Beyond Ethnicity: Applying Helms's White Racial Identity Development Model Among White Youth
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
JOURNAL OF RESEARCH ON ADOLESCENCE, ***(*), 1–20 Beyond Ethnicity: Applying Helms’s White Racial Identity Development Model Among White Youth Ursula Moffitt , and Leoandra Onnie Kara R. H. Dastrup Rogers University of Washington Northwestern University Research on racial identity among Youth of Color has expanded considerably in recent years, but a parallel examina- tion of racial identity among white youth has not occurred, reiterating whiteness as normative. We applied Janet Helms’s White Racial Identity Development (WRID) model among white U.S. youth (8–14 years old) to address this research gap. WRID centers racism and white supremacy, offering a framework to analyze white racial identity in the context of systemic inequity. Using longitudinal, qualitative analysis, we found age-related change over time, with some evidence of increasing resistance to racism. There was high participant variability, however, indicating that socio- cognitive abilities alone cannot predict anti-racist white identity development. We discuss implications for racial iden- tity research and social justice-orientated developmental science. Key words: White racial identity – racial identity development – whiteness – white youth – qualitative research – early ado- lescence Following the murders of George Floyd, Breonna identity among white youth has not occurred, reit- Taylor, and numerous other unarmed Black people erating whiteness as normative and invisible at the hands of U.S. police, the ensuing public out- (Helms, 2007; Rogers, 2019). cry has included demands for justice, and for the Silence with regard to whiteness in developmen- naming and dismantling of white supremacy. tal science is not neutral. U.S. society has been White supremacy refers to long-standing beliefs shaped by centuries of policies and norms enshrin- and practices that situate white1 people as norma- ing whiteness with advantage (e.g., Feagin & tive, while Black, Indigenous, and other People of Ducey, 2019), and the field of psychology has a Color (BIPOC) are minoritized and marginalized long history of privileging white scholars, partici- (Sue, 2006). Breaking down white supremacist pants, and perspectives (Dupree & Kraus, 2021; inequities requires myriad changes, including an Guthrie, 1998; Roberts, Bareket-Shavit, Dollins, interrogation of whiteness and white racial identity Goldie, & Mortenson, 2020). Examining how racial (e.g., Frankenberg, 1988; Winant, 2004). Although privilege and power shape the development of research on racial and ethnic identity development white youth acknowledges that race shapes the tra- among Youth of Color has seen a substantial and jectories of all humans, not just those who experi- overdue expansion in past decades (Uma~ na-Taylor ence racial oppression and marginalization (e.g., et al., 2014; Williams et al., 2020; Yip, Douglass, & Rogers, Moffitt, & Foo, 2021; Spencer, 2017). Such Sellers, 2014), a parallel examination of racial an approach disrupts the normativity of whiteness in the research context. Furthermore, examining racial identity among white youth makes space to We have no known conflict of interest to disclose. understand whether, how, and when they may also Funding for this manuscript was supported by a postdoctoral resist white supremacy. fellowship awarded to Dr. Moffitt from the National Science Foundation (NSF) SBE (#2005052), as well as by postdoctoral fel- The current study takes one step in this direc- lowships awarded to Dr. Rogers from the National Science tion. Drawing on longitudinal interview data with Foundation (NSF) SBE (#1303753) and the Spencer Foundation/ white children and adolescents (age 8–14 years), National Academy of Education (NAEd) Postdoctoral Fellow- we conduct a data-driven, hybrid inductive–deduc- ship. tive thematic analysis with the aim of adapting Requests for reprints should be sent to Ursula Moffitt, Depart- ment of Psychology, Northwestern University, 2029 Sheridan Janet Helms’s (1984, 1990, 2020) White Racial Iden- Road, Swift Hall, Evanston, IL 60208. E-mail: moffitt@northwest- tity Development (WRID) model. We engage this ern.edu model for two reasons: First, it focuses on racial In this manuscript, we intentionally write “white” in lower- identity specifically, meaning it centers whiteness case to push back on the notion that whiteness constitutes a cul- tural group, as well as in recognition of the increasing visibility of overt white supremacist hate groups, among whom the capi- © 2021 Society for Research on Adolescence talization of “white” has long been common practice. DOI: 10.1111/jora.12645
2 MOFFITT, ROGERS, AND DASTRUP in the U.S. context and not ethnic heritage. In often measured using the same tools for white doing so, it locates an anti-racist identity, rather youth and Youth of Color. The “universal” scales, than an “achieved” identity, as most adaptive; and such as the Multi-group Ethnic Identity Measure second, its structure allows for an investigation of (MEIM; Phinney, 1992; Roberts et al., 1999) and individual change over time. Although neither the Ethnic Identity Scale (EIS; Uma~ na-Taylor, Yazed- WRID model nor the current research assumes that jian, & Bamaca-Gomez, 2004), do not mention race, white racial identity develops as a by-product of power, or oppression (Helms & Talleyrand, 1997). maturation, the framework recognizes that personal Instead, they draw on the seminal work of Erikson and societal experiences can compel an individual (1968) and Marcia (1966) to assess exploration and toward shifts in racial identity—either progression commitment—“feeling good, happy, and proud” or regression. By analyzing change in the racial (Rivas-Drake et al., 2014, p. 77)—in relation to identity of white youth, we can see the role age- one’s ethnic group. Perhaps not surprisingly, white related socio-cognitive development may play as youth consistently score lower than Youth of Color children gain greater knowledge of racial inequity on these universal ERI measures (Phinney, 1992; while being socialized into a society dominated by Rogers, Kiang, et al., 2021). This could indicate that color-blind norms (e.g., Pauker, Apfelbaum, & Spit- ERI is of minimal (or relatively less) importance to zer, 2015). To situate our analysis, we review litera- white youth. It may also be that universal ERI mea- ture on ethnic and racial identity development and sures do not adequately capture the ways in which discuss the theoretical underpinnings of the WRID white youth identify with and make meaning model and related research. about race and their own whiteness. What does it mean to feel “committed” to being white in the context of a white supremacist society? Mixed- ETHNICITY, RACE, AND IDENTITY methods research has found that scoring high on DEVELOPMENT universal ERI measures can indicate explicitly In the United States, ethnicity is typically defined racist beliefs, explicitly anti-racist beliefs, or a by shared ancestry, customs, and traditions, strong identification with ethnic heritage (Gross- whereas race is defined primarily by phenotypical man & Charmaraman, 2009; Hughey, 2010). Unfor- traits, including skin color, hair texture, and facial tunately, universal ERI measures alone cannot features (Cokley, 2007). Yet, ethnicity is also racial- differentiate between these divergent interpreta- ized, with individuals categorized as white, regard- tions, making it very difficult to draw conclusions less of ethnic heritage, experiencing systemic (or comparisons) about ethnic and racial identity privileging in terms of power, status, and wealth among white youth. (e.g., Winant, 2004). Conversely, African and Carib- Perhaps more critically, the universal approach bean immigrants in the United States, for example, to ERI makes it difficult to assess an anti-racist- are racialized and experience discrimination “as if” oriented white identity (e.g., Helms, 2007). Under- they are Black, regardless of heritage (Waters, standing how white youth may come to see the 1999). Recognizing the interrelated nature of ethnic- inequitable ways in which power is afforded to ity and race, developmental psychologists put forth their racial group and the harm white privilege can the meta-construct of ethnic–racial identity (ERI) to cause necessitates a race-specific approach. This include racial and ethnic markers and experiences argument is not new. Prior to the advent of the (Uma~ na-Taylor et al., 2014). ERI refers to attitudes MEIM, racial identity theorists including Cross and beliefs about one’s membership in an ethnic or (1991), Helms (1984, 1990), and Sellers and col- racial group, as well as the process of exploring leagues (1998) developed group-specific racial iden- one’s heritage and achieving a committed sense of tity models based on this premise. These self in relation to ethnicity and race (Williams foundational frameworks pushed back on scientific et al., 2020). Among Youth of Color, the develop- racism in psychology, which often situated race as ment of ERI is a normative milestone (Yip et al., a biological construct used as an explanatory vari- 2014). Strong ERI can buffer the deleterious effects able to denigrate Black individuals (Yee, Fairchild, of racism and has been linked to adaptive psy- Weizmann, & Wyatt, 1993). In tandem with a chosocial outcomes across minoritized ethnic and model of Black racial identity, Helms (1984, 1990) racial groups (Rivas-Drake et al., 2014). put forth a theoretical model of white racial iden- Although the racialization of ethnicity occurs tity, naming racism as the system structuring the across groups, its meaning is distinct, as whiteness racial identity development process—for Black peo- confers a privileged status. Nonetheless, ERI is ple and white people.
WHITE RACIAL IDENTITY 3 Helms’ Model of White Racial Identity adopt assimilationist views, recognizing the exis- Development tence of racism and aiming to “help save” BIPOC individuals (Pseudo-independent), then begin The WRID model, despite early psychometric cri- engaging in a more active exploration of racism as tique of its measurement tool (Behrens & Rowe, systemic (Immersion/Emersion), and finally con- 1997; Rowe, Bennett, & Atkinson, 1994), remains front racism as a part of a multilayered system of highly relevant to research on white identity and intersectional oppression, while also feeling com- provides an alternative to relying on universal fortable within their white identity (Autonomy). models that conflate ethnicity and race. Centuries Importantly, Helms (1990, 2020) contends that of racist policies, research, and societal expectations because of the structure of the racial hierarchy and have situated white people as more human than normativity of whiteness, a white person may BIPOC individuals (Kteily & Bruneau, 2017), mak- remain in a given schema indefinitely, unless ing whiteness the baseline marker for humanity prompted by societal events and/or interpersonal (Feagin & Ducey, 2019). For this reason, Helms interactions to either progress forward or regress to contends that viewing white supremacy as “nor- earlier schemas. Thus, progression from Phase 1 to mal” while denying the importance of race is itself Phase 2 is not anticipated simply as a function of a racial identity. In fact, such color-blind ideology age. By examining among whom and under what predominates among many white adults (Bonilla- circumstances movement across schemas does Silva, 2013), who benefit from racism and do little occur, we can gain knowledge of individual and to question or resist the structures making racial contextual factors that prompt racial identity devel- privilege possible. It is therefore not a question of opment among white youth. if a given individual has a racial identity, but rather what that racial identity looks like and how Racial Identity Development Among White Youth it fits within our racially stratified society. Initially a stage model, Helms (2007, 2017, 2020) Existing research using the WRID model has shifted from stage-based language in later publica- focused on emerging adult and adult samples. For tions, largely in response to critique regarding the example, one longitudinal study with white coun- notion of linear development, as well as questions seling graduate students found that, overall, stu- about whether individuals are really in a single dents moved from Phase 1 to Phase 2 schemas stage at a time (e.g., Fasching-Varner, 2014; after taking part in a semester-long course on Spanierman & Soble, 2010). Helms (2020) now uses power, privilege, and oppression (Dass-Brailsford, the language of “schemas,” lenses through which a 2007). Early cross-sectional work found that stu- person views race and their racialized experiences, dents older than 20 years of age were more likely which are neither mutually exclusive nor rigidly to endorse Phase 2 schemas than younger college linear. The WRID model is made up of two phases, students (Pope-Davis & Ottavi, 1994), suggesting each of which contains three schemas (Figure 1) that WRID may occur without specific intervention. characterized by strategies, perspectives, and Other studies have found great variation across behaviors that either accommodate to norms of white college-aged participants, with higher racial inequity (Phase 1) or resist the socialization endorsement of Phase 2 schemas linked to out- into and participation in the racist status quo comes including less fear in response to racially (Phase 2). Though the schemas are not mutually charged situations (Siegel & Carter, 2014) and exclusive, the assumption remains that at a given lower support for self-segregation from People of time, a person is primarily within one phase, if not Color (Clauss-Ehlers & Carter, 2005). one schema, as the identity work required to To our knowledge, no scholars have applied this engage the anti-racist Phase 2 schemas necessitates theoretical model of racial identity development some degree of relinquishing the beliefs and behav- prior to emerging adulthood. A primary concern iors characteristic of Phase 1. when adapting such a model lies in its develop- In the three Phase 1 schemas, white individuals mental appropriateness: Do young people, during shift from willing or unintentional obliviousness middle childhood and early adolescence, have the about the role of race and the meaning of white- socio-cognitive skills needed to comprehend the ness (Contact) to a confused state of grappling with complex implications of racism and their own the recognition of whiteness (Disintegration) to a racial identities? By middle childhood (8–11 years), conscious embracing of the inequitable status quo children generally identify with a racial group, (Reintegration). In Phase 2, the individual may first while also showing racial constancy, or an
4 MOFFITT, ROGERS, AND DASTRUP Phase 2 Phase 1 Challenging the Status Internalized Racism Quo Pseudo- Immersion / Contact Disintegraon Reintegraon Autonomy independence Emersion FIGURE 1 Overview of WRID phases and schemas. Note. This flowchart is adapted from the theoretical WRID model as described by Helms (2020). WRID, White racial identity development. awareness of race as a fixed characteristic (Byrd, color-blind racial ideology upholds white supre- 2012). During this period, children tend to think in macy, and the intergenerational socialization of this concrete terms about race, gaining knowledge from norm perpetuates stagnant white racial identity peers, parents, teachers, and media about which development (Helms, 2020). groups are valued and devalued in society and Despite the pervasiveness of color-blindness, we where they fit in this structure (Williams et al., know that white youth are not ignorant about race 2020). and racism. Qualitative work has shown that some Though research on racial meaning making in white children and adolescents endorse color-blind middle childhood is limited, there is evidence that ideology in one breath while pointing out inequita- by early adolescence, increased life experience and ble treatment in the next (Hagerman, 2015; Rogers, cognitive development allow for racial perspective Moffitt, & Foo, 2021; Way et al., 2013). Interest- taking (Quintana, 1998, 2008), and a critical engage- ingly, a recent study with white adolescents in a ment with racial stereotypes and expectations “multicultural antibias course” (Thomann & Suye- (Way, Hern andez, Rogers, & Hughes, 2013). Dur- moto, 2018, p. 749) found that a deepening knowl- ing this period, some youth move beyond thinking edge of structural racism was coupled with greater about race in neutral terms, recognizing that indi- introspection, empathy, and reflections on white- vidual experiences are situated in a societal context ness, similar to research with college students of racial inequity (Quintana, 2008). However, Quin- (Dass-Brailsford, 2007). Yet, other research with tana (2008) argued that white youth show lower middle- to upper-class white high school seniors levels of racial perspective taking than same age found an overall decrease in support for educa- Youth of Color, in part because they are not tional equity following a semester-long course on prompted by experience to engage with race- social justice, with many youth viewing privilege relevant issues in the same way. That is, white as zero sum (Seiders, 2008). These mixed findings privilege cocoons white children from the kinds of underscore the need for greater research on racial race encounters, discrimination, and racism that identity development among white youth. can prompt racial identity development, even prior to adolescence. The Current Study Research on white racial socialization echoes The current analysis takes one step toward filling this, as most parents of white children either say this gap in the racial identity literature and nothing about race or express color-blind ideology responds to a call for more qualitative research (Chae, Rogers, & Yip, 2020; Hagerman, 2018; Perry, using Helms’s theoretical model (Spanierman & Skinner, & Abaied, 2019; Underhill, 2018). As white Soble, 2010). Specifically, we are interested in the youth move from childhood to adolescence, they extent to which the WRID model (Helms, 1990, also become less willing to talk about race (Apfel- 2020) is meaningful for understanding the progres- baum, Pauker, Ambady, Sommers, & Norton, sion of white identity during middle childhood 2008), and among youth who reported that their and early adolescence. To investigate this, we draw parents and teachers avoided conversations about on longitudinal interview data with white youth, race, their own willingness to discuss race was situating our analyses around the following lower than peers who received race-conscious research questions: socialization (Pauker et al., 2015). Color-blind mes- sages are not neutral (Bonilla-Silva, 2013). Color- (1) To what extent does the meaning making about blindness has been linked to increased racial biases race and white racial identity these youth and decreased recognition of overt racism (Apfel- engage in fit within the WRID schemas and baum, Norton, & Sommers, 2012). In other words, phases?
WHITE RACIAL IDENTITY 5 (2) What is the distribution of these youths’ coded as white at T1 and multiracial at T2 and were statements across WRID schemas and phases at therefore asked a different set of questions, pre- Time 1 and Time 2? Are there age-related pat- cluding longitudinal analyses; one who identified terns in this distribution? solely as Albanian and spoke about family and eth- Although we do not have specific hypotheses, nicity without directly answering the interview based on previous qualitative, race-focused questions; and a final child for whom the recording research with white youth (Rogers, Moffitt, & Foo device malfunctioned, leading to unusable data. 2021) , and taking into account the socio-cognitive The final analytic sample included 37 white youth: abilities of youth in middle childhood to early ado- 11 in middle childhood at T1 (Mage = 9.00, lescence (Quintana, 2008), we anticipate our partici- SD = 1.73) and 26 in early adolescence at T1 pants’ interviews will show evidence of greater (Mage = 11.62, SD = 0.50). There were 16 girls and endorsement of the Phase 1 schemas than Phase 2. 21 boys included in our final sample; each gender We also anticipate, however, that this model will was represented evenly in each age group. Partici- offer a meaningful organizational structure for pants received a university-themed pencil and $5 making sense of any variation we do find, includ- gift card at both T1 and T2. ing across time points and age groups. Thus, we do not intend to offer a proscriptive model map- Semi-Structured Interview Protocol ping out stages of white racial identity develop- ment. Instead, we investigate the ways in which The interviews were conducted in a private room our participants are engaging with their racial on school property and varied in length from 14 to identities, drawing on the interviews themselves to 81 min (M = 39.43, SD = 14.30). All interviews explore trends in the data. were semi-structured and explored participants’ meaning making on identity and subjective experi- ences (Rogers & Meltzoff, 2017; Rogers, Moffitt, & METHOD Jones 2021). Each participant completed a card- The data used in the current study were drawn sorting task, selecting cards that they felt applied from a larger research project examining self- to them with, “words we use to describe ourselves perceptions and social identities across childhood or other people” (including Asian/Black/His- and adolescence (see Rogers & Meltzoff, 2017 panic/white, boy/girl, son/daughter, student, ath- 2021). Participants were recruited from two public lete). All participants in our sample selected the elementary schools and one public middle school “white” card at both T1 and T2 without being in an urban, predominantly low-income area in the prompted by the interviewer, reflecting racial western United States. These schools were selected awareness and identification (Byrd, 2012; Williams due to their diverse student bodies, with a maxi- et al., 2020). mum of 40% of any one racial group. We could not Participants were asked to rank the perceived gather individual information on participants’ importance of each identity category for them per- socioeconomic status, though at least 70% of stu- sonally and then discuss the importance of these dents at each school received free or reduced-price identities and related experiences. In the current lunch. study, we focus on the race-related section of the interview (see Appendix S1 for a full list of race- related interview questions). This semi-structured Participants and Procedure interview protocol was expanded at T2 to include a Students from grades 2 through 6 were invited to question about intersectional identities: “How participate via information letters and parental con- important is being a white boy to you?” followed sent forms. Students with parental consent were by two hypothetical questions: “How might things interviewed at Time 1 (T1) in the fall of 2013 to be different if you were a Black girl?” and “How spring of 2014 and at Time 2 (T2) in the spring of might things be different if you were a Black boy?” 2016. A total of 242 youth took part in these inter- views. There was notable attrition by T2, largely CODING AND ANALYSIS due to school transfers and study approval guideli- nes preventing the tracking of students to different The recorded interviews were professionally tran- schools. A total of 109 youth participated at both scribed and checked for errors by trained research time points. Of those, 41 self-identified as white. assistants. Coding and analysis was conducted in We excluded four participants: two who identified NVivo–Qualtrics Software (version 12 for Mac).
6 MOFFITT, ROGERS, AND DASTRUP The first and third authors divided the 74 inter- the 23 data-driven codes, which were interpreted view transcripts (37 at T1 and T2), reading the race slightly differently in multiple occurrences, bring- sections and using memoing as a first step in the ing down the kappa score despite high coding over- analytic process (Birks, Chapman, & Francis, 2008). lap otherwise. The two codes accounting for the We took notes on what stood out about each inter- most discrepancies were Color-blindness and White- view, as well as on (a) conceptualizations of racial ness as normal, both of which are situated in the identity; (b) mentions of inequality, racism, or priv- Contact schema. The conceptual boundaries ilege; and (c) evidence of change across T1 and T2. between these codes were initially unclear, as both Based on our memos, we discussed the ways the relate to denying the importance of race. After data did and did not align with Helms’s (1990, returning to Helms’s (2020) theoretical model, 2020) theoretical model. The first author then con- descriptions were updated in the codebook and we ducted open coding (Salda~ na, 2015), using the decided to maintain each code to capture the great- memos to aid in creating data-driven codes rele- est nuance possible. Consensus was reached vant to the WRID process. This resulted in 23 through discussion. Following a recoding of these codes. Using these codes, all three authors then participants, a kappa score of .74 was achieved, worked to draft a full codebook based on compar- along with 98% overall coding agreement. The vast ison with the data and theory, nesting the data- majority of remaining discrepancies were due to driven codes within the six schemas of the WRID differing lengths of coded segments, meaning both model. The codes thus acted as subthemes, with coders situated a given statement in the same sub- the six schemas of the WRID model as themes (see theme, but one included a longer section of text Table 1). As such, the codes themselves were than the other. inductive and data-driven, but situating them Over the course of independently coding and within the schemas of the WRID model offered an discussing a total of nine interviews, we consoli- organizational structure, which guided the coding dated two codes and added two codes, such that and the subsequent analysis. The coding process the final codebook maintained 23 data-driven codes therefore followed a hybrid inductive–deductive across the six schemas. All of the codes are dis- model. cussed in the Results section. The finalized code- To refine the codebook, the first and third book was then used to complete comprehensive authors independently conducted comprehensive line-by-line coding of the remaining 28 participants line-by-line coding of three participants (six tran- (56 interviews). The full race-related section of scripts across T1 and T2), discussed each coding every interview was coded. decision, and made iterative updates to the code- book (Braun & Clarke, 2013). Coding was done at Researcher Positionality and Reflexivity the phrase level, and no data from the race-focused section of the interviews remained uncoded, mean- The identities of all authors were relevant to this ing all statements were situated within the theoreti- project (Denzin, Lincoln, & Giardina, 2006). The cal model and codebook. Each statement was second author, a Black woman from a middle-class coded into only one code for greatest clarity. In background, conducted all interviews. In a semi- doing so, we were not attempting to holistically structured interview, the interviewer and partici- position each child in a given schema or phase, but pants co-construct knowledge and make meaning rather to capture the multiplicity of racial identity in relationship with each other and within social through comprehensive coding of participants’ context (Fine, 2006; Gilligan, 2015). Thus, what is statements. Each interview included coded state- spoken and unspoken is influenced by social ments situated across multiple schemas. This pro- norms, positionality, and social desirability. In this cess was repeated with three additional research, norms of age and authority (adult-to- participants with the aim of establishing clear dis- child), social class, race, and our respective identi- tinctions between codes and an exhaustive code- ties and experiences shaped the process and out- book in line with both data and theory. come (DeCuir-Gunby, 2020). “Race-matching" has The first and third author then independently been used as a means to mitigate the influence of coded three more participants, at which point a race in the interview process (e.g., Mishler, 1986). kappa reliability score was calculated (Syed & Nel- While a suitable practice, race influences the inter- son, 2015). The relatively low kappa of .64 view process regardless of the racial identity con- prompted a thorough review of coding. We found figurations present (Rogers, Moffitt, & Jones, 2021). that a majority of discrepancies centered on two of Were white participants more likely to say, “race
TABLE 1 Overview of Phases, Schemas and Distributions Across T1 and T2, and Data-Driven Codes Phase and schema Theoretical explanation Data-driven code Exemplary quotation Phase 1 Contact Unawareness, passive denial or Colorblindness “It doesn’t matter what race you are” T1 T2 oblivion regarding impact of Whiteness as unimportant Child selected “white” as least important card in sorting task 54% 46% racism and role of whiteness Whiteness as normal “It’s just normal” No recognition of different “It doesn’t matter what color you are, you still get treated the same.” treatment Reference to ethnicity “. . .not the cultural but um, the actual ethnic backgrounds of like the European-Europe.” Dis-integration Confusion, shame, initial Learning race doesn’t matter “I learned that it doesn’t matter –er if you’re black or white. It all – it only grappling with role of matters if you’re the same.” T1 T2 whiteness, can include Active denial of whiteness as “It doesn’t affect me that much because I’m just going to do what I’m going to downplaying its importance important do I guess.” 18% 12% Racism as historical or “I know that it [slavery and discrimination] was terrible but that’s the past- elsewhere you can’t fix it.” Essentialist recognition of “Boy and girl are opposite and Black and white are opposite.” difference Shame and guilt about “I’m kind of ashamed of my past. . . because the old white Americans were all (historical) racism – used to be really, really racist.” Re-integration Conscious of white identity, Engaging in stereotyping “Some people I know they try to act ghetto.” T1 T2 embracing racial superiority Descriptions of “reverse “. . . the news is always talking about African American kids getting shot and if either implicitly or explicitly racism” a white kid gets shot by a cop, it’s like oh yeah, that happened.” 10% 6% Engaging in overt racism “if it was back in the old days when Martin Luther King did that I would actually be kind of happy because White people get to do more stuff than Black people. . .” Support for racial segregation “Because lots of people are white and if you don’t have that much white people in your school it would be feeling weird.” Phase 2 Pseudo- Assimilationist view of Race shouldn’t matter “. . .you’re just a person, so as long as you’re a good person then you’re fine.” Independent whiteness as standard to T1 T2 achieve, recognition of racism Emotional response to racism “It makes me sad” without acknowledging one’s or inequality 17% 28% role in it Recognition of differences in “I know that’s another thing people get bullied about, their skin color and treatment stuff.” Recognition of racialized “. . .you would be expected to play sports more.” expectations Valuing diversity “I think it’s really important to see different backgrounds.” Immersion/ Recognition of racism as Recognition of the impact of “. . .they were of color and they used to get made fun of for acting white and WHITE RACIAL IDENTITY Emersion systemic, awareness of power racism that kind of stigma tends to shape personalities. . .” 7
8 MOFFITT, ROGERS, AND DASTRUP doesn’t matter” because they were talking with a “I would be treated differently and how I was treated over my youth definitely “Well we’ve heard about all the police shootings and stuff and that really just Black person? Existing literature with white youth suggests not; instead, our research aligns with find- “I mean at this school, like I think some teachers, I get special treatment ings that white children are conscious of norms of racial silence (e.g., Apfelbaum et al., 2008; Perry, 2001; Rogers et al., 2012). At times, our conversa- affected how – my personality and who I’ve grown to be. . .” tions captured these norms explicitly, for instance, when one child told the interviewer that “asking about people’s colors is racist.” Such a response communicates with clarity the expectations of color-blindness. In this way, our data demonstrate that our participants were able and willing to doesn’t happen with young white males.” Overview of Phases, Schemas and Distributions Across T1 and T2, and Data-Driven Codes engage in race-related conversations. Such open- ness may have in fact been aided by the cross- racial setting, as white youth have likely learned norms of racial silence from white adults (e.g., Perry et al., 2019). With these social norms and racial dynamics because I’m white.” known to be present in the data collection process, Exemplary quotation all three authors were in ongoing discussion with each other and with members of our research team, which included undergraduate and graduate Black, Asian American, Latinx, Multiracial, and white TABLE 1 (Contd.) men and women. The first and third authors, who are white women from a working-class and middle-class background, respectively, conducted the bulk of the coding, though all three authors Description of police violence were in communication. In this way, we relied on Critique of white privilege an “interpretive community” (Marecek, Fine, & against Black people Kidder, 2001) throughout the research process, an intersectionality intentional strategy to push against the dominance Data-driven code Recognition of of any single perspective. RESULTS AND DISCUSSION We address our first research question by review- ing our data-driven codes and how they were situ- ated within Helms’s (1990, 2020) theoretical model. recognition of intersectionality We then draw primarily on a quantification of our and privilege, benefits and data to address the second research question, offer- Confrontation of racism, ing an overview of how participants’ coded state- Theoretical explanation ments are distributed and an examination of harm of whiteness change over time and age-related differences. Additionally, we use illustrative case studies to interpret our findings. WRID among White Children and Adolescents Our first research question centered on the extent to which our participants’ meaning making about 6% 2% T2 T2 race and their white racial identity fits within the Autonomy Phase and WRID schemas and phases. To answer this, we schema turned to Helms’s (1990, 2020) theoretical model. 0% 0% T1 T1 Rather than consolidating the 23 data-driven codes
WHITE RACIAL IDENTITY 9 into broader, data-driven themes, as we would that the meaning making participants engaged in have in a fully inductive thematic analysis (Clarke varied widely. & Braun, 2014), we engaged in an iterative process For instance, when asked to reflect on what of situating these codes within the WRID model being white means, many participants downplayed (see Table 1) and framework for analysis. Nesting the importance of race and did not recognize our codes within the schemas and phases of racialization. One 4th-grade boy noted, “I think Helms’s (1990, 2020) theoretical model allowed us being white means that it’s just the way you were to code our data at the statement level, while situ- made.” He then added, “If you weren’t made like ating our findings within an existing organizational this nothing would change at all; you’d still have and interpretive framework. Notably, the distribu- your same personality, right, it’s just the color of tion of the 23 data-driven codes was not equal your skin.” The first sentence in this excerpt was across schemas and phases. Instead, the first four coded as Whiteness as normal, while the second was schemas each contained between four and six coded as Color-blindness. This boy explains that codes, whereas the latter two contained three and whiteness is “just” about skin color and that being one, respectively, giving an initial indication of our white has little to do with personality or life expe- participants’ predominant engagement with Phase riences. Although his phrasing may differ from the 1 schemas. semantics employed by white adults, the message is similar: Race does not matter, we should focus Evidence of Phase 1 schemas. Con- on the individual. This type of statement reflects tact. Across participants and time points, 49% both the logic of late childhood and socialization (n = 417) of the total 843 coded statements were into a society in which whiteness is normative and grouped into one of the five data-driven codes color-blind ideology prevails. If race does not mat- nested within Contact, making it by far the most ter, then inequity cannot be understood as a pro- prevalent schema. One code in this schema was duct of racism (Bonilla-Silva, 2013). This child is Reference to ethnicity. After identifying as “white” not voicing these connections, and we are not argu- in the card-sorting task, an 8th-grade girl stated, ing that he is even aware of them. By nesting this “I’m actually a little bit Italian. . . My dad’s like coded statement into the Contact schema, however, pretty Italian.” Rather than interpreting this we can recognize that, without a push to move response as evidence of the overlap between race beyond this kind of reasoning, this child may and ethnicity, by situating our data-driven codes develop into an adult engaging the same type of within the WRID model, we can recognize the system-accommodating racial logic. implications of engaging ethnicity rather than race. Some participants’ color-blind statements were Helms (2020) describes this type of response as a also coupled with No recognition of different treat- conscious or unconscious strategy to avoid identi- ment, for instance, in this exchange with a 6th-grade fying as white. The participant’s reference to ethnic boy: identity in response to a question about her racial identity captures the avoidance central to the Con- I: Right, so what are some of the good tact schema, even if she did not deliberately things about being white? intend to do so. In this way, the WRID model P: Um I don’t know, everybody is pretty provides an interpretive lens, situating the socio- much treated the same these days, so I cultural relevance of participants’ (non)engage- don’t think that there is anyone who ment with race. benefits pretty much so. I think it A key aspect of the Contact schema is a lack of really – there aren’t any good things conscious reflection on one’s whiteness (Helms, about being any color, like it’s just the 1990, 2020). Although this showed up in different same. ways, it was most directly captured by Whiteness as unimportant, the sole action-based, binary code in Many participants answered this interview ques- our codebook, comprised of instances in which tion in a similar way, using color-blind logic to participants chose “white” as the least important claim all races are treated equally. This boy’s identity of those they were asked to consider. phrasing, like a number of his peers, indicates that Overall, 82% of participants did so, indicating a his interpretation of something “good” would consistent minimization of whiteness as a relevant mean something that creates “benefits” for his part of their identity. By examining the breadth of racial group. Although he is claiming there are no statements surrounding this choice, we could see privileges to “being any color,” his equation of
10 MOFFITT, ROGERS, AND DASTRUP “good” with racially stratified “benefits” is trou- people without doing anything beyond feel- bling. Helms’s (1990, 2020) theoretical model can ing guilty (p. 43). help us understand and situate the potential impact of this type of Contact reasoning, which Situating racism solely as historical may be one reflects not only racial naivete but also an internal- way to remain convinced that one has no responsi- ization of a racist status quo. bility beyond occasional feelings of guilt. Rather Disintegration. Five data-driven codes were sit- than drawing the links between past and present, uated within the Disintegration schema, though it participants with statements in these codes dis- accounted for only 15% (n = 124) of coded state- tanced themselves from racism by framing it as ments overall (compared with 49% in Contact). irrelevant to their lives today. Among children When nesting our data-driven codes, we returned who made such claims, the WRID model can help to Helms’s (1990, 2020) theory many times to ade- us recognize how they are making sense of what quately parse what fell into Disintegration, which is they learn from parents, teachers, and other marked primarily by the nebulous emotions of sources of socialization. While many participants guilt, shame, confusion, and ambiguity about one’s mentioned learning about historical racism, specifi- whiteness. In our data, this sense of ambiguity and cally with references to slavery, Jim Crow era racial confusion seemed to be captured in children’s ref- segregation, and the Civil Rights Movement, few erences to racism as something of the past, or discussed learning about its present forms, and something that may exist elsewhere, but not in none drew historical through lines. This lacuna is their present reality. For instance, after a 5th-grade telling regarding what is taught and what white boy said he had never seen kids treated unfairly youth internalize. because of their race, he went on to explain, Another strategy included in the Disintegration “Because that time is really far behind us and schema is Active denial of whiteness as important. [coughs] and it doesn’t – you know it just doesn’t This code differed from Whiteness as normal in the really matter.” Statements like this, which were Contact schema in that participants were recogniz- coded as Racism as historical or elsewhere were often ing their own whiteness while simultaneously coupled with statements coded into Shame and guilt arguing that it does not affect them. For example, about (historical) racism. For example, a 6th-grade after explaining that she does not “really see a dif- girl who had just claimed there was nothing hard ference between skin colors,” a 6th-grade girl went about being white added: on to explain that this is important because “then people won’t think that I’m like racist or something P: Unless like in 4th grade we were like that.” This explanation offers a stark example learning about slaves and I felt kind of of a strategy used to mitigate being perceived as a weird like during that time when we “bad white person” in a system understood in were learning about it; it was kind of interpersonal terms. More explicit than most, this like, uh oh, what did we do, what are girl explained that her color-blindness is not a pro- we learning? duct of obliviousness, but is intentional and self- She went on to say that it was hard to talk protective. about, “Because it’s like, like you feel bad sort of Reintegration. The least prevalent Phase 1 and like it’s so long ago and it’s different now so schema, 8% (n = 66) of all coded statements were it’s like hard to like, to like get why and. . . but included in one of the four data-driven codes in yeah. . .” Her reasoning reflects that of most state- Reintegration. This may reflect the nature of our ments included in this code—she felt bad while data; we only coded verbal statements made dur- learning about slavery because she felt implicated ing our interviews, rather than behavioral in this history as a white person, yet her focus responses or in vivo interactions. As Helms (2020) remains on herself and her own unease. As Helms explains, “White superiority can be expressed (2020) notes: overtly or covertly” and failing to challenge others’ racist acts is an example of “covert Reintegration” By feeling guilty about a situation that they (p. 48). Thus, it is possible that more participants cannot change, living in a racist society, peo- were engaging Reintegration than what we cap- ple using Disintegration convince themselves tured in our text-based coding of interview data. that they have no responsibility for racism Despite this limitation, we did find examples of and, thus, they seek ways to restore their the more overt side of this schema. Among the good feelings about themselves as White data-driven codes, the most common was
WHITE RACIAL IDENTITY 11 Descriptions of “reverse racism,” in which partici- 4th-grade boy offered the following reasoning when pants explained how white people are the ones asked why being white does not matter: being disadvantaged in school or in U.S. society. P: Because everybody should be treated For instance, when a 6th-grade boy was asked why the same because it doesn’t matter he selected “white” as the least important identity whatpeople see; it matters what you during the card-sorting task, he first explained, are. “Because. . . people kind of use that in bullying,” before then arguing that being white in fact matters I: Where did you learn that or how do a lot: you know that? P: Um, because um I kind of – cuz a lot P: I know that because I’ve seen people – of the troublemakers who are Black, I’ve seen white people pick on Black um they – they act kind to uh they act people and I don’t think it’s okay nice to the people, even to the people because it’s like we’re the same; it’s that are like, have the same color, even just different skin color. if they’re not friends and not um like This boy is simultaneously engaging the logic of in their group of troublemakers and color-blindness while also pinpointing the racism stuff, but then whenever they uh—but he has witnessed. Coded into Race shouldn’t matter with the other, with some of the white and Recognition of differences in treatment, these people they just um are really, are codes both capture the Pseudo-independent per- really mean. spective casting racism as an issue of “bad white With this explanation, he touches on the atti- people” (Helms, 2020) rather than as systemic. This tudes and perspectives characteristic of Disinte- boy contends that the way to stop racist bullying is gration, including anger toward BIPOC to focus on individual traits and downplay the individuals and the “belief that Whites are no meaning of race. Again, while the immediate out- more racist than other groups” (Helms, 2020, p. comes of this kind of reasoning necessarily look 49). That this boy feels excluded or even bullied different for a 4th-grader than for adults, the WRID by classmates is not acceptable, but the ways in model can help us make sense of both. This boy is which he racializes and makes sense of his expe- vying with dueling realities, on the one hand the rience also perpetuate harm. Helms (2020) color-blind socialization he is likely receiving, and describes this schema as stable and widespread on the other hand the fact that he is growing up in because it reflects the racial inequity shaping a society shaped by white supremacy, wherein daily life in the United States. The racism white racism persists (Uma~ na-Taylor, 2016). This duality youth perpetuate as they engage the Disintegra- underscores the likelihood of holding perspectives tion schema can have direct negative conse- spanning multiple schemas, which may be particu- quences for Youth of Color (e.g., Sladek, Uma~ na- larly common among white youth. The strength of Taylor, McDermott, Rivas-Drake, & Martinez- color-blind socialization may mean that some white Fuentes, 2020). In terms of their own racial devel- youth remain tethered to the Contact schema even opment, if white youth recognize this harm, they as they begin progressing in their racial identity may shift to Phase 2 schemas. development in other ways. This 4th-grade boy rec- ognizes his whiteness and acknowledges the exis- Evidence of Phase 2 schemas. Pseudo- tence of racism, but does not examine his own role independent. This was by a large margin the most within it, instead keeping his view outward. commonly engaged of the Phase 2 schemas, with An outward facing perspective is characteristic 23% (n = 195) of all 843 statements in one of the of this schema, in which individuals recognize the five data-driven codes. Using these codes, we were role of race in structuring society, without engag- able to capture examples of the internal conflict ing in work to bring about systemic change Helms (1990, 2020) describes as what can push (Helms, 2020). This can take multiple forms, white individuals toward greater racial identity including the explicit valuing of diversity or multi- development. This often included an acknowledg- culturalism without a concomitant recognition of ment of racism perpetuated against BIPOC individ- racial inequity. This occurred in our data primarily uals, though participants’ solutions and when participants talked about their schools, peer explanations tended to remain superficial, which groups, or friends, as in the following example is characteristic of this schema. For example, a from an 8th-grade girl, “And like I’m white and
12 MOFFITT, ROGERS, AND DASTRUP that’s it and my friend Rashida, she’s from Jamaica. Acknowledging the ways in which white supre- And, and like, it’s just good that I can be around macy shapes her school experiences, this partici- all these different people and not only just one pant is recognizing that racism not only type of person.” With this statement, she is casting disadvantages BIPOC individuals, but it advan- the diversity of her friend group as something pos- tages white people. Recognizing that positive treat- itive for her. This girl may have heard this type of ment from teachers may be a product of racial “benefits of diversity” rhetoric from her parents or inequity rather than individual merit represents the teachers. Regardless of source, this fits with critical reflection characteristic of this schema. Helms’s (2020) description of the Pseudo- One aspect of white privilege some participants independent schema as characterized by behaviors touched on was captured in the code Descriptions of and perspectives embracing a positive view of race police violence against Black people. This occurred in relations without engaging in anti-racist action. the following exchange with an 8th-grade boy who This tendency was more explicit in exchanges was asked about being treated differently because about witnessing interpersonal racism. For exam- he is white: ple, a 6th-grade girl said, “Sadly, but still it’s going P: Um. . . well we’ve all heard about all on sometimes” in response to whether she had the police shootings and stuff and that ever seen kids getting treated differently because of really just doesn’t happen with young their race. She was then asked, “So what usually white males. All the systems really are happens when something like that happens?” to kind of racist, a lot of them are really which she replied, “Oh I just hear it and I just walk racist in some cases. away, but they’re not saying it to me but I still think it’s wrong.” This aligns with what Helms This boy names the systemic racism undergird- (2020) calls an “Ain’t it a shame” strategy, which ing the ongoing police violence against Black peo- “allows the White person the illusion of sensitivity ple in the United States, pointing out that he does without requiring that he or she actually do any- not face the same threat. Such an acknowledgment thing about it” (p. 57). By and large, white youth of the ways in which systemic racism shapes one’s are not expected to “do anything” about racism, own experiences was very uncommon among our particularly in terms of interrogating their own role participants. Based on what we know of socio- within it; they are instead taught the values of cognitive development and racial perspective tak- being a good person and treating everyone equally. ing (Quintana, 1998, 2008), this was not surprising. While these values are positive, their individual Helms (2020) also points out that even among focus can passively work to maintain systemic white adults, Immersion/Emersion “rarely inequity—if I am a good person then I have done becomes a dominant schema” (p. 63). She explains all I can do and the bad things that happen must that the deep and critical examination of both one- be happening to bad people (Helms, 2020). If, how- self and society that characterizes this schema can ever, white youth are pushed to turn their gaze on be painful and isolating from other white people their own whiteness and to recognize the systemic who are not at a similar place in their own racial nature of racial inequity, they may begin engaging identity development (Helms, 1990, 2020). For the final two Phase 2 schemas. white youth, without explicit anti-racist socializa- Immersion/emersion. Overall, just 4% (n = 30) of tion, engaging the final Phase 2 schemas may be coded statements were situated within the three unlikely. data-driven codes in this schema. The Immersion/ Autonomy. The last schema of Phase 2 was by Emersion schema is characterized by a critical far the least represented, with only 1% (n = 11) of examination of whiteness and a move away from overall coded statements situated in the single an assimilationist stance (Helms, 2020). The most data-driven code, coming from a total of seven par- common way it was coded was as Critique of white ticipants. Individuals engaging Autonomy have privilege. For example, when an 8th-grade girl was reflected meaningfully on what whiteness means to asked about seeing kids get treated differently them and reached a level of comfort with their because of their race, she mentioned that teachers anti-racist white identity (Helms, 2020). Autonomy treat her differently because she is white. She is characterized by regular engagement with issues explained, “Like they treat me with more respect I related to racial diversity, including an acknowl- kind of think. . . And like, they like trust me a lot edgment of interconnected systems of oppression. more.” When asked how that makes her feel, she For this reason, the sole data-driven code situated replied, “I feel like it’s kind of unfair.” in this schema is Recognition of intersectionality. No
WHITE RACIAL IDENTITY 13 participant brought up intersectionality, or the Yet, the trend toward greater engagement of the ways in which their multiple privileged or Phase 2 schemas suggests age-related change, pos- oppressed identities are interconnected (Crenshaw, sibly indicating that some combination of lived 1994), without being prompted. All statements in experience and increased socio-cognitive capacity this schema were in response to the final questions may prompt white racial identity development. To from the T2 interviews. parse possible age-related differences in the shift For example, when an 8th-grade girl was asked toward greater engagement with Phase 2 schemas, how things might be different for her if she were a we divided our sample into two groups: middle white boy she responded, “Um, I’d have a ton of childhood (2nd-5th grade at T1) and early adoles- privilege. I’d be like the top of the food chain, I’d cence (6th grade at T1). We then compared these have so much privilege,” indicating that she recog- groups in terms of their change in overall percent- nizes her own racial privilege as a white person, age of Phase 2 statements. Due to the small sample that she is aware of male privilege, and that she sizes, no statistical analyses were conducted; the has reflected on how the two may intersect. The comparison offers a descriptive overview. Because other statements coded into this schema were simi- the percentages represent proportion of total coded lar—recognizing that race and gender overlap to statements, an increase in Phase 2 statements nec- influence experience. One 8th-grade boy reflected essarily means a decrease in Phase 1 statements. on the impact of cumulative experiences of racism Among participants in middle childhood, 7% he would likely experience if he were a Black boy, (n = 7) of their statements were in Phase 2 schemas noting, “if I were to just instantly swap it probably at T1 and 10% (n = 11) at T2. Among participants wouldn’t change me much. But if I were to have in early adolescence, 21% (n = 58) of their state- grown up as an African American it definitely ments were in Phase 2 schemas at T1 and 44% would have affected me.” Yet, no participant spoke (n = 160) at T2. This breakdown shows a marked at length about their own or others’ experiences at change across time and by age group. Specifically, the intersection of race and gender, and a critical the proportion of Phase 2 responses remained reflection on the interconnectedness of systems of stable and low among participants in middle child- oppression was notably absent. hood. However, among those in early adolescence at T1, we see a greater proportion of Phase 2 responses overall and the prevalence doubles from WRID Distribution and Age Variation T1 to T2. Taken together, there is evidence of both Our second research question explored the distri- within-person development and age-related differ- bution of participants’ coded statements across ences among these participants, with the overall schemas and phases at T1 and T2, as well as percentage of Phase 2 statements increasing over whether age-related variation was found. time, driven almost exclusively by the early adoles- Although there was a clear tendency toward cents in our sample. engaging Phase 1 schemas more than Phase 2, we To add texture and illustrate what the differ- were interested in whether there was any change ences in white racial identity can look like at the in this trend across the two time points. An over- person level across time points, we conclude with view of the proportions of coded statements two brief case studies. These case studies are nei- across schemas at T1 and T2 can be seen in Fig- ther exhaustive nor representative of our full sam- ure 2. Overall, 82% of our participants’ statements ple. Instead, they offer examples of trajectories at T1 were coded within the three Phase 1 sche- found within our data. The first describes a white mas; this dropped to 64% at T2. Thus, at both youth in middle childhood who remained primar- time points, the majority of our sample engaged ily in the Phase 1 schemas across time points, as in greater accommodation of than resistance to was the case for most of the younger participants. internalized norms of white supremacy. Despite The second is illustrative of a white adolescent the decrease from T1 to T2, the fact that the who was primarily engaging Phase 1 schemas at majority of participants’ statements were coded in T1 and primarily Phase 2 schemas at T2, which Phase 1 schemas at both time points was not sur- occurred more frequently among our older partici- prising, given the prevalence of color-blind social- pants. ization among white parents and teachers (Loyd & Gaither, 2018), and the systemic privileging of Case Study 1: Sam. Sam was in 2nd grade at white students and perspectives in U.S. schools T1, when 100% of the race-related section of his (Aldana & Byrd, 2015). interview was coded into Phase 1 schemas. In 4th
You can also read