design-formative-assessment

Design Formative Assessment

Build assessment-for-learning checkpoints that give teachers real-time data and give learners a clear signal of where they stand before it's too late to adjust.

Why This Is Best Practice

Adopted by: UK Assessment Reform Group policy, Dylan Wiliam's Embedded Formative Assessment adopted in 30+ countries, OECD learning frameworks Impact: Black & Wiliam's landmark 1998 meta-analysis (250 studies) showed formative assessment interventions produce effect sizes of 0.4–0.7 SD — among the highest of any educational intervention; Hattie's Visible Learning synthesis ranks feedback as the #1 influence on student achievement

Why best: Summative assessment tells learners how they did. Formative assessment tells them — and the teacher — what to do next, while there is still time to do it. The gap between "where learners are" and "where they need to be" is the formative assessment's only job to reveal.

Steps

1. Identify the learning gap to probe — Select one specific misconception or skill gap you expect at this point in the unit. Formative assessments probe one thing precisely; general "how are you doing" checks produce noise.
2. Choose the assessment technique — Match technique to purpose:
   • Hinge questions: multiple choice at a conceptual branch point — designed so each wrong answer reveals a specific misconception.
   • Exit tickets: 1–3 written questions at end of lesson, collected before students leave.
   • Mini-whiteboards: simultaneous reveal of responses across the class.
   • Cold calling + wait time: structured oral checks with 5+ seconds think time.
   • Pair discussion + teacher monitoring: teacher circulates and samples conversations.
3. Design a hinge question — Write one question where: the correct answer requires the target concept, each incorrect answer corresponds to a known misconception, and all options are plausible. Avoid "all of the above" or trick options.
4. Plan your response to each result — Before giving the assessment, decide: "If 80%+ correct → move on. If 60–80% correct → run structured partner discussion. If <60% correct → re-teach using a different representation." Unplanned formative assessment produces data that is not acted on.
5. Build in response time — Reserve 5–10 minutes per class to act on formative data. If there is no time to respond, the data is only useful for planning the next lesson.
6. Return feedback that is task-focused — Feedback on formative work should say what is correct, what is wrong, and what to do to close the gap — not scores or grades. Scored formative assessment activates ego-protective thinking and reduces the learning benefit.
7. Track class-level patterns — After each assessment, note the class distribution. Use this to form same-level groups for targeted follow-up activities rather than re-teaching the whole class.

Rules

• Formative assessments are not graded. Grading formative work changes the behavior — students play it safe rather than revealing genuine misunderstanding.
• Every formative assessment must trigger a planned instructional response — "assess and move on" is not formative assessment.
• Hinge questions must be designed so each distractor corresponds to a specific misconception — randomized wrong answers produce uninterpretable data.
• Share the learning objective before the assessment so learners can calibrate their own understanding against the target.
• Exit tickets are collected — not optional. Optional self-checks sample only confident students.

Examples

A hinge question in a fractions unit: "Which fraction is larger: 3/4 or 4/5?"

• A) 3/4 (reveals: student uses numerator only)
• B) 4/5 (correct)
• C) They are equal (reveals: student thinks larger denominator = larger fraction)
• D) Cannot tell without a common denominator (reveals: student knows cross-multiplication is needed but lacks confidence)

Each wrong answer tells the teacher exactly which re-teaching intervention to deploy.

Common Mistakes

• Grading formative work — Students learn to hide confusion rather than reveal it; the teacher loses the signal that makes formative assessment valuable.
• No planned response — Using formative assessment data only to update lesson plans for next week misses the core purpose: adjusting instruction now.
• One-size check-in — "Thumbs up if you understand" produces social conformity, not learning data; replace with simultaneous written or whiteboard responses.