Indigenous Diffuse Support and Descriptive Representation in the Canadian House of Commons

Abstract

Do Indigenous peoples in present-day Canada display lower levels of diffuse support than non-Indigenous settlers? Given settler colonial relations (both historic and contemporary) and Indigenous peoples’ own political thought, we can expect that Indigenous peoples would have even lower perceptions of state legitimacy than non-Indigenous peoples. However, there are conflicting expectations regarding whether the descriptive representation of Indigenous peoples in settler institutions is likely to make a difference: on one hand, Indigenous people may see themselves reflected in these institutions and consequently feel better represented; on the other hand, these forms of representation do not challenge the underlying colonial nature of these institutions. Using data from the 2019 and 2021 Canadian Election Studies, our statistical analysis demonstrates that: (1) diffuse support is significantly lower among Indigenous peoples than non-Indigenous peoples, including people of color; (2) Indigenous respondents across multiple peoples have similarly low levels of diffuse support, and (3) being represented by an Indigenous Member of Parliament does not change the levels of diffuse support among Indigenous peoples. Overall, our research highlights the outstanding challenges to achieving reconciliation through the Canadian state and points to ways large-N analyses may be made more robust.

Introduction

Indigenous-led protests against Canadian settler state institutions have become more visible in recent years, with mobilizations against government policy and legislation, court decisions, and especially police violence gaining significant media coverage. These events build on a long history of Indigenous ¹ organizing and the significant work of Indigenous nations, organizations, and governments to practice their self-determination. Such actions are based on an understandings of co-equal sovereignty and authority over their own peoples, with the mobilizations by Indigenous peoples ² also occurring alongside an increasing openness among many settlers to the work of reconciliation (The Environics Institute for Survey Research 2021). This openness is reflected in government rhetoric that speaks of mending historic harms and the importance of moving past “discriminatory” policies that adversely affect Indigenous peoples. Stated openness is not a panacea however, and the colonial state apparatus not only remains relevant, but colonial policies and genocidal state actions continue apace (National Inquiry into Missing and Murdered Indigenous Women and Girls 2019a).

Democratic states depend on citizens seeing the political system as legitimate. If citizens reject the legitimacy of their political institutions in large numbers, those institutions lose the moral authority to govern. The potential for a legitimacy crisis is especially pronounced for settler states such as Canada since their democratic systems are built on top of the Indigenous confederacies, nations, and communities that predate settlers in North America by thousands of years. Indigenous people also represent one of the fastest-growing groups in Canada (Statistics Canada 2021). As a result, low support for the legitimacy of state institutions among Indigenous people in Canada may represent a growing crisis for settler Canadian people, representatives, and institutions alike.

Considering this complex situation, we ask: (1) how do Indigenous people in present-day Canada view the legitimacy of the Canadian state and (2) does the descriptive representation of Indigenous people in colonial institutions change the perceived legitimacy of those institutions? We answer these questions using data from interviews with Indigenous participants in the 2019 and 2021 Canadian Election Studies (CES). To operationalize Easton’s (1975) concept of diffuse system support, we examine Indigenous respondents’ political trust and external efficacy to reflect their perceptions of state legitimacy. We then connect this analysis to the descriptive representation of Indigenous candidates in the 2019 and 2021 federal elections. Our findings suggest that Indigenous respondents have significantly lower levels of diffuse support for the Canadian state than both white and visible minority settlers. This aligns with expectations, given existing settler colonial relations of Indigenous dispossession and the elevation of settler sovereignty as superior to that of Indigenous sovereignty. We also find that the descriptive representation of Indigenous people in settler institutions does not change Indigenous peoples’ perceptions of the fundamental underlying legitimacy of the Canadian state. This suggests that while individual Indigenous representatives may seek office in settler institutions for a number of other good reasons, their election does not necessarily translate into greater perceptions of settler state legitimacy among members of Indigenous communities.

As we discuss in greater detail below, research into Indigenous peoples’ views is relatively under-developed in the field of political behavior – part of a broader pattern wherein “political scientists have largely ignored Indigenous political traditions,” which has led to a “disconnect between Indigenous politics and the discipline” (Ladner 2017, 165). This is beginning to change, with an increasing amount of research published which centers Indigenous politics and political traditions (Borrows 1997; 2010; 2016; Henderson 2020; Ladner 2003; Simpson 2014; Simpson 2008; 2011; Williams 2018a; Williams 2018b). However, there remain few statistical studies of Indigenous peoples’ own views of politics and their political behavior. ³ As a team of both Indigenous and settler researchers, our analysis attempts to center, but not coopt, Indigenous peoples’ own intellectual traditions in order to understand how these shape contemporary Indigenous-settler relations. It is an early attempt to develop a more fulsome understanding while engaging in the type of epistemological pluralism that we hope continues to flourish in political science. Further efforts are necessary, and we conclude by briefly looking ahead. While we identify several difficulties, we argue these can be overcome with attention to developing studies around Indigenous peoples’ own political knowledges.

Indigenous Peoples and Political Behavior in Canada

Almost everything we know about Indigenous political behavior in Canada concerns electoral participation – in particular, voter turnout. The gap between Indigenous and non-Indigenous voter turnout in federal and provincial elections is a historic, and persistent, phenomenon (Alfred, Pitawanakwat, and Price 2007; Bargiel 2012; Belanger 2009; Cairns 2003; Guérin 2003; Harell, Panagos, and Matthews 2010; Ladner and McCrossan 2007). Indigenous voters – like all other voters – have higher levels of turnout when they have more political resources – such as higher levels of education, income, age, and political knowledge (Fournier and Loewen 2011). Harell et al. (2010) also find support for the impact of these variables but argue a significant participation gap remains after accounting for political resources that is best explained through an understanding of Indigenous peoples’ contentious relationship with the Canadian state – an explanation that aligns with Simpson’s (2014) conceptualization of Indigenous refusal of settler politics as a way to affirm their own sovereignties. Similarly, Barsh et al. (1997) find that feelings of exclusion, lack of electoral information, and sociodemographic factors all contribute to lower First Nations voter turnout in Alberta, and Belanger (2009) further identifies six key barriers preventing greater First Nations voter turnout in Canadian elections: historical disenfranchisement, socioeconomic disadvantages, a lack of effective communication, geographical dispersion, feelings of exclusion, and views of voting as an abstraction.

A separate branch of research on Indigenous voter turnout examines the impact of political institutions – particularly, electoral institutions. Most recently, online voting has received attention as a potential way to increase Indigenous political engagement, with a specific focus on band council elections (Goodman, Gabel, and Budd 2018). The research on the impacts of online voting points to increased accessibility with mixed results regarding turnout among Indigenous voters, though there are important considerations for other forms of community engagement (Gabel et al. 2016; Gabel and Goodman 2018). Finally, an increasingly studied area of Indigenous political behavior concerns the impact of representation. Budd (2019) argues the guaranteed representation of Indigenous peoples in democratic institutions – accompanied by recognition of Indigenous sovereignties and nationhoods – is a key step necessary to increase Indigenous electoral participation. This argument finds support from Dabin et al. (2019), who demonstrate that “turnout was higher in Indigenous communities where Indigenous candidates were on the ballot”, with political parties running Indigenous candidates also receiving more support in Indigenous communities.

Focusing on voting behavior misses the spectrum of political attitudes held by Indigenous peoples in Canada. To our knowledge, these have largely not been studied. One notable exception comes from the Institute for Research on Public Policy’s Centre of Excellence on the Canadian Federation. In their survey research, they found little evidence for strong support of the Canadian state among Indigenous respondents (The Environics Institute for Survey Research 2021).

Diffuse Support and Indigenous Peoples

Examining diffuse support is an intuitive first step in beginning to fill this gap. As one of the most important indicators of the health of a democratic regime, diffuse support constitutes “a reservoir of favorable attitudes or good will that helps members to accept or tolerate outputs to which they are opposed” (Easton 1975, 444). This “reservoir” influences the extent to which citizens view the political regime as legitimate, or one which citizens view as having the moral authority to govern (Levi and Stoker 2000). Diffuse support is distinct from specific support – with the former referring to the political system as a whole and the latter to the political leaders of the day (Easton 1975).

Diffuse support is usually measured through a number of indicators, including external efficacy and political trust. External efficacy refers to “beliefs about the responsiveness of governmental authority and institutions to citizen demands” (Craig, Niemi, and Silver 1990, 290). Individuals with high external efficacy believe the political system would convert their preferences to political outcomes if pressured to do so. ⁴ By contrast, political trust is “a summary judgment that the system is responsive and will do what is right even in the absence of constant scrutiny” (Miller and Listhaug 1990, 358). In other words, while external efficacy refers to whether authorities are responsive to citizens’ influence attempts, political trust is concerned about outcomes that take place when regimes are left to their own devices.

As an outcome, diffuse support is often thought of as the combination of individual-level and system-level factors. At the individual level, the “social deprivation hypothesis” posits that individuals who are denied opportunities in society develop lower feelings of political trust and external efficacy (Abramson 1972; Iyengar 1978). According to this view, socioeconomic and educational differences strongly affect citizens’ views about the political system. Individuals with more socioeconomic resources and higher levels of education are more likely to vote, be interested in and participate in politics, and are less cynical about the political system (Hakhverdian, Van Der Brug, and De Vries 2012). At the system level, the “political reality” hypothesis suggests members of politically disadvantaged groups expect the political system to be less responsive to them because it has tended to be less responsive in the past (Abramson 1972; Iyengar 1978). This results in lower feelings of diffuse support. The nature of political disadvantage can be broad – ranging from women, who are underrepresented in elected politics, to individuals who support political parties which are excluded from power.

How might we expect Indigenous people in present-day Canada to feel about the legitimacy of the settler state? In answering this, we look to the past and present relationships between the state and Indigenous nations. Specifically, these relations are structured by settler colonialism, whereby sovereignty as ultimate governing authority is exclusively claimed by the (settler) state. This has not always been the case, however: in the past Indigenous peoples, nations, and communities rivaled the nascent state’s power, and asserted their own independence from it. These assertions of Indigenous sovereignty are found in the agreements struck between settlers and Indigenous peoples. Prominent examples include the numbered treaties across the prairies (for example, see Starblanket 2019) and those that cover much of what is today Ontario, and the Peace and Friendship treaties covering much of the land on which the Maritime provinces now rest.

Treaty relationships are embodied not only in written texts but also in physical and ceremonial forms. One of the most well-known is the Tekeni Teiohatatie Kahswentha, or the Two Row Wampum belt that we use to illustrate the relationships here. Wampum belts had been used between Indigenous nations to symbolize treaty relationships between peoples (Monture 2014, 13–14; Muller 2007). Made from beads derived from freshwater quahog shells, Tekeni Teiohatatie Kahswentha represents an agreement concluded between the Haudenosaunee and Dutch in 1613 that defined their relationship and the governance arrangements. It draws on the political traditions derived from the Haudenosaunee Great Law of Peace (Muller 2007), illustrating two vessels – one Haudenosaunee, one Dutch – traveling down a river alongside each other. As Monture (2014, 14) describes, “these two vessels coexist, they are considered separate but equal in status, never interfering in each other’s social or political affairs. To the Haudenosaunee of the seventeenth century, this was a declaration of sovereignty as well as recognition of the rights of outsiders within their territory.” As one of the first formal agreements between Indigenous nations and European settlers, the belt exemplifies the long history of Indigenous nations asserting their own sovereignty and independence vis-a-vis European settlers. This declaration of co-equal sovereignties is further affirmed in the negotiation with the British of the subsequent Tehontatenentsonterontahkhwa, or Covenant Chain alliance belt, which was agreed to on the same principles of Haudenosaunee sovereignty and co-equal co-existence.

One treaty cannot, and does not, represent all Indigenous peoples’ relationships with the settler state. What the Two Row Wampum shows, however, is the long history of claims to, and enactments of, sovereignty on the part of Indigenous peoples. Part of what is wrapped up in asking how Indigenous peoples view the legitimacy of the Canadian settler state, then, is this history and settler refusals of it. Instead, Canada used the Doctrine of Discovery to assume de jure sovereignty through its de facto control as a result of the power imbalance (McNeil 2019). We expect these historic and ongoing legacies to lead to low levels of diffuse support for settler institutions among Indigenous peoples in Canada.

Indigenous Descriptive Representation in Canada

Descriptive representation refers to a common ethnic, religious or gender identity between representatives and the represented (Pitkin 1967; Casellas and Wallace 2015). In this article, we examine descriptive representation as the relationship between Indigenous elected officials in the Canadian Parliament and the Indigenous constituents living within the geographic boundaries of their settler-defined electoral divisions. In many cases, “democratic representative institutions can be tools of oppression” for these communities because they have either formally prohibited members of these communities from holding elected office or have dramatically underrepresented the population of these communities (Dovi 2007, 297). By increasing the number of descriptive representatives in these institutions, historically excluded groups may be able to attain greater symbolic and substantive outcomes for their communities (Childs and Cowley 2011).

Indigenous nations and communities in relation to Canada are incredibly diverse, and assuming that Indigenous parliamentarians represent Indigenous communities risks essentializing these communities and homogenizing their views. In turn, diverse Indigenous representation is important in legislatures and our analysis. Descriptive representation that reflects the diversity of Indigenous peoples not only allows for Indigenous viewpoints and understandings to be included but also combats assumptions that all Indigenous peoples are the same, or that they can speak or represent one another. Indeed, Indigenous peoples and communities do not have “a fixed, knowable set of interests” or “a single, undisputed, authentic identity that can merely be received by a political representative as if the flow of meaning was all in one direction” (Saward 2006, 301, 313). Moreover, while some interests are well known, others remain “uncrystallized [and] not fully articulated” until a specific question arises, meaning members of these communities are themselves better placed than nonmembers to represent these interests (Mansbridge 1999, 628). Indigenous representatives in settler state institutions do not guarantee that the views of all Indigenous people will be represented in state decision-making; nevertheless, a degree of shared experience with the settler state does mean Indigenous elected officials may be better placed to ensure that at least some Indigenous voices are considered in settler government (Celis and Childs 2008).

Moreover, Indigenous peoples have drawn from long histories with governance systems of their own – both prior to and during the establishment of the Canadian state. Through complex structures such as confederacies and nations, Indigenous peoples took part in diplomacy, made decisions, and practiced politics in various forms – including through structures of consensus and representation. The Lnúk, ⁵ for example, divided the territory they steward into different districts. Each district was represented by an equal number of representatives who would then appoint a Chief to represent their district to a higher council (Henderson 2020). The district structures would handle local issues, concerns, and matters, while the higher council would focus on more regional issues, war, and alliances. The higher council was also responsible for questions relating to the Wabanaki Confederacy, which included the Lnúk along with the Maliseet, Passamaquoddy, Penobscot, and Abenaki (Henderson 2020). Each nation was equally represented in the Confederacy’s decision-making structure, which included consensus requirements when it came to issues of international affairs (Henderson 2020).

The Anishinaabeg, which includes the Algonquin, Michi Saagiig, Chippewa’ag, Odawa, and Bodewadami, also utilized consensus forms of government in the form of a dotem ⁶ system for representation. Each dotem reflected different responsibilities not only within Anishinaabeg communities but also by region and location. In relation to Canadian institutions, one could equate the migizi ⁷ dotem to something like the Prime Minister’s Office (PMO) and the ajijaak ⁸ dotem to fulfilling similar duties to that of the Ministry of Intergovernmental Affairs, or in the US context, the Department of the Interior (Cowie 2024, Sitting Eagle and Henry 2011). Each nation within the Anishinaabeg had equal standing when coming together and discussing intergovernmental, as well as international, decisions.

As these examples of governance systems show, Indigenous peoples had their own forms of descriptive representation – and utilized their own forms when negotiating and meeting within their own structures, but also when other nations arrived on their shores. Importantly, while representation existed within clans, societies, and decision-making processes, decision-making also often required forms of consensus that are unlike the oppositional forms of representative politics we see today.

These structures continued to be used following contact with Europeans, whether through alliances, wars, or treaties. Such approach continued, as previously highlighted, with the use of wampum belts as well as peace and friendship treaties – including following the formation of the United States and responsible government in the British colonies to the north and the Dominion of Canada (Borrows 1997; Cowie 2024; Simpson 2008; Williams 2018a). However, Indigenous representation in settler Canadian institutions was severely curtailed, and decision-making authority was usurped by settler Canadian representatives and institutions. This included Status Indians (First Nations) being labeled as wards of the Canadian state and not being granted the ability to fully represent themselves or participate in the Canadian state, and its institutions, unless they agreed to surrender their identity and treaty rights through enfranchisement. This was mirrored by the total exclusion of Inuit from representative liberal democratic structures until the 1950s. Exclusion of Indigenous self-determination came to a head with the Canadian state’s expansion into the prairies, where the Métis of the Red River Settlement area pushed back against Canada’s claim to territory that they believed they controlled.

Manitoba is also important in the context of Indigenous representation. Upon Manitoba’s formation, Métis were granted citizenship and thus an ability to not only participate but also vote in Canadian federal elections – leading to early Métis representation in the House of Commons (HOC). Indeed, as Table 1 shows, Métis have been at the forefront of a number of milestones for Indigenous descriptive representation in Canada:

Table 1. Historic Milestones for Indigenous descriptive representation in Canada

Table 1 documents significant exclusion followed initial openings to Indigenous peoples’ inclusion in democratic institutions. After 1871 in Manitoba, Canadian institutions and representatives quickly moved to repeal many of the protections originally agreed to in relation to the Métis. Officials expressed it was the will of the majority, especially as Anglo-Ontarians relocated to Manitoba (Ferguson 2021). In turn, many Métis fled their homelands in the Red River Settlement and sought to again counter the Canadian state during the Resistance of 1885 (Cowie 2021; Ferguson 2021). The Canadian state had the upper hand and quickly quashed the Métis resistance – leading to their submission and hiding of their identity. Métis representation would not return to the HOC for more than 60 years. Likewise, although First Nations men who individually owned land, and could prove they had for at least 3 years, were able to participate in the Canadian federal elections of 1886, 1891, and 1896, there is no clear indication any opted to seek election themselves (Kirkby 2019). Status First Nations would not, under Canadian law, be included for over another 50 years until 1960. Lastly, Inuit were almost entirely ignored until 1934 when the Canadian state introduced the Dominion Franchise Act, which barred Inuit from participation and voting (Cowie 2024; Cowie Forthcoming). It would not be until 1951 that Inuit were allowed to vote at all, with polling stations only being set up in Inuit communities for the first time in 1979.

In large part due to this history of exclusion and that of colonization, tensions remain around the question of voting in Canadian elections for Indigenous peoples. For First Nations, distrust remains, and there is concern that participating, whether voting or running for office, would further erode their own Nations and confederacies. Moreover, worries persisted that the unilateral decision to make Inuit and First Nations citizens, without having to give up their identity, was done to further legitimize the Canadian state and its institutions (Cowie 2024; Cowie Forthcoming). Although this is a view that many held, and still hold today, that does not mean individuals did not seek election – and in some cases in districts that reflect the traditional territory of their nation and/or community.

Indeed, Indigenous representatives have been pivotal in key moments in the development of the Canadian state. For example, in 1968 Leonard Marchand, a member of the Sylix nation and the first Status First Nations person elected to the HOC, was elected as the MP for Kamloops-Cariboo. Marchand would go on to assist with pushing the Pierre Elliott Trudeau government to formulate a land claims process following the Calder Case, 1973 (Cowie 2024). Almost 11 years following Marchand’s election win, Peter Ittinaur became the first Inuk elected to the HOC, serving the Inuit majority electoral district of Nunatsiaq. Ittinuar’s electoral win was important for Inuit rights and recognition. During the Constitutional debates and patriation of the Constitution Act, 1982, a comprehensive Inuit land claim process was pushed for by Ittinuar (Cowie 2024). The Pierre Elliott Trudeau government agreed with Ittinuar, leading him to not only cross the floor from the NDP to the LPC but also the eventual creation of the territory of Nunavut in 1999 (Cowie 2024). Finally, Ethel Blondin-Andrew, a Dene woman, won a seat in the 1988 election to sit as the MP for Western Arctic, becoming the first Indigenous woman to obtain a seat in the HOC. Blondin-Andrew’s district included a portion of the traditional territory of the Dene and while MP, self-government negotiations began, and an agreement formulated in relation to all but one Dene community. Clearly, representation in the HOC has led to concrete change in settler Canadian institutions.

As these historic firsts of Indigenous representation were occurring, it is important to note the overall number of Indigenous MPs being elected was also increasing. Additionally, Indigenous representation within Canada’s major federal parties was growing inside and outside of the HOC. The 1972 Canadian federal election was the first where two Indigenous MPs were elected at the same time since 1870 (Cowie 2024). The 1997 Canadian federal election not only saw that number increase to four but also the first election that witnessed all three Peoples (First Nation, Inuit, and Métis) having someone with the same identity elected to the HOC (Cowie 2024; Cowie Forthcoming). This increase has continued, leading to a record 81 self-identified Indigenous candidates in the 2021 Canadian federal election and a total of 13 self-Identified Indigenous MPs elected to the HOC (Cowie and Midzain-Gobin 2022).

Involvement in Parliament does not appear to be a panacea, however. While individual legislators are part of decision-making for the state, Westminster-style party government means that only those in cabinet have significant influence over government decisions. Instead, the relationship remains one of structural dispossession (Coulthard 2014; Pasternak 2017; Simpson 2016; Starblanket 2019), as we outline further below. Moreover, individual legislators represent whole electoral districts (and all those included within them), not the specific ethnic, cultural, and national communities to which they belong. Having individual legislators who are Indigenous does not fundamentally change this dynamic, especially when they are not in positions of leadership within party caucuses. Recent resignation speeches by NDP MP Mumilaaq Qaqqaq (Innu) and British Columbia NDP MLA Melanie Mark (Nisga’a) attest to this. Mark, in her resignation speech, stated “institutions fundamentally resist change … particularly colonial institutions like this Legislative Assembly and government at large” (quoted in Larsen 2023). In an interview with Teresa Qiatsuq after her resignation, Qaqqaq noted that she felt disconnected from community (CBC News 2021) and also stated that institutions are not places to help Indigenous people without “an immense amount of pressure” (Zingel 2021). The gendered nature of settler colonialism also produces specific pressures for Indigenous women who are elected as representatives – as highlighted by Qaqqaq’s and Mark’s high-profile resignations.

These sentiments from legislators are echoed by Indigenous youth, who understand legislatures as colonial institutions not positioned to support Indigenous aspirations (Alfred, Pitawanakwat, and Price 2007). In their view, since legislatures were established through – and meant to maintain and further – dispossession, they are illegitimate. This follows from a theory of misrecognition (Coulthard 2014) where seeking public office, and succeeding as a representative, does not undo the colonial relationship but rather makes one a participant. This is especially acute with such a small number of elected Indigenous representatives. If the settler state perpetuates itself by undermining Indigenous nations, laws, citizenships, and other governance processes, maintaining or upholding that state can be viewed as illegitimate – regardless of whether Indigenous or non-Indigenous individuals are seeking, or elected to, public office. Indigenous representatives cannot change this relationship themselves when the state itself extends beyond the legislature. Considering the circumstances, we do not expect descriptive representation to change the underlying view of the state’s legitimacy.

Hypotheses

Considering the complex relationship between the Canadian state and Indigenous peoples, two possible hypotheses emerge. On the one hand, Indigenous people may see settler state legislatures as ongoing infringements on their sovereignty. This is the null hypothesis for our research. On the other hand, the descriptive representation of Indigenous people in Canadian legislatures signals that at least some portion of Indigenous Canada thinks that these legislatures could be used as a tool to advance the interests of their communities. Formally, these two hypotheses can be stated as:

H₀: Descriptive representation in the Canadian Parliament has no effect on diffuse support among Indigenous respondents.

H₁: Descriptive representation in the Canadian Parliament is associated with higher levels of diffuse support among Indigenous respondents.

Data and Methods

Our goal is to understand whether Indigenous people in Canada have different perceptions of diffuse support than non-Indigenous people, as well as whether descriptive representation changes those attitudes. To examine these questions, we used candidate-level data from Cowie (2024) and voter-level data from the 2019 and 2021 Canadian Election Study (CES). The candidate-level data contains information on whether an Indigenous candidate ran (and was successful) in each of Canada’s 338 electoral districts during these two elections. The CES data we use has been developed and fielded by a non-Indigenous team and was not written from an Indigenous standpoint. Data that meets these two criteria and would align with Indigenous Data Sovereignty best practices (Kukutai and Taylor 2016) is simply unavailable. However, given its status and comprehensiveness, our use of the CES shows the sorts of analyses that are now possible using this type of data, highlighting the importance of the lessons in our conclusion.

The 2019 CES and 2021 CES are stratified online surveys of Canadians, each consisting of a campaign period and post-election wave. Both surveys were administered in English and French around the federal elections of 2019 and 2021, respectively (Stephenson et al. 2020, 2022). Prior to 2019, most Canadian surveys have lacked enough Indigenous respondents to examine whether the political attitudes of these respondents differ from other groups. The 2019 and 2021 CES surveys overcome this problem through a combined 19,000 respondents for the post-election survey (PES). ⁹ Approximately 5.3% of the sample, or 2,474 respondents, self-identified as Aboriginal, which is approximately the same as the 2016 Census estimate of Indigenous peoples comprising 4.9% of the Canadian population (Statistics Canada 2017). Among these respondents, 70% identified as First Nations, 28% as Métis, and 2% as Inuit. ¹⁰ This means that First Nations respondents in the sample are over-represented compared to their proportion of Indigenous peoples (and not broken down by individual nations), while Métis and Inuit respondents are under-represented. ¹¹

The dependent variable in our analysis is diffuse support, which is operationalized through external efficacy (two items) and political trust (four items) which come from the PES. The external efficacy items measure agreement with the statements that “Those elected to Parliament soon lose touch with the people” and “The government does not care much about what people like me think.” The political trust questions ask respondents to indicate how much confidence they have in several institutions, which for our purposes include the federal government, the courts, the public service, and the police. Respondents’ answers to these questions were combined into an index variable with a Cronbach’s Alpha value of 0.8 for the 2019 CES and 0.79 for the 2021 CES, indicating very good internal consistency. ¹² Since the dependent variable is continuous, we use OLS regression to model its relationship with our independent variables. ¹³

In the first analysis, our main independent variable of interest is self-reported ethnic identity. The CES asked respondents to indicate which ethnic or cultural groups they belong to through a closed-ended question, with up to two open-ended response options available for those who do not identify with any of the specified groups. We recorded both the open- and closed-ended responses into a new variable which categorized voters as either white (British, English, Irish, Scottish, Welsh, French, French Canadian, Québécois, as well as other European descent), visible minority, and Indigenous (First Nations, Inuit, and Métis). We followed the definition of “visible minority” provided in the Employment Equity Act, which refers to “persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour” (Government of Canada 1995) and which Statistics Canada (2015) says “consists mainly of the following groups: South Asian, Chinese, Black, Filipino, Latin American, Arab, Southeast Asian, West Asian, Korean and Japanese.” We use the modal response (white) as the reference category. ¹⁴

In the second analysis, our main quantity of interest is the interaction between respondents’ identity (i.e., whether they are Indigenous) and whether an Indigenous person was elected to represent them in the election immediately preceding their survey (2019 or 2021). ¹⁵ Both analyses control for several alternative explanations of respondents’ levels of diffuse support. Specifically, our model controls for respondents’ household incomes, political interest, internal efficacy, education, age (both as a raw number and as age squared, to account for potential non-linear effects), gender, party identification, urbanization and whether or not the respondent lives in Québec. Diffuse support should be higher among individuals from groups with more sociopolitical resources or those who should have more reason to trust the political system. This means that all-else-equal, we expect those who are (separately) wealthier, politically interested, internally efficacious, university-educated, older, male, and who live in larger cities to have higher levels of diffuse support. Since the Liberal Party of Canada (LPC) was in government both before and after the 2019 election, we also expect Liberal identifiers to have higher levels of diffuse support than identifiers of other parties. Finally, we expect Quebecers – who have traditionally had a more fraught relationship with the federal government than those who live in other provinces – to have lower levels of diffuse support for the political system, as consistent with previous research (Snagovsky 2020). We also include a dummy variable for the 2021 election to differentiate between respondents from the two survey years.

Descriptive statistics for the variables considered in our analysis are available in the supplementary appendix (Table A1).

Results

Our findings for respondents’ levels of diffuse support in relation to their group identification are presented in Table 2. Specification 1 shows the overall results, while specification 2 includes the interaction between identity and descriptive representation. These results are also presented graphically in Figure 1.

Table 2. Perceptions of diffuse support for CES respondents (Post-election survey)

Robust standard errors in parentheses; * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 1. Perceptions of diffuse support for CES respondents (postelection survey).

As predicted, Indigenous respondents had lower levels of diffuse support compared to white-Anglo and white-Franco respondents, who are the reference category, by between 0.06 and 0.07 units (depending on the specification). This coefficient is statistically significant at the 0.05 level. By contrast, and consistent with prior work, visible minority voters had higher levels of diffuse support compared to white voters by approximately 0.04 units (Bilodeau 2014; Gidengil and Roy 2015; Hwang 2017). As we discuss further below, the different effects for Indigenous respondents and visible minority respondents underscore that these groups cannot be treated as interchangeable, given that meaningful differences exist in the experiences of individuals from both backgrounds. From model 2, we can also see that the interaction term between Indigenous identity and descriptive representation is not statistically significant, indicating that Indigenous respondents do not have higher perceptions of diffuse support when represented in the House of Commons by another Indigenous person. The marginal effects of this interaction are presented in Figure 2 and show that the average estimates for each condition are very similar. As a result, Hypothesis 1 is not supported.

Figure 2. Marginal effects for interaction between identity and descriptive representation.

Most of the effects for the control variables are statistically significant and in the expected direction. The coefficients for income (0.02), internal efficacy (0.06), and university education (0.13) are all positive and statistically significant, as is the coefficient for residing in a large town/suburb (0.04). The coefficients for age and age squared are both statistically significant, and their direction indicates age and diffuse support have a curvilinear effect – declining toward middle age before increasing again as respondents get older. However, the magnitude of these effects is quite small. Identifying with any opposition party associated with lower perceptions of diffuse support compared to identifying with the governing Liberal Party – indeed, these effects were by far the largest in our two models (−0.66 for Conservative, −0.38 for NDP, −0.39 for Bloc Québécois, and −0.46 for Green Party). Contrary to expectations, living in Quebec and identifying as a woman were associated with higher perceptions of diffuse support. ¹⁶

While the 2019 CES asked Indigenous respondents whether they were First Nations, Inuit, or Métis, the 2021 CES did not. Instead, it asked all respondents to indicate “What are the ethnic or cultural origins of your ancestors?” in an open-ended, write-in style question. While some First Nations respondents did specify which nation they were from, and some Inuit and Métis respondents did identify themselves accordingly, most Indigenous respondents (just over 50%) did not. Moreover, because the CES questionnaire used a split-sample design, not all respondents were asked all questions. The result of these two factors means that while there were enough Indigenous respondents to examine diffuse support when grouped together, there were not enough degrees of freedom to permit distinguishing between First Nations, Inuit, and Métis respondents in a multivariate analysis. However, the bivariate results are presented in Figure 3.

Figure 3. Box plot of diffuse support for Indigenous respondents by group.

As the graph makes clear, First Nations, Inuit, and Métis respondents have very similar median levels of diffuse support. ¹⁷ This suggests that despite settler colonialism operating differently vis-à-vis each group, members of each group ultimately come to similar conclusions about the state.

Discussion

Our findings show that Indigenous people maintain lower levels of diffuse support for the Canadian state than Canadians. Being represented by an Indigenous Member of Parliament does not change this. While notable, especially compared to respondents belonging to a visible minority, we do not find these results surprising considering the distinct tradition of separateness and the understanding of ongoing Indigenous sovereignty among those belonging to Indigenous nations. Settler dispossession deliberately undermines Indigenous sovereignty, with the settler state simultaneously extending settler sovereignty over Indigenous peoples and territories while undermining Indigenous political authority. Given this, a distrust of the state seems natural. Indeed, the long history of Indigenous organizing against government actions, and centering of Indigenous self-determination, affirms the ongoing relevance of this position.

There are also more material reasons for why descriptive representation may not change Indigenous peoples’ levels of diffuse support. Even in a democratic polity such as Canada, the state continues to harm Indigenous peoples. Our reading aligns with a number of other analyses that see this as a structural issue (Coulthard 2014; Pasternak 2017, Cowie 2024). The settler state exists through and continues to enact colonial relations of dispossession, which distinguishes the colonization faced by Indigenous peoples from other forms of marginalization. Our model highlights this distinction by showing the difference in levels of diffuse support between Indigenous individuals and visible minority respondents. Despite a long and ongoing history of racism, direct exclusion, and discriminatory practices by the Canadian state, visible minority individuals also display higher levels of diffuse support than Indigenous peoples. This is in part attributable to the intellectual history of separateness among Indigenous peoples discussed above but is also specifically tied to the question of land as it relates to colonization.

There is a distinctive process of land accumulation as a core feature of settler colonialism (Coulthard 2014; Starblanket 2019) that distinguishes Indigenous peoples from visible minority respondents. Land means something very different for Indigenous peoples than non-Indigenous peoples: philosophies, cosmologies, and governance traditions have been developed over millennia with land and other nonhuman being as integral participants (L. Simpson 2011; Williams 2018a; Ladner 2003; Henderson 2020). Dispossession undermines these relationships in ways that cannot be resolved only by representation in decision-making bodies. Colonialism also affects non-white peoples, as historic examples such as redlining, internment, and other deliberate exclusionary and dispossessive policies that affect people’s relation to land illustrate. However, the structural processes of settler colonial land accumulation produce a specific relationship between Indigenous peoples and the settler state that continually reproduces alienation for Indigenous individuals.

Settler colonization also produces gendered violence enabled and enacted by the state and settler society. While we might expect the intersection of Indigeneity and gender to affect respondents’ perceptions of diffuse support, separate analyses suggest that Indigenous men and women have similar levels of diffuse support. This is somewhat surprising, as research on diffuse support typically finds that women tend to be more trusting of state institutions than men. However, as Arvin, Tuck and Morrill (2013) articulate, Indigenous women and queer/two-spirit people have experienced the effects of settler colonization not just as Indigenous peoples but also by virtue of one’s gender and sexuality. The political concerns of Indigenous women can differ from those of other women: “The feminist concerns of white women, women of color, and Indigenous women thus offer differ and conflict with one another. In other words, within the context of land and settler colonialism, the issues facing Indigenous women, as inseparable from the issues facing Indigenous peoples as a whole, are resolved via decolonization and sovereignty, not (just) parity” (2013, 10).

Part of the reason for the Indigenous feminist focus on decolonization and sovereignty arises from the direct violence faced by Indigenous women at the hands of settler society and the state. As the National Inquiry into Missing and Murdered Indigenous Women, Girls and 2SLGBTQ+ individuals finds, the targeting of Indigenous women and girls amounts to genocide (National Inquiry into Missing and Murdered Indigenous Women and Girls 2019a). Indigenous women are targeted specifically because of their governance roles and responsibilities – killing them enables the state to further undermine Indigenous sovereignty and nationhood (National Inquiry into Missing and Murdered Indigenous Women and Girls 2019b, Vol. 1a, Ch. 4; A. Simpson 2016). As Mohawk scholar Audra Simpson (2016, 7) writes, the violence “is not without explanation, their so called “disappearances” are consistent with this ongoing project of dispossession.” To our minds, as three men conducting this research, this helps contextualize the leading role Indigenous women have taken in organizing against settler colonialism, the difficulties in legislatures we noted Indigenous women have spoken about, and our findings. Indeed, it may be somewhat surprising that, given the genocidal outcomes of gendered settler colonialism, Indigenous women do not have even lower levels of diffuse support.

Our model deliberately includes measures relating to the police and criminal justice system. Both institutions have – and continue to maintain – fraught relationships with Indigenous peoples and nations, logically contributing to the lower levels of diffuse support reported by Indigenous respondents. While there are many institutions responsible for the violence of settler colonialism, policing plays an outsized role. Indeed, the RCMP, and Canadian police forces more broadly, have been direct tools of colonial power (Bell and Schreiner 2018), including by apprehending children for the purposes of returning them to residential schools (Truth and Reconciliation Commission of Canada 2015). Such colonial logics continue to be implemented as policing services are mobilized against Indigenous land defenders (Crosby and Monaghan 2018; Pasternak 2017) and further enact gendered violence against Indigenous women and girls (National Inquiry into Missing and Murdered Indigenous Women and Girls 2019b; Simpson 2016).

Other, more policy-specific areas also reflect these structural conditions faced by Indigenous communities. Boil water advisories remain across dozens of communities as of the time of this writing, down considerably over the previous decade (Indigenous Services Canada 2023), but continuing to illustrate the neglect the Canadian state has shown for First Nations and other Indigenous communities. Other policy areas such as Indigenous child welfare, criminal justice, education, housing, and healthcare face similar trends. Ultimately, Indigenous communities are not adequately served by the state, and government attempts to solve problems run up against structural dispossession, maintaining devasting conditions within Indigenous communities and further highlighting why communities may not see the state as legitimate.

Given the structural dispossession and genocide faced by Indigenous peoples, it is perhaps unsurprising that respondents’ perceptions of settler state legitimacy did not shift when we controlled for descriptive representation. Why might this be the case? First, there are relatively few Indigenous representatives, and Indigenous peoples may not see any single representative as sufficient to bring about meaningful change. Perhaps a critical mass of representatives may shift this since a large number of representatives could have an impact on legislation passed in the Canadian legislature. However, it is also possible that the baggage of colonial institutions may be too great and new models may be needed. The Royal Commission on Aboriginal Peoples (RCAP) recommended a distinct house of parliament, an Indigenous Peoples House in addition to the upper (Senate) and the lower (House of Commons) houses (Government of Canada 1996). Perhaps a distinct house with the ability to meaningfully shift the decisions made (and unmade) by the existing government will bring greater levels of perceived legitimacy?

Second, substantive representation may be more likely to bring about the shifts that descriptive representation does not. Having representatives enact transformative change across policy areas like those discussed above to ensure Indigenous peoples’ needs are met and address long-standing marginalization may produce a significant shift in perceptions of state legitimacy among Indigenous peoples. Of course, undoing state neglect is important regardless, so it should not only be undertaken as a legitimacy-making project.

Further, substantive representation can be viewed through a governance lens. At the core of RCAP’s recommendation was a recognition of the importance of providing Indigenous peoples with a meaningful role in decision-making. Our current approach to consultation and accommodation is not sufficient, but perhaps a substantive shift to place decision-making authority in the hands of Indigenous peoples could mark the start. It may also allow us to return to the principle of separateness as envisioned in agreements such as the treaties and wampum: Indigenous peoples would regain authority over decisions that impact them, with their sovereignty resting alongside that of the settler state, rather than being incorporated into it. Fulsome implementation of the United Nations Declaration on the Rights of Indigenous Peoples (UNDRIP) can provide such an opportunity. Implementation processes for UNDRIP are ongoing in the province of British Columbia and by the federal government. Promises of greater authority are being made, but we have yet to see if they will be realized.

Conclusion: Looking Ahead

How do Indigenous peoples’ perceptions of state legitimacy compare to those of non-Indigenous peoples? How does increased descriptive representation impact their perceptions? Our analysis shows Indigenous peoples have lower diffuse support than members of both white and other racialized communities. The findings mirror our expectations given the ongoing nature of settler colonialism and the simultaneous insistence of Indigenous peoples on their own enduring sovereignties. While intuitive, showing Indigenous peoples’ attitudes toward the state through large-N analysis of Indigenous respondents in Canada underscores the work to be done if meaningful reconciliation is to be achieved, and a shared path forward is uncovered.

To our knowledge, this is the first large-N analysis of Indigenous people’s political attitudes toward the Canadian state. As such, it fills an important gap in our national-level understanding of Indigenous political behavior. While not moving beyond a state-centered analysis of Indigenous politics, it allows us to see the work that remains if government actors are to institute credible—and meaningful—policies that develop a shared path forward for Indigenous peoples and settlers.

We also hope this study serves as a first step in encouraging the field of political behavior to center Indigenous peoples’ attitudes, which offers opportunities to move beyond the traditional focus on settler-colonial institutions. This work ought to proceed through research that furthers our understanding of Indigenous experiences. As we have shown using data from the CES, recent advances in sampling methodology make this type of analysis possible from large-scale surveys – even those not aimed at Indigenous peoples. However, while surveys that contrast Indigenous and non-Indigenous populations on measures such as educational attainment, wealth, health, and well-being may offer a view into policy areas that need further attention, they offer little support for a better understanding of Indigenous peoples’ own perspectives and views.

Such studies may even further a deficit framing that focuses on Indigenous lack rather than strengths. This is one of the key lessons from Walter and Andersen’s nayri kati (Good Numbers): that to be useful, the numbers produced through quantitative methods must embody an Indigenous standpoint and work from Indigenous strengths (2013, 82). Further survey development might fruitfully engage with nayri kati to move past the sorts of omissions and erasures that obscure Indigenous peoples’ own attitudes in the existing literature. Put bluntly, the field is not there yet. In closing, we examine some of the limitations of current statistics for this type of analysis, using our own study as a guide.

First, our analysis has been guided by settler concepts. We develop our collective understanding of political legitimacy as “diffuse system support” and draw on the concept of “descriptive representation” – both ways of understanding social and political relationships that have been primarily used to describe the settler state. Existing datasets make this type of research easier to do, thus incentivizing scholarship working from settler perspectives. To change this trend, survey development needs to be done not only in consultation with Indigenous experts but in genuine partnership.

Another limit to the existing data comes at the level of identity. The 2019 CES does a good job of dis-aggregating between distinct Indigenous Peoples, reporting responses by First Nations, Métis, and Inuit separately. This is not always the case in surveys, and positive on two fronts. Normatively, dis-aggregation helps to avoid the colonial move to homogenize distinct Indigenous peoples together, described as “Aboriginalism” (Alfred and Corntassel 2005). Analytically, dis-aggregation allows for greater precision in our model and specificity in our findings – for instance, we can identify whether levels of diffuse support differ between Métis and First Nations (or Inuit). Despite each people being dispossessed through settler colonialism, the specifics of that dispossession, and each peoples’ interactions and relationship to the settler state, differ. While our model shows no significant difference between the peoples’ perceptions of state legitimacy, this conclusion is only possible by having respondents self-identify.

For the 2021 CES, the decision to dis-aggregate identity was reversed: the identity variable asks whether respondents were Indigenous or not but does not further prompt respondents to identify if they are First Nations, Métis, or Inuit. The previous question (cps21_origin) asked for a respondent’s “ethnic or cultural origins” and offered five spaces to answer with. This additional question does allow for greater precision by being open-ended and linking to Statistics Canada’s census-linked list of examples of ethnic and cultural origins, which lists a number of distinct nations and Inuit regional groups. Normatively, this strikes us as a positive step: allowing for self-identification remains a best practice and having open-ended questions allows for greater precision for individuals to identify as they wish, not as the pre-selected categories allow. However, it is more difficult to assess analytically; indeed, there is some nonalignment between those who self-identify as Indigenous, and then the specific ethnic or cultural origins they noted. In total, 3% of Indigenous respondents did not then describe any ethnic or cultural identity the authors could credibly see as belonging to one of the Canadian administrative categories, while 41% simply provided no further information. This meant that we were unable to examine differences between First Nations, Inuit, and Métis political attitudes for close to half of Indigenous respondents in the 2021 CES.

Our approach, while systematic and rigorous, also leaves opportunities to disagree. For example, those respondents who identified themselves as “American Indian” were grouped into First Nation, while those who responded they were “Canadian” were included as outside the trichotomy we describe above. We can, and indeed, should, debate these categorizations not least because they touch on self-identification as an important political issue in settler states across the world. That administrative categories define state understandings of identity is a colonial problem at its roots (de Costa 2014), and we have not engaged with issues of Métis identity either (Gaudry 2018; Gaudry and Andersen 2016). For our analytical purpose, it also introduces an imprecision into the models and final results: if we are not sure of a particular individual’s background (including their understandings and background knowledges or common sense), or the strength of their connection with Indigenous communities, then including them in our assessments of Indigenous peoples’ feelings toward the settler state potentially muddies our analysis.

Connected to this, the way that the data was collected does not allow us to determine whether or not the individual respondents were from the same nation, or even People, as those who sought to be elected. This matters for our analysis, as stronger relationships being respondents and those seeking election may have an impact on descriptive representation’s effectiveness. This may also work in the reverse: tensions between nations may have a negative impact on respondent’s willingness to vote for a candidate, or their sense of being represented if the candidate is elected – and hence, it may impact their perception of the state’s legitimacy. Given the existing tools we have available, however, we cannot know. Likewise, we do not know the extent to which Indigenous respondents were familiar with the identities of candidates running in their constituencies. It may be that Indigenous respondents who did not know of their MP’s identity would see the state more favorably if they knew their MP was Indigenous. However, if these respondents are not interested in learning about who represents them in colonial institutions in the first place, this likewise does not bode well for the potential impact of descriptive representation in these bodies.

The geographic boundaries used by the CES also produced limits for our analysis. Respondents identify themselves in relation to the province in which they reside. This is an important aspect of our model because it helps us include the discussion of descriptive representation. However, provincial boundaries do not align with the national identities Indigenous peoples hold. For example, Inuit homelands span the territories of Nunavut and the Northwest Territories, and provinces of Quebec and Newfoundland and Labrador, or the Anishinaabe Nation extends across the provinces of Manitoba, Ontario, and Quebec. When it comes to understanding Indigenous peoples’ relationship to Canada, then, provincial boundaries, and the success of Indigenous candidates in an election in a different province may have important resonances, even if that does not fit with a typical analysis of descriptive representation. A Cree candidate elected in British Columbia may not mean the same thing as the election of a Coast Salish candidate but needed to be treated as such in our model. Models may be tweaked, but if we were to work from the existing data available, the changes would make any resulting findings insignificant or not worthwhile.

Each of these issues highlights why Walter and Andersen’s (2013) nayri kati is so important. Their conceptualization shows not only that statistical methods can be used by Indigenous peoples, but that they can be deeply effective when they work from Indigenous standpoints. This means not only that research programs and projects ought to be developed with Indigenous participation, but that to be especially effective, the data and models should be developed from Indigenous peoples’ based on their own knowledge and philosophies. This helps reduce the “recognition gaps” that appear in existing censuses, studies and data sets (Kukutai and Walter 2015; see also: Kukutai and Taylor 2016). That the question of identity became ensnared in colonial categories may not be entirely avoided, but Indigenous communities and organizations have developed expertise on how to ask questions that allow us the insights we feel the field needs to begin to open its eyes to. Organizations such as the First Nations Information Governance Centre or Inuit Tapiriit Kanatami hold this kind of expertise in the Canadian case, but others may also do so in other settler colonial contexts such as the University of Arizona’s Native Nations Institute in the United States of America. While the CES was not designed primarily to be of relevance for Indigenous peoples, it has been fielded in order to help us better understand Canadian political behavior. That the Canadian state’s existence stems from its treatment of Indigenous peoples on the territories it now claims strikes us as a good enough reason to seek to avoid the colonial trap of settler identity politics.