Abstract
Four studies examine whether demographic forms elicit social identity threat and institutional exclusion for Latine Americans before and after the 2020 Presidential election and leading up to the 2024 election. Using a mock-Census form and widely used ethnicity/racial designations, we manipulated the framing of Latine identity through a two-item (category listed separately) versus a single-item (combined) format, and also varied the inclusion of a citizenship question. Overall, Latine Americans who completed a two-item versus a single-item mock-Census form experienced greater social identity threat, institutional exclusion (e.g., threatened belonging), more negative external perceptions of Latine Americans, and lower Latine and American identification. Reactions to the citizenship question were particularly attenuated by study timepoint and sociopolitical context but also revealed a concern over confirming negative stereotypes about Latine Americans related to immigration and citizenship status. Results suggest Latine Americans face identity threat through demographic forms, and this may affect their sense of belonging within the United States.
Demographic categorizations in the U.S. Census, especially related to race/ethnicity, reflect governmental values and shape how people navigate racial and ethnic classifications. Since 1980, the U.S. Census has included one question to assess racial background (“What is Person 1’s race”) with five racial group options: White, Black or African American, Asian, American Indian or Alaskan Native, and Native Hawaiian or Pacific Islander. A separate question has been used for ethnicity (“Is Person 1 of Hispanic, Latino, or Spanish origin”), referring only to “Hispanic or Latino/a” heritage (Office of Management and Budget [OMB], 1997). The categorization structure used today reflects a sociopolitical history marred with concerns about exclusion, identity, citizenship, and belonging. 1 How might being listed separately from other groups under this classification structure on demographic forms create a sense of social identity threat for Latine 2 Americans?
We argue that the classification of “Hispanic/Latino/a” as the sole ethnic group on demographic forms is an experience of social identity threat and institutional exclusion. This framing threatens fundamental belonging needs and communicates that Latine people are not prototypical or valued members of U.S. society. Furthermore, proposals to reintroduce a citizenship question on the Census have similarly been met with concerns of questioning one’s American identity (Gamboa, 2020; Van Hook, 2019; Wang, 2018). Despite these concerns, there is a dearth of empirical data examining whether ethnicity/race framing and/or a citizenship question on demographic forms pose a social identity threat for Latine Americans.
Definitional Divides: Race and Ethnicity
Within the social sciences, race refers to a category of people who share certain inherited physical characteristics (e.g., skin color, facial features), whereas ethnicity refers to a group with shared social, cultural, and historical experiences (Barkan, 2016; Smedley, 1998, 2007). However, in practice, these definitions are inconsistently applied. For example, the U.S. racial category “Asian” refers to anyone with ancestry from any of the 48 countries in Asia (as recognized by the United Nations), encompassing wide variations in skin tones, languages, and cultural histories and practices (e.g., Espiritu, 2011; Goh et al., 2023). Nevertheless, “Asian” is designated as a race within the United States and not an ethnicity. In contrast, “Hispanic/Latino” referring to individuals with origins from Central or South America as well as the Iberian Peninsula and consisting of people with varying physical features but some shared linguistic and cultural practices is considered a panethnic group and not a racial group. Moreover, although the U.S. Census adheres to guidelines set by the OMB (1997) designating five official racial categories, previous Census forms have included “unofficially” recognized racial groups within the racial category list—but not “Hispanic/Latino” (e.g., 2020 U.S. Census race items also included “Chinese,” “Vietnamese,” “Korean,” Filipino”; see Figure 1). Together, these inconsistencies highlight the uneven application of classification standards and practices applied across different groups.
Studies 1 and 2: Ethnicity-Framing and Racial-Framing Stimuli.
Although the history of racial classification is rooted within racist ideology as a way of legitimizing prejudicial beliefs of a hierarchical classification of people (e.g., Bonilla-Silva, 1997), today, scholars recognize that racial categories reflect social rather than biological differences between groups (Fuentes et al., 2019). As such, these racial categories are socially developed constructs that can be negotiated and reinvented over time (e.g., Liebler et al., 2017).
The present work aims to explore the psychological experience of being classified within the existing structures designed to capture these identities.
Challenges of Latine Identity
Despite their ethnic classification, people with Hispanic/Latine ancestry are racialized within American society. Racialization is a social process in which racial identity is imposed onto people. Data from the Portrait of American Life Study find that although 42% of Latine people self-identify as White (when Latino/a/e was not an option), only 6% report being classified as White by other Americans (N. Vargas, 2015). This racialization process often results in unequal and discriminatory treatment (e.g., Cobas et al., 2015). Both historically and today, Latine individuals face exploitation (e.g., labor; Maldonado, 2009), economic exclusion (e.g., denied access to resources; De Haymes et al., 2000), and report high rates of discrimination (e.g., criticism for speaking Spanish, racial/ethnic slurs, microaggressions, deportation targeting; Findling et al., 2019; Lopez et al., 2018; Reichlin-Melnick, 2025). Latine Americans also experience negative stereotypes and negative health outcomes on par with other low-status racialized groups in the United States (e.g., H. E. Brown et al., 2018; Fox, 2004; Gonzales et al., 2002; Sui & Paul, 2017). Moreover, similar to research on other marginalized racial groups (e.g., Cobb et al., 2018; Meca et al., 2023; Rivas-Drake et al., 2014; Settles et al., 2010; Sellers et al., 1998), higher Latine identification is a predictor of greater well-being (e.g., French & Chavez, 2010; Torres & Ong, 2010).
Notably, many Latine Americans do not identify themselves along the U.S.-designated racial/ethnic classifications. For example, provided with racial options similar to the Census (i.e., White, Black, Asian, American Indian, something else), data from the Portrait of American Life Study finds over 50% of Latine respondents self-identify as “something else” (N. Vargas, 2015). Numerous factors may influence this identification including skin-tone, income level, racial/ethnic discrimination experience (see Golash-Boza & Darity, 2008; Stokes-Brown, 2012), immigration status, generational status, national origin (e.g., Landale & Oropesa, 2002; Jiménez, 2008), and gender (e.g., M. E. Campbell & Rogalin, 2006). Consistent with trends from previous years (e.g., Tafoya, 2004), data from the 2020 U.S. Census also show that 35.5% of Hispanic/Latine people selected “other race” and 8.1% did not respond to the race question at all (Peña et al., 2024). These patterns suggest that, perhaps due to exclusion embedded in the classification system leaving Latine respondents othered, unsure, or unwilling to respond, 43.6%, or approximately 27.9 million Hispanic/Latine people in the United States, were not accounted for within the racial categories provided by the 2020 U.S. Census (U.S. Census Bureau, 2020).
Thus, the political agenda behind the ethnic classification of “Hispanic/Latino/a,” to make “White” the de facto Latine racial categorization, appears to have fallen short (see Note 1). Some scholars have noted that the ethnic group status may even have had the opposite effect by perpetuating a Latine foreigner status (e.g., Gutierrez & Lechuga-Pena, 2022). Since 1993, the OMB has considered a “combined question” proposal to add “Hispanic/Latino” to the racial category list, and as of 2024, this proposal has been approved (Marks et al., 2024). While many groups have advocated for this restructuring (e.g., Navarro, 2023; A. Vargas, 2024), concerns have also been raised that this change risks obscuring the AfroLatine demographic, for example, who experience higher rates of discrimination compared to other Latine groups (e.g., AfroLatino Coalition, 2024; M. E. Campbell & Rogalin, 2006; Hernández, 2021). To date, there remains a dearth of empirical psychological evidence assessing personal responses to these different categorization systems.
Identity and Belonging Threats: Citizenship Questions
In 2018, the first Trump administration proposed reintroducing a citizenship question to the upcoming 2020 Census, citing in part that this question would support the Voting Rights Act and provide accurate data for policy decisions (Wang, 2018). Critics of this proposal cited concerns about data quality, predicting that the use of this question would deter participation in the Census from both documented and undocumented non-citizens (J. D. Brown et al., 2019). Indeed, by comparing response rates between the American Community Survey (including the citizenship question) and the 2010 Census (without the citizenship question), the 2020 U.S. Census may have experienced an overall 2.2% drop in participation (J. D. Brown et al., 2019). Although the Supreme Court rejected the Trump administration’s reasoning, the citizenship question continues to appear across other federal forms (e.g., American Community Survey; The White House, 2019).
Latine Americans, along with other multicultural individuals (e.g., those identifying with more than one cultural group), report being seen as “perpetual foreigners” (e.g., Sue et al., 2007). Research has shown White people are viewed as more prototypically American compared to racial/ethnic minority groups (e.g., Cheryan & Monin, 2005; Devos & Banaji, 2005). Thus, Latine Americans may experience identity denial wherein they are challenged or questioned about their identity (e.g., “Where are you really from?” Cheryan & Monin, 2005; Tsai et al., 2021). This denial has deleterious effects such as lowering sense of autonomy, increasing stress, and slowing down cortisol recovery (e.g., Albuja, Gaither, et al., 2019a; Albuja, Sanchez, & Gaither, 2019b). Thus, answering the citizenship question may be a psychologically stressful experience for Latine Americans.
Moreover, answering the citizenship question may trigger a specific type of identity threat. For example, Steele and Aronson (1995) first found that Black students taking an aptitude test experienced stereotype threat or fear of confirming a negative stereotype about one’s social group within a context-specific domain (e.g., Spencer et al., 1999). Building on this concept, the citizenship question may activate concerns about confirming negative stereotypes about Latine Americans related to immigration and documentation status. It remains unclear whether completing the citizenship question for Latine Americans creates an identity threat or not. However, given that recent increases in anti-immigration and xenophobic rhetoric have been related to increases in hate crimes (Edwards & Rushin, 2018) and prejudiced attitudes (Jardina & Piston, 2023; Newman et al., 2021), questioning of one’s citizenship status may be particularly salient.
Institutional Exclusion
Following Social Identity Theory, people define themselves in line with their social group memberships and Latine Americans may experience social identity threat when their group identity is devalued or excluded within this specific context (Crocker & Major, 1989; Tajfel & Turner, 1979, 1986). As defined by Riva and Eck (2016), social exclusion can be employed as an umbrella term referring to the array of negative experiences that cause psychological separation from others such as dehumanization, discrimination, and ostracism. We apply this framework of social exclusion to understand the experience of completing demographic forms, proposing that the formatting and the presence or absence of certain items may subtly signal exclusion. Extensive research has established that social exclusion threatens psychological needs, in particular, the need to belong, which is fundamental to group membership (Baumeister & Leary, 1995; Williams et al., 2000). Furthermore, exclusion may also threaten needs for meaningful existence (linked to feelings of dehumanization; Zadro et al., 2004), as well as needs for self-esteem, control, and leading to overall negative affect (e.g., Williams, 2009). Notably, people report feeling exclusion in cases as subtle as when conversation flow lags or even when people fail to make eye contact (e.g., Koudenburg et al., 2011; Pickett & Gardner, 2005; Wirth et al., 2010).
Institutional racism or “racially discriminatory policies and practices embedded in social institutions such as the government, the economy, the education system, the health care system, religious institutions, the family, and the media” (Needham et al., 2023) are among the nuanced ways that discrimination is practiced today (e.g., legislation and policies). Thus, studying how people respond to the presence or absence of different practices such as the framing of “Hispanic/Latino” and inclusion of a citizenship question may provide insight into the psychological effects of institutional racism (and further, structural racism). As such, we hypothesize that formatting “Hispanic/Latino” into a separate item category apart from other groups (i.e., as an ethnicity) will be experienced as social exclusion, that is, being kept apart from other groups. Similarly, we predict that the use of a citizenship question will also elicit social exclusion (i.e., threat to belonging; Riva & Eck, 2016).
In response to threat and discrimination, people may increase their group identification to maintain feelings of belonging and personal self-esteem (Rejection-Identification Model [RIM]; Branscombe et al., 1999). Thus, we further expect that when discrimination is perceived, either via exclusion from the ethnicity framing or via the use of a citizenship question, participants will report increased identification with their Latine or American identity. However, empirical support for the RIM model is mixed (e.g., Armenta & Hunt, 2009; Brugger, 2021; Romero & Roberts, 2003). An alternative perspective suggests that when faced with the uncertainty caused by discrimination, people are more likely to strongly identify with highly entitative groups—groups that are perceived as “real” by nature of having perceived clear boundaries (e.g., D. T. Campbell, 1958; Hogg et al., 2007). The lack of federal recognition of Latine American as a racial group may undermine the validity of “Hispanic/Latino” as a “real” group, which may result in lower identification with this group. Therefore, an alternative hypothesis is that Latine Americans will decrease identification with their Latine and American identities in response to exclusion cues via the mock-Census forms.
Study 1
We first test whether formatting “Hispanic/Latino” identity as a separate ethnicity (as in the 2020 Census) versus listing it alongside the other racial categories, and whether the presence or absence of a citizenship question, elicits identity threat among Latine Americans. Specifically, this study explores whether Latine Americans experience institutional exclusion via threat to one’s sense of belonging, impacting group identification. We report all manipulations, measures, and exclusions in this manuscript and the Supplemental Materials (SM). The study design and analyses were not pre-registered, but all study materials including data, codebooks, and analysis are available at Open Science Framework (OSF: https://osf.io/t857f/overview).
Method
Participants and Analysis Plan Transparency
An a priori power analysis conducted using G*Power (Faul et al., 2009) recommended 245 participants to achieve 80% power for two-way analyses of variance (ANOVAs) based on a small-medium effect size (f = 0.18, a = 0.05). Given the challenges of recruiting Latine participants, we recruited through both a university’s undergraduate participant and community SONA pool (n = 30) and Qualtrics Prime Panels (n = 242) in October 2020, prior to the 2020 Presidential election. To qualify, participants had to self-identify as Latine American and claim American citizenship (to ensure voting eligibility; see SM for additional voting-related analysis). SONA and Qualtrics samples were combined to maximize available data (see SM for additional analyses). All participants were compensated with either class credit, a $10 Amazon gift card, or an agreed upon amount in Qualtrics.
Participants were excluded from the final analysis if they did not identify as Hispanic/Latine (n = 7) or reported heritage exclusively from Spain, without any Latin American identification (n = 25). 3 As the experimental manipulation was subtle (i.e., ethnicity/race framing and citizenship question), participants were also excluded if they responded incorrectly to either of the two manipulation checks (n = 85; see section “Design and Procedure” for additional details). After exclusions, the final sample was N = 155, which resulted in limited statistical power for the planned analysis, restricting interpretability of results (f = 0.17, a = 0.05, 55.7% post hoc power; see SM, Table S2 for full model results). Study time constraints did not allow for additional data collection.
As no statistically significant interactions emerged from the planned analyses (Supplemental Table S2), we conducted independent samples t-tests to examine each manipulation separately. Two subsamples of the data were created for these analyses. One subsample (Dataset1a) tests the ethnicity/race-framing manipulation and excludes participants who failed the ethnicity/race-framing manipulation check only (n = 59; NDataset1a = 201, d = 0.50, a = 0.05, 93.4% post hoc power). The second subsample (Dataset1b) tests the citizenship question manipulation and excludes participants who failed the citizenship manipulation check only (n = 64; NDataset1b = 196, d = 0.39, a = 0.05, 77.5% post hoc power). These subsamples were analyzed independently.
Design and Procedure
Study 1 was an online experiment where Latine participants were given different versions of a mock-Census form. Using a between-subjects design, 2 (framing of “Hispanic/Latino” identity: ethnicity vs. race) × 2 (citizenship question: present vs. absent), this study had four total conditions. In the ethnicity-framing condition, “Hispanic/Latino” was listed separately from other groups, whereas in the race-framing condition, “Hispanic/Latino” was listed alongside other category options (see Figure 1). In the citizenship condition, the form either included a question about U.S. citizenship, or this question was omitted (see Figure 2).
Citizenship Question Stimuli.
Participants first viewed an 80-second informational video about the U.S. Census briefly discussing its designated purpose from the U.S. Census website (U.S. Census Bureau, 2020). Participants were then told that the researchers were interested in how people perceive the U.S. Census and that they would complete a mock-form that may reflect a past, present, or future version of the Census. Each experimentally manipulated mock-Census form included the official U.S. Census logo at the top and the instructions used by the 2020 Census, modified only to reflect the different study conditions (see Figures 1 and 2). Participants were then randomly assigned to one of the four possible conditions and completed their mock-Census form.
Immediately after completing the form, participants completed the manipulation check items. They were asked to report whether “Hispanic/Latino” had been listed as: (a) an ethnicity, (b) a racial group, and (c) do not remember. Participants were also asked whether there (a) was a citizenship question, (b) was not a citizenship question, and (c) do not remember. Participants were included in the final analysis if they answered correctly or responded, “I don’t remember.” 4 All participants then answered the same series of questions regarding psychological states.
Measures
Affect and Threatened Needs
To measure affect in response to the mock-Census forms, two measures were used. Participants answered an adapted four-item version of the State Trait Emotion Measure (Levine et al., 2011), rating how offended, stressed, annoyed, and angry they felt (1 = “Strongly disagree” to 7 = “Strongly agree”; αs > .88). In addition, the 10-item Positive and Negative Affect Scale (PANAS; Watson et al., 1988) was used to measure positive and negative affect (e.g., “Interested, Excited, Upset, Scared”; 1 = “Very slightly or not at all” to 5 = “Extremely”; αs > .80).
To measure psychological needs threatened by the mock-Census form, participants completed the Need Threat Scale (Williams, 2009, adapted two-item version). This measure included items assessing how much threat participants experienced toward their psychological needs. For belonging, items included, “I feel [rejected]; [like an outsider]”; for meaningful existence, “I feel [invisible]; [meaningless]”; for, self-esteem, “I feel [good about myself]; [that my self-esteem is high]” (reverse-coded items; Spearman–Brown coefficients > .74; Eisinga et al., 2013), and for control, “I feel [like I have the ability to alter events] (reverse-coded); [like decisions are out of my control]” (Spearman–Brown coefficients < .30). Due to low reliability, control subscale was excluded from analysis. Participants responded to all items from 1 = “Not at all” to 5 = “Extremely.”
Latine and American Identity
To assess whether the “Hispanic/Latino” framing or citizenship question affected how Latine Americans conceptualize their Latine and American identities, participants completed The Group Level Self Definition subscale measure of centrality (Leach et al., 2008). Three statements related to Latine identity and three statements to American identity (e.g., “I often think about the fact that I am [Latino/a/e; American]”; 1 =“Strongly disagree” to 5 = “Strongly agree”; αs > .85).
Participants were also assessed on the four-item Racial Collective Self-Esteem (CSE) public regard (Luhtanen & Crocker, 1992), as an exploratory measure of how they think others view Latine people (e.g., “Overall, my racial/ethnic group is considered good by others,” 1 = “Strongly disagree” to 7 = “Strongly agree”; αs > .75).
The three-item Identity Autonomy modified scale (Albuja et al., 2019b) was additionally used as an exploratory measure of how much participants felt they had the ability to racially identify as they desire (e.g., “I feel that I have a say in how I racially identify,” 1 = “Strongly disagree” to 7 = “Strongly agree”; αs > .66).
Latine and American Belonging
To assess perceived belonging, participants completed the Latinx Belonging Measure (Ocampo, 2018), which evaluates how much Latine individuals feel that they are accepted in the United States (e.g., “How strongly do you feel like you as a Latino/a/e belong in the United States [reverse coded]?” 1 =“Extremely” to 4 = “Not at all”; αs > .71). In addition, participants were also asked how much they as a Latino/a/e person perceive they are valued by other Americans (e.g., “Most Americans appreciate you as a Latino/a/e in the U.S.” [reverse coded], 1 = “Strongly agree” to 5 = “Strongly disagree”; αs > .90).
Impressions of the Census
To assess impressions of the mock-Census forms, participants were asked to rate the version of the mock-Census they saw on various positive traits (e.g., inclusionary, representative, fair, unbiased; αs > .86) and negative traits (e.g., exclusionary, offensive, discriminatory, prejudiced; αs > .85). In addition, participants responded to the statement, “If the Census looked like the example I saw, I would be inclined to complete it,” (1 = “Strongly disagree” to 5 = “Strongly agree”).
Demographic Items
Following the experimental manipulation mock-Census forms, participants reported demographic information including their age and gender identity. Racial/ethnic identity was additionally assessed (beyond the items used for the experimental purposes of the study) by responding to a single inclusive list of racial/ethnic groups, “What is your racial/ethnic background? (check all that apply).” Participants also responded to items about their heritage country, their generational status, and their political ideology (see Table 1).
Study 1 and Study 2 Participant Demographics.
Note. Demographics for t-test sample populations. abSee SM for list of racial groups and countries and additional subgroup analyses.
Results
Planned analyses are shown in Table S2 (SM). Below, we report independent sample t-tests for each experimental variable separately (ethnicity/race framing and citizenship question).
Independent Samples t-Tests: Ethnicity Versus Racial Framing
Results revealed that framing “Hispanic/Latino” identity as an ethnicity rather than a racial group resulted in significantly higher threat to belonging and meaningful existence needs, psBonferroni < .024, marginally higher negative affect, and marginally lower perceived acceptance of Latine people in the United States, psBonferroni < .073. There was no effect of “Hispanic/Latino” framing on other tested outcomes (see Table 2).
Independent Samples t-test Analysis: Study 1.
Note. Bonferroni correction used to control for family-wise error rate.
p < .08, *p < .05, **p < .01, ***p < .001.
Independent Samples t-Tests: Citizenship Question Versus No Citizenship Question
Results revealed that when the citizenship question was included rather than absent on the mock-Census form, participants reported significantly higher threatened meaningful existence, higher American identity centrality, and higher perceived valued from other Americans, psBonferroni < .043. In addition, the use of the citizenship question also resulted in marginally higher threat to belonging needs and identity autonomy, psBonferroni < .078. There was no effect of including the citizenship question on other tested outcomes (see Table 2).
Discussion—Study 1
Study 1 provides initial evidence in support of hypotheses that Latine Americans experience both the two-item ethnicity framing and the citizenship question as potential forms of institutional exclusion. The two-item ethnicity framing, compared to the single-item race framing, resulted in marginally greater negative affect and significantly higher threat to belonging and meaningful existence needs. Similarly, the presence of the citizenship question, compared to its absence, resulted in a marginally higher threat to belonging needs and significantly higher threat to meaningful existence needs. Together, this pattern of results is comparable to that of other work documenting the psychological experience of social exclusion (e.g., Williams, 2009).
We found no effect of the Latine ethnicity/race framing on either Latine or American identity centrality. Race framing was, however, associated with marginally higher perceived acceptance of Latine people in the United States. This result should be interpreted cautiously, as it may not be robust, but this might suggest that the way the U.S. Census frames “Hispanic/Latino” identity may influence perceptions of social acceptance. That is, a racial framing could lead to higher beliefs in acceptance of Latine people compared to an ethnicity framing. Results also offer partial support to previous research on RIM (Branscombe et al., 1999) in the context of the citizenship question. Specifically, marginally higher belonging threat in the presence of the citizenship question was accompanied by higher American identification, consistent with the idea that increased ingroup identification is used to buffer the negative effects of discrimination. However, this relationship is not conclusive via the current data.
Surprisingly, in the citizenship question condition, results also showed that participants reported higher perceived value of Latine people and marginally higher identity autonomy. One possible reason for this result may be that all participants were American citizens, and thus, claiming their American identity allowed them to reassert a high-status membership, which could increase identity autonomy (e.g., Sanchez et al., 2016) in line with current results. In addition, this first study was conducted during a contentious political election cycle and during the first Trump presidency, which may have influenced reactions to implications of the citizenship question.
Study 2
To explore the potential impact of broader social-political contexts on participant responses, Study 2 was conducted after the 2020 U.S. Presidential election, during the Biden presidency, compared to Study 1, which was conducted prior to the 2020 U.S. election during the first Trump presidency. Although not experimentally manipulated, this temporal comparison across sociopolitical climates provides an opportunity to test for potential identity salience effects. Furthermore, to test whether the citizenship question affects Latine people who do and do not claim American citizenship, we expanded our eligibility criteria in Study 2 to include Latine people in the United States of any citizenship status. Previous research suggests that in response to threat, individuals will perceptually enhance intragroup similarities to strengthen group boundaries, which buffers against outside stressors of threats to the group (e.g., Doosje et al., 1995; Ellemers et al., 1997). Thus, we aimed to explore additional self-identification constructs (i.e., perceived ingroup homogeneity) related to participants’ Latine and American ingroup memberships that may be affected by the racial/ethnicity framing and use of the citizenship question.
Participants and Analysis Plan Transparency
An a priori power analysis conducted using G*Power (Faul et al., 2009) recommended 199 participants to achieve 80% power for two-way ANOVAs (f = .20, a = .05). Participants were recruited and compensated as in Study 1 (university’s undergraduate and community participant pool, n = 51; Qualtrics Prime Panels, n = 168) in Spring 2021 (see Table 1 for demographics; see SM for subgroup analyses). All participants were pre-screened as Hispanic/Latine (could be in addition to any other racial/ethnic group). Claiming U.S. citizenship was not a study pre-requisite, and thus, participants were not asked to report their citizenship status (unless they were randomly assigned to this experimental condition). However, of the participants asked about their citizenship status, only 3 of the 85 asked reported that they were not U.S. citizens.
Study 2 used the same exclusion criteria as Study 1 (see SM for additional details and exclusions). Due to lower than expected final sample size and limited data collection time, statistical power was insufficient for the planned 2 × 2 ANOVA (N = 140, f = 0.17, a = 0.05, 51.5% post hoc power), constricting interpretability of results (see SM, Table S4). As no statistically significant interactions emerged from the planned analyses (see SM, Table S4), we again conducted independent t-tests to examine each manipulation separately. The first subsample (Dataset2a) was used to examine the ethnicity/race framing manipulation alone (Dataset 2a: NFraming = 143, d = .38, a = 0.05, post hoc power 72.6%) and the second subsample (Dataset2b) was used to examine the citizenship question manipulation alone (Dataset 2b: NCitizenship = 151, d = .29, a = 0.05, post hoc power 54.5%). As with Study 1, these subsamples were analyzed independently.
Design and Procedure
Study 2 followed the same experimental design and procedure as Study 1.
Measures
The same measures were used between Studies 1 and 2 (αs > .66). In addition, one new measure was added (Group Level Self Definition Measure) and the following removed for timing: PANAS, identity autonomy, and CSE Public regard.
The Group Level Self Definition Measure (Leach et al., 2008) assessed different self-identification concepts for both Latine and American ingroups such as identity centrality (e.g., “I often think about the fact that I am [Latino/a/e; American]”; αs > .75), solidarity (e.g., “I feel a bond with [Latino/a/e; American] people”; αs > .89), and satisfaction (e.g., “I am glad to be [Latino/a/e; American]”; αs > .89). In addition, items included measures of perceived self-stereotyping (e.g., “I am similar to the average [Latino/a/e; American] person”; Spearman–Brown coefficients > .83) and perceived ingroup homogeneity (“[Latino/a/e; American] people have a lot in common with each other”; Spearman–Brown coefficients > .81; 1 = “Strongly disagree” to 5 = “Strongly agree”).
Results
Planned analyses are shown in Table S4 (SM). Below, we report independent sample t-tests for each experimental variable separately (ethnicity/race framing and citizenship question).
Independent Sample t-Tests: Ethnicity Versus Racial Framing
Partially replicating Study 1, using an ethnicity-framing compared to a racial framing of “Hispanic/Latino” resulted in marginally higher threat to belonging needs, pBonferroni = .071. No other significant effects were found (see Table 3).
Independent Samples t-Test Analysis: Study 2.
Note. Bonferroni correction used to control for family-wise error rate.
p < .08, * p < .05, ** p < .01, *** p < .001.
Independent Sample t-Tests: Citizenship Question Versus No Citizenship Question
In contrast to Study 1, the presence versus absence of the citizenship question resulted in no significantly different outcomes between participants (see Table 3).
Discussion—Study 2
With the exception of marginally threatened belonging needs in the two-item ethnicity framing condition, no other effects replicated from Study 1. Between these two studies, this suggests that even across political contexts, ethnicity framing may be experienced as an institutional exclusion that threatens belonging needs, but this may not result in reconsideration of Latine and American group memberships. However, this marginal effect should be interpreted cautiously and warrants further research. Notably, there were no significant effects of the citizenship question in Study 2, which was conducted post-election during the early Biden presidency (Spring 2021). It is possible the citizenship question and its presence or absence may not have held the same social threat under a Biden presidency compared to the pre-election period under the Trump administration. Limitations of statistical power and differences between sample populations (e.g., average age) may have also contributed to these differences. Moreover, it is necessary to test the unique effects of the ethnicity/race framing and citizenship question separately, as these factors were crossed in the present study design. We assess this in Studies 3 and 4.
Study 3
Here, we explore whether exclusion-related responses to the ethnicity/race framing (two-item vs. single-item) persist when presented without the citizenship question and without the labeling of “ethnicity” and “race,” and within the context of a new election cycle (Fall 2024, Biden administration). We also measure social identity threat and beliefs about Latine Americans including perceived entitativity, prototypicality, and discrimination. This study was pre-registered, and all materials including data, codebooks, analysis code, and output are available at OSF (registration link is provided at https://osf.io/u29qk; also see SM for detailed hypotheses).
Participants
The pre-registration plan was followed for study recruitment and data exclusions, N = 350 (see Table 4 for participant demographics and SM for additional details including exclusions).
Study 3 and Study 4 Participant Demographics.
See SM for list of racial groups and countries included and subgroup analyses.
Design and Procedure
Study 3 followed the same design and procedure as Studies 1 and 2 with the following changes. Study 3 used a 2 (framing of “Hispanic/Latino” identity: single-item vs. two-item) × 2 (question labeling: label vs. no-label) between-subjects design resulting in four total conditions. In the single-item condition, “Hispanic/Latino” was listed alongside other demographic groups under one-item heading, whereas the two-item condition listed “Hispanic/Latino” under a separate item (see Figure 3). In the label condition, the form explicitly referenced the constructs (e.g., “What is your race?”), while the no-label condition omitted these terms entirely (i.e., “race” or “ethnicity” never appear on the form, see Figure 3).
Study 3: Two-Item Label and Single-Item Non-Labeled Stimuli.
As in Studies 1 and 2, participants first watched an informational video about the Census, were given instructions, and then randomly assigned to one of the four study conditions. Following completion of the mock-Census forms, all participants then answered manipulation check items and the same series of questions regarding psychological states and their perception of Latine Americans and their identification.
Measures
The following measures from Studies 1 and 2 were included for Study 3: State Trait Emotion (Levine et al., 2011), Need Threat Scale (Williams, 2009), Census impressions, adapted version of Latinx Acceptance and Value (Ocampo, 2018; e.g., “Latino/a/es belong; are appreciated in the U.S.”), Group-level self-definition (Leach et al., 2008), CSE Public regard (Luhtanen & Crocker, 1992); and identity autonomy (Albuja et al., 2019b; αs > .73; Spearman–Brown coefficients > .91). The following new measures were included (see SM for additional measures and results of political engagement).
Social Identity Threat
The Generalized Social Threat Measure (Ma et al., 2024) was adapted to measure direct social identity threat to both American identity and Latine identity (e.g., “This form is threatening to me as an [American; Latino/a/e]”) (1 = “Strongly disagree” to 7 = “Strongly agree”; α > .94)
Beliefs About Latine People
Items were adapted from Blanchard et al. (2020) to measure entitativity beliefs (e.g., “Latino/a/es are a unit”). In addition, two new items were created to measure perceived prototypicality of Latine people in the United States (e.g., “Latino/a/es are typical members of U.S. society”). Another two new items were created to measure perceived discrimination of Latine people in the United States (e.g., “Latino/a/es are discriminated against in the U.S.”). Finally, items were adapted from Bedyńska et al. (2019) to assess stereotype threat regarding citizenship status (e.g., “People assume Latino/a/es are not in the U.S. legally”). All items were scored 1 = “Strongly disagree” to 7 = “Strongly agree”; αs > .83; Spearman–Brown coefficients >. 87.
American and Latine Identification
Identity centrality was measured via a subscale of the Collective Self-Esteem (CSE) measure (Luhtanen & Crocker, 1992; e.g., “The racial/ethnic group I belong to is an important reflection of who I am”; 1 = “Strongly disagree” to 7 = “Strongly agree”; α > .77). Ethnic pride was measured with adapted items from Hernández et al. (2017) (e.g., “You like for people to know where your family is from”; 1 = “Strongly disagree” to 7 = “Strongly agree”; α > .86).
Demographic Items
For additional demographic variables (asked after the experimental mock-Census form manipulation), participants reported their age and gender identity and their racial/ethnic identity using an inclusive checklist. In addition, participants completed the same items about their parents’ racial/ethnic backgrounds and responded to items about their heritage country, their generational status, and their political ideology (see Table 4).
Results and Discussion
For the following, 2 × 2 multivariate analysis of variance (MANOVA) tests were first run to establish the omnibus tests for conceptually related dependent variables for hypotheses (see SM for details).
2 × 2 ANOVAs: Social Identity Threat and Exclusion
Two-way ANOVAs were run to test hypotheses that a two-item framing demographic form, compared to a single-item form, creates social identity threat and institutional exclusion, regardless of “ethnicity” or “race” labels (see Table 5). Supporting pre-registered hypotheses, participants in the two-item framing condition, compared to the one-item framing condition, reported significantly higher generalized social identity threat, higher threat to psychological needs (i.e., belonging, meaningful existence, self-esteem, control), and lower positive affect and impression of the Census. These effects were consistent regardless of “ethnicity” or “race” labels. These findings suggest that the structural separation of one identity category from all others in two-item demographic formats, not the specific labels, create a psychologically exclusionary experience for Latine Americans and threaten both Latine and American social identity.
2× 2 ANOVA: Framing (Single-Item vs. Two-Item Framing) × Label (Label vs. No-Label).
p < .08, *p < .05, **p < .01, ***p < .001.
2 × 2 ANOVAs: Beliefs About Latine People
Two-way ANOVAs were run to test hypotheses that the two-item framing, being more exclusionary than the single-item framing, would result in lower perceptions of Latine entitativity, prototypicality, acceptance, value, and higher perceived discrimination, regardless of label condition (see Table 6). Supporting pre-registered hypotheses, participants in the two-item framing condition, compared to the one-item framing condition, reported significantly lower perceived entitativity (i.e., viewing Latine people as a less cohesive group), lower perceived prototypicality of Latine people in the United States, and lower perceived acceptance and value of Latine people in the United States. Furthermore, participants also reported higher perceived discrimination and higher perceptions of stereotype threat related to citizenship status compared to those in the one-item framing condition, and all these effects were consistent regardless of “ethnicity” or “race” labels. Taken together, results suggest that the two-item framing not only threatens psychological needs and social identity, but it also shapes perceptions of Latine people in society. Specifically, the two-item framing prompts Latine Americans to perceive “Latinos” as a socially illegitimate group that is also seen as non-prototypical and unaccepted and unvalued within U.S. society.
2 × 2 ANOVA: Framing (Single-Item vs. Two-Item Framing) × Label (Label vs. No-Label).
p < .08, *p < .05, **p < .01, ***p < .001.
2 × 2 ANOVAs: Latine and American Identity
Two-way ANOVAs were run to test whether the two-item framing would also impact Latine and American group-level identification (see Table 7). In sum, the two-item framing, compared to the single-item framing, resulted in significantly lower collective-self-esteem public regard, identity centrality, and identity autonomy. In addition, among Latine identity measures, the two-item framing resulted in significantly lower identity satisfaction and perceptions of self-similarity, and marginally lower identity solidarity and higher perceived intragroup homogeneity. Among American identity measures, the two-identity framing resulted in significantly lower identity solidarity, satisfaction, perceived self-similarity, and intragroup homogeneity.
2 × 2 ANOVA: Framing (Single-Item vs. Two-Item Framing) × Label (Label vs. No-Label).
p < .08, *p < .05, **p < .01, ***p < .001.
These findings support an integrated bicultural identification framework (e.g., Meca et al., 2020) and are in line with previous research that people will identify less strongly with less entitative groups (two-item condition, e.g., D. T. Campbell, 1958; Hogg et al., 2007). This suggests using the two-item framing could negatively influence Latine American’s level of internal commitment to their Latine and American group memberships. Interestingly, perceived intragroup similarity (homogeneity) of both Latine (marginally) and American ingroups was highest in the two-item framing condition (e.g., “American people have a lot in common with each other”). This is consistent with work showing that people cognitively enhance intragroup similarities in response to threat to protect group boundaries (Doosje et al., 1995; Ellemers et al., 1997).
Finally, there were also main effects of the label use such that American identity satisfaction (marginally) and perceived group similarity (homogeneity) were lower when a label (e.g., “ethnicity” or “race”) was present. This result suggests the presence of these labels alone led participants to view Americans as less similar to one another. Overall, Study 3 supports previous research that Latine people see their Hispanic/Latine heritage as a part of both their ethnicity and race (Parker et al., 2015), such that the majority of effects in the present study did not depend on the specific label applied to the framing.
Study 4
In Study 4, we explore whether exclusion-related responses of the citizenship question persist when presented without questions about ethnicity/race, and within the context of a new election cycle (Fall 2024, Biden administration). We also measure social identity threat and beliefs about Latine Americans including perceived entitativity, prototypicality, and discrimination. This study was pre-registered, and all materials including data, codebooks, analysis code, and output are at OSF (registration link is provided at https://osf.io/t36qh; see SM also for detailed hypotheses).
Participants
The pre-registration plan was followed for study recruitment and data exclusions, N = 362 (see Table 4 for participants’ demographics and SM for additional details including exclusions).
Design and Procedure
Study 4 was an online study where participants were randomly assigned to receive one of three possible mock-Census forms: (a)“Respond-yes”—Participants were instructed to complete the form with the citizenship question (all participants indicated they were citizens); (b) “View-only”—Participants instructed to view the form and not complete it; and (c) “No Citizenship Question”—Participants received a form with no citizenship question (see Figure 4). No conditions contained questions about race/ethnicity.
Study 4: Citizenship Question Stimuli.
As in Studies 1, 2, and 3, participants first watched an informational video about the Census, were given instructions, and then randomly assigned to their study condition. Following exposure to the mock-Census, all participants completed manipulation check items and answered the same series of questions regarding psychological states and their perception of Latine Americans and their identification.
Measures
All measures from Study 3 were used in Study 4, αs > 79; Spearman–Brown coefficients >. 92.
Results and Discussion
In the following, 2 × 2 MANOVA tests were first run to establish the omnibus tests for conceptually related dependent variables for hypothesis (see SM for details).
One-Way ANOVAs: Social Identity Threat and Exclusion
One-way ANOVAs were run to test hypotheses that the citizenship question creates social identity threat and institutional exclusion. Results for all dependent variables showed no significant effects of study condition (see Table 8). Completing the citizenship question, compared to simply viewing it or not receiving it, did not result in any difference in reported direct threat to social identity. Replicating Study 2, there were no effects to psychological needs (belonging, meaningful existence, self-esteem, control), or changes in impressions of the Census. This suggests that other factors unique to Study 1, such as the sociopolitical climate or the recency and/or salience of the citizenship question debate and 2020 U.S. Census roll-out, may have influenced those specific results. As such, the citizenship question, at least when presented without the ethnicity/race framing, may not create a generalized threat to social identity or a sense of exclusion.
One-Way ANOVAs Citizenship Question.
Note. Superscripts mark significantly different means. Post hoc t-tests: Entitativity: t(359) = −1.83, p = .161; t(359) = 1.84, p = .158; t(359) = 3.63, p < .001. Perceived valued: t(359) = −0.83, p = .687; t(359) = 1.61, p = .244; t(359) = 2.42, p = .043. Stereotype threat: t(359) = −0.31, p = .949; t(359) = 3.37, p = .002; t(359) = 3.67, p < .001. Ethnic pride: t(359) = 1.56, p = .265; t(359) = −1.26, p = .422; t(359) = −2.78, p = .016. Latine solidarity: t(359) = −0.11, p = .993; t(359) = 2.43, p = .042; t(359) = 2.54, p = .031; Latine ingroup homogeneity: t(359) = −0.18, p = .983; t(359) = 2.20, p = .073, t(359) = 2.37, p = .048.
p < .08, *p < .05, **p < .01, ***p < .001.
One-Way ANOVAs: Beliefs About Latine Americans and Latine and American Identity
One-way ANOVAs were run to test hypotheses that responding to or viewing the citizenship question would differently affect perceptions of Latine entitativity, prototypicality, value, acceptance, stereotype threat, and group-level identification. Results showed significant effects for entitativity, stereotype threat, ethnic pride, Latine solidarity, and Latine ingroup homogeneity, and marginal effects of perceived value (see Table 8).
Participants in the “view-only” condition, who saw but did not complete the citizenship question, reported higher perceived entitativity, a trend toward higher perceived value of Latine people in the United States (marginal effect), but lower ethnic pride compared to those who did not receive the citizenship question. However, both the “respond-yes” and “view-only” conditions led to higher stereotype threat concerns, a specific type of social identity threat (e.g., “People assume Latino/a/es are not in the U.S. legally”) for Latine Americans compared to the no question condition. Clarifying Study 1, the present study suggests that viewing the citizenship question may reinforce external perceptions of “Latinos” as an entitative cohesive group (that may be valued by others), while at the same time undermine internal values about Latine group membership by lowering ethnic pride and creating stereotype threat. This distinction underscores a seemingly contradictory relationship between external perceptions and personal identification for Latine Americans. Yet, further research is needed to understand this relationship as well as the psychological differences between “viewing-only” and “responding” to the citizenship question. Given no differences in identity autonomy were observed between these conditions (e.g., Sanchez et al., 2016) and both conditions elicited stereotype threat, “responding-yes” to the citizenship question may not be a form of identity reassertion for Latine Americans who claim U.S. citizenship.
Moreover, in contrast to pre-registered hypotheses and results from Study 1, the citizenship question did not affect identification with an American ingroup. Instead, participants responded to the presence of the citizenship question (“respond-yes” or “view-only”), compared to its absence, by increasing their Latine solidarity and perception of Latine intragroup similarity. This suggests participants perceived the citizenship question as relevant to their Latine identity (e.g., Meca et al., 2020). As some form of the citizenship question led to both higher perceived stereotype threat and entitativity, additional research is needed to elucidate if related increases in identity solidarity stem from attempts to protect belonging needs (RIM; Branscombe et al., 1999; Study 1) and/or aim to reduce uncertainty (e.g., Hogg et al., 2007) as observed in Study 3. Finally, similar to Study 3, participants who reported higher stereotype threat also reported higher perceived ingroup similarity, suggesting cognitive strategies being employed to protect ingroup boundaries (e.g., Doosje et al., 1995; Ellemers et al., 1997).
General Discussion
Across four studies, we demonstrate that both the two-item ethnicity framing used in the 2020 U.S. Census and the proposed citizenship question contribute to social identity threat and institutional exclusion for Latine Americans. Utilizing the widely adopted ethnicity/racial framework, we find that presenting “Hispanic/Latino” identity with this two-item (listed separately from other categories)format,compared to a single-item (combined) format on a mock-Census form resulted in threats to both Latine and American social identity. Furthermore, the two-item format created an experience of institutional exclusion via threats to psychological needs such as belonging, meaningful existence, self-esteem, and control. These effects occurred regardless of whether forms used “ethnicity” or “race” labels, suggesting that the structural separation of “Hispanic/Latino” from other groups on these forms signals this exclusion.
Moreover, results also revealed that two-item framing affected Latine Americans’ perception of Latine people’s legitimacy as a group and belonging within American society. Participants exposed to the two-item format perceived “Hispanic/Latino” as a socially devalued group, reporting lower perceptions of Latine entitativity (i.e., group cohesion), as well as lower prototypicality, acceptance, and value within U.S. society. Furthermore, two-item framing also negatively impacted Latine and American group-level identification, which might suggest that beliefs about groups, such as perceived entitativity, may influence group identification (e.g., Hogg et al., 2007). These findings have important implications for Latine collective action. For example, institutional signals of exclusion or illegitimacy may weaken internal identification and in turn inhibit motivation for group-based engagement. The present work underscores the need for inclusive demographic practices and offers support for the recently accepted proposal to update the 2030 U.S. Census to reformat the questions about ethnicity and race into a single combined item (“What is your race or ethnicity?”; Marks et al., 2024).
Study results also found that the proposed citizenship question resulted in higher perceived stereotype threat related to immigration status. Moreover, the use of the citizenship question led to increased perceptions of Latine entitativity, value, and identification while simultaneously lowering ethnic pride. These results provide further evidence of how Latine identification may be shaped by perceived group entitativity (Hogg et al., 2007), which may both increase in response to threat (Branscombe et al., 1999). This pattern also suggests a complex dynamic wherein the citizenship question may enhance external perceptions of Latine cohesion while simultaneously undermining internal ethnic pride and triggering stereotype threat. Additional research is needed to elucidate differences between responding to and viewing the citizenship question and further if these results generalize to Latine people of different immigration and citizenship statuses.
Notably, results were attenuated by the social-political context (election season), particularly for the citizenship question. That is, participants also reported threats to belonging and meaningful existence in response to the citizenship question in Study 1 (during the 2020 election) not seen in Study 4 (during the 2024 election). These temporal differences underscore how changes in the sociopolitical context can shape the psychological salience of identity threat. The differences observed here may reflect changes in attention and discussion around the proposed citizenship question (2020 vs. 2024). More broadly, political election seasons may coincide with increased focus and attention on policy debates in general. However, discrepancies may also reflect broader shifts in political discourse and public attention to issues such as immigration specifically or overall changes in the electorate and policies across this time. As the current administration makes new proposals for the Census to exclude the count of those living in the United States without legal status (e.g., Wang, 2025), the present results underscore the psychological importance of an inclusive approach to large-scale categorical efforts (e.g., Wang, 2024). In sum, our findings suggest that the two-item framing communicates institutional exclusion, threatening psychological needs, whereas the citizenship question evokes stereotype-based identity threat rooted in concerns about immigration status.
Limitations and Future Directions
It is important to note that racial categorization as a construct was developed to reinforce racist ideologies and today does not reflect the current scientific understanding of biological (or lack thereof) differences between groups of people (e.g., Fuentes et al., 2019). As such, the present research reflects the experiences of some Latine groups, but it remains unclear whether AfroLatino people (as one example) would have the same psychological experience given concerns expressed over the changes to the 2030 Census (AfroLatino Coalition, 2024; see SM for additional exploratory analysis). Similarly, indigenous Latine people may face different forms of exclusion, miscategorization, and erasure. As such, while our results highlight the consequences of excluding Latine Americans from the current categorization systems (i.e., ethnicity-framing), they do not necessarily mean that a single-item categorization system is the most inclusive option. Additional research should also test whether findings extend to other institutions that use similar categorizations.
With the exception of small group subpopulation analyses (see SM), the current research is unable to identify high-powered subgroup differences between Latine Americans of different racial backgrounds, generational statuses, and countries of origin. Previous research shows these factors are predictive of racial identification and vary meaningfully across groups (e.g., Golash-Boza & Darity, 2008; Stokes-Brown, 2012). For example, individuals from countries with differing racial classification systems (e.g., Argentina vs. Brazil vs. Puerto Rico) may respond to Census items differently.
Conclusion
The current research explored Latine American’s experience of identity threat via institutional exclusion on a mock U.S. Census form. Overall, the commonly used two-item (ethnicity-framing) of “Hispanic/Latino” elicited higher threat to social identity, psychological needs (e.g., belonging, meaningful existence, control), resulted in negative perceptions of how Latine people are viewed in the United States, and decreased Latine and American identification. Although the proposed citizenship question was particularly threatening to psychological needs during the 2020 election season, it also triggered stereotype threat around immigration status and increased perceived entitativity and Latine identification. Taken together, these results suggest that current and proposed demographic categorization practices can have unintended psychological consequences for Latine Americans navigating questions of identity and belonging. These insights underscore the need for inclusive approaches in demographic enumeration practices not only in the Census but also across institutionally created forms.
Supplemental Material
sj-docx-1-psp-10.1177_01461672251387589 – Supplemental material for Enumeration or Exclusion? Demographic Forms and Latine Identity Threat
Supplemental material, sj-docx-1-psp-10.1177_01461672251387589 for Enumeration or Exclusion? Demographic Forms and Latine Identity Threat by Brenda C. Straka, Miguel Martinez and Sarah E. Gaither in Personality and Social Psychology Bulletin
Footnotes
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was in part funded by the Society for the Psychological Study of Social Issues as well as the Charles Lafitte Foundation.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
All data, analysis code, pre-registrations, and research materials are available on OSF: https://osf.io/u29qk and 
.
Supplemental Material
Supplemental material is available online with this article.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
