Computer Science Extended Essay: A Guide for 2027
By Michael Thompson · Education Specialist; 10 years teaching the IB at Bromsgrove School · Published 4 October 2026 · Updated 6 October 2026
A computer science extended essay asks you to answer a focused research question through evidence, analysis, and argument - not to build a product or write a programming tutorial. The 2027 assessment model applies to every student sitting the May 2027 session or later, and it brings a revised 30-mark framework that replaces the previous 34-mark model. Getting the subject choice right matters: computer science suits students who want to test something measurable - an algorithm's performance, a data structure's efficiency, a machine learning model's accuracy - and who can write about that test with precision. This guide covers the criteria, structure, topic selection, and the key differences from the internal assessment so you know exactly what you are producing before you start.
Key Takeaways
- New 30-mark criteria from May 2027: The updated assessment uses five criteria totalling 30 marks; the previous 34-mark model is used for the last time in November 2026.
- 4,000-word limit on the written essay: The word count covers the introduction, body, conclusion, and quotations - title page, contents, and references sit outside this count.
- Three mandatory reflection sessions: These culminate in a 20-30 minute viva voce with your supervisor, plus a 500-word reflective statement on the Reflection and Progress Form.
- Discussion and evaluation carries the most marks: At 8 out of 30, this criterion rewards critical appraisal of your method and honest acknowledgement of its limitations.
- The EE is a research essay, not a product: Unlike the IA - a computational solution to a real-world problem - the EE answers a question, so a build log or code project is not the right output.
- Subject-specific rules sit in the full EE guide: The IB Programme Resource Centre holds the complete guidance; ask your supervisor to share the relevant sections before you finalise your approach.
In This Article
- What Is the Extended Essay in Computer Science?
- Assessment Criteria and Marks for 2027
- Structure and Word Count Requirements
- Computer Science EE Topics and Research Question Ideas
- How the EE Differs from the Computer Science IA
- Common Pitfalls to Avoid in a Computer Science Extended Essay
- What to Do Next
1. What Is the Extended Essay in Computer Science?
The computer science extended essay is a compulsory, externally assessed component of the IB Diploma Programme core, answered through a single focused research question. Per the IB, the updated EE launches for first assessment in May 2027, and its new criteria apply to every student sitting that session or later.
> Note on the previous model: November 2026 is the last session assessed on the old 34-mark framework. If you are sitting May 2027 or later, the previous model does not apply to you.
Two research pathways are available. You can write a subject-focused EE anchored in one DP subject, or an interdisciplinary EE combining two DP subjects. Computer science can anchor either pathway, whether you pair it with mathematics to examine algorithmic complexity or keep it subject-focused on a single computing problem.
Computer science suits the EE format well for a practical reason: investigations in this subject tend to produce quantifiable, reproducible results. When you compare two algorithms or test a data structure's performance, you can define clear independent and dependent variables, run controlled tests, and present measurable outcomes. That makes it easier to build the line of argument the assessment criteria reward, and harder to drift into the vague, descriptive writing that undermines weaker essays across all subjects.
2. Assessment Criteria and Marks for 2027
The 2027 EE is marked out of 30, spread across five criteria. Per the IB EE updates page, the breakdown is:
| Criterion | Label | Marks |
|---|---|---|
| Framework for the essay | Criterion A | 6 |
| Knowledge and understanding | Criterion B | 6 |
| Analysis and line of argument | Criterion C | 6 |
| Discussion and evaluation | Criterion D | 8 |
| Reflection | Criterion E | 4 |
| Total | 30 |
The same five criteria apply in every subject, including the chemistry extended essay, and on both pathways (subject-focused or interdisciplinary). No grade boundaries for the 30-mark scale will be published until after the May 2027 session, with results due 6 July 2027.
One non-obvious point worth noting: Discussion and evaluation (Criterion D) carries the most marks at 8, yet many students treat it as an afterthought. In a computer science essay, that criterion is where you critically appraise your method and its limits, including things like hardware variation between test runs or bias in a dataset you relied on.
How each criterion plays out in a computer science essay:
- Criterion A (Framework for the essay): A focused, researchable question and a clearly justified method, for example choosing to compare two algorithms on a defined performance metric rather than describing them broadly.
- Criterion B (Knowledge and understanding): Accurate use of computing terminology and concepts, drawing on credible technical sources.
- Criterion C (Analysis and line of argument): Quantitative or qualitative analysis of your results, with a coherent argument that runs through each section.
- Criterion D (Discussion and evaluation): Critical appraisal of what your findings actually show, the limits of your test design, and sources of error such as hardware bottlenecks or dataset bias.
- Criterion E (Reflection): Your evaluation of what the EE experience taught you as a learner, assessed on the 500-word reflective statement you write on the Reflection and Progress Form at the end of the process.
Subject-specific guidance for computer science sits in the full EE guide on the IB Programme Resource Centre. Ask your supervisor to share the relevant sections before you settle on a method.
3. Structure and Word Count Requirements
The IB subject brief sets a hard upper limit of 4,000 words, covering the introduction, body, conclusion, and any quotations. The title page, contents page, and references and bibliography sit outside the count, but quotations do not - a detail that catches students who quote long passages verbatim.
Required elements
Per the subject brief, every submission must contain:
- Title page
- Contents page
- Introduction
- Body
- Conclusion
- References and bibliography
What each section must do
- Introduction - establishes the academic context for your question and states the research question clearly. It should also signal how the investigation will proceed.
- Body - presents your evidence, analysis, and argument. In a computer science essay this typically means benchmarking data, pseudocode, or comparative analysis, built section by section toward a position.
- Conclusion - answers the research question directly, based only on what your investigation found. It must also acknowledge the limitations of the approach.
Reflections and the viva voce
Students complete three mandatory formal reflection sessions with their school supervisor across the process. The final session is a viva voce lasting about 20-30 minutes. At the end of the whole process, you write a 500-word reflective statement, recorded on the Reflection and Progress Form (RPF).
Practical tip: build your contents page last. Page numbers shift during editing, and a contents page generated before final revisions is almost always wrong by submission day.
4. Computer Science EE Topics and Research Question Ideas
Many strong research questions in a computer science extended essay produce measurable, comparable results rather than a descriptive tour of a technology. A well-scoped empirical question gives you data to analyse and evaluate, which is what the analysis and discussion criteria (14 of the 30 marks) reward.
Mapping topics to the 2027 syllabus
The 2027 CS syllabus divides into two themes, and either is fair game for an EE:
- Theme A - Computer fundamentals, Networks, Databases, Machine learning
- Theme B - Computational thinking, Programming, Object-oriented programming, Abstract data types (HL only)
B.4 Abstract data types is the only whole topic reserved for HL, though HL also goes much deeper in other topics, notably machine learning and object-oriented programming. The EE is not tied to your SL or HL entry, so an SL student with the background can still draw on it. Confirm this with your supervisor and the full EE guide on the IB Programme Resource Centre before committing.
Worked example research question shapes
The questions below were written for this guide. They are illustrative, not model answers.
| Theme area | Example question (written for this guide) | Key metrics |
|---|---|---|
| Algorithms (A/B) | "How does obstacle density affect execution time and nodes explored when comparing A* and Dijkstra's algorithm on fixed-size weighted grids?" | Execution time, nodes explored, path cost |
| Database indexing (A) | "How do B-tree and hash indexes compare in equality-lookup latency and storage overhead as table size grows, using increasing subsets of a documented public dataset?" | Latency, storage overhead |
| Machine learning (A) | "How does training-set size affect classification accuracy and F1 score for a convolutional neural network tested on a documented image dataset?" | Accuracy, F1 score |
| Data compression (A) | "How do Brotli and gzip compare in compression ratio and processing time across a fixed corpus of structured and unstructured text files?" | Compression ratio, processing time |
Choosing a documented, publicly available dataset matters: it makes your experiment replicable, which is directly relevant to how the Framework criterion is assessed.
Checklist: is your research question ready?
Before finalising your question, run it through this checklist:
- Testable? If your essay makes a prediction, be clear what result would go against it; if it does not, be clear what evidence would change your answer.
- Variables defined? State your independent variable (what you change) and dependent variable (what you measure) explicitly.
- Data collectable and replicable? Someone else following your method should get comparable results.
- Scoped to 4,000 words? A question comparing three algorithms across five metrics is almost certainly too wide; narrow it to two approaches, one condition you change and the metrics that matter, and hold everything else constant.
- Requires computing knowledge to answer? If a non-specialist could answer it without understanding algorithms, data structures, or systems, the question is probably too general.
For any subject-specific method constraints, including whether a particular experimental approach is appropriate for a CS EE, check the full EE guide on the IB Programme Resource Centre with your supervisor. The subject brief names the criteria; the full guide sets out how to meet them in each subject.
5. How the EE Differs from the Computer Science IA
The computer science IA and the EE are both compulsory IB components, but they assess entirely different things. Conflating them is one of the more consequential mistakes a student can make at the planning stage.
Per the IB Computer Science subject brief, the IA asks you to develop a computational solution to a self-chosen real-world problem. It carries 35 allocated hours and is weighted 30% of the final grade at SL and 20% at HL. The output is a product: working software, documented design decisions, and a record of development.
The EE is the opposite in form. It is a research essay that answers a focused question. The output is an argument supported by evidence, not a piece of software.
That distinction has a practical consequence for your research question. A question that is resolved simply by building something and confirming it works does not generate the analysis, comparison, or critical evaluation the EE criteria require. "I built a sorting algorithm and it ran correctly" is a build log, not a finding. A question such as "How does merge sort compare with quicksort on nearly-sorted integer arrays of varying size?" (written for this guide) produces measurable evidence you can discuss and evaluate, which is what the EE demands.
A build log, a project write-up, or a design document submitted as an EE will be assessed as an essay and will score poorly against the criteria. The two components are assessed separately, by different processes, and neither can substitute for the other.
6. Common Pitfalls to Avoid in a Computer Science Extended Essay
A computer science extended essay fails in predictable ways. Knowing them in advance is more useful than discovering them at the viva voce.
A research question that cannot be tested. "Is machine learning useful?" has no falsifiable answer. A workable version names a specific model, dataset, and metric, for example: comparing classification accuracy on a fixed dataset using two different neural network architectures. If you cannot state what result would count as a negative finding, the question is too broad.
Describing rather than analysing. Narrating how a sorting algorithm works step by step is not analysis. The essay needs to evaluate what the data shows: why one approach outperformed another under specific conditions, and what that implies.
Skipping limitations. The Discussion and evaluation criterion carries 8 marks, the highest of the five, per the IB's updated assessment model. Examiners reward honest identification of constraints, whether that is the hardware environment used for benchmarking, dataset size, or the scope of testing. Omitting limitations reads as shallow, not confident.
Misreading the Reflection criterion. The IB EE subject brief is explicit: the 500-word reflective statement on the Reflection and Progress Form asks how the EE experience affected you as a learner. It is not a summary of your findings, and repeating your conclusion there earns little of the 4 marks Criterion E carries.
Ignoring subject-specific guidance. The subject brief covers structure and criteria. The full EE guide on the IB Programme Resource Centre contains additional subject-level rules. Ask your supervisor for access before you finalise your methodology.
7. What to Do Next
Your most useful move this week is a specific one: ask your supervisor to share the sciences section of the full EE guide from the IB Programme Resource Centre, which covers computer science alongside the other sciences. That section sets out what methods and sources are expected in a computer science essay, and reading it before you draft your research question will save you from building an approach that does not fit the criteria.
Once you have the guide, draft three candidate research questions using the checklist in section 4. Bring all three to your first formal reflection session rather than committing to one in advance. Supervisors can spot scope problems, testability gaps, and sourcing difficulties far more easily when they can compare options side by side.
Ask your supervisor for the IB Programme Resource Centre guide this week, then arrive at your first reflection session with three written candidates, not one.
FAQ
What is an extended essay?
The Extended Essay is a compulsory, externally assessed component of the IB Diploma Programme core in which students produce an independent research essay of up to 4,000 words; it is assessed against five criteria totalling 30 marks from the May 2027 session onward.
How long is the extended essay?
The written submission has an upper word limit of 4,000 words covering the introduction, body, conclusion, and any quotations; the title page, contents page, and references and bibliography sit outside this count.
Is computer science a good subject for the extended essay?
Computer science suits the EE format well because many questions can be answered with measurable, reproducible experiments - comparing algorithm performance or data structure efficiency, for example - which maps clearly onto the analysis, argument, and evaluation criteria.
What are the IB extended essay assessment criteria for 2027?
The five criteria and their marks are: Framework for the essay (6), Knowledge and understanding (6), Analysis and line of argument (6), Discussion and evaluation (8), and Reflection (4), totalling 30 marks, labelled Criterion A to E.
How does the computer science extended essay differ from the IA?
The IA is a computational solution to a real-world problem, allocated 35 hours and weighted 30% at SL or 20% at HL; the EE is a research essay answering a question, so a build log or product write-up is not an appropriate submission.
Are there grade boundaries for the new 30-mark extended essay?
No grade boundaries for the 30-mark scale will be available until after the May 2027 session; results for that session are due on 6 July 2027.
References
- DP extended essay updates - International Baccalaureate - https://www.ibo.org/university-admission/latest-curriculum-updates/dp-extended-essay-updates/
- Diploma Programme Subject Brief: Extended essay, first assessment 2027 (International Baccalaureate Organization, 2024) - https://www.ibo.org/globalassets/new-structure/university-admission/pdfs/ee_subjectbrief_en.pdf
- Diploma Programme Subject Brief - Computer science, first assessment 2027 (International Baccalaureate Organization 2024) - https://www.ibo.org/globalassets/new-structure/university-admission/pdfs/dp_comp_sci_subjectbrief_en.pdf