Plan record: Consistent marking: can predicted grades become a guide students can trust?

Harbour International School · DP · Led by Anna Vester · September 2026 to June 2028

Follows the IB's part order for programme development. Drawn from the record on 22 October 2026.

Focus

How might we make internal marking consistent across subjects, so predicted grades become a guide students can trust?

Serves the school priority: The same work gets the same grade, whoever marks it, in MYP and DP.

At the last comparison (12 Jul 2026), 15 of 22 predicted grades were within one grade of the final result. Biology was 9 of 12; History was 6 of 10.
Predictions sat high where the internal assessment was still a draft. In History, the essay mark and the predicted grade used different ideas of analysis.
The same criterion gets different marks in different rooms. Who agrees the standard has been an open question since 12 May 2026.

Question	Practice	Category	Code
What must be in place?	Assessment is run consistently and fairly	Learning: approaches to assessment	0404-03 to check
Who is responsible?	Leadership and teachers take part in timely professional learning	Environment: teacher support	0203-02 to check
How will the work be done?	Students and teachers use feedback to support stated outcomes	Learning: approaches to assessment	0404-01-0100 to check
Why this focus?	To choose	Purpose

For every DP student: predicted grades they can plan around. The measure: share of predicted grades within one grade of the final result.

Reading	Date	Value
Baseline	12 Jul 2026	68% (15 of 22)
Mocks against predictions	Feb 2027	to record
Results against predictions	Jul 2027	to record
Target	May 2028	80%

The measure readings above.
After the December cycle: the share of moderated samples within one band, per subject group.
Before each report round: whether every predicted grade was compared in a group of three subjects.

Harlen (2004), EPPI-Centre review of teachers' summative assessment: moderation through professional collaboration helps; reliable assessment needs time for teachers to meet.
Garry, McCool and O'Neill (2005), on anchoring in marking: a marker who has seen the first mark moves less than two independent markers. Hence "mark blind first, talk second".
What it can't say: one cycle won't prove scripts land within one band at Harbour.

Action	Lead (shares)	When	From decision	How we'll know	Status
Publish the internal assessment calendar draft	Anna Vester	3 days late	Anna owns the IA calendar; subject leaders flag clashes by Wednesday.	No week holds more than two major IA deadlines	Overdue
One joint MYP 5–DP 1 standardisation cycle in every subject group	Marco Ellison (Anna Vester)	12 Nov to 15 Dec 2026	Waits on the open decision: who agrees the standard across programmes (Jonas Berg)	Each subject group has one moderated sample	Waiting on Jonas
TOK exhibition internal standardisation	Anna Vester	By 1 Mar 2027	TOK exhibitions are standardised internally on one shared student example.	Teachers mark the shared example within one band	Moving
Predicted grade conversations	Helen Okonkwo	12 Jan to 1 Apr 2027	Predicted grades are compared across three subjects before reports go out.	Every predicted grade compared in a group of three before reports	Not started

Standardise one TOK exhibition example together (Mar 2026): Partly worked Two of three teachers now mark the example within one band.
Retire the old predicted-grade spreadsheet (Sep 2026): Didn't work Two departments still sent the old sheet in the spring round. Decided again: Predicted grades are compared across three subjects before reports go out.