Analytical Studies: Methods and References
Framework for the Use, Analysis, Interpretation and Dissemination of Police-reported Indigenous and Racialized Identity Data Through the Uniform Crime Reporting Survey

Skip to text

Text begins

Executive summary

In the 2021 Canadian federal budget, the Government of Canada recognized the critical need for data disaggregation to address racism, gender gaps and systemic barriers across the country. A five-year financial commitment was made to support the Disaggregated Data Action Plan (DDAP), a whole-of-government initiative led by Statistics Canada (Treasury Board of Canada Secretariat, 2023).

Disaggregating data refers to the process of breaking down large, aggregated datasets into finer categories based on relevant characteristics (Statistics Canada, 2024a). At Statistics Canada, the DDAP aims to produce more specific data on inequities across Canadian society.

Disaggregated data are essential for several reasons, including providing a better understanding of marginalized populations, helping with evidence-based policy making, monitoring progress and program evaluation, and ultimately aiming to bring fairness and inclusion to the decisions that affect the population of Canada (Treasury Board of Canada Secretariat, 2023; Statistics Canada, 2022b; United Nations, 2018).

This report discusses a joint initiative between Statistics Canada and the Canadian Association of Chiefs of Police, funded through the DDAP, to expand Statistics Canada’s Uniform Crime Reporting (UCR) Survey. The police-reported Indigenous and racialized identity data (PIRID) initiative was created to start collecting PIRID from victims and accused persons involved in criminal incidents, through Statistics Canada’s UCR Survey. The absence of these data at a national and systematic level represents a significant gap in criminal justice system data within the Canadian context.

This report, titled Framework for the Use, Analysis, Interpretation and Dissemination of Police-reported Indigenous and Racialized Identity Data Through the Uniform Crime Reporting Survey, is intended as a guide for users of the PIRID collected through the UCR Survey. Aspirational in nature, this document is evergreen—it sets an initial foundation for the responsible use of the data and will be updated as required.^Note 1

This framework has two main objectives:

1. Support the responsible and ethical use of PIRID by proposing five guiding principles, which are intended to help avoid further stigmatizing and marginalizing communities through the use of these data.
2. Equip data users with tools and guidance for careful and culturally competent data interpretation, ultimately contributing to the development of evidence to support decision making for the creation of more equitable outcomes in policing.

This report is organized into two main sections:

Section 1 introduces the UCR Survey expansion to enable the collection of Indigenous and racialized identity data and contextualizes the need for the framework.

Section 2 introduces five key principles for data users to support the responsible and ethical use of data collected under this initiative. It outlines important considerations and recommendations related to the use, analysis, interpretation and dissemination of the data, to ensure the implementation of the guiding principles. While supporting academic freedom in research, this section introduces users to Statistics Canada’s systematic approach to research—an approach that emphasizes the use of key checkpoints in the responsible use of disaggregated data.

Acknowledgments

Statistics Canada acknowledges and expresses gratitude to all those who participated in the development of the current framework, particularly the work undertaken by the Canadian Association of Chiefs of Police’s special purpose committee^Note 2 and the critical guidance of Indigenous and racialized community organizations, academics, police services, and others who participated in Statistics Canada’s consultative engagement.

Introduction

The police-reported Indigenous and racialized identity data initiative

On July 15, 2020, Statistics Canada and the Canadian Association of Chiefs of Police released a joint statement announcing their commitment to working collaboratively on collecting data about the Indigenous and racialized identity of victims and accused persons^Note 3 identified as involved in police-reported criminal incidents, as reported through the Uniform Crime Reporting (UCR) Survey.  

Increasing calls for better disaggregated data on individuals’ diverse experiences generally, and growing concerns about the different treatment of Indigenous and racialized persons in the Canadian criminal justice system more specifically, have revealed important gaps in the availability of disaggregated data about the population of Canada (Canadian Heritage, 2019; David & Mitchell, 2021; Millar & Owusu-Bempah, 2011; Samuels-Wortley, 2021). 

The police-reported Indigenous and racialized identity data (PIRID) initiative emerged as a concrete action toward responding to these calls, particularly given the longstanding overrepresentation of Indigenous and racialized (particularly Black) populations in the Canadian criminal justice system (National Inquiry into Missing and Murdered Indigenous Women and Girls, 2019; David & Mitchell, 2021; Millar & Owusu-Bempah, 2011; Samuels-Wortley, 2021; Statistics Canada, 2022a; Truth and Reconciliation Commission of Canada, 2015). The PIRID initiative aims to shed light on issues of systemic inequalities and discrimination in the criminal justice system, beginning with police as the entry point into the system.^Note 4

Among other things, these data can be used for

• identifying differences and inequities in policing pathways and outcomes, in combination with linked data
• providing quantitative indications to support evidence-based policies and programming
• developing targets and benchmarks to monitor progress and assess the effectiveness of policies and programs and their impacts on specific populations
• identifying the role of systemic issues in the inequitable experiences of Indigenous and racialized persons in policing and the criminal justice system more broadly, in combination with linked data
• providing communities with analytical findings to support programs and initiatives.

Changes to the Uniform Crime Reporting Survey under the police-reported Indigenous and racialized identity data initiative

The main outcome of the PIRID initiative is the expansion of Statistics Canada’s well-established UCR Survey to enable the collection of Indigenous and racialized identity data. The UCR Survey is an administrative survey that collects information from police records management systems on criminal incidents reported to police, to measure the volume and severity of crime in Canadian society and its characteristics. Effective February 2024, the UCR Survey was updated to enable the collection of data on the Indigenous and racialized identity of victims and accused persons, through the addition of new variables (Annex A).^{Note 5 Note 6}

The collection of Indigenous and racialized identity information through the UCR Survey follows Statistics Canada’s standardized population group categories.^{Note 7 Note 8} At Statistics Canada, a population group refers to whether a person reports being White, South Asian, Chinese, Black, Filipino, Arab, Latin American, Southeast Asian, West Asian, Korean, Japanese or another population group. These specific groups are the mark-in categories used in a question that collects data on the racialized population. While this document will use the terms “racialized identity” and “self-identification”, the population group variable and its derivations should not be considered a measure of self-identification but rather a reporting measure. Given the wording of the population group question, as well as elements such as individual perception and the response categories based on the Employment Equity Act it should be noted that what respondents report may not align with how they self-identify in everyday life. Statistics Canada’s population group standard was chosen as the proxy for measuring Indigenous and racialized identity through PIRID, given that consistent and comparable data were often cited as a fundamental issue. This method provides a foundation for common categories across Statistics Canada surveys and products and across most of the Canadian criminal justice system data.

Note that the concept of “race” emerges from social processes by which individuals are differentiated based on varied characteristics, generally leading to the marginalization of certain groups (Ontario Human Rights Commission, 2005). It follows that racialization refers to “the process by which societies construct races as real, different and unequal in ways that matter to economic, political and social life” (Commission on Systemic Racism in the Ontario Criminal Justice System, 1995, pp. 40-41). The concepts of “race” and “racialization” are socially constructed, and, as such, definitions and categorizations of these concepts may differ across geographies and cultures (Viano & Baker, 2020).

The terms “racialized persons” and “racialized identity” are therefore used broadly in this document in reference to those who are not Indigenous or White. While First Nations individuals, Métis and Inuit also experience racialization, it is important to consider their unique historical, political and legal context and their unique experiences. It is also essential to note that Indigenous peoples may not define themselves as racialized groups, but instead as peoples or nations (Ontario Human Rights Commission, 2018). It follows that Indigenous peoples need to be considered separately from other racialized identities (Ontario Human Rights Commission, 2003, 2018).

The collection of Indigenous and racialized identity information by police services is conducted using either an officer perception method, a self-identification method or both. The method of collection is selected at the discretion of individual police services.^Note 9

Building the Framework for the Use, Analysis, Interpretation and Dissemination of Police-reported Indigenous and Racialized Identity Data Through the Uniform Crime Reporting Survey

Purpose

The purpose of the PIRID initiative is to develop a standardized national approach for the collection of Indigenous and racialized identity information for victims and accused persons involved in criminal incidents reported by police to Statistics Canada through the UCR Survey (Reference Operational guidelines). These data provide a foundation for quantitative research and analysis, with policing data, to support informed decision making.

The PIRID initiative is grounded in the goal of producing data that can be used to create positive change. Significant feedback through engagement with community organizations, police services and academics, who raised concerns related to the historical and potential future misuse of data to further marginalize specific communities, groups and individuals, highlighted the need for robust safeguards, parameters, guidelines or protocols to help mitigate these concerns.^Note 10

The framework was developed as a critical guide for responsible data use. It aims to achieve two key objectives:

1. Support the responsible and ethical use of PIRID by proposing five guiding principles, which are intended to help avoid further stigmatizing and marginalizing communities through the use of these data.
2. Equip data users with tools and guidance for careful and culturally competent data interpretation, ultimately contributing to the development of evidence to support decision making for the creation of more equitable outcomes in policing.

While the current framework mainly refers to the use of PIRID within the context of a research project, the guidelines and approaches presented herein should be considered in all uses of the data.

Intended audience

The intended audience for the framework is UCR Survey data users who plan to use the new Indigenous and racialized identity data, in any format, at any point in the data cycle, and all of the derivatives of these data, including custom and published data. These users comprise researchers, analysts and anyone who uses these data or information. This includes Statistics Canada employees and others accessing the data through different means (e.g., through Statistics Canada’s research data centres, custom requests or publicly available aggregate statistics).

Key considerations and limitations

When using PIRID, data users should be aware of key considerations and limitations inherent in the UCR Survey data, the analytical methodologies for measuring systemic racism and the scope of the framework. By keeping the following considerations in mind, data users can approach PIRID analysis with a more nuanced and context-aware perspective, ultimately leading to more meaningful interpretations.

1. Acknowledging the limitations of Uniform Crime Reporting Survey

As stated above, the UCR Survey collects data about victims and accused persons involved in criminal incidents only. Although police interactions involving persons other than victims and accused persons (e.g., complainants, witnesses) and non-criminal incidents (e.g., wellness checks, traffic stops, street checks, use of force incidents), are still important for understanding inequitable treatment in the criminal justice system, they are outside the scope of the UCR Survey.

Approximately one in five police calls for service results in a police-reported criminal violation. Police call for service include all instances resulting in a police response for both criminal and non-criminal events. For example, they include “motor vehicle accidents and traffic matters, incidents of domestic violence, reports of suspicious individuals or circumstances, disturbance or disorderly conduct, attempted suicide, and calls related to intoxicated persons” (Mazowita & Rotenberg, 2019, p. 9). This means that the PIRID data collected by the UCR Survey only reflect 20% of calls for service, which is a notable limitation of understanding policing interactions. However, commencing the collection of PIRID through the UCR represents a starting point for understanding potential systemic racism in policing-reported crime, and may be expanded to other incident types in the future (Statistics Canada,2022a). Further, for the collection of PIRID, the UCR Survey uses standard Statistics Canada population group categories, which are used across surveys at Statistics Canada and the federal government more broadly to measure self-identification. While there are benefits to the universal use of population group standards, their broad categorization and the use of race, ethnicity and nationality to construct the population groups (Statistics Canada, 2024b) limit the ability to conduct more granular analysis. The limitations of these categories may obscure the diversity of experiences of Indigenous and racialized groups combined under one variable.

2. Measuring and addressing systemic racism

Addressing systemic racism and discrimination within the criminal justice system is complex. It demands multi-system, multi-sector and multi-pronged approaches. These approaches must tackle the issues of systemic racism within the criminal justice system; the socioeconomic issues that underlie the disproportionate disadvantages of some communities; and “multisectoral” systemic racism, that is, systemic racism across other critical social systems (e.g., education, housing and health care)^{Note 11} that may drive disproportionate criminal justice system involvement. Data collection and analysis represent only one avenue for supporting the identification, quantification and dismantling of systemic discriminatory issues and practices in policing.

3. Flags versus measures

It is important to acknowledge that although valuable, PIRID alone cannot conclusively determine the existence or cause of systemic racism in policing or the criminal justice system more broadly; rather, these data should be taken as flags for further study. They are limited to the policing sphere (based on the UCR Survey scope) (Text box 2). As stated above, addressing systemic issues of racism and discrimination within the criminal justice system demands multi-system and multi-sector approaches. References are made throughout this document to performing record linkages between PIRID and other datasets outside the UCR Survey scope. These include other criminal justice system datasets beyond policing, as well as datasets from other sectors where data provide information on social determinants of justice (e.g., health, housing and education). Leveraging such datasets to complement PIRID can help quantify measures of systemic racism across all aspects of the criminal justice system.

4. Scope of the framework

Within this context, to support the responsible use of the new UCR Survey (UCR 2.5) PIRID variables, the framework provides key analytical and methodological considerations that emphasize the importance of

• understanding the inherent limitations of PIRID, which influence what can and cannot be definitively concluded from data analysis
• acknowledging the impact of other internal and external factors that may contribute to contact with, and overrepresentation in, the criminal justice system, starting at the policing stage.

The current framework aims to balance freedom and flexibility of data use with the ethical imperative of not causing further marginalization and stigmatization of the communities represented in the data. The framework presented here offers approaches for the use, analysis, interpretation and dissemination of Indigenous and racialized identity data. It is essential to acknowledge that this framework is not exhaustive and that other ethical and methodological approaches can be applied. Researchers are therefore encouraged to consult additional guidelines and frameworks to complement the principles outlined herein, thereby helping to ensure the application of the highest ethical and methodological standards in their work.

The data collected through the UCR Survey are extracted from police records management systems. Police services collecting the data may have different rules or guidelines related to access, use, analysis, interpretation and dissemination of their data prior to data transmission to Statistics Canada.^Note 12

Framework for the Use, Analysis, Interpretation and Dissemination of Police-reported Indigenous and Racialized Identity Data Through the Uniform Crime Reporting Survey

The systematic approach to research

The framework recommends the use of Statistics Canada’s systematic approach to research in the development of projects using PIRID. For more information on the systematic approach to research, see Annex B. The systematic approach to research focuses on determining the “what, why, how and so what” of one’s research. It is based on a series of checkpoints to ensure that research with disaggregated data is based on the isolation of a specific question or problem that is in need of further study. Four key checkpoints are discussed within the framework:

1. issue identification
2. conceptualization
3. data and analysis
4. interpretation and communication of results.

Based on extensive feedback received through Statistics Canada’s police, community, and academic engagements (Report and Final Recommendations), the framework builds on the key checkpoints of the systematic approach to research to identify a number of important elements for consideration when using PIRID. Considering these elements will support the application of the guiding principles (see Text box 2 below) when using PIRID to ensure these data are used responsibly.  As a resource for PIRID users, a data user checklist is included in Annex C, to ensure the proper implementation of the framework in the use, analysis, interpretation and dissemination of PIRID.

Issue identification

The first step in the systematic approach refers to determining the issue with a supporting “problem statement” describing the purpose of intended activities and the specific questions being addressed.

When using PIRID, data users should specify clear objectives for how the data are going to be used, with a focus on the questions to be studied and the intended outcomes.

Data users should aim to determine objectives with the help of the communities that the data represent.^Note 15 This can be done by engaging directly with members of impacted communities from the outset of the project to develop a deeper knowledge of the community, their experiences and areas where research is needed. Through this engagement process, researchers can

• establish mechanisms, such as ongoing engagement and collaboration or the inclusion of qualitative methodologies, to more accurately reflect the lived experiences of the communities represented in the data
• determine the appropriate methodologies and analyses to use when working with these data
• identify approaches to include community engagement and participation throughout the analytical process, potentially including the development of a research question, to ensure accurate representation of the data and findings, and to enhance the utility and relevance of the final analytical product for the community
• consider how their findings and research relate to or align with existing community efforts (see Boilevin et al., 2018).

Conceptualization

The second step in the systematic approach to research is to determine the relevance and necessity of the intended outcomes of the project, through a review of the literature (academic papers, government and non-governmental organization reports, etc.), targeting the gaps the project proposes to fill and, if possible, engaging with individuals with lived experiences to assess for gaps or vital research questions. In this step, the key variables of study and how they are related to the objectives of the project can be determined. As well, contextual factors (e.g., historical and present context of the communities and individuals being represented, and legislative context) can be determined, including the nature of the analysis (e.g., presentation of descriptive statistics and multivariate modelling) and the level of disaggregation and intersectionality among the chosen key variables. The level of disaggregation and dimensions of intersectionality should also be driven by the project’s objectives.

Selecting key variables of study: Using officer perception versus self-identification variables

As stated at the outset, PIRID can be collected through two methods: officer perception and self-identification.

Officer perception: This variable represents an officer’s assessment of an individual’s Indigenous or racialized identity based on their observation of the person’s appearance and any other information known to the officer at the time of the interaction (Ontario Human Rights Commission, 2005).

This information is important because perceptions influence officers’ decision making during an interaction (Statistics Canada, 2022a; Correll, Park, Judd, & Wittenbrink, 2002). Perception data are collected for the purpose of monitoring potential bias in a specific service delivery, program or function. Officer perception data are most valuable when used to examine how systemic bias toward the perceived identity of victims and accused persons may impact decision making and policing outcomes (Ontario, 2018; Toronto Police Service, 2022).

Self-identification: This variable represents how an individual may self-identify in response to being asked their Indigenous or racialized identity during a police interaction. Self-identification may be based on many factors personal to the individual, such as history, family, community and culture, as well as reflect their personal preferences and experiences and can change over time.

This information provides insight into different experiences of members of Indigenous or racialized communities when accessing policing services. Self-identification data are most valuable when examining “whether individuals’ self-identified race and other identities are associated with their outcomes and experiences with police” (Toronto Police Service, 2022, p. 8). This form of data can also provide insights on systemic barriers that individuals from certain communities might experience when reaching out to police (e.g., reporting hate crimes or domestic violence) (Ontario, 2018).

It is recommended to consider the expected qualitative and quantitative differences between officer perception and self-reported identity data, it is imperative to analyze these variables independently of each other and not to combine or make direct comparisons between these variables.

Failure to properly consider the underlying constructs represented by these variables will likely lead to misleading results. Data users must carefully consider the purpose of each variable and select the appropriate one for their study to ensure valid findings. Analytic strategies should take missing data into consideration.

To maintain full transparency, data users must indicate whether results are derived from officer perception or self-identification Indigenous or racialized identity data.

Important notes on police-reported Indigenous and racialized identity data population group categories

As stated previously, Indigenous and racialized identity collected through the officer perception and self-identification methods are measured using Statistics Canada’s population group categories. When interpreting, analyzing and communicating the data with the public, data users should consider the following sensitivities.

Indigenous identity

Data users should employ language that acknowledges both the shared histories of Indigenous peoples in Canada and the diverse experiences of individual groups. At Statistics Canada, a distinctions-based analytical approach is used when examining Indigenous identity. The Government of Canada defines a distinctions-based approach as “working independently with First Nations Peoples, Inuit, Métis Peoples and Intersectional Peoples in recognition of their unique attributes” (Indigenous Services Canada, 2022; Indigenous Services Canada, 2023).  

Disaggregating data for First Nations, Métis and Inuit populations, when feasible, allows researchers to capture the specific outcomes of each group (see Annex A for what data are collected through the UCR Survey). It is important to note that Indigenous communities across Canada may self-identify specifically with one’s “clan, community, nationhood, or language family,” beyond the three distinct Indigenous groups (Ontario, 2018). These may be captured through the self-identification method in the open-field write-in option.

The term “pan-Indigenization” describes how individuals from different Indigenous groups often get grouped together under the broad umbrella term of “Indigenous,” where unique cultural, historical and linguistic differences are ignored. By acknowledging the potential for pan-Indigenization, researchers can be more mindful of the diversity across Indigenous communities and steer away from generalizations.

Avoiding pan-Indigenization requires clear differentiation between First Nations, Métis and Inuit identities, as well as how these identities may overlap. Statistics Canada acknowledges that data collection methods should enable the reporting of distinctions-based Indigenous identities where possible, allowing for a more nuanced understanding of experiences in the criminal justice system. Therefore, PIRID collection through the UCR Survey has been expanded such that police services can report perceived distinctions-based Indigenous identity information through a “detailed” officer perception variable, while allowing for the collection of partial Indigenous identity information through a “general” category variable when distinctions-based perception is not feasible. The self-identification collection method allows for the reporting of distinctions-based Indigenous identity using the Statistics Canada population group categories and an open-field response option.

Due to sample size limitations, analyses using Statistics Canada's population group categories may not allow for the full disaggregation of Indigenous groups. There will need to be a balance between the goals of the research, the data available and the conclusions that can be drawn. Further, disaggregating data for First Nations, Métis and Inuit populations may not be possible using officer perception PIRID, since officers may not be able to make a visual distinction between groups.

Black and other racialized populations

The Black^Note 16 population in Canada is made up of over 300 ethnic or cultural origins, speaking over 250 languages (Statistics Canada, 2024c). Despite the rich diversity of these populations in Canada, persons of African descent are often conflated under one racialized identity as “Black.”

PIRID collected through the UCR Survey use standard Statistics Canada population group categories, which include the category “Black.” While these standards are currently being assessed for future amendment to better reflect the groups they represent (Statistics Canada, 2024b), it is important to note that, currently, this is a limitation to the analytical capacity of PIRID relating to information on the Black population.

Similar concerns about self-identification apply to other racialized groups in Canada that have unique circumstances and encompass rich and diverse backgrounds, languages, ethnicities and cultures that might not be accurately reflected in the data fields.

It is recommended that data users acknowledge this limitation in their use of PIRID and do their due diligence in highlighting the diversity of the populations represented or those that are of specific interest in their research project. Data users should be cognizant of how the populations of interest in the study are shaped by time, geography, ethnicity and culture. This means data users should acknowledge the heterogeneity that exists within groups, taking note of the negative impact of generalizing statements.

Interpreting multiple self-identification variables in the Uniform Crime Reporting Survey

Although the officer perception collection method allows for the selection of only one identity category, multiple population group categories can be selected through the self-identification collection method, including an open-field response^Note 17 that allows victims and accused persons to specify their identity in their own words, in addition to or instead of selecting from the provided list (Annex A). The ability to select multiple options under the self-identification collection method can support analysis of data on persons with multiple racialized identities.

When analyzing PIRID, data users may want to recategorize self-identification variables when multiple categories are reported. Recategorization can be done through the creation of new “multiple identity” categories that combine certain groups together. This can lead to various interpretations of the data, depending on how the data user chooses to group and analyze multiple identities.

Understanding these different approaches is essential for accurate interpretation of the findings:

1. Data users can create a new “multiple identity” category to include respondents who identify with more than one Indigenous or racialized population group. Previous research suggests that multiracial individuals may have distinct experiences compared with those with a single racialized identity (Hernández, 2018). However, it is important to acknowledge that a “multiple identity” category may not fully capture the nuances of experiences within and across different multiracialized identities (British Columbia Ministry of Citizens’ Services, 2023).
2. In cases where sample sizes are adequate and a general “multiple identity” category might mask important differences, researchers can create more specific categories that combine specific population group identities, such as “Black and White,” “Inuit and First Nations,” “Métis and Black,” “South Asian and White,” “East Asian and White,” and “East Asian and Black.” This approach can help to identify and analyze the experiences of individuals with various combinations of multiracialized population group identities.
3. When a respondent identifies as both “White” and another racialized identity, researchers may consider recategorizing them to the racialized identity. This approach is based on the idea that individuals who identify as both White and a racialized identity may have experiences closer to those of the racialized identity because of the effects of racialization (British Columbia Ministry of Citizens’ Services, 2023; Ontario, 2018; Statistics Canada, 2022b). However, it is important to recognize that biracial or multiracial experiences are unique from monoracial experiences, and recategorizing individuals in this way will most likely capture less nuance of their racialization (British Columbia Ministry of Citizens’ Services, 2023). In many cases, retaining both categories can provide a more detailed and intersectional analysis.

If a data user opts to recategorize multiple identity groups, it is recommended that they clearly disclose how population group categories were used. Careful consideration should be given when combining categories, specifically attempting to strike a balance between analytical utility and generalization.

Harm prevention

Data hold an immense amount of power to cause harm to individuals and perpetuate inequality (British Columbia’s Office of the Human Rights Commissioner, 2020). Data about Indigenous and racialized populations have historically been used by governments to justify colonialism and the monitoring and subjugation of Indigenous and racialized populations in societies, in Canada and globally, and they continue to be an ongoing source of harm (Ontario, 2018; Rossiter & Ndekezi, 2021; United Nations, 2018). In the context of this document, harm prevention refers to steps taken to avoid using PIRID to perpetuate or reinforce such misuse, or steps taken to mitigate it.

Users of PIRID must be mindful of the potential unintended consequences of their work with the data and, to their best ability, avoid causing harm. The narrative created through data use should be intentional and ultimately benefit the communities it represents and society as a whole (United Nations, 2018). It is important that identity data be treated with sensitivity, given the ways they have been used to violate the human rights of Indigenous and Black and other racialized populations in Canada (Ontario, 2018). At the conceptualization step, data users should consider whether their use of PIRID will cause additional harm to individuals and communities represented in the analyses.

To prevent harm, data users have a critical responsibility to implement robust safeguards when working with PIRID, including the following:

• Community engagement: As a guiding principle of the framework, community engagement can be conceptualized as a method to reduce the potential for harm caused by data use. Involving communities across the project lifespan demonstrates respect to those represented by the data and can help ensure that the data are being used ethically and in ways that align with community values (Text box 3). Engaging with communities will also help data users use the data in a way that promotes positive change.
• Mitigation of stigma and profiling: Data users should ensure that research does not contribute to the reinforcement of existing stereotypes, profiling practices or the targeting of specific communities. It is essential to present findings in a way that promotes understanding.
• Reinforcement of the social construction of race and racialization: Data users should emphasize the social and historical context of race and racialization. While these concepts must be clarified as important for understanding societal disparities and informing issues around differential treatment, it is crucial to avoid perpetuating harmful stereotypes or biases and not attribute criminality as an inherent trait of certain groups (Foster et al., 2023; Toronto Police Service, 2022).
• Rich contextualization: To ensure a comprehensive understanding of the data and analytical findings, researchers should provide sufficient historical and contemporary context. This involves reviewing the academic literature, publications from community organizations and legislation governing policing. The unique context of the population under study must be reviewed and considered. If multiple groups are being examined, the context for each should be considered separately.
• Cultural competency: To enhance their understanding of the complexities of Indigenous and racialized identity data, data users are encouraged to undergo cultural awareness training (see Annex C for resource examples), focusing especially on the historical contexts that continue to impact these communities in Canada. This training can foster greater sensitivity and respect for the diverse experiences represented in the data. Cultural competency can also be gained when speaking directly to communities to help determine culturally appropriate questions and approaches to analysis.
• Cultural safety: To ensure respectful and ethical use of Indigenous and racialized identity data, data users are encouraged to engage in cultural safety training (see Curtis et al., 2019; So, Price, O’Mara, & Rodrigues, M. A., 2024). This training emphasizes understanding the power imbalances inherent in research and data collection, recognizing the potential for re-traumatization, and centering Indigenous and racialized perspectives. Cultural safety goes beyond awareness and competency by requiring data users to critically examine their own biases, assumptions, and positionality in relation to the communities represented in the data. This includes acknowledging the ongoing impacts of colonialism and systemic racism. Furthermore, fostering cultural safety involves establishing respectful relationships with communities, ensuring their meaningful participation in all stages of the research process, from data collection and analysis to interpretation and dissemination of findings. This collaborative approach aims to create a safe space for knowledge sharing and ensure that data is used in a way that benefits the communities involved.
• Anti-bias training: Data users should be aware of their own biases and how these can influence a project through question formulation, result interpretation and dissemination. To mitigate the impact of these biases, data users should consider taking courses on anti-bias specifically tailored to data analysis, as well as engage in ongoing consultative engagements with communities. See Annex C for training and education resources.

Intersectionality

Intersectionality, a term coined by legal scholar Dr. Kimberlé Crenshaw (1989), was originally used to highlight the intersection of marginalization Black women experience through multiple systems of oppression (e.g., racism and sexism). The concept of intersectionality has since been applied to broader contexts to understand how lives are shaped by the multitude of identities, relationships and social factors that influence an individual. These combine to create intersecting forms of privilege and oppression (Government of Canada, 2024).

When interpreting PIRID, it is crucial to be mindful of the complexities of intersectionality. Data users should avoid oversimplifying findings by conflating identities (e.g., pan-Indigenization; see Text box 4) and arbitrarily grouping racialized population groups (British Columbia Ministry of Citizens’ Services, 2023), while balancing statistical standards for analyzing small sample sizes. To provide a nuanced understanding, data users should strive, whenever possible, for disaggregate analysis and supplement PIRID with other data sources.

Moreover, to quantify and understand systemic issues of racism and discrimination in policing and the criminal justice system, it is important to understand how different sociodemographic factors interact with each other. That is, Indigenous or racialized identity should be considered alongside other factors, such as gender, geography, age, education and income, to examine their intersection with criminal justice system involvement.^Note 19 For example, researchers can use statistical models to explore how these factors interact and identify the strongest risk factors for involvement with the criminal justice system and corresponding outcomes (Text box 7).

Data and analysis

This step of the systematic approach to research encompasses analytical planning and the determination of methodologies to address research questions. This section of the framework outlines some important technical considerations when using PIRID, focusing on

• measures of disproportionality and disparity, including benchmarking and setting thresholds
• social determinants of justice and data linkage, including multivariate analysis
• statistical standards
• data quality.

Although the analytical considerations outlined in the framework reference commonly used analytical measures and methods, they should be considered in any analytical approach conducted using PIRID.

Measures of disproportionality and disparity

Disproportionality and disparity indexes are measures commonly used by public institutions to compare the outcomes of Indigenous and racialized communities across social sectors, including health, education and criminal justice systems (Ontario, 2018). Specifically, these measures can be used to identify potential unjust treatment and inequitable outcomes (Toronto Police Service, 2022, p. 8).

Disproportionalities and disparities occur when policies, programs and services unequally benefit and disadvantage groups across socioeconomic and demographic identities, resulting in different outcomes (Ontario, 2018)^Note 20. Ontario’s (2018) Data Standards for the Identification and Monitoring of Systemic Racism defines disproportionality and disparity indexes as follows:

• A disproportionality index is a measure that determines the over- or underrepresentation of a group of interest compared with their representation within a chosen reference group or “benchmark” population (Text box 5). For example, the population of all victims or all accused can be used as a benchmark for representation within particular UCR Survey violation types. Another frequently used benchmark is representation in the local resident population. Disproportionalities answer the following question: if everyone had an equal chance of being involved in a police incident, are certain groups overrepresented relative to their representation in the population? However, they cannot determine why overrepresentation exists.
• A disparity index is a measure that determines the differences in outcomes between a group of interest and a chosen reference group. For example, the proportion of an Indigenous or racialized group with a certain outcome (e.g., charges laid) might be compared (“benchmarked”) to the proportion of White individuals with the same outcome. Disparities indicate whether individuals from certain groups may receive different treatment at key decision-making points within the criminal justice system. Again, disparities cannot determine why overrepresentation exists, but can flag specific points for further research.

It is recommended that data users should ensure that they clearly identify and define the benchmarks used in their analysis and why they were selected.

Setting thresholds

Thresholds for disparity and disproportionality measures are predetermined values representing the level at which the measures are considered indicative of substantial disparity or disproportionality. These thresholds help prioritize areas requiring further research. Before analyzing PIRID, data users should determine appropriate thresholds: what magnitude of difference between group outcomes would indicate whether racial disparities or disproportionalities warrant addressing?

Determining the appropriate thresholds involves a delicate balance between sensitivity and specificity:

• Sensitivity: A low threshold might capture even small disparities or disproportionalities, ensuring no potential issues are missed. However, this can lead to flagging situations where the difference is minimal and may not warrant significant resources for research or programming.
• Specificity: A high threshold would flag only large disparities or disproportionalities, potentially overlooking less severe but concerning situations.

To address the above issues, a multi-tiered system could be considered with varying thresholds depending on the severity of the event and the magnitude of the disparity or disproportionality. The following is an example:

• Tier 1: Lower threshold for high-impact events: For the most severe events, a lower threshold might be established to ensure even small disparities receive attention.
• Tier 2: Mid-range threshold for broader concerns: For relatively less severe but still substantially concerning events, a mid-range threshold might be used for prompting additional research and investigation.
• Tier 3: Higher threshold for monitoring trends: For less impactful events, a higher threshold might be used to monitor potential trends and emerging issues.

To note, engaging with—and ensuring the active participation of—communities and seeking expert academic guidance when setting thresholds can lead to more effective solutions with greater buy-in from those directly impacted by racial disparities and disproportionalities.

For suggested reading on how to determine appropriate thresholds, see the Toronto Police Service’s Data Analysis Framework for Racial Equity and the Government of Ontario’s Data Standards for the Identification and Monitoring of Systemic Racism.

Considerations for benchmarking approaches

The purpose of PIRID (and disaggregated data more broadly) is to help to identify and resolve systemic racism by addressing issues within policies, procedures and practices of policing where they exist.

When interpreting ratios, it is important to consider that there are several drivers of under- or overrepresentation in police data that may be indicative of systemic racism. These drivers vary in terms of being internal or external to a police service, and being at either the system level or the individual level. Drivers internal to police services may include things such as individual officer bias (individual level) or institutional and cultural practices (system level). Drivers external to police services include things such as an individual’s choices and actions (individual level) and social and demographic factors (system level) (McCausland & Baldry, 2023; Ontario Association of Chiefs of Police, n.d.).

It is not possible to know what exactly causes these disproportionalities or disparities because of the complexity of these intertwining factors. In isolation, PIRID can only identify (flag) areas of concern where further investigation and intervention are needed. However, when used in conjunction with different methods and data sources, PIRID can help provide insights into systemic issues of racism within policing and the criminal justice system, and inform decision making.

When selecting analytical techniques, researchers should be mindful of their limitations. Below are some considerations for techniques where benchmarking is relevant.

Population distribution

Data users should consider the distribution of local resident population groups across different geographic areas when calculating disproportionality indices. Indigenous and racialized groups are not equally distributed across geographic areas, and large resident population benchmarks (e.g., national, provincial) are likely to mask the true extent of representation within police data. Instead, local resident population should be mapped as closely to police service delivery areas as possible. As such, disproportionality measures should use appropriate benchmark populations, with at least one that ideally represents the demographics of the area served by the police services under study. This would allow for a more accurate understanding of the representation within police data for a specific region. However, please note that only few police services have CMAs that closely align with service delivery areas so this may be a limitation to analysis to keep in mind.

Representation in decision point analysis

Decision point analysis examines the level of representation at specific points within policing data, and is considered a disparity index. Results can be used to provide insight into individual and offence characteristics that may impact outcomes, but causal statements are not supported. Further, attention must be given to geographic boundaries, since national-level indexes do not account for jurisdictional differences in police practices and reporting standards. While these measures in isolation do not provide insights into why there may be patterns of representation at different points in police-reported crime, linking to other data sources may shed light on decision making across the criminal justice process. By stringing together data from different decision points, one may be able to isolate specific decision-making points that are uniquely implicated in the outcomes experienced by certain groups.

Impact of cumulative disadvantage

While disproportionality and disparity analyses can identify where Indigenous and racialized groups may be under- or overrepresented in police-reported crime, the factors that may explain observed results cannot be isolated through these analytic techniques. Therefore, differences in outcomes cannot be attributed specifically to any one factor. Instead further work would be needed to identify causal attributes, including the consideration of aspects of policing and wider societal factors, such as those that increase the likelihood of individuals from certain communities coming into contact with police (Ontario Association of Chiefs of Police, n.d.; Text box 6).

It is recommended that caution should be taken when determining where disproportionalities and disparities exist, considering how geography, demographics, and policing legislation and practices may influence analysis. Beyond disproportionality and disparity indexes, data users should employ more rigorous methodologies to explore the various factors that mediate and moderate justice system risks and outcomes (see Text box 7).

Social determinants of justice and data linkage

Disproportionality and disparity on their own may not be conclusive evidence of systemic racial inequities (Ontario, 2018). Further analysis is necessary for determining the extent to which disproportionality and disparity may be attributed to, or influenced by, factors within the criminal justice system and external social determinants. While factors to the police sector, such as officer bias, institutional policies and organizational culture, are important, external factors such as individual behviour and choices, social conditions and demographic factors also contribute to representation (McCausland & Baldry, 2023; Ontario Association of Chiefs of Police, n.d).

To understand these complex interactions, the social determinants of justice are a valuable framework for analyzing PIRID in conjunction with other relevant datasets (Institute for Research on Public Policy, 2020; McCausland & Baldry, 2023). Social determinants of justice are social factors that shape individuals’ risk of involvement with the criminal justice system, their experiences within it and the outcomes they face. These factors are grouped into five categories: income, employment, stable housing, education and health (including mental health and addictions) (Institute for Research on Public Policy, 2020; Public Safety, 2022).^Note 21 Those who fare poorly by these measures are commonly referred to as being “at risk” because they are more likely to engage in activities or to experience precarious circumstances that lead them to encounters with the police.

By linking PIRID with datasets containing information on these social determinants, researchers can gain a more detailed understanding of the structural factors that contribute to racial disparities in police-reported crime. This can inform policy development and interventions aimed at addressing these disparities.^Note 22

Whereas disproportionality and disparity indexes are limited in what they can provide information on, these methods are simple to understand, and this is valuable for community engagement sessions. Services should combine complex and simpler analytical methods for the purposes of engagement. To create more robust statistics to analyze the role of multiple factors that may interact to result in a higher risk of being represented as a victim or an accused within police-reported crime, data users may consider employing some form of multivariate analysis (Text box 7) with data linkage, in combination with disproportionality and disparity indexes. Before performing data linkage, data users should consult with experts to ensure that their linkage method is valid. It is important to note that linking data can be seen as an intrusion of privacy and should be done with proportionality principles in mind.

Statistical standards

Statistics Canada’s (2024a) Disaggregated Data Action Plan has four guiding principles^Note 23:

• Data and analysis should be disaggregated at the lowest level of population detail possible, while respecting quality and confidentiality.
• Analysis should focus on intersectionality as opposed to binary interactions.
• Statistics Canada’s approved standards should be used for disaggregation across all programs.
• Data should be available at the lowest level of geography possible, proportional to the necessity of the data and the risk to privacy, fairness and transparency (see Statistics Canada’s Principles of Necessity and Proportionality for more information).

Small sample sizes may limit the ability to disaggregate data to the lowest level, because of confidentiality concerns. In such cases, combining categories might be necessary. However, other options should be explored first, such as expanding the scope or timeframe of the analysis to increase the overall sample size. For example, using appropriate statistical methods for small domains, such as small area estimation methods, may be considered. If combining categories is unavoidable, the combinations should be logical and meaningful, aiming to reflect any shared experiences and circumstances of the groups being combined (British Columbia Ministry of Citizens’ Services, 2023).

When combining categories, it is essential to highlight this decision and acknowledge the diversity in the Indigenous or racialized population that may be obscured. Data users should discuss the limitations of analyses based on these groupings and provide a broader context to help readers understand the potential differences that may be missed. This helps to ensure that the findings accurately represent the experiences of the various groups being studied. Data users may also consider taking the data findings back to the represented community groups to ensure insights are representative of lived experiences.

Data quality

It is recommended that to ensure high-quality analysis of PIRID, data users should adhere to the six key principles of Statistics Canada’s Quality Assurance Framework or their own institution’s data quality procedures.^Note 24

At Statistics Canada, general data quality is defined through the Quality Assurance Framework through six key dimensions:

• relevance
• accuracy
• timeliness
• accessibility
• interpretability
• coherence.

These dimensions are defined in Table 1 below, along with suggestions for how they can be applied in the context of PIRID.

Interpretation and communication of results

In this step of the systematic approach to research, data users determine the expected outcome of the study conclusions, implications for addressing the research questions, and considerations and limitations. In the context of the framework, important considerations related to PIRID include

• sufficient contextualization
• transparency and accessibility in reporting
• an action-oriented perspective
• continual and reciprocal engagement.

Sufficient contextualization

To prevent harm when using data about Indigenous and racialized communities, it is crucial to provide rich context. Data users should avoid drawing reductive conclusions without considering the historical and contemporary factors that shape current experiences, especially Canada’s colonial legacy (see Text box 8). Additionally, careful attention should be paid to acknowledging the broader systemic issues that contribute to disadvantages and increase the likelihood of interactions with the criminal justice system (British Columbia’s Office of the Human Rights Commissioner, 2020).

When considering the potential implications of publishing data, data users need to be mindful of the risks of perpetuating harmful narratives. This can be avoided by providing appropriate contextualization for the data and not presenting findings from a solely deficit-based perspective that may harm community members. Similarly, users should avoid generalizing results without considering the broader sociohistorical context that influences crime and victimization trends. Focus should be placed on articulating a balanced narrative of the data. By carefully contextualizing the findings, data users can help prevent further stigmatization and stereotyping of certain communities.

As a part of sufficient contextualization, it is also important that data users consider any relevant legislative or jurisdictional contexts that could impact policing practices and procedures. Where possible, such context should be provided to facilitate a more comprehensive interpretation of the data.

It is recommended that to be able to ensure readers have a nuanced and accurate understanding of the data, sufficient context in any analytical products should be provided.

Some examples and best practices for contextualizing results include

• publishing studies that reflect a balanced narrative of the experiences of Indigenous and racialized persons
• ensuring that a diverse range of identities and experiences from affected populations are represented in studies
• providing relevant contextual information to support the interpretation of the analysis, including
• historical and current experiences of systemic racism and discrimination
• impacts of colonialism on communities of interest
• community activism, including efforts to self-advocate and address concerns reflected in the data
• acknowledging and providing insights on the diversity among Indigenous and racialized groups, employing an intersectional lens to understand how multiple identities, geographies and cultures may influence individual experiences
• avoiding making generalizations about entire groups
• reporting data on Indigenous and racialized persons only when they are relevant and can be adequately contextualized
• reporting on any known legislative context of policing practices and procedures that may vary across jurisdictions and influence research findings
• engaging in community engagement to better understand the local sociohistorical context that may help explain, or at least, position in findings.

Transparency and accessibility in reporting

To ensure transparency and accessibility in PIRID products, data users should provide detailed information about the methodologies used in PIRID analysis and any limitations, including the following:

• Units of analysis: Clearly define the main units of analysis, what they measure, and how they were categorized and used.
• Analytical methodologies: Explain any complex or unique methods used and why they were chosen.
• Benchmarks and reference groups: Specify the benchmarks and reference groups used in the analysis, and why they were chosen.
• Thresholds: Clearly state the thresholds used to identify disproportionalities or disparities, and why they were chosen.
• Data quality and accountability: Describe the steps taken to ensure data quality and confidentiality. For more information, refer to the guidelines for ensuring data quality.
• Limitations: Acknowledge any data limitations and provide sufficient contextual and foundational information to the reader, whenever possible. This includes acknowledging all known limitations of the data and methodologies used and how they might affect the interpretation of findings.

An action-oriented perspective

Data users should consider the policy implications of their research using PIRID from the outset, keeping actionable items at the forefront of their data analysis where appropriate or relevant.^Note 25

It is recommended that based on the results of data analysis, data users should consider the ways findings can be used to support communities and how they can be operationalized by program and policy makers for social change.

When focusing on the action-oriented nature of research, data users may consider the factors that contribute to relevance (see Table 1).

According to Statistics Canada’s mandate and fundamental principles of official statistics, the information delivered must meet the test of practical utility—that is, it must be relevant to the population of Canada. Information may be considered relevant when responding to one or more of the following points:

• Does the analysis or data provide solid grounds that can be operationalized by program and policy makers for social change (actionable items) that can benefit communities?
• Does the analysis or data provide information that addresses research questions raised by other departments or stakeholders (e.g., cost-recovery projects)?
• Does the analysis or data provide information that is new (added value)?
• Does the analysis correspond to a legislative requirement?
• Does the analysis promote transparency?
• Does the analysis answer a question, or meet a need, identified though community engagement initiatives?

Continual and reciprocal community engagement

Building trust and establishing strong relationships through engagement are paramount, helping to ensure that data analyses accurately reflect the voices, concerns and lived experiences of the communities represented. Ultimately, this collaborative approach leads to more valuable and impactful analytical products and demonstrates respect to involved communities.

When resources permit, data users are encouraged to establish formal feedback mechanisms, such as community advisory panels. These panels can provide ongoing review and input throughout the data analysis process. Through ongoing engagement, data users can do the following:

• Amplify community voices: Reflect the data from the perspective of the communities being represented, empowering these communities.
• Ensure cultural sensitivity: Conduct analysis and interpretation in a manner that aligns with community values.
• Create a balanced narrative: Examine challenges while also highlighting strengths and understanding the population of interest.

Text box 3 outlines Statistics Canada’s practices related to Indigenous community engagement. In the context of interpreting and communicating results, these engagements can provide data users with the tools to do the following:

• Apply key Indigenous research ethics and methodologies on the advice of and through collaboration with Indigenous partners.
• Discuss practical aspects of community engagement with the research and the presentation of findings, such as language.
• Data users should be mindful that language can be used as a tool to reproduce or reinforce social inequities. Wording is very important, and stigmatizing language can perpetuate harmful stereotypes; therefore, caution should be taken.
• Present analysis in a way that is easy to understand and resonates with the community at its core.

For example, using Indigenous terminology in Indigenous research is important to ensure cultural respect and accurate representation, as well as to build trust between Indigenous communities and the research community.

Conclusion

This framework was created as a component of the PIRID initiative to help guide data users upon the release of the newest version of the UCR Survey (UCR 2.5). This version has been updated to include the Indigenous and racialized identity of accused persons and victims of reported criminal incidents in Canada. The framework was initially created based on engagement feedback regarding the need to support data users with the responsible and ethical use, interpretation, analysis and dissemination of these data to help safeguard against harm to affected communities caused by data use. The intended audience of this document is UCR Survey data users who plan to use the new Indigenous and racialized identity data, in any format, at any point in the data cycle, and all of the derivatives of these data, including custom and published data.

The framework has two main goals:

1. Support the responsible and ethical use of PIRID by proposing five guiding principles, which are intended to help avoid further stigmatizing and marginalizing communities through the use of these data.
2. Equip data users with tools and guidance for careful and culturally competent data interpretation, ultimately contributing to the development of evidence to support decision making for the creation of more equitable outcomes in policing.

Annexes

Annex A: New variables for the Uniform Crime Reporting Survey (UCR 2.5)

Uniform Crime Reporting (UCR) Survey version 2.5 extraction specifications

Effective February 13, 2024, the Canadian Centre for Justice and Community Safety Statistics (CCJCSS) is system ready to begin receiving a new version (v2.5) of the Uniform Crime Reporting (UCR) survey. This document provides information related to the standardized v2.5 format for data submission that is compatible with CCJCSS’s UCR processing system. We acknowledge that information related to operational guidelines related to the collection of Indigenous and racialized identity information is required to modify the records management systems used by police services. The work to determine operational guidelines, data standards and best practises is being undertaken by the Canadian Association of Chiefs of Police (CACP) Special Purpose Data Standards Working Group and the outcome of this work will be shared in the upcoming months. It is recommended that RMS updates do not take place until after these guidelines have been established and communicated.

Overview of changes

UCR v2.5 includes the addition of 13 new variables present on the victim file and 13 new variables present on the Charged/Suspect-Chargeable (CSC) file related to indigenous and racialized identity (population group). These variables include 2 officer perception variables (1 general and 1 detailed) and self-identification variables where up to 10 options can be selected, in addition to an open write-in field that can be filled in, as needed. In addition, a few modifications have been made to existing survey variables. The charges laid or recommended section variables (SECTION1, SECTION2, SECTION3 and SECTION4) on the CSC (Charged/Suspect-Chargeable) file have been increased from 6 spaces to 7 spaces to accommodate Criminal Code sections that are 7 spaces long (including period) ex: Criminal Code Section 320.102. Finally, the variables CSCSEX and VICSEX have been renamed to CSCGEN and VICGEN to be consistent with the change to gender in 2018. There are no changes to valid options or any other details pertaining to collection of gender data; this is simply a variable name update which may not impact Records Management Systems.

General information regarding the collection and reporting of Indigenous and racialized identity data through UCR v2.5

Police services adopting UCR v2.5 will have the flexibility to report data to the UCR based on either the officer-perception method, self-identification method, or both. Prior to making changes to police records management systems to enable reporting to the UCR, it should be determined which method of collection a police service is choosing to use (“opt-in” option). Reporting a valid response is mandatory regardless of collection method, according to the specifications outlined below.

• Officer perceived collection method: For police services opting to report Indigenous and racialized identity data collected using the officer perception method, selecting a population group is mandatory in the UCR, and only one selection is possible. See section below for detailed information regarding the officer-perception variables.
• Self-reported collection method: For police services opting to report Indigenous and racialized identity data collected using the self-identification method, reporting a valid response is mandatory, including the option to select ‘Unable to self-identify’ or ‘Refusal to self-identify’. Up to 10 selections are possible for this collection method, in addition to an open write-in field. See section below for detailed information regarding the self-reported variables.

Note regarding the alignment of the UCR v2.5 with other existing jurisdictional standards

Statistics Canada acknowledges and commends the excellent work that has and is being done in other jurisdictions, like Ontario, related to the development of standards and collection of race-based data to inform public sector policies and programs. We also recognize that other police jurisdictions across the country may be adopting or considering Ontario’s approach or wanting to explore a different path to collecting police-reported Indigenous and racialized identity data. The specifications and standards developed for the UCR v2.5 are aligned with standards adopted by jurisdictions currently collecting some form of race-based data, including Ontario’s Data Standards for the Identification and Monitoring of Systemic Racism. Statistics Canada will continue to work with police services and their Records Management System vendors to map the UCR v2.5 with standards being used or adopted.

Victim and CSC officer perceived population group (general and detailed):

Victim and CSC Officer Perceived Population Group is comprised of two variables: Officer Perceived Population Group (General) and Officer Perceived Population Group (Detailed). These two variables are present on both the victim file and CSC file. Officer perceived population Group (General) and Officer Perceived Population Group (Detailed) cannot conflict. Valid values and combinations are listed in the table below. For example, if Officer Perceived Population Group (General) is 10 ‘Indigenous’ then Officer Perceived Population Group (Detailed) must be either 11 ‘First Nations’, 12 ‘Métis’, 13 ‘Inuk (Inuit)’ or 00 ‘None selected/not specified’. Options presented in grey below are for RMS background programming use only and are not intended to be visible to officers as an option for selection. A valid population response is required. This is explained in more detail later in this document.

Officer Perceived Population Group (General)

Officer Perceived Population Group (Detailed)

Specifications for police services opting out of collecting victim and CSC officer perceived population group detailed and general Variables:

Should a police service not be in a position to collect Officer Perceived Population Group Detailed and  General Variables on the victim or CSC record (i.e., the police service does not intend to collect population group data using the officer perception method), then code 98 ‘Not collected by police service’ should be automated for both Officer Perceived Population Group Detailed and General Variables within the RMS for both variables for all records. A police service can only opt out of both variables; not one or the other.

Victim and CSC self-reported population group and corresponding open field:

The Victim and CSC Self-Reported Population Group is comprised of 10 identical variables with the same list of valid options and 1 corresponding open field that can be used if the victim/accused would like to specify a population group not included in the list of options.

Self-reported population group variables:

The Self-Reported Population Group Variables allow victims and CSCs to self-report up to 10 different population groups. The same code cannot be selected more than once. Included are options for situations where victims or CSCs are unable to provide this information or refuse to self-identify. Like the Officer Perceived Population Group Options presented above, the options in grey below are for RMS background programming only and are not intended to be visible as options for selection.

Self-Reported Population Group (1-10)

Self-Reported Population Group Corresponding Open field:

The corresponding open field write-in variable allows the victim/CSC to enter the Self-Reported Population Group of their choosing instead of, or in addition to, the picklist provided. There are no limitations on what can be written in this field but there is a maximum of 80 characters/numbers. This variable is present on both the victim file and CSC file.

Self-Reported Population Group Corresponding Open field

Source: Statistics Canada, Population group of person

There is no specific picklist for this variable as it is an open field write-in variable.

If this variable is used, one of the Self-Reported Population Group (1-10) variables should be ‘090 – Other; please specify’. This variable should be left blank if not used.

Specifications for police services opting out of collecting victim and CSC self-reported population group variables and corresponding open field:

Should a police service not be in a position to collect Self-Reported Population Group Variables and Corresponding Open field on the victim and/or CSC record (i.e., the police service does not intend to collect population group data using the self-identification method), then code 998 ‘Not collected by police service’ should be automated within the RMS for VSELPPG1/CSELPPG1. The remaining 9 variables should be automated to 999 ‘Not applicable’ and the corresponding open field should be left blank.

General requirements for 2.5 variables before and after police service’s 2.5 transition date

The new v2.5 population group variables outlined above are required to be populated following the police service’s UCR 2.5 transition date. However, these variables may be populated with “not collected by police service” if a police service has decided as an organization to opt out of collecting either officer perception or self-identification variable groups. Once a police service has transitioned to v2.5, it is optional to provide population group data for records reported prior to the transitioning. For example, if a person is charged after the police service’s transition date in relation to an incident reported prior to the transition date, the police service may be able and decide to provide population group data for the CSC. CCJCSS will accept population group data on the CSC and/or victim record for incidents reported prior to transitioning to 2.5 but this is not required. 

V2.5 Record Layout

Please see accompanying document titled: UCR v2.5 Record Layout_ENG for the new v2.5 record layout.

V2.5 UCR Manual Pages

The UCR manual pages are not yet available and will be shared once the operational guidelines, data standards and best practises have been developed with the support of the Canadian Association of Chiefs of Police (CACP) Special Purpose Committee.

Annex B: Systematic approach to research

A top recommendation from the Analytical Insights Working Group (AIWG) of the Disaggregated Data Action Plan (DDAP)

Background

In social, health, and economic research, an estimate for the total population of Canada is far from telling the whole story. By disaggregating data—that is, by breaking down summary statistics into more refined categories—we can uncover important differences within different groups of people. Most research topics can be readily disaggregated; however, there are instances where disaggregation has less relevance. Some analytic concepts may not lend themselves well to disaggregation. For example, while disaggregation may not be meaningful for a concept overview, few disaggregation may be appropriate in research products with new data and more disaggregation may be needed for analytical studies. In some cases, we may have limited experience and must first make stronger connections to the communities and groups we would like to write about. Overall, the level of disaggregation afforded by analytical research will depend broadly upon the nature of the research.

Disaggregated data analysis is not inherently complex. In fact, even complex analytical studies start with a good understanding of the data, descriptively. In all cases, disaggregated data analysis needs to be guided by a research question, informed by previous work and an analytical plan, and be purposeful throughout all the steps of a systematic approach to research.

Objective

The systematic approach to research (Figure 1) emphasizes that disaggregated data analysis is not a fishing expedition, nor is it about performing many cross-tabulations until one finds something statistically significant. Rather, disaggregated data analysis, like any rigorous analysis, must follow a series of systematic checkpoints. The purpose of this document is to provide a detailed description of the systematic approach to research (Figure 1), which can be used for a variety of purposes, including but not limited to providing training, giving presentations, and conducting research and analyses.

The AIWG acknowledges that analytical insights feed into all stages of the data development and analysis continuum and recommends using this approach and leverage all resources for thoughtful and ethical analysis of population data with a disaggregated data lens, where appropriate. Following a systematic approach to research facilitates identification of information needs, analytical gaps and relevant analytical questions, in addition to promoting engagement and communications throughout all stages. This approach is in line with the Necessity and Proportionality Framework that takes into account ethical considerations, such as privacy, fairness, and transparency.

The next section provides a detailed description of the systematic approach to research and focuses on best practices and guidelines throughout each step under this approach. Readers are also encouraged to view this presentation^Note 26 , which was offered at the Best Practices in Ethical Data Analysis Workshop organized by McGill University and the University of Waterloo.

Systematic approach to research

First of all, any rigorous analysis of data starts with four fundamental questions that researchers and analysts have to pose and answer. These are, What, Why, How, and So what.

• What? Is the issue. What problem or phenomenon are we investigating? What is the research question?
• Why? Is the relevance and necessity. Why is this issue important to study and necessary to address?
• How? Is the methodology. What data source(s) and which type(s) of analysis will I use to address this issue, while respecting considerations such as quality, privacy, and confidentiality?
• So what? Is the outcome. What are my conclusions, and who will benefit from this research? What are the limitations?

1. Issue Identification (What)

The first step of the systematic approach to research is the identification of an issue, the statement of a problem, and subsequently, the articulation of a clear purpose and a specific question or set of questions that need to be addressed. Formulating research questions is an essential part of sound and rigorous research. In this first step, the general analytic approach can be determined, such as descriptive versus correlational (e.g., multivariate) analysis or testing new variables for data development.

2. Conceptual Framework (Why)

Every sound research includes a literature review driven by research questions to understand what we already know about the topic and the important context to consider. What has been previously done in this area and what are the information needs or gaps? What are the key variables and relationships of interest? Which groups are most impacted by the issue and why?

The use of standards as well as geo-enabled data ensures consistent and coherent definitions for variables regardless of the data source researchers and analysts use. Both standards and geo-enabled data should be considered in the conceptual framework.

Overall, this step allows to verify whether the identified issue or the stated problem is necessary and relevant to be investigated, and it also helps develop or select a conceptual framework that is research-question driven and that provides an overall representation of the relationships among key variables about the topic. The latter facilitates the development of specific hypotheses (e.g., expected results) about the research questions.

In this step, it is essential to determine the nature of the analysis (e.g., descriptive or modelling or data development) as well as the level of disaggregation and intersectionality among the chosen key variables. The level of disaggregation and dimensions of intersectionality should be driven by engaging with implicated communities, as well as the purpose of the research. The nature of the research will also help with the choice of analytical products to reach various audiences.

3. Data and Analysis (How)

The third step focuses on methodological and analytical choices to address the research questions. In this step, various data sources are considered and the most appropriate and useful data source(s), given the purpose, research questions and the conceptual framework, are selected. Important checkpoints during this step include verifying if you have sufficient sample size to support the planned disaggregated analysis while respecting data quality, confidentiality, and statistical power.

With all the information gathered up to now, it is essential to pause and make an analytical plan, before diving into the analyses. The plan should include the purpose and objectives of the research, clear and concise research questions, a description of the data and variables to be used for analysis, and the planned methodology and statistical methods. Engaging stakeholders by sharing the analytical plan for feedback is an essential checkpoint during this step. We want to make sure that the planned research is appropriate, relevant, and sound.

After finalizing an analytical plan, whether any data development or integration is needed along with a feasibility study prior to a series of analytical products can be determined.

4. Interpretation and Communication of Results (So What)

The final step of the systematic approach to research involves appropriate, meaningful, and useful interpretation and communication of the results. Good communication of research findings, especially if based on more complex methods, always involves highlighting the key messages that matter most given the audience. The value-added of the research should be re-emphasized. In addition, a discussion on the implications of the findings along with a consideration of strengths and limitations (e.g., generalization, quality of disaggregated results) of the research is essential. Directions for future research and analysis may also be warranted in light of the results and conclusions.

We may choose to create and disseminate multiple products for the same analysis (e.g., an analytical report and an Infographic) in order to reach a wide variety of audiences.

5. Feedback loop

The feedback loop shown from the last step (So what) back to the first step (What) indicates that conclusions will always spur new issues and questions. Thus, the feedback loop reinforces the connection between research outcomes and new ideas.

6. Engagement and Communications

An essential element of the systematic approach to research is continuous engagement and communications with relevant partners and stakeholders, including people with lived experiences, throughout the entire process. Engagement and communications should occur as early as possible, ideally at the issue identification stage, and should be ongoing. Especially important checkpoints for review and feedback are the planning stage and before dissemination. Similar to analytical insights, engagement and communications feed into all stages of the data development and analysis continuum. It is an essential ongoing part of systematic approach to research, which may vary in levels given the desired opportunities for influence and involvement in research.

Top challenges

The AIWG acknowledges that there are various challenges that remain despite the best practices and guidelines under this approach. The top challenges include but are not limited to engagement, internal communications, ethics, terminology, reference groups, and sample size. The AIWG also recognizes the importance of capacity building, which can provide a new forum to identify and address barriers for conducting disaggregated and intersectional analyses.

Analytical capacity building

A pilot training in disaggregated data analysis has been developed to (1) provide participants with an understanding of the nature and importance of disaggregated data analysis, and an appreciation of disaggregated data analysis considerations at each step of the analytical process, (2) provide guidance to analysts who are being asked to disaggregate data and are currently unfamiliar with statistical standards, (3) understand methodological considerations, including data confidentiality and small sample sizes, as well as available tools and solutions, and (4) understand the important ethical considerations in disaggregated data analysis. The AIWG’s top recommendation of a systematic approach to research is the backbone of this analytical capacity training.

Annex C: Data user checklist

This checklist is a supplement to the Framework for the use, analysis, interpretation, and dissemination of police-reported Indigenous and racialized identity data (PIRID) through the Uniform Crime Reporting Survey (“the Framework”). Data users are expected to have read the Framework. This checklist aims to provide a high-level overview of the key elements of the Framework and, as such, can be consulted or used upon completion of a given project/research involving the data.

Five guiding principle:

• Community engagement
• Analytical rigour and an intersectional perspective
• Harm prevention
• Transparency and accessibility
• Action-oriented research

• The data user has read the Framework for the use, analysis, interpretation, and dissemination of police reported Indigenous and racialized identity data through the Uniform Crime Reporting Survey.

ISSUE IDENTIFICATION

• Concrete measures and mechanisms related to community engagement have been put in place to ensure adequate community participation and implementation of feedback from the beginning to end of the project. This may include:
•  Outreach to community leaders, gatekeepers, experts in the field, or relevant organizations to determine how best to engage with communities;^Note 27
• Engaging established community advisory groups.
• Project or research objectives were created considering feedback from community members represented by the data.

CONCEPTUALIZATION

• Selection of UCR units of analysis align with project objectives.
• How the UCR variables are operationalized for analysis is discussed in the report.
• Where both the officer perception and self-identification data are presented in the study, the variables were analyzed separately (not combined or compared).
• Where applicable, guidelines for interpreting multiple self-identification variables were followed.
• Methods of grouping or aggregating categories were disclosed.
• Robust safeguards were contemplated and put in place to prevent harm to community members represented by the data being used. This includes:
• Community engagement to support ethical use of the data.^Note 24
• Mitigating stigma and profiling through avoiding language that reinforces stereotypes, profiling practices or justifying discriminatory policies o practices.
• Reinforcing the social construct of race and racialization.
• Rich contextualization including the incorporation of historical and contemporary contexts, and considerations for unique circumstances and experiences.
• Considerations for one’s cultural competencies and acquiring necessary training.
• Awareness of one’s unconscious biases and acquiring any necessary training.
• An intersectional lens was applied, wherever possible and relevant, including considerations for:
• Indigenous identity
• Racialized identity
• Gender
• 2SLGBTQA+ identity
• Geography
• Age
• Other external mitigating factors that may coincide with a specific Indigenous or racialized identity.
• Data has not been used to delegitimize support to communities.
• Data has not been used for the oversight or evaluation of individual police personnel, or Indigenous and racialized communities.

DATA AND ANALYSIS

• Appropriate analytical techniques were chosen considering their limitations, including:
• Population distribution
• Representation in decision point analysis
• Impact of cumulative disadvantage
• Where possible, factors impacting data collection were considered in the analysis and interpretation of results, including:
• Geography
• Demographics
• Policing practices
• Effective methods were utilized in choosing benchmarks, including:
• Relevance to the population and objective of project
• Multiple benchmarks were used
• Sufficient context is provided related to the benchmarks used
• Appropriate thresholds were identified in a manner that balances sensitivity with specificity (multi-tiered system)
• Community participation and/or expert academic guidance was sought^Note 24
• Considerations for the social determinants of justice were employed where possible through:
• Use of multivariate analysis and data linkage
• Statistics Canada’s Disaggregated Data Action Plan guiding principles were followed in determining appropriate analytical standards.
• Quality of data and analysis adhere to Statistics Canada’s Quality Assurance Framework.

INTERPRETATION AND COMMUNICATION OF RESULTS

• Information about methodologies used and their limitations were communicated.
• Sufficient contextualization was provided.
• Assessments of how results can be used were represented, including how they contribute to relevance under the Statistics Canada Quality Assurance Framework
• Feedback was sought from communities represented by the data on the final analytical product.

References

Boilevin, L., Chapman, J., Deane, L., Doerksen, G. F., Joe, D., Leech-Crier, N., Marsh, S., McLeod, J., Neufeld, S., Pham, S., Shaver, L., Smith, P., Steward, M., Wilson, D., & Winter, P. (2018). Research 101: A manifesto for ethical research in the Downtown Eastside

British Columbia Ministry of Citizens’ Services. (2023). Guide on using categorical race & ethnicity variables.

British Columbia’s Office of the Human Rights Commissioner. (2020). Disaggregated demographic data collection in British Columbia: The grandmother perspective.

Buchanan, N. T., Perez, M., Prinstein, M. J., & Thurston, I. B. (2021). Upending racism in psychological science: Strategies to change how science is conducted, reported, reviewed, and disseminated. The American psychologist, 76(7), 1097–1112.

Canada School of Public Service. (n.d.). What is Indigenous data sovereignty?

Canadian Heritage. (2019). Building a foundation for change: Canada’s Anti-Racism Strategy 2019-2022.

Choe, C., Jung, S. E., & Oaxaca, R. L. (2020). Identification and decompositions in probit and logit models. Empirical Economics, 59(3), 1479-1492.

Clark, S. (2019). Overrepresentation of Indigenous people in the Canadian criminal justice system: Causes and responses. Department of Justice Canada.

Commission on Systemic Racism in the Ontario Criminal Justice System. (1995). Report of the Commission on Systemic Racism in the Ontario Criminal Justice System.

Correll, J., Park, B., Judd, C. M., & Wittenbrink, B. (2002). The police officer’s dilemma: Using ethnicity to disambiguate potentially threatening individuals. Journal of Personality and Social Psychology, 83(6), 1314-1329.

Crenshaw, K. (1989). Demarginalizing the intersection of race and sex: A Black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. University of Chicago Legal Forum, 1989(1), 139-167.

Crown-Indigenous Relations and Northern Affairs Canada. (2021). Indigenous peoples and communities. Government of Canada.

Curtis, E., Jones, R., Tipene-Leach, D., Walker, C., Loring, B., Paine, S., & Reid, P. (2019). Why cultural safety rather than cultural competency is required to achieve health equity: a literature review and recommended definition. International Journal for Equity in Health, 18(1). 

David, J. D., & Mitchell, M. (2021). Contacts with the police and the overrepresentation of Indigenous peoples in the Canadian criminal justice system. Canadian Journal of Criminology and Criminal Justice, 63(2), 23-45.

Donoghoe, M. W., & Marschner, I. C. (2015). Flexible regression models for rate differences, risk differences and relative risks. International Journal of Biostatistics, 11(1), 91-108.

First Nations Information Governance Centre. (n.d.). The OCAP® Principles.

Foster, L., Jacobs, L. A., Siu, B., & Azmi, S. (2018). Racial profiling and human rights in Canada: The new legal landscape. Toronto, ON: Irwin Law.

Foster, L., Park, S., McCague, H., Fletcher, M., & Sikdar, J. (2023). Black Canadian national survey: Final report 2023. Institute for Social Research, York University.

Government of Canada. (2023). 2023–2026 Data Strategy for the Federal Public Service.

Government of Canada. (2024). Government of Canada’s approach on Gender-based Analysis Plus. Canada.ca.

Government of Canada, Statistics Canada. (2022). The Daily — Indigenous population continues to grow and is much younger than the non-Indigenous population, although the pace of growth has slowed. 

Hart, M. A. (2021). Indigenous worldviews and knowledge. In M. A. Hart, S. J. Rowe, & A. E. Sinclair (Eds.), Indigenous research: Theories, practices, and relationships (pp. 3–24). Springer Publishing Company.

Hernández, T. (2018). Multiracials and civil rights: Mixed-race stories of Discrimination.

Indigenous Services Canada (2022). Engagement guide: Co-developing federal distinctions-based Indigenous health legislation. Government of Canada.

Indigenous Services Canada. (2023). Visions for distinctions-based Indigenous health legislation: Executive summary. Government of Canada.

Institute for Research on Public Policy. (2020). Rethinking Criminal Justice in Canada.

Inuit Tapiriit Kanatami. (2018). National Inuit Strategy on Research.

Mazowita, B., & Rotenberg, C. (2019). The Canadian Police Performance Metrics Framework: Standardized indicators for police services in Canada.

McCausland, R., & Baldry, E. (2023). Who does Australia lock up? The social determinants of justice. International Journal for Crime, Justice and Social Democracy, 12(3), 37-53.

Melvin, A. (2023). Postsecondary educational attainment and labour market outcomes among Indigenous peoples in Canada, findings from the 2021 Census. Insights on Canadian Society (Catalogue no. 75-006-X). Statistics Canada.

Millar, P., & Owusu-Bempah, A. (2011). Whitewashing criminal justice in Canada: Preventing research through data suppression. Canadian Journal of Law and Society, 26(3), 653-661.

Mittinty, M. N., & Lynch, J. (2023). Risk ratio regression—simple concept yet complex computation. International Journal of Epidemiology, 52(1), 309-314.

National Inquiry into Missing and Murdered Indigenous Women and Girls. (2019). Reclaiming power and place: The final report of the National Inquiry into Missing and Murdered Indigenous Women and Girls.

Ontario. (2018). Data Standards for the Identification and Monitoring of Systemic Racism.

Ontario Association of Chiefs of Police. (n.d.). Approach to benchmarking racial disproportionalities for policing insights. Unpublished.

Ontario Human Rights Commission. (2003). Paying the Price: The Human Cost of Racial Profiling.

Ontario Human Rights Commission. (2005). Policy and guidelines on racism and racial discrimination.

Ontario Human Rights Commission. (2018). Under Suspicion: Issues Raised by Indigenous Peoples.

Owusu-Bempah, A., & Jeffers, S. (2021). Black Youth and the Criminal Justice System: Summary Report of an Engagement Process in Canada. Ottawa: Department of Justice Canada.

Owusu-Bempah, A., & Jones, Z. (2023). Canada’s Black Justice Strategy: Framework.Department of Justice Canada.

Pedneault, C., Lee, S. C., & Jones, N. J. (2024). Reconvictions among adults sentenced to custody or community supervision across five provincial correctional programs, 2015/2016 to 2018/2019. Juristat (Catalogue no. 85-002-X). Statistics Canada.

Public Safety. (2022) Federal Framework to Reduce Recidivism.

Rossiter, J., & Ndekezi, T. (2021). Confronting racism with data: Why Canada needs disaggregated race-based data. Edmonton Social Planning Council.

Royal Commission on Aboriginal Peoples. (1996). Report of the Royal Commission on Aboriginal Peoples Vol. 1-5.

Samuels-Wortley, K. (2021). To serve and protect whom? Using composite counter-storytelling to explore Black and Indigenous youth experiences and perceptions of the police in Canada. Crime & Delinquency, 67(8), 1137-1164.

So, N., Price, K., O’Mara, P., & Rodrigues, M. A. (2024). The importance of cultural humility and cultural safety in health care. The Medical Journal of Australia, 220(1), 12–13.

Statistics Canada. (2021a). Population group of person.

Statistics Canada. (2021b). Criminal victimization in Canada, 2019.

Statistics Canada. (2022a). Report and draft recommendations: Police-reported Indigenous and racialized identity statistics via the Uniform Crime Reporting Survey.

Statistics Canada. (2022b). Visible minority and population group reference guide, Census of Population, 2021.

Statistics Canada. (2024a). Disaggregated Data Action Plan.

Statistics Canada. (2024b). Visible minority concept consultative engagement. 

Statistics Canada. (2024c). Black History Month… by the numbers.

Toronto Police Service. (2022). Data Analysis Framework for Racial Equity.

Treasury Board of Canada Secretariat. (2023). 2023-2026 Data Strategy for the Federal Public Service.

Tregle, B., Nix, J., & Alpert, G. (2018). Disparity does not mean bias: Making sense of observed racial disparities in fatal officer-involved shootings with multiple benchmarks. Journal of Crime and Justice, 42, 18-31.

Truth and Reconciliation Commission of Canada. (2015). Truth and Reconciliation Commission of Canada: Calls to Action.

United Nations. (2018). A human rights-based approach to data: Leaving no one behind in the 2030 Agenda for Sustainable Development.

Viano, S., & Baker, D. J. (2020). How administrative data collection and analysis can better reflect racial and ethnic identities. Review of Research in Education, 44(1), 301-331.