EXPLORING ETHNIC DIFFERENCES IN SELF-REPORTED PROTECTIVE BEHAVIOURS AGAINST COVID-19 MEDIATED BY PERCEPTIONS OF RISK By DANIEL MAULLON BScN, University of Ontario Institute of Technology, 2011 Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of MASTER OF NURSING in the FACULTY OF GRADUATE STUDIES TRINITY WESTERN UNIVERSITY September 2022 © Daniel Maullon, 2022 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS i Table of Contents List of Tables v List of Figures vi Abstract vii Acknowledgements vii Chapter 1: Introduction 1 Background 1 COVID-19 1 Interventions to Mitigate the Spread of COVID-19 3 Ethnic Group Differences with COVID-19 4 Confirmed Cases 4 Protective Behaviours 5 Perceptions of Risk 5 Definition of Terms 6 Self-Protective Behaviours 6 Perception of Exposure Risk 7 Ethnic Diversity 8 Thesis Description 9 Thesis Purpose and Objectives 9 Guiding Principals 10 Social Determinants of Health 10 Health Beliefs Model (HBM) 12 Perceived Severity 12 Perceived Benefits and Perceived Barriers 13 Cues to Action 13 Perceived Susceptibility Construct 14 Strengths and Limitations 14 Relevance and Significance 15 Research Gaps 15 Significance 16 Conclusion 16 Chapter 2: Literature Review 17 Search and Retrieval 17 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS Search Terms ii 17 Ethnic Diversity 18 Perceived Exposure Risk to COVID-19 19 Self-Protective Behaviours 19 COVID-19 19 Databases Strategies for Literature Review Inclusion/Exclusion Criteria Literature Review Results 19 20 20 21 Ethnic Differences in COVID-19 Rates 21 Socioeconomic Inequality 22 Perceived Risk and Self-Reported Protective Behaviours 23 Problem Statement 24 Conclusion 24 Chapter 3: Research Methods 26 Design 26 Rapid Review 26 Search Strategy 26 Data Extraction 29 Analysis 30 Survey Design 31 Sampling 31 Inclusion/Exclusion Criteria 31 Procedures 31 Measures 34 Amended Study Design 37 Survey Administration 37 Quantitative Data Analysis 39 Answering the Research Questions Ethics 42 43 Ethical Risks with Human Participation 43 Human Research Ethics Boards (HREB) 44 Conclusion 44 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS Chapter 4: Findings Descriptive Survey Results iii 46 47 Gender 47 Age 48 Income 48 Overall Health Status 49 Born in Canada 49 Vaccination Status 50 Research Questions #1: Ethnic Group Differences in Protective Behaviours 53 Rapid Review Results 53 Quantitative Survey Results 54 Research Questions #2: Ethnic Group Differences in Perceived Exposure Risk Rapid Review Results Quantitative Survey Results Research Question #3: Perceived Exposure Risk Explains Variation in Self-Reported Protective Behaviours 60 60 61 63 Rapid Review Results 63 Quantitative Survey Results 64 Research question #4: Self-Reported Behaviours are Explained by Ethnic Group Differences in Perceived Risk 70 Rapid Review Results 70 Quantitative Survey Results 71 Conclusion 71 Chapter 5: Discussion 73 Relationship Between Ethnic Diversity and Self-Reported Behaviours 73 Social Determinants of Health 77 Health Beliefs Model 78 Health Beliefs Model and Social Determinants of Health 79 Future Implications 81 Future Research in Canada 81 Nursing and Cultural Humility 82 Scientific Quality and Limitations 84 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS iv Representativeness 84 Sample Size 86 Rapid Review 86 Measures 87 Chapter 6: Conclusion and Recommendations 88 References 90 Appendix A 103 Appendix B 113 Appendix C 114 Appendix D 116 Appendix E 118 Appendix F 123 Appendix G 128 Appendix H 134 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS v List of Tables Table 1: Search Terms for Literature Review 18 Table 2 Inclusion and Exclusion Criteria for Literature Review 21 Table 3: Inclusion/Exclusion Criteria for Rapid Review 28 Table 4: COVID-19 Own Risk Appraisal Scale (CORAS) 35 Table 5: Questions pertaining to self-protective behaviours 36 Table 6: Cross tabulation of self-protective behaviours with demographics 51 Table 7: Descriptive statistics of the sample population 52 Table 8: Binomial logistic regression of hand-hygiene 55 Table 9: Binomial logistic regression of avoiding high-risk individuals 56 Table 10: Binomial logistic regression of avoiding public spaces 58 Table 11: Binomial logistic regression of avoiding eating at restaurants 59 Table 12: Binary logistic regression of wearing masks 60 Table 13: ANOVA results for linear regression between CORAS (dependent variable) and ethnic groups 62 Table 14: Linear regression coefficients for CORAS 63 Table 15: Binary logistic regression of hand-hygiene including CORAS 65 Table 16: Binary logistic regression of avoiding high-risk individuals including CORAS Table 17: Binary logistic regression of avoiding public spaces including CORAS 66 Table 18: Binary logistic regression of avoiding eating at restaurants including CORAS Table 19: Binary logistic regression of wearing masks including CORAS 68 67 70 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS vi List of Figures Figure 1: Mediation Model 10 Figure 2: PRISMA Diagram 29 Figure 3: Frequency of Ethnic Groups 53 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS vii Abstract This research explores ethnic differences in self-protective behaviours against COVID-19, and if identified, whether these differences are mediated by perceived risk towards COVID-19. I recruited 187 respondents to participate in an online survey about perceived risk and selfprotective behaviours against COVID-19. Using a mixed-methods design, I observed people belonging to an ethnically diverse population engaged in hand-hygiene, avoided high-risk individuals, avoided large crowds or public spaces, and avoided eating at restaurants more frequently relative to less diverse “White” populations. The likelihood of an individual wearing a mask in public was similar between both groups. There was no evidence of mediation from perceptions of risk on behaviours. Systemic barriers that ethnically diverse communities typically encounter, such as low income and multi-family homes, may play a role in the engagement of self-protective behaviours. More research is warranted to identify social determinants of health and their role in the engagement of self-protective behaviours. Keywords: COVID-19, self-protective behaviours, ethnicity, perception of risk ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS viii Acknowledgements I would like to acknowledge my supervisors Dr. Richard Sawatzky and Dr. Kendra Rieger for their patience and guidance throughout this chapter of my academic journey. I would also like to thank all the professors at the Trinity Western University School of Nursing for their guidance and dissemination of knowledge throughout these years. Additionally, I would like to thank Dr. Barbara Astle for encouraging me to pursue research in Nursing and always providing helpful advice during my graduate studies. Lastly, I would like to recognize my defense committee for, not only making me feel comfortable and calm, but helping me learn and grow from this experience. I would also like to give thanks and appreciation to my family and friends for their constant motivation and patience with me as I pursued this academic endeavor. More specifically, I would like to thank my wife Grace and our dog Molly. Without you both, none of this would have been possible. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 1 Ethnic Differences in Self-Reported Protective Behaviours Against COVID-19 Mediated by Perceptions of Risk Chapter 1: Introduction In Canada, a large percentage of confirmed COVID-19 cases have occurred within the most ethnically diverse communities, a phenomenon that also occurs in other multicultural communities around the world (Hou et al., 2020; Platt & Warwick, 2020). The use of nonpharmaceutical protective methods such as social distancing, hand-hygiene, and stay-at-home orders amongst others are effective methods to control the spread of COVID-19 in the community (Wu et al., 2020). However, the effectiveness of these methods is dependent on the population’s willingness to perform them. Studies have suggested that compliance with selfreported protective behaviours against COVID-19 improve with increased perception of exposure risk towards the disease (Clavel et al., 2021). Despite this, there is minimal research to determine if there is an ethnic component influencing self-reported protective behaviours. Utilizing crowdsourcing survey data from Canada, this research explored ethnic differences in self-reported protective behaviours towards COVID-19 and the perceived risk of transmission from the disease. Background COVID-19 COVID-19, first reported in China in late 2019, is caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) (Guan et al., 2020). Transmission of SARS-CoV-2 is primarily through droplet-contact transmission into mucosal entry-points such as the eyes, nose, and mouth. Upon entry into the body, SARS-CoV-2 binds to the receptors of cells and eventually enters the cell body to proliferate. While evading the human immune response, once it has proliferated within the cell, it will destroy the host while releasing ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 2 additional replications of itself, infecting neighbouring cells (Guan et al., 2020). COVID-19 can be asymptomatic or cause a host of mild to severe symptoms (Ma et al., 2021). Many confirmed cases of COVID-19 are asymptomatic or mild in severity; a systematic review of COVID-19 prevalence research by Ma et al. (2021) suggests that up to 40.5% of confirmed cases were from asymptomatic individuals. However, more serious cases can lead to significant symptoms, most commonly respiratory distress and sepsis, leading to hospitalization and death (Guan et al., 2020). In Canada, from January 2020 to June 2022, around 20% of all COVID-19 hospitalizations have resulted in the need for critical care or Intensive Care Unit [ICU] admission (Government of Canada, 2022). As of this writing, more than 40,000 of the 4 million confirmed cases in Canada have resulted in patient death (Government of Canada, 2022). Since 2020, COVID-19 has evolved into different Variants of Interest (VOI), each with unique properties that affected transmission, severity, and susceptibility to treatment (Fernandes et al., 2022). Between 2020-2022, hundreds of variants have been identified by the World Health Organization (WHO) but only a few of these VOIs affected the population on a global scale. The Alpha (first identified in the U.K. in December 2020), Beta (first identified in South Africa in December 2020), Delta (first identified in India in May 2021), and Omicron (first identified in South African in November 2021) VOIs became globally prominent COVID-19 strains at various points during the pandemic. Strains that became the dominant strain of COVID-19 were more transmissible than the previous strain which led to an increased number of cases in the community (Fernandes et al., 2022). On March 11, 2020, the WHO (Guan et al., 2020) declared COVID-19 a global pandemic and, as of this writing, there are more than 530 million confirmed cases with more than 6.3 million deaths worldwide (World Health Organization [WHO], 2022). Treatment of COVID-19 during the early stages of the pandemic (2020-early 2021) was challenging as there were no anti- ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 3 viral medications specific to COVID-19 deemed to be effective. During the initial year of the pandemic, treatment relied heavily on alleviating symptoms and supporting the body’s immune response to eliminate the virus. As COVID-19 spread across the globe, certain drugs were used experimentally, such as hydroxychloroquine, lopinavir, and interferon, with promising results (WHO Solidarity Trial Consortium, 2022). However, as the pandemic progressed and more research became available, hydroxychloroquine, lopinavir, and interferon amongst other treatment medications were no longer recommended due to the risk of adverse reactions (WHO Solidarity Trial Consortium, 2022). Remdesivir, an anti-viral that was originally developed to treat the Ebola virus, emerged as the primary anti-viral drug to treat moderate to severe COVID19 (WHO Solidarity Trial Consortium, 2022). In tandem, to manage respiratory symptoms, the use of corticosteroids (specifically Dexamethasone) has been recommended for treatment with strong evidence (Center for Disease Control and Prevention [CDC], 2022). As of April 2022, the combination or individual use of Remdesivir and corticosteroids for the management of COVID19 in patients is considered best practice (CDC, 2022). Interventions to Mitigate the Spread of COVID-19 To contain the virus, public health authorities in Canada have encouraged the population to engage in protective guidelines. These guidelines include, but are not limited to, social distancing, hand-hygiene with warm soapy water for 20 seconds, only leaving the house for essential purposes such as health emergencies and groceries, and mask-use in public places (Government of Canada, 2021a). Other, more drastic, guidelines have affected larger aspects of the economy such as work-from-home orders, non-essential store closures, and travel bans (Government of Canada, 2021a). The implementation of these individual (e.g., hand hygiene or social distancing) and systemic (e.g., border closures or stay-at-home orders) interventions have ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 4 been shown to reduce the spread of COVID-19 in communities within China, Canada, United States, and the United Kingdom (Guan et al., 2020). Ethnic Group Differences with COVID-19 Confirmed Cases Confirmed cases of COVID-19 have occurred through all of Canada; however, as of July 2022, Ontario (more than 1.3 million cases) and Quebec (more than 1 million cases) have the highest and second highest number respectively of COVID-19 cases in the country (Government of Canada, 2022). A report by Public Health Ontario (PHO) (2021) reported that between February 2020 to December 2021 the most ethnically diverse communities in Ontario accounted for around 70.2% of all COVID-19 cases in the province. In comparison, the least diverse communities in Ontario only accounted for 16.3% of COVID-19 in Ontario with the remaining number of cases occurring in moderately diverse communities (PHO, 2022). Similarly, rates of COVID-19 in Quebec are 1.7 times higher in ethnically diverse communities (i.e., greater than 46.5% of the community are of ethnic diverse backgrounds) compared to less ethnically diverse communities (i.e., less than 17.4% of population in the community are of ethnic diverse backgrounds) (Santé Montreal, 2022). Much of the ethnic disparity in confirmed cases can be attributed to the sociodemographic inequalities experienced by ethnically diverse groups such as low income, job opportunities, and education (Clavel et al., 2021; Platt & Warwick, 2020). In Ontario, the Black population is more likely than Caucasians to report a lower income (Houle & Statistics Canada., 2020) which, in turn, can result in living in multi-family homes, apartment buildings, or shelters, increasing the risk of contracting COVID-19 (Baena-Diéz et al., 2020; Hou et al., 2020). Despite certain ethnic groups reporting higher COVID-19 cases in Ontario and Quebec, individuals, regardless of ethnicity, can reduce the spread of the virus by practicing protective behaviours. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 5 Protective Behaviours COVID-19 is a highly contagious virus with few treatment options and strategies to prevent the transmission and spread of the virus are extremely crucial. Protective behaviours against the virus, such as social distancing, wearing a mask, and handwashing, are examples of strategies to protect the population from COVID-19 (Lai et al., 2020). Although protective behaviours are performed at an individual level, studies have shown that individuals who report being of a particular ethnic group may engage in protective behaviours more than others. The Black population, for example, is more likely to wear a mask in public than Caucasians. Furthermore, the Latinx and Asian-American ethnic groups were more likely to wear a mask than the Black population (Hearne & Niño, 2021). With this in mind, we should, theoretically, see an increased number of confirmed COVID-19 cases in the Caucasian population, yet the opposite is true. We, in fact, see a higher number of confirmed COVID-19 cases affecting other ethnic groups (PHO, 2022). Although ethnic-related social determinants of health such as income or housing situations have contributed to the rise of COVID-19 cases in ethnic groups (Bruine de Bruin & Bennett, 2020), the act of performing protective behaviours can also be attributed to an individual’s perception of risk towards COVID-19. Perceptions of Risk Despite the importance of identifying ethnic differences in self-reported protective behaviours against COVID-19, it is also vitally important to recognize the possibility of ethnic differences in perception of risk. According to Yildirim et al. (2021), increased perception of risk, vulnerability, and fear of COVID-19 correlates with more engagement in self-protective behaviours such as handwashing and avoidance of public transportation. Additionally, a survey of participants in the United States also shows a higher frequency of handwashing in participants who report a higher perception of risk towards COVID-19 (Bruine de Bruin & Bennett, 2020). ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 6 Regarding ethnic differences in perceived risk, the Caucasian ethnic group reports a lower perception of risk towards COVID-19 than the Black ethnic group (Clavel et al., 2021). More specifically, Asian-American and LatinX ethnic groups have a higher perception of risk towards COVID-19 than both Caucasian and Black ethnic groups (Niño et al., 2021). Being able to identify the perceived risk of a particular ethnic group towards a certain exposure can help determine the likelihood of that group practicing self-protective behaviours (Kan & Zhang, 2018). Definition of Terms Self-Protective Behaviours To reduce spread of the virus in the community, the implementation of self-protective behaviours has been considered one of the best interventions. Self-protective behaviours are a group of voluntary interventions that individuals perform to evade exposure to a particular risk— in this case, COVID-19. Throughout much of the pandemic, the Government of Canada (2021a) encouraged individuals to engage in self-protective behaviours such as performing hand-hygiene for 20 seconds with warm water and soap, having good indoor ventilation, cleaning or sanitizing surfaces in the home, and wearing a non-medical facemask or covering in public. For this research, I utilized the Government of Canada’s 2021 guidelines to protect against COVID-19 (Government of Canada, 2021a) to define self-protective behaviours. It is important to note the differences between self-protective behaviours and protective behaviours. Self-protective behaviours are voluntary actions which are designed to protect the sole individual performing them. Protective behaviours, however, incorporate voluntary or mandatory interventions that aim to not only protect oneself, but may also be performed with the intention to protect others (Jaspal et al., 2020). An example of a protective behaviours is the act of coughing or sneezing into a tissue or bend of the elbow whereby the purpose of this is to ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 7 prevent the spread of droplets to others. Wearing masks in public could be considered both a self-protective behaviour and protective behaviour—masks can prevent users from breathing in droplets caused by other and also prevent droplets caused by the user from reaching others (Hearne & Niño, 2021). For clarity, the questions in this research focus only on self-protective behaviours or self-protective behaviours that could also be protective barriers, as the concept of protective behaviours alone may introduce a participant bias caused by the perceived moral duty to protect others (Jaspal et al., 2020). Perception of Exposure Risk According to Cori et al. (2020), perception of risk is an individual’s intuitive and subjective view of the hazards they may be exposed to. It is important to understand perceived risk within a population because this may be associated with a willingness to engage in selfprotective behaviours (Wise et al., 2020). These perceptions may be directly or indirectly influenced by various social and contextual factors, such as the media, social media, societal influences, and ethnical/cultural norms (Cori et al., 2020); one, or all, of these factors can influence an individual’s perception of risk. According to Cori et al. (2020), “perception of risk” is influenced by four modalities: voluntariness, trust, visibility, and knowledge. The element of voluntariness pertains to the individual’s willingness to be exposed to the inherent risk. An example of this can be the act of going into the home with a confirmed COVID-19 case without the use of protective equipment (e.g., facemask). Knowledge refers to the level of understanding of the exposure or risk; a disease or virus with less familiarity or understanding (e.g., COVID-19) is perceived as a higher risk than a disease or virus that is more recognizable (e.g., influenza). Visibility pertains to an exposure that is visible versus one that is not. A less visible risk factor (e.g., COVID-19) is perceived as more hazardous than a visible one (e.g., a mosquito). Lastly, trust relates to ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 8 confidence towards others to effectively manage the risk; a power plant, for example, may not be seen as a high-risk exposure due to the public trust in the employees to safely manage it, whereby COVID-19 can be seen as high-risk if there is little trust in public health officials to control the spread of the virus. For this research, I focused only on the voluntariness modality because the effectiveness of COVID-19 preventive guidelines requires voluntary engagement from the public for them to be effective. Avoiding contact with the elderly, for example, may be an effective strategy to protect high-risk individuals; however, this can only occur if the public is willing to engage in this type of preventive strategy. Ethnic Diversity The term “ethnic diversity” is broad in nature and encompasses a plethora of characteristics. Statistics Canada (2021) used the following criteria in the 2021 census to help define ethnocultural diversity: place of birth (including place of birth for mother and father), immigration status, citizenship, religion, ethnic or cultural origins, Indigenous peoples (e.g., First Nations, Métis, and Inuit), visible minority groups, and language. Dinesen et al. (2020) support the characteristics defined by Statistics Canada by indicating that the term “ethnic diversity” is typically defined in research as a broad umbrella which encompasses various characteristics, including polarizing ethnic groups. In contrast, race, as its own definition, pertains specifically to the biological origins of a particular individual and is usually inclusive of physical traits such as skin or hair colour (Ross et al., 2020). Most research utilizes ethnic diversity in a broader sense as individual ethnic qualities appear to overlap such as religion, language, and place of birth (Dinesen et al., 2020). Akin to the research by Dinesen et al. (2020), for my research purposes, I refer to ethnic diversity as the collective set of racial backgrounds that distinctly identify as a different race from the population’s norm. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 9 Thesis Description Thesis Purpose and Research Questions The aim of this thesis was to explore whether there are ethnic differences in self-reported protective behaviours towards COVID-19, and, if so, whether these were associated with ethnic differences in perceived exposure risk within an ethnically diverse community in Canada. The corresponding research questions and subsequent mediation model were developed: 1) Are there ethnic group differences in protective behaviours? 2) Are there ethnic group differences in perceived exposure risk? 3) Does perceived exposure risk explain variation in self-reported protective behaviours? 4) If there are ethnic differences in protective behaviours, are these mediated by ethnic group differences in perceived exposure risk to COVID-19? ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 10 Figure 1 Mediation Model Sociodemographic Variables Health Status Vaccination Status (Covariates) Perceived Exposure Risk to 2 COVID-19 3 4 Self-Reported Protective Ethnic Groups Behaviours from COVID-19 1 Note: The numbers in the Figure refer to the numbering of the research questions Guiding Principals Social Determinants of Health According to the Government of Canada (2020), the determinants of health include various social, economic, and environmental factors that determine individual and population health. One of these factors, race/racism, can affect various aspects of an individual’s life experiences including employment, housing, culture, and access to health services, all of which are determinants of health in their own right (Government of Canada, 2020). The social determinants of health “refer to a specific group of social and economic factors within the broader determinants of health” (Government of Canada, 2020, para. 2). This specific group of factors includes income, education, and employment, which can be influenced by experiences of ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 11 racism or racial inequality (Government of Canada, 2020). For this thesis, I considered race as a determinant of health, and income and housing situation as social determinants of health, to inform my research. The sociodemographic questions used in our survey are adapted from Pinto et al.’s work (2016) on identifying inequities and health disparities in population health. The work by Pinto et al. (2016) focused on the social determinants of health and finding evidence of health disparities and inequities in the population that can be collected through administrative data and surveys. They suggest that many hospitals or organizations have difficulty identifying disparities from surveys due to the methods of survey administration, the format of the questions, or concerns of questions disrupting the therapeutic patient relationship (Pinto et al., 2016). The survey questions developed by Pinto et al. (2016) contain various sensitive, yet detailed, questions about sociodemographic factors including race, gender, language, religion, income, sexual orientation, housing, health status, and physical disabilities. The results of their study showed that the set of sociodemographic questions they developed had a high response rate (84%-100%) (Pinto et al., 2016). Only the question “In what year did you arrive in Canada?” had a lower response rate (74%). The utilization of this set of survey questions may be able to highlight which minority groups (e.g., low income, ethnically diverse groups) are being served by the health care system and which are not (Pinto et al., 2016). According to PHO (2022), ethnically diverse communities experience higher rates of confirmed COVID-19 cases than less ethnically diverse communities. As such, race may influence one’s susceptibility to COVID-19 whereby non-White groups are at higher risk of contracting the virus (Mackey et al., 2021). Race, on its own, has not been clearly identified as a determinant of susceptibility to COVID-19; instead, the combined social factors that are the result of race can lead to racial inequality with housing, education, income, and employment, ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 12 which can lead to a higher risk contracting COVID-19. For example, minority ethnic groups are more likely to work in factory jobs which require employees to work on-site as opposed to working from home, thus, increasing the risk of contracting COVID-19 in the workplace. Additionally, minority ethnic groups are more likely to live in multi-family homes which can spread the disease to more individuals within the household (Bonacini et al., 2021; Raifman & Raifman, 2020). I have used this premise to influence the development of my research questions and methods for this thesis. I have used the term “ethnic diversity” in my thesis, which encompasses the term “race.” Health Beliefs Model (HBM) One of the guiding principles used to inform my research is the Health Beliefs Model (HBM). The HBM, first developed by Rosenstock et al. (as cited in Rosenstock, 1974), helps predict health protective behaviours of a population by evaluating the perceived risk potential and disease prevention behaviours as major variables (Lee, 2021). The HBM includes five constructs: perceived severity, perceived benefits, perceived barriers, cues to action, and perceived susceptibility. I have chosen to focus my research based on the perceived severity and the perceived susceptibility constructs of the HBM, as the four modalities of risk perception (i.e., voluntariness, trust, visibility, and knowledge) can be influenced by the perceived severity and perceived susceptibility of the HBM. In contrast, the perceived benefits, perceived barriers, and the cues to action are more closely related to the perceptions regarding protective behaviours. Perceived Severity. Perceived severity involves the subjective cognitive belief of the seriousness of an exposure (Rosenstock, 1974). How an individual perceives the severity of an exposure involves their perception of the consequences. For example, an individual who considers themselves to have a strong immune system may not perceive COVID-19 to be a severe threat. In contrast, an individual who lives with high-risk individuals (e.g., elderly, ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 13 children, immunocompromised individuals) may consider COVID-19 to be a severe threat, even if that threat is not to themselves (Rosenstock, 1974). Perceived Benefits and Perceived Barriers. Rosenstock (1974) describes the perceived benefits and perceived barriers in the HBM as distinct, yet related constructs. The perceived benefits construct relates to how effective an individual believes an alternative action (a different action beyond what would have typically been done) towards an exposure will be in reducing their susceptibility to it (Rosenstock, 1974). During the COVID-19 pandemic, some members of the public chose not to wear masks as they did not believe that it would prevent them from contracting the virus (Erdemandi & Leach, 2021). In the areas where these beliefs were strong, the mortality rate from COVID-19 was higher (Erdemandi & Leach, 2021). With perceived barriers, this construct relates to how an individual perceives the negative effects of performing an alternative action (Rosenstock, 1974). For example, an individual may recognize the importance of taking the COVID-19 vaccine but may perceive the potential side effects of the vaccine as too severe for them to take it. Cues to Action. Cues to action are the triggers that cause an individual to engage in an alternative action (Rosenstock, 1974). Examples of cues, in relation to COVID-19, could be a positive case of a family member or a rising number of cases in the community. Rosenstock (1974) explains that the size of the cue (i.e., how much or little is required to influence action) is dependent on the other constructs of the HBM. To illustrate, an individual who perceives the severity of COVID-19 to be high and experiences little to no barriers to an alternative action may require a small cue (e.g., a family member who tested positive for COVID-19) to engage in social distancing. In contrast, an individual with a lower perceived risk towards COVID-19, and who experiences many barriers to an alternative action, may require a larger cue to action (e.g., the death of a family member from COVID-19). ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 14 Perceived Susceptibility Construct. The construct of perceived susceptibility refers to the belief in the likelihood of contracting a disease or condition (Champion & Skinner, 2008; Rosenstock, 1974). For example, an individual would need to believe they are at risk of getting COVID-19 prior to engaging in self-protective behaviours such as mask-wearing (Champion & Skinner, 2008; Rosenstock, 1974). Perceived susceptibility can be influenced by various factors such as accessibility to health care, being able to work from home, or living in a community with a smaller number of COVID-19 cases. Strengths and Limitations. Since the creation of the HBM in the 1950s, many researchers have applied the HBM framework to describe and influence change in social behaviour (Jones et al., 2014). One of the strengths with the use of the HBM in my research is the ability to provide insight for my analysis and explain my findings. Increased engagement with self-protective behaviours, for example, may be explained by a higher perceived risk which may be affected by social factors such as housing or employment status. Furthermore, decreased engagement with self-protective behaviours may be explained by social barriers, such as living with dependent, but high-risk, family members, in which the “perceived benefit” may not be enough to encourage self-protective behaviours. In absence of a guiding framework, it is more challenging to explain observations of the study. Despite the frequent utilization of the HBM in research studying social behaviours, the effectiveness of this model as an exploratory model are mixed. Champion and Skinner (2008) suggest that the HBM and the use of all five constructs are effective in predicting future selfprotective behaviour. The systematic review by Jones et al. (2014) suggests that behaviour change techniques (BCT), as opposed to theoretical application of HBM, may explain the variance in the reviewed studies. Another limitation of the model is when its collective constructs are broken down and applied separately. It is believed that doing so may reduce the ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 15 strength and relationship between the construct and reported outcome; however, further research is required to analyze this concept (Champion & Skinner, 2008; Jones et al., 2014). Relevance and Significance Research Gaps As COVID-19 continues to affect much of the world’s population, many studies are being conducted on the relationship between ethnicity and COVID-19. Particularly in the United States, many studies have reported ethnic differences in COVID-19 cases and risks. Although there is much research available outlining the ethnic differences in hospitalizations and risks (PHO, 2020; Platt & Warwick, 2020), there is minimal research exploring the ethnic differences in COVID-19 risk perception and its relation to self-reported behaviours. Furthermore, available research exploring ethnic differences with COVID-19 have typically consolidated ethnic groups into generalized categories such as Whites, Black, Latinx, and Asians. Particularly with the Asian categorization, where there are distinct ethnicities and cultures within the category itself, it can be difficult to identify ethnic-specific challenges if these groupings are too broad. To demonstrate, a study by Platt and Warwick (2020) suggests that the Pakistani population in the United Kingdom are at higher risk of contracting COVID-19 than the Chinese population. The characteristics and behaviours of both ethnicities are distinct, and interventions would differ amongst the two—yet many studies have combined these two under the “Asian” category (Webber, 2020). To address this limitation, it often is beneficial to break down ethnic categories into more specific features such as South Asian, East Asian, Black African, or Black Caribbean. By having more specific ethnic groups, it is possible to identify more ethnic-specific interventions to prevent COVID-19 (Webber, 2020). This explorative research can inform future work which focuses more on specific ethnic groups and help identify ethnic-specific interventions to encourage self-protective behaviours against COVID-19. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 16 Significance As of this writing, there are gaps in research related to the barriers experienced by people of an ethnically diverse population in engaging in self-protective behaviours against COVID-19. I believe that my research on this topic of interest can produce inquiry on how we observe an ethnically diverse population in engaging in self-protective behaviours against, not only COVID19, but other communicable diseases or viruses which we may encounter in the future. Furthermore, my findings can inspire future research questions regarding strategies to encourage self-protective behaviours in pandemics, especially with ethnically diverse communities. If this work is not conducted, we will not be able to understand the challenges and barriers experienced by people of an ethnically diverse population in engaging in self-protective behaviours. Conclusion COVID-19, caused by SARS-CoV-2, is a disease that can cause significant health-related issues such as respiratory distress and sepsis. The advent of COVID-19 in 2019 eventually led to a significant rise in confirmed cases globally and challenged healthcare systems around the world. To mitigate the spread of COVID-19 across the world, governments of various levels have implemented preventative strategies such as lockdowns, mask wearing in public, and encouragement of social distancing measures. The implementation of these measures is dependent on the individuals within the population willing to engage in them. In Ontario, it has been shown that COVID-19 is more prevalent in the most ethnically diverse communities, thus prompting the question as to why this is so. This thesis focused on describing differences in selfprotective behaviours between ethnically diverse and non-ethnically diverse populations and, if so, understanding whether these differences are mediated by ethnic differences in perceived risk to COVID-19. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 17 Chapter 2: Literature Review As the COVID-19 pandemic continues to persist as of this writing, more research is being conducted on the social effect this disease has had on the population. Many studies have explored how the existence of COVID-19 has altered the dynamics of self-perceived risks and how self-protective behaviours are practiced. The literature search for this thesis also focuses on protective behaviours and self-perceived risks but seeks to identify research that includes a cultural or ethnic component within the studies (see Appendix A for literature review matrix). I first explain my search strategy and the elements of the PICOT format that guided my search. Subsequently, I provide details of the search terms and inclusion/exclusion criteria used in my search strategy to identify relevant research for my literature review. Lastly, I provide the results of my literature review and apply them to my problem statement to provide the rationale behind the thesis’ purpose. Search and Retrieval The review question used to guide the search and retrieval process for this study is: “How do ethnic differences in perceived exposure risk influence self-reported behaviours in individuals during the COVID-19 pandemic?” The question followed the PICOT format with: (P) Ethnic groups, (I) Perceptions of exposure risk, (C) Other ethnic groups, (O) Self-reported behaviours, and (T) during the COVID-19 pandemic. Using this format allowed us to identify the key variables used in the study while also narrowing the included studies in the review to this specific question. We utilized the services of various scholarly journal databases to retrieve relevant literature to answer my review question. Search Terms Search terms (refer to Table 1 for search terms used for literature search) were generated with the assistance of the librarian at Trinity Western University (TWU) using the PICOT ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 18 question as guidance. I first identified my target key term, then subsequently determined synonyms that accurately represent them. The search terms were used to search any field in the databases. Relevant subject headings were mapped to the original search terms using automatic term mapping in MEDLINE and Pubmed and the default search algorithm for unqualified searches, including subject heading searched, on CINAHL. By default, Google Scholar does not use subject header searches. Table 1 Search Terms for Literature Review Search # 1 Key Word Search Terms COVID-19 2 ETHNIC GROUPS 3 PERCEPTION OF RISK SELF-PROTECTIVE BEHAVIORS COVID* OR CORONAVIRUS* OR SARS* CO* OR SEVERE ACUTE RESPIRATORY SYNDROME* OR "NCOV" OR "DELTA" OR "OMICRON" OR VARIANT* OR BETA* ETHNIC* OR RACE* OR "RACIAL* BACKG*" OR CULTUR* OR "CULTURAL NORM*" OR "CULTUR* BACKGRO*" OR ETHNOGRAPH* OR "SOCIODEMOGRAPHIC* RISK* OR PERCEIVED RISK OR RISK PERCEPTION* OR UNDERSTAND* RISK OR ATTITUDE* PROTECTIVE BEHAV* OR SOCIAL DISTANC* OR HAND WASH* OR RESTAURANT* OR GROUP ACTIVIT* OR ISOLAT* OR PREVENT* OR CONTROL* OR REDUC* OR QUARANTIN* #1 AND #2 AND #3 AND #4 4 5 COMBINED Ethnic Diversity. The population for this research revolves around ethnic diversity. For my research purposes, I refer to ethnic diversity as the collective set of racial backgrounds that distinctly identify as a different race from the population’s norm. For my search strategy, I determined synonyms that represent ethnic diversity: race, racial background, cultural background or norms, and ethnography. These search terms coincide with the definitions used in the 2021 Statistics Canada Census for immigration and ethnocultural diversity (Statistics Canada, ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 19 2021). As some research is analyzing the sociodemographic of a population altogether, I also included “Sociodemographic” as a keyword term for my searches. Perceived Exposure Risk to COVID-19. My preliminary searches regarding perceived exposure risk to COVID-19 returned minimal results, therefore I expanded my search to be broader. Exposure risk is one element of risk perception and most available research does not focus on individual elements of perceived risk. As a result, I broadened my search terms to be inclusive of perceived risks with the assumption that exposure risk is a variable that is captured naturally within the studies. These terms are risk perception, perceived risk, understanding of risk, and attitudes. Self-Protective Behaviours. The term “self-protective behaviours” pertains to the protective behaviours to protect oneself. However, as some research may not distinguish between protective behaviours and self-protective behaviours, I adjusted my search terms as such to be inclusive of this factor. The terms utilized for this search include self-protective behaviours, protective behaviours, and reported protective behaviours. COVID-19. The term COVID-19 refers to the disease process that is a result of the SARS-CoV-2 virus. Although COVID-19 is the most recognizable name during this pandemic, I included terms that are associated with the disease. These terms are COVID, SARS, coronavirus, and/or severe acute respiratory syndrome. Databases Searches for relevant research were performed primarily through the TWU library’s database. Three databases were utilized in this search: MEDLINE with full text, PubMed, and CINAHL. Google Scholar was the final database utilized to locate research pertaining to the research question. We limited our search to original scholarly and peer-reviewed studies conducted from 2019 onward that was conducted or translated in English. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 20 We used MEDLINE with full text and PubMed due to its large database of full-text scholarly journals related to health science and biomedicine (National Library of Medicine, 2022). CINAHL was utilized for its large database of scholarly journals related to the nursing profession (EBSCO, 2022). Lastly, Google Scholar was also used, as this database can be used to identify highly cited journals without restrictions or limitations on file types (Martin-Martin et al., 2017). Strategies for Literature Review Inclusion/Exclusion Criteria With a research focus on perceived risk and self-reported protective behaviours towards COVID-19, my inclusion criteria included data analysis involving perceived risk and protective behaviours. Additionally, I wanted to include research that involved the adult population only as adolescents and children may not understand the dynamics of the COVID-19 pandemic. Studies that focused primarily on the outcomes of having the COVID-19 pandemic (mortality rates and symptom severity) were excluded. Lastly, studies that focused on the perception of risk from healthcare providers, such as nurses, physicians, and personal support workers (PSWs), were excluded. Refer to Table 2 for list of inclusion and exclusion criteria. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 21 Table 2 Inclusion and Exclusion Criteria for Literature Review Inclusion criteria Data was collected on perceived risk to COVID Data was collected on protective behaviours against COVID Data was collected on ethnic group/race Exclusion criteria Patient outcomes as a result of COVID infection Sample population includes only adolescents or children Studies not based on original work (e.g., editorials or opinions) Research conducted from 2019 onwards Participants are 18 years or older Any research design Scholarly and peer-reviewed studies Journal written or translated in English Literature Review Results After an extensive search of the four databases, a total of 157 total articles were retrieved from the combined search terms. However, after inclusion and exclusion criteria were applied only 11 were applicable to this thesis and included in this review (see Appendix B for Literature Review database results). Ethnic Differences in COVID-19 Rates The effect of COVID-19 on the world’s economy and population has been thoroughly analyzed and documented in the past year. It has significantly affected the lives of many people of all races and ethnicities. Both developing and developed countries alike have experienced the challenges of containing the disease within their borders; however, in countries with a diverse ethnic population, such as Canada, it is becoming more evident that ethnically diverse groups are at higher risk than non-ethnically diverse groups. In the United States, for example, where the population is predominantly non-Aboriginal Caucasian, the Black and Latinx ethnic groups have ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 22 experienced higher rates of COVID-19 infection in comparison to the rest of the population. In Canada, not only are ethnically diverse groups at higher risk of contracting COVID-19, but their risk of death from the disease is also increased (PHO, 2022). Various reasons account for these health inequities but the underlying theme within the studies suggests that socio-economic inequalities such as housing, job opportunities, and income are the predominant root causes. Socioeconomic Inequality. The discrepancy in risk for contracting COVID-19 in Canada amongst White and ethnically diverse groups may be a result of the socioeconomic inequalities that exist in the country. Specifically, income, housing, and job opportunities have been suggested to be major factors in determining one’s risk of contracting COVID-19 (Baena-Díez et al., 2020). In Canada, a report on the economic impact of the COVID-19 pandemic in Canada shows that various ethnic groups prior to the pandemic were already experiencing high rates of income loss; during the pandemic, some of these rates increased (Hou et al., 2020). In contrast, non-ethnically diverse groups have reported a decrease in economic impact, demonstrating that the pandemic has not affected their financial stability as much as ethnically diverse groups (Hou et al., 2020). Although some groups, such as individuals of Japanese descent, have also reported a decrease in pandemic impact, the overall results of the report suggest an unequal distribution of hardships experienced by ethnically diverse groups in Canada (Hou et al., 2020). For many employers, the ability to perform their employees’ job duties from home became a viable option. This option allowed employers to maintain job output while minimizing the risk of spreading COVID-19 in the workplace (Bonacini et al., 2021; Raifman & Raifman, 2020). Unfortunately, ethnically diverse groups are less likely to be employed in a job that allows them to work from home, more often working in factories or in the healthcare sector that required on-site presence putting them at higher risk of contracting COVID-19 from a colleague (Bonacini et al., 2021; Raifman & Raifman, 2020). Employment in settings where exposure to ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 23 COVID-19 is high may lead to symptomatic or asymptomatic spread of the virus within the household (Bonacini et al., 2021). This increased exposure becomes a challenge as ethnically diverse groups typically reside in multi-family housing; with more individuals residing in a single space, this can lead to a higher incidence of COVID-19 throughout the home. Ethnically diverse groups in a low-income status within Canada often experience challenges with finding affordable housing for themselves or their family. Many of these people need to resort to affordable dwellings such as multi-family households, apartments, or shelters (Hou et al., 2020). Specifically in Spain, lower-income families had a higher incidence of COVID-19 and Baena-Diéz et al. (2020) suggests that one likely cause of this is related to overcrowding of homes (i.e., multi-family housing as a result of lower income). The ability to social distance, one of the self-protective behaviours against COVID-19, in these housing configurations is more challenging than in a single-family household and has led to high incidences of COVID-19 in the community (Bonacini et al., 2021; PHO, 2020). In Ontario, PHO (2020) reports that the most ethnically diverse communities in the province have the highest percentage of low-income families (20.7% of community) and the highest average of dwellers in the household (3.1 persons per home). Coincidently, this population group also accounts for the highest percentage of COVID-19 cases in the province. Perceived Risk and Self-Reported Protective Behaviours. The relationship between self-reported protective behaviours from COVID-19 and perceived exposure risk is an area of research that has been explored quite heavily throughout the pandemic. Based on the HBM, the intention of an individual to perform protective behaviours is based on their personal belief of risk of harm; therefore, if an individual has a high perception of risk for contracting COVID-19, theoretically, they will be more likely to perform protective behaviours. This is also evident in the literature in the United States, that those who perceived the risk of contracting COVID-19 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 24 were more likely to engage in protective behaviours such as handwashing and avoiding crowds (Bruine de Bruin & Bennet, 2020). Also, in Turkey, Yildirim et al. (2021) suggest that perceived vulnerability and fear, alongside perceived exposure risk, also led to increased engagement in protective behaviours. Problem Statement As the COVID-19 pandemic continues to affect the lives of many people across the globe, there are differences in COVID-19 rates amongst different ethnic groups. Compared to the White ethnic group, ethnically diverse groups have contracted COVID-19 at a higher rate (PHO, 2022). The risks associated with contracting COVID-19 may be attributed to socio-economic inequities against ethnically diverse groups such as lower income and housing situations. Protective behaviours can also prevent contracting COVID-19, with the individual’s perception of risk predicting whether these behaviours are implemented. However, there are gaps in the research pertaining to ethnic differences in self-reported protective behaviour and perceived exposure risk. For this research, I explored whether there are ethnic differences in self-reported protective behaviours and, if so, whether these differences can be explained by concurrent differences in perceived exposure risk. Conclusion I conducted a literature review of the most relevant research regarding self-protective behaviours against COVID-19. My literature review also explored the relationship between ethnic diversity and self-protective behaviours against COVID-19. The search terms used for my search strategy consisted of ethnic diversity, perceived exposure risk to COVID-19, selfprotective behaviours, and COVID-19. Using variations of each search term, I utilized multiple databases to locate studies that address the review question. The results of my literature review suggest that ethnically diverse communities are more affected by COVID-19 in comparison to ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 25 less diverse communities. This is more prominent in more diverse countries such as Canada and the United States where ethnically diverse communities within these countries are more prone to COVID-19. This may be explained by the socio-economic hardships that individuals within these communities may experience, such as low income, multi-family homes, and jobs that require on-site presence (e.g., factory worker or health care). Although the social determinants of health may explain some of the variance in cases, there are gaps in research to identify if selfprotective behaviours differ within ethnically diverse communities. As such, my thesis focused on understanding relationships between ethnic diversity and self-protective behaviours against COVID-19 and whether, if present, these relationships are explained by differences in perceived exposure risk. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 26 Chapter 3: Research Methods To understand how ethnic diversity is related to self-protective behaviours, I chose to follow a mixed methods study design that includes two components: 1) a review of current literature on the topic of ethnic diversity and self-protective behaviours, and 2) a cross-sectional survey. I chose to conduct a rapid review as an opportunity to synthesize current research on my topic, but also to compare to and inform the survey analysis and findings. The following section first describes the design of the rapid review component of my research. I subsequently explain the design of the survey, including sampling, recruitment, data collection, and analysis methods. I also describe challenges that I faced during the research process, which led to a revised study design. Lastly, I provide an overview of how I mitigated ethical risks and submitted applications to the Trinity Western University human research ethics board (HREB) to gain approval to proceed with my research. Design My study explores various relationships related to differences in ethnic groups, perception of exposure risks to COVID-19, and self-reported protective behaviours from COVID-19. This research specifically explores: 1) Are there are ethnic group differences in protective behaviours? 2) Are there are ethnic group differences in perceived exposure risk? 3) Does perceived exposure risk explains variation in self-reported protective behaviours? and 4) If there are ethnic differences in protective behaviours, are these mediated by ethnic group differences in perceived exposure risk to COVID-19? Rapid Review Search Strategy For my research synthesis, I followed the guidelines by the Collaboration of Environmental Evidence (2018) which begins with developing a search question followed by ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 27 creating a search protocol, conducting a systematic search using a repeatable strategy, article screening, critical appraisal and data extraction, data synthesis, and a written report. To guide my search strategy, I followed the question: From 2020-2022, were there ethnic group differences in self-reported protective behaviours towards COVID-19, and if so, were they affected by ethnic differences in perceived exposure risk? I followed a PICOT format for my research question with the population as ethnic groups, intervention as perceived risk towards COVID-19, comparison as other ethnic groups, outcome as self-protective behaviours towards COVID-19, and timing from 2020-2022. I have chosen this research question for my rapid review as it is closely related to the target of my study: to describe ethnic differences in selfprotective behaviours and, if so, are these mediated by ethnic differences in perceived risk. By creating a rapid review research question that is closely related to my research objectives, I endeavored to conduct a research synthesis with findings that are comparable to the findings of my collected survey data. Although the inclusion/exclusion criteria for the literature review in Chapter Two and rapid review are similar, the data extraction, synthesis, and results between the two reviews are different as the literature review focused more on identifying and explaining the scope of the issue while the rapid review focused on synthesizing the results of research conducted on the topic. I used the TWU OneSearch library, Medline, LitCovid, COVID-19 Evidence Network to Support Decision-Making (COVID-END), and Google Scholar databases to search for relevant research. I chose to use the databases LitCovid and COVID-END due to their focus on possessing research articles related to COVID-19. The TWU OneSearch library and Medline were also used primarily due to their large database of articles related to health. I avoided using grey literature as I wanted to focus on peer-reviewed articles. However, there is value with grey literature, especially during the COVID-19 pandemic, whereby rapid changes during the ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 28 pandemic are increasing the need and use of grey literature sources across the world which are historically quicker to publication than scholarly papers (Kousha et al., 2022). Search terms for my search strategy focused on four specific keywords: COVID-19, ethnic groups, perceived risk towards COVID-19, and protective behaviours against COVID-19 (see Appendix C for list of search strings used for rapid review). Search strings for the key terms COVID-19 were adapted from the Canadian Agency for Drugs and Technologies in Health (2020). For article selection, I utilized the PRISMA model diagram (Page et al., 2021) to facilitate selection of appropriate studies. Inclusion and exclusion criteria for articles (refer to Table 3 for inclusion and exclusion criteria for rapid review) were developed and used in the process of selecting appropriate articles. For article selections, I selected only scholarly peer-reviewed studies from 2020-2022. Additionally, I included only studies that specifically analyzed and reported ethnic differences in self-reported protective behaviours or perceived risk. Table 3 Inclusion/Exclusion Criteria for Rapid Review Inclusion Criteria Scholarly Peer-Reviewed Published between 2020-2022 Sample population includes participants over the age of 18 Article analyzes ethnic differences with perceived risk and protective behaviours towards COVID-19 Exclusion Criteria Patient outcomes as a result of COVID-19 infection Sample population includes only adolescents or children Studies not based on original work (e.g., editorials or opinions) Studies published in a language other than English Article does not analyze ethnic differences with perceived risk or protective behaviours towards COVID-19 Sample population inclusive of only one ethnic group ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 29 After retrieving articles from the various databases using the search terms, I screened the title of each article to determine relevancy for my research question. After, I read the abstract of each included article to further determine appropriateness for my research. Lastly, I conducted a full-text review and applied inclusion/exclusion criteria to further filter out articles from my synthesis. As a result, after removing duplicates, eight articles were included in the rapid review. Figure 2 PRISMA Diagram Data Extraction Articles that were included in the research synthesis were read and evaluated by the principal investigator. The data extraction and evaluation of included articles were then validated by the secondary investigator. Extraction of data included: title, authors, year of publication, country of research origin, sample characteristics, study design, methodology, research objectives, article appraisal. Additionally, data was extracted to answer the research questions: whether there are ethnic group differences in protective behaviours, whether there are ethnic group differences in perceived exposure risk, whether perceived exposure risk explains variation ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 30 in self-reported protective behaviours, and if there are ethnic differences in protective behaviours, are these mediated by ethnic group differences in perceived exposure risk to COVID-19. For quality appraisal of included articles, I used the Mixed-Methods Appraisal Tool (MMAT), an easy-to-use and reliable method of appraising articles for their strengths and weaknesses (Hong et al., 2018). Within the MMAT, I primarily assessed my selected articles under Section 4: Quantitative Descriptive Studies. In this section, I assessed the articles’ quality by reviewing the sampling strategy, sample representation, measurement appropriateness, risk of non-response bias, and statistical approach. Following the guidelines by Hong et al. (2018), which discourages ‘ranking’ the overall quality of an article, I provided a detailed presentation of each criterion of the MMAT to assess the quality of the selected articles (see Appendix D for MMAT). During each quality assessment, I reviewed the strengths, weaknesses, and gaps of each criterion within the MMAT. Analysis Analysis of the included articles for the rapid review followed a narrative review methodology as proposed by Popay et al. (2006). Furthermore, I utilized the HBM as a guiding framework to review and describe the synthesized data. Analysis of the rapid review focused on the four main research questions that were proposed in my research: 1) Are there are ethnic group differences in protective behaviours? 2) Are there are ethnic group differences in perceived exposure risk? 3) Does perceived exposure risk explains variation in self-reported protective behaviours? and 4) If there are ethnic differences in protective behaviours, are these mediated by ethnic group differences in perceived exposure risk to COVID-19? All the included studies in the rapid review were analyzed by the primary investigator and narratively synthesized to answer the primary research questions. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 31 Using my research questions to guide the rapid review, analysis of the articles included a full text read to identify results or findings which pertained to my questions. I first described the samples and research designs of each study. Next, I reviewed the results and findings. After identifying results and findings related to my research questions, I used a chart to organize and prepare for synthesis of the information (see Appendix E for data analysis of the rapid review studies). I subsequently synthesized the data and narratively interpreted the findings of my synthesis while comparing similarities and differences arising from the analysis of my survey. Survey Design We conducted an online cross-sectional survey with participant recruitment via two online platforms: Mechanical Turk (MTurk) and Facebook. Sampling Inclusion/Exclusion Criteria. For this research, Canadians over the age of 18 and who can read English were included in my sample population. I chose to include only adults who were 18 years old or older as both adaptations of the survey were validated through this population only (Breakwell et al. 2021; Zhao et al., 2020). Although participants who have received any COVID-19 vaccinations (1-dose, 2-dose, or booster shots) can develop a decreased perception of risk from the virus (Smith et al., 2020), most of the population in Canada is fully vaccinated (Government of Canada, 2021b), thus it would have been challenging to achieve my required sample size if I did not include this population. Lastly, as the survey is only delivered in English, only participants who can read English were included in the study. Procedures At the beginning of my research, I had initially endeavored to partner with a community hospital, Scarborough Health Network (SHN), and collect survey data from the multicultural community of Scarborough, Ontario. However, the research ethics board (REB) at SHN believed ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 32 the topic of COVID-19 and its relation to ethnic diversity too controversial and thus the proposal was rejected. As an alternative, I conducted my survey online using an online pool of potential participants: Amazon’s MTurk. My initial target sample size for my collected data was 330 with quota sampling goals of 55 respondents each for Black, South-Asian, and East-Asian ethnic groups. This was based on sample size guidelines for multivariable regression analyses, suggesting a minimum of 50 respondents plus 40 additional respondents per observed indicator (ethnic groups, income, age, gender, housing situation, overall health status, and perceived exposure risk) (Polit & Beck, 2017). Thus, we determined that having 330 total participants would be sufficient for minimizing the risk of a type II error (Polit & Beck, 2017). Amazon’s MTurk is a crowdsourcing website which invites users from around the world to register and participate in tasks called Human Intelligence Tasks (HITs). These tasks can encompass various objectives such as survey collection data, software development, or product surveys. Registered users of MTurk, called “workers” or “MTurkers”, can receive financial compensation for completing HITs (Amazon, 2021). Participants who completed my survey were eligible to collect a compensation of $0.50 through MTurk. According to Azambuja (2015), this is close to the recommended financial payment for short surveys on MTurk. According to Zhang and Gearhart (2020), the quality of data retrieved from MTurk is comparable, and at times better, than other traditional methods of data collection such as professional panels. Facebook allows users to advertise websites and links through their platform using Facebook Ads. Through Facebook Ads, a user pays a fee to advertise these links to a demographic of their choosing. Potential respondents were able to see the advertisement displayed in various areas of the Facebook platform where the user can subsequently click on the ad and participate in the study. Participants who completed the survey through Facebook did not receive financial compensation as I did not have a method of validating completion of the survey without compromising ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 33 confidentiality. According to Akers and Gordon (2018), Facebook can be an effective method of getting participants of a particular demographic due to the ability to create targeted ads. Additionally, Facebook fees are usually less expensive than other platform and can be a costeffective method of collecting respondents for a research study. In terms of differences in population demographics between MTurk and Facebook, participants on Facebook are typically older and have more education according to a United States study by Boas et al. (2020). In both online platforms, the Black and Latinx population is underrepresented with most users on both platforms reporting to be of a White population (Boas et al., 2020). The limitations of MTurk and Facebook for convenience sampling becomes evident with quota sampling of certain ethnic groups; it will take longer to achieve quota sampling goals when certain ethnic groups (i.e., Black, Latinx, or Asians) are not sufficiently represented. For the first three weeks of my data collection, I posted my survey solely on the MTurk platform. The survey was active on MTurk from September 28, 2021–October 20, 2021 (22 days). Through MTurk, I was able to recruit 107 respondents for my survey. The respondent rate during the second and third week of the MTurk survey was significantly lower than the initial week, thus, I opted to stop utilizing the MTurk services and started to use Facebook as a supplementary method for retrieving additional respondents. From November 9, 2021– November 23, 2021, the survey was advertised on Facebook. Through Facebook, I was able to recruit 80 additional participants for the study. In total, I recruited 187 total responses from both methods of data collection, below my intended goal of 330 responses. Due to target timelines to complete this research and increasing costs of advertisements with Facebook, I was unable to continue with my data collection despite not achieving my target sample size. With more time, I believe that we would be able to achieve not only our goal sample ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 34 size, but also recruit more participants from other ethnic groups to have a more representative sample. By doing so, this would have strengthened the validity and rigor of the observations. Measures The first set of survey questions focused on obtaining sociodemographic information from the participant (see Appendix F for sociodemographic questions with frequencies). Adapted from Pinto et al. (2016), these questions revolve around the social determinants of health and are inclusive enough to identify health inequities. I included all questions from Pinto et al. (2016) so as to provide flexibility in identifying potential covariates during my data analysis. The next section of the survey asked questions about the participant’s self-perceived risk of exposure to COVID-19. The set of six questions in this section are based on the COVID Own Risk Appraisal Scale (CORAS) by Jaspal et al. (2020). Jaspal et al. (2020) performed confirmatory factor analysis (CFA) and determined that certain questions within the survey had a poor fit or lack of validity (refer to Table 4 for set of CORAS questions used in the survey). This resulted in a set of six valid and reliable questions which are included in this study. Each question is in the form of a Likert scale from 1 (Strongly disagree; Very hard to do; Extremely unlikely) to 5 (Strongly agree; Very easy to do; Extremely Likely). Lower ratings on the CORAS suggest lower perceptions of risk to COVID-19 while higher ratings suggest higher perceptions of risk (Jaspal et al., 2020). According to Jaspal et al. (2020), convergent validity of the CORAS in comparison to another scale, the Fear of COVID-19 Scale, showed a positive correlation (Spearman’s rho = 0.54, p < 0.001) validating the use of CORAS. Additionally, the study by Breakwell et al. (2021) utilized the CORAS to predict preventive behaviours in the United Kingdom and provides confidence of adequate reliability and validity regarding the use of this tool for population-level appraisals of fear and risk perceptions. Although the CORAS has been validated only in the United Kingdom, I believe that the results of CORAS on a diverse ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 35 population present preliminary validity and thus CORAS should be appropriate for my sample population. Table 4 COVID-19 Own Risk Appraisal Scale (CORAS) (Extremely Unlikely) 1 2 3 4 (Extremely Likely) 5 Prefer not to answer (Strongly Agree) 1 2 3 4 (Strongly Disagree) 5 Prefer not to answer (Very Hard to Do) 1 2 3 4 What is your gut feeling about how likely you are to get infected with COVID-19? I think my chances of getting infected with COVID-19 are I am sure I will NOT get infected with COVID-19 I feel I am UNLIKELY to get infected with COVID19 (Very Easy to Prefer not to Do) answer 5 Picturing myself getting COVID-19 is something that I find The final section of the survey asked participants to rate their level of engagement of certain protective behaviours. This section had a set of five questions that were taken from the Understanding America Study (UAS) and chosen based on the recommendations from Health Canada, which advises individuals to avoid large crowds, perform hand-hygiene, and wear a face mask in public (refer to Table 5 for set of questions pertaining to self-protective behaviours). The UAS is a survey by University of Southern California (USC, 2020) which is nationally ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 36 distributed in the United States. The UAS has been administered biweekly to a panel of 3000+ participants using address-based probability sampling since May 2020. According to Flannelly et al. (2018), internal validity is improved through frequent administration (i.e., history or changes that occur throughout the experiment) as to have current and up-to-date results. The data retrieved by the frequent administration of UAS has been used in many studies to research a plethora of subjects including COVID-19 (USC, 2020). According to Angrisani et al. (2019) the UAS is comparable to other probability online surveys such as the Health and Retirement Study (HRS) and the Current Population Survey (CPS) in terms of reliability and response rates. The questions revolve around self-protective behaviours from COVID-19 and ask the participants about their engagement in certain protective behaviours within the past seven days. The participants will answer “Yes” or “No” on whether they have performed the behaviours specified in the question. Table 5 Questions pertaining to self-protective behaviours No Washed your hands with soap or used hand sanitizer several times per day Avoided contact with people who could be high-risk (e.g., elderly, children, immunocompromised) Avoided public spaces, gatherings, or crowds Avoided eating at restaurants Worn a mask or other face covering Yes Prefer not to answer ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 37 Amended Study Design I initially intended to perform a multivariable regression analysis with my collected data. However, my sample size of 187 was well below my desired sample size of 330. Additionally, a large percentage of the sample population was comprised of the White ethnic group (n=126, 67.4%) and, despite attempts to conduct quota sample techniques, I was unable to meet my quota (55 respondents each for Black, South-Asian, and East-Asian ethnic groups) for certain ethnic groups. As a result, I adjusted my methods using a mixed-methods strategy that involved the following stages: 1) I first conducted a rapid review of current literature on ethnic differences in self-reported protective behaviours mediated by perceived risk towards COVID-19. 2) I then conducted quantitative analyses of my collected data comparing perceived risk and selfprotective behaviours towards COVIDs between White and non-White ethnic groups. 3) And lastly, I analyzed and compared both results concurrently to answer my overarching research aim and objections. According to Dobbins (2017), a rapid review is a simplified version of a systematic review and is used to synthesize information in a timelier manner. For my study, this timeliness is advantageous as a rapid review requires less time to complete and less resources (can be completed by a single investigator) (Dobbins, 2017). Due to rapid changes in guidelines and positive cases occurring during the COVID-19 pandemic, time is of high importance, warranting the use of a rapid review. Although a rapid review can be informative and synthesize important information in a short period of time, Munn et al. (2018) suggest that due to shorter time commitments and fewer resources, a rapid review may be less detailed or extensive than a systematic review. Survey Administration The survey was created and administered online through the platform: SurveyMonkey (see Appendix G for printable copy of online survey). SurveyMonkey is a data collection service ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 38 which allows users to create free surveys. The initial page of the survey is an introduction page that outlined important information regarding participation. Items included in the introduction page are response anonymity statements, investigator contact information, research withdrawal information, and instructions on who to contact if the participants desire additional information. Additionally, I included a statement to notify participants that responses to the survey are completely anonymous and as such I do not collect any identifiable information such as name, phone number, or government-issued IDs. As this survey is anonymous, submitted responses could not be removed from the data. However, the participants were able to choose which questions they were willing to answer. Contact information of the principal investigator and the HREB of TWU were included should the participants have any questions or desired additional information. In total, there were 25 questions with a mixture of sociodemographic questions, Likert scales, and Yes/No questions. The survey required less than 10 minutes to complete which, according to Revilla and Ochoa (2017), was the ideal length for an online survey. Participants who successfully completed the survey through MTurk were compensated $0.50 through the MTurk platform. To utilize the services of MTurk, it was mandatory for all requesters to provide financial compensation to participants (Amazon, 2021). I chose $0.50 for compensation as it is in line with the recommended reward for short or quick HITs on MTurk (Azambuja, 2015). To be eligible to receive compensation for my survey, participants were required to input a completion identifier (see Appendix H for MTurk completion code) on the MTurk website. This identifier was only accessible on the final page of a completed survey; therefore, participants who inputted the incorrect code were not eligible for compensation as it would have been assumed that they did not complete the survey to completion. Each day, the principal researcher reviewed each submitted response on MTurk. Submitted responses that contained an invalid completion identifier were deemed ineligible for receiving compensation. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 39 Compensation to the participants was provided solely through MTurk. MTurk subsequently provided an invoice to the researchers, along with a 20% fee, for payment of services. Only MTurk possessed the personal information of each participant; however, the responses provided in the survey were only viewable by the researchers. Quantitative Data Analysis Following the rapid review, I conducted a descriptive statistical analysis and a series of quantitative analyses including cross-tabulations, logistic regressions, and linear regressions on the survey data to address each of the research questions. The dependent variables are those pertaining to Perception of Exposure Risk (section 2 of the questionnaire) and Self-reported Behaviours (section 3 of the questionnaire). The primary independent variable was ethnic group. However, due to the small sample size and lack of ethnic diversity in the sample, I collapsed respondents’ self-reported ethnicity into two ethnic groups: White ethnic group and Ethnicdiverse (non-White) group. Religious beliefs, born in Canada, vaccination status, and general health status were included as potential covariates in all regression analyses. Although the focus of the study was to analyze the potential relationship between ethnic groups and self-reported protective behaviours, it was important to analyze the other independent variables to identify relevant covariates. To explore my research questions, I first completed a descriptive analysis of the sociodemographic variables of the sample population. Cross-tabulations with corresponding chisquare tests and odds ratios were conducted to compare the ethnic groups with each of the covariates, including gender, age, income, born in Canada, described overall health, and vaccination status. Percentages of the sample population were calculated with a corresponding pvalue as a measure of statistical significance. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 40 I followed the guidelines by Hosmer et al. (2013) for purposeful variable selection, which is a process for creating a lean and parsimonious analysis model (Heinze et al., 2018). Purposeful variable selection assists descriptive research in explaining the outcomes between independent and dependent variables. A model with too many variables may dilute the associations between the dependent and independent variables, whereby a model with too few variables may not have outcomes that accurately describe the observations (Heinze et al., 2018). Firstly, I conducted a cross-tabulation with corresponding chi-square tests between the ethnic groups and each of the sociodemographic variables (independent samples T-test was used for the age variable). Sociodemographic variables observed to have a significant effect (p=<0.25; Hosmer et al., 2013) on the ethnic groups were considered potential confounders and included in the final analysis model. Secondly, I analyzed the associations between CORAS and the sociodemographic variables by conducting linear regressions with CORAS as the dependent variable and sociodemographic variables as the independent variable. Sociodemographic variables with more than two categories were dummy coded; this included income, described overall health status, and vaccination status. Any observed statistically significant associations (at p<.25) between the independent and dependent variables were considered potential confounders and included in the final analysis model (Hosmer et al., 2013). Lastly, I conducted chi-square analysis of selfreported protective behaviours with the sociodemographic variables and any observed statistically significant associations (at p<.25) were included in the final multivariable regression analyses. By conducting purposeful variable selection, I aimed to create an analysis model that best described the outcomes between ethnic groups, self-perceived risk towards COVID-19, and reported self-protective behaviours against COVID-19. Altogether, the final analysis model consisted of ethnic groups, religious beliefs, born in Canada, overall health status, vaccination status, and CORAS scores. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 41 Prior to conducting analyses on the survey data, I performed a frequency analysis of the covariates to identify missing cases. In a binomial logistic regression, variables with missing cases would have been removed from the data analysis in its entirety and the number of valid responses would have decreased (i.e., listwise deletion) and may lead to a biased estimate (Polit & Beck, 2017). Based on the survey data, the religious variable contained 12 missing cases because of respondents answering “Other” or “Prefer not to answer”. To address this, I reviewed the answers from the “Other” section and recoded two of the responses (Agnostic and Methodist) to indicate a religious belief and the other two responses (Norse Paganism and Non-secular humanism) as not having a religious belief. For the respondents who answered “Prefer not to answer”, we consolidated their response with the “No” category. This resulted in the religious variable containing 176 valid responses with no missing cases. For the income variable, the number of missing cases was 28 and for the age variable there were 35 missing cases. Due to the nature of these two variables, I could not recode any of the responses or consolidate them into another category, thus they have been excluded from the final analysis model. With consideration of the potential confounding effect of income and age on self-protective behaviours, I conducted a sensitivity analysis through a binomial logistic regression of the final analysis model with the income and age variables, narratively describing the results and differences with the original model. After identifying all the variables to be included in the final analyses, I conducted a multivariable logistic regression analysis to identify relationships between ethnic groups, with sociodemographic variables as covariates, and the self-reported behaviours. A separate logistic regression was conducted for each self-reported behaviour variable (i.e., hand-hygiene, avoiding large crowds or public spaces, avoiding eating at restaurants, avoiding high-risk individuals, and ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 42 wearing masks in public). Assumptions related to logistic regression (linearity, multicollinearity, and outliers) were examined. Assumptions for linearity were met through the Box-Tidwell (Box & Tidwell, 1962) procedure. No multicollinearity was identified (p-value < .8; Shrestha, 2020). Outliers were assessed via SPSS and one case was identified as an outlier for the analysis of hand hygiene (SD=2.63) and wearing masks (SD=2.85); after assessment the case was included in the analysis. Then, I performed a multivariable linear regression analysis to examine the association between ethnic groups and covariates as independent variables and perceived exposure risk as the dependent variable. Assumptions related to linear regression models (linearity, normality, homoscedasticity, and multicollinearity) were examined and met (Casson & Farmer, 2014). After analyzing the relationship between ethnic groups and perceived exposure risk, I analyzed the effect of perceived exposure risk on self-reported protective behaviours, which included ethnic groups and sociodemographic variables as covariates through binomial logistic regressions. Lastly, I analyzed the relationships between the variables and determined if self-reported behaviours are mediated by ethnic differences in perceived exposure risk. Answering the Research Questions After completing the rapid review and analysis of the survey data, we compared the results and, using the rapid review to inform our survey data results, consolidated the information for a narrative discussion of the findings. To answer the research question “Whether there are ethnic group differences in protective behaviours,” we extracted the data from the rapid review and used it to inform the findings from the binomial logistic regression pertaining to ethnic groups and self-protective behaviours. To answer the question “Whether there are ethnic group differences in perceived exposure risk,” we used the findings of the rapid review to inform the results of the linear regression between ethnic groups and the CORAS. To answer the question ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 43 “Whether perceived exposure risk explains variation in self-reported protective behaviours,” we used the rapid review to inform the results of the binomial logistic regression between ethnic groups (including the CORAS scores) and self-protective behaviours. Lastly, we only used the results of the rapid review to answer the question “If there are ethnic differences in protective behaviours, whether this is mediated by ethnic group differences in perceived exposure risk to COVID-19.” As part of my mediation model for this thesis, a direct effect between ethnic differences and self-protective behaviours needs to be identified prior to analyzing for an indirect effect (i.e., mediating factors) (Hayes, 2009). As the results of our survey data did not find a mediating factor, we are only using the results of the rapid review to answer the research question. Ethics Ethical Risks with Human Participation The study collected information about a participant’s demographics, perception of risk, and self-reported behaviours against COVID-19 via an online survey. I did not believe that the participants would be exposed to any physical risks associated with participating in the study. There is, however, some level of psychological harm that could have arisen from completing the survey such as feelings of embarrassment, stress, or guilt. For example, a participant may feel guilty if they mentioned that they do not engage in self-protective behaviours such as maskwearing. Additionally, feelings of stigma may have altered the way a participant answers a question (social desirability bias). For example, a participant may have indicated that they perform hand-hygiene as it may be the most responsible thing to do but may not be what they practiced at home. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 44 To mitigate these risks, prior to the survey, I indicated that no personal identifiers would be collected from the participant. These included names, addresses, government-issued IDs (e.g., health card numbers, driver’s license numbers, or passport information). This strategy encouraged the participant to be truthful and alleviate any fears that responses would be attributed to them. Regarding inclusion criteria, I only included participants who were able to read English. Due to time constraints and time sensitivities related to the COVID-19 pandemic, a translation of the survey into other languages was not feasible. To mitigate this, I checked the survey questions and explanations provided were readable at a grade eight grammar level. Human Research Ethics Boards (HREB) As this study included human participants, it was approved by the TWU Human Research Ethics Board (HREB) on June 21, 2021. The initial iteration of data collection included the use of a community hospital’s internet website, Scarborough Health Network (SHN), for advertisement of the survey. However, SHN rejected the proposal as it was deemed too polarizing of a subject which can lead to racial discrimination towards certain ethnic groups. Conclusion At the beginning of my research, I initially wanted to conduct a survey in a multi-cultural community in Ontario: Scarborough. I endeavored to partner with the largest community hospital in that city, Scarborough Health Network (SHN), however, I was unable to get approval from their research ethics board. As a result, I chose to collect data via MTurk, an online database of potential participants that is hosted by the online retailer Amazon. I believed I would have been able to achieve my target number of participants of 330 with MTurk; unfortunately, I were unable to achieve this during the 22 days I used MTurk. As a result, I advertised my survey on ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 45 Facebook to supplement my current collection for an additional 14 days. Despite using various avenues to collect data, I still fell short of my desired target of 330 participants with only 187 confirmed respondents to the survey. With a smaller sample size, I chose to adjust my strategies of doing solely a quantitative analysis of my data and decided to conduct a mixed methods design by completing a rapid review of current literature on my research topic and subsequently comparing the findings with my analyzed data collection. For the rapid review, I developed a research question following the PICOT format to guide my search strategy. After collecting, analyzing, and carefully selecting my research articles using the PRISMA diagram, I conducted full-text reviews of the articles to assess their data, strength, and quality. Once the finalized list of included research articles was determined, I completed a research synthesis of the findings and narratively reviewed the information in this thesis. I utilized the rapid review findings to inform the survey findings and interpretation. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 46 Chapter 4: Findings I initially conducted a rapid review of current research regarding ethnic diversity and self-reported protective behaviours against COVID-19 and subsequently completed the quantitative analysis of the survey data from my sample population. A total of eight studies were included in the rapid review. Out of the eight articles, six included a sample population from the United States with the other two articles originating from the United Kingdom (see Appendix E for a data analysis of rapid review studies). In the studies from the United Kingdom, the sample population consisted of White and Black, Asian and Minority Ethnic (BAME) groups while the studies from the United States consisted of White, Black, Asian, and Latinx groups. All of the studies in the rapid review were published in 2021 with all studies utilizing online surveys for data collection (five studies used cross-sectional surveys and three studies used survey data from another source). The results of the rapid suggest that non-White groups (i.e., BAME groups) engage more in self-protective behaviours and have higher perception of risk towards COVID-19 than White groups. I used the results of my rapid review to inform the results of the data analysis of my sample population; therefore, I integrated the findings of both the rapid review and my sample population data in my report. For the descriptive analysis, I provide details about the sample population’s gender, income, and overall health status, which were identified as common covariates analyzed in the articles used within the rapid review (see Table 6 for covariate analysis). To explain my findings, I used the research questions within my mediation model to guide my analysis: 1) Are there are ethnic group differences in protective behaviours? 2) Are there are ethnic group differences in perceived exposure risk? 3) Does perceived exposure risk explain variation in self-reported protective behaviours? and 4) If there are ethnic differences in ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 47 protective behaviours, are these mediated by ethnic group differences in perceived exposure risk to COVID-19? Descriptive Survey Results For my survey data, a total of 187 total respondents were retrieved from MTurk and Facebook. Respondents that selected “Prefer not to answer” or an invalid race to question 6: Which of the following best describes your race? were deemed ‘Missing Cases’ as I would not have been able to categorize them in one of the groups. In total, I retained 176 valid responses for my study. Out of 176 responses, the responses were separated into two groups: White ethnic group (n=127) and Ethnically diverse group (n=49). Refer to Figure 3 for a graph showing the frequencies of the ethnic group responses. Respondents who reported their race as “White/European (e.g., English, Italian, Portuguese, Russian)” were sorted into the White ethnic group while all other responses were categorized into the Ethnically diverse group (non-White). Refer to Table 7 for descriptive analysis of the sample population. Gender For gender, 108 respondents were male (White group=76; Ethnically diverse group=32) while 67 were female (White ethnic group=50; Ethnically diverse group=17). One respondent reported non-binary as a gender. In my Chi-square descriptive analysis, I conducted four different analyses with adjustments to the non-binary response: converted the response to ‘Male’, converted the response to ‘Female’, considered the response as a ‘Missing Case’, and left the response as is. Each of these analyses revealed similar results. To avoid excluding this response, I therefore opted to arbitrarily convert the response to ‘Male’ in my final analysis as there are more male respondents in the sample population. Cross tabulation results of the gender analysis ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 48 between the ethnic groups are as followed: there were 50 females (39.4% of White group) and 76 males (60.6% of White group) in the White group while the Ethnically diverse group had 17 females (34.7% of Ethnically diverse group) and 32 males (65.3% of Ethnically diverse group). The Chi-square difference test comparing the White and Ethnically diverse groups suggests that the two groups were similar in regard to gender (χ2(1) = .328, p=.567). Age Age was calculated in SPSS by retrieving the data from the variable “What year were you born?” and subtracting the year from 2021 (the period when the data collected). The question was optional and respondents were instructed to leave the section blank if they did not want to answer. Respondents who did not answer the question were deemed as “Missing Cases” (N=27) and were not included in the final analysis model. In total, 149 respondents answered the question and I subsequently calculated the age and created a continuous variable in SPSS to utilize for analysis. Collectively, the mean age for the sample population was 45.14 (SD: 17.01). The mean age for the White group was 48.66 (SD: 17.03) and for the Ethnically diverse group it was 35.07 (SD: 12.11). The difference had a small p-value of less than .001. Income For income, respondents were asked to indicate their total household income with answers ranging from less than $10,000 to greater than $150,000 in increments of $10,000. Respondents who responded with “Prefer not to answer” or “Do not know” were deemed as “Missing Cases” (White group=18; 14.2% missing cases; Ethnically diverse group=4, 8.2% missing cases) and thus omitted from the analysis. In SPSS, I consolidated the responses into three groups: <$10,000-$49,000; $50,000-$99,000; and >$100,000. In the White group, there is a fairly equal distribution of respondents in each category (<$10,000-$49,000 [N]=35; ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 49 32.1%;$50,000-$99,000 [N]=37; 33.9%; >$100,000 [N]=37, 33.9%). In the Ethnically diverse group, there is a higher percentage of respondents in the $50,000-$99,000 category ([N]=20, 44.4%) than in the other categories (<$10,000-$49,000 [N]=17, 37.8%; >$100,000 [N]=8, 17.8%). The difference between the White and Ethnically diverse groups had a p-value of .045. Overall Health Status For overall health status, respondents were asked to indicate their overall health status using a Likert-scale type question ranging from Poor to Excellent. Respondents were also given the option to indicate “Do not know” or “Prefer not to answer”. Respondents who indicated “Do not know” or “Prefer not to answer” were deemed as a “Missing case” (N=1) and were not included in the analysis. I used an ordinal logistic regression to compare the overall health status of the White ethnic group and Ethnically diverse groups. The results showed reported overall health status between the White ethnic group and the Ethnically diverse group were similar (p=.505). Born in Canada For those born in Canada, respondents were asked to indicate if they were born in Canada with a Yes or No answer. In the White group, 108 respondents (85.4% of White group) were born in Canada while 19 respondents (14.6% of White group) reported “No”. In the Ethnically diverse group, 30 respondents (61.2% of Ethnically diverse group) reported being born in Canada while 19 respondents (38.8% of Ethnically diverse group) reported “No”. The differences between the White group and the Ethnically diverse group had a small p-value of .001. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 50 Vaccination Status For vaccination status, respondents were asked to indicate if they are fully vaccinated (two or more doses), partially vaccinated (one dose), or not vaccinated. In the White group, 86 respondents (69.4% of White group) reported being fully vaccinated; 7 respondents reported (5.6% of White group) being partially vaccinated; and 31 respondents (25% of White group) reported not being vaccinated. In the Ethnically diverse group, 41 respondents (83.7% of Ethnically diverse group) reported being fully vaccinated; 5 respondents (10.2 % of Ethnically diverse group) reported being partially vaccinated; and 3 respondents (6.1% of Ethnically diverse) reported not being vaccinated. The differences between the White group and Ethnically diverse group had a small p-value (.015). ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 51 Table 6 Cross tabulation of self-protective behaviours with demographics Washed Hands with Soap and water No Yes Were you born in Canada? No 2.3 17.9 Yes 13.3 67 Religious Beliefs? No 11 33 Yes 4.6 51 Income <10,000-49,000 5.2 28.8 50,000-99,000 3.9 33.3 >100,000 2.6 26.1 General Health Status Excellent 2.3 11 Very Good 4.1 25 Good 6.4 35 Fair 1.7 10.5 Poor 0.6 3.5 Vaccination Status Fully Vaccinated 8.2 65.9 Partially 0.6 6.5 Vaccinated Not vaccinated 5.9 12.9 Gender Male 11 50.3 Female 4.6 34.1 pvalue Avoided high-risk individuals No pvalue Yes Avoided Public Spaces No pvalue Avoided Restaurants Yes No pvalue Yes Wore a mask in public No pvalue Yes 0.446 5.8 28.3 14.5 51.4 0.439 27.3 6.4 52.3 14 0.748 31.8 8.1 48 12.1 .988 1.4 1.2 69.4 19 0.226 0.003 15.6 18.5 28.3 37.6 0.727 15.1 18.6 29.1 37.2 0.904 17.9 22. 26. 34.1 .83 7.5 4. 36.4 52. 0.044 0.593 1.5 7.8 12.4 23.5 29.4 16.3 0.057 9.9 6.6 13.2 24.3 3.9 15.1 0.007 8.5 13.1 15. 25.5 24.2 13.7 .21 3.9 2.6 1.3 3.1 34.6 27.5 0.428 4.7 8.1 16.3 4.10 1.20 8.7 2.9 25 8.1 2.9 0.768 4.1 9.9 15.2 3.50 0.6 9.4 18.7 26.3 8.80 3.50 0.767 5.2 12.8 16.9 4.7 . 8.1 16.3 24.4 7.6 4.1 .281 1.7 3.5 2.9 2.3 1.2 11.6 25.6 38.4 9.9 2.9 0.333 20 2.4 54.1 4.7 0.002 20.1 0.6 53.8 6.5 <.001 28.8 1.2 45.3 5.9 .128 2.4 1.2 71.8 5.9 11.2 7.6 12.4 6.5 9.4 9.4 7.1 11.8 11.6 22.5 27.2 38.7 20.9 12.8 40.1 26.2 12.7 27.2 26 34.1 8.1 3.5 53.2 35.3 0.997 0.013 .291 .348 .845 .332 <.001 .394 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 52 Table 7 Descriptive statistics of the sample population Characteristics Gender Male Female Age Mean years (SD) Min-Max (years) Missing cases Income <$10,000-$49,000 $50,000-$99,999 >$100,000 Missing cases Were you born in Canada? Yes No Described Overall Health Excellent Very Good Good Fair Poor Missing cases Vaccination status Fully Vaccinated Partially Vaccinated Not Vaccinated Prefer not to answer Total (n=176) % White Group (n=127) % Ethnic Group (n=49) % p-value 61.93 38.07 60.63 39.37 65.31 34.69 .567 45.14 (17.01) 20-85 15.3 48.66 (17.03) 20-85 15.7 35.07 (12.11) 20-67 14.3 29.55 32.39 25.57 12.49 27.56 29.13 29.13 14.18 34.69 40.82 16.33 8.16 78.41 21.59 85.04 14.96 61.22 38.78 .001 13.07 29.55 40.91 11.93 3.98 0.56 13.39 26.77 44.09 12.60 3.15 0 12.24 36.73 32.65 10.20 6.12 2.06 .505 72.16 6.82 19.32 1.70 67.72 5.51 24.41 2.36 83.67 10.20 6.12 0 .015 <.001 .045 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 53 Figure 3 Frequency of Ethnic Groups Note. Frequency numbers are shown above each ethnic group Research Question #1: Ethnic Group Differences in Protective Behaviours Rapid Review Results Out of the eight included studies, five reviewed ethnic differences in protective behaviours. Protective behaviours in this subset of articles included hand-hygiene (Barrett & Cheung, 2021), social distancing (Breakwell et al., 2021; Orom et al., 2021), and mask usage (Hearne & Niño, 2021). Regarding ethnic group differences in protective behaviours against COVID-19, some of the studies included in the rapid review suggest that there are differences between the Whiteethnic group in comparison to non-White ethnic groups. Breakwell et al. (2021) found that participants in the Black, Asian, and Minority Ethnic (BAME) group, compared to the White British population, are associated with higher compliance with COVID-19 preventive behaviours (β=.22; p=<.001) on the COVID-19 Preventive Behaviours Index (CPBI) questionnaire. The ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 54 CBPI contains questions regarding preventive behaviours including mask wearing, social distancing, and hand-hygiene and asks participants to rate their likelihood of engaging in those behaviours using a Likert scale from 1 (not likely at all) to 5 (very likely) (as cited in Breakwell et al., 2021). More specifically, it is found that non-White ethnic groups are associated with higher engagement with mask wearing in public in comparison to White ethnic groups (Garfin et al., 2021; Hearne & Nino, 2021). Furthermore, in the study by Orom et al. (2021), a higher percentage of respondents in the Black group (77%) reported avoiding others in person in comparison to the White group (59.35%; p=<.01). In contrast, the article by Barrett and Cheung (2021) found no ethnic group differences with protective behaviours. However, the sample population for this particular study consisted primarily of university students with the majority of participants being between the ages of 1825; according to Bruine de Bruin and Bennett (2020), younger-aged individuals are less likely to engage in protective behaviours. Quantitative Survey Results For hand-hygiene, a total of 173 responses were recorded (White group: 124 total responses; Ethnic-diverse group: 49 total responses). In the White group, 100 respondents reported ‘Yes’ to engaging in hand-hygiene with 24 respondents reporting ‘No’. For the Ethnically diverse group, 46 respondents reported ‘Yes’ to engaging in hand-hygiene with 3 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results of the regression showed that the model explained 20.7% (Nagelkerke R²) of variance in reporting hand-hygiene practices and correctly classified 85.8% of cases. The results in Table 8 showed that the Ethnically diverse group were 4.04 times (95% CI, .844 to 19.33; ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 55 p=.081) more likely to report washing their hands with soap or hand sanitizer several times a day compared to the White group. However, when income and age were included in the model, the odds ratio decreased to a value of 1.475 (95% CI, .248 to 8.76) and a larger p-value of .669 (N=136). For the covariates, those who reported having a religious belief were 4.42 times (95% CI, 1.616 to 12.072; p=.004) more likely to report washing their hands with soap or hand sanitizer several times a day compared to those who report not having a religious belief. Table 8 Binomial logistic regression of hand-hygiene B Ethnic Group (White 1.396 group = referent) Religious Affiliation 1.485 (No = referent) Were you born in Canada? 0.483 (Yes = referent) Vaccination Status Vaccinated 0.958 Partially Vaccinated 1.402 Not Vaccinated (referent) In general, would you say your health is: Excellent -0.734 Very Good -0.274 Good 0.018 Fair -0.002 Poor (referent) Note. N=169. OR = odds ratio. 1 pvalue 0.081 4.038 95 CI for OR Lower Upper 0.844 19.328 8.385 1 0.004 4.417 1.616 12.072 0.825 0.342 1 0.559 1.620 0.322 8.164 0.545 1.196 3.096 1.375 1 1 0.078 0.241 2.607 4.065 0.897 0.390 7.583 42.389 1.361 1.273 1.259 1.374 0.291 0.046 0.000 0.000 1 1 1 1 0.590 0.829 0.989 0.999 0.480 0.760 1.018 0.998 0.033 0.063 0.086 0.067 6.910 9.217 12.008 14.754 S.E. Wald df 0.799 3.053 0.513 OR For avoiding high-risk individuals, a total of 173 responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 73 respondents reported ‘Yes’ to “Avoided high-risk individuals” with 51 respondents reporting ‘No’. For the Ethnically diverse group, 41 respondents reported ‘Yes’ to avoiding high-risk ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 56 individuals with 8 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results showed that the model explained 15.2% (Nagelkerke R²) of variance and correctly classified 71.6% of cases. The results of Table 9 showed that Ethnically diverse groups were 2.69 times (95% CI, 1.09 to 6.64; p=.032) more likely to avoid high-risk individuals than the White group. However, when income and age were included in the model, the odds ratio decreased to a value of 1.484 (95% CI, .509 to 4.33) and a larger p-value of .469 (N=136). For the covariates, fully vaccinated respondents were 3.88 times (95% CI, 1.61 to 9.33; p=.002) more likely to avoid high-risk individuals compared to nonvaccinated respondents. Table 9 Binomial logistic regression of avoiding high-risk individuals B Ethnic Group (White 0.988 group = referent) Religious Affiliation 0.139 (No = referent) Were you born in Canada? -0.256 (Yes = referent) Vaccination Status Vaccinated 1.356 Partially Vaccinated 0.966 Not Vaccinated (referent) In general, would you say your health is: Excellent -0.502 Very Good -0.118 Good -0.532 Fair -0.410 Poor (referent) Note. N=169. OR = odds ratio. 1 pvalue 0.032 2.687 95 CI for OR Lower Upper 1.087 6.643 0.152 1 0.697 1.149 0.571 2.312 0.478 0.286 1 0.593 0.774 0.303 1.976 0.448 0.760 9.159 1.616 1 1 0.002 0.204 3.879 2.627 1.612 0.593 9.332 11.649 1.015 0.960 0.935 1.029 0.244 0.015 0.324 0.159 1 1 1 1 0.621 0.902 0.569 0.690 0.606 0.889 0.587 0.664 0.083 0.135 0.094 0.088 4.424 5.838 3.667 4.984 S.E. Wald df 0.462 4.581 0.357 OR ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 57 For avoiding public spaces, gatherings, or crowds, a total of 172 responses were recorded (White group: 123 total responses; Ethnically diverse group: 49 total responses). In the White group, 76 respondents reported ‘Yes’ to avoiding public spaces, gatherings, or crowds with 47 respondents reporting ‘No’. For the Ethnically diverse group, 38 respondents reported ‘Yes’ to avoiding public spaces, gatherings, or crowds with 11 respondents reporting ‘No’. After listwise deletion, a total of 168 valid cases were included in the binomial logistic regression. The results showed that the model explained 33% (Nagelkerke R²) of variance and correctly classified 74.5% of cases. The results of Table 10 showed that the Ethnically diverse group had 1.6 times greater odds of avoiding public spaces, gatherings, or crowds than the White group. However, the confidence interval was large (95% CI, .674 to 3.8; p=.287) and, when income and age were included in the model, the odds ratio decreased to a value of 1.43 (95% CI, .455 to 4.51) and an even larger p-value of .540 (N=135). For covariates, fully vaccinated individuals were 4.75 times (95% CI, 1.93 to 11.68; p=.001) more likely to avoid public spaces, gatherings, or crowds compared to non-vaccinated respondents while partially vaccinated individuals were 17.66 times (95% CI, 1.91 to 163.55; p=.011) avoid public spaces, gatherings, or crowds than non-vaccinated respondents. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 58 Table 10 Binary logistic regression of avoiding public spaces B Ethnic Group (White 0.470 group = referent) Religious Affiliation 0.059 (No = referent) Were you born in Canada? -0.157 (Yes = referent) Vaccination Status Vaccinated 1.558 Partially Vaccinated 2.871 Not Vaccinated (referent) In general, would you say your health is: Excellent -1.250 Very Good -1.261 Good -1.335 Fair -1.002 Poor (referent) Note. N=168. OR = odds ratio. 1 pvalue 0.287 1.600 95 CI for OR Lower Upper 0.674 3.795 0.026 1 0.872 1.061 0.517 2.174 0.477 0.109 1 0.741 0.854 0.335 2.176 0.459 1.136 11.54 6.393 1 1 0.001 0.011 4.751 17.66 1.933 1.907 11.68 163.55 1.287 1.219 1.210 1.282 0.943 1.070 1.217 0.610 1 1 1 1 0.331 0.301 0.270 0.435 0.287 0.283 0.263 0.367 0.023 0.026 0.025 0.030 3.570 3.091 2.820 4.535 S.E. Wald df 0.441 1.136 0.366 OR For avoiding eating at restaurants, a total of 173 responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 70 respondents reported ‘Yes’ to avoiding eating at restaurants with 54 respondents reporting ‘No’. For the Ethnically diverse group, 34 respondents reported ‘Yes’ to avoiding eating at restaurants with 15 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results showed that the model explained 9.9% (Nagelkerke R²) of variance and correctly classified 60.9% of cases. The results of Table 11 showed that the Ethnically diverse group was 1.69 times (95% CI, .767 to 3.7; p=.194) more likely to avoid large crowds. However, when income and age were included in the model, the ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 59 odds ratio decreased to a value of 1.41 (95% CI, .525 to 3.81) and a larger p-value of .494 (N=136). Table 11 Binomial logistic regression of avoiding eating at restaurants B Ethnic Group (White 0.522 group = referent) Religious Affiliation 0.152 (No = referent) Were you born in Canada? -0.084 (Yes = referent) Vaccination Status Vaccinated 0.334 Partially Vaccinated 1.265 Not Vaccinated (referent) In general, would you say your health is: Excellent -20.717 Very Good -20.879 Good -20.629 Fair -20.514 Poor (referent) Note. N=169. OR = odds ratio. 1 pvalue 0.194 1.685 95 CI for OR Lower Upper 0.767 3.699 0.202 1 0.653 1.164 0.600 2.257 0.441 0.036 1 0.849 0.919 0.387 2.182 0.437 0.899 0.582 1.981 1 1 0.446 0.159 1.396 3.544 0.592 0.608 3.289 20.650 1 1 1 1 0.999 0.999 0.999 0.999 S.E. Wald df 0.401 1.689 0.338 14831.21 14831.21 14831.21 14831.21 OR For wearing masks, a total of 173 responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 107 respondents reported ‘Yes’ to wearing a mask in public with 17 respondents reporting ‘No’. For the Ethnically diverse group, 46 respondents reported ‘Yes’ to avoiding large crowds with 3 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results showed that the model explained 43.9% (Nagelkerke R²) of variance and correctly classified 89.3% of cases. The results of Table 12 showed that the Ethnically diverse group was 1.13 times (95% CI, .199 to 6.38; p=.893) more likely to wear a ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 60 mask. However, when income and age were included in the model, the results were opposite with the White group showing an odds ratio value of 2.42 (95% CI, .206 to 28.57) with a p-value of .482 (N=136). For covariates, fully vaccinated respondents were 30.04 times (95 CI, 6.26 to 14416; p=<.001) more likely to wear a mask than those who were not vaccinated and those with a religious belief were 3.93 (95 CI, 1.08 to 14.27; p=.038) times more likely to wear a mask than those without a religious belief. Table 12 Binary logistic regression of wearing masks B Ethnic Group (White 0.119 group = referent) Religious Affiliation 1.368 (No = referent) Were you born in Canada? 0.623 (Yes = referent) Vaccination Status Vaccinated 3.402 Partially Vaccinated 1.013 Not Vaccinated (referent) In general, would you say your health is: Excellent -0.021 Very Good 0.363 Good 1.721 Fair -0.883 Poor (referent) Note. N=169. OR = odds ratio. *<.001 S.E. Wald df p-value OR 0.885 0.018 1 0.893 1.127 95 CI for OR Lower Upper 0.199 6.378 0.658 4.322 1 0.038 3.928 1.081 14.269 1.239 0.253 1 0.615 1.864 0.164 21.148 0.800 1.016 18.08 0.994 1 1 *0.001 0.319 30.04 2.753 6.259 0.376 144.16 20.16 1.377 1.268 1.317 1.396 0.000 0.082 1.708 0.400 1 1 1 1 0.988 0.775 0.191 0.527 0.979 1.437 5.588 0.414 0.066 0.120 0.423 0.027 14.566 17.239 73.773 6.380 Research question #2: Ethnic Group Differences in Perceived Exposure Risk Rapid review results To answer the question about ethnic differences in perceived exposure risk, a total of six articles assessed ethnicity in relation to perceived exposure risk to COVID-19. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 61 Based on these studies, the results of ethnic differences in perceived exposure risk to COVID-19 appear to be mixed. Some studies report that the Black ethnic groups have a higher perception of risk towards COVID-19 compared to the White ethnic group (Kumar & Encinosa, 2021; Orom et al., 2021). In contrast, the articles by Reiter and Katz (2021) and Garfin et al. (2021) suggest that non-Latinx Blacks have a lower perceived susceptibility to COVID-19 when compared to White ethnic groups. Lastly, both articles by Breakwell et al. (2021) and Nino et al. (2021) suggest that Black and White ethnic groups do not differ in their perceived exposure risk to COVID-19. Quantitative survey results The data collected for the CORAS was retrieved in a Likert scale format with respondents being able to respond to the question between 1 to 5. After respondents completed the CORAS section, I summed the score from each CORAS question to create a total score (continuous variable). A complete CORAS score ranged from 5-25, thus scores less than 5 were deemed as missing cases. A total of 175 responses were retrieved for the CORAS score (White group: 126; Ethnically diverse group: 49), but after listwise deletion, 171 valid responses were included in the analysis. A multivariable linear regression analysis was conducted to understand the differences in CORAS scores between White and Ethnically diverse groups. Both the White group and Ethnically diverse groups showed normal distribution through visual inspection of a normal probability plot. However, in the White group, two cases were deemed as outliers (CORAS score = 25). The analysis was run with and without the outliers and there was minimal difference between the results, therefore the outliers were retained in the data. Included in the analysis were the covariates religious beliefs, born in Canada, overall health status, and vaccination status. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 62 Sensitivity analysis was also conducted with the income and age variables. There was independence of residuals, as assessed by a Durbin-Watson statistic of 2.06 Homoscedasticity was confirmed via visual inspection of a scatterplot of standardized residuals versus standardized predicted values. Residuals were normally distributed as assessed by visual inspection of a normal probability plot. The linear regression results indicate that the ethnic group variable, along with the covariates, accounted for 18.6% of the variance in the CORAS score with an adjusted R² of 14%, a small effect size (Ferguson, 2016). Refer to Table 13 for ANOVA scores which show that Ethnic groups, along with the covariates, significantly predicted CORAS score, F(9,161)=4.09, p=<.001. The coefficients of the linear regression (refer to Table 14 for coefficient values for the independent variables) show an intercept of 14.27 (SE=1.924). Being part of the Ethnically diverse group would predict an increase of .248 (p-value=.751) in the CORAS score. Sensitivity analysis of the CORAS scores, including the age and income variables as covariates, also predicts CORAS score, F(12,123)=2.38, p=.009. Regarding the coefficients, the intercept increases to 15.54 (SE=2.695), with the Ethnically diverse group predicting an increase of .244 (p-value=.801). As such, the inclusion of age and income variables result in a similar outcome. Table 13 ANOVA results for linear regression between CORAS (dependent variable) and ethnic groups Sum of Squares df Mean Square F p-value Regression 671.450 9 74.606 <.001* Residual 2939.404 161 18.257 Total 3610.854 170 Note. *less than .001 4.09 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 63 Table 14 Linear regression coefficients for CORAS Constant Ethnic group (White group = referent) Religious affiliation Yes No (referent) Were you born in Canada? Yes No (referent) Vaccination status Fully Vaccinated Partially Vaccinated (only 1 of 2 doses) Not Vaccinated (referent) Overall Health Status Excellent Very Good Good Fair Poor (referent) Note. N=171. SE=Standard Error Unstandardized Coefficients B SE 14.27 1.924 0.248 0.780 Standardized Coefficients Beta t p-value 0.024 7.420 0.318 0.000 0.751 -0.571 0.683 -0.061 -0.835 0.405 -0.285 0.857 -0.025 -0.333 0.740 1.856 2.673 0.875 1.506 0.178 0.149 2.121 1.775 0.035 0.078 -3.357 -2.953 -0.500 1.791 1.899 1.778 1.740 1.924 -0.249 -0.294 -0.053 0.128 -1.768 -1.661 -0.288 0.931 0.079 0.099 0.774 0.353 Research Question #3: Perceived Exposure Risk Explains Variation in Self-reported Protective Behaviours Rapid Review Results Only one of the included articles addressed the relationship between perceived exposure risk and self-protective behaviours. Breakwell et al. (2021) suggested that increased perception of risk towards COVID-19 results in more compliance with COVID-19 protective behaviours (CPB). Based on their research, they found that perceived risk towards COVID-19 also correlates with fear of the virus, therefore increased perceived risk of COVID-19 leads to increased fear and subsequently more compliance with self-protective behaviours (Breakwell et ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 64 al., 2021). There are various factors which can influence one’s perceived risk towards COVID19, such as trust in science and political figures. Breakwell et al. (2021) found that groups that have a higher trust in science have an increased perceived risk towards COVID-19 which led to more engagement with self-protective behaviours. Furthermore, trust in science is mediated by in-group trust which is further mediated by trust in political figures (Breakwell et al., 2021). Although perceived exposure risk may explain variations in self-reported behaviours, Breakwell et al. (2021) suggest that further research is required on the topic as there may be differences within the BAME group which can alter the results. Quantitative Survey Results A binomial logistic regression was conducted between CORAS scores and each of the self-protective behaviours. For each binomial logistic regression, the covariates income, overall health status, age, born in Canada, and vaccination status were included with each analysis. For hand-hygiene, a total of 173 valid responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 100 respondents reported ‘Yes’ to engaging in hand-hygiene with 24 respondents reporting ‘No’. For the Ethnically diverse group, 46 respondents reported ‘Yes’ to engaging in hand-hygiene with 3 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The model explained 21.6% (Nagelkerke R²) of variance in reporting hand-hygiene practices and correctly classified 87.7% of cases. Hosmer and Lemeshow test indicated a good fit of the model (p=.141). The results in Table 15 showed that increases in CORAS scores increased the respondents were 1.12 times (95% CI, 992 to 1.26; p=.068) more likely to report washing their hands with soap or hand sanitizer several times a day. However, by ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 65 including the income and age variables, the odds ratio increased to 1.19 (95% CI, 1 to 1.41, pvalue=.044). Table 15 Binary logistic regression of hand-hygiene including CORAS B SE Ethnic Group (White 1.336 0.797 group = referent) Religious Affiliation 1.537 0.524 (No = referent) Were you born in 0.498 0.821 Canada? (Yes = referent) Vaccination Status Vaccinated 0.638 0.590 Partially vaccinated 1.144 1.218 Not vaccinated (referent) In general, would you say your health is: Excellent -0.533 1.482 Very good -0.038 1.404 Good -0.070 1.383 Fair -0.295 1.506 Poor (referent) CORAS 0.111 0.061 Note. N=169. SE=Standard Error. OR=Odds Ratio. 1 pvalue 0.093 3.805 95 C.I. for OR Lower Upper 0.799 18.128 8.586 1 0.003 4.650 1.663 12.999 0.368 1 0.544 1.646 0.329 8.232 1.169 0.882 1 1 0.280 0.348 1.893 3.138 0.595 0.289 6.020 34.130 0.129 0.001 0.003 0.038 1 1 1 1 0.719 0.978 0.960 0.845 0.587 0.963 0.932 0.744 0.032 0.061 0.062 0.039 10.716 15.078 14.020 14.261 3.335 1 0.068 1.118 0.992 1.26 Wald df 2.815 OR For avoiding high-risk individuals, a total of 173 responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 73 respondents reported ‘Yes’ to “Avoided high-risk individuals” with 51 respondents reporting ‘No’. For the Ethnically diverse group, 41 respondents reported ‘Yes’ to avoiding high-risk individuals with 8 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results showed that the model explained 16.7% (Nagelkerke R²) of variance and correctly classified 70.4% of cases. The Hosmer and ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 66 Lemeshow test indicated a good fit of the model (p=.999). The results of Table 16 showed that as CORAS scores increased, respondents were 1.06 times (95% CI, .976 to 1.15; p=.341) more likely to avoid high-risk individuals. After including the income and age variables for sensitivity analysis, the odds ratio decreased to 1.04 (95% CI, .944 to 1.14, p-value=.451). Table 16 Binary logistic regression of avoiding high-risk individuals including CORAS B SE Ethnic Group (White 0.961 0.462 group = referent) Religious Affiliation 0.167 0.360 (No = referent) Were you born in -0.234 0.480 Canada? (Yes = referent) Vaccination Status Vaccinated 1.238 0.457 Partially vaccinated 0.825 0.780 Not vaccinated (referent) In general, would you say your health is: Excellent -0.311 1.051 Very good 0.073 0.999 Good -0.504 0.964 Fair -0.499 1.058 Poor (referent) CORAS 0.059 0.042 Note. N=169. SE=Standard Error. OR=Odds Ratio. Wald df Sig. OR 4.335 1 0.037 2.615 95 C.I. for OR Lower Upper 1.058 6.464 0.214 1 0.644 1.181 0.583 2.394 0.239 1 0.625 0.791 0.309 2.025 7.340 1.118 1 1 0.007 0.290 3.448 2.282 1.408 0.494 8.443 10.530 0.088 0.005 0.273 0.223 1 1 1 1 0.767 0.942 0.601 0.637 0.733 1.076 0.604 0.607 0.093 0.152 0.091 0.076 5.749 7.626 3.995 4.830 1.955 1 0.162 1.061 0.976 1.153 For avoiding public spaces, gatherings, or crowds, a total of 172 responses were recorded (White group: 123 total responses; Ethnically diverse group: 49 total responses). In the White group, 76 respondents reported ‘Yes’ to avoiding public spaces, gatherings, or crowds with 47 respondents reporting ‘No’. For the Ethnically diverse group, 38 respondents reported ‘Yes’ to avoiding public spaces, gatherings, or crowds with 11 respondents reporting ‘No’. After listwise ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 67 deletion, a total of 168 valid cases were included in the binomial logistic regression. The results of the binomial logistic regression showed that the model explained 18.9% (Nagelkerke R²) of variance and correctly classified 72.6% of cases. Hosmer and Lemeshow test indicated a good fit of the model (p=.647). The results of Table 17 showed that as CORAS scores increased, respondents were 1.07 times (95% CI, .981 to 1.17; p=.125) more likely to avoid public spaces, gatherings, or crowds. After including the income and age variables for sensitivity analysis, the odds ratio increased to 1.08 (95% CI, .972 to 1.21, p-value=.141). Table 17 Binary logistic regression of avoiding public spaces including CORAS B SE Ethnic Group (White 0.467 0.444 group = referent) Religious Affiliation 0.093 0.371 (No = referent) Were you born in -0.131 0.480 Canada? (Yes = referent) Vaccination Status Vaccinated 1.417 0.471 Partially vaccinated 2.766 1.147 Not vaccinated (referent) In general, would you say your health is: Excellent -1.070 1.342 Very good -1.133 1.274 Good -1.381 1.262 Fair -1.184 1.338 Poor (referent) CORAS 0.069 0.045 Note. N=168. SE=Standard Error. OR=Odds Ratio. 1 pvalue 0.293 1.595 95 C.I. for OR Lower Upper 0.669 3.805 0.063 1 0.802 1.098 0.531 2.271 0.074 1 0.785 0.878 0.343 2.247 9.074 5.819 1 1 0.003 0.016 4.126 15.897 1.641 1.680 10.377 150.457 0.635 0.791 1.197 0.783 1 1 1 1 0.426 0.374 0.274 0.376 0.343 0.322 0.251 0.306 0.025 0.026 0.021 0.022 4.764 3.913 2.983 4.216 2.355 1 0.125 1.072 0.981 1.171 Wald df 1.108 OR For avoiding eating at restaurants, a total of 173 responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 70 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 68 respondents reported ‘Yes’ to avoiding eating at restaurants with 54 respondents reporting ‘No’. For the Ethnically diverse group, 34 respondents reported ‘Yes’ to avoiding eating at restaurants with 15 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results showed that the model explained 11.4% (Nagelkerke R²) of variance and correctly classified 63.3% of cases. The Hosmer and Lemeshow test indicated a good fit of the model (p=.968). The results in Table 18 showed that as CORAS scores increased respondents were 1.06 times (95% CI, .976 to 1.15; p=.168) more likely to avoid large crowds. After including the income and age variables for sensitivity analysis, the odds ratio decreased to 1.02 (95% CI, .932 to 1.13, p-value=.616). Table 18 Binary logistic regression of avoiding eating at restaurants including CORAS B SE Ethnic Group (White 0.522 0.403 group = referent) Religious Affiliation 0.192 0.342 (No = referent) Were you born in -0.077 0.444 Canada? (Yes = referent) Vaccination Status Vaccinated 0.213 0.449 Partially vaccinated 1.151 0.909 Not vaccinated (referent) In general, would you say your health is: Excellent -20.590 14605.81 Very good -20.777 14605.81 Good -20.671 14605.81 Fair -20.674 14605.81 Poor (referent) CORAS 0.057 0.041 Note. N=169. SE=Standard Error. OR=Odds Ratio. Wald df Sig. OR 1.679 1 0.195 1.686 95 C.I. for OR Lower Upper 0.765 3.713 0.316 1 0.574 1.212 0.620 2.369 0.030 1 0.862 0.926 0.388 2.210 0.224 1.605 1 1 0.636 0.205 1.237 3.162 0.513 0.533 2.985 18.766 1 1 1 1 0.999 0.999 0.999 0.999 1 0.168 1.059 0.976 1.148 1.9 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 69 For wearing masks, a total of 173 responses were recorded (White group: 124 total responses; Ethnically diverse group: 49 total responses). In the White group, 107 respondents reported ‘Yes’ to avoiding large crowds with 17 respondents reporting ‘No’. For the Ethnically diverse group, 46 respondents reported ‘Yes’ to avoiding large crowds with 3 respondents reporting ‘No’. After listwise deletion, a total of 169 valid cases were included in the binomial logistic regression. The results showed that the model explained 47.5% (Nagelkerke R²) of variance and correctly classified 91.1% of cases. The Hosmer and Lemeshow test indicated a good fit of the model (p=.989). The results in Table 19 showed that CORAS scores increased respondents were 1.15 times (95% CI, .997 to 1.36; p=.054) more likely to wear a mask. After including the income and age variables for sensitivity analysis, the odds ratio increased to 1.19 (95% CI, .942 to 1.33, p-value=.201). ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 70 Table 19 Binary logistic regression of wearing masks including CORAS B SE Wald Ethnic Group (White -0.155 0.928 0.028 group = referent) Religious Affiliation 1.271 0.670 3.595 (No = referent) Were you born in 0.942 1.335 0.498 Canada? (Yes = referent) Vaccination Status Vaccinated 3.275 0.835 15.367 Partially vaccinated 0.633 1.026 0.380 Not vaccinated (referent) In general, would you say your health is: Excellent 0.351 1.524 0.053 Very good 0.942 1.457 0.418 Good 1.954 1.461 1.789 Fair -1.259 1.541 0.667 Poor (referent) CORAS 0.136 0.071 3.701 Note. N=169. SE=Standard Error. OR=Odds Ratio. df Sig. OR 1 0.868 0.857 95 C.I. for OR Lower Upper 0.139 5.282 1 0.058 3.563 0.958 13.251 1 0.480 2.566 0.187 35.150 1 1 0.000 0.538 26.431 1.882 5.141 0.252 135.881 14.073 1 1 1 1 0.818 0.518 0.181 0.414 1.421 2.566 7.059 0.284 0.072 0.147 0.403 0.014 28.185 44.637 123.706 5.820 1 0.054 1.146 0.997 1.317 Research question #4: Self-Reported Behaviours are Explained by Ethnic Group Differences in Perceived Risk Rapid Review Results Regarding self-protective behaviours explained by ethnic differences in perceived risk, the articles by Orom et al. (2021) and Breakwell et al. (2021) are the only articles to specifically address the topic. Orom et al. (2021) suggest that Black ethnic groups engaged in more protective behaviours compared to White ethnic groups due to increased fear of infecting others as well as the importance of protecting distal others. Orom et al. (2021) suggest that ethnic difference in perceived risk towards COVID-19 may influence the self-protective behaviours that ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 71 an individual may engage in. They believe that this perception of risk is mediated by the level of control to mitigate the spread of the virus within one’s community. The White group, for example, may have a greater sense of control in their community to prevent the spread of COVID-19 and thus will perceive the risk of COVID-19 to be lower. For Black and Latinx communities, the sense of control is much lower, thus the perception of risk is higher and leads to more engagement with self-protective behaviours. Breakwell et al. (2021) show similar results with the BAME group showing higher levels of perceived risk leading to increased engagement in self-protective behaviours. Breakwell et al. (2021) acknowledge that their results may not be statistically significant, though further research to divulge on the topic would be valuable. Quantitative Survey Results In my analysis, when controlled for confounders, ethnic differences did not show a direct effect on self-protective behaviours; therefore, ethnic differences in perceived risk did not have a mediating factor on self-protective behaviours. Conclusion For my research, I initially conducted a rapid review and subsequently conducted a logistic and linear regression of my data. Using my research questions to guide my research, I used the results of the rapid review to help inform the results of my data analysis. I observed ethnic groups demonstrating self-protective behaviours more than non-ethnic groups. In terms of self-perceived risk, ethnic groups and non-ethnic groups show similar perceptions of risk towards COVID-19. I also found that perception of risk may influence one’s desire to engage in self-protective behaviours against COVID-19; however, I observed this only through the rapid review and not through my data analysis. Lastly, self-protective behaviours may be mediated by ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS ethnic differences in perceived risk as per my rapid review, but I did not observe a mediating effect of this within my own data. 72 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 73 Chapter 5: Discussion COVID-19 is a global pandemic that has affected the lives of many individuals across the world. Within Canada, the response by the government to control and mitigate the spread of the virus was challenging with the implementation of lockdowns and mandates such as wearing masks. Adherence to COVID-19 protective guidelines by the public was an important step in reducing the spread of COVID-19; however, not all members of the public adhered to these guidelines. According to Chum et al. (2021), in 2020, some members of the public negatively perceive and did not follow Ontario’s COVID-19 prevention strategies such as mask mandates and closure of businesses, even as cases continued to rise. Although there may be a plethora of reasons why individuals did not adhere to these guidelines, my thesis focused on a gap in research revolving around the role of ethnic diversity on self-protective behaviours against COVID-19. Relationship Between Ethnic Diversity and Self-Reported Behaviours The results of my research synthesis showed that there is a potential correlation between Ethnically diverse groups and the likelihood of engagement in self-protective behaviours against COVID-19. Some studies from the research synthesis showed that BAME groups were more likely to engage in self-protective behaviours such as mask wearing in public and social distancing. The results of my survey appear to suggest similar results. The Ethnically diverse group was more likely to engage in all measures of self-protective behaviours: hand-hygiene practices, avoiding high-risk individuals, avoiding public spaces, avoiding eating at restaurants, and wearing masks in comparison to the White group. All of the variables for self-protective behaviours, except for avoiding high-risk individuals with a relatively low p-value, were similar between the White group and the Ethnically diverse group. The similarities between the groups ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 74 may be related to the low sample size and the disproportionate number of White group respondents (N=127) compared to the Ethnically diverse group (N=49). In our survey data, the White group consisted of 72% of the total sample population. According to Rusticus and Lovato (2014), unequal sample sizes decrease the power of the sample size and can lead to a decreased ability to detect effects. I believe that the unequal sample size could be attributed to the online platforms utilized to administer the survey (i.e., Mturk and Facebook); according to Boas et al. (2020), there is a higher percentage (73%) of White users on these platforms in comparison to other ethnic groups. Being able to recruit enough participants into the Ethnically diverse groups from these platforms was challenging due to the time sensitivity of the project. For future research, more equitable study groups may be able to result in more significant findings. The results of my survey data, which suggests Ethnically diverse groups are more likely to engage in hand-hygiene, avoid high-risk individuals, avoid eating at restaurants, and avoid large crowds or public spaces, appear to be supported by the results of the rapid review. For the utilization of masks, however, I was surprised to note the results between the two groups were quite similar; in comparison, the results from the rapid review suggest that non-White groups were more likely to wear masks than White groups. I believe that the difference between the results of the rapid review and my survey data is related to the country-specific mask mandates that were implemented, or not implemented, during the pandemic. The rapid review consisted primarily of studies based in the United States or United Kingdom; none of the studies were based in Canada. The results of our survey data showed that likelihood of the White group and Ethnically diverse group in wearing masks in public is similar; however, based on the timing of the survey administration (August 2021-November 2021), all provinces had an active mask mandate in place, thus all citizens, regardless of ethnicity, were required to wear masks in public ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 75 or indoors (Canadian Institute of Health Information, 2022). In the United States, however, only 39 states implemented a mask mandate, typically lasting less than 200 days, between mid-2020 through early-2021. Only six states had an active mask mandate in place past June 1, 2022 (Ballotpedia, 2022). There is a difference between Canada and the United States regarding mask use with the population in Canada utilizing masks earlier and more consistently than the population in the United States (Jang et al., 2021), which may, in part, explain the differences between the rapid review and my survey data. Beyond ethnic diversity, I observed other factors that may play a role in determining whether participants engage in self-protective behaviours. Those who reported having a religious belief were more likely to report engaging in hand-hygiene and wearing masks in comparison to those who do not have a religious belief. These results are also supported by the research conducted by Franz and Dhanani (2021) whereby they found that increased religiousness in participants from the United States led to a higher perception of risk and engagement in selfprotective behaviours against COVID-19. The increased perception of risk and engagement with self-protective behaviours may be due to social traits typically exhibited by religious individuals; those with religious beliefs are more likely to listen or acknowledge advice from religious leaders or doctors than those who do not have a religious belief (Franz & Dhanani, 2021). Although my research does not focus on the role of religion and self-protective behaviours against COVID-19, it is worth noting that this topic may be important to conduct future research on, especially within a Canadian context. Vaccination status also played a significant role in the results, whereby fully vaccinated individuals were more likely to avoid high-risk individuals, avoid large gatherings, and wear a mask. This result was expected as increased perceived risk leads to a higher intention to be ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 76 vaccination (Al-Amer et al., 2022). It is important to acknowledge that the survey data were collected during a time when only two doses of the vaccine were considered effective for COVID-19. Considering the existence of other COVID-19 VOIs, specifically the Omicron VOI, booster doses (beyond the second dose) became recommended by public health officials across the world (Chenchula et al., 2022). I recognize that the progression of the COVID-19 pandemic, along with the development and updates related to vaccines, VOIs, and public health guidelines, could have affected the results of the data significantly. Using a cross-sectional survey, I could only collect data based on a certain point in time. However, a longitudinal survey for this type of research could be able to capture the rapid changes that occurred during the pandemic (Polit & Beck, 2017). Gender can also affect the results of self-protective behaviours with females 1.24 times more likely than males to have a higher perceived threat towards their individual health from COVID-19 (Niño et al., 2021). In the study by Niño et al. (2021), the proportion of females to males was 50 whereas in my survey data the proportion of females to males is 38. Although the descriptive analysis showed no significant differences between the groups, and gender was not included in the final model analysis, it is worth noting the comparative difference in numbers between males and females within my survey data for future research. I observed that ethnic diversity on its own may not be sufficient to determine whether individuals will engage in selfprotective behaviours. I believe that it is crucial to define and understand the term “Ethnic diversity” and determine how broad the term is to be used. For example, my research focused more on the racial backgrounds of respondents, whereby Dinesen et al. (2020) consider Ethnic diversity to “entail a shared language, religion, nationality, and phenotype” (p. 443). The result of my research shows that, although Ethnic diversity, within solely the definition of racial ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 77 background or racial origin, may in part explain differences in engagement with self-protective behaviours, other factors such as religion may also influence these behaviours. This coincides with the research by Perry et al. (2020) where they found that religious commitment led to increased compliance with COVID-19 mitigation strategies such as handwashing or avoiding large crowds. Social Determinants of Health Although my research synthesis and data analysis may suggest ethnic differences in selfprotective behaviours, there may be factors that affect these differences that extend beyond ethnic diversity. Ethnically diverse groups may be more inclined to perform self-protective behaviours compared to non-ethnically diverse groups because they experience a higher risk of contracting COVID-19 due to the racialized barriers they encounter. This is further supported by the research by Orom et al. (2021) where they found that White individuals perceived lower risk towards COVID-19 compared to BAME groups as they found a greater sense of control to prevent the spread of the virus within their community. This sense of security involving these communities may be attributed to greater access to health care, smaller population densities, and higher income (Orom et al., 2021). According to Turner-Musa et al. (2020), Black, Latinx, and Native Americans in the United States have a higher risk of susceptibility and mortality from COVID-19 than Whites due to economic and social barriers that disproportionately affect minority groups. These factors include, but are not limited to, housing situations, access to health care, income, and education (Turner-Musa et al., 2020). The disproportionate number of cases affecting non-White ethnic groups is visible within a Canadian context as well where a higher concentration of confirmed COVID-19 cases can be observed in lower-income, higher-housing density, and higher ethnically diverse communities throughout Canada (Xia et al., 2022). ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 78 Although lower income, housing density, and ethnicity are separate disadvantages they are inherently dependent on one another with race being a mediating factor (Block et al., 2019). Despite Canada being a multi-cultural country, racialized groups (non-White) continue to make significantly less income than non-racialized groups (Block et al., 2019). Furthermore, racialized groups are more likely to work in essential jobs that require them to work on-site as opposed to working from home, thus increasing the likelihood of contracting COVID-19 from the workplace (Bonacini et al., 2021; Raifman & Raifman, 2020). Ethnic groups receiving less income are more likely to live in multi-family homes which can increase the spread of the virus among multiple family members (Baena-Diéz et al., 2020; Hou et al., 2020). Although my research focuses primarily on ethnic diversity, I recognize the presence of racial disparities within perception of risk and self-protective behaviours. I believe that future research is needed on the topic to better understand its role on COVID-19. Health Beliefs Model The conceptual model that was used to guide my research was the HBM. Using this model, I believed that, as participants perceived a higher threat of getting COVID-19, I would see a subsequent increase in engagement with self-protective behaviours. In my research synthesis, Breakwell et al. (2021) observed this in their study, suggesting that as the perceived risk towards COVID-19 increases, participation in self-protective behaviours also increased. This also aligns with the survey by Bruin de bruin and Bennet (2020) who found that participants with increased fear towards COVID-19 were more likely to implement self-protective behaviours. For my research data, I did not observe a significant increase in protective behaviours as CORAS increased. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 79 Using the HBM to inform my findings, I anticipated seeing an increase in self-protective behaviours against the virus as the perception of risk towards COVID-19 increased. Although my research synthesis results suggested an increase in protective behaviours as the perception of risk increased, I observed in the survey data analysis that self-protective behaviours did not increase as the reported CORAS increased. With these results, I recognized that perception of risk alone cannot determine whether an individual engages in a self-protective behaviour or not; other factors need to be considered within the HBM in order to predict whether these behaviours are acted upon. The other constructs from the HBM (perceived benefits, perceived barriers, cues to action, and perceived susceptibility) should be taken into consideration as suggested by Champion and Skinner (2008). To better understand the likelihood that an individual will engage in self-reported behaviours, using the HBM as guidance, I need to consider the environmental situation in which they live in and their personal cognitive perception of the disease. For example, an individual who faces racialized sociodemographic barriers such as low income may not be able to avoid high-risk individuals if they live in a multi-family home. Despite having a higher perception of risk towards COVID-19, the perception of barriers (e.g., unable to afford a single-family home) that prevents someone from engaging in self-protective behaviours may outweigh the benefits (e.g., being able to social distance or isolate from close family members). Alternatively, an individual who does has a lower perception of risk towards COVID-19 may still wear a mask in public as the perception of benefit (e.g., being able to go out in public) outweighs the barrier (e.g., staying home). Health Beliefs Model and Social Determinants of Health The constructs within the HBM can be heavily influenced by the social determinants of health. Although race and the biological heredity can play a significant role in the overall health ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 80 of an individual, there are various other social and economic barriers that marginalized racial groups experience that can affect the constructs of the HBM. Poor access to healthcare services for BAME groups, for example, can increase the perception of risk towards COVID-19 (Breakwell et al., 2021) but the perceived barriers that come alongside low-income communities, such as lower availability of work-from-home jobs or food insecurity (Raifman & Raifman, 2020), may prevent this population from engaging in self-protective behaviours (Rosenstock, 1974). The combination of the HBM and the Social Determinants of Health may explain why the results of my data analysis show similar CORAS scores between White and Ethnically diverse groups but differences in reported self-protective behaviours. The White group, despite reporting the similar CORAS scores to the Ethnically diverse group, are less likely to engage in self-protective behaviours. This may be explained by sociodemographic factors that provide the White population a greater sense of security against COVID-19. The White population in Canada has typically reported higher income, greater access to healthcare, and decreased housing density compared to non-White ethnic groups (Sundaram et al., 2021). Sundaram et al. (2021) report that visible minority groups are at higher risk of contracting COVID-19 due to greater housing densities, immigration status, and low-income service jobs that provide lower protection from the virus. By understanding and mitigating the systemic racial barriers that exist in Canada, I may be able to improve the perception of COVID-19 or other potential exposures experienced by ethnically diverse groups and increase the engagement of self-protective behaviours. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 81 Future Implications Future Research in Canada The development and spread of COVID-19 across the world have highlighted existing racial disparities that have led to the higher confirmed cases in ethnically diverse communities. In my study, I observed that ethnically diverse groups were more likely to engage in various selfprotective behaviours against COVID-19. However, despite this, there are more confirmed cases within ethnically diverse communities in Ontario (Xia et al., 2022). The discrepancy in confirmed cases between ethnically diverse communities and non-ethnically diverse communities is too great to ignore and poses the question as to why this event is occurring when the data suggests that we should, theoretically, see fewer cases in these communities. These observations from my study suggest the need to better understand the social determinants of health and the barriers encountered by ethnically diverse groups in Canada. I used the HBM as a guiding principle in my study and I discussed how the constructs within the HBM may be affected by the social determinants of health. Currently, there is minimal research available that describes any association between the HBM and the social determinants of health. I believe that future research may be appropriate to better understand if there are relationships between the HBM and the social determinants of health. Even though my study was challenged by limitation in sample size and representation of ethnic groups, we did observe differences between the Ethnically diverse and White group. These findings are suggestive of the need to further understand ways to encourage self-protective behaviours during pandemic situations. Although the focus of our study revolved around ethnic diversity, the influence of religion or sociodemographic factors that influenced the results of our study cannot be ignored. Respondents with religious beliefs were more likely to wash their hands ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 82 or wear a mask in public. In Canada, there is a gap in research to identify the role of religion and its effect on self-protective behaviours; it may be prudent for future researchers to acknowledge this gap and seek opportunities to address it. In summary, we conducted this research in the context of COVID-19, but future research on ethnic diversity and self-protective behaviours may be applicable for mitigation strategies in future pandemics. Nursing and Cultural Humility The premise of my study revolved around observing any differences in the engagement of self-protective behaviours against COVID-19 between ethnically diverse populations and nonethnically diverse populations. While we observed differences between both groups, the findings of my study also suggest the need to further explore how healthcare providers, or nurses, can better approach individuals of an ethnically diverse population to protect themselves from COVID-19 or future diseases. A common term utilized in healthcare circles is “cultural competence” where healthcare providers attend to the culturally diverse backgrounds of patients or individuals (Lekas et al., 2020); this definition, however, is not enough to address the systemic barriers experienced by those of an ethnically diverse population. It is more important to recognize a shift towards cultural humility, as opposed to competence, as a means to look past the biases or barriers experienced by this population and more effectively address their needs. Cultural humility is the concept of nursing self-reflection to understand the holistic cultural needs of a patient or research participant (Yeager & Bauer-Wu, 2013). It consists of being able to understand the bias and cultural distinctions of the self and respecting these differences of others as opposed to judgement and stereotyping. Cultural humility requires reflexivity of one’s own biases, beliefs, and values that come from personal experiences; it is ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 83 important for nurses to identify these barriers before trying to understand another person’s points of view (Yeager & Baurer-Wu, 2013). When attempting to work with other individuals, either as patients or research participants, it is vital for the nurse to understand that person’s beliefs, culture, values, and life experiences that have shaped them into the person they are today. For example, we may identify certain ethnic groups to have specific characteristics or traits that we believe are vital to their livelihood. However, with cultural humility, it is important for us to know that not all individuals of a particular ethnic group will inherit these features. It is stereotypical to assume that all individuals of a particular culture or ethnicity are the same and are to be treated the same. This is the difference between cultural humility and cultural competence whereby the latter assumes a complete cultural understanding of an individual and attending to the needs of these cultural differences. Cultural humility is exhibited when we look beyond these cultural assumptions and understand the holistic aspect of the person; this includes the culture, race, personal beliefs, or life experiences (Lekas et al., 2020). Familial relationships, for example, play an important role for certain ethnic groups. To illustrate, in the Filipino culture the elderly members of the family are typically taken care of by their children within the same household. A nurse exemplifies cultural humility by understanding that an individual from this population may not exhibit the same values as described; not all individuals in the Filipino population will have strong familial relationships. It would be inappropriate for a nurse to encourage an individual to communicate with their family which he or she is estranged from; it is important to understand the person’s past and experiences to perform an appropriate action. Cultural humility, in the context of COVID-19, will be an important aspect of nursing care particularly regarding suggestions or actions given to patients or families to keep them safe from the virus. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 84 When advising patients or families of actions to reduce the risk of contracting COVID19, it is important for the nurse to understand the various values of that individual or family. These not only include the ethnic or cultural differences, but the unique experiences of that individual which may be in contrast of the typical characteristics of their culture or ethnicity (Lekas, 2020; Yeager & Baurer-Wu, 2013). The results of our survey data showed that people of an ethnically diverse population are more likely to engage in self-protective behaviours. However, it would be inappropriate to assume that all individuals in this population engage in self-protective behaviours or that individuals from the White population do not engage in selfprotective behaviours. There are various factors that can influence one’s desire to engage in selfprotective behaviours; these may include religious beliefs, personal beliefs about the potential risk, or even environmental factors (i.e., employment or housing). Scientific Quality and Limitations Representativeness There are potential limitations with my study. The study lacked representation from various demographic groups such as race, females, and educational status. Having little representation from certain demographics groups can affect the generalizability of the research’s findings with the target population. I had initially intended to conduct quota sampling with my sample population whereby I would achieve at least 55 separate respondents in the following groups: Black, South-Asian, and East-Asian ethnic groups. I was unable to achieve the target quota and thus consolidated the ethnic groups into a single group for data analysis. For gender, there were more males (61.9) than females (38.1) in the sample population. Additionally, one respondent reported to be “Non-binary”; which we included in the “Male” category (more males in the sample population than females) after sensitivity analysis showed no effect on the results. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 85 However, it categorized this individual into a binary when they did not identify as such. I did not ask respondents to indicate their educational status in my survey; this variable could have been a potential confounder in my analysis. Lastly, due to time constraints, my survey was only available to respondents who were able to read English. The data from my research was collected using an online survey administered through MTurk and Facebook. Online surveys can be easy, quick, and affordable methods of collecting data from the target population; however, certain population groups can be under-represented using these methods. Only respondents who had access to internet can participate in this study; this may lead to an under-representation of the lower-income population or people in rural areas that do not have access to internet (Boas et al., 2020). Furthermore, only respondents who are registered users of MTurk or Facebook were able to access the online survey. Additionally, the majority of users on MTurk and Facebook are of the White ethnic group (Boas et al., 2020). This made recruitment of respondents in the Ethnically diverse group challenging as there is a smaller population of non-White users on these online platforms. As a result, for my sample groups, I needed to consolidate all non-White respondents into a single group: Ethnically diverse group. Although we initially endeavoured to collect recruit respondents into separate ethnic groups, the consolidation of the non-White respondents into a single group was a necessary step in the research process, even at the detriment of data loss due to over-generalization. An underrepresentation of the sample population affects the generalizability of the outcomes to the target population (Polit & Beck, 2017). It is important for future research to consider that consolidating ethnic groups into a single category (e.g., not differentiating East Asians from South-East Asians) can lead to overgeneralization of the data and loss of vital information. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 86 My data collection methods are also subject to self-selection bias. This relates to a respondent choosing to participate in the survey rather than being selected or randomly chosen. COVID-19 has been a polarizing subject with certain groups of the population having extreme viewpoints regarding the pandemic. Some groups may be against the mitigation strategies against COVID-19 while some other groups may be strong advocates for them. This can lead to extreme responses to the questions of the survey and produce outliers. The article by Schaurer and Weib (2020) suggests caution with online non-probability sampling methods related to COVID-19 as their results showed extreme opinions for and against COVID-19 self-protective behaviours. Sample Size I initially calculated that a sample size of 330 would be required for my survey. Despite using two data collection methods, I was only able to achieve 187 respondents for my survey. I was unable to extend the survey longer due to the cost in making the survey available online and time constraints related to the thesis. The low sample size from my research can affect the reliability of the results and can lead to false-positives, or over-estimate the magnitude of the results (Hackshaw, 2008). Lower reliability, or less consistency, can decrease the strength of the research (Polit & Beck, 2017). Rapid Review Rapid reviews are a quick and affordable method of conducting a synthesis of research and studies (Dobbins, 2017). However, due to the continuous changes occurring with the COVID-19 pandemic, a lot of research studies were conducted without publication (i.e., grey literature). For my rapid review, one of my inclusion criteria was scholarly peer-reviewed articles, thus grey literature was not included. As described by Kousha et al. (2022), grey ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 87 literature, especially in the context of COVID-19, is a valuable source of information for researchers. A lot of credible information is available on non-academic websites such as organizations or news articles; this information is missed if grey literature is excluded in research (Kousha et al., 2022). Researchers should strongly consider the inclusion of grey literature in the development of future research, especially when studying a rapidly changing pandemic such as COVID-19. Measures One of the challenges with the measures of the study revolved around the appropriateness and robustness of the survey questions. For the questions in the survey, all the questions allowed participants to select “Prefer not to answer” or “Do not know”. The absence of responses led to listwise deletion during the binomial logistic regression analyses of the self-protective behaviours. We compensated for this by conducting a sensitivity analysis in the binomial logistic regression and the linear regressions of the survey data. However, inclusion of the variables incomes and age in the final analysis model could have improved the strength and generalization of the findings. Another challenge revolved around the robustness and appropriateness of the questions. For self-protective behaviours, the questions were adapted from the UAS (USC, 2020) but the survey did not ask other questions about practicing social distancing, cleaning or sanitizing the living environment, or being able to, or choosing to, work from home. These aspects of selfprotective behaviours are also recommended from Government of Canada (2021a) and the absence of these specific questions could have been beneficial for the study. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 88 Chapter 6: Conclusion and Recommendations COVID-19 has been a significant challenge for countries all around the world. It has impacted the lives of many people, affecting not only the economy, but also the social dynamics that are associated with human interaction. Transmission of COVID-19 is primarily through contact or droplet secretions and, as a result, governments and health organizations, such as the WHO, have provided guidelines to reduce the spread of COVID-19 in the community. These protective guidelines included changes in behaviours such as washing hands and social distancing; these have been shown to be effective in mitigating the spread of COVID-19 in the community. Unfortunately, many communities in Ontario have experienced high rates of COVID-19 infections despite the implementation of these guidelines. Of these communities, the most ethnically diverse communities were attributed with the highest rate of confirmed COVID19 cases in the province. For this research, I explored if there were ethnic differences in selfreported protective behaviours against COVID-19 and, if so, was this mediated by differences in perceived risk. I sought to address my research questions by conducting an online survey administered via the online platforms MTurk and Facebook. Using a rapid review of current research to help interpret the results of my data analysis, I observed that Ethnically diverse groups are more likely to perform certain self-protective behaviours than the White ethnic group despite similar perceptions of risk. This aligned with the results of my rapid review which suggests that Ethnically diverse groups are more likely to engage in self-protective behaviours. However, I did not observe, from my survey data, that these behaviours are mediated by ethnic differences in perception of risk towards COVID-19. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 89 Although my analysis did not reveal evidence of mediation from ethnic differences in perceived risk towards COVID-19, the results show that self-protective behaviours may be associated with other sociodemographic factors such as income, religious beliefs, age, and vaccination status. I believe that these confounders correlate with the constructs of the HBM whereby systemic barriers, which Ethnically diverse groups encounter more than non-Ethnically diverse groups, may influence their engagement in self-protective behaviours. Systemic barriers such as access to healthcare, living in multi-family homes, income, and working conditions are factors which can influence an individual’s ability to engage in self-protective behaviours. Despite the limitations and barriers to my study, I believe that my research can be used to identify research gaps and opportunities to address ethnic-diversity-related barriers during a pandemic. For the current, and future, pandemics, there are systemic barriers that need to be addressed in the ethnically diverse population in order to reduce the likelihood of contracting a disease or virus. With further research regarding the relationship between the HBM and the social determinants of health we may be able to better understand their potential relationships and how we can mitigate the risks encountered by ethnically diverse populations. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 90 References Al‐Amer, R., Maneze, D., Everett, B., Montayre, J., Villarosa, A. R., Dwekat, E., & Salamonson, Y. (2022). COVID‐19 vaccination intention in the first year of the pandemic: A systematic review. Journal of Clinical Nursing, 31(2), 62–86. https://doi.org/10.1111/jocn.15951 Amazon. (2021). FAQs. Amazon Mechanical Turk. https://www.mturk.com/help Angrisani, M., Finley, B., Kapteyn, A. (2019). Can internet match high-quality traditional surveys? Comparing the Health and Retirement Study and its Online Version. The Econometrics of Complex Survey Data, 39, 3–33. https://doi.org/10.1108/S0731905320190000039001 Azambuja, R. (2015). Conducting surveys on MTurk. https://www.ufrgs.br/gpmc/wpcontent/uploads/2015/11/Conducting-Surveys-on-MTurk.pdf Baena-Diéz, J. M., Barroso, M., Cordeiro-Coelho, S. I., Diáz, J. L., & Grau, M. (2020). Impact of COVID-19 outbreak by income: Hitting hardest the most deprived. Journal of Public Health (United Kingdom), 42(4), 698–703. https://doi.org/10.1093/pubmed/fdaa136 Ballotpedia. (2022). State-level mask requirements in response to the coronavirus (COVID-19) pandemic, 2020-2022. https://ballotpedia.org/Statelevel_mask_requirements_in_response_to_the_coronavirus_(COVID19)_pandemic,_2020-2022 Barrett, C., & Cheung, K. L. (2021). Knowledge, socio-cognitive perceptions and the practice of hand-hygiene and social distancing during the COVID-19 pandemic: A cross-sectional study of UK university students. BMC Public Health, 21(1), 1–18. https://doi.org/10.1186/s12889-021-10461-0b ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 91 Block, S., Galabuzi-Grace, E., & Tranjan, R. (2019). Canada's colour coded income inequality. Canadian Centre for Policy Alternatives. https://policyalternatives.ca/sites/default/files/uploads/publications/National20Office/2019/ 12/Canada's20Colour20Coded20Income20Inequality.pdf Boas, T. C., Christenson, D. P., & Glick, D. M. (2020). Recruiting large online samples in the United States and India: Facebook, Mechanical Turk, and Qualtrics. Political Science Research and Methods, 8(2), 232–250. https://doi.org/10.1017/psrm.2018.28 Bonacini, L., Gallo, G., & Scicchitano, S. (2021). Working from home and income inequality: Risks of a ‘new normal’ with COVID-19. Journal of Population Economics, 34(1), 303– 360. https://doi.org/10.1007/s00148-020-00800-7 Box, G. E. P., & Tidwell, P. W. (1962). Transformation of the independent variables. Technometrics, 4, 531–550. https://doi.org/10.2307/1266288 Breakwell, G. M., Fino, E., & Jaspal, R. (2021). COVID-19 preventive behaviours in White British and Black, Asian and Minority Ethnic (BAME) people in the UK. Journal of Health Psychology, 27(6), 1–17. https://doi.org/10.1177/13591053211017208 Bruine de Bruin, W., & Bennett, D. (2020). Relationships between initial COVID-19 risk perceptions and protective health behaviours: A national survey. American Journal of Preventive Medicine, 59(2), 157–167. https://doi.org/10.1016/j.amepre.2020.05.001 Canadian Agency for Drugs and Technologies in Health. (2020). CADTH COVID-19 Search Strings. https://covid.cadth.ca/literature-searching-tools/cadth-covid-19-search-strings/ Canadian Institute of Health Information. (2022, June 9). COVID-19 Intervention Timeline in Canada. https://www.cihi.ca/en/covid-19-intervention-timeline-in-canada ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 92 Casson, R. J., & Farmer, L. D. (2014). Understanding and checking the assumptions of linear regression: A primer for medical researchers. Clinical & Experimental Ophthalmology, 42(6), 590–596. https://doi.org/10.1111/ceo.12358 Center for Disease Control and Prevention. (2022, April 22). COVID-19 treatments and medications. https://www.cdc.gov/coronavirus/2019-ncov/your-health/treatments-forsevere-illness.html Champion, V., & Skinner, C. S. (2008). “The health belief model.” In K. Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior and health education (4th ed., pp. 45–65). JosseyBass. Chenchula, S., Karunakaran, P., Sharma, S., & Chavan, M. (2022). Current evidence on efficacy of COVID‐19 booster dose vaccination against the Omicron variant: A systematic review. Journal of Medical Virology, 94(7), 2969–2976. https://doi.org/10.1002/jmv.27697 Chum, A., Nielsen, A., Bellows, Z., Farrell, E., Durette, P. N., Banda, J. M., & Cupchik, G. (2021). Changes in public response associated with various COVID-19 restrictions in Ontario, Canada: Observational infoveillance study using social media time series data. Journal of Medical Internet Research, 23(8), e28716. https://doi.org/10.2196/28716 Clavel, N., Fellow, P., Badr, J., Gautier, L., & Lavoie-Tremblay, M. (2021). Risk perceptions, knowledge and behaviours of general and high-risk adult populations towards COVID-19: A systematic scoping review. Public Health Reviews, 42, 1–12. https://doi.org/10.1101/2021.02.09.21250257 Collaboration for Environmental Evidence. (2018). Guidelines and Standards for Evidence synthesis in Environmental Management. https://environmentalevidence.org/informationfor-authors/ ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 93 Cori, L., Bianchi, F., Cadum, E., & Anthonj, C. (2020). Risk perception and COVID-19. International Journal of Environmental Research and Public Health, 17(9), 3114. https://doi.org/10.3390/ijerph17093114 Dinesen, P. T., Schaeffer, M., & Sonderskov, K. M. (2020). Ethnic diversity and social trust: A narrative and meta-analytical review. Annual Review of Political Science, 23, 441–465. https://doi.org/10.1146/annurev-polisci-052918-020708 Dobbins, M. (2017). Rapid review guidebook: Steps for conducting a rapid review. https://www.nccmt.ca/uploads/media/media/0001/01/a816af720e4d587e13da6bb307df8c9 07a5dff9a.pdf. EBSCO. (2022). CINAHL database. https://www.ebsco.com/products/research-databases/cinahldatabase Erdemandi, M., & Leach, J. (2021). ‘Masks Don’t Work:’ Ideology associations and the geospatial propagation of COVID-19 disinformation on Twitter. SSRN. https://dx.doi.org/10.2139/ssrn.3898097 Ferguson, C. J. (2016). An effect size primer: A guide for clinicians and researchers. Professional Psychology: Research and Practice, 40(5), 532–538. https://psycnet.apa.org/doi/10.1037/a0015808 Fernandes, Q., Inchakalody, V. P., Merhi, M., Mestiri, S., Taib, N., Moustafa Abo El-Ella, D., Takwa, B., Raza, A., Al-Zaidan, L., Mohsen, M. O., Ali Yousef Al-Nesf, M., Ait Hssain, A., Yassine, H. M., Bachmann, M. F., Uddin, S., & Dermime, S. (2022). Emerging COVID-19 variants and their impact on SARS-CoV-2 diagnosis, therapeutics and vaccines. Annals of Medicine, 54(1), 524–540. https://doi.org/10.1080/07853890.2022.2031274 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 94 Flannelly, K. J., Flannelly, L. T., & Jankowski, K. R. (2018). Threats to the internal validity of experimental and quasi-experimental research in healthcare. Journal of Health Care Chaplaincy, 24(3), 107-130. https://doi.org/10.1080/08854726.2017.1421019 Franz, B., & Dhanani, L. Y. (2021). Beyond political affiliation: An examination of the relationships between social factors and perceptions of and responses to COVID19. Journal of Behavioural Medicine, 44(5), 641–652. https://doi.org/10.1007/s10865-02100226-w Government of Canada. (2020). Social determinants of health and health inequalities. https://www.canada.ca/en/public-health/services/health-promotion/population-health/whatdetermines-health.html Government of Canada. (2021a). Coronavirus disease (COVID-19): Prevention and risks. https://www.canada.ca/en/public-health/services/diseases/2019-novel-coronavirusinfection/prevention-risks.html Government of Canada. (2021b). COVID-19 vaccination in Canada. https://healthinfobase.canada.ca/covid-19/vaccination-coverage/ Government of Canada. (2022). COVID-19 daily epidemiology update. https://healthinfobase.canada.ca/covid-19/?stat=rate&measure=cases_total&map=pt#a2 Guan, J., Wei, Y., Zhao, Y., & Chen, F. (2020). Modeling the transmission dynamics of COVID19 epidemic: A systematic review. Journal of Biomedical Research, 34(6), 422–430. https://doi.org/10.7555/JBR.34.20200119 Hackshaw, A. (2008). Small studies: Strengths and limitations. European Respiratory Journal, 32(5), 1141–1143. DOI: 10.1183/09031936.00136408 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 95 Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication monographs, 76(4), 408–420. https://doi.org/10.1080/03637750903310360 Hearne, B. N., & Niño, M. D. (2021). Understanding how race, ethnicity, and gender shape mask-wearing adherence during the COVID-19 pandemic: Evidence from the COVID impact survey. Journal of Racial and Ethnic Health Disparities, 9(1), 176–183. https://doi.org/10.1007/s40615-020-00941-1 Heinze, G., Wallisch, C., & Dunkler, D. (2018). Variable selection–a review and recommendations for the practicing statistician. Biometrical Journal, 60(3), 431–449. https://doi.org/10.1002/bimj.201700067 Hong, Q. N., Fàbregues, S., Bartlett, G., Boardman, F., Cargo, M., Dagenais, P., Gagnon, M.-P., Griffiths, F., Nicolau, B., O’Cathain, A., Rousseau, M.-C., Vedal, I., & Pluye, P. (2018). The Mixed Methods Appraisal Tool (MMAT) version 2018 for information professionals and researchers. Education for Information, 34(4), 285–291. DOI: 10.3233/EFI-180221 Hosmer, D. W. J., Lemeshow, S., & Sturdivant, R. X. (2013). “Model‐building strategies and methods for logistic regression.” In D. W. J. Hosmer, S. Lemeshow, & R. X. Sturdivant (Eds.), Applied logistic regression (pp. 89–151). Wiley. https://doi.org/doi:10.1002/9781118548387.ch4 Hou, F., Frank, K., & Schimmele, C. M. (2020, July 6). Economic impact of COVID-19 among visible minority groups. https://www.researchgate.net/publication/342747625 Houle, R., & Statistics Canada. (2020). Changes in the socioeconomic situation of Canada’s Black population, 2001 to 2016. https://www150.statcan.gc.ca/n1/pub/89-657-x/89-657x2020001-eng.htm ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 96 Jang, H., Rempel, E., Roth, D., Carenini, G., & Janjua, N. Z. (2021). Tracking COVID-19 discourse on twitter in North America: Infodemiology study using topic modeling and aspect-based sentiment analysis. Journal of Medical Internet Research, 23(2), e25431. doi: 10.2196/25431 Jaspal, R., Fino, E., & Breakwell, G. M. (2020). The COVID-19 Own Risk Appraisal Scale (CORAS): Development and validation in two samples from the United Kingdom. Journal of Health Psychology, 27(4), 790–804. https://doi.org/10.1177/1359105320967429 Jones, C. J., Smith, H., & Llewellyn, C. (2014). Evaluating the effectiveness of health belief model interventions in improving adherence: A systematic review. Health Psychology Review, 8(3), 253–269. https://doi.org/10.1080/17437199.2013.802623 Kan, T., & Zhang, J. (2018). Factors influencing seasonal influenza vaccination behaviour among elderly people: A systematic review. Public Health, 156, 67–78. https://doi.org/10.1016/j.puhe.2017.12.007 Kousha, K., Thelwall, M., & Bickley, M. (2022). The high scholarly value of grey literature before and during Covid-19. Scientometrics, 127, 3489–3504. https://doi.org/10.1007/s11192-022-04398-3 Kumar, V., & Encinosa, W. (2021). Racial disparities in the perceived risk of COVID-19 and in getting needed medical care. Journal of Racial and Ethnic Health Disparities, 1–10. https://doi.org/10.1007/s40615-021-01191-5 Lai, C. C., Shih, T. P., Ko, W. C., Tang, H. J., & Hsueh, P. R. (2020). Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and coronavirus disease-2019 (COVID-19): The epidemic and the challenges. International Journal of Antimicrobial Agents, 55(3), 1–9. https://doi.org/10.1016/j.ijantimicag.2020.105924 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 97 Lee, J.W. (2021). COVID-19 Preventive action intention determinants-focused on the expanded Health Belief Model. Annals of the Romanian Society of Cell Biology, 25(1), 1425–1432. https://www.annalsofrscb.ro/index.php/journal/article/view/257 Lekas, H. M., Pahl, K., & Fuller Lewis, C. (2020). Rethinking cultural competence: Shifting to cultural humility. Health Services Insights, 13. https://doi.org/10.1177%2F1178632920970580 Ma, Q., Liu, J., Liu, Q., Kang, L., Liu, R., Jing, W., Wu, Y., & Liu, M. (2021). Global percentage of asymptomatic SARS-CoV-2 infections among the tested population and individuals with confirmed COVID-19 diagnosis: A systematic review and meta-analysis. JAMA Network Open, 4(12), 1–18. DOI:10.1001/jamanetworkopen.2021.37257 Mackey, K., Ayers, C. K., Kondo, K. K., Saha, S., Advani, S. M., Young, S., Spencer, H., Rusek, M., Anderson, J., Veazie, S., Smith, M., & Kansagara, D. (2021). Racial and ethnic disparities in COVID-19–related infections, hospitalizations, and deaths: A systematic review. Annals of Internal Medicine, 174(3), 362–373. https://doi.org/10.7326/M20-6306 Martin-Martin, A., Orduña-Malea, E., Harzing, A. W., & López-Cózar, E. D. (2017). Can we use Google Scholar to identify highly-cited documents? Journal of Informetrics, 11(1), 152163. https://doi.org/10.1016/j.joi.2016.11.008 Munn, Z., Peters, M. D., Stern, C., Tufanaru, C., McArthur, A., & Aromataris, E. (2018). Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Medical Research Methodology, 18(1), 1–7. https://doi.org/10.1186/s12874-018-0611-x National Library of Medicine. (2022, May 2). MEDLINE: Overview. Medline. https://www.nlm.nih.gov/medline/medline_overview.html ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 98 Niño, M., Harris, C., Drawve, G., & Fitzpatrick, K. M. (2021). Race and ethnicity, gender, and age on perceived threats and fear of COVID-19: Evidence from two national data sources. SSM – Population Health, 13, 1–8. https://doi.org/10.1016/j.ssmph.2020.100717 Orom, H., Allard, N. C., Kiviniemi, M. T., Hay, J. L., Waters, E. A., Schofield, E., Thomas. S. N., & Tuman, M. (2021). Racial/ethnic differences in prosocial beliefs and prevention behaviour during the COVID-19 pandemic. Journal of Racial and Ethnic Health Disparities, 9, 1807–1817. https://doi.org/10.1007/s40615-021-01117-1 Page, M. J., Moher, D., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … McKenzie, J. E. (2021). PRISMA 2020 explanation and elaboration: Updated guidance and exemplars for reporting systematic reviews. BMJ, 372, 160. https://doi.org/10.1136/bmj.n160 Perry, S. L., Whitehead, A. L., & Grubbs, J. B. (2020). Culture wars and COVID-19 conduct: Christian nationalism, religiosity, and Americans’ behaviour during the coronavirus pandemic. Journal for the Scientific Study of Religion, 59(3), 405–416. https://doi.org/10.1111/jssr.12677 Pinto, A. D., Glattstein-Young, G., Mohamed, A., Bloch, G., Leung, F. H., & Glazier, R. H. (2016). Building a foundation to reduce health inequities: Routine collection of sociodemographic data in primary care. Journal of the American Board of Family Medicine, 29(3), 348–355. doi: 10.3122/jabfm.2016.03.150280 Platt, L., & Warwick, R. (2020). Are some ethnic groups more vulnerable to COVID-19 than others? Institute for Fiscal Studies. https://www.ifs.org.uk/inequality/wp- ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 99 content/uploads/2020/04/Are-some-ethnic-groups-more-vulnerable-to-COVID-19-thanothers-V2-IFS-Briefing-Note.pdf Polit, D., & Beck, C. T. (2017). Nursing research: Generating and assessing evidence for nursing practice (10th ed). Wolters Kluwer. Popay, J., Roberts, H., Sowden, A., Petticrew, M., Arai, L., Rodgers, M., Britten, N., Roen, K., & Duffy, S. (2006). Guidance on the conduct of narrative synthesis in systematic reviews. Academic, 1(1), 1-92. http://dx.doi.org/10.13140/2.1.1018.4643 Public Health Ontario. (2021). COVID-19 in Ontario – A focus on diversity: February 26, 2020 to December 13, 2021. https://www.publichealthontario.ca//media/documents/ncov/epi/2020/06/covid-19-epi-diversity.htm Raifman, M., & Raifman, J. (2020). Disparities in the population at risk of severe illness from COVID-19 by race/ethnicity and income. American Journal of Preventive Medicine, 59(1), 137–139. https://doi.org/10.1016/j.amepre.2020.04.003 Reiter, P. L., & Katz, M. L. (2021). Racial/Ethnic differences in knowledge, attitudes, and beliefs about COVID-19 among adults in the United States. Frontiers in Public Health, 9. https://doi.org/10.3389/fpubh.2021.653498 Revilla, M., & Ochoa, C. (2017). Ideal and maximum length for a web survey. International Journal of Market Research, 59(5), 557–565. https://doi.org/10.2501/IJMR-2017-039 Rosenstock, I. M. (1974). Historical origins of the Health Belief Model. Health Education Monographs, 2(4), 328–335. https://doi.org/10.1177/109019817400200403 Ross, P. T., Hart-Johnson, T., Santen, S. A., & Zaidi, N. L. B. (2020). Considerations for using race and ethnicity as quantitative variables in medical education research. Perspectives on Medical Education, 9(5), 318–323. https://doi.org/10.1007/s40037-020-00602-3 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 100 Rusticus, S. A., & Lovato, C. Y. (2014). Impact of sample size and variability on the power and type I error rates of equivalence tests: A simulation study. Practical Assessment, Research, and Evaluation, 19(1), 11. https://doi.org/10.7275/4s9m-4e81 Santé Montréal. (July 13, 2022). Inequality during the pandemic: Racialized populations. https://santemontreal.qc.ca/en/public/coronavirus-covid-19/situation-of-the-coronaviruscovid-19-in-montreal/survey-of-the-health/racialized-populations/ Schaurer, I., & Weib, B. (2020, June). Investigating selection bias of online surveys on coronavirus-related behavioural outcomes. Survey Research Methods 14(2), 103–108. https://doi.org/10.1007/s40037-020-00602-3 Shrestha, N. (2020). Detecting multicollinearity in regression analysis. American Journal of Applied Mathematics and Statistics, 8(2), 39–42. DOI: 10.12691/ajams-8-2-1 Smith, L. E., Mottershaw, A. L., Egan, M., Waller, J., Marteau, T. M., & Rubin, G. J. (2020). The impact of believing you have had COVID-19 on self-reported behaviour: Crosssectional survey. PloS ONE 15(1), e0248076. https://doi.org/10.1371/journal.pone.0240399 Statistics Canada. (2021). Dictionary, census of population, 2021. https://www12.statcan.gc.ca/census-recensement/2021/ref/dict/az/indexeng.cfm?T=18#results Sundaram, M. E., Calzavara, A., Mishra, S., Kustra, R., Chan, A. K., Hamilton, M. A., Djebli, M., Rosella, L. C., Watson, T., Chen, H., Chen, B, Baral, S. D., & Kwong, J. C. (2021). The individual and social determinants of COVID-19 diagnosis in Ontario, Canada: A population-wide study. Canadian Medical Association Journal, 193(20). https://doi.org/10.1101/2020.11.09.20223792 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 101 Turner-Musa, J., Ajayi, O., & Kemp, L. (2020). Examining social determinants of health, stigma, and COVID-19 disparities. Healthcare, (8)2. https://doi.org/10.3390/healthcare8020168 University of Southern California. (2020). Understanding America Study. https://uasdata.usc.edu/index.php Webber, R. (2020). COVID-19 and race: Protecting data or saving lives? International Journal of Market Research, 62(5), 528–537. https://doi.org/10.1177/1470785320946589 WHO Solidarity Trial Consortium. (2022). Remdesivir and three other drugs for hospitalised patients with COVID-19: Final results of the WHO Solidarity randomised trial and updated meta-analyses. The Lancet, 399(10339), 1941–1953. https://doi.org/10.1016/S01406736(22)00519-0Wise, T., Zbozinek, T. D., Michelini, G., Hagan, C. C., & Mobbs, D. (2020). Changes in risk perception and self-reported protective behaviour during the first week of the COVID-19 pandemic in the United States. Royal Society Open Science, 7(9). https://doi.org/10.31234/osf.io/dz428 Wu, J., Tang, B., Bragazzi, N. L., Nah, K., & McCarthy, Z. (2020). Quantifying the role of social distancing, personal protection and case detection in mitigating COVID-19 outbreak in Ontario, Canada. Journal of Mathematics in Industry, 10(1). https://doi.org/10.1186/s13362-020-00083-3 World Health Organization (2022). WHO coronavirus (COVID-19) dashboard. https://covid19.who.int/ Xia, Y., Ma, H., Moloney, G., García, H. A. V., Sirski, M., Janjua, N. Z., ... & Maheu-Giroux, M. (2022). Geographic concentration of SARS-CoV-2 cases by social determinants of health in metropolitan areas in Canada: a cross-sectional study. CMAJ, 194(6), 195-204. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIOURS 102 Yeager, K. A., & Bauer-Wu, S. (2013). Cultural humility: Essential foundation for clinical researchers. Applied Nursing Research, 26(4), 251–256. https://doi.org/10.1016/j.apnr.2013.06.008 Yildirim, M., Gecer, E., & Akgul, O. (2021). The impacts of vulnerability, perceived risk, and fear on preventive behaviours against COVID-19. Psychology, Health and Medicine, 26(1), 35–43. https://doi.org/10.1080/13548506.2020.1776891 Zhao, E., Wu, Q., Crimmins, E., & Ailshire, J. (2020). Media trust and infection mitigating behaviours during the COVID-19 pandemic in the USA. BMJ Global Health, 5(10), 1–10. http://dx.doi.org/10.1136/ bmjgh-2020-003323 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 103 Appendix A Review Matrix – Quantitative Authors, Title, Journal Year Published Purpose (of author’s research) Dependent Independent # of Subjects Subject characteristics Sample Design Source or Instrument Year Data Collected Conclusions Li, S, Feng, B, Liao, W., & Pan, W. Internet Use, Risk Awareness , and Demograp hic Characteris tics Associated with Engageme nt in Preventive Behaviours and Testing: CrossSectional Survey on COVID-19 in the United States 2020 To examine how internet use, risk awareness, and demographic characteristics are associated with engagement in preventive behaviours and testing during the COVID-19 pandemic in the United States Engagement in preventive behaviours and testing Internet use; Risk Awareness; Demographics (sex, age, ethnicity, income, education level, marital status, and employment status 979 Gender Female (): 47.6 Online Crosssectional surveys Participants retrieved from Amazon Mechanical Turk (crowdsour cing website). April 1014 2020 Compared with whites, African Americans and Asians more frequently wore a facemask in public (OR 1.81, 95 CI 1.26-2.59, P<.001; OR 2.47, 95 CI 1.65-3.69, P<.001, respectively) and stayed home (OR 1.88, 95 CI 1.28-2.77, P=.001; OR 2.23, 95 CI 1.47-3.37, P<.001, respectively). In addition, compared with whites, Asians covered their noses and mouths when sneezing and coughing more often (OR 1.78, 95 CI 1.132.80, P=.01), and kept social distance more often (OR 1.61, 95 CI 1.04-2.48, P=.03). African Americans reported more frequent effort in cleaning frequently touched surfaces than whites (OR 2.00, 95 CI 1.362.94, P<.001). Age (mean): 36.94 years Marital Status Single (): 45.9 Employment status Working ():81.0 Race () White: 69.2 Black American: 11.4 Hispanic/Latin American: 4.7 Asian or AsianAmerican: 9.8 Other: 4.9 Household income (US $) () <10,000: 4.6 10,001-20,000: 7.8 20,001-40,000: 18.7 40,001-60,000: 22.7 60,001-80,000: 20.5 80,001-100,000: 10.8 100,001-120,000: 6.5 >120,000: 8.4 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS Smith, L. E., Mottersha w, A. L., Egan, M., Waller, J., Marteau, T. M., & Rubin, G. J.. The impact of believing you have had COVID-19 on selfreported behaviour: Crosssectional survey. 2020 To investigate whether people who think they have had COVID-19 are less likely to report engaging with lockdown measures compared with those who think they have not had COVID19. Self-reported adherence to social distancing measures (going out for essential shopping, nonessential shopping, and meeting up with friends/family ; total out-ofhome activity), worry about COVID-19 and perceived risk of COVID-19 to oneself and people in the UK Perceived immunity to COVID-19 6,149 104 Gender () Male: 47 Female 53 Online Crosssectional surveys Recruited from Predictiv’s research panel April 2022 2020 People who believed that they had had COVID-19 were: more likely to agree that they had some immunity to COVID-19; less likely to report adhering to lockdown measures; less worried about COVID-19; and less likely to know that cough and high temperature / fever are two of the most common symptoms of COVID-19. Online crosssectional survey The survey was prepared on the Qualtrics platform 2020 On univariate logistic regression analyses, NPI adherence was associated with a belief that NPIs would reduce personal risk Age () 18 to 24 years: 23 25 to 34 years: 20 35 to 44 years: 17 45 to 54 years: 12 55 years and over: 28 Have a child? () No: 42.7 Yes: 57.3 Employment Status () Not working: 33.7 Working: 66.3 Working in key sector () No: 62.7 Yes: 27.3 Highest educational or professional qualification () GCSE/vocational/Alevel/No formal qualifications: 72 Degreee or higher (bachelors, Masters, PhD: 28 Kantor, B. & Kantor, J. Nonpharmaceu tical 2020 Understand public attitudes and beliefs regarding various NPIs Performed in last week, frequency; Baseline characteristics (Social Distancing; Required by government to 1,005 Sex () Men 494 (48.8) Women 518 (51.2) Age, y, () 18–30 250 (24.5) ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS Interventio ns for Pandemic COVID19: A CrossSectional Investigati on of US General Public Beliefs, Attitudes, and Actions and selfreported adoption of NPIs, and to explore associations between NPI performance and the baseline characteristics of respondents. remain at home) and attitudes (Slows the spread of COVID-19, level of agreement) 105 31–40 204 (20.0) 41–50 146 (14.3) 51–60 198 (19.4) >60 222 (21.8) Education level () $150,000 44 (4.3) Location () and distributed to a representati ve US sample stratified by age, sex, and race, through Prolific Academic a platform for academic survey research. of developing COVID19 [OR 3.06, 95 CI [1.25, 7.48], p = 0.014] and with a belief that the NPIs were not difficult to perform [OR 1.79, 95 CI [1.38, 2.31], p < 0.0001]. Adherence was also associated with self-described religiosity [OR 1.85, 95 CI [1.42, 2.39], p < 0.0001]; full-time employment [OR 1.35, 95 CI [1.02, 1.78], p = 0.035]; worry regarding a family member contracting COVID-19 [OR 1.47, 95 CI [1.11, 1.93], p = 0.007]; and belief that the media was not exaggerating the severity of the pandemic [OR 1.44, 95 CI [1.09, 1.91], p = 0.012]. ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 106 Urban 743 (72.8) Rural 277 (27.2) Parsons Leigh, J., Fiest, K., BrundinMather, R., Plotnikoff, K., Soo, A., Sypes, E. E., ... & Stelfox, H. T. A national crosssectional survey of public 106ercepti on of the COVID-19 pandemic: Selfreported beliefs, knowledge , and behaviours . Hearne, B. & Nino, M. Understan ding How Race, Ethnicity, and Gender Shape 2020 2020 Examine perceptions, knowledge acquisition, and behaviours across Canada. (3) more comp rehen sivel y inves tigati ng the role Self-reported beliefs; Selfreported knowledge acquisition; Self-reported behaviours Demographic characteristics 1,996 See Table 1 of article Online crosssectional survey Administer ed through Ipsos throughout Canada 2020 The vast majority of respondents in my study reported practicing a high level of physical distancing, and a surprisingly high number felt that they could maintain this for a long period of time (6 months or more) if necessary. It is evident that most were motivated to limit social and physical interactions as a means to protect themselves and others from becoming infected with COVID19. Mask-wearing adherence Race/Ethnicity (White [reference], Black, Latino/a, and Asian/AsianAmerican 4,688 Race (mean) Black: 0.13 Latina/o: 0.18 Asian: 0.03 White: 0.67 Secondary analysis using pooled data from COVID Impact Survey (CIS) Pooled data from COVID Impact Survey (CIS) 2020 Findings demonstrate that all three historically marginalized groups (Black, Latina/o, and Asian) included in this study were more likely to report wearing a mask during this time period. Gender (mean) Male: 0.49 Age (mean) ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS MaskWearing Adherence During the COVID-19 Pandemic: Evidence from the COVID Impact Survey Atchison, C. J., Bowman, L., Vrinten, C., Redd, R., Pristera, P., Eaton, J. W., & Ward, H. Perception s and behavioura of race and ethni city on mask weari ng durin g the COV ID19 pandemic and (2) examining whether gender intersects with race and ethnicity to differently influence mask-wearing patterns. 2020 18-29: 0.14 30-44: 0.30 45-59: 0.23 60+: 0.33 Household Income (mean): <$40,000: 0.36 $40,000->$75,000: 0.29 $75,000-150,000+: 0.35 (3) adopt Risk ion of socia ldista ncing meas ures, (2) abilit y to work from home, and (3) 107 perceptions; preventive behaviours; willingness and ability to self-isolate Sociodemograp hic variables 2,108 Demographic and socio-economic Age (years) 18-24: 218 (11.1) 25-34: 294 (14.4) 35-44: 396 (19.3) 45-54: 355 (17.5) 55-69: 519 (24.2) 70 or above: 326 (13.5) Sex Male: 987 (48.0) Female: 1094 (50.7) Online crosssectional survey YouGov, market research company with a UK panel of 800,000+ individuals March 2020 Respondents from ethnic minority backgrounds perceived themselves to be less able to selfisolate than respondents from White backgrounds (84.8 vs. 92.1, aOR:0.47; 95 CI:0.27,0.82), although they were equally willing to do so ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS l responses of the general public during the COVID-19 pandemic: A crosssectional survey of UK Adults. willingness and (4) ability to self-isolate. 108 Prefer not to say: 27 (1.3) Ethnicity White: 1985 (93.9) Asian/Asian British: 48 (2.4) Black / African / Caribbean / Black British: 20 (1.0) Other ethnic group, including mixed/multiple ethnic groups: 39 (1.9) Prefer not to say: 16 (0.77) Marital status Married, civil partnership, or living as married: 1283 (60.3) Separated, divorced, or widowed: 270 (12.2) Never married: 545 (27.1) Prefer not to say: 10 (0.45) Education No formal qualification: 121 (5.5) Secondary-level qualification: 859 (42.1) Post-secondary-level, below bachelor: 148 (6.9) Bachelor-level or above: 664 (30.8) ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 109 Other technical, professional or higher qualification: 245 (11.2) Don’t Know: 32 (1.6) Prefer not to say: 39 (2.0) Household income <£20,000: 440 (20.7) £20,000 to £29,999: 355 (16.8) £30,000-£49,999: 472 (22.4) £50,000 and over: 429 (20.6) Don’t know: 103 (5.1) Prefer not to say: 309 (14.4) Yildirim, M& Akgul, O.The Impacts of Vulnerabili ty, Perceived Risk, and Fear on Preventive Behaviours Against COVID-19 2020 To unravel whether the age, gender, education level, fear, personal risk, and vulnerability acted as predictors of preventive behaviour among adults Preventive behaviours Fear, Perceived personal risk, vulnerability 4,536 Gender Female: 3,165 Male:1,371 Education Level High school or below:488 Bachelor’s degree: 2,873 Master’s degree: 937 Doctoral degree: 238 Marital Status Married: 1,765 Single: 2,637 Divorces/widowed: 134 Socioeconomic level Lower than average: 565 Crosssectional online survey Social networking sites MarchApril 2020 The results suggest that vulnerability, perceived risk, and fear can significantly increase engagement in preventive behaviours during the novel coronavirus pandemic ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 110 Average: 3,490 Higher than average: 481 Number of children None: 2,982 1: 537 2: 627 3: 288 3 or more: 102 Number of siblings 1 or none: 257 2: 1,378 3: 1,358 4: 760 5 or more: 783 Household members: 1: 241 2: 588 3: 1,018 4: 1,345 5 or more: 1,344 Wise, T., Zbozinek, T., Michelini, G., Hagan, C. & Mobbs, D. Changes in Risk Perception and Protective Behaviour During the First Week 2020 Understand how perceived risk to COVID-19 affects protective behaviours from the virus. Protective behaviours (avoiding social interactions, buying more food and water, staying home, buying sanitary products, washing hands, travelling less) perceptions of infection likelihood and severity 1591 See Wise et al. (2020) for demographic information Online survey Subjects were recruited through Prolific (13) between 3/11/20, the day when the WHO declared COVID-19 a pandemic, and 3/16/20 March 2020 higher perceived personal risk predicts engagement in protective behaviours such as hand washing and social distancing, as shown in studies of prior pandemics ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 111 of the COVID-19 Pandemic in the United States Bruine de Bruin, W. & Bennett, D. Relationshi ps Between Initial COVID-19 Risk Perception s and Protective Health Behaviours ”A National Survey 2020 Authors examined perceived risks for COVID-19 infection and infection fatality and whether these risk perceptions were associated with protective behaviours. Protective behaviours Risk of infection and infection fatality 6,684 See Bruine de Bruin & Bennet (2020) for demographic information Secondary Analysis Data retrieved from Understand ing America Scale (UAS) March 2020 People perceiving greater risks were more likely to implement protective behaviours— especially later (versus earlier) in March 2020 Review Matrix – Qualitative/Systematic Review Authors; Title; Year Denford, S., Morton, K. S., Lambert, H., Zhang, J., Smith, L. E., Rubin, J. G., ... & Methodo logy Method Phenomena of Interest Geographical Cultural Participants Descripti ve thematic Analysis Semistructured interviews via Zoom The aim of this study was to gain a better understanding of how people from lowincome and Black, Asian, United States Participants were between the ages of 18 and 65 years and from Black African and Black Caribbean (N = 4), Asian9 20 participants. BAME Background Group White background Data Analysis Authors Conclusion Individuals in BAME communities face more pressures and barriers to adhere to social distancing guidelines. Some of these barriers include low income Reviewers Comments ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS Yardley, L. Understan ding patterns of adherence to COVID19 mitigation measures: A qualitative interview study. 2020 Clavel, N., Badr, J., Gautier, L., & LavoieTremblay, M. Risk Perceptions, Knowledge and Behaviours of General and HighRisk Adult Populations towards COVID-19: A Systematic Scoping Review. 2021 and Minority Ethnic (BAME) communities are adhering to social distancing and self-isolation measures during the COVID-19 pandemic, and to explore in detail the reasons underpinning this behaviour. System atic Review Conducted a comprehensiv e search of the following electronic databases: MEDLINEOvid, EMBASEOvid, PsycINFOOvid, Web of Science, and CINAHL (EBSCO). T Map the early evidence on risk perceptions, knowledge, and behaviours of general and high-risk adult populations towards COVID-19. and White (N = 7) ethnic groups. The average (mean) Index of Multiple Deprivation decile was 4.15. Four participants reported that they had had COVID-19, or symptoms of COVID-19 in the household. Africa, Asia, Europe, and North America 112 or need to support others. Participants also justified breaking social distancing rules as they did not feel that they, or their household, were vulnerable. Finding suggests that people, especially those in less-affected areas, might have underestimated the infectivity of the virus. Perceived susceptibility combined with perceived severity plays a vital role in motivating health protection behaviours (53) and may facilitate or reduce the transmission of a virus during pandemics ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 113 Appendix B Literature Database Results Key word Ethnic Groups Perceived exposure risk to COVID Other ethnic groups Self-reported protective behaviours COVID-19 Pandemic Synonyms/Search Term Ethnicity Cultural background Cultural norms Race Racial Background Risk Perceived Risk Risk Perceptions Understanding of risk Attitudes Ethnicity Cultural background Cultural norms Race Racial Background Self-protective behaviours Reported protective behaviours COVID Coronavirus SARS CoV Severe Acute Respiratory Syndrome Combine search term results: Research after inclusion/exclusion criteria: MEDLINE 2,134,732 PubMed 2,160,024 CINAHL 365,092 3,195,529 3,181,952 1,302,345 2,134,732 2,160,024 365,092 2,957 274,878 1,510 119,012 122,916 43,207 9 0 123 5 3 0 Google Scholar 22 6 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 114 Appendix C Search Strings for Rapid Review Database Search Search # 1 Key Word TWU Oneseach COVID-19 COVID* OR CORONAVIRUS* OR SARS* CO* OR SEVERE ACUTE RESPIRATORY SYNDROME* OR "NCOV" OR "DELTA" OR "OMICRON" OR VARIANT* OR BETA* Medline 1. COVID-19/ or exp COVID19 Testing/ or COVID-19 Vaccines/ or SARS-CoV-2/ 2. (coronavirus/ or betacoronavirus/ or coronavirus infections/) and (disease outbreaks/ or epidemics/ or pandemics/) 3. (nCoV* or 2019nCoV or 19nCoV or COVID19* or COVID or SARS-COV-2 or SARSCOV-2 or SARS-COV2 or SARSCOV2 or SARS coronavirus 2 or Severe Acute Respiratory Syndrome Coronavirus 2 or Severe Acute Respiratory Syndrome Corona Virus 2).ti,ab,kf,nm,ot,ox,rx,px. 4. ((new or novel or “19” or “2019” or Wuhan or Hubei or China or Chinese) adj3 (coronavirus* or corona virus* or betacoronavirus* or CoV or HCoV)).ti,ab,kf,ot. 5. (longCOVID* or postCOVID* or postcoronavirus* or postSARS*).ti,ab,kf,ot. 6. ((coronavirus* or corona virus* or betacoronavirus*) adj3 (pandemic* or epidemic* or outbreak* or crisis)).ti,ab,kf,ot. 7. ((Wuhan or Hubei) adj5 pneumonia).ti,ab,kf,ot. 8. or/1-7 LitCOVID Sciencedirect COVID OR CORONAVIRUS 2 OR SARS-COV2 Journal of racial and ethnic health disparities ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS Search # Key Word 2 ETHNIC GROUPS 3 PERCEPTION OF RISK 4 5 SELFPROTECTIVE BEHAVIORS COMBINED TWU Onesearch ETHNIC* OR RACE* OR "RACIAL* BACKG*" OR CULTUR* OR "CULTURAL NORM*" OR "CULTUR* BACKGRO*" OR ETHNOGRAPH* OR "SOCIODEMOGRAPHIC* "PERCEIVED RISK*" OR RISK PERCEPTION* OR UNDERSTAND* RISK PROTECTIVE BEHAV* OR SOCIAL DISTANC* OR HAND WASH* OR ISOLAT* OR PREVENT* OR INFECTION CONTROL* OR INFECTION REDUC* OR QUARANTIN* OR NONPHARMACEUTICAL INTERVENTION* #1 AND #2 AND #3 AND #4 Medline ETHNIC* OR RACE* OR "RACIAL* BACKG*" OR CULTUR* OR "CULTURAL NORM*" OR "CULTUR* BACKGRO*" OR ETHNOGRAPH* OR "SOCIODEMOGRAPHIC* "PERCEIVED RISK*" OR RISK PERCEPTION* OR UNDERSTAND* RISK PROTECTIVE BEHAV* OR SOCIAL DISTANC* OR HAND WASH* OR ISOLAT* OR PREVENT* OR INFECTION CONTROL* OR INFECTION REDUC* OR QUARANTIN* OR NONPHARMACEUTICAL INTERVENTION* #1 AND #2 AND #3 AND #4 LitCOVID ("ETHNICITY" OR "RACE" OR "RACIAL BACKGROUND" OR CULTUR* OR "CULTURAL NORM" OR "CULTUR* BACKGRO* OR ETHNOGRAPH* OR "SOCIODEMOGRAPHIC*) AND ("PERCEIVED RISK*" OR RISK PERCEPTION* OR UNDERSTAND* RISK) AND ("PROTECTIVE BEHAVIOR" OR SOCIAL DISTANC* OR HAND WASH* OR ISOLAT* OR PREVENTION OR INFECTION CONTROL* OR INFECTION REDUCTION OR QUARANTINE ) 115 Sciencedirect Journal of racial and ethnic health disparities ETHNICITY OR RACE "ETHNICITY" OR "RACE" OR "RACIAL BACKGROUND" OR CULTUR* OR "CULTURAL NORM" OR "CULTUR* BACKGRO* OR ETHNOGRAPH* OR "SOCIODEMOGRAPHIC* PERCEIVED RISK OR PERCEPTION OF RISK "PERCEIVED RISK*" OR RISK PERCEPTION* OR UNDERSTAND* RISK NON PHARMACEUTICAL INTERVENTIONS OR PROTECTIVE BEHAVIORS "PROTECTIVE BEHAVIOR" OR SOCIAL DISTANC* OR HAND WASH* OR ISOLAT* OR PREVENTION OR INFECTION CONTROL* OR INFECTION REDUCTION OR QUARANTINE OR Non-pharmaceutical Interventions NOT Vaccine #1 AND #2 AND #3 AND #4 #2 AND #3 AND #4 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 116 Appendix D Mixed-Methods Appraisal Tool (MMAT) for Rapid Review First author Year Title S1. Are there clear research questions? S2. Do the collected data allow to address the research questions? Is the sampling strategy relevant to address the research question? Is the sample representative of the target population? Are the measurements appropriate? Is the risk of nonresponse bias low? Is the statistical analysis appropriate to answer the research question? Barret, C. 2021 Yes Yes Yes Yes Yes No Yes Breakwell , G. 2021 Yes Yes No Yes Yes No Yes Garfin, D.R. 2021 Knowledge, sociocognitive perceptions and the practice of handhygiene and social distancing during the COVID-19 pandemic: a crosssectional study of UK university students. COVID-19 preventive behaviours in White British and Black, Asian and Minority Ethnic (BAME) people in the UK. Risk perceptions and health behaviours as COVID-19 emerged in the United States: Results from a probability-based nationally representative sample. Can't tell Yes Yes Yes Yes Yes Yes ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS S1. Are there clear research questions? S2. Do the collected data allow to address the research questions? 117 First author Year Title Is the sampling strategy relevant to address the research question? Is the sample representative of the target population? Are the measurements appropriate? Is the risk of nonresponse bias low? Is the statistical analysis appropriate to answer the research question? Yes Hearne, B. 2021 Understanding how race, ethnicity, and gender shape maskwearing adherence during the COVID19 pandemic: evidence from the COVID impact survey. Yes Yes Yes Can't tell Yes No Kumar, V. 2021 Yes Yes Yes Can't tell Yes Yes Yes Nino, M. 2021 Yes Yes Yes No Yes No Yes Orom, H. 2021 Yes Yes Yes Can't tell Yes No Yes Reiter, P.L. 2021 Racial Disparities in the Perceived Risk of COVID-19 and in Getting Needed Medical Care. Race and ethnicity, gender, and age on perceived threats and fear of COVID19: Evidence from two national data sources. Racial/ethnic differences in prosocial beliefs and prevention behaviour during the COVID-19 pandemic. Racial/Ethnic differences in knowledge, attitudes, and beliefs about COVID-19 among adults in the United States. Yes Yes Yes Can't tell Yes No Yes ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 118 Appendix E Data Analysis of Rapid Review Studies First author Year Title Barret, C. 2021 Knowledge, socio-cognitive perceptions and the practice of hand-hygiene and social distancing during the COVID-19 pandemic: a crosssectional study of UK university students. Breakwell, G. 2021 COVID-19 preventive behaviours in White British and Black, Asian and Minority Ethnic (BAME) people in the UK. Methods/Study Design & Objectives Using a crosssectional online survey in May 2020, the study aimed to investigate knowledge and socio-cognitive perceptions, and their associations with selfprotective behaviours in UK university students A sample of 478 individuals in the UK participated in an online cross-sectional survey aims at investigating ethnic difference in patters of COVID-19 preventive behaviour and the social psychological factors associated with them Findings Ethnicity had a direct effect on CPB (BAME reported higher CPB) and an indirect effect on it through political trust, ingroup power. Ethnicity was not significantly related to COVID-19 fear. COVID-19 fear and trust in science were positively associated with CPB. Data for Question 1: Whether there are ethnic group differences in protective behaviours Ethnicity was not significantly related to COVID-19 fear Data for Question 2: Whether there are ethnic group differences in perceived exposure risk COVID-19 fear and trust in science were positively associated with CPB Data for Question 3: Whether perceived exposure risk explains variation in selfreported protective behaviours Data for Question 4: If there are ethnic differences in protective behaviours, whether is this mitigated by ethnic group differences in perceived exposure risk to COVID-19 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS First author Year Title Methods/Study Design & Objectives Findings Garfin, D.R. 2021 Risk perceptions and health behaviours as COVID-19 emerged in the United States: Results from a probability-based nationally representative sample. Using a national probabilitybased sample of 6,514 Americans, with three separate, consecutive representative cohorts between March 18, 2020 and April 18, 2020, the study aimed to identifying psychosocial correlates with self-protective behaviours Wearing face masks and/or gloves was positively associated with Black (β = .14, p < .001) and Hispanic (β = .07, p = .002) ethnicity, Hearne, B. 2021 Understanding how race, ethnicity, and gender shape mask-wearing adherence during the COVID-19 pandemic: evidence from the COVID impact survey. Using an online, nationally representative cross-sectional survey of 4688 respondents, the aim was to investigate the role of race and ethnicity on mask wearing and examining whether gender intersects with race and ethnicity to differently influence maskwearing patterns. Compared with White respondents, results revealed Black, Latina/o, and Asian respondents were more likely to report wearing a mask in response to the coronavirus. Moreover, results show White men were least likely to wear a mask from late April 2020 to early June 2020. 119 Data for Question 1: Whether there are ethnic group differences in protective behaviours Blacks saw themselves as less susceptible than whites Data for Question 2: Whether there are ethnic group differences in perceived exposure risk Data for Question 3: Whether perceived exposure risk explains variation in selfreported protective behaviours Data for Question 4: If there are ethnic differences in protective behaviours, whether is this mitigated by ethnic group differences in perceived exposure risk to COVID-19 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS First author Kumar, V. Year Title 2021 Racial Disparities in the Perceived Risk of COVID19 and in Getting Needed Medical Care. Methods/Study Design & Objectives Using a nationally representative survey in the U.S. in May 2020, they examined disparities in perceived risks of COVID-19 and getting medical care Findings Black respondents were 15 percentage points more likely than White respondents to believe the pandemic would not end by Summer 2020 (92 vs 77, p < .01), and were 19 percentage points more likely than any other race to feel a need to protect their family from COVID-19 (81 vs 62, p < .01). Latinx respondents were 10 percentage points more fearful than White respondents of catching COVID-19 in public places (55 vs 45, p < 0.01). 120 Data for Question 1: Whether there are ethnic group differences in protective behaviours In contrast, White people were the most optimistic about public places. The perceived probability of catching a COVID-19 infection in public places was lowest in the White population (44.8), and it was highest in the Latinx population (54.6) (p < 0.05, compared to White people). Data for Question 2: Whether there are ethnic group differences in perceived exposure risk Data for Question 3: Whether perceived exposure risk explains variation in selfreported protective behaviours Data for Question 4: If there are ethnic differences in protective behaviours, whether is this mitigated by ethnic group differences in perceived exposure risk to COVID-19 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS First author Year Title Methods/Study Design & Objectives Findings Nino, M. 2021 Race and ethnicity, gender, and age on perceived threats and fear of COVID-19: Evidence from two national data sources. A nationally representative panel of randomly selected U.S. adults responded to an online survey to test whether race and ethnicity, gender, and age were associated with six perceived threat and fear outcomes related to COVID-19 Racial and ethnic minorities are more likely to report high fear of coronavirus, perceive coronavirus as a major threat to population, and perceive coronavirus as a major threat to individual health. Orom, H. 2021 Racial/ethnic differences in prosocial beliefs and prevention behaviour during the COVID-19 pandemic. A nationally representative sample of 410 completed a survey about COVID-19 beliefs and prevention behaviours. The purpose of the study was to understand to what extent COVID-19 prosocial beliefs and behaviour differ by race/ethnicity Compared to White respondents, Black respondents perceived the risk of COVID19 to be greater to the US population; and both Black and Latinx respondents thought it was more important to protect a variety of nonclose others (e.g., people in their city or state). Black and 121 Data for Question 1: Whether there are ethnic group differences in protective behaviours Regression estimates show Latina/os were more likely to report coronavirus to be a high or very high threat to themselves and their families when compared to Whites, whereas Black and Asian American respondents do not differ from Whites. Compared to White respondents, Black respondents perceived the risk of COVID19 to be greater to the US population Data for Question 2: Whether there are ethnic group differences in perceived exposure risk Data for Question 3: Whether perceived exposure risk explains variation in selfreported protective behaviours There were indirect effects of Black vs. White race on engaging in protective behaviours through greater perceived risk to others and beliefs in the importance of protecting distal others Data for Question 4: If there are ethnic differences in protective behaviours, whether is this mitigated by ethnic group differences in perceived exposure risk to COVID-19 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS Reiter, P.L. 2021 Racial/Ethnic differences in knowledge, attitudes, and beliefs about COVID-19 among adults in the United States. Using an online survey with a convenience sample of 2,006 adults the aim of the research was to understand racial/ethnic differences in knowledge, attitudes, and beliefs about COVID-19 Latinx respondents engaged in several prevention behaviours, including social distancing, to a greater extent than White respondents. There were indirect effects of Black vs. White race on engaging in protective behaviours through greater perceived risk to others and beliefs in the importance of protecting distal others. For beliefs and attitudes, non-Latinx blacks ( = −0.09) and non-Latinx participants of another race ( = −0.05) reported lower perceived likelihood of getting COVID19 in the future compared to non-Latinx whites 122 In multivariable analyses, nonLatinx blacks ( = −0.09) and non-Latinx participants of another race ( = −0.05) reported lower perceived likelihood of getting COVID19 in the future compared to non-Latinx whites ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 123 Appendix F Sociodemographic Data from Survey Sociodemographic Race Mixed Heritage/Other Aboriginal (eg, Inuit, First Nations, Non-status Indian, Metis, Aboriginal person from outside Canada Frequency Percent Valid Percent Cumulative Percent 7 2 3.7 1.1 3.7 1.1 3.7 4.8 Asian–East (eg, Chinese, Japanese, Korean) Asian–South (eg, Indian, Pakistani, Sri Lankan, Indo-Caribbean) Asian–South East (eg, Malaysian, Filipino, Vietnamese) Black–Africa (eg, Ghanaian, Kenyan, Somali) Black–Caribbean region (eg, Barbadian, Jamaican) Middle Eastern (eg, Egyptian, Iranian, Lebanese) White/European (eg, English, Italian, Portuguese, Russian) Prefer not to answer 12 15 5 1 1 7 126 11 6.4 8.0 2.7 0.5 0.5 3.7 67.4 5.9 6.4 8.0 2.7 0.5 0.5 3.7 67.4 5.9 11.2 19.3 21.9 22.5 23.0 26.7 94.1 100.0 Language spoken at home Other (please specify) Arabic Bengali Chinese (Cantonese) Chinese (Mandarin) English Farsi (Persian) French Hindi Punjabi Russian Tagalog 2 1 1 1 2 164 1 11 1 1 1 1 1.1 0.5 0.5 0.5 1.1 87.7 0.5 5.9 0.5 0.5 0.5 0.5 1.1 0.5 0.5 0.5 1.1 87.7 0.5 5.9 0.5 0.5 0.5 0.5 1.1 1.6 2.1 2.7 3.7 91.4 92.0 97.9 98.4 98.9 99.5 100.0 Ability to speak English Very well Well Not well Not at all Prefer not to answer 171 14 1 0 1 91.4 7.5 0.5 0.0 0.5 91.4 7.5 0.5 0.0 0.5 91.4 98.9 99.5 99.5 100.0 Language preferred to read health care information English French 178 8 95.2 4.3 95.2 4.3 95.2 99.5 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS Spanish 124 1 0.5 0.5 100.0 147 40 78.6 21.4 78.6 21.4 78.6 100.0 Year arrived in Canada 1959 1967 1970 1972 1973 1974 1976 1981 1983 1984 1985 1988 1993 1994 1995 1997 1998 1999 2002 2005 2006 2008 2009 2010 2012 2013 2015 2018 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 3 1 1 2 1 2 1 1 1 2 5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 1.1 0.5 0.5 0.5 1.1 0.5 0.5 0.5 0.5 0.5 1.6 0.5 0.5 1.1 0.5 1.1 0.5 0.5 0.5 1.1 2.7 2.6 2.6 2.6 2.6 2.6 2.6 2.6 5.1 2.6 2.6 2.6 5.1 2.6 2.6 2.6 2.6 2.6 7.7 2.6 2.6 5.1 2.6 5.1 2.6 2.6 2.6 5.1 12.8 2.6 5.1 7.7 10.3 12.8 15.4 17.9 23.1 25.6 28.2 30.8 35.9 38.5 41.0 43.6 46.2 48.7 56.4 59.0 61.5 66.7 69.2 74.4 76.9 79.5 82.1 87.2 100.0 Year of birth 1937 1945 1946 1948 1950 1951 2 2 2 2 4 5 1.1 1.1 1.1 1.1 2.1 2.7 1.3 1.3 1.3 1.3 2.6 3.3 1.3 2.6 3.9 5.3 7.9 11.2 Born in Canada? Yes No ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 1952 1953 1954 1955 1956 1957 1958 1959 1961 1962 1964 1965 1966 1967 1968 1969 1970 1971 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 125 1 1 4 2 3 4 2 1 2 2 3 4 2 2 2 1 3 1 1 3 1 1 2 2 2 2 5 5 2 7 6 5 7 4 5 4 7 4 2 5 3 4 1 0.5 0.5 2.1 1.1 1.6 2.1 1.1 0.5 1.1 1.1 1.6 2.1 1.1 1.1 1.1 0.5 1.6 0.5 0.5 1.6 0.5 0.5 1.1 1.1 1.1 1.1 2.7 2.7 1.1 3.7 3.2 2.7 3.7 2.1 2.7 2.1 3.7 2.1 1.1 2.7 1.6 2.1 0.5 0.7 0.7 2.6 1.3 2.0 2.6 1.3 0.7 1.3 1.3 2.0 2.6 1.3 1.3 1.3 0.7 2.0 0.7 0.7 2.0 0.7 0.7 1.3 1.3 1.3 1.3 3.3 3.3 1.3 4.6 3.9 3.3 4.6 2.6 3.3 2.6 4.6 2.6 1.3 3.3 2.0 2.6 0.7 11.8 12.5 15.1 16.4 18.4 21.1 22.4 23.0 24.3 25.7 27.6 30.3 31.6 32.9 34.2 34.9 36.8 37.5 38.2 40.1 40.8 41.4 42.8 44.1 45.4 46.7 50.0 53.3 54.6 59.2 63.2 66.4 71.1 73.7 77.0 79.6 84.2 86.8 88.2 91.4 93.4 96.1 96.7 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 2000 2001 2002 126 1 1 3 0.5 0.5 1.6 0.7 0.7 2.0 97.4 98.0 100.0 4 55 16 2 4 2 10 35 6 11 7 9 10 3 5 8 2.1 29.4 8.6 1.1 2.1 1.1 5.3 18.7 3.2 5.9 3.7 4.8 5.3 1.6 2.7 4.3 2.1 29.4 8.6 1.1 2.1 1.1 5.3 18.7 3.2 5.9 3.7 4.8 5.3 1.6 2.7 4.3 2.1 31.6 40.1 41.2 43.3 44.4 49.7 68.4 71.7 77.5 81.3 86.1 91.4 93.0 95.7 100.0 Gender Other (please specify) Female Male Prefer not to answer 1 73 112 1 0.5 39.0 59.9 0.5 0.5 39.0 59.9 0.5 0.5 39.6 99.5 100.0 Sexual orientation Heterosexual (“straight”) Gay Bisexual Queer Questioning Prefer not to answer Do not know 168 3 10 1 1 2 2 89.8 1.6 5.3 0.5 0.5 1.1 1.1 89.8 1.6 5.3 0.5 0.5 1.1 1.1 89.8 91.4 96.8 97.3 97.9 98.9 100.0 Total household income <$10,000 $10,000 to $19,999 $20,000 to $29,999 $30,000 to $39,999 $40,000 to $49,999 $50,000 to $59,999 4 11 14 7 19 8 2.1 5.9 7.5 3.7 10.2 4.3 2.1 5.9 7.5 3.7 10.2 4.3 2.1 8.0 15.5 19.3 29.4 33.7 Religious affiliation Other (please specify) I do not have a religious or spiritual affiliation. Atheism Baha’i faith Buddhism Christian Orthodox Christian, not included elsewhere on this list Christianity Hinduism Islam Judaism Protestant Roman Catholic Sikhism Spiritual Prefer not to answer ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 127 $60,000 to $79,999 $80,000 to $99,999 $100,000 to $150,000 >$150,000 Prefer not to answer Do not know 28 23 35 10 22 6 15.0 12.3 18.7 5.3 11.8 3.2 15.0 12.3 18.7 5.3 11.8 3.2 48.7 61.0 79.7 85.0 96.8 100.0 Number of people income supports 0 1 2 3 4 5 6 7 2 38 55 19 22 6 4 1 1.1 20.3 29.4 10.2 11.8 3.2 2.1 0.5 1.4 25.9 37.4 12.9 15.0 4.1 2.7 0.7 1.4 27.2 64.6 77.6 92.5 96.6 99.3 100.0 62 112 13 33.2 59.9 7.0 33.2 59.9 7.0 33.2 93.0 100.0 25 56 76 22 7 1 13.4 29.9 40.6 11.8 3.7 0.5 13.4 29.9 40.6 11.8 3.7 0.5 13.4 43.3 84.0 95.7 99.5 100.0 134 12 38 3 71.7 6.4 20.3 1.6 71.7 6.4 20.3 1.6 71.7 78.1 98.4 100.0 Housing Rent Own Living with family or friends General health status Excellent Very Good Good Fair Poor Do not know Vaccination status Yes - Fully Vaccinated Yes - Partially Vaccinated (only 1 of 2 doses) No Prefer not to answer ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 128 Appendix G In what language do you feel more comfortable speaking in at home? Select one only. 1. American Sign Language 2. Arabic 3. Bengali 4. Chinese (Cantonese) 5. Chinese (Mandarin) 6. Cree 7. Dari 8. English 9. Farsi (Persian) 10. French 11. German 12. Greek 13. Gujarati 14. Hebrew 15. Hindi 16. Hungarian 17. Italian 18. Korean 19. Oji-Cree 20. Ojibwe 21. Polish 22. Portuguese 23. Punjabi 24. Russian 25. Somali 26. Spanish 27. Tagalog 28. Tamil 29. Urdu 30. Vietnamese 31. Other __________ How would you rate your ability to speak English? 1. Very well 2. Well 3. Not well 4. Not at all 5. Unsure 6. Prefer not to answer 7. Do not know In what language would you prefer to read health care information? 1. American Sign Language 2. Arabic 3. Bengali 4. Chinese (Cantonese) 5. Chinese (Mandarin) 6. Cree 7. Dari 8. English 9. Farsi (Persian) 10. French 11. German ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 129 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. Greek Gujarati Hebrew Hindi Hungarian Italian Korean Oji-Cree Ojibwe Polish Portuguese Punjabi Russian Somali Spanish Tagalog Tamil Urdu Vietnamese Other __________ Were you born in Canada? 1. Yes 2. No What year did you arrive in Canada? (leave blank if you prefer not to answer) What year were you born? (leave blank if you prefer not to answer) Which of the following best describes your race? 1. Aboriginal (eg, Inuit, First Nations, Non-status Indian, Metis, Aboriginal person from outside Canada 2. Asian–East (eg, Chinese, Japanese, Korean) 3. Asian–South (eg, Indian, Pakistani, Sri Lankan, Indo-Caribbean) 4. Asian–South East (eg, Malaysian, Filipino, Vietnamese) 5. Black–Africa (eg, Ghanaian, Kenyan, Somali) 6. Black–Caribbean region (eg, Barbadian, Jamaican) 7. Black–North America 8. Latin American (eg, Argentinean, Chilean, Salvadoran) 9. Middle Eastern (eg, Egyptian, Iranian, Lebanese) 10. White/European (eg, English, Italian, Portuguese, Russian) 11. Prefer not to answer 12. Mixed Heritage/Other __________ What is your religious or spiritual affiliation? 1. I do not have a religious or spiritual affiliation. Animism or 2. Shamanism 3. Atheism 4. Baha’i faith 5. Buddhism 6. Christian Orthodox ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 130 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Christian, not included elsewhere on this list Christianity Confucianism Hinduism Islam Jainism Judaism Native spirituality Protestant Rastafarianism Roman Catholic Sikhism Spiritual Unitarianism Wicca Zoroastrianism Prefer not to answer Do not know Other __________ Do you have any of the following disabilities? Check all that apply No disabilities Physical disability Chronic illness Sensory disability (ie, hearing or vision loss) Developmental disability Drug or alcohol dependence Learning disability Mental illness Prefer not to answer Do not know Other What is your gender? 1. Female 2. Male 3. Transexual Male 4. Transexual Female 5. Prefer not to answer 6. Do not know 7. Other What is your sexual orientation? 1. Heterosexual (“straight”) 2. Gay 3. Lesbian 4. Bisexual 5. Two-spirit 6. Queer 7. Questioning 8. Prefer not to answer 9. Do not know What was your total family income before taxes last year? 1. < $10,000 2. $10,000 to $19,999 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 131 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. $20,000 to $29,999 $30,000 to $39,999 $40,000 to $49,999 $50,000 to $59,999 $60,000 to $79,999 $80,000 to $99,999 $100,000 to $150,000 >$150,000 Prefer not to answer Do not know How many people does this income support? (leave blank if you 'Do not know' or 'Prefer not to answer' What type of housing do you live in? 1. Rent 2. Own 3. Living with family or friends 4. Temporary housing (eg, shelter, hostel) 5. Homeless 6. Correctional Facility 7. Prefer not to answer 8. Other In general, would you say your health is: 1. Excellent 2. Very Good 3. Good 4. Fair 5. Poor 6. Prefer not to answer 7. Do not know Have you been vaccinated with any COVID-19 vaccines? 1. Yes - Fully Vaccinated 2. Yes - Partially Vaccinated (only 1 of 2 doses) 3. No 4. Prefer not to answer 5. I don't know ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 132 COVID Own Risk Appraisal Scale (CORAS) (Extremely Unlikely) 1 2 3 4 (Extremely Likely) 5 Prefer not to answer (Strongly Agree) 1 2 3 4 (Strongly Disagree) 5 Prefer not to answer (Very Hard to Do) 1 2 3 4 What is your gut feeling about how likely you are to get infected with COVID-19? I think my chances of getting infected with COVID-19 are I am sure I will NOT get infected with COVID-19 I feel I am UNLIKELY to get infected with COVID19 (Very Easy to Prefer not to Do) answer 5 Picturing myself getting COVID-19 is something that I find Which of the following have you done in the last seven days to keep yourself safe from COVID-19? Only consider actions that you took or decisions that you made personally No Washed your hands with soap or used hand sanitizer several times per day Avoided contact with people who could be high-risk (e.g.,: elderly, children, immunocompromised) Avoided public spaces, gatherings, or crowds Avoided eating at restaurants Worn a mask or other face covering Yes Prefer not to answer ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 133 Thank you for your participation in this study. If you have any questions regarding your participation in the study or the results, please contact Daniel Maullon at daniel.maullon@mytwu.ca For additional information regarding COVID-19, please visit the Government of Canada COVID-19 information webpage: https://www.canada.ca/en/public-health/services/diseases/2019-novel-coronavirusinfection/prevention-risks.html Your unique Amazon Mechanical Turk completion code is: DRM-EEDSRPBACMPR-21 ETHNIC DIFFERENCES IN PROTECTIVE BEHAVIORS 134 Appendix H Image 1. MTurk HIT introduction page