Study
GCSE Business AI marking, whole mock papers
Accuracy: 23 of 23 GCSE Business mock scripts were within 4 marks of an independent examiner on a 43 mark paper. Average gap of 2.3 marks.
- Scripts
- 23
- Paper worth
- 43 marks
- Handwritten
- 23 of 23
- Examiner
- Independent
- Engine version
- 2026.11.04
- Tested
- 6 November 2026
By DeepMark Labs Limited. Last updated 8 November 2026.
The result
An independent examiner marked a class set of GCSE Business mock papers by hand. DeepMark then marked the same scanned scripts. On a 43 mark paper, DeepMark’s total was within 4 marks of the examiner on 23 of 23 scripts and within 2 marks on 13. The average gap was 2.3 marks, and 2 scripts got exactly the same total.
Gap from the examiner's mark
Script by script
- Same mark as the examiner's mark
- Within 4 marks
- Scripts within 4 marks
- Scripts further away
- Larger dots hold more scripts.
Ranking correlation: 0.93. DeepMark and the examiner put strong and weak scripts in almost the same order. This measures order; the average gap is the accuracy figure.
Where the gap comes from
| Section | Marks | Average gap | DeepMark on average | Higher | Same | Lower |
|---|---|---|---|---|---|---|
| Short answers, 1 to 4 marks | 25 | 1.4 marks | 1.0 lower | 4 | 7 | 12 |
| Longer answers, 6 and 12 marks | 18 | 2.0 marks | 1.5 higher | 18 | 2 | 3 |
On the longer 6 and 12 mark answers, marks tend to run high: DeepMark gave more than the examiner on 18 of 23 scripts, by 1.5 marks on average. On the short answers it was slightly stricter, about 1 mark lower on average. The two partly cancel out in the totals, so a teacher reviewing the longer answers will catch most of the difference.
Accuracy over time
How we tested it
- 1A former examiner and practising teacher with no connection to DeepMark marked one class set of GCSE Business mock papers by hand, before DeepMark saw them. The examiner recorded a mark for each section and a total.
- 2The paper was worth 43 marks: short answers worth 1 to 4 marks, then longer answers worth 6 and 12 marks.
- 3We scanned the handwritten scripts and marked them with DeepMark, the same way a teacher would. Reading the handwriting is part of what is measured.
- 4We compared DeepMark's section marks and total with the examiner's for every script. One student was absent, and the examiner's marks for 2 scripts could not be confirmed, so those were left out.
- 5Student names were removed and each student was given a number.
What this does and does not show
The examiner’s mark is one experienced marker’s judgement, not a published definitive mark. Examiners differ from each other too: two trained examiners agree on an individual 6 mark question only 46% to 75% of the time (Ofqual, Marking Consistency Metrics, 2018), so some of the gap here is ordinary marker disagreement.
This is 23 scripts from one class on one paper, and the examiner recorded section totals rather than question marks. Treat it as an early result for GCSE Business, alongside the larger studies on this site.
Every mark
Every script in the study, with the examiner’s total and DeepMark’s total. We publish marks only, never student writing.
Every mark, answer by answer (23 scripts)Show
| Script | Paper worth | Examiner's mark | DeepMark mark | Difference |
|---|---|---|---|---|
| 1 | 43 | 32 | 35 | 3 marks higher |
| 2 | 43 | 32 | 32 | Exact |
| 3 | 43 | 31 | 28 | 3 marks lower |
| 4 | 43 | 31 | 27 | 4 marks lower |
| 5 | 43 | 30 | 32 | 2 marks higher |
| 6 | 43 | 30 | 31 | 1 mark higher |
| 7 | 43 | 29 | 29 | Exact |
| 8 | 43 | 29 | 27 | 2 marks lower |
| 9 | 43 | 29 | 26 | 3 marks lower |
| 10 | 43 | 27 | 30 | 3 marks higher |
| 11 | 43 | 26 | 28 | 2 marks higher |
| 12 | 43 | 25 | 27 | 2 marks higher |
| 13 | 43 | 25 | 27 | 2 marks higher |
| 14 | 43 | 25 | 21 | 4 marks lower |
| 15 | 43 | 24 | 25 | 1 mark higher |
| 16 | 43 | 20 | 24 | 4 marks higher |
| 17 | 43 | 20 | 21 | 1 mark higher |
| 18 | 43 | 20 | 18 | 2 marks lower |
| 19 | 43 | 18 | 16 | 2 marks lower |
| 20 | 43 | 17 | 21 | 4 marks higher |
| 21 | 43 | 16 | 20 | 4 marks higher |
| 22 | 43 | 13 | 16 | 3 marks higher |
| 23 | 43 | 5 | 6 | 1 mark higher |
How to read these numbers
- Average gap
- How far DeepMark's mark was from the examiner's mark, on average, ignoring direction. A gap of 1.3 marks means a typical answer was a little over 1 mark away.
- Matched exactly
- DeepMark gave the same whole mark as the examiner's mark. This is the strictest test, and only possible because every DeepMark mark is a whole mark.
- Within 1 mark
- DeepMark's mark was the examiner's mark or 1 mark either side.
- Ranking correlation
- How closely DeepMark puts answers in the same order as the examiner's mark, from 0 to 1. It measures order, not accuracy, so we always show it beside the average gap, never instead of it.
Questions teachers ask
How accurate is DeepMark on GCSE Business?
On 23 handwritten GCSE Business mock scripts worth 43 marks, DeepMark's total was within 4 marks of an independent examiner on 23 and within 2 marks on 13. The average gap was 2.3 marks.
Who marked the scripts?
A former examiner and practising teacher with no connection to DeepMark marked the class set by hand before DeepMark saw it. DeepMark then marked the same scanned scripts.
Where should a teacher check?
On the longer 6 and 12 mark answers, marks tend to run high: DeepMark gave more than the examiner on 18 of 23 scripts, by 1.5 marks on average. On the short answers it was slightly lower, so the two partly cancel out in the totals.
Other studies
- AI marking repeatability264 answers, 76 papers, 185 questions, 48 teachers, 1,320 runs
- GCSE Geography AI marking, 1 to 9 marks104 exemplars on questions worth 1 to 9 marks
- GCSE History AI marking, up to 16 marks57 exemplars on questions worth 4, 8 and 16 marks
- GCSE English Language reading AI marking, 15 marks31 exemplars on 15 mark reading questions
- A level History AI marking, 25 marks44 exemplars