The short answer
Use multiple sources of evidence to choose a next step your team can act on. The starting point is a focused question and evidence the team can use.
Ask a useful question first
Begin with a question about a practice or opportunity the team can influence. For example: Which students are regularly participating in grade-level tasks, and what supports are available when participation breaks down?
Write the decision at the top of the meeting agenda before opening a dashboard. A question such as “Should we adjust the discussion routine in this unit?” needs different evidence from “Where should we allocate coaching support next term?” Name the group, setting, and period you are considering. Without those boundaries, a team can spend the whole meeting comparing numbers that are interesting but do not help it choose a next action.
Ask what you would do differently depending on the answer. If every possible finding leads to the same training session, the team may already have selected its solution. Reframe the question so the evidence can genuinely change the plan. Also identify what the available information cannot tell you. A participation count can show who spoke during an observation, but it cannot by itself explain the quality of each student’s reasoning or why someone remained silent.
Use more than one source
Program records, student work, classroom observations, and conversations with educators each reveal different parts of the system. Look for patterns across them. Avoid treating one indicator as a complete explanation or attributing an outcome to a single cause.
Build a small evidence set around the question rather than gathering every available measure. For a discussion question, that might include a lesson task, a consistent participation note, and an educator reflection. The task shows what was asked; the observation records part of what happened; the reflection adds context. Treat disagreements among sources as questions to investigate. They may reveal a difference in tasks, conditions, or interpretation rather than an error to erase.
Check the comparison before interpreting the pattern. Were the tasks similarly demanding? Did students have comparable time and supports? Are you looking at the same group over time, or at different groups? Record missing information and changes in conditions alongside the result. In a small group, a single absence or a different observation window can noticeably change a percentage. Preserve the underlying counts and context instead of presenting a precise-looking number without its limitations.
Connect evidence to action
Choose a small change, a responsible owner, and a date to look again. Track both implementation—whether the practice happened—and student experience. Improvement cycles work best when the evidence is manageable enough for the team to use.
Use separate columns for the observation, a possible explanation, and the next question. “Three students did most of the explaining” is an observation within a particular lesson. “The roles were unclear” is a possible explanation. “What happens when partners have explicit, rotating roles?” is a question the team might test. Keeping these statements separate prevents an interpretation from becoming an established fact simply because it appears in a meeting document.
Define an implementation check and a student-experience check for the selected change. For a new discussion routine, the first might ask whether the directions and roles were used; the second might ask whether each partner had a meaningful chance to explain. If implementation was inconsistent, investigate the support or design before judging the approach. If the routine was used but participation remained uneven, examine the task, grouping, time, and language resources before deciding what to adjust.
Put it to work
Your team’s next steps.
- Write the decision you want the evidence to inform.
- Check that the groups, measures, and timeframes are comparable.
- Separate what you observed from why you think it happened.
- Choose a change the team can influence and note competing explanations.
- Review whether the change was implemented before interpreting student results.
An example in practice
What could this look like?
A fictional grade-level team asks whether every learner has an opportunity to explain reasoning. It examines two comparable lesson tasks, participation notes, and educator reflections. The team notices that discussion structures differ, but records that as a possible explanation rather than a proven cause. It chooses a shared routine and a common observation prompt for its next review. No student names or invented outcome percentages are needed to make the example useful.
A practical planning tool
Turn a data meeting into a decision
Try this suggested structure with a bounded question and a small, comparable evidence set. The goal is a documented decision with its reasoning and limitations. It is not a larger dashboard or a claim that one short improvement cycle establishes cause and effect.
Describe before explaining
Give the team time to name what is directly visible in the evidence. Record the source and timeframe next to each observation. Ask participants to point to the information behind a statement. Keep possible causes in a separate column so the group can see where it is moving beyond description.
Look for another plausible explanation
Before agreeing on a cause, name at least one other explanation and what evidence would help distinguish them. A task change, absence pattern, observation method, or different support could matter. You may not be able to resolve every question in the meeting; record the uncertainty instead of filling it with confidence.
Specify the next decision point
Choose a manageable action, an owner, and a review date. Record what will be checked about implementation and student opportunity, and how the evidence will be collected consistently. At the next meeting, compare the observations with the original question. Decide whether you need an adjustment, a longer trial, or better evidence before drawing a conclusion.
Questions teams ask.
How much data is enough to start?
Start with enough relevant information to describe a bounded problem and its limitations. A small evidence set can support a tentative planning decision, while a consequential program decision may require broader review. Do not treat the convenience of available data as proof that it is adequate for every decision.
Should we use percentages or counts?
Use the form that makes the pattern and its scale understandable. A percentage without the group size can hide how few observations it represents. Keep counts, timeframes, and definitions available, and follow district rules for protecting small groups and student information when deciding what may be shared.
What if the evidence contradicts our initial idea?
Document that finding and revisit the explanation. An improvement cycle should allow a team to change its mind. Check whether the practice was actually implemented and whether the measures fit the question before deciding to continue, adapt, or stop the action.
What informs this approach?
Learning Forward’s Evidence standard calls for multiple sources to inform professional learning and for evidence to guide decisions about its impact. This supports using several kinds of information rather than treating one indicator as the whole story. Learning Forward: Evidence standard.
What should the team be careful about?
Small groups and incomplete records can produce misleading comparisons. Keep student information in approved district systems, use aggregate or de-identified material for shared planning, and record limitations alongside conclusions.
Put the idea into practice.
Explore Data-Informed Program Improvement for support applying this work, or continue with How do we move from compliance to classroom practice?
Use the Beyond the Checkbox toolkit to plan your next step, or explore the three-tier framework with your team.
For current Texas program requirements, consult the Texas Education Agency’s bilingual and ESL program guidance. These guides focus on improvement planning, not legal advice.
