6 Steps to Equitable Data Analysis

It’s 8:40 a.m. on Monday, and a group of third-grade teachers are looking at a spreadsheet displaying student names and reading levels for each quarter of the year. Data fills the page. After the meeting, we feel confident about our analysis and next steps.

But something was missing in the discussion: We didn’t spend any time considering students’ identities and backgrounds, nor how we can improve equitable outcomes.

For teachers and administrators, it’s easy to overlook culture and identity when analyzing data. However, ignoring students’ diversity markers means pretending that their identity doesn’t matter. In order to close gaps in student outcomes, we must name equity as an essential component to data analysis.

6 Steps to a More Equitable and Culturally Proficient Data Analysis

1. Research identity: We can’t consider equitable outcomes if we don’t know who students are. Using tools such as surveys, calls home, and classroom team-building activities, we need to hear students’ stories.

One group of teachers at my school realized how little they knew about a small group of students who were struggling. They each made a research plan for getting to know three kids, designing questions and topics to explore in the coming week. When they next met, they shared their learnings. They were surprised about how much students shared about their home lives, career aspirations, and more. As one teacher said, “Wow—we need to start asking these questions in the first weeks of school.”

Educators bring our own history and perspective to this work, so we also need to know ourselves and our implicit biases. In order to heighten your self-awareness, take Harvard’s Implicit Association Test or complete the Diversity Awareness Profile. Research systemic racism and implicit bias, and reflect on your own identity and various forms of privilege. It also can be helpful to spend some time completing and reflecting on these activities with your team prior to looking at data.

2. Preempt implicit bias: Research suggests that priming your team with positive representations of different groups can influence their perspectives. For example, you might read an article by a Latina woman or look at pictures of successful black leaders prior to reviewing writing samples by students sharing similar identity markers. You can explicitly tell the team that we carry biases into data analysis, and that one research-based approach to counteract those biases is to read about, discuss, or highlight examples of, for example, a racial minority group that portray that group in a positive light.

Additionally, create a low-stress space where you’ll do your data analysis. Although easier said than done in many buildings, when possible have snacks, calm lighting, the right temperature, and regular breaks. Happier conditions will mitigate the chance of implicit bias.

3. Frame and challenge: What are your team’s ground rules for looking at data? Many teams I support use the norm, “Talk about students the way we talk about our own children.” Be wary of the data analysis meeting where everyone reads through student writing samples and shares the “funny” ones aloud.

Facilitators need to be prepared to calmly challenge assumptions and stereotypes. We can’t sit silently when we hear a colleague blame home circumstances, the previous year’s teacher, or a learning disability as reasons a student can’t make the same progress as everyone else. When possible, respond with “I don’t see it that way” or “Can you explain what you mean by that?” to start a dialogue that unpacks people’s perspectives. It will likely take time to dismantle a deficit mindset, but doing so is essential for promoting equity.

4. Set intentions: Have each participant name an intention prior to analyzing data to generate personal accountability. One team member might set the intention to “prioritize student strengths,” while another might intend to “leave with a plan for better supporting the Latinx students in my class.” Whatever the intention, this focus will help give direction and purpose to the analysis, as well as promote a positive mindset.

5. Pick the right data: Standardized assessments should not be your only metric. There’s plenty of research reminding us that standardized tests were born out of the eugenics movement and to this day contain significant bias favoring affluent white students. Make sure your team spends time with varied types of assessment data, like essays, short stories, projects, science labs, and notes from classroom discussions. Additionally, consider flexibility when creating scoring rubrics—do they only include “have to” criteria, or do they provide space for student choice? The goal should be for all students to have the opportunity to shine.

6. Strategically sort: Disaggregate data so that it’s broken down by race, gender, IEP status, home language, etc. You should be able to see outcomes for groups of students and quickly call out disparities, if any exist. Sometimes this is in a spreadsheet, and other times it simply means putting exit tickets into different piles.

Conversely, sometimes hiding students’ identities at first can help. For example, cover up student names when reviewing writing samples in order to counteract assumptions based on what you already “know” about a student’s academic level. This also helps us avoid applying a student’s negative behaviors to their academic performance.

It’s easy to depersonalize data, but every data point is a child. Ultimately, equity-driven data analysis means doing whatever it takes to remove bias and stereotypes. Taking even some of the small steps above will not only strengthen the quality of data analysis, but will begin to dismantle systemic inequity in the school community.