One way to mitigate personal biases is to use data as a tool to help define the problem and find solutions. Data, however, is not immune to bias; recognizing its limitations and using it as a tool rather than absolute fact can help apply an equity lens.
Applying an equity lens pertains to all choice points in your data process. From what you collect to how you analyze and disaggregate the data and how you present it all affects how decisions are made in response. The right data allows for a focus on eliminating disparities and supporting solutions so that all people are allowed equitable opportunities to reach their full potential. Inequities exist due to a myriad of systemic social, economic or environmental factors which should be constantly brought forth through the data processes. By applying an equity lens to data gathering and dissemination, you can help identify, assess, and reflect on the potential impact a solution has or the impact a solution already underway is having on communities with different needs in their social context.
- What Do the Numbers and Text Really Mean: Using Data to End Health Disparities and Strengthen Communities is a publication by Community Science for the National Partnership for Action to End Health Disparities. The guide is designed to help users understand how an equity lens can be applied to creating a data-driven strategy.
Disaggregating data creates the opportunity to mitigate negative impacts, enhance positive impacts, and (ideally) prioritize solutions for the populations that may need it most. In some cases, trends can mask disparities, or hide the fact that some groups are struggling even though the population at large is making progress. Put another way, solutions cannot be community-centric if don’t pay attention to community nuances; if you don’t know who you’re impacting, how do you know you’re making an impact? In addition, disaggregating data can help frame larger conversations about the need for solutions that are responsive to the needs of different racial, ethnic, and socioeconomic groups, and possibly the need to redirect services or funding to priority populations.
Disaggregated data helps you answer different questions than data that stays at a full population level. For example, instead of asking:
- What is the level of student achievement in our school?
You can ask:
- Is there an achievement gap among different demographic groups? Is the gap getting bigger or smaller?
- Are students at some grade levels or in some classrooms doing better in core subjects?
- Are lower income students overrepresented in special-education classes or underrepresented in gifted and talented classes?
Sometimes disaggregated data is available just by asking. For example, if you plan to leverage an existing database, such as a health epidemiology database, education achievement database, or patient records database, it can be helpful to explain the types of questions you hope to answer to ensure you get the data at the level you need to answer them. Other times, databases that already exist are either lacking key demographic information or were collected with samples too small to disaggregate. In these cases, you may need to collect additional data to supplement what is more easily available.
- This post by Race Matters Institute outlines a series of questions to ask when disaggregating data.
- Hiding in Plain Sight via the Huffington Post looks at the importance of disaggregating education data, and how the StrivePartnership used disaggregated data as a baseline for improvement.
- Data-driven Equity in Urban Schools provides information on how data can help enhance equity in education, including a look at: types of data, how to analyze and use data with an equity lens, how to select a technology tool to support data-driven decision making, and data disaggregation tools.
- Data Ferret is an online data analysis and extraction tool that allows users to customize local, state, and federal data, creating tables, graphs, and maps to visualize the data.
- The Kids Count Data Center is an online resource that includes data on child well-being. The data can be disaggregated along multiple dimensions and is available at both the local and national level.
- Community Commons includes data sets and data visualizations for a number of topics, including vulnerable populations. Their map and data center allows users to select a tool by channel, e.g., economy, environment, or equity.
Planning processes that seek to solve complex problems typically turn to data, sometimes early to define the problem, sometimes later to understand which solutions can make a difference, and sometimes throughout. At any point in time, it is likely that some or most of the data you need will not be available. The choice point has two steps in it. First:
- Is this information important enough for making an equity decision that you need to collect, analyze, and report on it?
And if yes:
- What information can you collect, analyze and report on that will help you make an equity decision?
- Who can help you plan what data to collect, collect it, and help you analyze and interpret it?
Data collection processes are filled with choice points, from the decision the types of data to collection to the design of the data collection tool to the group of people collecting and analyzing the data. Even how the data is brought back into your process can influence its use (see our Complex Decision-Making Toolkit for more information on bringing data into a decision-making process).
To increase the likelihood that your data will help you make an equity decision:
- Do your background research: What are the questions your group wants answered? What information is available? What is missing?
- Go deep: Meet with key informants who can provide first-hand knowledge about a context which can also be used to frame a larger data collection strategy such as a community survey or youth focus groups.
- Consider different data collection methods: In some cases, traditional methods of data collection may not be effective or inclusive; alternative methods, such as passive data collection (e.g., neighborhood newsletters, social events or community meetings, blogs, and local music) may result in better community data because the things that get talked about in these circles are community issues and provide community context in a way that is difficult to capture in a survey.
- Develop your data collection instrument: Design it with the analysis in mind, thinking about how you need to be able to break down the data to tell you a meaningful, actionable story. Be careful to ensure it is culturally and linguistically (education level) appropriate.
- Pilot your data collection methodology: Engage those most affected by the problem to help you refine the tool. Does the terminology make sense? Is it culturally-appropriate? You may want to consider paying someone to serve as a content expert.
- Data Practices and Five Guiding Questions and this blog post from Anchoring Success can help groups consider how their data relates to five guiding questions to increase or develop a focus on equity.
- Spark’s other toolkits explore some of the more innovative data collection and analysis strategies, which may be of use when answering questions related to complex, dynamic problems and solutions. For more information on collecting and analyzing data, look at the methods pages in: The Developmental Evaluation Toolkit; The Advocate’s Evaluation Toolkit; and Tools for Adaptive Planning
Bringing data into the decision-making process can be done in ways that facilitate its use or result in the findings being tabled and the group moving on. Some of the issues to keep in mind:
- Since data can play such an important role, it’s worth considering effective ways to present data to a diverse audience. In some cases, data visualization can unintentionally reinforce stereotypes and perpetuate inequality, as highlighted in this blog post by Stephanie Evergreen. In other cases, how the data is presented may not be accessible, particularly for persons with visual impairments. Color Brewer 2.0 is a useful tool for creating color schemes that are colorblind safe and WebAIM has a color contrast checker to ensure the contrast between text and a background color is readable.
- Bringing in new data can sometimes be a challenge, depending on the composition of your group. Who brings in the information (who is seen as credible), as well as how the information is interpreted and processed (within existing cultural frames or cognitive biases) can affect how information is or isn’t used. Techniques such as joint fact-finding or a neutral third party may help, as can having a facilitated dialogue on how to process the information and develop a shared interpretation.