Assessment and Feedback
From Assessment Data to Instruction: A Teacher Decision Guide
Decision first: Assessment data are useful when the evidence matches a specific instructional decision. One score should not become a permanent description of a student. Check what the task measured, who could access it, what data are missing, and whether the next lesson can test the interpretation.
Resource type: Decision guide and worked example
Preparation time: 20–30 minutes for one assessment set
Best for: Teacher teams planning the next lesson or short cycle
Important limit: Use approved systems and the minimum student information needed. Follow local rules for records, access, sharing, retention, accommodations, and high-stakes decisions.
Begin with the decision
“Use the data” is not a decision. Write one action the evidence could change:
- Reteach a prerequisite to the whole class.
- Provide a short, temporary small-group lesson.
- Change the model or example used next.
- Add guided practice before independent work.
- Check whether the assessment itself created a barrier.
- Move on while planning spaced retrieval of earlier learning.
Next, ask what evidence is sufficient for that action. A quick exit ticket may guide tomorrow’s grouping; by itself, it is not sufficient for placement, a disability inference, course recommendation, or another consequential decision.
Match evidence to the question
On a small screen, scroll this table horizontally to see all columns.
| Instructional question | Useful evidence | Insufficient by itself |
|---|---|---|
| Can students select the correct operation in a word problem? | Two or more items with different surface features, student work, and a brief explanation | One multiple-choice answer with no reasoning |
| Which part of an argument needs reteaching? | A short response scored separately for claim, evidence, and reasoning | A single total essay score |
| Can students retrieve a concept after a delay? | A low-stakes prompt several days later with comparable conditions | Correct work immediately after a modeled example |
| Did an accommodation or format affect access? | Student plan, observed interaction with the format, work samples, and support-team input | A score difference attributed to disability, language, effort, or anxiety without investigation |
Check the evidence before interpreting it
Item and score meaning
Confirm what each item required and whether the scoring rule distinguishes the knowledge or reasoning needed for the next decision. A total score can hide different error patterns. Review the actual work alongside the dashboard category.
Small samples
One or two items provide unstable evidence. A student may guess, misread, omit, or encounter an unusual example. Use another prompt, a brief conference, or later work before treating the pattern as established. In small groups, percentages can look more precise than the underlying count; report both.
Missing data
When possible, record why evidence is missing: absence, technical failure, interrupted administration, skipped item, unavailable accommodation, or no scorable response. Do not automatically interpret missing work as zero understanding. Decide whether to collect replacement evidence, proceed with uncertainty, or use a different source.
Accessibility and language
Check reading load, vocabulary, directions, visual layout, response mode, timing, assistive technology, and required accommodations. For multilingual learners, distinguish the language demanded by the objective from language that is incidental to it. A wrong answer may reflect the target concept, the language of the item, or both.
Opportunity to learn
Ask whether students had enough instruction and practice on the assessed target in the tested form. Evidence cannot fairly guide reteaching if the content, representation, or item type was not taught or accessible.
Do not turn a result into a permanent label
Describe the evidence in time-bound, task-bound language:
On a small screen, scroll this table horizontally to see all columns.
| Avoid | Use instead |
|---|---|
| “Low student” | “On Tuesday’s three fraction-comparison items, the student used denominator size without comparing quantity.” |
| “Advanced reader” | “In this passage, the student integrated two details to support the stated inference.” |
| “Unmotivated” | “The response was blank after the student spent four minutes rereading the directions; the cause is not established.” |
| “Not ready for algebra” | “The current sample shows accurate substitution but inconsistent distribution across a negative factor.” |
Descriptions of observable work support a testable instructional response. Labels encourage fixed expectations and often conceal uncertainty.
Use temporary groups, not fixed tracks
- Group by the immediate learning need, not a broad ability category.
- Set a review point, such as after two lessons or the next evidence check.
- Keep grade-level or course-level learning available while adding prerequisite support.
- Vary group membership by target and evidence.
- Avoid public names such as “high,” “low,” or animal/color systems students can rank.
- Check whether the group assignment changes access to rich tasks, teacher attention, or future opportunity.
Worked teacher decision: equivalent fractions
Decision
A grade 5 teacher must decide whether tomorrow’s lesson should reteach fraction magnitude to the whole class, provide a temporary group on equivalence, or proceed to addition with unlike denominators.
Evidence collected
An exit ticket contains three items:
- Place 3/4 on a number line from 0 to 1.
- Choose whether 4/6 and 2/3 are equivalent and explain.
- Compare 5/8 and 5/6 and justify without using a decimal calculator.
Initial pattern
Six of 24 students correctly place 3/4 but treat 4/6 and 2/3 as different because the numerators and denominators differ. Four other students omit item 3. Two of the omissions are from students whose required text-to-speech setup was unavailable during the ticket.
Checks before acting
- The equivalence item and explanation directly fit the next decision.
- Six responses share an observable strategy error across written work.
- The four omissions are not coded as lack of understanding; two occurred under invalid access conditions.
- The teacher checks yesterday’s work and confirms that four of the six students used the same non-equivalence rule twice.
- A one-item follow-up asks the other two students to build 3/4 and 6/8 with fraction strips and explain the relationship.
Action
The class begins with a five-minute number-line retrieval prompt because several students need more magnitude practice. A temporary group of four students receives a 12-minute equivalence lesson using area models and multiplicative relationships. The two students awaiting follow-up complete the model task before group placement. Students with invalid access conditions redo the evidence check with the required support.
Review point
After two lessons, every student answers one equivalence item and explains the multiplicative relationship. The temporary group ends or changes based on that result. The original exit ticket is not used as a permanent group label.
When evidence should change instruction
Change the next step when several sources support the same interpretable pattern. The pattern should relate to an important prerequisite or objective, and the response should be feasible to test. Keep the plan when a difference reflects random variation, an irrelevant item feature, or a barrier that should be repaired before interpreting knowledge. Collect more evidence when the sample is too small, contradictory, missing, or gathered under invalid conditions.
Privacy and sharing
Use the least identifiable data needed for planning. Discuss student work in approved settings with staff who have a legitimate role. Do not export named records to an unapproved spreadsheet, dashboard, transcription service, or generative tool. Apply retention and deletion rules, and format team displays so a small subgroup or distinctive detail does not expose a student. The student data privacy checklist provides a fuller pre-tool review.
A six-question decision record
- What exact decision could this evidence change?
- What did the task and scoring rule actually measure?
- Which responses are missing, inaccessible, or too few to interpret?
- What alternative explanation must be checked?
- What short instructional response will test the interpretation?
- When will the group, strategy, or conclusion be reviewed?
Sources and scope
- Institute of Education Sciences, Using Student Achievement Data to Support Instructional Decision Making.
- Institute of Education Sciences, Organizing Instruction and Study to Improve Student Learning.
- U.S. Department of Education Student Privacy Policy Office.
- CAST Universal Design for Learning Guidelines.
Assessment evidence supports professional judgment; it does not eliminate uncertainty. High-stakes, diagnostic, placement, disability, or legal decisions require the authorized process and evidence appropriate to that decision.