Open Chapter 2 of many theses and you will find the same paragraph pattern repeated: “Smith (2018) studied… and showed… Garcia (2019) surveyed… and found… Chen et al. (2021) analysed… and concluded…”. Each paragraph summarises one paper. After twenty pages the reader holds twenty summaries but still cannot tell where the field agrees, where it argues, and where the author of the thesis stands.
A literature review is not an annotated bibliography. It is an argument: given what is known, my question is the one worth asking next. Every source you cite is a brick in that argument, not an entry in a list.
Summary versus synthesis
| Summary | Synthesis | |
|---|---|---|
| Organising unit | Each author, each paper | Each idea, each question |
| Sentence subject | Author names | Concepts, findings |
| Relation between sources | Placed side by side | Compared, contrasted, differences explained |
| Author’s voice | Absent | Evaluating, commenting, guiding |
| End of each section | Move to the next paper | Point to what remains unknown |
The synthesis matrix: your most important working tool
Before drafting, build a spreadsheet: one row per source, one column per dimension you care about. Typical columns are setting (country, level), participants and sample size, method, key constructs, main findings, limitations and your own notes. For a project on teacher burnout:
| Source | Setting | Method | Findings on workload | Findings on support |
|---|---|---|---|---|
| Paper 1 | Urban primary | Survey, n = 800 | Strongly associated with burnout | Principal support buffers the link |
| Paper 2 | Rural secondary | Interviews, n = 20 | Not prominent | Professional isolation is the main issue |
| Paper 3 | Mixed levels, two-year panel | Repeated survey | Predicts next-year burnout | Not measured |
Read the matrix by column instead of by row and themes emerge on their own: workload is consistently linked to burnout in urban settings but not prominent in rural ones; support is measured in incompatible ways; almost no longitudinal study measures both. Each of those observations can become a paragraph.
Anatomy of a synthesis paragraph
- Topic sentence making a claim about the literature, with no author name as subject.
- Evidence from several sources, grouping those that agree into one citation.
- Contrast: which sources diverge and why they might (setting, method, measurement).
- Evaluation: how strong the evidence is.
- Bridge to what is missing or to the next paragraph.
A rewritten paragraph
Before: “Smith (2018) studied 800 primary teachers and found that workload was related to burnout. Garcia (2020) interviewed 20 rural secondary teachers and found that they felt isolated. Park (2021) followed teachers for two years and found that workload predicted burnout.”
After: “Workload is the most consistently documented predictor of burnout, appearing in large cross-sectional surveys (Smith, 2018) and two-year panel data alike (Park, 2021). This evidence, however, comes mainly from urban schools. Rural teachers rarely foreground workload, describing professional isolation as their main burden instead (Garcia, 2020) — suggesting that the pathways to burnout may differ by context, something existing quantitative designs have not tested.”
The second version uses the same three sources but does four things: it makes a claim, weighs the evidence, flags a contextual limit and sets up a gap.
Reporting verbs carry stance
Your choice of verb signals how you judge a source. “Demonstrate” and “show” imply you accept the conclusion; “suggest” and “propose” are more tentative; “argue” is neutral; “claim” often implies you are not convinced. Using a single verb for every source throws away one of your argumentative tools.
Four common types of gap
- Contextual gap: the phenomenon has been studied elsewhere but not in your population or region — and there is reason to think results would differ.
- Methodological gap: mostly cross-sectional studies with few longitudinal or experimental ones; mostly self-report with little objective measurement.
- Conflicting findings: studies disagree and nobody has explained why.
- Conceptual gap: two bodies of literature study related constructs without being connected.
“No study has examined X in country Y” is a weak gap unless you explain why country Y’s context might change the answer.
How many sources is enough?
There is no standard number, but there is a signal: when you read a new paper and already know nearly every source it cites, you are approaching saturation. Keeping a search log (database, search string, date, number of hits) lets you answer when an examiner asks how you searched. Use Zotero or Mendeley to manage sources and tag them with your matrix columns.
A revision checklist for your review chapter
- Highlight the subject of every paragraph’s opening sentence: how many begin with an author’s name?
- Every paragraph compares at least two sources.
- Every section ends with what remains unknown.
- At least one passage evaluates the strength of the evidence.
- The gap at the end of the chapter matches your research question.
Next step: choose your ten most important sources, build a five-column matrix in one evening, then read it by column and turn each observation into a topic sentence. Those sentences are the skeleton of your new review chapter.
Câu hỏi thường gặp
What is the difference between a literature review and a summary?
A summary retells each source, while a literature review organises sources by idea, compares and contrasts them, and builds towards a research gap. A review is an argument, not a list.
What is a synthesis matrix?
It is a table with one row per source and one column per dimension, such as setting, method and findings. Reading it by column reveals agreements, contradictions and gaps.
How do I identify a research gap?
Look for missing contexts, weak designs, unexplained contradictory findings or unconnected bodies of literature. Explain why filling that gap would matter.
How many sources should a literature review have?
There is no fixed number. When new papers mostly cite sources you already know, you are close to adequate coverage. A search log helps demonstrate how thorough you were.
Should I start sentences with author names in a literature review?
Occasionally, but if most paragraph openings begin with an author’s name, the review is probably summarising. Make concepts or findings the subject instead.
