OCR Creative iMedia R093 · Exam guide

How the R093 exam is marked

What earns the marks on each kind of R093 question, from one-mark recalls to the nine-mark document questions, summarised in our own words from OCR's specification, mark schemes and examiner reports. Then exactly what SubjectHub's automarking checks, and where it stops.

How is R093 marked?

Short questions are point-marked: a mark for each correct point, and in describe and explain questions a second mark for developing it. The six- and nine-mark questions are levels-marked: the examiner reads the whole answer, decides which of three levels it fits best, then places it within that level. A work plan built from a brief is point-marked against the brief.

The paper at a glance

1 hour 30 minutes, 70 marks, 40% of the Cambridge National in Creative iMedia (J834).
Section A, 10 marks: 7 to 10 multiple-choice, closed and short-answer questions that recall knowledge, with at least one from each topic area.
Section B, 60 marks: one scenario that develops through the paper, with closed and short-answer questions and three extended answers. Two of the extended answers test analysis and evaluation; the third tests knowledge and its application.
On the four papers we analysed, Section B's extended answers were two nine-mark questions and one six-mark question, and June 2024 also had a six-mark work plan to build.

Point-marked questions

Most of the paper is marked point by point, and the command word tells you how many points a mark needs.

Identify, state, name, choose: one mark for each correct, relevant point. Short answers are fine.
Describe: give the feature and the detail that describes it accurately.
Explain: usually one mark for the point and one for the reason or effect. Words like ‘because’, ‘so’ and ‘this means that’ carry the second mark. ‘Explain two ways’ for four marks is two of these.
Complete, label, annotate: fill in what is missing until the sentence, table or diagram is finished.
In Section B, the point must fit the scenario. A correct fact about a different kind of product earns little; the same fact applied to this client's audience and purpose earns the marks.

Levels-marked questions

Six- and nine-mark questions are marked in three levels. A nine-mark question has Level 1 (1 to 3 marks), Level 2 (4 to 6) and Level 3 (7 to 9); a six-mark question has 1 to 2, 3 to 4 and 5 to 6. The examiner starts at the top level and works down until the description fits the answer as a whole, then decides the mark: the bottom of the level for an answer only just there, the top for one that meets the description consistently.

The levels move up with the same few qualities: how far the answer uses the scenario, how many relevant points it makes, how well it explains them rather than just stating them, and how consistently it uses the right terminology. A short list of correct points with no explanation stays in Level 1, however many points there are.

The nine-mark questions are about documents

On every R093 paper we have analysed, both nine-mark questions were about a pre-production document. They come in two forms:

Discuss the suitability of a document for a named user, in writing: a storyboard for a video editor (sample paper), a digital mood board for a content creator (January 2024), a visualisation diagram for an animator (June 2024) and a script for a director (January 2025).
Improve the effectiveness of a draft document for a named user, by drawing the improved version: a visualisation diagram (sample paper), a storyboard for camera operators (January 2024), a wireframe for a web developer (June 2024) and a digital mind map for a graphic designer (January 2025).

Across the mark schemes, the levels for these questions look for four things:

The improvements: a range of them for Level 3, some for Level 2, a few for Level 1.
The components they cover: for a storyboard, shot types, angles, movement, timings, lighting, sound, transitions and locations; for a mind map, the central node, the nodes and sub-nodes, the links and how ideas are grouped; for a wireframe, the page structure, images, text, navigation, links and interactions.
The conventions: applied consistently and effectively for the top level, adequately for the middle, in a limited way at the bottom.
The annotations: at Level 3, most of them show how the document now suits the needs of its named user. Annotations that only describe what was changed belong lower down.

OCR's examiner reports for January 2024 and January 2025 make the same point: the candidates who missed the highest marks did not link their improvements to the user of the document. ‘I added timings’ describes a change; ‘I added timings so the camera operator knows how long to hold each shot’ explains why it helps the user. The topic pages show how for mood boards, mind maps and visualisation diagrams and for storyboards, scripts and wireframes.

Work plans: points, not levels

When the June 2024 paper asked candidates to create a work plan from a client brief, it was marked by points: marks for giving the tasks the right durations, for keeping to the brief's rule on how many tasks could run at once, for contingency in the right places, and for no overlaps. The last mark only counted if the task column actually named the tasks in the brief. Read every number and rule in the brief and make sure the plan shows it. See work plans, flow charts and asset logs.

What SubjectHub's automarking checks

SubjectHub marks R093 answers as soon as they are submitted, following the same structure as OCR: points where OCR uses points, three levels where OCR uses levels. Here is exactly what it does.

Closed questions

Multiple choice, true or false, matching and similar questions are marked instantly by code against the answer key.

Short and extended written answers

Written answers are marked by AI against the question's own mark scheme. For six- and nine-mark answers it is given the levels and the indicative content, decides the level first on a best-fit reading of the whole answer, and only then the mark within it. The student sees the mark, what was credited and what to do next.

Pre-production documents: the checks before any marking

Storyboards, mind maps, mood boards, visualisation diagrams, wireframes and work plans are built in the browser. Before any AI is involved, code gives no marks to an empty document, to an improve question where the draft has been handed back unchanged, and to placeholders such as ‘Idea 1’. Answers that copy the question back, stuff keywords, or include a message asking the marker for marks also get no marks and are flagged for the teacher.

The levels, and the caps that hold them

For storyboards and the canvas documents (mood boards, visualisation diagrams and wireframes), the AI reads a description of the document, such as each scene's technical details or where each element sits on the page, and picks the OCR level. It does this three times independently and the middle result is kept, so one odd judgement cannot decide the mark. Code then keeps the mark inside the chosen level and applies caps for anything the levels rule out:

The brief: each thing the brief asks for is checked. A document that is not about this brief cannot go above Level 1. One that misses most of what the brief asks for is held low (Level 1 for a mood board, visualisation diagram or wireframe; the bottom of Level 2 for a storyboard missing half or more), and missing any one item means it cannot get full marks. Choices that work against the feel the brief asks for, such as ‘warm, honest and hand-crafted’, hold it to the middle of Level 2.
The annotation strand: with no annotations a document cannot pass the middle of Level 2. If the annotations only describe what was done, or only some explain how the choice helps the named user, it cannot pass the top of Level 2: Level 3 needs most of them to explain the benefit to that user. The same note pasted on every element earns nothing and is flagged for the teacher.
The conventions: a storyboard must use its technical details consistently in every scene, and its timings must add up to the length in the brief, for full marks. A wireframe with no navigation, links or interactions is treated as a visualisation of content, as OCR's mark scheme does, and stops at the top of Level 2.

Improve questions are marked on the improvement

On an improve question, what was already in the draft earns nothing: only the changes the student makes count. No real change means no marks. A few improvements hold the mark at the top of Level 1, and some improvements without the range Level 3 needs hold it at the top of Level 2. A problem from the draft that is still there stops the answer getting full marks, and one the code can see is plainly still there (or, on a mind map, two or more) holds it at the top of Level 2.

Mind maps: the AI classifies, code counts

For mind maps, the AI never awards marks. It classifies each branch (is it a real idea for this brief, which of the brief's key areas does it cover, how many genuine sub-ideas does it have, is anything under the wrong branch) and each note (does it explain how the change helps the user, or only say what changed). Code then works out the mark from the structure and how many of the brief's key areas are covered and, on improve questions, applies the same annotation and improvement caps.

Work plans: no AI at all

A work plan is marked by code against points written for each brief, as OCR marks them: each task's duration and position, the order the brief requires, how many tasks run at once, contingency where the brief asks for it, the deadline milestone, and OCR's rule that a plan only earns its marks when its tasks are the ones in the brief. The same plan gets the same mark every time.

What the student sees, and what the teacher can do

Every document answer comes back with its level and mark, a short summary, what held the mark back, a checklist of what was and was not done, and one next step. Automarking gives a first mark, not the last word: teachers see every answer and can change any mark, unusual answers are flagged for them, and a school that turns AI marking off has these answers sent to the teacher to mark instead.

The marking was tested against sets of answers written to each level of OCR's mark schemes, from full marks to empty, plus deliberate attempts to game it, and the rules above are the result. You can try it on the free R093 mini-mock with an account.

Questions about R093 marking

How many levels are there in R093 mark schemes?

Three. On a nine-mark question Level 1 is 1 to 3 marks, Level 2 is 4 to 6 and Level 3 is 7 to 9; on a six-mark question the levels are 1 to 2, 3 to 4 and 5 to 6. An answer with nothing worth credit gets 0.

How do you get Level 3 on an R093 nine-mark document question?

Make a range of improvements that cover a range of the document's components, apply its conventions consistently, and annotate most of your changes to explain how each one helps the named user, such as the camera operator or graphic designer. OCR's examiner reports say the candidates who missed the highest marks had not linked their improvements to that user.

Are the R093 nine-mark questions always about pre-production documents?

They have been so far. On OCR's sample paper and the January 2024, June 2024 and January 2025 papers, both nine-mark questions asked candidates to discuss whether a document suited its user or to improve one: storyboards, a digital mood board, visualisation diagrams, a wireframe, a digital mind map and a script. The specification does not promise this, so revise every topic.

Does SubjectHub mark R093 the way OCR does?

It follows OCR's mark schemes: points for short answers, three levels for extended answers and documents, and points for a work plan built from a brief. The AI picks a level and code keeps the mark inside it and applies caps for anything the levels rule out, such as annotations that do not explain the benefit to the user. Teachers see every mark and can change it.

Back to the R093 revision hub. R093 and Cambridge Nationals are OCR qualifications; this guide is SubjectHub's own summary and is not endorsed by OCR.