1. Define the boundary of the review
- The question the literature has to speak to
- Date range - and a justification if you exclude older work
- Source types allowed: peer-reviewed only, or also reports and theses
- Databases searched and the search terms used, if your rubric wants the method documented
2. Build a synthesis matrix
Make a table with sources as rows and themes as columns: research question, sample, method, key finding, limitation. Once it is filled, the columns tell you the themes and the disagreements. That table becomes your structure.
3. Organise by theme, debate or method
Group sources that speak to the same point together, then write about the point rather than the papers. Sources become citations inside your sentence instead of the subject of it.
Source-led (avoid): Smith (2019) found X. Jones (2020) found Y. Lee (2021) found Z.
Theme-led (use): Studies using self-report measures consistently find X (Smith, 2019; Lee, 2021), while the two observational studies report the opposite effect (Jones, 2020), suggesting the measure rather than the population drives the result.4. Name the gap explicitly
The review should end where your study begins. State plainly what has not been examined - a population, a method, a timeframe, a contradiction nobody has resolved - and let that sentence lead into your research question.
5. Check for review-specific problems
- Every paragraph makes a point, not a summary
- Contradictions between studies are discussed, not averaged away
- Methodological quality is weighed, not just findings reported
- Citation density is even - no theme resting on a single source
Key takeaways
- Set explicit boundaries for scope, dates and source types
- Use a synthesis matrix to find themes before writing
- Write about themes and debates, not one source per paragraph
- End by naming the gap your own work fills