AfL KAT 03 · Support
Assessment for Learning
What this KAT is for
Reach for this when you need to know what students understand right now — not at the end of the unit. The goal is feedback that moves learning forward, both teacher-to-student and student-to-self. The fastest formative cycle wins: the sooner a student sees what they got wrong and why, the faster they correct.
Familiar examples: walking past a desk and realising five students misread the same word problem; a P3 oral practice where you wish every student got a turn, not just two; marking 40 compositions and seeing the same vocabulary mistake again and again. Technology gives you eyes on every student’s thinking at once, and gives students immediate feedback when their attention is highest.
MOE defintion:
Allow teachers to capture, analyse, summarise and visualise learning data to provide students with targeted feedback and resources to move their learning forward.
How to choose a tool for Assessment for Learning
Not every quiz tool does AfL. A graded quiz with no feedback is summative in function, regardless of intent. Assessment becomes formative only when the student receives information they can act on — not just a number. To do AfL work, the tool needs to do at least one of these three things:
1 Signal 01 · Fast cycle
Feedback within the lesson, not the week — where possible.
The sooner a student receives feedback while the task is still in their head, the stronger the connection between the comment and the thinking that produced it. Same-lesson feedback is gold; same-week is good practice. Later feedback can still be formative — but it takes more deliberate effort to re-engage students with the work and close the loop.
2 Signal 02 · Targeted feedback
Specific to the mistake, not generic
"Try again" doesn’t move learning. "You may have confused 末 with 未 — check the length of the top stroke" does. The tool should let teachers (or auto-feedback) say something the student can act on.
3 Signal 03 · Visible to the learner
Students see their own progress
A teacher who sees the data but never shows it to students is doing assessment of learning. AfL needs students to know what they’re aiming for, where they are, and what to do next.
Tools by AfL category
Grouped by what kind of AfL the tool does, not by subject.
Asynchronous structured SLS Feedback Assistants & quizzes
MOE-native tools that capture written responses, generate AI-drafted feedback, and let teachers review-and-edit before students see. The first port of call for AfL at Maris Stella.
SLS quizzes with auto-feedback MOE— MCQ, MRQ, fill-in-the-blanks, click-and-drop, free response — all with conditional feedback on each option.
SLS Feedback Assistants (FA-Math, LangFA-EL, ShortAnsFA) MOE — subject-specific AI feedback — the canonical Amplification-level AfL tool.
SLS Learning Assistant (LEA) MOE — teacher-activated AI chatbot built into SLS lessons. Choose a "recipe" (discussion facilitator, perspective builder, ideas generator) and a knowledge base; the chatbot asks Socratic guiding questions instead of giving direct answers. Surfaces reasoning and misconceptions during teacher-led inquiry.
SLS Learning Process Dashboard (for Adaptive learning) MOE— tracks overall progress of students by topics and individual students.
SLS audio recording for oral assessment MOE— every student records — perfect for 看图说话, English oral, MTL — teacher leaves targeted audio or text comments.
SLS MCQ with hint-on-error feedback MOE — immediate corrective feedback and a worked example — students don’t bake in misconceptions.
SLS rubrics in assignments MOE— make success criteria visible to students before they start.
SLS exit-ticket quiz(3 diagnostic questions) MOE— end-of-lesson check that informs the next lesson’s reteach.
Live polling Live formative assessment tools
Whole-class snapshots in seconds. Best for in-the-moment checks during teacher-led lessons — every student responds, you see the room's understanding, decide what to do next while the lesson is still warm.
ClassPoint Norms Free— adds live interactive activities (MCQ, short answer, word cloud, draw-on-slide) directly inside PowerPoint slides. Students join with a class code; responses appear on your slide while you teach. Best for in-the-moment comprehension checks during teacher-led lessons.
Pear Deck Norms Free— turns Google Slides or PowerPoint into interactive sessions where every student responds to embedded prompts on their own device. Strong for reflection prompts, draw-your-answer tasks, and anonymous responses on sensitive topics.
Wayground Norms Free— gamified quizzes at student-paced or live-race speed. Heavy on visuals and feedback-on-error. Best for independent practice, revision, or low-stakes review homework.
Kahoot Norms Free— game-show-style quizzes with time-pressured questions and a live leaderboard between rounds. Best for high-energy whole-class openers or closers. Use sparingly — novelty fades fast.
Mentimeter quick polls Free— live misconception check; pose a diagnostic question, see the bar chart on the board, teach into the gap.
Plickers Free— no devices needed for students — paper cards, teacher’s phone reads the room in seconds.
Desmos Classroom dashboard Free— teacher sees every student’s answer live during a card sort or polygraph — formative for Math.
AI dialogue AI-mediated dialogue tools (P4-P6 with care)
AI takes the first pass at a student's open-ended thinking, the teacher reviews afterward. Best for assessing reasoning, not just answers. SLS LEA (above) is the MOE-native version of this category and should be your first port of call.
Prerequisite
Students must complete the SLS AIEd Literacy Module before engaging in any AI-facilitated activities in the classroom. The module builds foundational understanding of how AI works, its benefits and limitations, and how to use it responsibly. Do not assign AI-dialogue tools to students who have not yet completed it.
Snorkl Norms Free— students explain their thinking by recording voice + drawing on a digital whiteboard. AI gives immediate feedback on their explanation against teacher-set criteria — what's right, what's a misconception, what to revise. Best when you want to assess reasoning, not just answers: Math working-out, Science explanations, 看图说话, oral fluency.
MagicSchool AI — Writing Feedback Tool, Rubric Generator, Norms Freestudents submit written work; the AI generates rubric-aligned feedback that the teacher reviews before release. Same pattern as SLS ShortAnsFA but works outside SLS — useful when the writing happens in Google Docs, Word, or on paper-then-photographed.
SchoolAI — Spaces & Mission Control Norms Free— students work with a teacher-configured AI tutor in a chat-style "Space"; the teacher monitors all conversations live from a dashboard ("Mission Control") that flags misconceptions and suggests prompts. Best for diagnostic conversations, Oral conversation practice and surfacing what students don't know they don't know.
Diagnostic Diagnostic check tools (no AI)
Quick teacher-controlled diagnostic checks without AI in the loop. Branching logic and rubric-based feedback do the work.
Google Forms/Microsoft Forms VPN with section-based branching Free— use "Go to section based on answer" on multiple-choice questions to route students to different feedback sections based on their response. A student who gets it wrong sees a re-teach section with worked examples; a student who gets it right sees an extension question. Diagnostic routing in any subject.
Google Classroom rubric + comment bank Free— reusable specific feedback at scale — write the comment once, apply many times.
Khan Academy mastery checks Free— auto-graded, immediate, and tracks student progress over time. Strong for Math practice.
NoRedInk / Quill.org Free— diagnostic grammar tools that show exactly which rule each student is missing.
Lesson exemplars · Maris Stella teachers
Add your lesson — submit via Good Practices.
Design tips
Complete the SLS AIEd Literacy Module first. Students must finish the module before engaging in any AI-facilitated activity — LEA, Snorkl, MagicSchool, SchoolAI, or any other AI-dialogue tool. The module equips them to understand how AI works, its limitations, and how to use it responsibly. This is a prerequisite, not optional enrichment.
Start with the SLS Feedback Assistants. ShortAnsFA, FA-Math, LangFA-EL, LangFA-CL, and LEA are all MOE-native, evaluated for safety, and integrated with your class data. Reach for external tools only when SLS doesn't have what you need.
Match the AfL category to the question. Live polling for "did the room get it?" — async feedback for "did this individual get it?" — AI dialogue for "can this student reason it out?" — diagnostic branching for "where exactly is this student stuck?" Different questions, different tools.
Auto-feedback is only as good as the wrong-answer hints. Invest the time in writing diagnostic distractors and the feedback for each — not just plausible-looking wrong answers.
Show students the rubric BEFORE the task. Make "what good looks like" visible from the start. Hidden criteria is not assessment, it's a guessing game.
Act on data within the same week. Otherwise it becomes summative-after-the-fact. The point of AfL is that it changes what you do next.
Treat AI-assisted feedback as a draft. Teacher judgement is the final layer — especially for language nuance, cultural context, and student-specific encouragement. This applies to ShortAnsFA, MagicSchool, SchoolAI, and Snorkl alike.
Don't let "data" replace conversation. A heat map tells you who is stuck; a 30-second conversation tells you why. Use the tool to find who to talk to.
Beware game-mode fatigue. Kahoot and Wayground lose their motivational pull if used every lesson. One game-style session per week is the usual sweet spot.