HOW FORMATIVE ASSESSMENT WORKS · CHECK → COMPARE → ADJUST → RETEST · eduKateSG
Measure While Learning Can Still Change
A final examination tells us something important, but it arrives late. By the time the score appears, the teaching sequence has ended, the revision window has narrowed or closed, and the learner cannot go back into that same performance event to repair the mistake.
Formative assessment exists earlier in the loop. It asks what the learner knows now while there is still enough time to change what happens next.
Formative assessment is the deliberate collection and interpretation of learning evidence during the learning process so teaching, practice, feedback or student action can be adjusted before the final outcome is fixed.
This makes formative assessment less about the object called “a quiz” and more about the function of the information. The same five-question test can be formative if it changes the next lesson, or merely summative if the score is recorded and everyone moves on unchanged.
This article sits beside How Diagnostic Assessment Works, How Feedback Works, How Error Correction Works, How Learning Goals Work, How Deliberate Practice Works and the new Mastery Learning and Progress Tracking owners in this batch.
The 50-Second Read
- Formative assessment happens while change is still possible. The point is to improve learning before final evaluation.
- Evidence must change a decision. If nothing changes, the assessment is only measurement.
- Checks can be tiny. One question, one explanation, one confidence rating or one comparison can be formative.
- Targets matter. The learner and teacher need a reference state against which evidence can be interpreted.
- All students should think. Asking one volunteer is weak evidence about a whole class.
- Errors are signals. Their pattern should influence reteaching, practice and grouping.
- Feedback closes only part of the loop. The learner must then act and be retested.
- Formative assessment should not become constant surveillance. Measure enough to make better decisions, not everything that can be measured.
- The final goal is self-formative learning. Students should increasingly collect and interpret their own evidence.
1. Summative and Formative Are Functions, Not Just Test Types
A final examination is usually summative because it summarises achievement after a period of learning. A classroom quiz is often called formative because it happens earlier. But timing alone is not enough.
If the quiz is marked, recorded and forgotten, it has not changed the learning process. If the examination paper is later analysed and used to redesign the next term, parts of that same evidence can become formative for the next cycle.
The defining question is: what does the evidence do?
2. Formative Assessment Needs a Learning Goal
Evidence has meaning only relative to a target. A student answers 7/10 correctly, but what capability was being tested?
If the goal is to select the correct method in mixed algebra, a quiz of ten identical equation questions gives weak evidence. If the goal is causal explanation in Science, fact-recall questions are misaligned.
How Learning Goals Work supplies the reference state. Formative assessment asks whether the learner is moving toward it.
3. Use Evidence That Matches the Capability
A learning goal should be assessed in the form in which it must eventually be used. Retrieval goals need cue-free recall. Method-selection goals need mixed questions. Writing goals need actual writing. Scientific reasoning goals need evidence, variables and explanation.
Misaligned assessment can produce false reassurance. The learner looks strong on recognition questions and weak on independent production because the check measured the wrong layer.
Formative assessment is only as informative as its alignment.
4. Checks for Understanding Can Be Very Small
Teachers do not need a formal test every time they need information. A well-chosen question can reveal a misconception.
Ask every student to write the next algebra step. Ask for a one-sentence explanation of why a variable must be controlled. Ask students to choose between two possible inferences and justify the choice. Ask for a confidence rating before revealing the answer.
Small checks are powerful when they are designed around the exact distinction the teacher needs to see.
5. One Volunteer Is Not the Class
A teacher asks, “Does everyone understand?” One confident student nods. The lesson continues.
This is weak evidence because the sample is self-selected and tiny. Formative assessment should make enough of the class state visible to support a decision.
Use mini-whiteboards, short written responses, polls, cold call after think time, exit tickets or pair explanation. The exact tool matters less than increasing observability across the group.
6. Ask Questions That Separate Misconceptions
A strong formative question is not merely difficult. It is discriminating. Different answer choices or solution routes reveal different models.
For percentage, include alternatives corresponding to common base errors. For Science, contrast a descriptive answer with a causal answer. For comprehension, present literal evidence beside an inference that overreaches.
The wrong answer then tells us something specific about why the learner is wrong.
7. Diagnostic and Formative Assessment Overlap
Diagnostic assessment identifies the mechanism. Formative assessment repeatedly checks whether learning is moving and whether the intervention is working.
The same evidence can serve both purposes. A short mixed Mathematics set may reveal that method selection is the first weak link, then later show whether targeted interleaving improved it.
How Diagnostic Assessment Works therefore sits at the start of the loop; formative assessment keeps watching the loop move.
8. Feedback Is the Response Layer
Assessment produces evidence. Feedback interprets the difference and points toward action.
How Feedback Works explains how that signal should be specific, actionable and proportionate.
Without feedback or another adjustment, formative assessment can become an endless stream of mini-scores that changes nothing.
9. Correction Is the Action Layer
Once the gap is identified, something needs to change. The teacher reteaches, the student revises a paragraph, the tutor introduces a worked example, the learner practises retrieval or a misconception is contrasted with the correct model.
How Error Correction Works turns the measurement into route replacement.
10. Retest Closes the Formative Loop
After reteaching or correction, assess again. Did the gap close?
If the student now succeeds only with the same example visible, the intervention has not yet transferred. If performance survives after delay and in a changed question, evidence of learning is stronger.
The complete loop is:
Goal → attempt → evidence → interpretation → intervention → retest → update.
11. Formative Assessment Should Be Low-Stakes Enough for Honesty
If every check contributes heavily to grades, students may hide uncertainty, avoid risky answers or focus on marks rather than information.
Low-stakes checks encourage honest exposure of current state. A wrong answer becomes useful because there is still time to act.
This does not mean standards disappear. It means the function of the check is improvement rather than judgement.
12. Confidence Ratings Add a Second Signal
Ask students not only for an answer but how confident they are. Four states appear: confident-correct, confident-wrong, uncertain-correct and uncertain-wrong.
Confident-wrong answers may reveal stable misconceptions. Uncertain-correct answers show competence exceeding belief.
This helps Academic Confidence become calibrated and gives the teacher better information than correctness alone.
13. Formative Assessment Can Be Embedded in Explanation
Teachers do not need to stop teaching to assess. Explanation can contain prediction points.
Before revealing the next step, ask learners to predict it. Before showing an experimental outcome, ask for a hypothesis. Before completing a paragraph, ask which evidence should come next.
The prediction gives immediate evidence about the learner’s internal model and makes the explanation interactive.
14. Worked Examples Can Contain Formative Checks
A worked example becomes more formative when selected steps are removed. Ask students to complete the next line, explain why the expert chose the method or identify a deliberate error.
How Worked Examples Work provides the scaffold. Formative assessment tests whether the learner is ready for the scaffold to fade.
15. Retrieval Practice Is Formative When It Changes the Plan
Retrieval gives honest evidence of what is available without notes. If the result changes the next revision decision, it is formative.
A student retrieves 40% of a Science mechanism. The learner then schedules relearning and another spaced return. The assessment has shaped the learning process.
How Retrieval Practice Works therefore functions as both learning and measurement.
16. Exit Tickets Can Create a Next-Lesson Signal
A short end-of-lesson response can reveal what deserves attention tomorrow. Ask one diagnostic question, one explanation or one “muddiest point.”
The value appears only if the teacher uses the signal. If half the class misses the same relationship, the next lesson should respond before moving on.
Exit tickets are useful because they connect today’s state to tomorrow’s design.
17. Homework Can Be Formative Evidence
Homework is not only completion. It can show whether learning survives outside the classroom and without immediate teacher support.
If classwork is strong but homework collapses, the missing variable may be independence, environment, memory or task capture.
How Homework Routines Work helps interpret the evidence rather than assuming every incomplete task is an academic failure.
18. Discussion Can Be Formative—If Participation Is Broad Enough
Class discussion reveals reasoning, vocabulary and misconceptions. But a few articulate students can dominate the signal.
Use think-pair-share, written preparation, structured turn-taking or random selection after think time so more learners generate an answer.
The aim is not to maximise public speaking. It is to increase visibility into the class’s thinking.
19. Peer Assessment Can Reveal Criteria
Students sometimes see weaknesses more clearly in another person’s work. With clear criteria, peer assessment can help them understand quality.
Keep the task narrow: identify one unsupported claim, one missing causal link or one algebra step that does not preserve equality.
The teacher remains responsible for accuracy and classroom safety. Peer assessment is useful as a learning activity, not simply unpaid marking.
20. Self-Assessment Is the Long-Term Destination
Students should increasingly be able to ask: What is the goal? What evidence do I have? Where is the gap? What should I do next?
This is formative assessment moving inside the learner. How Independent Learning Works reaches maturity when the student can generate and interpret useful learning evidence without waiting for an adult.
21. Formative Assessment in Mathematics
Mathematics supports fast formative checks because intermediate working is visible. Ask every student for the first step, classification or estimate.
Use distractors representing common misconceptions. Separate method selection from execution. Analyse whether students can solve without the topic heading.
The small-group model in Secondary 1 Mathematics Tutor Clementi | Small Groups Tutorials provides high-resolution formative visibility because the tutor can observe each learner repeatedly within the same lesson.
22. Formative Assessment in English Comprehension
Use one passage and vary the question function. Ask students to identify evidence before writing an inference. Track pronoun reference. Compare two possible answers and explain why one overreaches.
The assessment should reveal which reasoning stage is weak rather than merely produce another score.
23. Formative Assessment in Writing
A full composition is expensive to assess. Formative writing can work at smaller scales.
Ask for a paragraph plan, one evidence-to-claim link, one opening, one revised sentence or one conclusion. These micro-products let the teacher inspect a target before the learner invests an hour in the whole piece.
Then use full writing to test whether the repaired component survives integration.
24. Formative Assessment in Vocabulary
Vocabulary checks should test more than recognition. Ask for meaning, context, collocation, near-synonym distinction or sentence use.
Spaced retrieval can reveal which words are becoming durable and which need another cycle.
25. Formative Assessment in Science
Science assessment should distinguish recall, causal understanding, application, data interpretation and experimental reasoning.
One well-designed scenario can ask students to predict, justify, identify variables and explain a mechanism. The pattern of responses shows where the model is incomplete.
26. Formative Assessment During Revision
Revision should contain frequent low-stakes measurement because the learner is deciding where to spend scarce time.
How Revision Works uses retrieval checks, mixed sets and timed sections to update priorities.
A revision plan that never measures is operating blind.
27. Past Papers Can Be Formative Before the Final Exam
A past paper is often treated as a summative mock, but its deeper value is formative when the analysis changes later training.
Where did marks disappear? Which errors repeat? Where did time go? Did one difficult question destabilise the rest? Which topics remain cue-dependent?
The next week’s practice should reflect the answers.
28. Do Not Measure So Often That Learning Becomes Testing
Assessment has a cost. Every check uses class time, attention and emotional bandwidth. Too much measurement can crowd out explanation, exploration, reading and deliberate practice.
The purpose is enough visibility to make better decisions. If the teacher already knows the state with high confidence, another check may add little.
29. Do Not Collect Data Nobody Uses
Digital platforms can generate large amounts of student data. More data does not automatically produce formative value.
Ask what decision the metric changes. If a dashboard shows ten indicators but nobody changes teaching, practice or support, the system has visibility without control.
Measure what matters enough to act on.
30. Formative Assessment Should Protect Motivation
Frequent low-stakes checks can support motivation when students see them as information. They can damage motivation when every check feels like a public ranking.
Keep purpose explicit: we are checking so we know what to teach and practise next. Use private response systems where appropriate. Highlight progress across attempts.
31. Formative Assessment and Mastery Learning
Mastery learning depends on formative evidence because the system needs to know whether the learner is ready to build the next layer.
A learner who has not stabilised the prerequisite receives corrective work before moving on. A learner who demonstrates mastery can progress or stretch.
The next article, How Mastery Learning Works | Don’t Build on an Unstable Floor, turns formative signals into progression decisions.
32. Formative Assessment and Progress Tracking
One check shows state. Repeated checks show change.
Progress tracking connects formative evidence across time: fewer sign errors, stronger retrieval after delay, reduced help dependence, improved timing.
The final article in this batch, How Progress Tracking Works | Seeing Change Before the Final Grade, turns isolated measurements into a trajectory.
33. The Formative Assessment Loop
- Goal: define the target capability.
- Elicit: create a task that reveals relevant thinking.
- Observe: collect enough evidence across learners.
- Interpret: identify the meaningful gap or misconception.
- Respond: reteach, regroup, practise, feedback or extend.
- Retest: check whether the response changed performance.
- Update: move forward, continue repair or reopen diagnosis.
34. A Student Formative Audit
- What am I trying to learn?
- What evidence do I have right now?
- Can I retrieve it without notes?
- Can I explain it?
- Can I use it in a changed question?
- What error keeps appearing?
- What feedback changes my next attempt?
- When will I retest?
35. A Parent Formative Audit
- Am I looking only at final grades?
- What leading signals show learning earlier?
- Does my child know current weak areas?
- Are quizzes being corrected and retested?
- Can the child explain what the teacher’s feedback means?
- Am I turning every low-stakes check into a high-stakes event?
36. A Teacher or Tutor Formative Audit
- Is the check aligned with the learning goal?
- Does it reveal thinking rather than only final answer?
- Am I sampling enough learners?
- Can common wrong responses distinguish misconceptions?
- Will the evidence change the next teaching decision?
- Do I retest after intervention?
- Am I helping students become self-assessors?
37. A Four-Week Formative Build
Week 1 — Clarify targets and checks. For one subject, define three important capabilities and design one short check for each.
Week 2 — Improve observability. Replace “Does everyone understand?” with responses from all learners. Track the most common wrong model.
Week 3 — Link evidence to action. Every significant pattern must trigger reteaching, targeted practice, grouping or extension. Retest later.
Week 4 — Transfer assessment inward. Students predict performance, self-check against criteria and propose the next action before receiving adult judgement.
38. What Not to Do
- Do not call a quiz formative simply because it is short.
- Do not assess a capability with the wrong task type.
- Do not infer class understanding from one volunteer.
- Do not collect scores without changing instruction or practice.
- Do not make every check high stakes.
- Do not overwhelm students with constant testing.
- Do not collect data nobody uses.
- Do not give feedback without a chance to act on it.
- Do not skip retest.
- Do not keep assessment ownership entirely with adults forever.
Frequently Asked Questions
What is formative assessment?
It is assessment used during learning to collect evidence that changes what teachers or students do next before final performance is fixed.
Is a quiz formative assessment?
It can be. The defining feature is whether the evidence from the quiz changes teaching, practice, feedback or student action and is followed by another opportunity to improve.
What are examples of formative assessment?
Checks for understanding, mini-whiteboards, exit tickets, retrieval questions, short quizzes, confidence ratings, draft feedback, error analysis, peer assessment and self-assessment can all be formative when they guide subsequent learning.
How is formative assessment different from diagnostic assessment?
Diagnostic assessment focuses on identifying the mechanism behind a current problem. Formative assessment repeatedly measures learning during instruction so the process can be adjusted.
Should formative assessment be graded?
It can carry some marks in some systems, but lower stakes often support more honest evidence and willingness to expose uncertainty. The main purpose should remain improvement.
Return: Measure Early Enough to Matter
Every educational system measures eventually. Examinations, reports and grades tell us what the learner could produce at a particular point.
Formative assessment asks for the same courage earlier.
Show us what is missing now. Show us which misconception is still alive. Show us which student is following only because the worked example remains visible. Show us whether last week’s correction survived.
Measure while the future is still editable.
That is the defining value. Formative assessment is not a smaller examination. It is a control signal placed early enough that the teacher, tutor and learner still have time to steer.
When the loop matures, students begin doing this for themselves. They test, interpret, adjust and retest. Measurement becomes less about being judged and more about knowing where to go next.
Continue: How Diagnostic Assessment Works · How Feedback Works · How Error Correction Works · How Learning Goals Work · How Deliberate Practice Works · Secondary 1 Mathematics Tutor Clementi | Small Groups Tutorials.