Literature Review Matrix: How to Organize 100+ Sources
The difference between a chaotic literature review and a coherent one is almost always organization. After a few dozen PDFs, no human brain can hold study characteristics in working memory — you start to confuse authors, forget methodologies, and describe Smith's randomized trial as if it were Jones's qualitative study. The solution is the same tool doctoral students have used for decades: a literature review matrix.
What a matrix is
A literature review matrix is a table — usually a spreadsheet — with one row per source and one column per attribute you care about. Done well, it lets you synthesize across studies rather than describe them one at a time.
Garrard's Health Sciences Literature Review Made Easy popularized the matrix method in 1999, but the underlying logic is older: it is how Cochrane reviewers have always extracted studies for systematic review. The matrix is not just for systematic reviews; it works equally well for narrative, scoping, and integrative reviews.
What columns to include
Start with these eight columns for any review:
- Citation (Author, Year — e.g., Thomas & Harden, 2008)
- Title
- Research question or aim
- Methodology (design, paradigm)
- Sample (who, how many, how sampled)
- Setting and country
- Key findings (bulleted, in your words)
- Relevance to my review (one line)
Add columns specific to your review:
- For intervention reviews: intervention, comparator, outcomes, effect sizes
- For qualitative reviews: analytic approach, theoretical framework, themes
- For scoping reviews: concepts, populations, gaps identified
- For methodology reviews: tool used, psychometrics, validity evidence
Our free literature review matrix template has these pre-built. Download it, or drop rows into the free Subthesis Literature Matrix for an interactive version.
How to fill it in efficiently
- Extract once, properly. Re-reading a paper because you skipped a field is the single biggest time-waster. Use a checklist.
- Quote sparingly. Paraphrase findings in your own words. If you quote, include a page number immediately.
- Standardize language. "Randomized controlled trial," not "RCT" in one row and "random trial" in another.
- Date the entry. When your supervisor asks, "is this current?" a timestamp answers immediately.
- Flag concerns. A "notes" column for "small sample," "author retraction," or "can't access full text" prevents surprises at synthesis.
Use the matrix to actually synthesize
A matrix is useless if you only fill it in. Once populated, use it to:
- Sort and filter. Group by methodology to see what designs dominate. Group by year to see how the field evolved.
- Identify themes. When five rows all say "clinician burnout correlated with staffing ratios," that is a theme, not a coincidence.
- Spot gaps. An empty quadrant (e.g., no qualitative studies in a population) is a gap worth naming.
- Draft synthesis paragraphs. Each theme becomes a paragraph; each row contributes a sentence. Transitions write themselves.
See our process of synthesis page for the thematic and narrative synthesis techniques that work on matrix data.
Common mistakes
- Too many columns. Thirty columns no one fills in is worse than eight that are always complete. Cut ruthlessly.
- Copy-pasting abstracts. The abstract is the author's summary, not yours. Paraphrase.
- Not reading the full paper. Abstracts omit sample size, effect size, and half the limitations. Read the full paper before extracting.
- Using Word. Tables in Word cannot be sorted. Use Excel, Google Sheets, or Airtable.
- No version control. When the matrix corrupts — and it will — you will want yesterday's copy.
When to use Excel vs a dedicated tool
For fewer than 40 sources, a spreadsheet is fine. Above 40, friction starts to dominate: filtering breaks, formulas get weird, collaboration is painful. Move to a dedicated tool — Covidence for systematic reviews, the Subthesis Literature Matrix for narrative and scoping reviews, or Airtable for custom workflows.
Pairing with other review assets
A matrix is one part of a reviewer's toolkit. Pair it with:
- A search strategy documentation form — so you know what you searched
- A PRISMA flow diagram — so you know what you excluded and why
- A data extraction form for systematic reviews with stricter rigor requirements (see our data extraction forms post)
Matrix-first, synthesis-second is the rhythm of a sane literature review.