Data-Informed SEL: A Practical Data Cycle for Schools

By the Transforming Education editorial team. Reviewed for accuracy against the sources listed below.

Data-informed SEL means a school looks regularly at evidence about students' social and emotional development and about its own SEL practice, and changes what adults do as a result. It does not mean a dashboard, a vendor report or a score next to a child's name. This guide lays out a cycle that a school team can run in a year, the data worth collecting, a meeting protocol, and the guardrails that keep the work honest.

Why a cycle, not a report

CASEL's Guide to Schoolwide SEL names systems for continuous improvement as one of the indicators of schoolwide SEL: implementation and outcome data are collected and used to improve SEL systems, practices and policies with a focus on equity. Its Focus Area 4 defines continuous improvement as a deliberate and structured process to address problems of practice and improve outcomes, and describes two cycle lengths: longer cycles run in structured team meetings that review the overall SEL plan, and shorter cycles that let schools test new ideas. It also notes that continuous improvement does not replace rigorous research. A school data cycle tells you whether your plan is working here; it does not prove a program works everywhere.

The same guidance assigns administrators a specific job: build an improvement-focused culture, build trust that data will not be used to judge staff, provide access to good data, model reflective practice, and make sure staff have authority to act on what they learn. Without the trust piece, the rest of this page fails.

The data cycle: collect, review, act, reflect

The short cycle (a grade-level team trying morning check-ins for six weeks and looking at tardies and student check-in responses) nests inside the long cycle (the school SEL team reviewing fall and spring survey results against the annual plan). Both need a calendar slot before the year starts.

What to collect

The CASEL guidance distinguishes implementation data (did we do what we planned?) from outcome data (did anything change for students?). Collecting outcome data without implementation data is the most common failure, because a flat result cannot be interpreted. If self-management scores did not move, you need to know whether the self-management lessons were taught.

School level

Washoe County School District shows what a mature version looks like. Its SEL snapshot describes user-friendly data tools, protocols and trainings for SEL decisions, an early warning risk index to flag students at risk of falling off the graduation path, and a public data site. Its competency items ride inside the annual student climate survey of about 24,000 students in grades 5 to 9 and 11, with aggregate results returned to schools through data profile books, professional development and data summits.

Classroom level

• A two-question check-in at the start of class (how are you arriving, what do you need) tallied weekly. See SEL check-in questions.
• Exit tickets after explicit lessons: one skill named, one place to use it.
• Student goal sheets reviewed every two weeks.
• A teacher's own tally of which signature practices (welcoming routine, engaging strategy, optimistic closure) actually ran that week.
• Observation notes from a peer or coach using the school's SEL look-fors.

Try this: For one month, have every teacher log one line a week: "SEL lesson taught: yes/no. One student I checked in with: initials." The aggregate tells the SEL team more about Tier 1 than any survey.

A 45-minute team protocol

Structured protocols keep a data conversation from sliding into either defensiveness or storytelling. The Center for Leadership and Educational Equity offers a free protocol, ATLAS: Looking at Data, in its protocol library. The version below is our own and fits a 45-minute meeting.

1. Predict (5 minutes). Before anyone sees the data, each person writes what they expect to see and why. This surfaces assumptions and makes surprises visible.
2. Describe (10 minutes). Only factual statements: "Grade 7 self-efficacy averaged 3.4 in fall and 3.1 in spring." "Twelve of 18 classrooms logged the lesson weekly." No "because" allowed yet.
3. Interpret (10 minutes). Now offer explanations, each tied to a specific observation. Include at least one explanation that is about adult practice or school conditions, not about students.
4. Question (5 minutes). What would we need to know to choose between explanations? Often the answer is "ask students," and that becomes the first action.
5. Decide (10 minutes). One or two actions, each with an owner, a start date and a check date. Write them where everyone can see them.
6. Reflect (5 minutes). Did the protocol help? Did any voice dominate? Adjust next time.

Rules that make it work: data arrive at least two days before the meeting; results are shown at school and grade level, never by named teacher; the facilitator is not the principal; and the meeting ends with actions, not with a plan to meet again.

Equity and subgroup analysis

Disaggregation is where SEL data earn their keep, and where they can do harm. CASEL's Jeremy Taylor and RAND's Laura Hamilton wrote in The Hechinger Report that SEL assessment can reveal disparities in how students' needs are being met and help identify root causes. The CORE Districts' survey, analyzed by PACE, illustrates the payoff: girls' self-efficacy fell sharply relative to boys' through middle school, economically disadvantaged students reported lower skills across the board with some gaps narrowing in high school, and white students reported higher levels than African American and Latinx students. No school discovers those patterns without disaggregating.

Three cautions apply.

First, the PACE authors warn that self-report differences across groups may reflect different normative standards rather than real skill differences. A 2022 study in Scientific Reports found that adolescents with higher-achieving peers rated their own self-regulation lower, while task-based measures did not show the pattern. A gap in survey scores is a reason to look harder, not a measurement of a deficit.

Second, protect small groups. Set a minimum reporting size (many districts use 10 or more students) and suppress cells below it. A subgroup of four students is identifiable to anyone in the building.

Third, frame findings as questions about adult practice and conditions. "Our Black students report lower social awareness" is a dead end. "Which classroom and school experiences differ for our Black students, and what do they tell us about belonging?" is a plan. Washoe's CDI example shows the move: after its survey showed students felt weakest in expressing feelings, the district targeted staff development to that skill.

Avoiding misuse

The guardrails below come from the measurement research and from the districts that have run SEL data systems longest.

• No high-stakes use. Angela Duckworth and David Yeager concluded in Measurement Matters that no available measure is suitable for between-school accountability judgments and that indicators become corrupted once decisions depend on them. The CORE Districts weighted their SEL measures at eight percent of a school's score and, as PACE notes, dropped the accountability use after ESSA passed.
• No labeling. A handful of items cannot diagnose a child, and a situational problem can be misattributed to the child. CASEL's practitioner guidance, as reported by K-12 Dive, states that screening for deficits is not an appropriate use of SEL assessment tools. Washoe describes its use as low-stakes decision-making, discussion and trend monitoring.
• No teacher evaluation. The CASEL continuous improvement guidance asks leaders to build trust that data will not be used to judge staff. The first time a survey result appears in an evaluation, the data stop being honest.
• No surprise disclosures. Decide and publish how data will and will not be used. Surveys touching the eight protected areas under the Protection of Pupil Rights Amendment, such as mental or psychological problems, need consent planning; competency items usually do not, but families should still be able to opt out.
• No single source. Taylor and Hamilton advise pairing SEL assessment with climate and implementation measures. If the survey, attendance and teacher observations disagree, the disagreement is the finding.

Note: Keep SEL scores off report cards and out of discipline files. A score that follows a student into next year's classroom has become a label, whatever the intent.

Examples from practice

• Austin Independent School District raised local funds for a two-person SEL research team that produced regular leadership reports, including one measuring a curriculum's relationship to absences, discipline incidents, grades and test scores, and another comparing program longevity with implementation quality, with mixed results. CASEL's CDI report credits those reports with building buy-in and raising funds.
• Anchorage School District found staff survey participation had dropped sharply, used the data with CASEL to reset training priorities, and saw participation rise from 30 percent to 79 percent once principals saw the surveys guiding their work.
• Long Beach Unified, within the CORE network, lets each school respond to its own dashboard. The 74 reported responses ranging from new curricula to parent nights to a math course that alternates algebra with SEL lessons.
• Boston Public Schools uses mid-year and end-of-year climate surveys inside its SEAD coaching cycle. The Rennie Center's 2025 report describes one school where students reported a 9 percent increase in belonging by mid-year and teacher-reported climate indicators improved 17 percent between mid-year and year's end.

Common pitfalls

• Collecting outcome data without implementation data, then blaming the program.
• Reviewing data once a year, in June, when nothing can change.
• Letting the vendor dashboard define the questions.
• Reporting by teacher name.
• Treating a 0.1 change on a five-point scale as a trend.
• Skipping the student conversation. Students can usually explain a number faster than a regression can.

Start small

A school that gives one survey twice, pulls attendance and discipline monthly, logs its lessons, and runs the 45-minute protocol three times a year is doing data-informed SEL. Add complexity only when the team has acted on what it already has. For background on the measurement debate, see Measuring MESH; for examples of districts that built data systems over a decade, see the case study compilation.