Mothers' and fathers' racial socialization in African American families: Implications for youth

Mothers’ and Fathers’ Racial Socialization in African American Families:

Implications for Youth

Susan M. McHale, Ann C. Crouter, Ji-Yeon Kim, Linda M. Burton, Kelly D. Davis, Aryn M. Dotterer,and Dena P. Swanson

The Pennsylvania State University

Mothers’ and fathers’ cultural socialization and bias preparation with older (M 5 13.9 years) and younger(M 5 10.31 years) siblings were studied in 162 two-parent, African American families. Analyses examinedwhether parental warmth and offspring age and gender were linked to parental practices and whether parents’warmth, spouses’ racial socialization, or youth age or gender moderated links between racial socialization andyouth outcomes. Parental warmth was linked to parents’ socialization. Mothers engaged in more socializationwith older offspring, and fathers more with sons. Mothers’ cultural socialization was positively related to youthethnic identity and fathers’ was negatively related to youth depression symptoms. Youth exhibited a lower locusof control when mothers were high but fathers were low in racial socialization.

Racial socialization refers to the means throughwhich ‘‘parents shape children’s learning about theirown race and about relations between ethnicgroups’’ (Hughes & Johnson, 2001, p. 981) and is acentral dynamic in most African American families(Coard & Sellers, 2005; Sanders Thompson, 1994;Stevenson, 1995; Thornton, Chatters, Taylor & Allen,1990). One reason why research on racial socializa-tion is important is because it illuminates the ways inwhich sociocultural factors are manifested in thefamily lives of children and youth: African Americanparents face childrearing demands that are commonto all parents, but also demands that are unique totheir sociocultural group, given its history and po-sition as a disadvantaged minority in the larger so-cial structure (e.g., Boykin & Toms, 1985; Peters,1985). The study of racial socialization is also im-portant because it expands the scope of past researchon African American youth and families, which hastended to focus on dysfunction and pathology (Hill,Murry, & Anderson, 2005; McLoyd, 1998).

Grounding our work in ecological and familysystems perspectives, in the present study we ex-amined the family contexts of racial socialization andits implications for youth. A family systems per-spective directs attention to similarities and differ-ences in the experiences of different family membersand to reciprocal relations between subsystems in thefamily, including the mother – offspring and father –offspring subsystems (Whitchurch and Constantine1993). An ecological perspective, in turn, highlightsthe contextual embeddedness of individuals’ experi-ences. In the case of family relationships, dimensionsof relationships may go together in different ways indifferent family contexts (e.g., Parke & Buriel, 1998),and objectively similar experiences may have differ-ent implications for individuals in different settings(Bronfenbrenner & Crouter, 1983). These ideas pro-vided the foundation for this study.

Specifically, focusing on a sample of two-parentAfrican American families with two youth in middlechildhood and adolescence, this study was directed attwo goals. First, to illuminate the family relationshipcontexts of racial socialization, we measured the linksbetween mothers’ and fathers’ warmth toward theiroffspring and their racial socialization practices, andwe tested whether parents differed in their socializa-tion of daughters versus sons or older versus youngeroffspring. As we describe below, although parents’individual characteristics and experiences have beenstudied as correlates of racial socialization, we knowalmost nothing about the relationship contexts ofparents’ practices. Further, given inconsistencies in

r 2006 by the Society for Research in Child Development, Inc.All rights reserved. 0009-3920/2006/7705-0020

This work was funded by a Grant from the National Institute ofChild Health and Human Development, RO1-HD32336-02, SusanM. McHale and Ann C. Crouter, Co-Principal Investigators. Wethank Marni Kan, Ashleigh May, Jaime Marks, Carolyn Ransford,Cindy Shearer, Shawn Whiteman, Megan Winchell, SandeeHemman, and Kristen Johnston and Temple University’s SurveyResearch Center for their help in conducting this study and theparticipating families for their time and insights about their familylives.

Correspondence concerning this article should be addressed toSusan M. McHale, Department of Human Development and Fa-mily Studies, 105 White, Pennsylvania State University, UniversityPark, PA 16802. Electronic mail may be sent to [email protected].

Child Development, September/October 2006, Volume 77, Number 5, Pages 1387 – 1402

extant data on age and gender differences in youth’sracial socialization (Hughes, Rodriguez, Smith, John-son, & Stevenson, in press), we compared parents’practices with two offspring; such within-familycomparisons provide for a more sensitive test of ageand gender differences in experiences with parentsthan do between-family comparisons of unrelatedyouth (McHale, Crouter, & Whiteman, 2003).

Our second research goal was to assess the linksbetween mothers’ and fathers’ racial socializationand youth’s ethnic identity development, locus ofcontrol, and symptoms of depression. A small bodyof work has begun to examine the associations be-tween racial socialization and youth functioning, butthe findings are inconsistent (Hughes et al., in press).From an ecological perspective, such inconsistenciesmay arise because racial socialization has differentimplications in different social settings. To test thispossibility, we built on prior research to explore thefamily and relationship conditions under which thelinkages between racial socialization and youthfunctioning were more or less apparent. In additionto examining youth age and gender, we also studiedparental warmth and the racial socialization practicesof the other parent as potential moderators of thelinks between mothers’ and fathers’ racial socializa-tion practices and youth psychosocial functioning.

Studies of racial socialization highlight its multi-dimensionality (e.g., Boykin & Toms, 1985; Hughes& Chen, 1997; Stevenson, 1995). Most con-ceptualizations of racial socialization encompassparents’ efforts to foster children’s knowledge andappreciation of cultural values and practices; pre-pare children for experiences, such as discrimina-tion, that stem from their minority status; andsocialize children for experiences in the larger so-ciety, including teaching respect for a diversity ofcultural groups. Given that more is known about theformer two kinds of dynamics, in this paper we focuson parents’ cultural socialization, or efforts to fosteryouth’s knowledge and appreciation of AfricanAmerican culture, as well as their preparation for bias,that is, efforts to build awareness of and providestrategies for coping with prejudice and discrimina-tion. In the following pages, we review the literatureon racial socialization in African American familiesthat provides the basis for our research goals.

Explaining Variation in Racial Socialization

A body of work reveals that parents vary con-siderably in the extent and nature of their racial so-cialization practices. Most of this work examines the

personal characteristics and experiences of parentsthat are linked to their parenting practices. One lineof study focuses on the associations between parents’socio-demographic characteristics and their sociali-zation practices. Consistent findings are that women,older parents, and parents with higher levels ofeducation and income engage in more cultural so-cialization practices such as providing books aboutAfrican American history or taking children to cul-tural events (e.g., Hughes & Chen, 1997; Thornton etal., 1990). A smaller set of studies has examinedparents’ racial orientations and experiences andshows that factors such as parents’ experiences ofdiscrimination and their racial identity are positivelyrelated to their socialization practices (Hughes &Chen, 1997; Thomas & Speight, 1999). Findings thatparents’ individual characteristics and experiencesare related to their racial socialization mean that, atleast in some families, mothers and fathers maydiffer in their parenting practices.

Our family systems and ecological frameworkshighlight that parents’ socialization practices takeplace within a relationship context. As such, it isimportant to learn how racial socialization is con-nected to other parent – offspring relationship dy-namics. In this study, we focused on parentalwarmth as a potentially important correlate of racialsocialization. Parental warmth is a key dimension ofparent – child relationships, and one that has beenlinked to a range of youth outcomes (e.g., Darling &Steinberg, 1993; Parke & Buriel, 1998). To the extentthat racial socialization practices evolve in the con-text of emotionally involved and engaged parenting,they may be related to the level of warmth in theparent – child relationship. Indeed, focusing on asample of preschool children and their caregivers,Caughy, O’Campo, Randolph, and Nickerson (2002)found modest positive associations between indicesof involved parenting and a range of racial sociali-zation practices. In addition, Frabutt, Walker, andMacKinnon-Lewis (2002) found that moderate levelsof socialization regarding discrimination were linkedto mothers’ positive parenting of young adolescents.We built on this work, focusing on a sample ofschool- and adolescent-age youth, and testing theselinkages separately for mothers and fathers.

A systems perspective directs attention to re-ciprocal influences in family members’ relationshipexperiences, including how children’s characteristicsaffect the parenting practices of their mothers andfathers. Across the transition to adolescence, for ex-ample, parents may see that an expanding socialworld exposes their offspring to individuals fromother ethnic backgrounds as well as to racism and

1388 McHale et al.

discrimination and that their offspring are moreaware of race-based differential treatment. In addi-tion, cognitive developmental advances mean thatyouth may become better able to understand par-ents’ messages about race relations. Finally, the sal-ience of identity formation as a developmental taskmay foster youth’s interest in cultural values, tradi-tions, and achievements. These developmentalchanges may elicit more and different racial sociali-zation practices on the parts of parents as youthmature (Hughes & Chen, 1997; Hughes & Johnson,2001; Quintana, 1998). Extant data suggest that agedifferences in racial socialization are most apparentwhen samples include adolescents or youth from abroad age span (e.g., Fatimilehin, 1999); further, ageeffects are more consistent in parental socializationabout discrimination than in cultural socialization(e.g., Hughes & Chen, 1997). In our analysis of therole of offspring age in parents’ practices, we ex-panded on extant research in two ways. First, giventhat studies to date have focused mostly on mothersor asked youth about ‘‘parents’’ in general, we as-sessed the role of offspring age in the socializationpractices of both mothers and fathers. Second, ourwithin-family design allowed us to test whethermothers’ and fathers’ practices varied for older andyounger offspring in the same family. Because such awithin-family approach controls for other char-acteristics, it provides for a more sensitive test ofmother – father and offspring age differences in par-ental practices.

Parents may also differ in the socialization stra-tegies they use with their sons versus their daugh-ters. These differences may emerge because boys andgirls face different challenges as they are growing up,and also because parents hold different expectationsfor their sons’ and daughters’ adult roles and ex-periences. For instance, the stereotypically femininerole of kin keeper may mean that girls are subject tomore socialization around cultural traditions andvalues (Hagestad, 1986). In contrast, minority males’more frequent experiences of racial discrimination(e.g., Fischer & Shaw, 1999; Stevenson, Cameron,Herrero-Taylor, & Davis, 2002) may make parentsmore concerned with preparing sons as comparedwith daughters for such experiences. Indeed, al-though findings of gender differences in racial so-cialization are less consistent than findings of agedifferences, Hughes et al. (in press) concluded thatboys may receive more socialization messages aboutdiscrimination, and girls, more messages about ra-cial pride. A limitation of this work, however, is thatmost studies have focused on mothers or askedyouth about their ‘‘parents’’ in a general way. Thus,

an important direction for research is to comparemothers’ and fathers’ racial socialization practiceswith boys and girls. Research on parents’ genderroles suggests that parents, particularly fathers, maybe more involved in socialization activities withoffspring of their own gender (McHale et al., 2003);in the case of racial socialization as well, parents may‘‘specialize’’ in the socialization of same-sex off-spring. To test this idea we used a within-familydesign, which allowed us to compare how mothersand fathers from the same families socialized theirsons versus their daughters.

Links Between Racial Socialization and YouthPsychosocial Functioning

A growing body of work examines the potentialimplications of parents’ racial socialization practicesfor their offspring. One set of studies is grounded inthe hypothesis that parents who highlight issues ofrace and ethnicity in their socialization practices willhave offspring with stronger or more mature ethnicidentity. By making culture salient and by providinginstruction about cultural practices and theachievements of group members, parents build prideand knowledge of cultural traditions and values thatunderlie youth’s ethnic identity; children may alsomodel their parents’ identification with their culturalgroup (e.g., Sanders Thompson, 1999). Althoughsome studies have not found linkages between par-ents’ racial socialization and youth’s ethnic identity(e.g., Phinney & Chavira, 1995), most studies havedetected such associations in samples of AfricanAmerican children and adolescents (e.g., Marshall,1995; Spencer, 1983; Stevenson, 1995) as well as inretrospective reports by African American adults(Sanders Thompson, 1994, 1999). One possibility isthat racial socialization’s effects on ethnic identityvary as a function of the specifics of parents’ mes-sages. Based on their review of the literature, forexample, Hughes et al. (in press) concluded that thelink between cultural socialization and youth ethnicidentity is among the most well-documented im-plications of parents’ racial socialization; in contrast,the link between preparation for bias and ethnicidentity is less clear cut. As noted, however, moststudies focus on mothers’ racial socialization, and wemay find that different patterns emerge when fathersare included: Given that African American maleshave more experience with discrimination than doAfrican American females (Stevenson et al., 2002),fathers’ role in preparing offspring for bias may bemore important in youth ethnic identity than ismothers’.

Mothers’ and Fathers’ Racial Socialization 1389

Another set of studies on the implications of racialsocialization addresses the links between parentalpractices and youth adjustment. Working within arisk and resilience framework, scholars have arguedthat racial socialization builds pride in and a sense ofbelonging to a racial group and may help youth todevelop strategies for coping with racism. In theseways, racial socialization may protect youth from thesequelae of negative experiences associated withminority status such as personal and vicarious ex-periences of discrimination, negative images ofAfrican Americans, and inequalities in the distribu-tion of opportunities and resources in the larger so-ciety (Hughes & Chen, 1997; Miller, 1999; Stevenson,Reed, Bodison, & Bishop, 1997). Importantly, how-ever, some forms of racial socialization may give riseto adjustment problems as when, for example, youthcome to expect unfair treatment or see limits to theiropportunities (Rumbaut, 1994; Stevenson et al.,1997). As such, research on the adjustment implica-tions of racial socialization must take into accountthe nature of parents’ socialization messages(Hughes et al., in press).

In the face of its theoretical and practical sig-nificance, the empirical literature on the links be-tween racial socialization and youth functioningremains small, and the results are inconsistent. Asnoted, some investigators have found negative ad-justment correlates of socialization that promoteyouth’s orientations toward the majority culture(Constantine & Blackmon, 2002) or that highlightracial barriers and discrimination (Ogbu, 1974;Rumbaut, 1994), and a handful have found no youthwell-being correlates of racial socialization (e.g.,Johnson, 2001; Phinney & Chavira, 1995). Most stud-ies of cultural socialization, however, documentmodest positive associations between racial sociali-zation and positive adjustment (Bowman & Howard,1985; Murry & Brody, 2002; Scott, 2003), and negativeassociations between racial socialization and adjust-ment problems (Constantine & Blackmon, 2002;Stevenson et al., 1997).

There are several limitations to this literature,however. First, virtually all work has focused onmothers’ socialization and, with some importantexceptions (e.g., Johnson, Brown, Wayumba, Choi, &Hunter, 2005), we know little about the implicationsof African American fathers’ socialization practicesfor youth. Findings from the larger literature on fa-thers, however, highlight fathers’ role in connectingoffspring to the world outside the family (Parke &Buriel, 1998); in addition, as noted, prior work in-dicates that African American males have differentexperiences than African American females vis-a-vis

the majority culture (Fischer & Shaw, 1999; Ste-venson et al., 2002). By studying the implications ofboth mothers’ and fathers’ practices in the presentstudy, we may find that fathers play a uniquely im-portant role in racial socialization.

Another limitation is extant studies’ reliance onself-reports from one family member (Coard & Sell-ers, 2005; Hughes et al., in press). When parents oryouth provide information on both racial socializa-tion practices and youth ‘‘outcomes,’’ associationsmay be inflated because of reporter bias. For thisreason, we relied on information on racial socializa-tion collected from mothers and fathers, and in-formation about ethnic identity and adjustmentcollected from youth.

From an ecological perspective, an important di-rection for research on racial socialization is to learnmore about the conditions under which parents’ prac-tices have implications for youth. Inconsistent find-ings across studies may emerge because racialsocialization has different implications for youthwith different characteristics and from different so-cial settings. We studied offspring age and gender astwo child characteristics that may moderate the linksbetween racial socialization and youth outcomes.With respect to age, racial socialization may havemore positive implications for older offspring whohave the maturity to understand parents’ messagesand who are in a position to use those messages asthey negotiate the world beyond the family. Withrespect to gender, parents may specialize in racialsocialization with same-sex offspring, and messagesfrom same-sex parents may have special credibilityfor youth.

To shed light on how other family dynamicsprovide a context for racial socialization, we alsostudied each parent’s warmth and the racial sociali-zation practices of the other parent as potentialmoderators of mothers’ and of fathers’ socializationpractices. Darling and Steinberg’s (1993) hypothesisabout the role of the emotional tone of the parent –offspring relationship in child socialization was thegrounding of our focus on parental warmth as amoderator of the link between racial socializationand youth functioning. Specifically, these authorsproposed that parental practices are more effectivewhen the emotional climate of the parent – child re-lationship is positive, because such a climate pro-motes youth’s willingness to be socialized. As such,the combination of high parental warmth and highinvolvement in racial socialization should give riseto stronger ethnic identity and more positive well-being in youth because youth will be more likelyto attend to and assimilate parents’ socialization

1390 McHale et al.

messages, for example, about pride in their heritageand effective strategies for coping with unfairtreatment.

Research on co-parenting provides another sce-nario on the conditions under which racial sociali-zation will have implications for offspring. Co-parenting is a family systems construct that refersto the extent to which mothers and fathers worktogether in the tasks of childrearing, including sup-porting one another in their parenting roles, backingup one another in their childrearing decisions anddisciplinary practices, and conveying consistent so-cialization messages to their offspring (Feinberg,2003). An emerging body of research documents thatsuch teamwork by parents is linked to positive youthadjustment (e.g., Margolin, Oliver, & Medina, 2001).Although the co-parenting construct has not yet beenapplied to the study of racial socialization, the ex-isting literature led us to expect that, when twoparents convey the same racial socialization mes-sages to offspring, the message will be more readilylearned. Mismatches between parents’ practices, incontrast, may reflect differing perspectives or evenovert conflicts between parents about the importanceof racial socialization; in some families, one parentmay also intensify his or her racial socialization ac-tivities when the other parent is seen as uninvolved.Grounding our ideas in extant research and theoryon co-parenting, we tested whether the implicationsof parents’ racial socialization practices depended onthe practices of the spouse. First, we expected that,when both parents were highly involved in racialsocialization, youth would report stronger ethnicidentities. In addition, given findings from prior re-search on the negative adjustment implications ofinterparental disagreements, we also anticipated thatmismatches between fathers’ and mothers’ practiceswould be linked to lower levels of locus of controland higher levels of depression symptoms in youth.Importantly, our focus on the patterning of mothers’and fathers’ racial socialization practices is con-sistent with the call to study racial socialization as amultidimensional phenomenon (Coard & Sellers,2005).

Method

Participants

The data came from fathers, mothers, and twooffspring in 162 families that were participating inthe first phase of a short-term longitudinal study ofgender socialization and development in two-parent,Black/African American families. Given the goals ofthe larger investigation, a study of family gender

socialization, we did not seek a representative sam-ple. Rather, we sought families that self-identified asbeing Black or African American and included botha mother and father who were living together andrearing at least two middle-childhood – adolescence-aged offspring.

To generate the sample, we targeted two con-tiguous urban centers on the eastern seaboard thatcensus data revealed had substantial African Amer-ican populations; within the zip code areas fromwhich we recruited, 38% of households includedmarried African American parents with at least onechild between 6 and 17 years of age (U.S. Census,2001). We used two strategies for recruiting families.First, we hired African Americans residing in thetargeted communities to recruit families by postingflyers in local businesses, providing information onthe study to local churches, and distributing flyers atyouth activities. Interested families then contactedthe recruiters who passed on their names to theproject office. Approximately half the sample wasrecruited using this procedure. We also purchased amarketing firm list that included names and ad-dresses of African American students in grades 4 – 7who lived in the geographic region of interest. Wesent letters to families that described the study andincluded an 800 number to call and a postcard toreturn to the project office if the family fit the studycriteria and was interested in participating. Becausethe marketing firm could not determine familystructure or either presence or age of siblings, how-ever, many letters went to families that did not meetproject criteria. Of 1,796 letters sent, 131 were re-turned by the post office as undeliverable. Of the 142families that expressed interest, 93 were eligible and49 ineligible. Of those eligible, 86 of the families re-cruited via the marketing list were interviewed and 7were not either because the family could not be lo-cated based on the information provided on theirpostcards or because the family was too busy.

The characteristics of the 162 participating fa-milies are shown in Table 1. Families were generallyworking and middle class based on family incomeand parent education, and the majority (80%) in-cluded two or three children (M 5 2.81, SD 5 1.18,range 5 2 – 8 children); in families with more thantwo children, we focused on the two who were in ourage range and proximal in birth order. Most parentswere employed (98% of both mothers and fathers),with mothers working an average of 33.48 hr/week(SD 5 17.46) and fathers working an average of44.22 hr (SD 5 18.00). Mothers’ job prestige averaged48.31 (SD 5 11.87) and fathers’ averaged 47.12(SD 5 13.33) based on the National Opinion Research

Mothers’ and Fathers’ Racial Socialization 1391

Council’s codes; jobs in this range include teacher’saide (48.75), dental assistant (47.80), and police/de-tective (47.68). Although the average family incomein this sample was almost $90,000, the range wasconsiderable ($3,500 – $525,000). We also calculatedincome-to-needs ratios. The average family had anincome 4.43 times higher than the poverty level forfamilies of their size, but again, the range was con-siderable (SD 5 2.96). Families also came from awide range of communities. School data, for ex-ample, indicated that the schools youth attendedvaried in their racial composition from less than 5%to 100% minority enrollment; the average youth inthis sample attended a school in which about twothirds of the students were African American.

With respect to target youth, as Table 1 indicates,youth were in middle childhood and adolescence.The sample was approximately equally dividedbased on the sex constellation of the sibling dyad andincluded n 5 165 girls (older and younger siblingscombined) and n 5 159 boys. As Table 1 also shows,the majority of youth were biologically related toboth parents, although more often related to mothersthan fathers; in six families, youth were living withcouples other than their biological, step, or adoptiveparents (e.g., grandparents; aunts and uncles). Withrespect to couple relationships, 15 sets of parentswere cohabiting and the rest were married. In fa-

milies in which youth were not living with always-married parents, couples had been living togetherfor at least 4 years.

Procedures

Mothers, fathers, and two youth from each familywere interviewed in their homes by a team of twointerviewers, almost all of whom were AfricanAmerican. The interviews began with a brief de-scription of the study and a review of informedconsent procedures. Then, family members wereinterviewed individually. To maintain interest, in-terviewers used a variety of procedures, includingcard sorts, response cards, and questionnaires; in thecase of children under age 13 and family memberswith reading difficulties, all questions were pre-sented orally. For all measures, except the index ofdepression symptoms, family members were askedto report on their experiences during the past year.Parent interviews lasted about 2 hr and youth inter-views lasted about 1 hr. Following completion of theinterviews, families were sent a $200 honorarium.

In the recruitment process, families self-identifiedas Black or African American. At the beginning ofeach individual interview, family members wereasked whether they preferred to refer to themselvesas Black or as African American, and that term wasused to refer to the family member during the re-mainder of the interview. In addition, youth whowere not growing up with two biological parentswere asked how they wished to refer to their parentfigures, and these terms were used during the re-mainder of the interview.

Measures

Racial socialization was assessed using two sub-scales from a measure developed by Hughes andChen (1997) that assessed cultural socialization (5items; e.g., ‘‘I’ve read or provided Black historybooks to my child’’) and preparation for bias (7items; e.g., ‘‘I’ve talked to my children about ra-cism’’). For each item, parents used a 6-point ratingscale that ranged from never to very often, to describetheir socialization practices; at separate points in theinterview, each parent reported on their practiceswith their older and younger child. Reliability of thescales was satisfactory: Cronbach’s as ranged from.80 (mothers’ reports of cultural socialization witholder child) to .90 (fathers’ preparation for bias witholder child). Parents’ reports of racial socializationwith their two children were highly correlated, ran-ging from r 5 .75, po.01 for mothers’ preparation for

Table 1

Means, Standard Deviations, and Percentages for Background Char-

acteristics of the Sample (N 5 162)

Variables M SD %

Fathers’ educationa 14.27 2.37

Mothers’ educationa 14.67 1.84

Family incomeb $90,105.00 $57,805.00

Parents’ years living together 15.62 6.00

Older siblings’ age 13.95 1.99

Younger siblings’ age 10.31 1.10

Fathers’ age 43.70 6.93

Mothers’ age 41.00 5.65

% Biological mothers of

older siblings

92.60

% Biological mothers of

younger siblings

95.70

% Biological fathers of

older siblings

79.00

% Biological fathers of

younger siblings

85.20

aNote. 12 5 high school graduate, 14 5 some college, 16 5 collegegraduate.bSum of mothers’ plus fathers’ income; 10 mothers and 14 fathersrefused to report their income.

1392 McHale et al.

bias with their older and younger children, to r 5 .88,po.01 for fathers’ cultural socialization.

Parental warmth was measured using an 8-itemscale from the parent version of the Child’s Report ofParental Behavior Inventory (Schwarz, Barton-Henry, & Pruzinsky, 1985). For each item (e.g., ‘‘I ama person who makes my child feel better after talkingover his/her worries with me’’), parents used a 5-point scale (1 5 not at all to 5 5 very much) to describetheir relationships during the past year with theirolder and younger child, respectively. Cronbach’s asranged from .83 (mothers’ reports of warmth witholder child) to .89 (fathers’ reports of warmth witholder child), and mothers’ and fathers’ ratings werenot highly correlated, r 5 .24, po.01 and r 5 .12, ns,for older and younger offspring.

Youth ethnic identity was assessed using the Mul-tigroup Ethnic Identity Measure (Phinney, 1992). Foreach of 10 items, youth used a 4-point scale(1 5 strongly disagree; 4 5 strongly agree) to rate howwell the item described them over the past year (e.g.,‘‘I feel a special connection to other African Amer-icans/Blacks’’; ‘‘I am happy that I am AfricanAmerican/Black’’). Although the ten items originallycame from two subscales (Affirmation and Belong-ing and Ethnic Behaviors and Practices), the scaleswere highly correlated, r 5 .51, po.01 for older sib-lings and r 5 .62, po.01 for younger siblings, andtherefore combined such that high scores reflect astronger and more positive orientation to AfricanAmericans and their culture. Cronbach’s as were .73for older and .80 for younger siblings’ ethnic identity,and siblings’ reports were uncorrelated, r 5 � .01, ns.

Youth locus of control was assessed using the 21-item version of the Nowicki – Strickland Internal –External Control Scale for Children (Nowicki &Strickland, 1973). On this measure, youth indicatedwhether they agreed or disagreed (1 5 yes, 2 5 no)with statements such as, ‘‘Most problems will solvethemselves if I just don’t fool with them.’’ Cron-bach’s as were .71 for older and .68 for youngersiblings, and siblings’ reports were correlated,r 5 .25, po.01.

Youth depressive symptoms were assessed using the10-item version of the Children’s Depression In-ventory (Kovacs, 1981). For each item, youth choseone of three statements that best represented theirfeelings over the past week, e.g., ‘‘I am sad once inawhile,’’ ‘‘I am sad many times,’’ or ‘‘I am sad all thetime.’’ Reliabilities for the depression scale were .76and .72 for older and younger siblings, respectively,and siblings’ reports were uncorrelated, r 5 .06, ns.Owing to skewness on this measure we used squareroot-transformed scores.

Family background information, including familymembers’ birth dates and levels of education, in-formation about the couple relationship (marital sta-tus, duration of co-residence), information aboutyouth’s relationship to each parent and to one another,and information about parents’ employment (status,occupation, income), was provided by parents.

Results

The results are organized around our research goals.First, we tested whether parental warmth was linkedto mothers’ and fathers’ cultural socialization andpreparation for bias, and whether practices varied asa function of offspring’s age or gender. We then ex-amined the links between mothers’ and fathers’ ra-cial socialization practices and youth’s ethnicidentity, locus of control, and depression symptoms,and we tested the potential moderating roles ofparental warmth, the other parent’s racial socializa-tion, offspring age, and offspring gender in theselinkages.

Links Between Racial Socialization and Parental Warmth

In a preliminary step, we used one-way analysesof variance (ANOVAs), with parent as a within-groups factor, to compare fathers’ and mothers’racial socialization practices. Given the high corre-lation between parents’ practices with their olderand younger offspring and our focus on offspringage in later analyses, for this descriptive purpose, weused the means of each parent’s preparation for biasand cultural socialization with their two offspring asthe dependent variables in these analyses. A sig-nificant parent effect emerged for cultural socializa-tion, F(1, 161) 5 45.08, po.001, M 5 17.95, SD 5 5.94and M 5 21.38, SD 5 5.13 for fathers and mothers,respectively, and a trend emerged for preparation forbias, F(1, 161) 5 2.77, po.10, M 5 26.73, SD 5 7.27and M 5 27.79, SD 5 6.26 for fathers and mothers. Asthese means indicate, parents’ ratings were above themidpoint of the 6-point rating scale, indicating thaton average, parents ‘‘sometimes’’ engaged in thesepractices, and mothers reported more frequent racialsocialization than fathers, particularly in the case ofcultural socialization. In a second preliminary step,we examined the correlations between mothers’ andfathers’ socialization practices. These analyses re-vealed only modest positive associations betweenmothers’ and fathers’ reports of cultural socializa-tion, r 5 .32, po.01, and preparation for bias, r 5 .28,po.01. In some families, mothers and fathers maydiffer in their practices, and in other families parents’practices may be more similar.

Mothers’ and Fathers’ Racial Socialization 1393

To study how parental warmth and youth age andgender were linked to parents’ racial socialization, wetested a series of multilevel models (MLM), examiningmothers’ and fathers’ cultural socialization and prep-aration for bias in separate analyses, using theHierarchical Linear and Nonlinear Modeling Statis-tical Package (HLM version 5.05; Raudenbush, Bryk,Cheong, & Congdon, 2003). MLM extends multipleregression to take into account the nonindependenceassociated with nested data, in our case, sibling data.Prior studies of age and gender differences in youth’sracial socialization experiences have relied on be-tween-family comparisons; an MLM approach, incontrast, allowed us to test whether gender and agedifferences between siblings were linked to differencesbetween siblings’ racial socialization experiences.Specifically, we took advantage of the MLM frame-work, structuring our analysis to focus on the corre-lates of within-family differences; this focus is notinherent in multilevel models, but as described below,requires only a minor elaboration (see Jacobs, Lanza,Osgood, Eccles, & Wigfield, 2002).

At Level 1, the within-family model, explanatoryvariables were unique to each sibling and includedmothers’ and fathers’ warmth as well as offspring’sgender (0 5 female, 1 5 male) and age. In this sample,youth age was confounded with birth order, r 5 .74,po.001, and thus both factors could not be includedin the models. At Level 2, the between-family model, weincluded variables that were common to both sib-lings. Specifically, as control variables at this level,we included mother and father age, mother and fa-ther education, and a dummy code to index off-spring relatedness (0 5 full siblings, 1 5 other). Wealso included mean age and mean gender of off-spring at Level 2 so that we could separate within-family age and gender effects from between-familyeffects. The results for the model tests are shown inTable 2, and the equations for the model appear inthe Appendix. In reporting the results, we focus oneffects at po.05, but consider trend level effects(po.10) when these were consistent with resultsfrom prior research or with a larger pattern of sig-nificant findings.

Table 2

Coefficients, Standard Errors, and T Ratios for Links Between Parental Warmth and Mothers’ (M) and Fathers’ (F) Racial Socialization

Mothers’ cultural

socialization

Mothers’ preparation for

bias

Fathers’ cultural

socialization

Fathers’ preparation for

bias

B SE t ratio B SE t ratio B SE t ratio B SE t ratio

Fixed effects

Intercept 21.69�� 0.45 47.82 28.34�� 0.59 47.97 17.57�� 0.52 34.10 26.24�� 0.64 40.77

Level 1

Youth age 0.14� 0.06 2.45 0.97�� 0.10 10.11 � 0.03 0.06 � 0.54 0.08 0.10 0.88

Youth gender � 0.28 0.32 � 0.88 � 0.27 0.53 � 0.52 0.65� 0.34 1.92 1.30� 0.52 2.50

M warmth 0.89� 0.42 2.12 2.29�� 0.64 3.59 0.10 0.45 0.22 0.84 0.65 1.29

F warmth 0.17 0.35 0.49 0.04 0.53 0.07 1.38�� 0.37 3.71 2.52�� 0.54 4.63

Level 2

Youth agea 0.10 0.32 0.32 � 0.80w 0.41 � 1.96 0.13 0.37 0.35 0.24 0.45 0.53

Youth gendera � 2.19w 1.15 � 1.90 � 1.55 1.48 � 1.05 � 2.74� 1.32 � 2.08 � 3.47� 1.63 � 2.14

M age 0.10 0.12 0.80 � 0.14 0.15 � 0.91 � 0.11 0.14 � 0.77 � 0.03 0.17 � 0.20

F age � 0.08 0.09 � 0.85 0.14 0.12 1.17 0.23� 0.11 2.10 0.23w 0.13 1.76

M education 0.46w 0.24 1.91 0.56w 0.30 1.83 � 0.16 0.28 � 0.57 � 0.22 0.34 � 0.65

F education 0.32w 0.19 1.73 0.16 0.23 0.71 0.63�� 0.21 2.94 0.23 0.26 0.89

Biological relatednessb � 0.93 1.04 � 0.89 � 2.31w 1.30 � 1.77 0.38 1.20 0.31 � 0.80 1.45 � 0.55

Variance SD Variance SD Variance SD Variance SD

Variance components

Residual 4.07 2.02 10.98 3.31 4.47 2.11 10.72 3.27

Intercept 21.58�� 4.65 31.33�� 5.60 28.88�� 5.37 40.14�� 6.34

aNote. Family-level age and gender were included as covariates to capture within-family differences beyond between-family effects.b0 5 full sibling; 1 5 other (e.g., half sibling, step sibling).wpo.10; �po.05; ��po.01.

1394 McHale et al.

Parental warmth, youth characteristics, and mothers’racial socialization. Maternal warmth was a positiveand significant correlate of mothers’ cultural socia-lization (see Table 2). Offspring age was also sig-nificant at Level 1, indicating that mothers engagedin more cultural socialization with older as com-pared with younger siblings. Among the controlvariables, both mothers’ and fathers’ education werelinked, at trend level, to maternal cultural socializa-tion. Turning to mothers’ preparation for bias, theanalyses again revealed a positive effect for maternalwarmth, a positive effect for offspring age, and atrend-level positive effect for the control variable,maternal education.

Parental warmth, youth characteristics, and fathers’racial socialization. Fathers’ reports of warmth werepositively and significantly related to their culturalsocialization practices (see Table 2). There also was aneffect at Level 1 for offspring gender, indicating thatfathers engaged in more cultural socialization withtheir sons than with their daughters. In addition, atLevel 2, fathers’ education and age were significantpositive covariates. With respect to fathers’ prepara-tion for bias, paternal warmth was again a positiveand significant predictor and a gender effect in-dicated that fathers engaged in more bias prepara-tion with their sons than with their daughters.Fathers’ age also was a trend level positive predictor.

Parents’ Racial Socialization and Youth PsychosocialFunctioning

To address our second goal, we used an MLMstrategy to test a series of two-level models focusedon the implications of parents’ cultural socializationand preparation for bias for youth’s ethnic identity,locus of control, and depression symptoms. In aneffort to simplify presentation of the results and so asto exclude nonsignificant interaction terms (whichcontribute to increases in standard errors; Aiken &West, 1991), we ran a preliminary set of analyses thattested the potential moderating roles of parentalwarmth, spouses’ racial socialization, offspringgender, and offspring age. Contrary to our hypoth-esis, these tests revealed no significant interactionsbetween racial socialization and parental warmth,that is, no evidence of stronger links between par-ental practices and youth functioning under condi-tions of high parental warmth. In contrast, we foundsome evidence for a moderating role of spouses’racial socialization. Accordingly, parental warmthwas included only as a control variable, but we re-tained the interaction between mothers’ racialsocialization and fathers’ racial socialization in the

models. These preliminary analyses also revealed noevidence that offspring gender moderated the linksbetween racial socialization and youth functioning,but some evidence of the moderating effects of age.Accordingly, in the final models we included off-spring gender only as a control variable, but retainedthe interaction between youth age and racial socia-lization. Finally, there were no significant three-way interactions, so these were excluded from themodels.

We conducted an additional set of preliminaryanalyses examining the implications of mothers’ andfathers’ socialization practices in separate models toascertain whether including indices of both parents’practices in the same model altered the pattern ofresults (e.g., due to multicollinearity, suppressor ef-fects). The results for paternal and maternal practicesdid not differ when these factors were examined inseparate versus combined models, therefore, we re-port only the results from the models that includemeasures of both parents’ practices as predictors.

To study the implications of racial socializationpractices for youth’s ethnic identity, locus of control,and depression symptoms, the Level 1 model in-cluded offspring’s age and gender, mothers’ and fa-thers’ warmth, mothers’ and fathers’ racialsocialization, the two-way interactions between off-spring age and each parent’s racial socializationpractices, and the two-way interaction between fa-thers’ and mothers’ practices. The Level 2 modelincluded the control variables common to both sib-lings (parental education, parental age, and thedummy code for offspring relatedness). The equa-tions for this model are shown in the appendix, andthe results are shown in Tables 3 (cultural socializa-tion) and 4 (preparation for bias).

Cultural socialization and youth functioning. Begin-ning with youth ethnic identity, the findings revealedthat mothers’ cultural socialization was a positiveand significant correlate (Table 3). No other effectswere significant. Turning to youth’s locus of control,the analyses revealed a significant interaction be-tween mothers’ cultural socialization and fathers’cultural socialization. As Figure 1 shows, and asimple slopes test revealed, when fathers engaged inlow levels of cultural socialization, youth scoredhigher in locus of control if their mothers also en-gaged in low levels of cultural socialization,B 5 � .13, po.01. When fathers engaged in high lev-els of cultural socialization, in contrast, youth locusof control did not vary as a function of maternalpractices. In other words, consistent with ourprediction about interparental congruence, youthreported a higher locus of control when both

Mothers’ and Fathers’ Racial Socialization 1395

mothers and fathers were highly involved or lessinvolved in cultural socialization. Inconsistent withour prediction, youth also reported higher locus ofcontrol when fathers but not mothers engaged in

more cultural socialization. In addition to this inter-action, these analyses revealed that girls, older off-spring, and youth whose fathers had higher levels ofeducation reported a higher locus of control. Finally,with respect to youth’s depression symptoms, wefound that fathers’ cultural socialization was a ne-gative correlate: When fathers reported more cul-tural socialization, youth reported fewer depressionsymptoms.

Preparation for bias and youth functioning. In termsof ethnic identity, the findings revealed that fathers’socialization was a positive, trend-level predictor(Table 4). Findings for youth’s locus of control re-vealed that mothers’ bias preparation was negativelyrelated to youth locus of control, but this main effectwas qualified by an interaction between mothers’socialization and fathers’ socialization. As Figure 2shows, consistent with the findings for cultural so-cialization and with our prediction about inter-parental consistency, youth reported lower locus ofcontrol when their mothers but not their fathers were

Table 3

Coefficients, Standard Errors, and T Ratios for Links Between Mothers’ (M) and Fathers’ (F) Cultural Socialization (Cultural) and Youth Outcomes

Ethnic identity Locus of control Depressiona

B SE t ratio B SE t ratio B SE t ratio

Fixed effects

Intercept 3.26�� .04 77.00 33.94�� .32 105.93 1.04�� .08 13.13

Level 1

Youth age � 0.01 .01 � 0.98 0.37�� .08 4.64 � 0.02 .02 � 0.95

Youth genderb � 0.09 .06 � 1.62 � 0.86� .40 � 2.16 � 0.00 .10 � 0.04

M warmth 0.01 .05 0.11 0.03 .41 0.06 � 0.02 .10 � 0.16

F warmth 0.04 .05 0.90 0.53 .36 1.46 � 0.08 .09 � 0.85

M cultural 0.02�� .01 3.05 � 0.04 .04 � 0.89 0.01 .01 0.65

F cultural � 0.00 .01 � 0.72 0.01 .04 0.23 � 0.02� .01 � 2.36

M � F cultural 0.00 .00 0.01 0.01� .01 2.06 0.00 .00 0.03

Age � M cultural 0.00 .00 1.13 0.01 .01 0.56 0.00 .00 0.56

Age � F cultural 0.00 .00 0.38 � 0.02 .01 � 1.60 0.00 .00 0.51

Level 2

M age � 0.01 .01 � 0.69 0.06 .07 0.91 0.00 .02 0.11

F age � 0.00 .01 � 0.32 � 0.03 .05 � 0.49 0.00 .01 0.27

M education � 0.03 .02 � 1.44 0.10 .14 0.71 0.01 .03 0.43

F education 0.01 .01 0.82 0.37�� .11 3.37 � 0.04w .03 � 1.72

Biological relatednessc � 0.13w .08 � 1.73 0.87 .60 1.46 � 0.13 .14 � 0.89

Variance SD Variance SD Variance SD

Variance components

Residual .24 .49 10.04 3.17 .74 .86

Intercept .00 .02 2.64�� 1.63 .06 .25

Note. aSquare root transformation used to correct skewness.bGirls are the reference group.c0 5 full sibling; 1 5 other (e.g., half sibling, step sibling).wpo.10; �po.05; ��po.01.

30

32

34

36

38

Low High

Mothers' Cultural Socialization

Yo

uth

Lo

cus

of C

on

tro

l

Low Fathers' Cultural SocializationHigh Fathers' Cultural Socialization

Figure 1. Interaction between mothers’ and fathers’ cultural so-cialization predicting youth locus of control.

1396 McHale et al.

highly involved in bias preparation; simple slopestests revealed that the line describing fathers withlow levels of bias preparation was significant,B 5 � .12, po.05, but the line describing fathers with

high bias preparation was not. In this model, fathers’education again emerged as a positive predictor ofyouth locus of control, and boys and younger off-spring reported lower levels of locus of control. Al-though a trend level interaction between youth ageand fathers’ bias preparation emerged in this ana-lysis, follow-up simple slopes tests were not sig-nificant. Finally, in the case of youth’s depressionsymptoms, mothers’ preparation for bias was a posi-tive, trend level predictor: When mothers reportedhigher levels of bias preparation, youth reportedmore depression symptoms. Given that this findingonly reached trend level and does not replicate priorwork, however, this result should be viewed withcaution. Among control variables, fathers’ educationwas the only significant correlate: Youth reportedhigher levels of depression when fathers had feweryears of education.

Summary. Both maternal and paternal warmthwere positively related to their own racial socializationpractices, suggesting that both cultural socialization

Table 4

Coefficients, Standard Errors, and T Ratios for Links Between Mothers’ (M) and Fathers’ (F) Preparation for Bias (Bias) and Youth Outcomes

Ethnic identity Locus of control Depressiona

B SE t ratio B SE t ratio B SE t ratio

Fixed effects

Intercept 3.27�� .04 76.39 34.03�� .32 105.92 1.04�� .08 13.09

Level 1

Youth age � 0.01 .01 � 1.04 0.45�� .08 5.35 � 0.03 .02 � 1.39

Youth genderb � 0.10w .06 � 1.79 � 0.82� .39 � 2.10 0.02 .10 0.16

M warmth � 0.01 .06 � 0.11 0.18 .42 0.44 � 0.04 .10 � 0.40

F warmth 0.01 .05 0.21 0.54 .37 1.46 � 0.09 .09 � 0.98

M bias 0.00 .00 0.66 � 0.07� .03 � 2.07 0.02w .01 1.90

F bias 0.01w .00 1.66 0.02 .03 0.54 � 0.01 .01 � 1.61

M � F bias � 0.00 .00 � 0.74 0.01� .00 2.00 � 0.00 .00 � 1.15

Age � M bias 0.00 .00 0.21 � 0.01 .01 � 1.16 0.00 .00 0.31

Age � F bias 0.00 .00 0.34 � 0.02w .01 � 1.75 0.00 .00 0.76

Level 2

M age � 0.00 .01 � 0.36 0.03 .07 0.38 0.01 .02 0.45

F age � 0.01 .01 � 0.83 � 0.01 .05 � 0.15 � 0.00 .01 � 0.03

M education � 0.02 .02 � 1.04 0.14 .14 0.97 0.01 .03 0.21

F education 0.01 .01 1.02 0.36�� .11 3.25 � 0.05� .03 � 1.98

Biological relatednessc � 0.14w .08 � 1.88 0.78 .61 1.29 � 0.13 .14 � 0.87

Variance SD Variance SD Variance SD

Variance components

Residual 0.24 .49 9.65 3.11 .73 .86

Intercept 0.00 .04 3.02�� 1.74 .08w .28

Note. aSquare root transformation used to correct skewness.bGirl is reference group.c0 5 full sibling; 1 5 other (e.g., half sibling, step sibling).wpo.10; �po.05; ��po.01.

30

32

34

36

38

Low High

Mothers' Preparation for Bias

Yo

uth

Lo

cus

of C

on

tro

l

Low Fathers' Prep for BiasHigh Fathers' Prep for Bias

Figure 2. Interaction between mothers’ and fathers’ preparation forbias predicting youth locus of control.

Mothers’ and Fathers’ Racial Socialization 1397

and bias preparation occur within a larger context ofpositive parenting. We found within-family differencesin parents’ practices, however: Mothers engaged inmore racial socialization with their older as com-pared with their younger offspring, and fathers en-gaged in more racial socialization with their sonsthan with their daughters. These patterns emergedwhen parents’ age and education were controlled.Parents’ racial socialization practices, in turn, werelinked to youth psychosocial functioning. In the caseof youth ethnic identity, mothers’ cultural socializa-tion and fathers’ bias preparation (the latter at trendlevel) were associated with stronger, more matureidentity, and these linkages were not moderated byyouth or family characteristics. Findings for youthdepression symptoms were less consistent: Fathers’cultural socialization was negatively related to youthdepression symptoms, but mothers’ bias preparationwas positively related, albeit at trend level, to youthsymptoms, and there were no significant moderatorsof these linkages. The links between parents’ racialsocialization and youth locus of control, in contrast,were moderated by the spouse’s socialization prac-tices: High levels of maternal socialization in com-bination with low levels of paternal socialization(both bias preparation and cultural socialization)were linked to lower levels of youth locus of control.Across the ‘‘outcome’’ measures, we did not findstrong evidence that parental warmth, youth gender,or youth age moderated the link between racial so-cialization practices and youth functioning.

Discussion

An emerging literature has established racial social-ization as an important dynamic in African Amer-ican families. We built on this foundation to studythe racial socialization practices of mothers and fa-thers in two-parent African American families. Ourdescriptive findings were consistent with a familysystems perspective in documenting within-familydifferences in mothers’ and fathers’ socializationpractices with older and younger sons and daugh-ters: In the case of racial socialization dynamics, fam-ily members may have quite different experienceswithin their shared family environment. A primarygoal of our work was to assess the implications ofracial socialization for youth’s ethnic identity andpsychological adjustment. Working from an ecolo-gical perspective, we studied the family conditionsunder which the linkages between racial socializa-tion and youth functioning were more or lessapparent. Our findings suggest that, in the caseof two-parent families, the implications of racial

socialization for youth’s psychosocial functioningmay be better understood when both parents’practices are taken into account. Below, we reviewour results, considering both how they extend un-derstanding of racial socialization dynamics in fam-ilies and what directions they suggest for future re-search.

Explaining Variation in Parents’ Racial Socialization

Our findings suggested that parents’ racial social-ization may be part of a larger pattern of positiveparent – child relationships: Parental warmth waslinked to cultural socialization and bias preparationfor both mothers and fathers, and these links wereevident when parent education and age and off-spring age and gender were controlled. These find-ings replicate those reported by Caughy et al. (2002),who studied preschoolers’ primary caregivers(mostly mothers) and Frabutt et al. (2002), who stud-ied young adolescents’ mothers. Our study relied onparents’ reports of both their parent – child relation-ship quality and their racial socialization practices.As such, these linkages may have emerged becauseparents see racial socialization as an element of po-sitive parenting. Future studies should include in-formation from other sources (e.g., from youth orobjective reporters) to further establish how racialsocialization is connected to larger parent – offspringrelationship dynamics.

An important question in the racial socializationliterature has been whether and how parents’ prac-tices vary as a function of offspring age and gender.Findings to date are inconsistent, in part due tomethodological limitations of prior research (Hugheset al., in press). Our findings revealed that mothers,but not fathers, engaged in more preparation for biasand cultural socialization with older as comparedwith younger offspring. In contrast, fathers’ but notmothers’ socialization practices varied by offspringgender, with fathers engaging in more racial social-ization with sons than with daughters. Importantly,these findings reflected within-family comparisons ofparents’ socialization of two offspring. Our findingson age differences in maternal practices are con-sistent with prior research in which the most clear-cut age differences have emerged in the domain ofbias preparation (e.g., Hughes & Chen, 1997). Mostprior studies, however, have failed to distinguishmothers’ from fathers’ practices. Given that motherstend to be the primary caregivers, they may be moreresponsive to the developmental needs and cap-abilities of their children than are fathers. Priorfindings regarding differences in parents’ practices

1398 McHale et al.

as a function of offspring gender are less consistentthan findings for age effects, and extant studies havenot differentiated men’s and women’s socializationpractices with sons and daughters. Our findingssuggest that when fathers have the opportunity to doso (because they have both a daughter and a son),they engage in more racial socialization with theirsons. These results are in keeping with prior workshowing that fathers exhibit more gender-differ-entiated socialization of children than mothers(McHale et al., 2003).

Taken together, our findings underscore the im-portance of distinguishing between mothers’ andfathers’ practices. Further, they suggest that, al-though age differences in mothers’ bias preparationhave been evident in prior literature, other offspringage and gender effects may be detected only in largesamples or studies with complex research designssuch as ours (i.e., within-family comparisons). Tobetter understand how child characteristics mayelicit racial socialization practices, future in-vestigators should ask mothers and fathers in anopen-ended way about their reasons for the sociali-zation strategies they use with daughters and sonsand their beliefs about how their practices influenceoffspring at different periods of development.

Implications of Racial Socialization for Youth

A primary goal of our study was to assess theimplications of parents’ practices for youth’s ethnicidentity and psychological well-being. In doing so,we addressed several methodological limitations ofprior work. First, we moved beyond correlationsbetween self-reports to examine parent reports oftheir socialization practices and youth reports of theirethnic identity and well-being. Further, in contrast tothe field’s focus on the socialization practices ofmothers or generic ‘‘parents,’’ we collected in-formation about the practices of mothers and fathersfrom the same families so that we could determinewhether the implications of maternal and paternalsocialization practices were similar or different. Ourwork also addressed an important gap in the litera-ture by studying youth gender and age as potentialmoderators of the links between racial socializationand youth functioning.

A primary contribution of this study was its ex-amination of the family conditions under which ra-cial socialization practices were linked to youthfunctioning. To this end, we studied parentalwarmth and the racial socialization practices of theother parent as potential moderators of socializa-tionFyouth functioning linkages. In the face of

theory about the significance of the emotional cli-mate of parent – offspring relationships (Darling &Steinberg, 1993), we found no evidence that parentalwarmth moderated the links between racial sociali-zation and youth outcomes. Importantly, with bothfactors in the model, racial socialization proved to bea more consistent correlate of youth functioning thanparental warmth. In keeping with scholars’ call fortreating racial socialization as a multidimensionalphenomenon (e.g., Coard & Sellers, 2005), however,we did find evidence that the combination of mothers’and fathers’ socialization practices had implicationsfor youth: In two-parent families, the implications ofracial socialization for youth adjustment, specificallylocus of control, may be better understood whenboth parents’ practices are taken into account.

To study the implications of mothers’ and fathers’practices, we first tested the links between parents’socialization and youth’s ethnic identity. Replicatingprior work (see Hughes et al., in press), we foundthat mothers’ cultural socialization was a positivecorrelate. As noted, mothers were more involved incultural socialization than fathers, and the salience ofmothers’ cultural socialization, coupled with its linksto the ‘‘kin-keeper’’ role, may underlie the sig-nificance of mothers’ practices in this domain. Priorfindings on the links between bias preparation andethnic identity are less consistent than findings forcultural socialization (Hughes et al., in press). Fa-thers have not been an explicit focus of this work,however, and our results showed that fathers’ biaspreparation was linked, at trend level, to youth eth-nic identity. Prior work highlighting males’ morefrequent experiences of discrimination (e.g., Ste-venson et al., 2002) may mean that fathers have morereason than mothers to see bias preparation as im-portant, and they also may have more credibilitywith their children in this domain of parenting.

In contrast to the main effects of racial socializa-tion on youth ethnic identity, links between racialsocialization and youth locus of control were mod-erated by other family dynamics. Specifically, ourfindings suggested that the implications of mothers’practices were contingent on those of fathers: Youthreported lower locus of control in families in whichmothers were highly involved in racial socializationbut fathers were not, and this pattern emerged forboth cultural socialization and bias preparation.These results are consistent with an emerging bodyof work on co-parenting that highlights the im-portance of interparental consistency for youth well-being (e.g., Feinberg, 2003), and with family systems’notions about linkages between dyads within a fam-ily (Whitchurch & Constantine, 1993). In contrast

Mothers’ and Fathers’ Racial Socialization 1399

to the negative implications of the high mother/lowfather socialization pattern, when fathers werehighly involved in racial socialization but motherswere not, youth reported high levels of locus ofcontrol. Because mothers report higher levels of ra-cial socialization, on average, than fathers do, youthmay perceive few differences between mothers andfathers who display the latter pattern. It is also pos-sible that the high mother – low father pattern reflectsproblematic co-parenting dynamics, such as motherscompensating for fathers’ lack of involvement orinterparental disagreement about how to socializeoffspring. In future studies, parents should be askeddirectly about co-parenting in the domain of racialsocialization.

With respect to youth depression symptoms, theresults revealed that fathers who engaged in morecultural socialization had children who were lessdepressed. As was the case with locus of control,however, mothers’ bias preparation was negativelyrelated (at trend level) to youth adjustment. Al-though cultural socialization has been shown to be aprotective factor, some prior work indicates that ra-cial socialization that highlights barriers to oppor-tunity may be linked to poorer youth well-being(e.g., Constantine & Blackmon, 2002; Ogbu, 1974).The different valences of mothers’ versus fathers’socialization for youth depression symptoms mayalso arise because mothers tend to engage in moreracial socialization, on average, than fathers, andthus, mothers who score high on this dimension maybe overly involved. The fact that a similar patternemerged in the case of locus of control (i.e., mothers’racial socialization was negatively related and fa-thers’ racial socialization was positively related toyouth adjustment) also may mean that a more com-plex process was in play: Youth in this samplewere generally well adjusted and limited variabilityin depression symptoms may have precludedour detecting the mother socialization by fathersocialization interactions that emerged for locus ofcontrol.

Taken together, our findings imply that genderedpatterns in parents’ racial socialization and theirimplications merit future research scrutiny. A bodyof work describes differences in mothers’ and fa-thers’ parental roles (McHale et al., 2003). Givendifferences in their individual experiences and or-ientations to parenting, it would be surprising if menand women approached the tasks of racial sociali-zation in precisely the same way. Qualitative datafrom youth on their experiences with mothers andfathers may reveal more mother – father differencesthan are apparent when rating scales are used.

Limitations and Directions for Study

Although our findings contribute to the literatureon racial socialization, our study is not withoutlimitations. First, our study goals directed attentionto a particular type of family, that is, families thatincluded a mother and father and at least two chil-dren in middle childhood and adolescence. Had wetargeted a randomly selected and representativesample of African American families and studied theracial socialization efforts of the caregivers in thosefamilies (e.g., mothers and grandmothers or otheradults in the home), a different pattern of findingsmay have emerged.

Our lack of detail on family members’ ethnicity isanother limitation of this study. Our interest in familydynamics directed our focus on constructs such asracial socialization rather than status variables such asethnicity. Variability in ethnicity among individualswho self-identify as Black or African American isconsiderable, however. In future work, researchersshould examine whether ethnicity moderates thelinks between cultural dynamics like racial socializa-tion and youth and family relationship outcomes.

A final limitation is the cross-sectional and cor-relational nature of our research design. The existingliterature led us to frame our research to study theimplications of racial socialization for youth, but ourdesign does not allow for such causal inferences.Longitudinal research is essential to shed light on thedirection of effect of these dynamics, that is, whetherhigher levels of racial socialization give rise to posi-tive youth functioning or the other way around. In-tervention studies, in which parents are trained inthe use of effective racial socialization and the im-portance of co-parenting is highlighted, are anotherdirection for research that may have both practicaland theoretical significance.

In the face of its limitations, this study provides animportant contribution to the literature on racial so-cialization by setting this dynamic in its larger familycontext. Our findings add to a small but importantliterature in documenting the characteristics andexperiences of family members that give rise to racialsocialization practices and how these practices, inturn, have implications for youth and family re-lationships. This research also speaks to the literatureon family socialization influences, more generally, inits investigation of normative dynamics in AfricanAmerican families. Finally, our examination of theracial socialization practices of two parents with twooffspring provides insights into how families operateas systems in the process of child and adolescentsocialization.

1400 McHale et al.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: testingand interpreting interactions. Thousand Oaks, CA: Sage.

Bowman, P. J., & Howard, D. (1985). Race-related sociali-zation, motivation, and academic achievement: A studyof Black youth in three-generation families. Journal of theAmerican Academy of Child Psychiatry, 24, 134 – 141.

Boykin, A. W., & Toms, F. D. (1985). Black child socializa-tion: A conceptual framework. In H. P. McAdoo & J. L.McAdoo (Eds.), Black children: Social, educational, andparental environments (pp. 33 – 52). Newbury Park, CA:Sage.

Bronfenbrenner, U., & Crouter, A. C. (1983). The evolutionof environmental models in developmental research. InP. Mussen (Ed.), Handbook of child psychology (Vol. 1, pp.358 – 414). New York: Wiley.

Caughy, M., O’Campo, P. J., Randolph, S. M., & Nickerson,K. (2002). The influence of racial socialization practiceson the cognitive and behavioral competence ofAfrican American preschoolers. Child Development, 73,1611 – 1625.

Coard, S. I., & Sellers, R. M. (2005). African American fa-milies as contexts for racial socialization. In V. C.McLoyd, N. E. Hill, & K. A. Dodge (Eds.), AfricanAmerican family life (pp. 264 – 284). New York: Guilford.

Constantine, M. G., & Blackmon, S. M. (2002). Black ado-lescents’ racial socialization experiences their relationsto home, school, and peer self-esteem. Journal of BlackStudies, 32, 322 – 335.

Darling, N., & Steinberg, L. (1993). Parenting style ascontext: An integrative model. Psychological Bulletin, 113,487 – 496.

Fatimilehin, I. A. (1999). Of jewel heritage: Racial sociali-zation and racial identity attitudes amongst adolescentsof mixed African-Caribbean/White parentage. Journal ofAdolescence, 22, 308 – 318.

Feinberg, M. E. (2003). The internal structure and ecologi-cal context of coparenting: A framework for researchand intervention. Parenting: Science and Practice, 3,95 – 131.

Fischer, A. R., & Shaw, C. M. (1999). African Americans’mental health and perceptions of racist discrimination:The moderating effects of racial socialization experi-ences and self esteem. Journal of Counseling Psychology,46, 395 – 407.

Frabutt, J. M., Walker, A. M., & MacKinnon-Lewis, C.(2002). Racial socialization messages and the quality ofmother/child interactions in African American families.Journal of Early Adolescence, 22, 200 – 217.

Hagestad, G. O. (1986). The family: Women and grand-parents as kinkeepers. In A. Pifer & L. Bronte (Eds.), Ouraging society (pp. 141 – 160). New York: Norton.

Hill, N. E., Murry, V. M., & Anderson, V. D. (2005). Socio-cultural contexts of African American families. In V. C.McLoyd, N. E. Hill, & K. A. Dodge (Eds.), AfricanAmerican family life (pp. 21 – 44). New York: Guilford.

Hughes, D., & Chen, L. (1997). When and what parents tellchildren about race: An examination of race-related

socialization among African American families. AppliedDevelopmental Science, 1, 200 – 214.

Hughes, D., & Johnson, D. (2001). Correlates in children’sexperiences of parents’ racial socialization behaviors.Journal of Marriage and Family, 63, 981 – 995.

Hughes, D., Rodriguez, J., Smith, E. P., Johnson, D., &Stevenson, H. C. (in press). Parents’ ethnic/racial so-cialization practices: A review of research and directionsfor future study. Developmental Psychology.

Jacobs, J. E., Lanza, S., Osgood, D. W., Eccles, J. S., & Wig-field, A. (2002). Changes in children’s self-competence andvalues: Gender and domain differences across gradesone through twelve. Child Development, 73, 509 – 527.

Johnson, D. (2001). Parental characteristics, racial stressand racial socialization processes as preditors of racialcoping in middle childhood. In A. M. Neal-Barnett, J. M.Contreras, & K. A. Kerns (Eds.), Forging links: AfricanAmerican children clinical developmental perspectives(pp. 57 – 74). Westport, CT: Praeger.

Johnson, D., Brown, D., Wayumba, E., Choi, Y. K., &Hunter, A. G. (2005). Correlates of racial socializationamong African American fathers. Paper presented at thebiennial meeting of the Society for Research in ChildDevelopment, Atlanta, GA.

Kovacs, M. (1981). The children’s depression inventory: A self-rated depression scale for school-aged youngsters. Un-published manuscript. Pittsburgh, PA: University ofPittsburgh School of Medicine.

Margolin, G., Oliver, P. H., & Medina, A. M. (2001). Con-ceptual issues in understanding the relation betweeninterparental conflict and child adjustment. In J. H.Grych & F. D. Fincham (Eds.), Interparental conflict andchild development (pp. 9 – 38). Cambridge, UK: Cam-bridge University Press.

Marshall, S. (1995). Ethnic socialization of African Amer-ican children: Implications for parenting, identity de-velopment, and academic achievement. Journal of Youthand Adolescence, 24, 377 – 396.

McHale, S. M., Crouter, A. C., & Whiteman, S. D. (2003).The family contexts of gender development in child-hood and adolescence. Social Development, 12, 125 – 148.

McLoyd, V. C. (1998). Changing demographics in theAmerican population: Implications for research onminority children and adolescents. In V. C. McLoyd & L.Steinberg (Eds.), Studying minority adolescents: Con-ceptual, methodological, and theoretical issues (pp. 3 – 28).Mahwah, NJ: Lawrence Erlbaum.

Miller, D. B. (1999). Racial socialization and racial identity:Can they promote resiliency for African Americanadolescents? Adolescence, 34, 494 – 501.

Murry, V. M., & Brody, G. H. (2002). Parenting and racialsocialization processes in rural, single-parents families:An examination of self-worth and self-regulatory behav-ior of Black children. In H. P. McAdoo (Ed.), Black chil-dren: Social, environmental, and parental environments (2nded., pp. 97 – 115). Thousand Oaks, CA: Sage.

Nowicki, S. Jr., & Strickland, B. R. (1973). A locus of controlscale for children. Journal of Consulting and Clinical Psy-chology, 40, 148 – 154.

Mothers’ and Fathers’ Racial Socialization 1401

Ogbu, J. U. (1974). The next generation: An ethnography ofeducation, in an urban neighborhood. New York: Academic.

Parke, R., & Buriel, R. (1998). Socialization in thefamily: Ethnic and ecological perspectives. In W. Damon& N. Eisenberg (Eds.), Handbook of child psychology:Social, emotional, and personality development (Vol. 4,pp. 463 – 552). New York: Wiley.

Peters, M. F. (1985). Ethnic socialization of young Blackchildren. In H. P. McAdoo & J. L. McAdoo (Eds.), Blackchildren: Social, educational, and parental environments (pp.159 – 173). Newbury Park, CA: Sage.

Phinney, J. S. (1992). The multigroup ethnic identity mea-sure: A new scale for use with diverse groups. Journal ofAdolescent Research, 7, 156 – 176.

Phinney, J. S., & Chavira, V. (1995). Parental ethnic socia-lization and adolescent coping with problems, related toethnicity. Journal of Research on Adolescence, 5, 31 – 54.

Quintana, S. M. (1998). Children’s developmental under-standing of ethnicity and race. Applied and PreventivePsychology, 7, 27 – 45.

Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., & Congdon,R. (2003). HLM 5.05: Hierarchical linear and nonlinearmodeling. Chicago: Scientific Software International.

Rumbaut, R. (1994). The crucible within: Ethnic identity,self esteem, and segmented assimilation among childrenof immigrants. International Migration Review, 28,748 – 794.

Sanders Thompson, V. L. (1994). Socialization to race andits relationship to racial identification among AfricanAmericans. Journal of Black Psychology, 20, 175 – 188.

Sanders Thompson, V. L. (1999). Variables affecting racial-identity salience among African Americans. Journal ofSocial Psychology, 139, 748 – 761.

Schwarz, J. C., Barton-Henry, M. L., & Pruzinsky, T. (1985).Assessing child-rearing behavior: A comparison of rat-ings made by mother, father, child, and sibling on theCRPBI. Child Development, 56, 462 – 479.

Scott, L. D. (2003). The relation of racial identity and racialsocialization to coping with discrimination amongAfrican American adolescents. Journal of Black Studies,33, 520 – 538.

Spencer, M. B. (1983). Children’s cultural values and par-ental child rearing strategies. Developmental Review, 3,351 – 370.

Stevenson, H. C. (1995). Relationships of adolescent per-ceptions of racial socialization to racial identity. Journalof Black Psychology, 21, 49 – 70.

Stevenson, H. C., Camercon, R., Herrero-Taylor, T., & Da-vis, G. Y. (2002). Development of the teenager experienceof racial socialization scale: Correlates of race-relatedsocialization frequency from the perspective of Blackyouth. Journal of Black Psychology, 28, 84 – 106.

Stevenson, H. C., Reed, J., Bodison, P., & Bishop, A. (1997).Racism stress management: Racial socialization beliefsand the experience of depression and anger in AfricanAmerican youth. Youth and Society, 29, 197 – 222.

Thomas, A. J., & Speight, S. L. (1999). Racial identity andracial socialization attitudes of African American par-ents. Journal of Black Psychology, 25, 152 – 170.

Thornton, M. C., Chatters, L. M., Taylor, J. R., & Allen, W.R. (1990). Sociodemographic and environmental corre-lates of racial socialization by Black parents. Child De-velopment, 61, 401 – 409.

U.S. Census Bureau. (2001). 2000 Census Population andHousing. Retrieved from http://www.census.gov/Press-Release/www/2001/sumfile.htm

Whitchurch, G. G., & Constantine, L. L. (1993). Systemstheory. In P. G. Boss, W. J. Doherty, R. LaRossa, W. R.Schumm, & S. K. Steinmetz (Eds.), Sourcebook of familytheories and methods: A contextual approach (pp. 325 – 355).New York: Plenum.

Appendix

Links Between Parental Warmth, Youth Age, and Genderand Racial Socialization

Level 1:

yij ¼ b0j þ b1jðyouth ageÞ þ b2jðyouth genderÞþ b3jðmothers0 warmthÞ þ b4jðfathers warmthÞ þ eij

Level 2:

b0 ¼g00 þ g01ðyouth mean ageÞþ g02ðyouth mean genderÞ þ g03ðmothers0 ageÞþ g04ðfathers0 ageÞþ g05ðmothers educationÞ þ g06ðfathers0 educationÞþ g07ðsibling relatednessÞ þ u0j

b1j ¼ g10; b2j ¼ g20; b3j ¼ g30; b4j ¼ g40

Parents’ Racial Socialization and Youth PsychosocialFunctioning

Level 1:

yij ¼ b0j þ b1jðyouth ageÞ þ b2jðyouth genderÞþ b3jðmothers warmthÞ þ b

4jðfathers warmthÞ

þ b5jðmothers0 racial socializationÞþ b6jðfathers racial socializationÞþ b7jðmothers racial socialization

� fathers0 racial socializationÞþ b8jðyouth age �mothers0 racial socializationÞþ b9jðyouth age � fathers0 racial socializationÞ þ eij

Level 2:

b0j ¼g00 þ g01ðmothers0 ageÞþ g02ðfathers0 ageÞ þ g03ðmothers0 educationÞþ g04ðfathers0 educationÞ þ g05ðsibling relatednessÞ þ u0j

b1j ¼ g10; b2j ¼ g20; b3j ¼ g30; b4j ¼ g40; b5j ¼ g50; b6j ¼ g6; b7j

¼ g70; b8j ¼ g80; b9j ¼ g90

1402 McHale et al.