What standardisation is for
Criterion-based assessment depends on teachers reading the descriptors the same way. They do not, unless they practise. A standardisation meeting is that practice: teachers of the same subject and year mark the same samples independently, compare the levels they gave, and work out why they differ. The aim is not for everyone to agree on the day. It is for everyone to leave marking the way the group decided, and for the decision to be written down.
The IB expects schools to standardise internally wherever more than one teacher assesses the same subject in the same year, and to standardise the personal project and community project. Beyond the requirement, it is the single cheapest thing a school can do to make its grades mean something. A school that standardises well can explain any grade it gives; one that does not is guessing more than it knows.
Before the meeting: samples, task and clarifications
A meeting without preparation becomes a discussion of teaching. The preparation is short and specific.
- Choose one task that every teacher in the group has set, or a task close enough that the criteria and clarifications are shared.
- Pull four to six samples across the range: one that is clearly strong, one clearly weak, and the rest from the middle where the disagreements live. Remove names.
- Circulate the task sheet, the criterion with its strands, and the task-specific clarifications at least two days ahead, so the meeting is spent comparing judgements rather than reading.
- Ask everyone to mark independently before arriving and to write one line per sample saying which strand decided the band. The line is the useful part.
- Name a chair who is not the head of department, so the department head's marking is compared like everyone else's.
During the meeting: compare, then argue by strand
Start by collecting the levels for every sample without discussion; a table on the board with teachers down the side and samples across the top. Disagreements of one level within a band are normal and need no time. Disagreements across a band are the agenda.
Take each of those in turn. The rule that keeps the meeting to an hour is that the argument is about a strand, not about the student or the task. Which strand did the higher marker see met that the lower marker did not? Read the descriptor for that strand aloud, read the work again, and decide. Most cross-band disagreements come down to one teacher crediting effort or content that the strand does not ask for.
When the group settles a sample, record the agreed level, the strand that decided it, and one sentence of reasoning. When the group cannot settle it, record that too, and pick the band the descriptors support rather than the band the majority gave. A vote is not a judgement.
After the meeting: anchors, re-marking and the record
The output of a standardisation meeting is a set of anchor samples: pieces of work the department agrees sit at particular levels, with the reasoning attached. Keep them. A new teacher who joins in January should be able to read the anchors and mark like the department by February.
If the meeting showed that one teacher was consistently a band adrift, the work already marked by that teacher is re-marked or moderated before it is reported. That is uncomfortable and it is the point. Standardisation that changes nothing has not happened.
Write a short record: date, task, criterion, who attended, the agreed levels, the decisions about clarifications, and any change to marking that follows. It is the evidence an evaluation visit asks for when it asks how the school ensures consistency of assessment, and it is far easier to keep as you go than to reconstruct.
How often, and with whom
Once a term per subject group is a realistic floor for a department with several teachers in a year; once a year is too rare, because teachers drift apart within a term. The meeting need not be long. Four samples and one criterion, well prepared, fits in fifty minutes. The habit matters more than the length.
Cross-subject standardisation is worth doing occasionally, because the general grade boundaries assume that a 5 in one subject group describes the same quality as a 5 in another. A session where a science teacher and a humanities teacher read each other's anchors is the only way to test that assumption inside a school.
For the personal project, standardisation is a requirement and the samples should include projects from every supervisor. Supervisors mark with the affection of eight months' work behind them; the meeting is where the affection is corrected.
A one-page routine
For a department that has never done this, the routine below is enough to start. Refine it once it has run twice.
- Week 1 — choose the task and criterion; collect and anonymise four to six samples; circulate task, criterion, clarifications.
- Week 2 — everyone marks independently and writes one line per sample naming the deciding strand.
- Meeting, 50 minutes — table the levels; discuss only cross-band disagreements, strand by strand; record agreed levels and reasoning.
- Same week — file the anchors and the record; agree any re-marking; update the task-specific clarifications if the meeting showed they were unclear.
- Next term — repeat with a different criterion, so all four are standardised across the year.
