How to write a literature review in psychology

A psychology literature review is not a survey of everything ever published on your topic. It is an argument: this is what the field knows, this is how much confidence that knowledge deserves, and this is the specific gap your study fills. Examiners read hundreds of these chapters, and they can usually tell within two pages whether you have synthesised the literature or just stacked summaries in a row.

Psychology adds complications of its own. The field runs from cognitive neuroscience to social theory, several of its best-known findings have been through a replication reckoning, and APA conventions set precise expectations for how you report and attribute evidence. Your review has to handle all of that while staying readable.

This guide works through the process end to end: where to search, which source types examiners expect, how to structure the chapter thematically, when PRISMA applies, and how to keep every claim in an AI-assisted draft tied to a paper you actually hold.

What a psychology literature review needs to do

Start by naming the kind of review you are writing, because examiners judge each kind by different standards. A narrative review, the default for most thesis chapters, is judged on the quality of its argument and the fairness of its coverage. A systematic review is judged on protocol: predefined criteria, documented searches, and reproducible screening. A meta-analysis adds quantitative pooling and is judged on statistical method as much as on coverage.

Whatever the form, the review has four jobs. Define your constructs precisely, because self-efficacy is not self-esteem and examiners notice when terms drift. Trace the theoretical development that produced your hypotheses. Appraise the methods behind the findings you rely on, not just the findings themselves. And locate the gap your study fills in terms specific enough that a reader could check whether you filled it.

Where to search: the best databases for psychology research

No single database covers psychology. APA PsycInfo is the disciplinary core, but clinical and neuroscience topics need PubMed, and citation chasing works best in the big cross-disciplinary indexes. A defensible search touches several of these:

CiteDash's Literature Finder runs your question across an internal corpus plus live OpenAlex, PubMed, Semantic Scholar and arXiv in a single search. That covers the peer-reviewed core and the citation trail quickly; you can then run PsycInfo through your university library to pick up anything indexed nowhere else.

• APA PsycInfo: the core psychology index, strongest for behavioural and social research, with a controlled vocabulary that makes searches precise
• PubMed and MEDLINE: essential for clinical psychology, psychiatry, and neuroscience questions
• Scopus or Web of Science: cross-disciplinary coverage and citation tracking in both directions
• ProQuest Dissertations and Theses: unpublished studies that matter when publication bias is a concern
• PsyArXiv: preprints and the newest work, read with the caution unreviewed findings deserve
• Google Scholar: broad but noisy, useful for catching what the curated databases miss

Build a documented search strategy

Whatever you find, an examiner will eventually ask how you found it. Work from a written search strategy: the concepts in your question, the synonyms and controlled vocabulary terms for each concept, and how you combined them. Combine synonyms with OR inside each concept block and join the blocks with AND. Log the databases, the exact strings, the date limits, and the date you ran each search.

This discipline pays off twice. It makes your search defensible at the viva, and it makes the search repeatable in the final months of writing up, when you need to check what has appeared since you started.

Screen in two passes: titles and abstracts first, then full text for what survives. Note why you excluded borderline papers, even in a narrative review. The habit costs minutes and buys you a defensible answer to the most common viva question about coverage: why this study is in and that one is out.

Source types psychology examiners expect to see

Psychology has a looser evidence hierarchy than medicine, but examiners still expect a deliberate mix rather than whatever the first search returned:

Grey literature needs care. Conference posters, unpublished measures, and blog summaries can point you toward sources, but they are weak foundations for claims, and examiners treat them accordingly.

• Meta-analyses and systematic reviews to anchor what is established about your effect
• Recent empirical articles that carry the argument, typically from the last five to ten years
• Registered reports and replication studies wherever they exist for the effects you rely on
• Theoretical papers that define your constructs and the models behind your hypotheses
• Classic studies cited from the primary source, not from a textbook summary of them
• Psychometric validation papers for every measure you plan to use

Structure a psychology literature review thematically

Chronological structure almost always reads as a list, and paper-by-paper structure always does. Organise by theme instead: by construct, by theoretical debate, or by methodological approach. The chapter as a whole should funnel from broad theory to your constructs, then to the specific relationship you study, and finally to the gap.

End every thematic section with evaluation rather than summary. Writing that a set of studies collectively supports a moderate association but nearly all rely on cross-sectional self-report is synthesis; restating each study's findings in turn is not.

Follow APA conventions as you write: past tense for findings, precise attribution of claims to authors, and measured language. Absolute words like proves invite an examiner to go looking for the counterexample.

At paragraph level, the move examiners want is claim first, evidence second, evaluation third. Open with the point the paragraph proves, support it with the studies that carry it, then say what the evidence does and does not allow you to conclude. Paragraphs that open with an author name usually signal summary rather than argument.

Take the replication crisis seriously

Examiners increasingly expect a psychology review to show awareness of replication issues. Some well-known effects in the field have failed to replicate, and building your rationale on a single small study is the kind of decision that gets probed at the viva. You do not need to relitigate the field's debates, but your weighting of evidence should reflect them.

Practical habits help here: check whether the key effects in your argument have been replicated or meta-analysed, prefer preregistered studies where they exist, discuss effect sizes rather than significance alone, and say so plainly when an evidence base is thin.

Measurement deserves the same scrutiny. Psychology's constructs are only as good as the scales that operationalise them, so note the reliability and validity evidence for the measures behind your key findings, and flag results built on ad hoc or unvalidated instruments. A review that tracks measurement quality reads as far more expert than one that treats every questionnaire as equivalent.

Check for retractions too. Psychology has had prominent ones, and a retracted study in your reference list undermines the whole chapter. Run your anchor citations through a free retraction check; inside CiteDash, retracted papers carry a badge in your library and are blocked from citation entirely.

When you need PRISMA: systematic reviews in psychology

If you describe your review as systematic, examiners will hold you to it. That means a protocol written before screening: the databases you will search, the exact strings, the inclusion and exclusion criteria, and how disagreements get resolved. It also means reporting a PRISMA flow diagram with the numbers at each stage: records identified, deduplicated, screened, excluded with reasons, and included.

Frame the question first. PICO works for intervention questions; SPIDER suits qualitative and mixed questions. Record screening decisions as you make them, because reconstructing them afterwards is slow and unconvincing. If this is the review you are running, our guide to a PRISMA systematic review with AI covers screening and the flow diagram in detail.

Plan the synthesis before you finish screening. Decide what you will extract from each included study: design, sample, measures, effect direction, and quality judgement, and record it in a structured way from the first paper onward. Extraction done consistently is what turns a pile of included studies into a review you can actually write.

A worked outline for a psychology literature review

Here is an outline for a thesis on academic self-efficacy and procrastination, adaptable to most psychology topics:

Each theme section is a small argument with its own arc: what the field claims, how good the evidence is, and what remains open. The gap section should follow from the critique, not appear beside it.

• Introduction: scope of the review, search approach, and a roadmap of the sections
• Theoretical framework: social cognitive theory and how self-efficacy is defined and measured
• Theme 1: self-efficacy and academic performance, what is established and on what evidence
• Theme 2: procrastination, its measurement, and the competing explanatory models
• Theme 3: the self-efficacy and procrastination relationship, including contradictory findings
• Methodological critique: designs, samples, and measures across the evidence base
• Synthesis and gap: what remains unknown and the precise question your study answers

How CiteDash grounds a psychology literature review

The risk with AI assistance is well documented: general chatbots fabricate plausible references and misattribute findings. CiteDash is built so that cannot happen. Every citation is a database object that resolves to a real paper; there are no free-text references anywhere in the system.

The workflow maps onto this chapter directly. Search with the Literature Finder, save papers to your library, upload PDFs or fetch details from a DOI, and use Ask and extract to pull methods, samples, and findings from a paper with sentence-level provenance, so every extracted claim binds to the exact passage it came from. Then build an evidence matrix in the Synthesis Lab: papers down one axis, with constructs, designs, samples, and findings across the other. If your review is systematic, the same matrix supports PRISMA-friendly workflows.

When you draft, the Fact Checker verifies each claim against the full-text PDFs you hold before the text reaches you. A claim it cannot verify is flagged rather than passed through, which is exactly the discipline a psychology review needs.

Common mistakes examiners flag in psychology reviews

The recurring problems are predictable, which makes them avoidable:

The common thread is distance from the sources. Keep the search documented, read what you cite, weigh evidence in proportion to its quality, and make every claim resolvable to a paper you hold. Do that and the review reads the way examiners want it to: as the argument that makes your study necessary.

• Organising by paper or by year instead of by theme or construct
• Citing from abstracts without reading the full text, which shows the moment an examiner probes
• Treating a single underpowered study as settled evidence
• Ignoring failed replications or contradictory findings for effects central to the rationale
• Citing classic studies second-hand from textbooks or other reviews
• Describing methods without evaluating them, so the critique never happens