Why This Matters

"Good effort — level 5."

You've written something like that. Most teachers have, especially at 11pm with fifteen essays still in the pile. It feels like feedback. It isn't — a student reading it has no idea what to do differently next time, and you, if a colleague asked you to justify that level tomorrow, might struggle too. Feedback is the highest-leverage, lowest-cost thing you do as a teacher; a good comment can change what a student does this week. And the moment right after — turning your judgement into a specific level you can defend — is the moment that makes new MYP teachers most anxious. This article is about doing both well.

Understand the Idea

Useful feedback tends to answer three different questions, sometimes called feed-up, feedback, and feed-forward. Feed-up answers "where am I going?" — it reminds the student what the target actually is. Feedback answers "how am I doing?" — it names the gap between where they are and that target, using evidence from their actual work. Feed-forward answers "what next?" — it names one specific, doable action. Comments that only do the first two ("here's the goal, here's how you fell short") leave a student informed but stuck. The feed-forward piece is what makes feedback actually move learning, and it's the piece most often missing.

The single biggest lever for whether feedback works is specificity. "Be more analytical" is not an action a student can take on Tuesday afternoon — it's a restatement of the goal, dressed up as advice. "In your second paragraph, you describe what happened but don't say why it mattered to the people involved — try adding one sentence after your evidence that explains its significance" is an action. The difference isn't tone or effort; it's whether the sentence tells the student a specific thing to do, tied to a specific place in their own work.

Once you've given feedback, you still have to arrive at a level — and this is where best-fit judgement (introduced in Article 22) does its real work. You're matching the whole piece of evidence to the achievement level descriptor it resembles most closely, strand by strand where relevant (Article 21), not counting up points. What makes a level defensible isn't certainty — even experienced teachers disagree at first pass — it's being able to point to specific evidence in the work and specific language in the descriptor and show how they connect.

That's also why standardisation and moderation exist as team practices, not just compliance exercises. Standardisation is when a department looks at the same piece of work together, before or alongside live marking, to calibrate what "meets this descriptor" looks like in practice. Moderation is checking marked work afterward for consistency across teachers. Neither is about forcing everyone to agree by splitting the difference — it's about making sure disagreements get resolved by returning to the evidence and the descriptor, not by compromise. 📌 Official source: current requirements and recommended processes for standardisation and moderation are set by the IB and may vary by school context — check your current MYP guidance and your school's assessment policy.

⚠️ Common trap: Two teachers disagree on a level and simply average their two numbers to "split the difference." That resolves the argument but not the judgement — neither teacher has actually reread the descriptor evidence, so the agreed number may not be defensible to either of them, or to a student who asks why.

See It in Practice

Same piece of student work — a short investigative paragraph from the "change and industrialisation" unit — with three different feedback comments.

Ineffective: "Nice work, but needs more depth. Level 5." Why it fails: No feed-up (what does "depth" mean here?), vague feedback (no evidence cited), no feed-forward (no action named).

Adequate: "You've described the source well, but you need to evaluate it more. Level 5." Why it's only adequate: Correctly identifies the gap (description vs. evaluation, echoing the strand distinction from Article 21) but still gives no specific action and no pointer to where in the text this shows up.

Effective: "Your third sentence describes the source accurately — good detail on who wrote it and when. But you stop there. Try adding: 'This means the source can tell us ___, but probably can't tell us ___, because ___.' That sentence structure will push you from describing the source into evaluating it, which is what this strand is actually asking for. Level 5, upper band — the description is strong, but the paragraph needs that evaluative move to reach the top band."

Notice the third comment does all three jobs: it points to specific evidence (the third sentence), it names the specific gap, and it hands the student a concrete sentence frame to use next time — plus it connects the level directly to strand language rather than stating it in isolation.

Two-teacher moderation disagreement, resolved properly: Teacher A reads the same paragraph and awards a level 3, reasoning "there's barely any evaluation here, it's mostly description." Teacher B awards a level 6, reasoning "the description itself is genuinely strong and there's a glimmer of evaluative language in the last line." That's roughly a two-level gap — worth taking seriously, not smoothing over.

Instead of averaging to a level 4 or 5, the two teachers reread the actual descriptor language for bands 3–4, 5–6, and 7–8 together, then reread the paragraph against each band specifically. They notice band 5–6 language (illustratively) describes "some evaluation, not yet consistent" — and the paragraph has exactly one evaluative sentence, not zero and not several. They land on level 5 together — not as a compromise number, but because rereading the evidence against the actual descriptor showed that band fit best. The number happens to sit near the middle of their two original guesses, but it was earned by evidence, not manufactured by averaging.

Apply It

🛑 Do this now.
  1. Take two invented (or real, anonymized) student work samples for the same task — make one noticeably stronger than the other, but not by an enormous margin.
  2. For each, write a full piece of feedback that includes feed-up, feedback, and feed-forward — following the "effective" model above. Name specific evidence in the sample and one concrete next action.
  3. For each sample, assign a best-fit level and write exactly one sentence justifying it, citing both the descriptor band language and the specific evidence in the work that matches it.
  4. Imagine handing both your feedback and your one-sentence justification to a colleague and asking: "Would you have landed here too?" Note anywhere your justification sentence feels thin — that's usually a sign you matched the level too quickly.

Artefact to produce: Two full feedback comments plus two one-sentence level justifications, written as if you were about to defend them in a real moderation conversation.

Avoid This

Feedback that restates the goal instead of naming an action. Symptom: your comments could be pasted onto any student's work unchanged. Fix: cite something specific from this piece of work every time.

Assigning a level first and reverse-engineering the justification. Symptom: your written rationale is vague ("just felt like a 5") when asked to explain it. Fix: read the descriptor bands first, then match — not the other way around.

Resolving moderation disagreements by averaging. Symptom: two teachers land on wildly different levels and simply split the difference to move on. Fix: return to the descriptor and the evidence together; let the number emerge from that, even if it takes longer.

Feedback with no student action at all. Symptom: comments read like a verdict, not a next step. Fix: every comment should contain at least one thing the student can actually do this week.

Check Yourself

Take one piece of feedback you gave recently. Does it name specific evidence from the student's own work, and does it end with something the student can actually do? If either answer is no, rewrite it using the effective model above before your next marking session.

Continue Learning

You've planned it, taught it, assessed it. Chapter 06 asks whether you can do it again — better, and alone. Next step → Chapter 06 moves from working through each stage with support to running the full plan-teach-assess cycle independently. Downloadable resource: Feedback Frames + Level Justification template. 📌 Official source: check current IB guidance on standardisation, moderation, and level determination, alongside your school's assessment policy, before finalizing any reported grade.