Guide written by the AI for Research editorial team · Last reviewed
A literature review matrix, also called a synthesis matrix, is a table with one row per source and one column per question you ask of every source: its aim, design, sample, findings, limitations and themes. Filling it in makes you read each paper the same way. Reading down the columns then shows where studies agree, where they differ and what nobody has studied, which is the raw material of synthesis [1][2]. The builder above gives you an editable matrix in the browser, fills the citation from a DOI or title, and exports it as CSV or a table.
What is a literature review matrix?
It is a grid that puts the same questions to every source, so you can compare sources instead of summarizing them one at a time. Garrard's Matrix Method made it a standard step in health sciences reviews [1], and Webster and Watson describe a concept matrix (concepts across, articles down) for information systems reviews [2]. The idea works in any field.
A summary describes one paper. Synthesis says what several papers show together: "three of the four surveys found X, but the only trial found Y". The matrix is the step in between.
Which columns should a literature review matrix have?
Start with the default columns, then add the ones your question needs. The defaults cover what most reviews compare.
| Column | What to write | Example (illustrative) |
|---|---|---|
| Citation | Author, year, title, journal, DOI | Study A (2021). Peer mentoring for new teachers. |
| Aim | The study's question, in one line | Does peer mentoring reduce first-year teacher burnout? |
| Design | Method and design | Cluster randomized trial |
| Sample/setting | Who, how many, where | 42 schools, 310 teachers, urban districts |
| Key findings | The main result, with numbers | Burnout scores lower at 12 months; no change in retention |
| Limitations | What weakens the result | Self-reported burnout; high attrition |
| Theme(s) | Your labels, separated by ; | Mentoring; Retention |
Add columns such as Theory used, Country, Measure or Funding when they matter to your question. Rename or remove any column from its header; the builder saves your changes in this browser.
How do you fill in a review matrix?
Read each source with the columns in front of you, write short notes in your own words, and add a page number next to anything you may quote.
- Add the source. Paste a DOI or title and choose the match. The builder fills the citation (authors, year, title, journal and DOI), and fills Authors, Year or Title columns too if you add them.
- Fill one row at a time. Keep each cell short: a phrase or a number, not a paragraph.
- Tag themes as you go. Write one or more labels in Theme(s), separated by semicolons. Reuse the same words so rows group together.
- Leave gaps visible. If a paper does not report its sample size, write "not reported". Missing information is a finding.
Worked examples across fields (illustrative)
These rows are invented to show the format; "Study A" and so on are not real papers.
Education: teacher burnout
| Citation | Design | Key findings | Theme(s) |
|---|---|---|---|
| Study A (2021) | Cluster randomized trial | Lower burnout at 12 months | Mentoring; Wellbeing |
| Study B (2019) | Survey, 1,200 teachers | Workload predicted burnout more than pay | Workload |
| Study C (2022) | Interviews, 24 teachers | Mentors helped most with behavior management | Mentoring |
Reading down Theme(s): two studies on mentoring, one on workload. Grouping by theme puts Study A and Study C side by side, a trial and an interview study that point the same way.
Engineering: sensors for bridge monitoring
| Citation | Sample/setting | Key findings | Limitations |
|---|---|---|---|
| Study D (2020) | 3 steel bridges, 18 months | Fiber-optic sensors detected strain changes early | One climate zone |
| Study E (2023) | Lab tests on 40 beams | Wireless sensors drifted after 6 months | Lab only |
Reading down Limitations shows a gap: no field study compares the two sensor types in the same bridges.
Business: remote work and productivity
A matrix with an added Measure column showed that studies measuring output found small gains, while studies using self-reports found larger ones. That contrast is a synthesis point no single paper makes.
How do you go from a matrix to written synthesis?
Write one paragraph per theme, not one per paper, and use the columns to say how the studies compare.
- Group the matrix by theme (tick Group by theme).
- For each theme, read down Design, Sample and Key findings. Where do the studies agree? Where do they differ, and does design or setting explain it?
- Write a topic sentence that states the pattern, then cite the studies that support it, then the exceptions.
- Use the Limitations column to say how strong the evidence is.
From the education rows above: Mentoring was linked to lower burnout in both a trial (Study A, 2021) and an interview study (Study C, 2022), though neither followed teachers past the first year.
For structured approaches to synthesis, such as narrative synthesis, see Popay et al. [4] and Snyder's overview of review methods [3]. For the full writing process, read how to write a literature review.
What is a synthesis matrix for an essay?
For a synthesis essay, rows are sources and columns are the ideas your essay discusses; each cell says what that source says about that idea. Choose the Synthesis essay (sources × ideas) template. It has columns for the main claim, three ideas you rename to your own, a quote and page, how you will use the source, and themes. Reading down an idea column tells you which sources agree, which disagree and where you need another source. The synthesis essay guide shows how to turn the grid into paragraphs.
Common mistakes
- Writing summaries in the cells. Long cells are hard to compare. Write phrases, numbers and labels.
- Different words for the same theme. "Mentoring", "mentorship" and "peer support" will not group together. Settle on one label.
- Only one kind of column. Findings alone hide why studies differ. Keep Design, Sample and Limitations.
- Copying the abstract. The abstract is the authors' summary. Fill the cells from the methods and results.
- No page numbers. Note the page next to anything you may quote, or you will hunt for it later.
- Keeping the only copy in a browser. The matrix is saved in this browser only. Download the CSV regularly.
How to build your matrix with AI for Research
The builder is free, runs in your browser and needs no account. DOI and title lookup uses the same index as the citation generator. Export the matrix as CSV for a spreadsheet, or copy it as a Markdown or HTML table to paste into a document.
When you are signed in, the next step is the evidence matrix: add PDFs from your research library, choose your columns, and AI for Research fills each cell from the paper, with the quote and page behind every answer so you can check it. Citations, search and the free tools cost nothing; signed-in AI features use credits, and the Free plan has 45 credits a month. To draft a review from real papers, try the literature review generator.
Next steps
- How to write a literature review
- How to organize research papers
- Boolean search string builder: find the papers to put in the matrix
- Evidence matrix in the app
Sources
- Garrard, J. Health Sciences Literature Review Made Easy: The Matrix Method. Consult the edition available through your library.
- Webster, J., & Watson, R. T. (2002). Analyzing the past to prepare for the future: writing a literature review. MIS Quarterly, 26(2), xiii–xxiii. https://www.jstor.org/stable/4132319
- Snyder, H. (2019). Literature review as a research methodology: an overview and guidelines. Journal of Business Research, 104, 333–339. https://doi.org/10.1016/j.jbusres.2019.07.039
- Popay, J., et al. Guidance on the Conduct of Narrative Synthesis in Systematic Reviews. Consult the report for publication details.