Background: “Of Color” Identities

Our investigation focuses on Latinas and Asian American women—two panethnic groups whose members include large percentages of foreign-born individuals (US Census 2022). Ongoing processes of racialization and experiences with discrimination impact Latinos and Asian Americans, especially as it relates to racialization based on perceived foreignness (Takaki Reference Takaki1989; Kim Reference Kim1999; Chang Reference Chang2001; Haney Lopez Reference Haney Lopez2006; Wong et al. Reference Wong, Ramakrishnan, Lee and Junn2011; Junn and Masuoka 2013; Omi and Winant 2015).Footnote ² Scholars have also uncovered gender dynamics within these communities (Hardy-Fanta Reference Hardy-Fanta1993; Glenn 1999; Lien et al. 2004; Uchida 2008; Bejarano, Manzano, and Montoya Reference Bejarano, Manzano and Montoya2011; Hardy-Fanta et al. Reference Hardy-Fanta, Lien, Pinderhughes and Sierra2016; Phillips and Lee Reference Phillips and Lee2018; Bejarano et al. 2021). Scholarship dedicated to understanding the politics of Latinas and Asian American women is expanding (Bejarano Reference Bejarano2014; Harvie 2016; Phillips and Lee 2020; Phillips 2021; Cargile 2021), thereby addressing what Cathy Cohen (Reference Cohen and Carroll2003: 191) described as “the invisibility of women of color in studies of American politics.”

Our research is about potential public opinion effects of identifying as a woman of color among Latina and Asian American women. Historically, women have mobilized politically along the lines of gender, race/ethnicity, and their intersections (Terborg-Penn 1995; Giddings Reference Giddings1996; Cohen, Jones, and Tronto Reference Cohen, Jones and Tronto1997; Gray White 1999; Hill Collins Reference Hill Collins2000; Lien Reference Lien2001; Bejarano and Smooth Reference Bejarano and Smooth2022). In the 1950s, 1960s, and 1970s, social movements challenged the boundaries around racial/ethnic, gender, immigrant, panethnic, and other identities, contesting social norms and seeking to alter power arrangements (Freeman 1975; Espiritu 1992; Robnett Reference Robnett1997; Franklin 2000; Evans Reference Evans2003; Roth Reference Roth2004; Beltran Reference Beltran2010; Omi and Winant 2015). In the 1970s, self-identified “women of color” and “Third World Women” created new alliances and organizations that were self-consciously distinct from the group “white women” and actively worked to politicize these new alliances by making space for both the unique experiences of any one group (i.e., Black women, Asian women) and working together to decentralize any one racial/ethnic group of women (Moraga and Anzaldúa 1979; Hong 2018; Romney Reference Romney2021). For example, Romney describes 1970s activism that involved “women of color who came together across ethnicities and race and ‘operated together as a political category’” (2021: 11). This rich history of centering the marginalized informs our expectations of current WOC attitudes.

Today, the term “women of color” is more commonly invoked in the context of electoral politics and holds relevance for those women who might claim membership within its boundaries (Greene et al. Reference Greene, Matos and Sanbonmatsu2022; Matos et al. Reference Matos, Greene and Sanbonmatsu2023). Like the term “people of color” (Pérez Reference Pérez2021; Starr and Freeland 2022), the meaning of “women of color” is contested. And because it is a contested identity, women who self-identify or are perceived as other than non-Hispanic white may not automatically self-identify as a “woman of color.” In fact, Greene et al. (Reference Greene, Matos and Sanbonmatsu2022) and Matos et al. (Reference Matos, Greene and Sanbonmatsu2023) found variation across race/ethnicity in whether women identify as such. We found that virtually all Black women identified as “women of color”; but such was not the case for other groups of women (Matos et al. Reference Matos, Greene and Sanbonmatsu2023).

The term WOC is often most legible for Black women given the history of a Black-white racialized system in the United States. However, that is not to argue that there is not contestation within the “Black woman” category based on other racialized markers, like immigration. For example, non-African American Black women might experience their Blackness in distinct ways and in ways that may not be legible to others given the history of the United States. Jacqui M. Alexander and Chandra Mohanty speak to their own experiences as a Black Caribbean woman and a Brown Indian woman, respectively, when they state that their experiences “could be recognized and acknowledged only to the extent that they resembled those of African American women” (Alexander 1997: xiv). Furthermore, they state that their “foreign-ness” muted “the legitimacy of [their] claims to the experiences of different racisms” (Alexander 1997: xiv). Hence, any work that uses an intersectional perspective must contend with other racialized markers for women other than non-Hispanic white given their own socio-political history in the United States, which for many includes immigration.

As Masuoka and Junn (Reference Masuoka and Junn2013) observe, the traditional concept of a color line that separates Black and white Americans on the dimension of hierarchy fails to capture the ways that immigration informs U.S. racial and ethnic categories. Applying their racial prism framework, they propose that a diamond pattern that situates Asian Americans and Latinos in relationship to Blacks and whites better captures the racial hierarchy of the United States historically (Masuoka and Junn Reference Masuoka and Junn2013). They note: “The common experience for Asian Americans and Latinos of being asked the question “Where are you from?” typifies the context of the tension between minority-group status and perceptions of Americanness” (Masuoka and Junn Reference Masuoka and Junn2013: 27). However, within-group heterogeneity is not accounted for in Masuoka and Junn’s conceptualization of the racial hierarchy in the United States, as noted by Alexander and Mohanty (Reference Alexander, Mohanty, Alexander and Mohanty1997) as well as scholars who focus on Afro-Latinas’ experiences (Flores and Jiménez Román Reference Flores and Jiménez Román2009; Jiménez Román and Flores Reference Jiménez Román and Flores2010). In general, patterns of identity and the identity terms themselves are unstable as race and immigration categories are contested and shifting.

In Greene et al. (Reference Greene, Matos and Sanbonmatsu2022), for example, we observed that women can take different “pathways” to a WOC ID. We argued (2022: 14):

In other words, nativity matters as well as women’s lived experiences in the United States.

WOC and Policy Preferences among Latina and Asian American Women

To the degree that Asian American women and Latinas identify as WOC, then, they are, we argue, identifying with other marginalized groups of women who are not of the “default” racial group. In contrast to the 1970s usage, the WOC term is arguably more frequently heard in conversations today about electoral politics and voter demographics than grassroots or international politics (Matos et al. Reference Matos, Greene and Sanbonmatsu2023). But as in the past, WOC ID is a designation that brings women together across major racial/ethnic group boundaries.

The WOC term that animated activism in the 1970s arose from leftist politics and was situated within a global perspective (Matos et al. Reference Matos, Greene and Sanbonmatsu2023). Matos et al. (Reference Matos, Greene and Sanbonmatsu2023: 551) observe that for the 1970s activists, “to identify as a WOC was to do identity politics.” Hence, we propose that WOC self-identification has implications for the policy views of Latinas and Asian American women. As a raced and gendered identity, being a WOC may be related to the likelihood of supporting an agenda that can improve the lives of historically marginalized groups. Our thinking about the political implications of this identity builds on the growing field of intersectional studies (Crenshaw Reference Crenshaw1989; Crenshaw Reference Crenshaw1991; Simien Reference Simien2005; Simien Reference Simien2006; Hancock Reference Hancock2011; Tormos Reference Tormos2017; Montoya et al. Reference Montoya, Bejarano, Brown and Gershon2021). Intersectionality theories originated with Black women’s perspectives and activism; Black women see the struggles of Blacks and women as inherently connected because the categories themselves are mutually constitutive (Alexander-Floyd Reference Alexander-Floyd2012; Giddings Reference Giddings1996; Harris-Perry Reference Harris-Perry2011; King Reference King1988; Nash Reference Nash2011).

Black feminist activists of the 1960s and 1970s articulated the simultaneity of these oppressions as well as the goal of improving the lives of marginalized groups broadly, beyond Black women, Black people, and women as a group. For example, the Combahee River Collective (CRC 1977) noted that “The inclusiveness of our politics makes us concerned with any situation that impinges upon the lives of women, Third World and working people. We are of course particularly committed to working on those struggles in which race, sex, and class are simultaneous factors in oppression.” When King (Reference King1988) theorized that “interactive oppressions” mark Black women’s lives and politics, she observed that “the necessity of addressing all oppressions is one of the hallmarks of black feminist thought” (1988: 43; also see Nash Reference Nash2011). And King cited civil rights activist Fanny Lou Hamer’s observation that “when I liberate myself, I’m liberating other people” (1988: 43). As Romney (Reference Romney2021: 13) notes about the Third World Women’s Alliance, theirs was “a movement not against men but a movement for justice that was pro-woman, pro-people of color, pro-children, and pro-family.”

Empirical studies of elites find evidence of the political manifestations of intersectional identities. For example, in a study of congresswomen’s websites, Brown and Gershon found that “All the legislators examined implicitly connected their identity to an advocacy issue, but minority congresswomen are most likely to include more than one marginalized identity (such as their race, class, and/or gender) to illustrate their concern for a disadvantaged subpopulation” (Brown and Gershon Reference Brown and Gershon2016: 101). They argued that “women legislators of color are more likely to view race/ethnicity and gender as intersectional forms of representation” (2016: 103).

Our interest lies with the general public rather than activists or elites. But drawing on past research about how gender, race, and their intersection affect politics, we theorize that Asian American women and Latinas who self-identify as WOC are standing in solidarity with other marginalized communities. For these reasons, we hypothesize that WOC ID will be associated with greater support for policies that benefit marginalized outgroups.

“Of Color” Identities and Public Opinion Studies

Public opinion studies of Latinas and Asian American women have examined such factors as nativity, language, generation, socioeconomic status, and linked fate (Lien et al. 2004; Bejarano Reference Bejarano2014; VanSickle-Ward and Pantoja Reference VanSickle-Ward, Pantoja, Brown and Gershon2016; Bejarano et al. 2021; Carey and Lizotte Reference Carey and Lizotte2023). Traditional political factors such as ideology and partisanship play a role in opinion formation as well (e.g., Bejarano Reference Bejarano2014), although both Latinos and Asian Americans are less incorporated into the two-party system than other groups (Hajnal and Lee Reference Hajnal and Lee2011). We contribute to this body of work with our examination of how WOC ID is related to Latinas and Asian American women’s attitudes.

We build on recent developments in the study of “of color” identities (Matos and Sanbonmatsu Reference Matos and Sanbonmatsu2024). Our expectations are similar to those of Pérez (Reference Pérez2021) who proposed that identity as a person of color, or “POC ID,” leads people to support issues that are perceived as disproportionately affecting other racial/ethnic groups. In addition to “policies broadly impacting minorities” (2021: 128), Pérez (Reference Pérez2021) examined “support for policies narrowly impacting minorities—issue domains that are more narrowly associated with a specific minority group” (2021: 128). For example, Pérez considered how non-Latinos reacted to undocumented immigrants, how non-Asian Americans reacted to legal immigration, and how non-Blacks reacted to #BlackLivesmatter. Pérez (Reference Pérez2021) found that higher levels of POC ID “heighten solidarity with PoC, pro-minority sentiment, pro-Democrat feelings, support for policies broadly affecting minorities, and even support for policies narrowly impacting some minorities” (2021: 131).

We depart from Pérez (Reference Pérez2021) in two ways. First, we focus on WOC rather than POC. Second, while Pérez (Reference Pérez2021: 67) asserts that all people who are “non-White” identify as POC, we believe, in contrast, that people who could be categorized as “of color” may not self-identify with the POC group. We see a parallel choice to self-identify as WOC and do not assume WOC ID among Latinas and Asian American women. While Pérez (Reference Pérez2021) sees racial/ethnic group identity as nested within a POC ID, we are agnostic about the hierarchy of the WOC ID.

Meanwhile, we depart from recent work about WOC. The study by Carey and Lizotte (Reference Carey and Lizotte2023), also using the 2020 CMPS, argues that WOC-linked fate leads to common goals that benefit all WOC, finding that “WOC linked fate appears to promote support for policies perceived as beneficial to women of color” (2023: 16). Unlike Carey and Lizotte, we measure the effect of claiming identity as WOC rather than linked fate (Dawson Reference Dawson1994). We also theorize differently about how the WOC category is related to policy attitudes. While Carey and Lizotte posit that WOC-linked fate will lead to more support for policies that benefit WOC, we theorize that policies beyond one’s own group will be affected; in other words, we expect to find that WOC ID leads to support for policies that benefit members of an outgroup rather than benefiting oneself as an ingroup member of WOC.

The Limits of “Of Color” Identities

Importantly, we also contend that there are limits to a WOC ID and that political attitudes and racialized ideologies have the potential to cut across WOC ID and solidarity with outgroups via policy preferences. Given the socio-political history of the United States, we expect racial attitudes to play a moderating role on WOC ID and progressive policy preferences.

Racial attitudes have long played a role in American public opinion (Kinder and Sanders Reference Kinder and Sanders1996; Sears et al. Reference Sears, Sidanius and Bobo2000). Scholars have found racial resentment, as one measure of racial attitudes in the United States, to influence both racial and non-racial policy preferences. They have argued that racial resentment captures a “belief in longstanding and entrenched American values of racial individualism and deservingness” (Matos Reference Matos2020: 425). Further, expressions of racial resentment exist among all racial and ethnic groups (Nunnally and Carter Reference Nunnally and Carter2012; Orey Reference Orey2004; Smith Reference Smith2014). Identifying as a “woman of color” does not preclude being racially resentful.

Racial resentment has been largely conceptualized as a more subtle way to get at anti-Black attitudes. For that reason, we expect a strong connection to attitudes toward the ban on chokeholds and the issue of police violence aimed at Black Americans (Jefferson et al. Reference Jefferson, Neuner and Pasek2021; Riley and Peterson Reference Riley and Peterson2020). However, although racial resentment was largely designed to measure whites’ views towards Blacks, subsequent scholarship has tapped the concept to capture racial attitudes more broadly and a wide range of policy questions (e.g., Enders and Scott Reference Enders and Scott2019; Lucas and Mohamed Reference Lucas and Mohamed2021; Matos Reference Matos2020; Tesler Reference Tesler2012; Tokeshi Reference Tokeshi2023). For example, Wallace and Zepeda-Millán (Reference Wallace and Zepeda-Millán2020) find that racial resentment explains attitudes toward Trump’s border policies including child separation. Meanwhile, racial resentment has been used in models of anti-Muslim attitudes (Lajevardi et al. Reference Lajevardi2023). Hence, we do expect it to be important for all of the issues we investigate, which are heavily racialized.

Equally important, we expect that Latinas and Asian American women—regardless of WOC ID—with higher levels of racial resentment will hold less liberal attitudes on outgroup policy issues, given their apparent lack of understanding of the structural nature of issues surrounding policy reform, police brutality, immigration, and anti-Muslim politics. Importantly, expressions of racial resentment will be correlated with less liberal attitudes on outgroup issues due to the recipients of the outgroup in question, as they could be seen not only as outsiders but as also threatening the ingroup’s status (Davis and Wilson Reference Davis and Wilson2022).

Research and Methods

In order to test our expectations about the relationship between WOC ID and policy attitudes for Latinas and Asian American women, we turn to the most comprehensive and representative dataset available, the 2020 CMPS. Our main independent variable is WOC ID. We measure self-identification with a question we added to the 2020 CMPS: “Do you think of yourself as a woman of color?” In the 2020 CMPS, 36% of Latinas and 58% of Asian American women answered affirmatively (Greene et al. Reference Greene, Matos and Sanbonmatsu2022). The higher rate of “of color” identity for the Asian American sample compared with Latinas is similar to Starr and Freeland’s (Reference Starr and Freeland2023) finding using a national survey of POC ID.

Our hypotheses about Latinas and Asian American women concern policy questions that specified a marginalized outgroup. Most of the policy questions on the 2020 CMPS could arguably directly benefit Latinas and Asian American women. However, we were able to identify three policy questions that met our criteria for both Latinas and Asian American women and a fourth policy question for the Asian American women sample only.Footnote ³

Criminal justice and criminal justice reform, our first policy area, is a racialized policy space that disproportionately impacts Blacks and Latinos, with implications for public opinion (Tuch and Weitzer Reference Tuch and Weitzer1997; Walker, Spohn, and DeLone 2000; Hurwitz et. al Reference Hurwitz, Peffley and Mondak2015; Edwards et al. Reference Edwards, Esposito and Lee2018; Corral Reference Corral2020). The #BlackLivesMatter movement has brought public attention to police brutality, police misconduct, and disproportionate surveillance of both Black and non-Black communities alike. Eric Garner’s death in 2014 due to a police chokehold imprinted the words “I can’t breathe” on a national discourse around police brutality. Unfortunately, these same words would be uttered as George Floyd died from a police officer kneeling on his neck in 2020 spurring some of the largest protests worldwide against violence against Black people.

Because Blackness is diasporic, we have selected a 2020 CMPS question that specifically references the “killing and abuse of African Americans” (italics ours); we therefore categorize it as an outgroup issue for both Latinas and Asian American women, especially as some Latinas are also racially Black. The question wording from the 2020 CMPS survey is as follows, with the response option that was coded as the most liberal response in parentheses:

The second policy domain concerns Muslims. Most Muslims in the United States trace their origins to the Middle East/North Africa (MENA) region, and MENA people have been considered “white” by the US Census (Kayyali Reference Kayyali2013; Strmic-Pawl, Jackson, and Garner Reference Strmic-Pawl, Jackson and Garner2018; Hephzibah et al. 2018).Footnote ⁴ Identifying as Muslim is a matter of religious affiliation not racial identity. However, Muslims are a racialized group (Selod and Embrick Reference Selod and Embrick2013; Lajevardi Reference Lajevardi2020; Oskooii et al. Reference Oskooii, Dana and Barreto2021). In fact, a MENA category is scheduled to make its debut in the 2030 US Census (Marks et al. Reference Marks, Jones and Battle2024). We include two 2020 CMPS policy questions related to Muslims. The question about patrolling and securing “Muslim neighborhoods before they become radicalized” reflects Islamophobia prevalent since 9/11 and the subsequent passage of the USA PATRIOT Act. As the question wording indicates, this policy debate racializes Muslims and casts them as security threats. As Lajevardi (Reference Lajevardi2020) and others have documented, media coverage of Muslim Americans has grown more negative since 9/11. Meanwhile, President Trump’s “Muslim ban” to prevent travel to the United States from predominantly Muslim countries—one pillar of his white nationalist policy agenda—was one of his first acts as president during his first term (Lajevardi Reference Lajevardi2020). The question wording from the CMPS is as follows:

The third policy area, for Asian American women only, is about child separation at the border. U.S. immigration policies reflect beliefs about race and give rise to racial categories, forestalling some racial projects while making others possible (e.g., Haney Lopez Reference Haney Lopez2006; Hochschild and Powell Reference Hochschild and Powell2008; Masuoka and Junn Reference Masuoka and Junn2013; Omi and Winant 2015; Kim Reference Kim2023). One contemporary policy debate that was included on the 2020 CMPS is a question about migration at the U.S.-Mexico border specifically. This policy issue attracted widespread attention during the Trump era, when President Trump embraced a policy of separating children from their families at the border, thereby treating the separated children as unaccompanied minors (Wallace and Zepeda-Millán Reference Wallace and Zepeda-Millán2020). Although Trump would change direction after public condemnation, the damage inflicted on the families would be lasting because the federal government failed to track the children (Dickerson Reference Dickerson2022; Wallace and Zepeda-Millán Reference Wallace and Zepeda-Millán2020). Because this policy is about the southern border and its impact on Latin American and Caribbean migration, we categorize this immigration issue as one that does not disproportionately impact Asian American women. The question wording is:

Because our research is designed to examine how WOC ID is related to policies that benefit outgroups, we limit the analysis to monoracial women (i.e., those who identify solely as either Latina or Asian American, without reporting membership in additional racial/ethnic groups).Footnote ⁵ Because two policy questions have a disproportionate impact on self-identified Muslims, we exclude women who identify as Muslim from our analysis.Footnote ⁶ In the end, our unweighted N stands at 1,537 for Latinas and 1,886 for Asian American women.Footnote ⁷

In the first set of models, we include three demographic covariates: age, education, and income.Footnote ⁸ In a second set of models, we include additional control variables. In light of past work on public policy issues concerning race, immigration, criminal justice, and policies targeting Muslims, we add controls for nativity (U.S.-born), parental status, importance of racial identification, self-reported skin color, party identification, ideology, and political interest (e.g., Bejarano Reference Bejarano2014; Wallace and Zepeda-Millán Reference Wallace and Zepeda-Millán2020; Oskooii et al. Reference Oskooii, Dana and Barreto2021; Silver et al. Reference Silver, Goff and Iceland2022; Tokeshi Reference Tokeshi2023).Footnote ⁹

Partisanship and ideology are expected to play a role for all four policies. With respect to #BlackLivesMatter and criminal justice policies, the parties’ reputations on crime and longstanding party positions on civil rights matters are relevant, as are the distinct positions taken in the 2020 presidential election by President Trump and presidential candidate Joe Biden (Carmines and Stimson Reference Carmine and Stimson1989; Petrocik Reference Petrocik1996; Stephens-Dougan Reference Stephens-Dougan2016). Trump’s Muslim travel ban further solidified a partisan dimension to policies concerning Muslims (Lajevardi Reference Lajevardi2020). Likewise, the Trump administration’s controversial child separation policy at the border alongside immigration issues writ large leads us to expect a partisan relationship (Wallace and Zepeda-Millán Reference Wallace and Zepeda-Millán2020; Matos Reference Matos2020).

Finally, we include a model that controls for racial resentment. Racial resentment is coded from 0 to 1 (from lowest to highest resentment). The racial resentment items are: “Irish, Italians, Jewish and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors”; “Generations of slavery and discrimination have created conditions that make it difficult for Blacks to work their way out of the lower class”; “Over the past few years, Blacks have gotten less than they deserve”; “It’s really a matter of some people not trying hard enough; if Blacks would only try harder they could be just as well off as Whites.” Among Asian American women (r =.-.27) and Latinas (r =- .20) identifying as WOC, and racial resentment are negatively correlated. Descriptively, 5% of Asian American women and 3.9% of Latinas scored a “1” on the “0” to “1” continuous scale of racial resentment, with 1 representing the most racially resentful, and 61% of Asian American women scored 0.5 or higher on the scale, as did 62% of Latinas.

Analysis

In Table 1, we display the means and standard deviations for our key dependent variables for each sample. Response options took the form of a five-category variable from strongly agree (or strongly support) to strongly disagree (or strongly oppose). All dependent variables were recoded such that higher values represent the more liberal responses. For both groups of women, police reforms “to stop the killing and abuse of African Americans” such as the ban on chokeholds and strangleholds, yielded the most liberal responses (in favor of a ban), with mean values of over 4 on the 5-point scale. For Asian American women, opposition to separating parents and children at the border was also in the liberal direction with a mean approaching 4 on the 5-point scale (with 5 representing strong opposition).

Table 1. The Policy Views of Latinas and Asian American Women

Source: 2020 CMPS.

Analyses are limited to women who are monoracial and do not self-identify as Muslim.

The exact question wordings are: “We need police reform to stop the killing and abuse of African Americans. We need to ban chokeholds and strangleholds, and greatly reduce police shooting at suspects.” (strongly support); “We need to empower law enforcement to patrol and secure Muslim neighborhoods before they become radicalized.” (strongly disagree); “Ban people from predominantly Muslim countries from entering the US.” (strongly disagree); “We should end the policy of separating immigrant children from their parents at the border.” (strongly support).

The response options are coded 1 to 5 from the most conservative to most liberal response. Cell entries are mean values with the standard deviation in parentheses.

Meanwhile, for both groups of women, views on the two policies related to Muslims were less liberal on average. The average response to the surveillance question of Muslim neighborhoods was about 3 (or neutral), whereas the Muslim ban attracted some support, with a mean response closer to 2 on the 5-point scale for both groups of women (with 5 the most liberal response). Thus, we find variation within both groups of women and across the four policy areas.

We turn to a multivariate analysis with the policy questions as the dependent variables. The main independent variable of interest, WOC ID, is coded “1” if the woman self-identified as WOC and “0” if she said “no” or “unsure.” We use a least squares regression model with the policy questions scaled such that higher values represent more liberal responses. We begin the analysis with our sample of Latinas and a simple model with the demographic covariates of age, education, and income.

Figure 1 provides initial support for our hypotheses.Footnote ¹⁰ Those Latinas who self-identify as WOC are more likely than other Latinas to support a ban on chokeholds of Blacks, oppose surveillance of Muslim neighborhoods, and oppose a Muslim travel ban. This positive relationship is in the direction that we hypothesized, with WOC ID leading to greater support for policies that would benefit marginalized outgroups.

Figure 1. WOC ID and Policy Views (simple model). Note: 2020 CMPS. This figure depicts the effect of self-identifying as a woman of color on policy views, controlling for demographic factors (age, education, and income). Full results appear in the appendix

We pursue a similar analytic strategy for Asian American women, adding the fourth policy question about child separation at the border—a policy question unlikely to directly impact Asian American women. In Figure 1, we find that in a simple model, WOC ID is positively related to all four policy questions as we hypothesized.Footnote ¹¹ Asian American women who claim the WOC identity are more likely than other Asian American women to support policies that disproportionately affect marginalized outgroups. Self-identifying as WOC is related to Asian American women supporting bans on chokeholds of African Americans, opposing increased surveillance of Muslim neighborhoods, opposing a Muslim travel ban, and opposing separating children from their families at the border.

To more rigorously test our hypotheses, we add a host of additional explanatory variables to the model with results for Latinas displayed in Figure 2.Footnote ¹² We include nativity, parental status, importance of racial identification, self-reported skin color, party identification, ideology, and political interest.

Figure 2. WOC ID and Policy Views (full model). Note: 2020 CMPS. This figure depicts the effect of self-identifying as a woman of color on policy views, controlling for other factors (e.g., demographic and political variables). Full results appear in the appendix

The results of this fully specified model for Latinas appear in Figure 2. Even after adding these controls to the model, WOC ID remains positively signed and significant: as hypothesized, Latinas who take up the WOC identity are more supportive of all three policies that benefit marginalized outgroups. The control variables work mostly as expected with the political variables having a more consistent relationship across the three policy areas. Ideology and partisanship matter, meaning that those respondents who are Democrats and liberal take positions in the liberal direction. Older Latinas are more opposed to the Muslim ban than younger Latinas. Lajevardi et al. (Reference Lajevardi2023) find similar effects in relation to age and the Muslim ban. Latinas with children at home are more supportive of the ban on chokeholds. This may reflect the vulnerability of Latino youth to surveillance and stereotypes about criminality (Beckman and Rodriguez Reference Beckman and Rodriguez2021; Christiani Reference Christiani2020; Welch et al., Reference Welch, Price and Yankey2002) and indicate that chokeholds are an issue for the Latino community writ large and especially Afro-Latinos. Afro-Latinas/os tend to live in closer proximity to African Americans, as well as other Black diasporic communities (Logan Reference Logan, Jiménez Román and Flores2010), and thus are subjected to the same kinds of anti-Blackness that all diaspora members receive in the United States (Greer Reference Greer2013).

We hold constant the full range of explanatory variables for Asian American women similar to the model for Latinas (Figure 2).Footnote ¹³ In three of the four models, WOC ID is statistically significant and positive: Asian American women who self-identify as WOC are more likely than other Asian American women to oppose surveillance of Muslim neighborhoods, oppose the Muslim ban, and oppose separating children from their families at the border, consistent with our expectations. Democratic partisanship and being liberal are positively related to all policies. Older Asian American women are more opposed to the Muslim ban than younger Asian American women. In only one case, the surveillance of Muslim neighborhoods, was the importance of racial identity significant, but this effect is in the opposite direction than we expected: Asian American women who place greater importance on their identity as Asian Americans are less opposed to surveillance. We suspect there is intragroup heterogeneity in how Asian American women view this surveillance. Although we do not test this empirically, it could be the case that for some Asian American women whose Asian American identity is important, they may want to differentiate themselves from Muslims or other outgroups (Mishra Reference Mishra2016). This finding opens the door for future research. While the inclusion of ideology diminishes the WOC ID effect in some cases, ideology and WOC ID are distinct measures. The two variables are positively, but not highly, correlated.Footnote ¹⁴

Together, the models show strong support for our hypotheses. The relationships are in the direction we expected given the link between WOC ID and political solidarity: WOC ID is associated with the policy direction favoring outgroups.

We also investigated the results with the inclusion of a measure for racial resentment (Kinder and Sanders Reference Kinder and Sanders1996) (Tables 2, 3).Footnote ¹⁵ For both Latinas and Asian American women, we found racial resentment is among the strongest and most consistent predictors of opposition to liberal policies towards marginalized outgroups (Tables 2, 3). In all models, Latinas and Asian American women with higher levels of racial resentment were less likely to stand with marginalized outgroups. In other words, those who were racially resentful were least likely to oppose a Muslim ban, law enforcement, and surveillance in Muslim neighborhoods, child separation at the southern border, and police chokeholds. Once racial resentment was included, the WOC ID measure was no longer statistically significant in two of the models for Latinas (chokeholds and Muslim neighborhoods), and two of the models for Asian American women (the anti-Muslim policies).Footnote ¹⁶

Table 2. WOC ID and Policy Views, Latinas (full model with racial resentment)

*** p < 0.001; ** p < 0.01; * p < 0.05.

Source: 2020 CMPS.

Analyses are limited to women who are monoracial and do not self-identify as Muslim.

The dependent variables are coded 1 to 5 from the most conservative to most liberal response.

WOC ID is self-identification as a “woman of color,” coded 1, 0 otherwise; education is coded 1 to 5 (from less than high school to post-graduate education); income is coded 1 to 12 (from low to high income); child at home is coded 1 for child under 18 at home, 0 otherwise; racial identity importance is coded 1 to 5 (from not at all important to extremely important); skin color (self-rating) is coded 1 to 10 (from lightest to darkest; US born is coded 1 if the respondent was born in the United States, 0 otherwise; party identification is a 7-point scale (from strong Republican to strong Democrat); ideology is coded 1 to 5 (from very conservative to very liberal); political interest is coded 1 to 4 (from not at all to very interested in politics); racial resentment is coded from 0 to 1 (from lowest to highest resentment).

Table 3. WOC ID and Policy Views, Asian American women (full model with racial resentment)

*** p < 0.001; ** p < 0.01; * p < 0.05.

Source: 2020 CMPS.

Analyses are limited to women who are monoracial and do not self-identify as Muslim.

The dependent variables are coded 1 to 5 from the most conservative to most liberal response.

WOC ID is self-identification as a “woman of color,” coded 1, 0 otherwise; education is coded 1 to 5 (from less than high school to post-graduate education); income is coded 1 to 12 (from low to high income); child at home is coded 1 for child under 18 at home, 0 otherwise; racial identity importance is coded 1 to 5 (from not at all important to extremely important); skin color (self-rating) is coded 1 to 10 (from lightest to darkest; US born is coded 1 if the respondent was born in the United States, 0 otherwise; party identification is a 7-point scale (from strong Republican to strong Democrat); ideology is coded 1 to 5 (from very conservative to very liberal); political interest is coded 1 to 4 (from not at all to very interested in politics); racial resentment is coded from 0 to 1 (from lowest to highest resentment).