Socio-cultural correlates of self-reported experiences of discrimination related to COVID-19 in a culturally diverse sample of Canadian adults

Minorities and marginalized groups have increasingly become the target of discriminatory actions related to the COVID-19 pandemic. Detailed information about the manifestation of COVID-related discrimination is required to develop preventive actions that are not stigmatizing for such groups. The present study investigates experiences of perceived discrimination related to COVID-19 and its socio-cultural correlates in a culturally diverse sample of adults in Quebec (Canada). An online survey was completed by 3273 Quebec residents (49 % 18−39 years old; 57 % female; 49 % White). We used multivariate binomial logistic regression models to assess prevalence of COVID-related discrimination and to investigate socio-cultural correlates of reasons and contexts of discrimination. COVID-related discrimination was reported by 16.58 % of participants. Non-white participants, health-care workers and younger participants were more likely to experience discrimination than White, unemployed and older participants, respectively. Discrimination was reported primarily in association with participants’ ethno-cultural group, age, occupation and physical health and in the context of public spaces. Participants of East-Asian descent and essential workers were more likely to report discrimination because of their ethnicity and occupation, respectively. Although young people experienced discrimination across more contexts, older participants were primarily discriminated in the context of grocery stores and because of their age. Our findings indicate that health communication actions informed by a social pedagogy approach should target public beliefs related to the association of COVID-19 with ethnicity, age and occupation, to minimize pandemic-related discrimination. Visible minorities, health-care workers and seniors should be protected and supported, especially in public spaces.


Introduction
The COVID-19 pandemic is affecting global health, societies and economies around the world, and fueling fear, discrimination and xenophobic discourses in the general population. There is considerable intersectionality between the pandemic's impact on health and issues of structural inequality and racism in many countries. Discrimination refers to the unjust treatment of an individual based on a critical lens to develop strategies aimed at creating an egalitarian and inclusive environment. It stresses the importance of acknowledging and critically examining the power imbalances that exist in the larger socio-cultural and political context to fight oppression, racism and discrimination in society. It informs the concept of cultural competence, the goal of which is to create a healthcare system and workforce capable of delivering care to every patient that considers and integrates all facets of culture (e.g., race, ethnicity, culture, or language proficiency) (Sue & Sue, 2003). In a crisis context, such as during the present pandemic, this social pedagogy lens should inform health communication approaches to avoid promoting prejudices and the targeting of specific groups through media, social media and institutional messaging. The lessons learned from the present context are crucial also to inform future preventive actions.
visible minority to obtain a more culturally diverse sample, in line with the study objectives. Participants were required to self-identify with one of the following ethnic groups: White, Chinese, South Asian (e.g. East Indian, Pakistani, Sri Lankan, etc.), Black (African, Afro-American, etc.), Filipino, Latin American, Southeast Asian (e.g. Cambodian, Indonesian, Laotian, Vietnamese, etc), Arab (Maghreb, Middle East), West Asian (e.g., Afghan, Iranian, etc.), Japanese, Korean, Other. We preserved White, Black, Arab and South-Asian categories per participant responses; for statistical reasons, we re-grouped the remaining categories into broader ethno-cultural groups. Specifically, Chinese (n = 223), Japanese (n = 15) and Korean participants (n = 11) were grouped together as East Asians (N = 249); Filipino (n = 11) and Southeast Asian participants (e.g., Cambodian, Indonesian, Laotian, Vietnamese, etc; n = 108) were all categorized as Southeast Asians (N = 119). The "Other" cohort (N = 61) included participants who self-identified as West Asians (n = 30), Latin American (n = 27) and those who originally identified as "other" (n = 4). The survey was presented as a study about COVID-19 and social distancing. Participants completed the survey in either French or English between June 1, 2020 and June 23, 2020. Participation was voluntary and confidential. All participants received between 50 cents to $2 in compensation, depending upon length of time taken to complete the survey (average completion time of 12 min) and provided electronic informed consent. A total of 8825 invitation emails were sent. The response rate was 37 %. The study was approved by the Institutional Review Board of the Faculty of Medicine at McGill University.

Table 1
Sociodemographic characteristics of participants for the total sample (n = 3273), and separately for participants who reported COVID-related discrimination and for those who did not report it.  D. Miconi et al. International Journal of Intercultural Relations 81 (2021) 176-192 Measures COVID-related discrimination Participants reported reasons for perceived discrimination (if any) in the last month because of their presumed COVID-19 status, based on a questionnaire developed by Williams, Yu, Jackson, and Anderson (1997) and adapted to the present health emergency context. Specific reasons and contexts of discrimination for the present study were selected following a discussion and consensus reached within the research team on reasons and contexts most relevant to investigate during the pandemic. This was informed by current public health measures (e.g., shutdown of non-essential businesses, school closures, social distancing) and clinical observations in the field. Participants were asked: "Have you been discriminated against because of your presumed COVID-19 status for any of the following reasons in the past month? Check all that apply". The list of reasons included (yes/no response format): Age, Gender, Physical Health, Immigration Status, Race/ethnicity, Occupation, Income, Neighborhood you live in. Reasons of discrimination were summed to get to the total number of reasons of discrimination reported by each participant (range 0-8). Participants who reported at least one incident of COVID-related discrimination in the last month were asked: "Have you experienced discrimination because of your presumed COVID-19 status in any of the following situations in the past month? Check all that apply (yes/no response format): at work or looking for a job, looking for an apartment or house, online, public transportation, supermarkets/pharmacy/grocery stores, in the streets/parks and health and social services. Contexts of discrimination were summed to get to the total number of contexts in which a participant experienced discrimination in the last month (range 0-7).

Socio-cultural variables
Participants self-reported age (18− 39, 40− 59, 60+), gender (male, female, other), education (high school or less, technical degree or some college/university, university degree and above), immigrant generation (first-, second-and third-generation immigrant and above), race/ethnicity (White, East Asian, South Asian, Black, South East Asian, Arab, Other), employment status/occupation (unemployed prior to COVID-19, employeddesignated as an essential worker by the Quebec government, employednot designated as an essential worker, essential health care worker, and unemployed because of COVID-19), geographical location (GMA vs elsewhere in Table 2 Results of a multivariate binomial regression model looking at the associations between socio-cultural variables and COVID-related discrimination (0 = no reported experiences of COVID-related discrimination; 1 = at least one reported experience of COVID-related discrimination) (n = 3273).   Miconi et al. International Journal of Intercultural Relations 81 (2021) 176-192 the province) and exposure to the virus (exposed vs non-exposed).

Data analysis
Descriptive information for the sample was summarized using counts and proportions for categorical variables. Less than 5% of data were missing for each variable and pairwise deletion was used in analyses. Next, we used multivariate binomial logistic regression on the total sample to assess if specific socio-cultural characteristics were associated with higher risk of experiencing COVID-related discrimination. Then, researchers used multivariate binomial logistic regression on the subsample of participants who experienced discrimination to assess if and how the socio-cultural variables of interest (i.e., age, ethno-cultural group, employment status/occupation and geographical location) were associated with each reason and context of discrimination. Prior to conducting multivariate binomial regression models for each reason and context of discrimination, age was recoded from three to two categories by merging participants aged between 40 years and more in one group, for issues related to statistical power (i.e., few participants aged 60+). Poisson generalized linear regression models with a log link function were used to assess differences in the total number of reported reasons and contexts of discrimination across socio-cultural groups. We used robust standard errors to adjust for violation of the distribution assumption that the variance equals the mean (Cameron & Trivedi, 2009). All multivariate regression models controlled for gender, education, immigrant status and exposure to COVID-19, as such factors have been linked to experiences of discrimination across multiple reasons and contexts within an intersectionality perspective (Ellermann, 2020;Ifatunji & Harnois, 2016;McMahon et al., 2016;Potter et al., 2019;Singh, Bhutani, & Fatima, 2020). The threshold for statistical significance was set to 0.05 (two-sided tests). R software was used in all analyses (R Core Team, 2017).

Overall prevalence and correlates of COVID-related discrimination
The prevalence of reported COVID-related discrimination in the overall sample was 16.58 % (n = 541) (see Table 1). Participants who experienced COVID-related discrimination were more likely to be men, young, with a lower education and healthcare workers compared to those who did not report discrimination. Non-White participants were also at higher risk of experiencing discrimination, with East Asian participants being four times more likely than White participants, followed by South Asian and South-East Asian participants. Participants who did not report or perceive as much discrimination did not differ in terms of geographical location (GMA vs the rest of the province), or immigrant status (see Table 2).

Prevalence of COVID-19 related discrimination among those who experienced discrimination
Overall, 66 % of discriminated participants experienced discrimination for one reason, 21 % for two reasons and the remaining 13 % for three or more reasons (range 1-7). In terms of contexts, among participants who reported discrimination, 23 % did not report any context of discrimination, 46 % reported one, 19 % reported two and the remaining 12 % reported three or more (range 0-6). Among those that reported COVID-related discrimination, the most common reported reasons were race/ethnicity and age, followed by occupation and physical health. Participants were more likely to report COVID-related discrimination in stores and in streets/parks than other locations. Of interest, although only 66 participants over 60 reported having being discriminated (10 % of total sample), 77    Note. *p < .05; ** p < .01; ***p < .001. All models included gender, education, immigrant status and exposure to COVID-19 as control variables. Separate models were run for each COVID-related discrimination reason (0 = not discriminated; 1=discriminated).

Table 6
Results of multivariate binomial regression models on the association between age, ethno-cultural group, occupation and geographical location and reported contexts of COVID-related discrimination (n = 541 Note. *p < .05; ** p < .01; ***p < .001. All models included gender, education, immigrant status and exposure to COVID-19 as control variables. Separate models were implemented for each COVIDrelated discrimination context (0 = not discriminated; 1=discriminated).
% of them experienced discrimination because of their age, and 48 % of them were discriminated in stores/supermarkets (see Tables 3  and 4).

Correlates of perceived discrimination among those who experienced COVID-related discrimination
Results of binomial regression models to investigate the impact of age, ethno-cultural group, employment status/occupation and geographical location on reported reasons of COVID-related discrimination are reported in Table 5. Age was statistically significantly associated with gender as reason of discrimination, in that younger participants were more likely to experience COVID-related discrimination because of their gender. One's ethnocultural group was significantly associated with race/ethnicity, age, immigrant status and residential neighbourhood as reasons of COVID-related discrimination. Specifically, East Asian participants reported the highest level of discrimination because of ethnicity, followed by South East Asian, South Asian and Black participants. East Asian and Black participants were less likely than White participants to be discriminated because of age, and East-Asian participants were also less likely to be discriminated because of the neighbourhood they lived in. South Asian, Arab and Black participants were more likely to report discrimination because of their immigrant status. Employment status was significantly associated with age and occupation as reasons of discrimination. Specifically, essential workers, and health-care workers in particular, were at higher risk of reporting COVID-related discrimination because of their occupation. Unemployed participants were more likely to report COVID-related discrimination because of their age. Finally, participants' geographical location was significantly associated with neighbourhood as reason of discrimination. Participants living in the GMA were twice as likely than participants living in the rest of the province to report experiencing discrimination because of the neighbourhood they lived in. No statistical significant associations emerged between the socio-cultural variables of interest and income or physical health.
In terms of contexts of discrimination (Table 6), age was significantly associated with discrimination experienced online, looking for a job and looking for an apartment, in that younger participants (aged 18− 39) were more likely to experience discrimination at work, looking for an apartment, and online. Participants's ethno-cultural group was associated with discrimination at work or looking for a job. Specifically, participants identifying as Arab or with the "Other" minority group reported more COVID-related discrimination at work or looking for a job than White participants. Participants' employment status was associated with discrimination experienced at the workplace/looking for a job and in supermarkets/grocery stores. In terms of odd-ratios, participants who were unemployed prior to the pandemic reported more COVID-related discrimination in stores/supermarkets but reported less discrimination at work or looking for a job than essential workers and participants who lost their job because of the epidemic. Participants' geographical location was significantly associated with discrimination experienced in the health and social services and in street\parks, in that participants who were living in the GMA were twice as likely to have experienced COVID-related discrimination in health and social services, but less likely than participants not living in the GMA to experience COVID-related discrimination in streets/parks.

Table 7
Results of Poisson regression models on the association between age, ethno-cultural group, occupation and geographical location and total counts of reasons and contexts of COVID-related discrimination (n = 541). Note. *p < .05; ** p < .01; ***p < .001. All models included gender, education, immigrant status and exposure to COVID-19 as control variables.
D. Miconi et al. International Journal of Intercultural Relations 81 (2021) 176-192 Correlates of sums of self-reported reasons and contexts of discrimination among those who experienced COVID-related discrimination Results from multivariate Poisson regression models indicated that the total number of reported reasons of COVID-related discrimination did not vary across age, ethno-cultural group, occupation or geographical location (Table 7). Only age was significantly associated with total number of reported reasons of COVID-related discrimination. Younger participants (aged 18− 39) were more likely to experience discrimination across multiple contexts (see Table 7).

Discussion
The present study documented experiences of COVID-related discrimination during the pandemic in an ethno-culturally diverse adult sample in Quebec (Canada). Results build on existing literature on the deleterious effects of COVID-related discrimination on mental health (Liu et al., 2020;Miconi et al., 2021) and social distancing intentions (Frounfelker et al., in press). Our results are not surprising in light of the rise in xenophobic and discriminatory incidents observed during the COVID-19 pandemic (He et al., 2020;Ransing et al., 2020) and underline the need of empirical research to inform public policy and pedagogical interventions. Such interventions should aim to mitigate the social and individual negative effects of the pandemic in terms of mental health and community inter-relations now and in the near future.

Prevalence of COVID-19 related discrimination
Our findings confirmed that COVID-related discrimination is an emerging but concerning reality during the present pandemic (Devakumar et al., 2020;Ransing et al., 2020). Overall, racialized minorities, health-care workers and younger participants were more likely to report COVID-related discrimination. Our findings confirm that ethnicity, occupation and age are important socio-cultural determinants and reasons of COVID-related discrimination during the current pandemic (He et al., 2020;Previtali et al., 2020;Singh & Subedi, 2020). Participants who were exposed to COVID-19 were also more likely to experience disease-related discrimination, in line with prior literature on the association between infection during an epidemic and stigma (Lee et al., 2007;McMahon et al., 2016;. In addition, men and less educated participants were also at higher risk of experiencing discrimination. These findings can be understood within an intersectionality perspective suggesting that race, gender and socio-economic status shape the experience of perceived discrimination, which is the result of a combination of different daily realities (Ifatunji & Harnois, 2016;Potter et al., 2019).

COVID-related discrimination and ethno-cultural origin
Our findings are in line with reports of discriminatory incidents that have been populating the news since the start of the pandemic, and confirm that participants with an Asian background, especially East Asians, are at a very high risk of incurring into COVID-related discrimination because of their race/ethnicity (He et al., 2020;Liu et al., 2020). Public messaging and political leaders may have played an active role in fueling rather than fighting this type of racial discrimination during the pandemic by associating in their speeches the virus with a specific geographical location, such as China (Sheth, 2020).
The fact that East Asian participants were less likely to report age or neighbourhood as reasons of discrimination may be related to the fact that race/ethnicity was for them the main reason of experiences of discrimination. The same explanation applies for Black participants, who were also less likely to experience discrimination because of age. The fact that South-Asian, Arab and Black participants were more likely to report discrimination because of their immigrant status may be linked to the demographic characteristics of recent immigrants in Quebec (Statistics Canada, 2017) which are also reflected in our sample, with the majority of South Asian, Arab and Black participants being first generation immigrants, hence more likely to have a precarious or temporary status in Quebec. Arab and other minority groups were at an increased risk of being discriminated at work or while looking for a job. These ethno-cultural groups are over-represented among essential workers in Quebec during the pandemic, as a result of pre-existing social and structural inequalities (Center for Research-Action on Race Relations (CRARR) (2020)). Taken together, these results suggest that although participants with an Asian descent are the ethno-cultural group at the highest risk of COVID-related discrimination, other racialized minorities are still experiencing COVID-related discrimination in links with their ethnic background and other cultural reasons, such as one's immigrant status. Although some initiatives were taken by public authorities in Quebec to ensure health coverage and protection of immigrant and racialized groups at the beginning of the pandemic, our findings indicate that such efforts were not completely effective and that more work is needed to ensure the safety of immigrant groups and ethnic minorities during the pandemic.

COVID-related discrimination and age
Younger participants reported more experiences of COVID-related discrimination overall and across contexts than older participants, mirroring a more active and dynamic life-style during the pandemic, which is likely to expose them more to situations in which discrimination may occur. This may also reflect a pitfall of ageism, in that young people may underestimate the risks and feel invincible and not responsible during the pandemic (Previtali et al., 2020), as showed for instance by the behaviors of young students who were partying and gathering together during spring break (Miller, 2020). Nonetheless, seniors were primarily discriminated in stores and because of their age. Our findings are in line with news reports on the outbreak of "Ageism" during the pandemic (Previtali et al., 2020) and highlight the need to protect not just young people but also seniors from discriminatory experiences, especially in stores.

COVID-related discrimination and employment status/occupation
In line with our expectations, health care workers were at higher risk of COVID-related discrimination and more likely to attribute it to their occupation. This confirms prior findings on social stigma and discrimination experienced by frontline workers across the globe (McKay et al., 2020;Singh & Subedi, 2020;The Economist, 2020;World Health Organization, 2020c). Reducing the risks of experiencing stigma and discrimination for frontline health workers is of crucial importance to protect their mental wellbeing and help in controlling this public health crisis effectively.

COVID-related discrimination and geographical location
The unique patterns of discrimination in the GMA are noteworthy and cause for further reflection. For instance, neighbourhood was more frequently reported as a reason of discrimination among participants living in the GMA than other areas of Quebec. Although this may simply be a product of the urban structure of the GMA and greater emphasis on neighborhood boundaries and characteristics than elsewhere in the province, such a finding could be related to the phenomena of public authorities' sharing information about rates of infection in the GMA by neighbourhood/municipality; this practice highlighted geographical disparities in COVID infection based on neighborhood characteristics such as socioeconomic status and cultural diversity (Santé Montreal, 2020).
In addition, the fact that participants living in the Montreal area were twice as likely to experience discriminaton in health and social services than participants living in the rest of the province raises questions as to whether health and social services in Montreal lack the cultural competence and inclusivity necessary to promote access to high-quality health-care to a culturally diverse population. Last, the fact that people living in the GMA were less likely than people living in the rest of the province to experience discrimination in the streets and parks suggests that highly diverse urban settings may be relatively protective from discrimination when minorities become the demographic majority (Hunt, Wise, Jipguep, Cozier, & Rosenberg, 2007;Welch, Sigelman, Bledsoe, & Combs, 2001).

Implications for public health messaging and programs
Our findings indicate that anti-discriminatory actions should address and modify public beliefs related to the association of COVID-19 with race/ethnicity, age, occupation, physical health and neigbourhood. In the short term, our findings also suggest that health communication actions with an anti-discriminatory objective should primarily target public spaces (e.g., stores, streets), but also the workplace and health and social services. Public health messaging oriented towards solidarity and how the general population can protect vulnerable individuals such as members of visible minorities and seniors, as well as the promotion of positive images of healthcare and essential workers among the general public seem potential avenues to reduce the stigma and discrimination associated with one's ethnic origin, occupation and age. From a prevention standpoint -in terms of future directions -our work points to the urgent need for the development, implementation and evaluation of multi-sectoral, community-based anti-discrimination programs which are grounded in principles of social pedagogy and anti-oppression (Nelson & Venkatesh, in press). In the specific context of the COVID-19 pandemic, social pedagogy refers to the reflexive and inclusive creation of dialogic spaces that promote pluralism, critical thinking and digital literacy. When applied alongside Freire (2017) principles of anti-oppression, social pedagogy in a pandemic context can ensure a re-alignment of the development of public policy that takes into account the perspectives of those communities that are marginalised through discrimination and also including key stakeholders with relevant expertise in health (mental, physical or otherwise), public safety, community development and social services. This also means that interventions grounded in social pedagogy should not only be designed to improve learning outcomes related to knowledge acquisition (e.g., digital literacy, building resilience aginst misinformation) and cognitive outcomes (e.g., transfer of learning in ecologically valid contexts), but also catalyse concrete outcomes associated with reduction in perpetrating discrimination (Logie & Turan, 2020). Such primary prevention programming could potentially include four unique components: 1) creating consultation groups led by community stakeholders with input from partners in public policy, public safety, education, social services, and health, 2) training, mentoring and equipping of marginalised community groups with critical digital literacy skills that contribute to a reduction in perpetration of discriminatory rhetoric, 3) showcasing community-led efforts on how to improve engagement, reduced aggression, and creating pluralist dialogic platforms to respond to discrimination; and 4) creating a robust evaluation framework which links the measured impacts of the prevention programming on individual, group and stakeholder outcomes.
In addition, it is important to monitor discriminatory phenomena in time during the pandemic, as the situation is changing rapidly. Studies aimed to investigate why COVID-19 related discrimination is experienced more in some contexts by specific social groups are warranted, as such knowledge could help to further the development of timely discrimination prevention policies and programs.
Undoubetely, age emerges as an important determinant of contexts and reasons of COVID-related discrimination. Our findings warn about the use of chronological age for public policies and messaging and suggests that interventions should be developed more on the basis of individual differences across age groups and considering the intersection of different social, economic and health factors (Ayalon et al., 2020;Previtali et al., 2020).
Emphasizing the efficacy and reduction of risk of infection due to preventive measures such as wearing a mask, hand washing and physical distancing over risks could reduce the stigma attached to essential workers and individuals with pre-existing health conditions. It is essential to disseminate accurate information on COVID-19 and risks of infection and transmission of the virus and emphasize the precautions taken by health care and essential workers to reduce such risks (Singh & Subedi, 2020). Increasing agency and empathy to promote solidarity and support within communities has proven to be an effective strategy to mitigate the negative outcomes of prior epidemics (Eaton & Kalichman, 2020; Logie & Turan, 2020).
Training health professionals in cultural competence can improve one's awareness around communication styles, explanatory styles as well as facilitate collaborations with interpreter services in the health domain. These are all fundamental aspects to enable effective work in cross-cultural situations. In addition, an acknowledgement of systemic racism and of the associated structural violence is often warranted to establish trust and help patients feel safe while receiving services. Cultural competence and cultural safety are complementary in health and social services and together constitute an essential step to overcome access barriers to care for cultural minority groups in ethnically diverse cities such as the GMA. The framing of public messaging regarding infection rates by neighbourhoods and municipalities needs to be carefully assessed to make sure that inclusiveness and solidarity rather than fear and isolation are underlined by policies and the media (Eaton & Kalichman, 2020). Our findings also indicate that health and social professionals should be aware of the increased risk of experiencing COVID-related discrimination of ethnic minorities, health-care workers, young people and seniors. Professionals working in the health and social services should actively support and protect these populations from discrimination and stigma during and in the aftermath of the pandemic.

Limitations and future directions
There are several limitations to this study. First, our study used a convenience sample and included a majority of participants with some college or a university degree, therefore findings cannot be generalized to the larger Quebec population nor to less educated populations. Second, differences may exist within the broad ethno-cultural groups used in the present study. Studies including larger samples and collecting more fine-grained ethno-cultural information are warranted. Third, we relied on self-reports and thus social desirability and response biases need to be taken into account. Fourth, participants who present several risk factors (e.g., essential healthcare worker of Asian descent) may be at increased risk of experiencing COVID-related discrimination but because of sample size limitations we could not test such interaction effects. Intersectionality may be very useful to understand the dynamics and relative risk of discrimination (Logie & Turan, 2020;Weber, 2010) and future studies are needed to shed light on this issue during the present pandemic. Last, our results cannot be generalized to different countries nor to other Canadian provinces, and more research on regional and trans-national differences is needed.

Conclusion
In spite of these limitations, our study provides the first empirical evidence of experiences of COVID-related discrimination during the current pandemic in a culturally diverse adult sample in the Canadian context. COVID-related discrimination is a concerning phenomenon during the present pandemic that deserves timely prevention and intervention. As it is multi-determined and assumes multiple forms, it is important to acknowledge its prevalence and complexity and to situate it within an intersectionality framework. It is crucial to protect minority groups who have already been disproportionately impacted by the pandemic due to systemic inequalities. Likewise, young people, seniors and essential health-care workers should be supported at the individual and collective level. The reduction of overall levels of discrimination -COVID-related and notwould help building a more just and equitable society and reduce the long-term negative impacts of the pandemic on individuals and communities.

Funding
Our work is one of the critical research programs being supported by the McGill Interdisciplinary Initiative in Infection and Immunity (MI4) with seed funding from the MUHC Foundation. The research presented in this paper is that of the authors and does not reflect official policy of the McGill Interdisciplinary Initiative in Infection and Immunity.

Declaration of Competing Interest
The authors reported no declarations of interest.