A literature review chapter fails in one of two ways: it becomes an annotated list of summaries with no argument, or it becomes an argument with no defensible search behind it. This guide walks a South African master’s or honours student through the chapter step by step — including the South African source layer most generic guides skip — with the expected output stated at every step.
How to write a literature review for a South African dissertation, step by step
Step 1: Fix the review’s job before you search
A dissertation literature review has one job: to show the gap your study fills and the conceptual tools you will fill it with. Write a single sentence of the form: “This review establishes what is known about X in context Y, and shows that Z remains unanswered.” Everything you read gets tested against that sentence. Expected output: the sentence, taped above your desk. If you cannot write it yet, your research question needs work first — go back to our research proposal guide before reading another paper.
Step 2: Design a search you can defend to an examiner
Examiners increasingly ask how you searched, not just what you found. Decide and record: which databases and platforms, which search terms and combinations, which years, which inclusion rules. Your university library’s databases are the backbone — and add the South African layer that international guides ignore: your institution’s own repository and other universities’ repositories for recent South African theses, South African journal platforms available through your library, and the policy and statistics sources where South African evidence actually lives — government departments, the CHE, Stats SA. For the data side of that layer, our guide to accessing South African data shows the access procedures. Expected output: a half-page search protocol — databases, terms, dates, limits — that goes into your methods or review chapter.
Step 3: Screen hard, and log what you exclude
Read titles and abstracts against your Step 1 sentence and sort into three piles: core (directly about your question), context (frames it), and out. Be ruthless with “interesting but out” — that pile is where review chapters go to die. Keep a simple screening log: how many found, how many kept, why the rest fell away. You are not required to run a formal systematic review for a standard dissertation chapter, but borrowing its discipline — stated criteria, logged decisions — is what “a defensible search” means. Expected output: a core list of manageable size and a one-paragraph account of how you got there.
Step 4: Build a synthesis matrix — the step that prevents the summary-list failure
Open a spreadsheet. One row per source; columns for: full reference, context/sample, method, key finding, limitation, and — the column that does the real work — “speaks to which theme”. Fill it as you read, not after. When twenty rows exist, the themes column hands you your chapter structure: the recurring conversations in the literature become your section headings, and every section can now cite three to six sources in genuine dialogue with each other. Expected output: the matrix. It is also your insurance policy — six months from now, when an examiner asks why you excluded a famous paper, the matrix remembers what you actually read.
Step 5: Write theme by theme, arguing as you go
Structure each section the same way: what the literature agrees on, where it disagrees or thins out, and what that means for your study. The test of synthesis is the sentence shape — “Author A found X in urban schools, while Author B found the opposite pattern in rural clinics, suggesting the effect is context-dependent” is synthesis; “Author A found X. Author B found Y.” is a list. Aim for most paragraphs to cite more than one source. South African context earns its place here too: where local findings diverge from the international literature, that divergence is often exactly the gap your study addresses. Expected output: drafted theme sections, each ending with one sentence linking the theme to your study.
Step 6: Close with the gap, explicitly
The final section of the chapter states the gap in plain terms and hands over to your methodology: what is known, what is not, and why your design is the right way to find out. If your review cannot produce this paragraph, the chapter is not finished — it is the whole point of the exercise. Expected output: a gap statement your methodology chapter’s first page can quote.
Step 7: Get the referencing right the first time
Every source in the matrix needs a complete, correctly formatted reference before you write about it — retrofitting fifty references the week before submission is how reference lists end up mismatched with in-text citations. A worked Harvard example of the commonest type, the journal article:
Mokoena, T. & Naidoo, R. 2024. Postgraduate throughput at South African universities: a cohort analysis. South African Journal of Higher Education, 38(2):45–67.
Check the exact punctuation against your faculty’s guide — UCT, Wits and Stellenbosch each publish their own Harvard variant, and the differences are real; our Harvard vs APA comparison covers how to identify which rules govern you. A reference manager captures sources as you read; Tesify keeps the review’s structure, drafts and citations in one project and generates the reference list from what you actually cited. Expected output: a reference list that reconciles one-to-one with your in-text citations.
Take notes that become paragraphs, not quotes that become plagiarism
The mechanical habit that separates a synthesised review from a stitched-together one happens at the reading stage. For every core source, write three sentences in your own words before you close the PDF: what the study did, what it found, and what it means for your question. Those sentences — not highlighted quotations — are the raw material of your chapter, and because they were yours from the start, the paraphrasing problem that fills similarity reports never arises. Record page numbers for anything you might quote directly, and mark verbatim phrases unmistakably in your notes; the most common route into accidental plagiarism is a note file where the writer can no longer tell their own words from the source’s. Six weeks later, drafting a theme section means assembling and sharpening sentences you already wrote — which is also the only version of literature-review writing that fits into the one-hour evening sessions a working student actually has.
How long should the literature review be?
There is no national rule — length is set by your faculty’s dissertation guidelines and the conventions of your discipline. The useful internal test is proportion: a review that dwarfs your findings chapter usually signals summary-listing, and one thinner than your introduction usually signals a shallow search. Ask your supervisor for two or three recent passed dissertations in your department and measure against those, not against a generic number from the internet.
The three failure patterns examiners flag most
The historical tour. Starting at the oldest paper and marching forward in time reads like a chronicle, not an argument. Organise by theme, and use chronology only inside a theme when the development itself matters.
The missing South African layer. A review built entirely on North American and European studies, examining a South African question, hands the examiner an easy criticism. Local theses, local journals and local policy documents exist; a defensible search finds them, and where genuinely little local work exists, saying so — with the search evidence to back it — strengthens your gap statement rather than weakening it.
The suspiciously perfect consensus. If your review reports that everyone agrees, you have either stopped reading too early or filtered out the disagreement. Disagreement is where research questions live; hunt for it deliberately.
FAQ
How many sources does a master’s literature review need?
Enough to cover the core conversations around your question — quality of synthesis beats raw count, and your faculty’s guidelines and recent passed dissertations in your department are the realistic benchmark. A thin search shows; so does padding.
Can I cite other students’ dissertations?
Yes — completed, examined theses in university repositories are citable scholarly work, and for South African topics they are often the most current evidence available. Cite them as theses, with the institution named.
Do I need a formal systematic review?
Not for a standard dissertation chapter — a narrative review with a documented search protocol is the norm. If your degree requires a systematic review as the study itself, that is a different method with its own reporting rules; confirm with your supervisor which one you are writing.
How recent must my sources be?
Cover the current conversation — roughly the last five to ten years — plus the foundational works your field still argues with. An examiner reads the reference list dates as a map of how alive your search was.
Can I use AI to help with my literature review?
For structure, summarising your own notes and tightening prose — within your university’s disclosed-use rules, yes; see our guide to AI policies at South African universities. Never let it invent or verify sources: every reference must be one you opened yourself.
Should the literature review come before or after data collection?
The main review is written before — it justifies the study and passes through your proposal and ethics process. Expect to update it before submission, because the literature moved while you were in the field.
What is the difference between a literature review and a theoretical framework?
The review maps what is known; the framework selects the concepts and relationships your study will actually use. In many South African dissertations the framework closes the review chapter, built from the themes the review established.
My topic has almost no South African literature. Is that a problem?
It is a finding. Document the search that proved the scarcity, use regional and international literature for the conceptual work, and present the local gap as part of your study’s justification.