Question-level analysis after IGCSE mocks
A question-level analysis (QLA) records each student's mark on each question part, not just their total. You can then see which questions, topics and skills lost the most marks across a class or cohort. To run one, mark and record per question, compare each question's average with its maximum, group the weakest questions by topic and skill, and plan reteaching around the top three gaps. Students within a few marks of a grade boundary become your intervention list.
Why totals aren't enough
Two students both score 58/100 on 0580 Paper 4. One lost marks spread across the paper. The other lost almost everything on vectors and transformations. They need completely different next steps, and only per-question data shows the difference.
Across a cohort, a QLA also separates teaching gaps (everyone lost marks on 6(c)) from individual gaps (a few students did).
What to record
At a minimum, for each student:
| Column | Example |
|---|---|
| Student | (name or ID) |
| Class / teacher | 11B / Mr A |
| Tier or route | Extended |
| Mark per question part | Q1a = 2/2, Q1b = 1/3, … |
| Total | 58/100 |
| Mock grade (estimate) | B (June 2025 thresholds) |
And for each question part, once per paper:
| Column | Example |
|---|---|
| Question | 6(c) |
| Max marks | 3 |
| Topic | Vectors |
| Skill or AO | Problem solving |
| Command word | Show that |
Mapping questions to topics takes about 20 minutes per paper. Do it once and reuse it every time you set that paper.
The three views that matter
1. Percentage of marks gained per question
Divide the class average by the maximum for each question part and sort from lowest. The bottom five are your first reteaching candidates.
2. Topic and skill summary
Group the question percentages by topic and by skill. Patterns appear quickly. For example: "Students can carry out the calculations but lose marks on 'explain' questions across all topics." That's a skill gap, not a content gap. The fix is practice with command words, not reteaching the content.
3. Boundary students
List students within a few marks of the next grade, using the right threshold table (see IGCSE mock grade boundaries). Then check which questions they lost marks on. Small, targeted gains matter most here.
Reading the patterns
| Pattern | Likely cause | Response |
|---|---|---|
| Low marks on one question across all classes | Topic not taught well, or the question was unusually hard | Reteach the topic. Check the examiner report to see whether students worldwide found it hard too. |
| Low marks in one class only | Teaching or marking difference | Check moderation first, then the teaching sequence. |
| High method marks, low accuracy marks | Careless arithmetic, rounding, units | Short daily practice and a final-answer checklist. |
| Low marks on extended "explain" questions | Answer structure, command words | Model answers and structure scaffolds. |
| Blank answers late in the paper | Timing | Timed practice, and advice on question order. |
The Principal Examiner Report for Teachers for the same paper helps here. If examiners report that candidates worldwide struggled with a question, your students' low score may say more about the question than about your teaching.
From analysis to a reteaching plan
Keep it small enough to actually happen:
- Three whole-class priorities: the lowest-scoring topics or skills.
- One lesson or starter sequence per priority, ideally before the next assessment.
- Small-group intervention for boundary students, built around their specific lost marks.
- Individual feedback telling each student their two biggest gains: "On question 6, you lost 5 marks on vectors. Practise these three questions."
- A re-test of the priority questions 2–4 weeks later to check the reteaching worked.
The bottleneck: getting per-question data
Teachers know QLA is worth doing. They skip it because typing every student's mark for every question part into a spreadsheet takes hours after marking is done.
If the marking itself records per-question marks, most of the data entry disappears. The Learning Mate marks each question and part separately, so every student's per-question marks are available to review as soon as marking finishes. Totals export to Excel, and the built-in analytics show results across students and exams. Each student also gets written feedback on each question, which covers step 4 above.
FAQ
What is a question-level analysis?
A record of every student's mark on every question part of an assessment. It's used to find which questions, topics and skills lost the most marks across a class or cohort.
How long does a QLA take?
Mapping a paper's questions to topics takes around 20 minutes and can be reused. The slow part is entering marks per question. A marking tool that records marks per question part saves most of that time.
What should I do with mock QLA results?
Choose three whole-class priorities, plan short reteaching for each, give boundary students targeted support, and re-test those questions a few weeks later.
Should I compare my QLA with the examiner report?
Yes. If the Principal Examiner Report says candidates generally struggled with a question, your students' result on it is less likely to reflect a gap specific to your teaching.
Try it
Mark a class set with marks recorded for every question part. Start with one paper at thelearningmate.com/trial, with no account needed.