Wherever data shapes decisions and drives change, equity must remain at the forefront. At Mathematica, we’re committed to the use of equitable and culturally responsive practices in data collection, interpretation, and use. Here are some tips to help you infuse equitable data collection practices into your work.
Data processing consists of a sequence of stages called the data lifecycle. From creation through eventual disposal, the stages include context-setting, planning, collection, access, analysis, and reporting. Our Education to Workforce Indicator Framework identifies seven core data equity principles to incorporate equity throughout each of these stages. The principles, developed in partnership with Mirror Group, and examples of how to apply them in practice follow.
Evaluate data practices to determine whether they might contribute to greater equity, merely reinforce the status quo, or cause harm to marginalized communities. Ensure that community members are involved in data governance, institutional review, and advisory structures.
Minimize the collection of sensitive and personally identifiable information unless it is critical to achieving the project’s intended benefits.
Clearly describe the methods and algorithms used to analyze the data. Acknowledge the extent to which the methods and algorithms might be inaccurate and biased and specify how they will be used to inform decision making.
Understand and acknowledge that data represent peoples’ lived experiences, protect data from improper use and exposure, and return the data to community partners to promote equity and earn public trust.
Develop a list of data elements to be collected and any linked data sets. Decide how you will store data, who will have access to data, how you will use data and for how long, and what you will do with the data after analysis is complete. Decisions about data access can be made by a governance body that represents individuals who provide their data, which should include local leaders who represent affected communities.
Disclose data privacy and security processes to participants when collecting data.
Store data in a secure location that is only accessible to authorized users.
Disparities in individuals’ experience might be hidden in aggregate data. Therefore, data analysis might require multiple levels of disaggregation to capture intersections among individuals’ lived experiences.
Work with community members to determine which characteristics to measure during data collection or to link into the data (if already available), and how to label these characteristics in collection tools as well as eventual reporting (for example, Hispanic, Latino/a, Latinx).
Disaggregate both outcome and systems data at multiple levels to illuminate disparities.
When reporting disparities by subgroup, connect them to systems and root causes, not people.
Understanding the social and historical contexts and root causes of disparities, including how structural conditions, past policies, programs, and institutions might have promoted or perpetuated racial inequities, is critical to developing solutions to address them.
Identify key historical events, policies, and processes that provide context for observed present-day disparities. You can conduct a historical analysis through an equity audit, an environmental scan, or organizational reflection, such as a visual timeline activity that maps trends in outcome data and compares them with policies and other changes over time.
When conducting root cause analyses, vet research questions and data collection plans with the groups of people most affected by the identified problem. Seek community reactions to, and interpretation of, findings to illuminate root causes that might not have surfaced during analysis.
One way to guard against inadvertently reinforcing historical biases, deficit narratives, and power imbalances is to question default methods for data collection and analysis. Gathering multiple sources and types of information can help counter the bias in any one data source.
At the outset of a data project, conduct an implicit bias test or group reflection activity among the proposed team to identify individual and institutional biases and discuss ways to mitigate those biases throughout the project life span. To increase cultural competency, learn about the history, power structures, and systematic barriers that exist in priority communities, as well as the community’s past experiences with data collection efforts.
Ensure data teams reflect diverse lived experiences, particularly those experiences being studied by the data project.
Carefully consider whether findings reinforce negative stereotypes or deficit narratives.
When considering aspects of data visualizations—including labels, coloring, order of appearance, and type of graphics and icons—ensure that they are accessible to multiple audiences and do not reinforce stereotypes and deficit narratives.
Build teams with people who have a diversity of lived experiences to decrease the likelihood that implicit bias might appear in data visualizations.
As stated in all of the principles, engaging community members who have lived experience is key to centering equity throughout the data life cycle. Data users should follow best practices for effective community engagement.
Identify what you mean by “priority communities”— that is, who is directly and indirectly affected by the issue being focused on. Do not assume that racial, ethnic, or socioeconomic diversity within a group will provide lived experience that is relevant to the project. Collaborate with community members to decide on the key issues and which perspectives to prioritize.
Recruit members of priority communities to participate in initiative teams or advisory councils.
Add depth to findings through anecdotal and contextual information gleaned from the lived experiences of community members.
Addressing power imbalances within your data collection teams will enhance the reliability and usability of the data collected. To make data collection teams more authentically inclusive, consider the following questions:
Grounded in the principles of simplicity and effectiveness, this resource provides a foundational guide for countries seeking to foster equity through their data practices—even those with resource constraints.
Read MoreLearn more about how to think about these questions and how to make instrument design, staff training, and participant recruitment practices more inclusive so that your methods reflect the perspectives of people most affected by the research.
Learn more about our data analytics solutions that address all the phases of the data lifecycle.