EPQ Literature Review: How to Write One That Earns AO2 Marks

By Michael Thompson · Education Specialist; 10 years teaching the IB at Bromsgrove School · Published 24 October 2026 · Updated 25 October 2026

An EPQ literature review is where you demonstrate AO2 - the ability to research, critically select and use a range of sources to inform your own thinking, not just collect information. Most students who lose marks here do so for the same reasons: lists of URLs instead of formatted citations, sources that never appear in the argument, and evaluation that describes rather than judges. This guide sets out what the section needs to contain, how to structure it by theme rather than source, and what the OCR and AQA moderators' reports say separates strong work from weak. Follow the steps and the review becomes the part of your project that shows you can think.

Key Takeaways

In This Article

  1. What the EPQ literature review is actually for
  2. What strong and weak EPQ research review looks like - in moderators' words
  3. How to structure your EPQ literature review
  4. Choosing and evaluating EPQ sources
  5. Primary research in an EPQ: when to include it and when not to
  6. Common mistakes that cost AO2 marks
  7. How long should an EPQ literature review be?
  8. What to do next

1. What the EPQ literature review is actually for

The EPQ literature review is not a summary of what you have read. It is the part of your written report where you demonstrate AO2, the assessment objective the boards call "Use Resources." Most students write it like a reading list with commentary. That misreading costs marks.

AO2 in plain language. AQA's specification sets out the AO2 requirement as: "Research, critically select, organise and use information, and select and use a range of resources. Analyse data, apply relevantly and demonstrate understanding of any links, connections and complexities of the topic." Every word in that sentence matters. Critically select, not collect. Organise and use, not list and summarise.

OCR moderators draw the same line differently. Their Summer 2024 report lists 'Research is information collection and assembly' among the misconceptions they meet, and answers it: candidates "should collect ideas, arguments, and evidence, but they should also engage with these", and research material "should be used to inform and support candidates' own thinking". The OCR Summer 2024 moderators' report is direct: quantity of sources gathered is not the measure. Depth of engagement with them is.

The review sits inside your written report. For AQA, a research-based report runs to approximately 5,000 words; OCR's dissertation equivalent is a similar length. The review is one section of that document, not a standalone piece. How much space it deserves depends on how central secondary sources are to your argument, which later sections of this guide address.

2. What strong and weak EPQ research review looks like - in moderators' words

The clearest picture of what separates a good EPQ research review from a poor one comes from the boards themselves. The OCR Summer 2024 moderators' report is unusually direct about this, and it is worth reading its exact language.

Strong candidates, according to OCR, 'narrowed and deepened research as they progressed'. They 'produced well-formatted in-text citations and bibliographies', 'refined a research question which was targeted and focused', and 'made and justified changes to their plans and topics'. Their evaluations showed 'genuine self-reflection and insight'.

Weak candidates did the opposite on every count. They 'researched broadly and shallowly OR narrowly and shallowly'. They 'omitted in-text citations and bibliographies OR used hyperlinks or URLs instead of formatting references'. They 'selected a generic topic', 'stuck to topics and plans even when they weren't on track', and produced evaluations that were 'descriptive or lacked awareness of strengths and weaknesses'.

One counter-intuitive point from that same report: a bibliography of hyperlinks is not just aesthetically wrong. OCR notes that hyperlinks 'cannot be followed once the page is printed', which means a moderator reviewing a printed submission cannot verify a single source. A list of URLs actively undermines your credibility.

On the AQA side, the June 2023 report flags a specific over-marking pattern: high AO2 marks were awarded at some centres 'with little evidence of critical' selection. The implication is direct. Critical selection must be visible in the text of your review. If it is not there on the page, no one can credit it, regardless of the reading you actually did.

The consistent thread across both boards: research is expected to shape your thinking, not just fill a reference list.

3. How to structure your EPQ literature review

The most common structural mistake in an EPQ literature review is writing it as an annotated list: source one says X, source two says Y, source three says Z. That pattern tells the reader what exists, but it does not show that you can think across sources. Group your sources by theme instead, not one by one or in order of publication.

The five-component shape

As a structure written for this guide, a well-built EPQ literature review moves through five components in order:

  1. Introduction - one or two paragraphs that state the topic, the research question, and what the review sets out to cover.
  2. Main Themes and Evidence - the body, organised by argument or theme rather than by source.
  3. Critical Evaluation - assessment of methodologies, sample sizes, bias, and credibility across your sources.
  4. Research Gaps - a clear statement of what current research does not resolve, contradicts, or leaves open.
  5. Project Direction - how the review feeds into your method and refines your question.

That fifth component is the one most students drop. Ending on gaps alone misses the point: the review should explain why you are approaching your question the way you are.

What to do inside each theme

For every theme in the main body, work through these five moves:

This is the structure that earns AO2 marks, because it demonstrates synthesis and critical selection rather than information assembly.

Worked example paragraph (written for this guide, not a board exemplar or real student work)

Theme: the effect of evening phone use on sleep onset

"Marlow et al. (2021) and Okafor and Lind (2015) both find a significant association between screen exposure in the hour before sleep and delayed sleep onset in adolescents, suggesting the relationship is robust across different age cohorts. However, Marlow et al. relied on self-reported screen time, which is known to underestimate actual usage, while Okafor and Lind drew on a larger nationally representative dataset, making their findings more generalisable. Neither study isolates blue-light exposure from social and cognitive stimulation as a causal mechanism, which means the question of whether screen type or content drives the effect remains open. This gap shapes my method: rather than asking whether phones affect sleep, my project asks which aspect of phone use (notification response, passive scrolling, or blue-light exposure) has the strongest association with sleep-onset latency in the 16-18 age group."*

The two studies cited in this paragraph are invented for illustration; they are not real publications.

Notice how the paragraph names agreement, surfaces a limitation of one source (self-reported data), identifies the gap, and then connects it directly to the project's method. That final sentence is what most EPQ literature reviews are missing.

Practical takeaway: draft your themes first as headings, then slot sources under each one. If a source only fits under one theme and you cannot connect it to any other, consider whether it belongs in the review at all.

4. Choosing and evaluating EPQ sources

The source mix

A credible EPQ literature review draws on more than one type of source. Relying on a single category (say, news articles alone) signals shallow research; piling on academic papers without checking their relevance signals the opposite problem. Aim for a mix across these categories:

The non-obvious trade-off: a recent government report will often contain more up-to-date statistics than a peer-reviewed article published three years earlier, because journal publication lags. Knowing which type of source wins on currency versus which wins on methodological rigour is itself evidence of EPQ source evaluation skill.

A six-point evaluation checklist (written for this guide)

The OCR moderators' report names the CRAAP tool as an acceptable evaluation framework. The checklist below maps to the same logic:

QuestionWhat to ask
PurposeWhy was this published? To inform, sell, campaign, or report?
AuthorWho wrote it, and what is their institutional position or expertise?
DateIs it current enough for your specific question?
EvidenceWhat data or method underpins the claim? Is that method appropriate?
BiasDoes the author or funder have a stake in the conclusion?
UsefulnessWhat does this source specifically add to your question, beyond background?

Work through all six points for every significant source, not just the ones you suspect are weak.

What to keep in your log

The OCR moderators' report lists "a source analysis document" and "a bibliography" as acceptable portfolio evidence of research. Keep a running source analysis document alongside your project record (the Project Progression Record if you are with OCR), updated each time you add a source. Do not leave this until the write-up stage.

The same report identifies good practice in candidates who separated their reference list from their bibliography: the reference list contains sources cited directly in the report; the bibliography covers sources read but not directly quoted. Maintaining that distinction from the start is far easier than reconstructing it at the end.

5. Primary research in an EPQ: when to include it and when not to

Primary research is not compulsory in an EPQ. The OCR Summer 2024 moderators' report is explicit: it should "always contribute to project outcomes." If it does not serve your research question, leave it out.

When primary research does fit the question, the forms that tend to work are:

Surveys are the weak link. The OCR report found that effective surveys "were outweighed by the ineffective surveys" in the 2024 sample, and states they "should be conducted only where they can meaningfully add to the outcome." A survey of thirty classmates rarely adds analytical weight to a research question that secondary sources can answer more rigorously.

The counter-intuitive point: if your primary research does not produce the results you hoped for, that is not automatically a problem. The OCR report confirms it is "still possible to showcase it by analysing why it was not effective." Unpicking why a survey generated unusable data, or why an interviewee's response shifted your thinking, can itself demonstrate AO2 analytical thinking.

Whatever form your primary research takes, document it carefully. OCR candidates should record it in the Project Progression Record (PPR) with dates and copies of any correspondence. AQA candidates should log it in the Production Log and Assessment Record with the same level of detail. An undocumented interview is difficult for a moderator to credit.

6. Common mistakes that cost AO2 marks

Six errors come up again and again in EPQ research reviews, several of them named in the moderators' reports. They are distinct problems, but they share the same root cause: treating the review as a filing exercise rather than an argument.

A bibliography of URLs or hyperlinks. The OCR moderators' report for Summer 2024 is direct on this: hyperlinks cannot be followed once a page is printed. Formatted in-text citations and a proper bibliography are required. A list of links is not a bibliography.

Sources catalogued but never used. If a source appears in your bibliography but is never cited in your argument, it signals padding, not research. The review must show sources actively shaping your thinking, not sitting in a list for appearance.

One source per paragraph, no synthesis. Running through your sources one at a time, summarising each in turn, produces description, not analysis. Moderators want to see you connecting sources: where they agree, where they conflict, and what that means for your question.

No evaluation of credibility or reliability. Explaining what a source says is not the same as evaluating it. You need to address why it is trustworthy, or where its limitations lie, for every source you cite.

Research that stops when writing starts. The OCR report notes that strong candidates narrowed and deepened their research as the project progressed. Treating the research phase as finished once drafting begins is one of the clearest markers of a weaker submission.

Implicit rather than explicit critical selection. The AQA June 2023 report flags that high AO2 marks awarded with little evidence of critical selection represent over-marking. The practical consequence: your critical selection must be visible on the page. If you chose one source over another, or rejected a line of evidence, say so and say why.

7. How long should an EPQ literature review be?

No exam board sets a word count specifically for the epq literature review. What the boards do specify is the length of the written report as a whole: AQA's specification states that a research-based written report should be approximately 5,000 words, and OCR's specification gives the same figure for a dissertation outcome.

How you divide that total across sections is your decision, not the board's.

As practical advice written for this guide, the literature review is typically one of the longer sections of the report. A working target of roughly 25-35% of your total written word count is reasonable, which translates to around 1,250-1,750 words in a 5,000-word report. Some centres publish their own indicative range for the review; if yours does, follow it, remembering that it is centre guidance rather than a board requirement.

The counter-intuitive point worth knowing: a tightly argued 1,200-word review that evaluates five sources with genuine critical depth is likely to earn more AO2 credit than a 3,000-word review that summarises twelve sources without interrogating any of them. Length follows from the complexity of your question and the quality of your sources, not the other way round.

Depth and quality of evaluation matter more than hitting a word count. If you find yourself padding to reach a number, that is a signal to find a better source or push your analysis further, not to add more summary.

8. What to do next

Open your source analysis document this week and write two sentences on each of your three most-used sources: one sentence on why the source is useful to your specific research question, and one sentence identifying a credibility concern (author bias, age of the data, limited sample, and so on). Those two sentences cover the two things Cambridge OCR moderators say a source evaluation should include, usefulness and credibility, and they double as usable evidence for your log or portfolio.

Before you write a word, confirm where your centre wants source evaluations kept. Each board has its own record of the project (AQA's Production Log and Assessment Record, OCR's Project Progression Record (PPR), Pearson's activity log and records of research, WJEC's Candidate Production Record), and your supervisor will tell you whether evaluations belong there, in a separate source analysis document or in the report itself. Wherever they go, keep them where a moderator can find them.

Check with your supervisor which format applies to you, then start those three source evaluations before your next scheduled check-in.

FAQ

What is an EPQ literature review?

An EPQ literature review is the section of your written report where you evaluate existing research on your topic, assessing each source for usefulness, credibility and reliability, and showing how the sources inform your question and method - not a summary or a list of what you read.

Does an EPQ need a literature review?

No board mandates a section called 'literature review' by name, but AO2 - which requires you to research, critically select and use a range of sources - must be evidenced in your written report, and a structured review is the standard way to do that.

How long should an EPQ literature review be?

No exam board sets a word count for the review itself; AQA specifies approximately 5,000 words for the whole research-based written report, and OCR gives the same figure for a dissertation outcome, so the review's length is a matter of proportion and depth rather than a fixed rule.

How do I evaluate sources for an EPQ?

OCR moderators confirm that source evaluations must address both the usefulness of each source to your specific project and its credibility and reliability; the CRAAP tool is cited by OCR as an acceptable evaluation framework, and OCR identifies a source analysis document as acceptable portfolio evidence.

Should I include primary research in my EPQ literature review?

Primary research is not compulsory; OCR moderators say it should only be included where it meaningfully contributes to the outcome, with interviews, expert correspondence and site visits identified as more commonly effective than surveys.

Why do EPQ students lose AO2 marks on the literature review?

The failures the OCR and AQA reports name are URLs instead of formatted citations, evaluations that describe rather than judge, uncritical selection of sources, and research that stays shallow or leaves gaps unfilled once writing begins; listing sources that never appear in the argument is a further common weakness.

References