Four criteria, one subject group at a time
Each MYP subject group has its own four criteria, labelled A to D, and they are not the same four across subjects. In language and literature they are analysing, organising, producing text and using language. In mathematics they are knowing and understanding, investigating patterns, communicating, and applying mathematics in real-life contexts. In sciences, criterion B is inquiring and designing and criterion D is reflecting on the impacts of science. A teacher moving between subject groups should expect to relearn the criteria, not carry them across.
Each criterion is broken into strands, usually three or four, numbered with roman numerals. The strands are what a task actually asks a student to show. A task that touches only strand i of criterion B is not evidence for criterion B as a whole, and the achievement level you award has to reflect that.
The criteria are published for years 1, 3 and 5 of the programme, with the descriptors becoming more demanding at each stage. A school teaching year 2 or year 4 uses the descriptors from the year below or the year above, and should say which in its assessment policy.
The 0–8 scale and its four bands
Every criterion is marked on a scale of 0 to 8. The descriptors are written in bands: 1–2, 3–4, 5–6 and 7–8, with 0 reserved for work that does not reach the lowest descriptor. Within a band, the lower number means the student's work fits the band's description but not securely; the higher number means it fits it comfortably.
The consequence that surprises people: the scale is not a percentage and does not behave like one. A 4 is not half marks. It means the work matches the second band, which for most criteria describes adequate, partly developed work. Parents used to percentages read a 4 out of 8 as fifty per cent and a 6 as seventy-five; neither reading is right, and it is worth saying so on the report.
- 0 — the work does not reach the standard described by any of the descriptors below.
- 1–2 — limited: the student shows the beginnings of what the criterion asks for.
- 3–4 — adequate: the strands are present, with gaps or a lack of depth.
- 5–6 — substantial: the work meets the criterion with some consistency and some sophistication.
- 7–8 — excellent: consistent, thorough, with the qualities the top descriptor names present throughout.
Best fit, not arithmetic
The MYP asks teachers to make a best-fit judgement. Read the whole body of evidence for a criterion, read the four band descriptors, and decide which band describes the work best; then decide whether it sits at the top or the bottom of that band. It is a judgement about a description, not a calculation.
This is why averaging is wrong, and explicitly not permitted. A student who scored 2, 8 and 8 on three tasks for criterion C has not earned a 6. If the 2 was in September and the 8s were in May, the evidence says the student now works at the top band, and best fit says 8. If the 2 was last week, the picture is different. Recency matters, task demand matters, and a low score on a task that assessed one strand should not drag down a criterion the student has since shown in full.
Task-specific clarifications make this workable. For each summative task, the teacher rewrites the band descriptors in terms of that task: what does substantial organising look like in this particular essay? Clarifications are written before the task is set, shared with students, and used again when the task is marked. Without them, two teachers will read the same descriptor differently and never know it.
How often each criterion must be seen
For a final judgement to be defensible, each criterion has to be assessed at least twice in each year of the programme. Twice is the floor, not the target. A criterion assessed twice, with one of the two tasks covering a single strand, gives a teacher very little to make a best-fit judgement from, and very little to show a parent who asks how the grade was reached.
Planning across the year is the fix. Lay the summative tasks out against the four criteria before term starts, check that each criterion appears at least twice with all its strands covered at least once, and check that no single unit is trying to assess all four. Two criteria assessed properly in a unit is better evidence than four assessed in passing, and the unit planner guide on this site says the same from the planning side.
From 32 to a grade from 1 to 7
At the reporting point, the four criterion levels are added to give a total out of 32, and the total is converted to a final grade from 1 to 7 using the general grade boundaries the IB publishes. The boundaries are the same for every subject group, which is part of the point: a 5 in mathematics and a 5 in the arts are meant to describe the same quality of achievement.
The current general boundaries are: 1–5 gives a grade 1, 6–9 a grade 2, 10–14 a grade 3, 15–18 a grade 4, 19–23 a grade 5, 24–27 a grade 6 and 28–32 a grade 7. Each final grade also has a descriptor, and the descriptor is what a report should explain, not the number. The authority for the boundaries and the descriptors is the IB's MYP: From principles into practice; check the current edition rather than a table remembered from training.
A grade 4, then, is a total of 15 to 18, which is a student whose four criterion levels average roughly 4: adequate work across the board, or substantial in two areas and limited in two others. The total hides that difference, which is why the criterion levels should travel with the grade wherever it is reported.
Where schools most often go wrong
The same handful of problems appear whenever MYP assessment is reviewed, and none of them is about not knowing the rules.
- Converting percentages to levels. A test marked out of 50 and mapped to 0–8 by dividing is not a criterion-based judgement, and it produces levels that do not match the descriptors.
- Assessing all four criteria on every task. The judgements get thin and the task gets overloaded.
- Averaging levels at the end of the year because it feels fairer. It is not fairer; it ignores growth and punishes early attempts.
- Writing task-specific clarifications after marking, to justify levels already given.
- Reporting a grade without the criterion levels beside it, so nobody outside the classroom can see what the grade is made of.
- Standardising once a year, or never. Two teachers in the same subject drift apart within a term unless they mark samples together.
What to tell students and parents
Students do better under criterion-based assessment when they can read the descriptors themselves. Put the task-specific clarifications on the task sheet, mark against them visibly, and let students see which strand cost them the band. A comment that names the strand is worth more than a comment that names the level.
For parents, the message is that a level describes the quality of work against a published standard, that it is not a percentage, and that the final grade is a best-fit judgement across the year rather than a running average. Most confusion at report time comes from those three points, and all three can be explained in a paragraph on the report itself.
