Latino adolescents’ ethnic identity. Journal of Latina/o Psychology, 1, 14 –35. http://dx.doi.org/10.1037/2168-1678.1.S.14

Umaña-Taylor, A. J., Quintana, S. M., Lee, R. M., Cross, W. E., Jr., Rivas-Drake, D., Schwartz, S. J., . . . the Ethnic and Racial Identity in the 21st Century Study Group. (2014). Ethnic and racial identity during adolescence and into young adulthood: An integrated conceptualization. Child Development, 85, 21–39. http://dx.doi.org/10.1111/cdev.12196

U.S. Census Bureau. (2013). Historical income tables: Households (Table H-3). Retrieved from https://www.census.gov/hhes/www/income/data/ historical/household/

Wade, B. H., & Younge, S. (2015). The racial framing of the Barack Obama Presidency: A mixed methods analysis of HBCU student per- ceptions. In D. Cunnigen & M. A. Bruce (Eds.), Race in the Age of Obama: Pt. 2 (pp. 171–199). http://dx.doi.org/10.1108/S0195- 744920140000019009

Whitesell, N. R., Mitchell, C. M., & Spicer, P. (2009). A longitudinal study of self-esteem, cultural identity, and academic success among American Indian adolescents. Cultural Diversity and Ethnic Minority Psychology, 15, 38 –50. http://dx.doi.org/10.1037/a0013456

Wong, C. A., Eccles, J. S., & Sameroff, A. (2003). The influence of ethnic discrimination and ethnic identification on African American adoles- cents’ school and socioemotional adjustment. Journal of Personality, 71, 1197–1232. http://dx.doi.org/10.1111/1467-6494.7106012

Yip, T., Douglass, S., & Sellers, R. M. (2014). Racial and ethnic identity. In F. L. Leong, L. Comas-Diaz, G. C. Nagayama Hall, V. C. McLoyd, & J. E. Trimble (Eds.), APA handbook of multicultural psychology: Vol. 1. Theory and research (pp. 179 –205). Washington, DC: American Psychology Association.

Received October 28, 2015 Revision received December 27, 2016

Accepted January 4, 2017 �

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1141RACIAL CENTRALITY IN AFRICAN AMERICANS