Introduction
Following the May 2020 police killing of George Floyd in Minneapolis, protesters gathered in cities across the USA, Canada, and around the world to protest institutional racism and police violence. While the 2020 protests brought renewed attention to the issue of racism—particularly anti-Black racism—and policing, police violence and anti-racism activism are not new to Canada. The Montreal North police shooting of an unarmed youth, Freddy Villanueva (Maughan et al., Reference Maughan, Kilpatrick and Lift2008), sparked anti-racism protests in 2008 (Canadian Broadcast Corporation (CBC), 2010). After the incident, internal reports of policing in Montreal North—a neighbourhood with a large immigrant and minority population—detailed instances of officers making racist comments and described youth living in constant fear of police. As a result of these investigations, the Montreal police promised to create a public database for tracking of the racial backgrounds of people who were stopped or investigated by police (CBC, 2017b)—a database that ultimately revealed powerful evidence of racial profiling in Canada (Feith, Reference Feith2019; Livingstone, Reference Livingstone2019). The problem of racial profiling may even be worsening: Research shows that, in Toronto, the gap in police stops between White and Black youth actually widened between 2003 and 2012 (Meng, Reference Meng2017).
In our article, we first review the literature on the relationship between race, police violence, and political behaviour in the USA. Existing American studies suggest that perceptions of injustice by police can mobilize protest activity among disempowered social group members who are disproportionately impacted by police surveillance and violence (Walker, Reference Walker2014; Reference Walker2020; Lerman and Weaver, Reference Lerman and Weaver2014; Rios, Reference Rios2011). Although sociologists and criminologists have considered racial differences in attitudes towards the police in Canada (Wortley and Owusu-Bempah, Reference Wortley and Owusu-Bempah2009; O’Connor, Reference O’Connor2008; Sprott and Doob, Reference Sprott and Doob2014; Cao, Reference Cao2011), much of this research depends on data collected in the 1990s and early 2000s. Furthermore, although there is excellent qualitative research on race and policy making in Canada (Livingstone, Reference Livingstone2024; Livingstone et al., Reference Livingstone2018), and some research showing that negative attitudes towards the police have serious social consequences—for instance, Canadians who doubt the legitimacy of the police and have low trust in the police are less likely to report crimes (Schulenberg et al., Reference Schulenberg, Chenier, Buffone and Wojciechowski2017)—quantitative scholars have otherwise largely neglected the political consequences of attitudes towards the police, including whether confidence in the police is related to protest activity. Perhaps ironically, although political science is concerned with the legitimate use of force, relatively few academic political science studies consider the police or perceptions of the police in Canada (although see Bourgault and Gow (Reference Bourgault and Iain Gow2002) for an analysis of the Sûreté du Québec). Furthermore, Canadian political science has, at least historically, been notably silent on the topic of race (Thompson, Reference Thompson2008), focusing on certain identities (such as national, linguistic or regional identities) over others (such as racial or Indigenous identities) (Nath, Reference Nath2011).
Our article addresses two gaps in the Canadian literature. First, we offer an updated look at the relationship between race and on confidence in the police in Canada, independent of other potentially related factors, such as immigration status and socioeconomic status. Second, we clarify the relationship between confidence in the police and protest behaviour. We do so using three datasets collected by Statistics Canada: the General Social Survey (GSS) on Canadians’ Safety Cycle 34 (GSS34; April 2019–March 2020; n = 22,277), the GSS on Social Identity Cycle 35 (GSS35; August 2020–December 2021; n = 32, 079) and the Impacts of COVID-19 on Canadians’ Experiences of Discrimination (ICCED) (August 2020; n = 35,718).Footnote 1
To clarify who has confidence in the police, we first delve into the univariate relationships between confidence in the police and race, income and immigration status. Difference of means tests reveal that Black Canadians and Indigenous peoples consistently express significantly less confidence in police as compared with White Canadians and that Canadians living in poverty express less confidence in police compared with those living above the poverty line. Black respondents consistently express the lowest confidence in police. Average confidence differences (compared with White Canadians) are less distinct for the category grouping all other People of Colour (POC). Statistically significant differences between immigrants compared with respondents born in Canada are unstable across surveys. But all our variables’ group differences are statistically significant in the multivariate ordinary least squares (OLS) model. We test the multivariate relationship a few ways for robustness, and the results hold across modelling strategies. Regressing confidence in the police on race reveals that Black, Indigenous and other POC express lower confidence in the police, independent of income and immigration status, and controlling for other variables (gender, region, generational group and education). Regressing protest on confidence in the police reveals a significant association between feelings towards the police and this important form of political engagement: those with the least confidence in the police are the most likely to protest. It should be noted that, although our research is motivated by the issue of racial injustice, our outcome measure asks respondents whether they protested in general in the preceding 12 months at the time of survey,Footnote 2 without asking respondents to specify what they were protesting. As such, we do not use our data to make claims about the relationship between race and specific types of protest.
Race, Policing and Protest
There is a longstanding finding that White Americans perceive the police to be more legitimate or to express greater confidence in the police compared with Black Americans (Hurwitz and Peffley, Reference Hurwitz and Peffley2010). There is also clear evidence that police violence and policy threats that disproportionately impact Black communities mobilize Black Americans’ political participation in non-electoral politics. For instance, Lerman and Weaver (Reference Lerman and Weaver2014) find that being stopped by police is positively associated with engagement in civic associations. Other studies show that marginalized youth who have been harassed by the police are more likely to engage in protest activities (Rios, Reference Rios2011). Similarly, Williamson, Trump and Einstein (Reference Williamson, Trump and Levine Einstein2018) created a new dataset that brings together data on Black Lives Matter protests with contextual information about the areas where the protests took place to show that Black Lives Matter protests were more likely to be organized in cities with higher rates of police-caused deaths.
Although direct contact with the carceral state (that is, through incarceration) reduces Black Americans’ political interest (Drakulich et al., Reference Drakulich, Hagan, Johnson and Wozniak2017) and participation (Lerman and Weaver, Reference Lerman and Weaver2014; Walker, Reference Walker2020), proximal contact motivates political participation in non-electoral activities (Walker, Reference Walker2014; Reference Walker2020). Proximal contact is defined as that of “individuals who have not had personal contact with the criminal justice system but who have had indirect contact” such as having family members who have had interactions with the police (Walker, Reference Walker2014: 810). Drakulich et al. (Reference Drakulich, Hagan, Johnson and Wozniak2017) also show that other forms of indirect contact—being stopped by the police without being arrested—increases political interest. The theory behind the causal mechanism is that indirect contact increases feelings that the criminal justice system unjustly punishes or surveils members of one’s social in-group. This activates feelings of group threat that can mobilize political engagement without the suppressing effects of incarceration.
While state surveillance and carceral policy threats disproportionately impact Black Americans, there is evidence from the USA that similar processes also mobilize other affected communities of colour. For instance, Cho, Gimpel and Wu (Reference Cho, Gimpel and Wu2006) find that news about the Patriot Act had a strong effect in mobilizing the political participation of Arab Americans. A substantial body of literature suggests that group-based threats in general—including but not limited to the threat of surveillance and police violence—can mobilize political engagement, particularly among communities of colour. Evidence shows that threatening policy environments have mobilized political engagement among Latin communities in the USA (Ramírez, Reference Ramírez2013; Merolla et al., Reference Merolla, Pantoja, Cargile and Mora2013), including mobilizing participation in rallies to protest anti-immigrant legislation (Zepeda-Millán, Reference Zepeda-Milla’n2017). Similarly, policy threats have mobilized political participation among Asian American immigrant communities (Ong, Reference Ong2011) (for a review of this literature see Walker (Reference Walker2020)).
Evidence also shows that the mobilizing effects of group-based policy threats are not limited to those personally impacted by threatening policies or government procedures. Rather, group-based policy threats can have spillover effects mobilizing participation among members of disproportionately affected communities. Although not personally impacted, members of disproportionately affected communities “may view their experiences in terms of group-based grievances” (Walker, Reference Walker2020: 122). Anti-immigration policies not only mobilize the political participation of undocumented workers but also motivate the political engagement of documented workers and citizens alike (Aranda, Menjívar and Donato, Reference Aranda, Menjívar and Donato2014).
Canadian scholarship shows that Indigenous and Black Canadians are disproportionately impacted by the criminal justice system. For instance, Monchalin (Reference Monchalin2016) notes that, although Indigenous peoples made up 4.3 per cent of Canada’s total population in 2011, Indigenous adults accounted for nearly one-third (28%) of custody sentences. And while the connection between attitudes towards the police and political behaviour is less well-studied in Canada, there is a growing effort to understand how race and political action shape political decision making. For instance, Livingstone (Reference Livingstone2024) shows that a culture of Black political activism and multi-racial coalition building in Ontario resulted in more race-conscious public policies. For instance, to address the perceived problem of “youth gangs” (which disproportionately affected Black youth), Ontario developed out-of-school programs for children and adolescents. By contrast, in Quebec, race became a stand-in for conflicts over culture and ethnicity, which limited the effectiveness of Black political activism on policy making related to Black youth and resulted in more top-down, less race-conscious policies.
With respect to police violence and attitudes towards police in Canada more specifically, a number of recent high-profile incidents have sparked public outrage. Canadian police departments have historically not collected data on the race of those shot by the police, a serious gap in the statistics on crime and violence in Canada (c.f., Zay, Reference Zay1963). However, in 2017 the Canadian Broadcast Corporation (CBC) published Deadly Force, a country-wide database which tabulates every person who was killed by the police since 2000. The CBC’s data shows that police violence disproportionately affects communities of colour: Black Canadians (who constitute 3.4% of Canada’s population) represent 9 per cent of total fatalities, while Indigenous peoples (who constitute 4.8% of Canada’s population) represented 15 per cent of fatalities (Marcoux and Nicholson, Reference Marcoux and Nicholson2018; Kim, Reference Kim2019). That same year, a Royal Canadian Mounted Police (RCMP) briefing note written for Public Safety Minister Ralph Goodale revealed that, in the past decade, over one-third of RCMP shooting victims across Canada were Indigenous (Freeze, Reference Freeze2019). In 2017, Quebec’s Human Rights Commission also ruled that the Montreal police department and an officer had racially profiled a youth in Montreal and ordered the department and officer to pay a total of CA$17,000 in damages (CBC, 2017a).
In 2018, the Ontario Human Rights Commission published a report revealing that Black Torontonians were nearly 20 times more likely to be fatally shot by the Toronto Police Service than their White counterparts (Ontario Human Rights Commission, 2018). In the same year, a Quebec Superior Court Justice granted Black plaintiffs the right to proceed with a class-action lawsuit against the City of Montreal for racial profiling by police (CBC, 2019). Most recently, a 2020 breakdown of use-of-force tactics revealed that the RCMP in New Brunswick had brandished guns at a five-to-one rate higher than non-lethal weapons such as Tasers, batons, pepper spray and the like (Boynton, Reference Boynton2020). The report came on the heels of a series of two high-profile police shootings of Indigenous people in New Brunswick (The Canadian Press, 2020).
A number of Canadian studies in sociology and criminology have considered group-based variation in attitudes towards the police. Research shows that older Canadians and women tend to have more confidence in the police (Hu et al., Reference Hu, Dai, DeValve and Lejeune2020). There is also strong evidence that attitudes towards the police vary by race, although unfortunately much of the existing research in this area relies on data collected in the 1990s and early 2000s. For instance, a survey of Torontonians that oversampled Black and Chinese respondents reveals that Black Torontonians expressed the most negative evaluations of police (Wortley and Owusu-Bempah, Reference Wortley and Owusu-Bempah2009). The study also showed that in Toronto, 80 per cent of Black respondents indicated that they believe the police treat Black Canadians worse than White Canadians. By contrast, 59 per cent of White respondents and 50 per cent of Chinese respondents expressed the same view (Wortley and Owusu-Bempah, Reference Wortley and Owusu-Bempah2009). A more detailed study of Chinese attitudes towards police in the Metro Toronto affirms these findings, revealing that Chinese residents express a great deal of ambivalence towards the police and on average express only a marginal level of satisfaction (Chow, Reference Chow2002).
The findings from the Toronto study are substantiated by nationally representative data collected for Statistics Canada’s General Social Survey (GSS) in 1990. GSS data show that POC across Canada express lower evaluations of the police than do White Canadians (O’Connor, Reference O’Connor2008; Cao, Reference Cao2011). Using the same data but considering variation across groups and regions, Sprott and Doob (Reference Sprott and Doob2014) show that, in Quebec and Ontario, Black respondents are more likely to report more negative interpersonal interactions with police than White respondents. By contrast, in Alberta, Indigenous respondents are more likely to report more negative interpersonal interactions than White respondents. Cheng (Reference Cheng2015) also finds that people of colour, and Indigenous peoples in particular, express the lowest satisfaction with local police. To our knowledge, no quantitative researchers have yet investigated the relationship between attitudes towards the police and protest activity in Canada. Finally, existing scholarship shows that both race and immigrant status independently impact attitudes towards the police, and thus, any study seeking to clarify the relationship between race and attitudes towards the police must account for immigrant status (Jung, Greene and Sprott, Reference Jung, Greene and Sprott2021).
Drawing on this literature, we hypothesize that, in Canada, social group members who experience disproportionate police surveillance and violence, including Black Canadians, Indigenous peoples and other POC, people living in poverty, and immigrants will express significantly lower levels of confidence in the police (Hypothesis (H) 1). Extrapolating from proximal contact US literature, we hypothesize that low levels of confidence in the police are associated with a higher likelihood of participating in protest activity in Canada (H2).
Methods
To test our hypotheses, we use three datasets collected by Statistics Canada: the General Social Survey (GSS) on Canadians’ Safety Cycle 34 (April 2019–March 2020; n = 22,277; dataset version 3), the Impacts of COVID-19 on Canadians’ Experiences of Discrimination (ICCED; August 2020; n = 35,718; dataset version 1) and the GSS on Social Identity Cycle 35 (August 2020–December 2021; n = 32,079; dataset version 2). Showing that our main findings replicate across three separate Statistics Canada studies increases confidence in our results. We chose to use Statistics Canada data because its very large sample sizes ensured that we had enough respondents to identify the relationship between attitudes towards the police and distinct racial groups, rather than, for example, collapsing Black, Indigenous and POC into a single category.
Although difficult to access,Footnote 3 there are benefits to using Statistics Canada data in addition to the very large sample sizes. For instance, the General Social Surveys are carried out by Statistics Canada using a “common telephone frame” that combines landline and cellular telephone numbers associated with one address, including cellphone-only households (Statistics Canada, 2019). Statistics Canada has respondent consent to link administrative records, such as tax reports, with the organization’s surveys (exempting respondents who object to linking their personal records). As such, the GSS income data are drawn from tax or other administrative files, rather than relying on self-reported income. By contrast, the ICCED was conducted online, as part of a broader series on coronavirus disease 2019 (COVID-19)’s impacts (Statistics Canada’s Questionnaire Design Resource Center, 2022). This non-probabilistic sample of self-selected volunteers was crowd-sourced and managed by Statistics Canada.
For all datasets, missing values on predictor variables were by imputed using the MICE package in R to avoid unnecessary data loss (Azur et al., Reference Azur, Stuart, Frangakis and Leaf2011; Zhang, Reference Zhang2016) (for a summary of the missing values see the supplementary material, Table S6). Missing values on the main independent variables and outcome variables were dealt with using list-wise deletion.Footnote 4 Statistics Canada generated weights for the GSS datasets, but not for the ICCED. We used an iterative raking procedure to create survey weights to adjust for region, gender, age group, education level, and race on the basis of population data taken from the 2021 Canadian census using the survey package in R. We applied these weights to the ICCED dataset’s analysis.Footnote 5 We also re-ran the regression using our raked weights on all three datasets, using the 2016 census population distributions on pre-2021 surveys, and provide that version in the appendix as a robustness check (Supplementary Materials S4).
In our analysis, we consider two outcomes of interest: confidence in police and protesting. The measure of confidence in police is obtained in all three datasets through the same question-wording. Responses to the confidence items were measured using Likert-type scales that ranged from a great deal to none at all, which we transform into a numeric variable (from 0 (none) to 1 (a great deal)).Footnote 6
Protest is operationalized using a dummy variable indicating whether a respondent had protested at least once in the past 12 months (protested = 1). The measure of protest is only available in the GSS35 (August 2020–December 2021; protest event recall starts in August 2019). We make use of the Canadian Election Study (CES) 2021, which also measured protest activity this way, in the Supplementary Material.Footnote 7 It should be noted that while this variable asks respondents whether or not they protested during the past 12 months, neither survey asks respondents what type of protest they participated in.
Our main independent variables of interest are race, with separate categories indicating whether a respondent is White, Black, Indigenous or another POC (defined as all other Canadians who are neither White, Black or Indigenous), immigration status, and living in poverty, which is defined by an income under CA$20,000. For each of these variables of interest, we treat the modal category (the category with the largest number of respondents) as the reference category, that is: White Canadians, Canadians who did not immigrate, and Canadians who are not living in poverty. We control for level of education (less than college, some college or trade or completion of a four-year university degree), region (British Columbia, West, Ontario, North, Quebec and Atlantic) and generation (Generation Z and Millennial, Generation X, Boomer or Silent generation). We explicitly note the reference categories in the regression results’ tables and interpretation.
Findings
Who has confidence in the police?
Difference of means tests reveal that, as hypothesized, Black Canadians and Indigenous people have significantly lower confidence in police compared with White Canadians (Figure 1). Black Canadians average confidence in police estimates are: GSS34 −0.037, ICCED −0.258, GSS35 −0.078; Indigenous respondent estimates are −0.097, −0.132 and −0.105. The average for other people of colour is more variable—only statistically significantly lower average confidence in the ICCED survey (−0.081). Similarly, immigrants expressed average levels of confidence in police, which fluctuate in their statistical significance difference compared with those born in Canada (ICCED −0.0096 and GSS35 0.044, while the GSS34 0.008 estimate is not significantly different from zero). Those living in poverty (which is Canada Revenue Agency (CRA) tax-reported income verified data) expressed significantly lower levels of confidence in police compared with those living above the poverty line.Footnote 8
Difference of Means in Confidence in Police by Social Group Membership, across All Datasets.
We model these variables (four-category race, income and place of birth), regressing them on police confidence. The OLS model in Table 1 controls for gender, generational cohort, geographical region and education levels (see S3 in the Supplementary Material for the full results). Across the three surveys, Black, Indigenous and other POC all express significant lower confidence in the police compared with their White counterparts.
OLS Dependent Variable: Police Confidence
Note: * p < 0.1; ** p < 0.05; *** p < 0.01.
The estimates range from −0.076 association from the ICCED’s Black respondents to −0.015 from the GSS34’s Indigenous respondents. Black respondents’ negative association to confidence in police is the strongest relative to White respondents in two of the three surveys, and practically overlaps with Indigenous peoples’ lowest confidence in the GSS34. The estimates for other POC are about half the magnitude in the first two surveys but not statistically different from White respondents in the GSS35. Respondents whose verified income puts them under the poverty line also expressed significantly lower confidence in police: Estimates ranged between −0.008 (p < 0.05) in the GSS35 and −0.012 (p < 0.01) in the GSS35. Birthplace differences are also statistically significant, with more than 99 per cent confidence intervals across surveys; foreign-born respondents expressed less confidence in the police compared with Canadian-born respondents in all three surveys. The adjusted R 2 across the models regressing sociodemographic variables on attitudes towards the police was low, ranging between 0.03 and 0.09, suggesting that, although attitudes towards the police are indeed related to these respondent attributes, many other factors explain variation in attitudes towards the police.
Unfortunately, we are unable to provide tests for alternative explanations, given a lack of alternative theories theories and data-collection restrictions from Statistics Canada’s survey design. For example, only the GSS34 asks respondents whether they have interacted with the police in the last 12 months. However, we cannot use this item in our analysis because the number of people reporting contact is very small—and, furthermore, there are no follow-up questions to clarify whether the contact was negative or positive (for example, feeling heard when reporting a crime or having a dispute solved by police). Moreover, the GSS34 did not ask about protest behaviour. Scholars with access to large Canadian survey samples would do well to measure these questions and proximal contact items for future alternative hypothesis testing.
Do attitudes towards the police motivate political participation?
The GSS Cycle 35 (2020) includes an item asking respondents about their protesting habits “in the last 12 months,” which means the window of protest participation recall spans August 2019–December 2021. Using logistic regression, we regressed protesting on confidence in the police, race, poverty and immigration status and a vector of control variables to reveal that, as hypothesized, low confidence in the police is associated with greater protest activity (Table 2; refer to S5 in the Supplementary Material for fully specified model). The results are different for each racial or ethnic group: Black and Indigenous respondents are more likely to protest than White respondents, although the estimate is not statistically significant for Black respondents (holding all other variables constant). Among other non-Indigenous POC, the probability of not protesting is statistically significantly different to White respondents. We also analysed the CES 2021 protest data in the Supplementary Material (S15). The CES 2021 has a smaller sample size (n = 7401) but has the benefit of including party identity and ideology. Furthermore, the CES 2021 was conducted between September 23 to October 4, 2021, overlapping with the latter part of the GSS35’s sampling period. We find striking resemblance in the estimate effects in terms of directionality and statistical significance, buttressing the GSS35 results.
GSS v35 Protest Model
Note: * p < 0.05; ** p < 0.01; *** p < 0.001.
For ease of interpretation, we have plotted the predicted probability of protesting given respondents’ confidence in police, holding the other model variables constant at their means. Figure 2 shows how, when police confidence increases (from zero, indicating no confidence, to one, indicating total confidence), the predicted probability of protesting drops from 0.17 given no confidence in police to almost no protest activity (0.02) given full confidence. See S3 in the Supplementary Material for its comparison the protest model using CES 2021 data.
Effect of Confidence in Police on Protesting in the GSS35.
Discussion and Conclusion
Recent incidents of police violence have prompted massive protests across the USA, Canada and the world. In June 2020, two separate police killings in New Brunswick—of Chantel Moore, a young Tla-o-qui-aht mother, and Rodney Levi, a Mi’kmaw father from Metepenagiag First Nation—prompted protests across New Brunswick and in Halifax (CBC News, 2020; Bailey, Reference Bailey2020). COVID-related wellness checks and mental health crises that ended in deaths after police interactions, including in the case of Regis Korchinski-Paquet, an Afro-Indigenous woman living in Toronto, also called into question why police are in charge of health service interactions (Britneff, Reference Britneff2020). Protests, public debates and activism by the Liberal party’s Black caucus may have motivated the Liberal government’s mention of police reform in their September 2020 Throne Speech (Wherry, Reference Wherry2020; Berthiaume, Reference Berthiaume2020).
However, despite Trudeau’s recognition that there was systematic racism within the RCMP and across other institutions (Tunney, Reference Tunney2020), in a federal system, that provincial and municipal policing units are not necessarily responsive to a Prime Minister’s expressed concerns. Therefore, federal speeches or even reforms may have minimal impacts on the day-to-day interactions between many Canadians and their local police. Systematic police reform requires coordinated national, provincial and municipal efforts. Unfortunately, in certain provinces, policy makers are reluctant to accept the idea that racism is institutionalized or systemic, complicating the political calculus for reform (c.f. Livingstone, Celemencki and Calixte, Reference Livingstone, Celemencki and Calixte2014).
Our work contributes to the existing scholarship by offering an up-to-date look at confidence in police and showing that confidence in police varies systematically by social group membership. Echoing, supporting and extending the period of observation of previous studies (Wortley and Owusu-Bempah, Reference Wortley and Owusu-Bempah2009; O’Connor, Reference O’Connor2008; Sprott and Doob, Reference Sprott and Doob2014), we find that a number of disempowered social group members express significantly lower levels of police confidence. Notably, Black Canadians and Indigenous peoples (and, in some surveys, POC), as well as those living in poverty express significantly lower levels of confidence in police relative to White Canadians and to Canadians living above the poverty line.
Our present work does have limitations. For instance, there is an important and growing literature on attitudes towards the police that goes beyond feelings towards the police, including research on attitudes towards the use of violence against police in the USA (Maguire et al., Reference Maguire, Barak, Cross and Lugo2018, Reference Maguire, Barak, Wells and Katz2020; Tyler et al., Reference Tyler, Barak, Maguire and Wells2018) as well as Americans’ reactions to police use of force (Maguire et al., Reference Maguire, Tyler, Khade and Mora2024). Our present work is more limited in scope, considering only feelings towards the police, using a scale that asks respondents to rate how much confidence they have towards the police. However, future research in Canada should consider the relationship between race and other attitudes towards the police. It would be especially helpful for future studies to collect data on direct contact with police beyond carceral interactions, such as proximal contact, and qualitative recall of how police interactions unfolded. Additionally, survey experiments testing the relation between confidence in the police and willingness to protest, varying when the subject of the protest is police misconduct, would increase our understanding of voters’ complex decision-making processes.
While our findings offer an important update to the existing Canadian sociology and criminology literature on variation in Canadians’ attitudes towards police, our most important contribution is to the political science canon, where less attention has been paid to the political consequences of attitudes toward the police. Our research corroborates existing American research on the relationship between attitudes towards the police and political behaviour (for example, Gillion, Reference Gillion2013; Walker, Reference Walker2020). Specifically, we find a significant association between attitudes toward the police and protest activity. This finding supports the intuition that Canadians who are more dissatisfied with the police are more likely to demand democratic accountability by turning to protest activity. Of course, we must again stress a scope condition—the survey item we used asks respondents about whether they protested but not what they were protesting. As such, we cannot say whether attitudes towards the police are associated with particular types of protest. Our analysis is thus about protest in general, and not about a specific type of protest, such as Black Lives Matter protests. Furthermore, because our work focuses on correlations, we cannot say conclusively whether disaffection with the police causes protesting. We hope that the findings in our article mobilize future research on the causal effect of political attitudes, social group membership, and protest in Canada.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0008423925100863.
Financial support
We thank the following institutions and programmes for funding our research throughout the research’s development: Social Sciences and Humanities Research Council (#430-2021-00294), McGill University Mobility and Simon Fraser University’s Visiting Research Scholar and Political Science department funding. This research was conducted at the Quebec Interuniversity Centre for Social Statistics (QICSS), part of the Canadian Research Data Centre Network (CRDCN). This service is provided through the support of QICSS’ Member Universities, the province of Quebec, the Canada Foundation for Innovation, the Canadian Institutes of Health Research, the Social Sciences and Humanities Research Council, the Fonds de Recherche du Québec, and Statistics Canada. All views expressed in this work are our own.
Competing interests
The authors report there are no competing interests to declare.
Open access
