A pile of good notes can still produce a weak literature review if the sources never meet. A synthesis matrix fixes that by putting sources beside one another under shared themes. Clean each note so you can trace it, group the notes by the question they help answer, and then compare down each theme.
The immediate payoff is a clearer outline. The wider payoff is stronger research judgment: you can see where evidence agrees, where methods explain a difference, and where your argument would be stronger if you found another source.
Know what the matrix is meant to reveal
A summary tells you what one source says. Synthesis explains how several sources relate to one another.
A matrix makes those relationships visible by placing sources in rows and themes, topics, or methods in columns. Stanford's synthesis-matrix activity uses the structure to compare agreement, disagreement, divergence, and method.
A basic matrix looks like this:
| Source | Theme A | Theme B | Theme C | Method or context |
|---|---|---|---|---|
| Source 1 | Note + locator | — | Note + locator | Survey, first-year students |
| Source 2 | Note + locator | Note + locator | — | Interviews, mature students |
| Source 3 | — | Note + locator | Note + locator | Experiment, secondary school |
A dash means the source does not address that theme. Do not fill a cell just to make the table look complete. An empty cell can show that a source has a different focus or that the evidence for a theme is thin.
Make every useful note traceable
Do not begin by pasting every highlight into a spreadsheet. First make sure each note can lead you back to the original evidence.
Keep these details together:
- a short source ID, such as
Chen 2024 - the idea or finding in your own words
- the page, section, paragraph, table, or figure
- the note type: paraphrase, summary, or exact quotation
- why the note matters to your research question
- one or more possible themes
These details stop you searching through the paper again while drafting. They also separate the author's claim from your interpretation. The Monash guide to making research notes recommends keeping bibliographic details and page numbers beside notes, paraphrasing key ideas, and marking your own comments clearly. An NC State synthesis-matrix guide similarly recommends recording a page number or a searchable phrase.
Here is a constructed note using a fictional source:
Chen 2024 — paraphrase — p. 18: Students reported fewer interruptions when notifications were disabled, but the study measured self-reported focus rather than exam performance.
Possible themes: notification control; measurement limits
My comment: useful contrast with studies that measure test scores
If you save an exact quotation, add quotation marks immediately. Otherwise, copied language can blend into your paraphrase before you draft.
Clean the notes source by source
Create one row for each source. Use columns that make the papers easy to compare later:
- short source ID
- full reference or link
- research question or purpose
- population or material studied
- method
- main finding or argument
- important limitation
- page-level notes
This table is organised by source because its job is accuracy. Check each row:
- Confirm that you can identify the source quickly.
- Add a locator to every note you may use.
- Separate the source's findings from your reactions.
- Record the method and context when they affect the result.
- Remove notes that do not help answer your research question.
Do not use this table as the literature-review outline. Following it from top to bottom often creates one paragraph about Source A, then Source B, then Source C. The Monash literature-review process warns that this can become a “laundry list” and recommends reorganising the material around themes so relationships become visible.
Group the notes by the question they answer
Now scan across the sources. Look for ideas that repeat, conflict, qualify one another, or answer the same part of your research question.
For a project about smartphone use and student learning, possible themes might include:
- notification interruptions
- time on task
- self-regulation
- sleep
- academic performance
- study design or measurement
Choose columns that help answer the research question. Technology is probably too broad because almost every note could fit. A detail that appears once may not need its own column.
Test a theme by completing this sentence:
This theme matters because it helps explain ______.
If you cannot finish the sentence clearly, change or remove the theme.
Your first set of columns is not final. Merge themes that keep receiving the same notes. Split a crowded theme when it contains two questions. Add method or context when sources define or measure the same idea differently. Both Stanford's matrix activity and the Monash literature-review guide treat method and context as useful ways to explain divergence.
Keep each cell short enough to compare
A matrix cell should help you scan across sources. It does not need to summarise the whole paper.
Use this pattern:
Finding or position; important qualification; locator
Here is a constructed matrix using fictional sources:
| Source | Notification interruptions | Academic performance | Method or context |
|---|---|---|---|
| Chen 2024 | Disabling alerts reduced self-reported interruptions; p. 18 | Not measured | Survey; first-year university students |
| Ali 2023 | Frequent alerts linked with more task switching; pp. 41–42 | Lower quiz scores only in high-interruption condition; p. 47 | Controlled experiment |
| Rivera 2022 | Students described alerts as disruptive but kept them enabled; p. 9 | No direct performance measure | Interviews |
Keep the wording in your own voice unless the exact words matter. Mark quotations clearly and keep the locator beside them. A pasted paragraph hides the comparison the matrix is meant to show.
Read down a theme to find the relationship
Choose one column and ask:
- Where do the sources agree?
- Where do they disagree?
- Do they study the same population?
- Do they define the theme in the same way?
- Do they use comparable methods?
- Does one source limit or qualify another?
- Could time, place, discipline, sample, or measurement explain the difference?
- What remains uncertain?
Summary would report that Chen found one result, Ali another, and Rivera a third. Synthesis explains the relationship. In this constructed example, all three sources identify interruptions, but only one measures performance. The interview study may help explain why students keep a behaviour they describe as disruptive.
That difference is the point of the literature review. The Texas A&M guide to literature reviews explains that a review should compare and appraise sources, then synthesise them into an account of the scholarly conversation.
Before calling two sources contradictory, compare their methods and contexts. A survey of perceived focus does not answer the same question as an experiment measuring quiz scores.
Turn one strong column into a paragraph plan
Build the paragraph around the idea, not the authors:
- State the pattern. Tell the reader what the sources show together.
- Show the agreement. Bring together sources that support the pattern.
- Add the qualification. Introduce evidence that complicates or contradicts it.
- Explain the difference. Compare methods, populations, definitions, or contexts.
- State the implication or gap. Explain what the comparison means for your question.
Here is a constructed plan using the fictional matrix:
Pattern: Notifications consistently interrupt study activity, but evidence about academic outcomes is less complete.
Agreement: Chen and Ali both identify interruption or task switching.
Qualification: Only Ali measures quiz performance directly.
Explanation: The studies use different outcomes and research designs.
Gap: More evidence is needed to connect everyday notification behaviour with longer-term course performance.
The sources appear together because they help explain the same theme.
Fix the problems that make a matrix hard to use
- Every column summarises a source. Keep source facts in the inventory; add themes that answer the research question.
- A note has no locator. Return to the source now and add the page, section, paragraph, figure, or table.
- A quotation is not labelled. Add quotation marks and the locator before copied words blend into your paraphrase.
- You are forcing every source into every theme. Use a dash when the source does not discuss it.
- You created too many themes. Merge columns that repeatedly contain the same notes.
- You recorded disagreement but did not explain it. Compare definitions, populations, methods, dates, and evidence.
- The matrix never changes. Revise the columns as your understanding grows.
Your matrix is ready when every usable claim leads back to a source, the themes answer parts of the research question, and you can state the main pattern in each strong column. It is not the finished literature review. It is the bridge between reading and argument—the place where separate notes start becoming your explanation of what the evidence means together.