Introduction
Whose rights do we recognize? Liberal democratic states are expected to provide residents with protections, which include both negative (or civil) rights, such as limitations against the abuse of police powers, and some range of positive (or social) rights, such as access to public benefits. Indeed, various thinkers have argued that civil, political and social rights, together, are indispensable to full citizenship or membership in a polity (Marshall, Reference Marshall1950). Such rights are commonly deemed to apply equally, without respect to individuals’ ascriptive backgrounds. Thus, in the words of the Canadian Charter of Rights and Freedoms, “Every individual is equal before and under the law and has the right to the equal protection and equal benefit… in particular, without discrimination based on race, national or ethnic origin, colour, religion, sex, age or mental or physical disability” (Section 15(1)). Further, while the Charter links certain rights to citizenship, such as the vote (Section 3) and the ability to enter, remain in and leave Canada (Section 6), other provisions apply to “everyone,” such as “life, liberty and security of the person” (Section 7) or “the right to be secure against unreasonable search or seizure” (Section 8).
Research has shown, however, that while the general public might support universal equal rights and protections in the abstract, the public regularly expresses categorical inequality in the application of rights to some groups relative to others (Brooks and Manza, Reference Brooks and Manza2007). Categorical inequalities are those that divide people by socially recognized traits that form the basis for persistent economic and political disparities, such as gender, race, immigrant status or class origins (Tilly, Reference Tilly1998; Massey, Reference Massey2007) producing status hierarchies. Such inequalities are well studied in the US context, with a particular focus on race. The US public is especially likely to view Black Americans as less deserving of social rights (Gilens, Reference Gilens2009) but also to view Latinos unfavourably (Fox, Reference Fox2004). In terms of the latter group, there is a lively debate among researchers as to whether ethnoracial minority background and immigration status are conflated to produce sentiments of racialized illegality irrespective of actual citizenship or legal status (Pérez, Reference Pérez2016; Masuoka and Junn, Reference Masuoka and Junn2013) or, instead, whether the US public holds genuine moral beliefs about legal status, including around what Levy and Wright (Reference Levy and Wright2020) call “civic fairness.” Civic fairness views ascriptive distinctions based on ethnoracial background as wrong but not distinctions between citizens and legal immigrants, on the one hand, and undocumented or “illegal” immigrants, on the other.
Such questions about categorical inequalities and rights recognition have been much less studied in Canada, especially at the intersection of race and legal status. Canada is, of course, home to many fewer undocumented immigrants than the United States. But empirically, there are perhaps hundreds of thousands of people without status living in Canada (Smith and Kim, Reference Smith2022: 2), and many more hold temporary, precarious visas as workers or students or have uncertain status as asylum-seekers.Footnote 1 Theoretically, examining the intersection of race and legal status in public attitudes can provide insight into the construction of membership boundaries in a diverse country like Canada. Does citizenship status overcome the categorical inequalities of race? Further, while we have some evidence that, in Canada, would-be beneficiaries’ race, ethnicity, indigeneity and immigrant background affect opinion on social rights and benefits (Harell et al., Reference Harell, Soroka and Iyengar2016, Reference Harell, Banting, Kymlicka and Wallace2022), we know almost nothing about the effect of race and legal status on civil rights (though see Banting et al., Reference Banting, Harell and Kymlicka2022). Scholars of public opinion have tended to study public attitudes toward rights in silos that match the negative and positive rights distinction (though see Sniderman et al., Reference Sniderman, Fletcher, Tetlock and Russell1996), with greater attention to redistribution or social rights. We ask: does the Canadian public make distinctions about what sorts of people should benefit from what sorts of rights?
Surprisingly, little empirical work, in Canada or other liberal democracies, directly compares categorical inequalities in public opinion across negative and positive rights. An older literature has investigated public views on civil liberties, such as freedom of speech, assembly and protection against the arbitrary use of state powers (for a review, see Sullivan and Transue, Reference Sullivan and Transue1999), but it has not delved into differential public support for negative rights based on rights-holders’ ascriptive background or legal status. Thus, despite a robust literature on attitudes toward racial profiling, mostly drawing on US data, few studies directly compare how citizens view the same police action when the target is from a higher or lower status group (though see Peffley and Hurwitz, Reference Peffley and Hurwitz2010). After the 9/11 terrorist attacks, some surveys revealed the public's willingness to deny basic civil liberties to people of Muslim faith and/or Arab descent (for example, Davis, Reference Davis2007; Schildkraut, Reference Schildkraut2009), but overall, empirical research across liberal democracies has overlooked the intersection of rights, race and non-citizenship in public opinion, as well as the question of whether the public's views of who enjoys rights carry across the negative/positive rights distinction.Footnote 2
In this article, we investigate the degree to which the public draws membership boundaries, based on race and legal status, around who can enjoy civil and social rights. We leverage two separate vignette survey experiments conducted in Canada in 2020 and 2022 that vary ethnoracial background and consider extremes in legal status: full citizens versus those without legal status on the territory. Our findings suggest that intersectional hierarchies, based on a combination of race and legal status, influence Canadian attitudes. Intriguingly, the ways in which categorical inequalities play out vary across social rights and civil rights (or what many Canadians might call Charter rights).Footnote 3 With respect to social rights, White citizens tend to have an advantage, with more people agreeing that the government should provide someone going hungry with extra assistance, although the effect disappears when the White citizens’ origins are specified. For Black individuals, holding Canadian citizenship reduces but does not always eliminate differences in support for social benefits compared to White citizens. In all cases, non-citizens without status are heavily penalized, showing how legal status affects social rights. This pattern varies for negative rights. Here, respondents are more likely to interpret a policing scenario with a Black citizen as being unfair and requiring action, suggesting that the public's interpretation of negative rights violations may be viewed more easily as a racial inequity issue. But again, citizenship matters: respondents are more willing to support action being taken for Black citizens than non-citizens, even if Charter provisions enshrine equal protections for “everyone.” These survey data also challenge the notion that international rights regimes and human rights norms have decoupled rights from citizenship (Soysal, Reference Soysal1994; Jacobson, Reference Jacobson1997), at least in the minds of ordinary Canadians.
Which Rights and For Whom?
There is an extensive social science literature evaluating public opinion toward redistribution and civil liberties, but relatively little of the empirical work engages the more normative literature on the nature of rights and their relationship to inclusion and inequality. Such norms imbue our legal structures, like the Charter, and political rhetoric. As Bauböck underscores, political science must deal with normative questions not just as a matter of research context, but also research content, such that “the attempt to purify the social sciences of normative questions is misguided” (Reference Bauböck, Porta and Keating2008: 48). A goal of this article is to bring ideas from political theory in conversation with careful empirical work on public opinion. We thus hope to respond, in a modest way, to calls to inform controversial political topics such as migration with sustained dialogue between “realist” empirical social sciences and “idealist” normative theory (Carens, Reference Carens1996; Bauböck, Reference Bauböck, Porta and Keating2008). Normative theory provides rich conceptualizations of rights, be they negative, positive or human rights, and it helps elucidate the complexity and contested nature of rights claims, especially in relation to migration status. At the same time, as Carens (Reference Carens1996) cautions, an idealist approach to normative theory, one inattentive to the facts on the ground, risks making the theorist irrelevant in helping to answer the question of how to act in a non-ideal world. We aim to bring real-world data into this conversation, documenting how categorical inequalities influence public opinion in Canada.
Rights and Membership
In debates on rights, we highlight long-standing distinctions between negative (civil) and positive (social) rights. Whereas “[t]he holder of a negative right is entitled to non-interference” or freedom from the state, “the holder of a positive right is entitled to provision of some good or service,” that is, a claim to some social right or public benefit (Wenar, Reference Wenar and Zalta2021: section 2.1.8). For some, negative rights are “basic” civil liberties, more foundational than social rights; the former are also arguably less costly for governments to provide than the latter. As such, residents of liberal democracies might support and extend (at least in principle) negative rights broadly, to everyone, even as they may practise, and even defend, welfare chauvinism around social rights, only allocating public benefits to members. Other scholars, however, question negative and positive rights distinctions, arguing against the idea that one is easier or less costly than the other or that one should be circumscribed to a certain group of people. For these scholars, “basic” rights should entail, jointly, the right to physical security, a minimum level of sustenance, and right to liberty (Shue, Reference Shue1996), such that there may be “just as strong a moral case for fulfilling a person's right to adequate nutrition as there is for protecting that person's right not to be assaulted” (Wenar, Reference Wenar and Zalta2021: section 2.1.8). The United Nations Declaration of Human Rights contains both, although it does enumerate more negative rights (for example, freedom from torture, right to assembly, equality before the law, presumption of innocence) than positive ones (for example, the right to an adequate standard of living).
In tension with the UN Declaration, many citizenship theorists have bundled civil and social rights, but rather than viewing them as universal, they link them together through common citizenship. T. H. Marshall famously argued that “basic human equality… has been enriched with new substance and invested with a formidable array of rights. It has been clearly identified with the status of citizenship” (Reference Marshall1950: 9). Citizens’ equality (and class inclusion) was realized, according to Marshall, through the state's extension of civil, political and social rights over time. Philosopher Will Kymlicka (Reference Kymlicka2015: 4) articulates the membership challenge as one of justice and solidarity such that “justice to strangers is humanitarian” but social justice and equality among members of the nation-state “is rooted in an ethic of social membership” requiring social rights and a robust welfare state. There is thus theoretical disagreement about the degree to which civil and social rights should be thought of as distinct, and whether such rights are universal human rights or more properly bounded by the nation-state, applicable only to citizen members.
Parallel to theorists’ debates, a voluminous empirical literature has investigated the degree to which the public actually supports civil or social rights. Classic studies of public opinion in liberal democracies probe whether people would permit groups that they do not like to speak out or to protest, and whether governments or police may violate civil liberties in particular circumstances (Stouffer, Reference Stouffer1955; McClosky and Brill, Reference McClosky and Brill1983; Sullivan et al., Reference Sullivan, Piereson and Marcus1993; Sniderman et al., Reference Sniderman, Fletcher, Tetlock and Russell1996). The upshot is that ordinary people—and a substantial number of elites—support violating civil rights in a range of specific situations even as they back civil rights in the abstract. For example, asked shortly after the new Canadian Charter entered into force whether they found it reasonable or a violation of a person's rights for the police to search “a young man seen near a house where the police know that drugs are sold,” 59 per cent of Canadian voters (and 57 per cent of elites) found a police search reasonable in 1987 (Sniderman et al., Reference Sniderman, Fletcher, Russell, Tetlock and Gaines1991: 358). More recent data, from 2017, finds that Canadians are virtually indistinguishable from Americans in the percentage who agree that the government can record telephone calls and emails without a warrant (31 and 30 per cent, respectively) or that the police can search property without a warrant (22 and 21 per cent, respectively) (Harell and Hinckley, Reference Harell and Hinckley2017).
A more voluminous literature explores the contours of public opinion toward social rights, redistribution and the welfare state (Brooks and Manza, Reference Brooks and Manza2007; Van Oorschot et al., Reference Van Oorschot, Roosma, Meuleman and Reeskens2017). As with views on negative rights, scholars have found that general support for helping people in poverty or facing hardship can dissipate when respondents are asked about social benefits provided to distinct sub-groups of individuals (Gilens, Reference Gilens2009; Harell et al., Reference Harell, Soroka and Iyengar2016). Two of the most studied target groups—ones for which majority populations express greater hesitancy in providing social rights—are racialized minorities and immigrants.
Categorical Inequality and Rights: Race and Legal Status
Whose rights does the public recognize? To tackle this question, we mobilize the concept of categorical inequality (Tilly, Reference Tilly1998; Massey, Reference Massey2007). By “category,” we refer to processes by which people are placed into social groups or given labels that tend to homogenize those within the group, overshadowing “individual differences in attributes, propensities, or performances,” while also obscuring group members’ similarities to those outside the group (Tilly, Reference Tilly1998: 7; see also Krueger, Reference Krueger, Smelser and Baltes2001). Social categorization can be horizontal, without inherent hierarchy, such as in distinguishing “Canadians” from “Americans,” and it can be temporary, with modest consequences, as when social psychologists conduct minimal group experiments. However, when there are durable inequities and vertical status hierarchies in treatment, resources, or perceptions between social categories, such as between “citizens” and “non-citizens,” we speak of categorical inequality.Footnote 4 Scholars debate the mechanisms that produce categorical inequality, which might lie in social closure, institutionalization and exploitation (Tilly, Reference Tilly1998; Masuoka and Junn, Reference Masuoka and Junn2013); inherent human tendencies toward cognitive grouping and in-group preference combined with socialization (Fiske and Taylor, Reference Fiske and Taylor2013; Massey, Reference Massey2007; Pérez, Reference Pérez2016; Rhodes and Baron, Reference Rhodes and Baron2019); or the overlay between cultural schema and social inequality (Lamont and Molnár, Reference Lamont and Molnár2002).
We focus not on the mechanisms producing categorical inequalities, but the possible consequences of social categorization into two widely recognized, recurring bases of categorical inequality: ethnoracial background and citizenship/legal status. Charles Tilly, drawing on a plethora of historical data, argued that across time and place, “we discover and rediscover paired, recognized, organized, unequal categories such as black/white, male/female… and citizen/noncitizen” (Reference Tilly1998: 8). These categories structure the way people think about who deserves to enjoy the full benefits of membership in a state. They are also viewed somewhat differently by political theory. The Black/White distinction is an ascriptive category grounded in processes of racialization that is overwhelmingly perceived as an illiberal basis on which to make distinctions. The citizen/non-citizen distinction is a legal category that, although historically imbued with explicit racial biases, is viewed by many—but not all—liberal theorists as a legitimate rights-marker in some circumstances.Footnote 5
The literature on race and social benefits is particularly vast, notably in the United States, where Americans’ support for welfare shows strong racial bias, especially but not exclusively toward African Americans (Bobo, Reference Bobo1988, Reference Bobo1999; Gilens, Reference Gilens1995, Reference Gilens1996, Reference Gilens2009; Fox, Reference Fox2004). There is also growing evidence in Canada demonstrating some forms of bias, notably that anti-Indigenous bias moderates support for cash benefits for Indigenous people (Harell et al., Reference Harell, Soroka and Ladner2014, Reference Harell, Soroka and Iyengar2017). Little evidence was found, however, of a corresponding effect toward Black Canadians (Harell et al., Reference Harell, Soroka and Iyengar2017). This is not to say that racial attitudes do not matter for public opinion in Canada (see, for example, Besco and Matthews, Reference Besco and Matthews2022) or that racial inequality in Canada is not a persistently experienced phenomenon, whether we consider economic outcomes, politics or interactions with police and the criminal justice system (Banting and Thompson, Reference Banting and Thompson2021; Henry et al., Reference Henry, Rees and Tator2010; Tolley, Reference Tolley2016).
With respect to immigration background and legal status, research indicates that, in Europe, the public tends to discriminate against foreign-born people in allocating social benefits (Reeskens and van Oorschot, Reference Reeskens and van Oorschot2012; Reeskens and van der Meer, Reference Reeskens and van der Meer2019). In the United States, researchers also report that legal status—with a particular focus on undocumented status—shapes survey respondents’ attitudes toward immigration (Levy and Wright, Reference Levy and Wright2020) as well as their willingness to extend social rights (Brown and Jones, Reference Brown and Jones2016; Bloemraad et al., Reference Bloemraad, Silva and Voss2016). But exclusions are not just based on irregular status. Bloemraad, Silva and Voss (Reference Bloemraad, Silva and Voss2016: 1663) report that sizable groups of registered voters in California wanted to restrict welfare benefits and food assistance to only citizens, 40 and 36 per cent of respondents, respectively, excluding even legal permanent residents.
There is less research on the effects of legal status on public opinion in Canada, perhaps due to less contested views about immigration, lower numbers of out-of-status residents, and the generally high human capital profile of newcomers. Scholars have looked at media debates about extending government-funded health insurance to refugees (Villegas and Blower, Reference Villegas and Blower2019) as well as the relationship between the public's attitudes toward immigration and the welfare state, finding that those most supportive of redistribution are also those most willing to share social benefits with immigrants (Soroka et al., Reference Soroka, Wright, Bloemraad, Johnston, Goodyear-Grant, Johnston, Kymlicka and Myles2018). Recent work also explores support for immigrants’ right to protest, showing baseline gaps in support (Banting et al., Reference Banting, Harell and Kymlicka2022). However, to our knowledge, no one has yet examined Canadian public opinion on extending social and civil rights to non-citizens in a way that distinguishes immigration status from ethnoracial origin.
Further, compared to the work on social rights, there is less research on categorical inequalities and the application or enforcement of civil liberties, a gap we seek to fill. The research that has been done raises the possibility of racial, religious and non-citizen biases. In the United States, there is evidence of racial bias in support for the death penalty (Soss et al., Reference Soss, Langbein and Metelko2003; Bobo and Johnson, Reference Bobo and Johnson2004; Peffley and Hurwitz, Reference Peffley and Hurwitz2010); that White people evaluate Black suspects more negatively (Peffley et al., Reference Peffley, Hurwitz and Sniderman1997); and that White judges tend to be harsher in their punishments of Black defendants (Welch et al., Reference Welch, Combs and Gruhl1988). There are hints of similar patterns around Whites’ attitudes toward undocumented migrants in the US (Haynes et al., Reference Haynes, Merolla and Karthick Ramakrishnan2016; Pérez, Reference Pérez2016; Wright et al., Reference Wright, Levy and Citrin2016; Levy and Wright, Reference Levy and Wright2020) or toward Arab Americans, Muslims and Middle Easterners after the 2001 terrorist attacks (Davis, Reference Davis2007; Schildkraut, Reference Schildkraut2009), but little of this research tests opinion on civil liberties directly. Data on the Canadian context is scarce. We do know that Black Canadians report being stopped and searched by police much more than White or Asian-origin Canadians and that they express less trust in police than non-racialized Canadian residents (Cotter, Reference Cotter2022; Wortley and Owusu-Bempah, Reference Wortley and Owusu-Bempah2011, Reference Wortley and Owusu-Bempah2022). Interestingly, Wortley and Owusu-Bempah (Reference Wortley and Owusu-Bempah2022) show that while Black people's negative contact with the police in Canada remains fairly stable over time, White people's awareness of this discrimination has risen dramatically from 1994 to 2019.
In short, research documents that categorical inequalities based on race and citizenship status exist, but much of this work has been conducted in the United States. Less clear, across North America, is how these categorical differences might translate to an unequal recognition of civil and social rights among the public. Further, we believe that differences related to citizenship status need to be disentangled from ethnic and racial origin, and we suspect that these distinctions may play out differently depending on the type of right under consideration.
Data and Methods
To explore these questions, we ran two custom-designed vignette experiments embedded within large representative surveys collected in Canada in 2020 and 2022. Study 1 data are drawn from a two-wave panel. An initial online sample of 8,170 Canadian adults (citizens and permanent residents aged 18 years or older) was collected from May 5–12, 2020 through Dynata using independent quotas for age, gender, language and province based on the 2016 Canadian Census. The vignette experiment was embedded in a second wave of the survey, collected from June 19–29, 2020, in which all respondents from wave 1 were eligible to participate. All surveys were programmed using the Qualtrics online platform.Footnote 6 In total, 4,382 respondents completed the second wave (return to sample of 54%).
Study 2 was collected from May 5–20, 2022. It was an omnibus single-wave survey of 9,928 respondents, in which a random sample of 3,385 respondents received the vignette experiments. Data collection and weighting followed the same procedures as the 2020 data collection, although respondents were drawn instead from the LEO panel. Both Study 1 and 2 were run in English and French.
Our analysis centres on a pair of vignette experiments, run in each study, that presented both a positive and negative rights scenario. For social (positive) rights, we replicate a food insecurity vignette first used by Bloemraad, Silva and Voss (Reference Bloemraad, Silva and Voss2016) in California. For civil (negative) rights, we created an original scenario about possible targeted police questioning, a topic that not only probes respondents’ views on civil rights but is especially relevant in a context where racial bias in policing has become the subject of much public debate. All respondents saw both vignettes in random order.
The vignettes included a 2×2×3 design with manipulations for race (White/Black) and extremes in legal status (citizen versus undocumented), and which included a framing preface experiment (blank versus two priming frames). Here, we hone in on whether categorical differences affect public opinion, so we focus only on respondents who did not receive any framing before the vignette. This includes almost 2500 respondents (n=1467 from Study 1 and n=970 from Study 2).Footnote 7 Our experimental manipulations for categorical inequality were chosen for the theoretical reasons discussed above and to provide comparative findings to US studies; the range of manipulations was constrained by sample size and power considerations.Footnote 8
The racial background and legal status manipulations are embedded within each vignette. The racial manipulation is a combination of a common name and country of origin related to White, European descent (Irish/French) or Caribbean-origin, Afro-descent (Haitian/Jamaican) individuals. Since Canada is a bilingual country, we ensured that implied language capacity was similar (for example, portraying two French speakers in the food security scenario and two English speakers in the policing scenario that are constant across racial cues).Footnote 9 In Study 1, the text and manipulations read as follows:
Food Insecurity: {Name-French/Haitian} [Status = is a Canadian citizen/ came from France on a temporary visa that has expired, leaving her without valid immigration papers/ is a Canadian citizen, originally from Haiti/ came to Canada from Haiti on a temporary visa that has expired, leaving her without valid immigration papers.] {Name} has a very low income. She must often skip meals and eats smaller portions than is healthy. Sometimes, a whole day or longer goes by without her eating anything.
Police Targeting: {Name-Irish/Jamaican} {Status = is a Canadian citizen/ came from Ireland on a temporary visa that has expired, leaving her without valid immigration papers/ is a Canadian citizen, originally from Jamaica/ came to Canada from Jamaica on a temporary visa that has expired, leaving her without valid immigration papers.] {Name} lives in a high crime area in her community. She has never done anything wrong, but the police have repeatedly stopped her when she has been out in the evenings. The police officers regularly ask her to explain what she is doing out.
For Study 1, which was pre-registered (see Bloemraad et al., Reference Bloemraad, Harell, Esses and Wallace2021), the baseline category is the Canadian citizen with a common Anglo-Saxon or French name whose origins are not specified and may be presumed as White. We focus on the pre-registered status hierarchy hypothesis: we expect people to be more willing to support rights when presented with co-citizens (particularly those assumed to be White) and least supportive of non-citizens who have overstayed their visa (especially those perceived as racial minorities). Furthermore, we suspect that the race effect will be greater in the condition with non-citizens. In other words, we expect more racial discrimination when the person is portrayed as categorically different on another criterion (lack of legal status). In Study 2, we sought to reproduce the findings in Study 1, using identical vignettes, but the baseline category was modified to be a presumably White citizen explicitly of immigrant origin. The prompt therefore became “is a Canadian citizen, originally from France/Ireland” to mirror the Black citizen prompt more closely.
Each vignette was followed by a series of questions that asked respondents to evaluate the nature of the rights violation and what should be done about it. Respondents were provided with a five-point disagree/agree scale, recoded to run from -1 to 1 where 0 is the neutral “neither agree nor disagree.” In the food insecurity scenario, we replicate the question wording and outcome variable language used in a prior US study (Bloemraad et al., Reference Bloemraad, Silva and Voss2016), asking: “Should [NAME] receive extra benefits from the government to ensure she has enough to eat?”Footnote 10 In the police scenario, we evaluate a scale of four items that assess whether the scenario is a violation of civil rights and what should be done about it. The four items, designed with the vignette specifically for this study, are:
The police should be able to question [NAME] any time she is on the street. (reverse coded)
[NAME] is being treated unfairly by the police.
[NAME] should file a complaint about her treatment so that the officers can be punished.
The police are just doing their job when they question [NAME]. (reverse coded)
These four items are highly correlated, surpassing our pre-registration threshold for scale creation in both studies, with Cronbach's alpha=.877 (Study 1) and .869 (Study 2). We, therefore, use the scale as our dependent variable to evaluate negative rights recognition. Additional information about the instrument and variable coding is available in Appendix G.
Analysis: Whose Rights Are Recognized?
Overall, when respondents in our studies read about a woman going hungry, the majority of people agreed (35%) or strongly agreed (29%) that the individual should receive extra benefits from the government so that she has enough to eat. Similarly, when presented with a scenario in which a woman was repeatedly stopped by police in the evening, despite having done nothing wrong, most respondents agreed (34%) or strongly agreed (31%) that she was being treated unfairly. Many respondents, though not all, endorsed the provision of positive, social rights or recognized a violation of negative rights. We wonder: was this endorsement or recognition uniform, irrespective of racialized minority or legal status?
To test our core status hierarchy hypothesis, we regress our treatment categories onto our outcome variables of interest, controlling for the order that the vignette was viewed. Figure 1 provides the predicted level of agreement for each outcome in Study 1 and 2 based on the status and racial background of the profile viewed, and when necessary, linear hypothesis tests of differences are reported in the text when comparing the regression model's coefficient estimates for pairs of treatments to each other. For simplicity, we report 95 per cent confidence intervals in the image, with full models provided in Appendix A. Our expectation was that we would see a hierarchy of support for both positive and negative rights, with citizens receiving the highest level of support (especially when White) and undocumented profiles receiving the lowest level of support (especially when Black).
Predicted levels of agreement for positive and negative rights by treatment
Note: Predicted agreement in each study on a 5-point agreement scale from -1 (strongly disagree) to 1 (strongly agree), with 95% confidence intervals. Models include a control for vignette order.
The data analysis shows support for status hierarchy, although not uniformly in line with our hypotheses. Legal status is a clear basis of categorical inequality; the impact of race is partial and variable. We also find that patterns of rights support vary depending on whether we consider positive or negative rights.
In terms of social rights, in both Study 1 and Study 2, respondents are much more likely to support providing public benefits to citizens, whether White or Black, than visa-overstayers.Footnote 11 In Study 1, we also find a significant distinction between White citizens and Black citizens with immigrant origins (F=6.67, p=.010) but no parallel racial distinction among profiles of a person without valid papers (F=0.59, p=.443). In Study 2, racial distinctions among citizens disappears when we compare two first-generation citizens (F=.49, p=.483), but a new—and unexpected—bias emerges in favour of the Black undocumented profile relative to the White visa-overstayer (F=18.81, p=.000).Footnote 12
One interpretation of the small change in results from Study 1 to Study 2 would be that the distinction among differentially racialized citizens in Study 1 is driven not by race but by a distinction respondents are making between a (presumably White) Canadian citizen who is likely native-born (as no other origins are mentioned in the vignette) and a (presumably Black) first-generation Canadian whose immigrant origins are made explicit. Indeed, when both citizens are described as having immigrant roots (Study 2), the racial distinction among citizens disappears.Footnote 13 However, this difference between Study 1 and Study 2 may instead be driven by the general increase in support for all Black profiles in Study 2. Below, we discuss possible timing effects on the 2020 and 2022 surveys. For now, we note that the estimated level of support for social benefits for the White citizen in Study 1 is statistically indistinguishable from the level of support for the White first-generation citizen in Study 2. This suggests that the wording change, adding in Irish or French origins, is not primarily at work in mitigating differential assessments of White and Black citizens. In comparison, support for providing social benefits to both the Black citizen and the Black undocumented resident profiles increases from Study 1 to Study 2 (p<.05).Footnote 14 Constant across the two studies is the significant preference for allocating social rights to citizens, independent of race.
When we turn to attitudes toward negative rights—freedom from state harassment in the police identification check scenario—we again see some evidence of a legal status difference in Study 1, but this effect is clearly racialized. Respondents were more likely to see police action as wrong when the Black citizen was stopped compared to the Black undocumented person (F=12.46, p=0.000). Likewise, the White citizen receives stronger agreement that the police are acting inappropriately than the White undocumented person (F=4.34, p=.037). At the same time, respondents clearly agree that police were violating the Black citizen's rights more than the White citizen's (F=9.18, p=.003). The salience of race is also evident in that when the undocumented person was described as Jamaican, respondents were equally as likely to see police wrongdoing as when they were presented with a vignette about a White Canadian citizen (F=.20, p=.653). Mean agreement of police wrongdoing appears to increase within each treatment group except for White citizens between Study 1 and Study 2, though these differences are not significant in a pooled model (see Appendix B). Still, Black citizens again receive the highest level of agreement on our scale of police wrongdoing. However, we find no significant distinction among the other three profiles in Study 2.Footnote 15 This suggests that respondents were particularly attentive to the combination of citizenship and racialized minority status in assessing civil rights violations, rather than citizenship or race as independent categorical distinctions.
The combined results of the policing scenarios in Study 1 and Study 2 point to an awareness of racial bias in the enjoyment of negative rights, one where respondents in Canada view police-citizen interactions with greater scrutiny when a person is presumably Black. Yet, importantly, this awareness of racialized jeopardy does not extend fully to all Black people, including those without legal resident status. One interpretation is that those without papers are not viewed as having Charter rights in the same way as racialized citizens.
The Relevance of Timing
The results from Study 1 and 2 point to categorical inequalities in Canadians’ recognition of both positive and negative rights. Citizenship status clearly matters for the recognition of social rights in the minds of the public, reflecting what we have seen in the comparative literature about the use of immigration status as a deservingness heuristic for access to social benefits (van Oorschot, Reference Van Oorschot2006; Reeskens and van de Meer, Reference Reeskens and van der Meer2019). Whether in 2020 or 2022, citizens were far more likely to be given social benefits than non-citizens lacking status in Canada. In this respect, public opinion largely matches much social benefit provision in Canada, but it stands at odds with human rights declarations or aspirations calling for no person to go hungry and all people to have an adequate standard of living. The findings for categorical inequalities on race are less consistent. In Study 1, we found evidence for a bias toward the White citizen compared to the Black citizen of immigrant origin, as hypothesized. This can be interpreted as evidence of White racial privilege among citizens. But the effect disappeared in Study 2, largely, we suspect, because of increasing support for the Black citizen, rather than a discounting of the White citizen's immigrant background in Study 2. We did not find support for the hypothesis that respondents would be least generous toward a Black undocumented person: for those lacking status, race was not a statistically significant difference in 2020, while in 2022, the White visa-overstayer was viewed least sympathetically of all profiles.
Citizenship status also matters for the recognition of negative rights, but not as uniformly as for social rights. Unlike with social rights, where the law on the books largely excludes out-of-status migrants from publicly funded health insurance, income assistance or other social benefits, basic civil rights against unreasonable police action apply equally to all persons in Canada. Such negative rights are effectively human rights, but our data show that the public does not adopt this approach, making a distinction between Black citizens and everyone else.
The data also offer evidence for the public's awareness of the racialized nature of policing in Canada. Recall that Study 1 was fielded in June 2020, shortly after the murder of George Floyd by police officers in Minneapolis, an event that led to large-scale protests over police brutality, racial profiling and racial bias in police enforcement. This became a highly salient topic not just in the United States but in numerous countries around the world including Canada. For example, Black Lives Matter organizations have been active in Canada since 2014Footnote 16, and the murder of George Floyd led to a series of protests in CanadaFootnote 17 as well as statements by political leaders, including Prime Minister Trudeau taking a knee at an anti-racism protest on Parliament Hill on June 5, 2020, while this survey was in the field.Footnote 18 Yet, the results from Study 1 are not unique to that period. In fact, agreement of wrongdoing is equally present in 2022 for vignettes featuring a Black citizen as well as for both undocumented profiles (see Appendix B). This may speak to the lasting effect of public debates and social movement mobilization around race-based bias in policing.
In sum, categorical inequalities in Canadians’ recognition of rights are evident, but they vary by the type of right under consideration and are stronger for legal status extremes than race, at least in the positive and negative rights scenarios that we test here. So, while we find evidence of some racial inequalities favouring White citizens in the positive rights scenario, racialized differences are more modest than would be expected from the largely American literature on welfare and race. And, to the extent that race mattered for negative rights, we interpret our results as evidence of greater consciousness about racial bias by the police among the public, including in Canada, although such concerns are refracted by citizenship status.
Extensions: Respondents’ Background
Is there heterogeneity among the Canadian population in how people recognize rights? Our core analysis focused on the Canadian population as a whole. Yet, we might suspect that the effect of categorical inequalities on attitudes is most likely to occur among Canadians who are less likely to face discrimination based on their legal status or their ethnoracial background, namely native-born citizens and respondents who identify as White. Expectations are less clear for immigrants and non-White respondents. On the one hand, we might expect immigrants to show in-group favouritism and demonstrate less differentiation on our outcome measures. Yet, on the other hand, we could also imagine immigrants penalizing the non-status individual because the vast majority of immigrants in our sample would have immigrated through regularized, legal channels. In that sense, they might not show any particular sympathy for those who lack status. We imagine less equivocation for racialized respondents, particularly those who identify as Black. For other non-Black racialized minorities, we expect that they may be more sensitive to the discrimination faced by Black people in Canada, and so, express more sympathy for access to social rights for a racialized individual or be conscious of racial bias in policing, but anti-Black bias is not necessarily restricted to White Canadians.
In Figure 2, we estimate predicted levels of agreement for White, Canadian-born respondents compared to foreign-born and racial minorities.Footnote 19 We have collapsed Study 1 and 2 together in order to ensure a sufficient sample of immigrants (identified by foreign birth, n=422) and racial minority respondents (self-identified by those who selected ethnoracial categories other than White, n=364). Ideally, we would also model the latter specifically for those who reported they identify as Black, but we lack a sufficient sample to run this model (n=52). This means that we cannot fully test if there is an in-group effect in the Black treatment categories.
Predicted levels of agreement for positive and negative rights, by respondent background
Note: Predicted agreement in each study on a 5-point agreement scale from -1 (strongly disagree) to 1 (strongly agree), with 95% confidence intervals. Model includes a control for vignette order and study.
What is striking about Figure 2 is the extent to which the categorical inequalities documented in Figure 1 remain in the subgroup analysis. Citizenship status again emerges as the main inequality in access to social benefits. In the pooled sample, we see a small preference for White citizens, but separate analysis of White, native-born respondents suggests this is driven primarily by Study 1. The general pattern is reproduced among foreign-born respondents as well as racial minority respondents. The only other obvious difference is that White, native-born respondents provided the most lukewarm support for the non-status White profile. Support was generally higher for this category among racial minority and foreign-born respondents, suggesting they identify less of a categorical distinction between the two non-status profiles based on race. Interestingly, rather than White racialized in-group bias, native-born White respondents appear particularly unsympathetic to an Irish or French undocumented migrant.
When it comes to the negative rights scenario, we see again an identical pattern across sub-groups: respondents are most likely to recognize a negative rights violation for Black citizens. Here, among racial minority and foreign-born respondents, the predicted level of agreement is not statistically distinct from the neutral midpoint though caution is warranted given the larger confidence intervals. Strikingly, even though racial bias in policing likely affects all racialized individuals, irrespective of citizenship, the subgroup analysis suggests that respondents react most strongly to the violation of Black citizens’ negative rights, not just a Black person, or even any person, despite Charter or human rights protections for all individuals. In other words, we largely find similar patterns across these subgroups. Further analyses with interaction terms show no significant moderating effect of being a racial minority or being foreign-born on our treatment effects (see Appendix D).
Conclusion
To whom do we owe what sorts of rights? It is not uncommon for political theorists, legal scholars and public officials to view negative rights or liberties as more fundamental—or at least less contested—than rights to welfare state benefits. This would suggest, as inscribed in Canadian and international law, that civil rights should apply to everyone, irrespective of ethnoracial background or legal status. The case of positive rights is more contested, with some seeing freedom from hunger and basic welfare support as a universal human right, while others link government benefits to citizenship. Among citizens, however, distinctions by race are considered illegitimate, even as a voluminous US literature and some Canadian studies point to race-based bias in public opinion on redistribution.
Our data show that Canadian residents view both positive and negative rights to be more applicable to citizens as compared to immigrants who have fallen out of status. There is no evidence of universalism in rights. The small literature documenting the categorical inequality of legal status, especially being undocumented, has overwhelmingly focused on the United States. This study shows that Canadians, too, view rights more as something reserved for citizens, and less as a set of human rights for all. We note, however, that our study contrasts legal extremes: profiles of full legal citizens compared to those without current status to reside in Canada. Future research should explore other categories, such as scenarios with non-citizens who are permanent residents, people on temporary work or student visas and asylum-seekers or refugees. It is possible, in line with arguments about “civic fairness” (Levy and Wright, Reference Levy and Wright2020), that Canadians are particularly troubled by—and thus withhold rights from—those who are out-of-status rather than all non-citizens.
We also found that race matters in the recognition of rights, but not consistently in the status hierarchy that we hypothesized nor as robustly as suggested by studies conducted in the United States. We had expected a preference for White over Black citizens, and an even bigger racialized preference for White over Black non-status migrants. In the case of social rights, we found a modest racialized preference for White citizens in Study 1, but this disappeared in Study 2; for vignettes with an undocumented migrant, we found, if anything, more generosity toward a person of Jamaican or Haitian background who was out of status and hungry than a similarly situated woman from Ireland or France. Further, these results were quite consistent among sub-groups in the Canadian population: native-born Whites, racialized minority residents and the foreign-born. Future research should extend these findings to other salient groups, notably among Black Canadians themselves, as well as groups such as linguistic communities in Canada. Canadian public opinion does not demonstrate as consistent racial biases as oft-reported in the United States, contextualizing the generalizability of group identity models predicated primarily on race (for example, Masuoka and Junn, Reference Masuoka and Junn2013).
It would be very premature, however, to conclude that race is not a basis of categorical inequality in Canada. We know that Black Canadians experience inequality in the labour market, politics and their everyday experiences with policing. It is possible that Canadian respondents engage in more social desirability than US counterparts, self-censoring when responding to surveys. Our study also only compared two racialized social categories, defined as women of presumed European and Afro-descent origins. Further research that probes the intersections of ethnoracial background and immigrant status should consider additional categories, including East Asian, South Asian and Arab-descent communities, as well as variation by gender and language. We also note the possibility of respondents questioning the validity of our experimental manipulation of an undocumented person of European descent, consciously or unconsciously producing particularly negative responses in that scenario. Empirically, there are French and Irish visa-overstayers living in Canada, but the public might not include such people in their cognitive schemes of irregular-status migrants. Alternatively, respondents might take the existence of a White visa-overstayer at face value but view them unsympathetically, since they perceive Ireland and France to be rich countries with robust welfare states and expect nationals from these countries who face food insecurity to return home if they need help, an expectation not extended to those from Haiti or Jamaica, given greater poverty there. If this is the case, it further underscores the idea that the public makes judgements about social rights based on membership categories delineated by citizenship (being Irish or French), rather than immediate need.Footnote 20
When it comes to negative rights, racial distinctions made respondents more likely to agree that a policing scenario was problematic when the victim was implied to be a Black Canadian citizen. We read this result to show that respondents recognize that racialized inequities in policing exist, even in Canada, and we find suggestive evidence that mobilization against anti-Black racism might be having a modest effect on public opinion, given slight changes in reported attitudes from Study 1 to Study 2. At the same time, we underscore the limits of this recognition: concern is primarily for Black citizens of Canada, not uniformly for those who may be perceived as Black.
These findings largely hold across the Canadian population, with little variation depending on whether a respondent was born in Canada or not or whether the person self-identifies as White or not. In other words, there seems to be general recognition that rights—whether positive or negative—should be guaranteed for citizens. The public is far more lukewarm toward extending even the most basic of freedoms—from food insecurity or being harassed by the police—to those who lack status in Canada. While our study only focuses on two examples of positive and negative rights and two types of categorical inequalities, the results clearly demonstrate that Canadians do distinguish when recognizing rights. Further research is needed to assess the extent to which other categorical inequalities exist—the documented bias against Indigenous peoples in Canada comes immediately to mind—and to what extent the findings here extend to other negative and positive rights scenarios.
These findings also speak to the importance of bringing normative and empirical political science into conversation. It is common to distinguish between scholarly work asking what society or people ought to do and research about what actually happens or exists in the world. But much of political contestation, especially on questions of migration and citizenship, centres on moral claims about who ought to have rights, whether rights are enshrined in law, written into legislation or offered through public institutions. Why can some enjoy rights, and not others? These moral schemes are deeply embedded in debates around immigration (Yazdiha, Reference Yazdiha, Dromi, Hitlin and Luft2023). Political theorists help us to clarify the stakes and values embedded in such debates. Today, among theorists who grapple seriously with the legitimacy of providing rights to some while excluding others, appeals to human rights—as a matter of justice and universality—is a prominent response; this is a claim also articulated regularly by social movements advocating for migrants (Voss et al., Reference Voss, Silva and Bloemraad2020). But given our findings, theorists would do well to consider the tension between idealist and realist perspectives (Carens, Reference Carens1996). As Rainer Bauböck notes, “The theorist is generally not held accountable for her advocacy of policies,” yet a fruitful dialogue between theorists and empiricists would expose normative ideas “to the full force of critique from… empirically grounded research that analyse[s] the application context” (Reference Bauböck, Porta and Keating2008: 59). Our study makes clear that Canadians distinguish across types of rights, despite moves in the normative literature to embed them all within a more universalistic human rights framework. Scholars who care about reducing categorical inequalities, especially toward those who lack full citizenship, need to explore how and when moral claims of universal rights can—both in the short and long-term—become more accepted by Canadians.
Supplementary Material
The supplementary material for this article can be found at https://doi.org/10.1017/S0008423924000076.
Acknowledgements
This research was supported by a Catalyst Grant from the Canadian Institute for Advanced Research (CIFAR) in collaboration with the Consortium on Electoral Democracy (C-Dem), which is supported by a SSHRC Partnership Grant (#895-2019-1022). We would like to thank Victoria Esses, Audrey Macklin, Catherine Dauvergne, Keith Banting and Will Kymlicka for helpful comments on previous versions, as well as Rebecca Wallace, Laura French Bourgeois and Sabrina Spence for assistance.
Competing Interests
The authors declare none.
Open access
