A full spreadsheet is not a literature review. A literature review matrix is a table with one row per source and the same columns for every row, and its job is to show where studies agree, disagree, and say nothing. Most students stop when the table is full and then write the review source by source. The patterns in the columns are what turn a review into an argument. This guide covers what a matrix is, how it differs from a synthesis matrix and an annotated bibliography, 6 steps to build one, a worked example with real studies, a template to copy, and how to turn the table into paragraphs.
What is a literature review matrix?
A literature review matrix is a table that records the same details about every source you read, one source per row, so studies can be compared side by side. It is also called a literature matrix, a lit review matrix, or a review matrix. Adler University Library’s Literature Review Guide describes it as a chart/spreadsheet nd says it is useful for spotting patterns across the research findings on a topic.
A matrix does three jobs. It stores your reading, so you do not reread the same paper. It asks every source the same questions. And it makes patterns visible. The University of Scranton’s Writing Center notes that a matrix puts aims, methods, and findings in one chart, which helps you spot where a gap might be. It adds that you do not have to use one, but it becomes more useful as your sources multiply.
It works the same way whether the review sits in a research proposal, a dissertation prospectus, or a full dissertation chapter.
A matrix stores what each source says. The review has to say what the sources add up to.
How is a literature review matrix different from a synthesis matrix or an annotated bibliography?
The three tools overlap, and universities use the names loosely. The practical difference is what each one is organized around.
| Tool | Organized around | Best for | What you end up with |
|---|---|---|---|
| Annotated bibliography | One entry per source | Learning a topic and showing what you read | A list of summaries or evaluations |
| Literature review matrix | One row per source, same columns for each | Comparing design, sample, and findings | A table you can sort and scan |
| Synthesis matrix | The points or themes you plan to discuss | Planning paragraphs | A map from each point to its sources |
An annotated bibliography, according to the Purdue OWL, gives a summary, an evaluation, or both for each source. It is built around single sources, so patterns across them stay hidden until you compare the entries yourself. A synthesis matrix flips the layout.
Many programs use “literature matrix” and “synthesis matrix” for the same thing, so check which format your advisor expects. If you are not sure, build one matrix and add a theme column in Step 4. It does both jobs.
How do you make a literature review matrix?
Most matrices fail in the setup, not in the filling. Six steps fix that, and the order matters because each step narrows the next.
Step 1: Write the question your review must answer
The question decides the columns. A review of whether practice testing beats restudying needs a column for the delay before the final test. A review of a theory needs a column for the framework each author used. Brandeis’s handout advises choosing column headings that support your research question and goals. A fixed list copied from someone else’s project will not. If the question is still shifting, settle that first, because a broad topic needs refining before the reading starts.
Step 2: Choose 6 to 9 columns
Keep the first few columns the same in every matrix: who wrote the source, when, and what it set out to do. The nine columns below cover most projects.
| Column | What goes in it | Why it matters |
|---|---|---|
| Source | Author, year, and a short ID | Lets you sort by date and cite quickly |
| Research question or purpose | What the authors set out to answer | Shows whether the study speaks to your question |
| Theory or framework | The lens the authors used | Reveals which schools of thought your sources come from |
| Design and method | Experiment, survey, interviews, review, and so on | Lets you compare how the evidence was produced |
| Sample and context | Who or what was studied, and where | Shows who the findings actually cover |
| Key findings | The results, in your own words | The evidence you will cite |
| Limitations | Weaknesses the authors or you noticed | Feeds your critique and your gap |
| Theme | A short label for the idea this source supports | Lets you sort rows into paragraphs |
| Relevance to my study | One line on why this source is in your review | Keeps filler sources out |
A column you never read down is clutter, so nine is a ceiling, not a target. Disciplines also differ. The Scranton Writing Center notes that matrix categories usually depend on the discipline, so a humanities review might swap “Sample and context” for “Primary texts or archives.”
Step 3: Read first, then fill each row
Fill each row from the full text, not the abstract, and write in your own words. Brandeis warns against trying to complete the matrix before you have read the articles. Mark anything the authors do not report as “NR,” so a blank cell always means “not read yet.” Note page numbers next to anything you might quote.
Short, regular sessions suit this work, and the 45-minute rule for ABD candidates is a good model. Keep your search record somewhere separate. The research methodology example shows how to log a search strategy and exclusion criteria. Reading and extraction also belong in your plan for how long a dissertation takes.
Step 4: Add a theme column and sort by it
Once several rows are filled, give each source a short theme label and sort by it. Write each label as a claim in waiting, such as “testing helps most at long delays,” not just “testing.” Sort by year as well. Brandeis recommends ordering sources by date to see changes in research over time.
Step 5: Read down the columns for patterns
A row tells you about one source. A column tells you about the field. Read each column from top to bottom and look for five patterns.
| Pattern | What to look for | What it becomes in your review |
|---|---|---|
| Agreement | Several rows reaching the same finding | A claim you can state with confidence |
| Conflict | Rows that disagree | A paragraph explaining why: design, sample, or measure |
| Method pattern | Most rows use the same design | A limit on what the field’s evidence can show |
| Sample pattern | The same population every time | Where your study could add something |
| Empty cells | Questions no row answers | A possible research gap |
Empty cells are where a research gap may sit, but test the gap before you build on it. A gap that looks original can break under a simple test, and the matrix is the place to run that test.
Step 6: Write one paragraph per theme
The UNC Writing Center separates summary from synthesis. A summary recaps a source. A synthesis reorganizes the information so the sources connect. The matrix gives you the connections, so each paragraph can follow one pattern: state the claim, give evidence from two or three rows, name the conflict or limit, and say what it leaves open. That is the same argument-first approach as a literature review that actually argues something.
The finished theme list is also the skeleton of the chapter. The Chapter & Outline Map is built for that handoff.
What does a literature review matrix example look like?
The example below uses four real sources on one question: does practice testing help learning more than restudying? Your field will differ, but the mechanics are the same. It is an illustration, not a complete review of the field, so check each paper before you cite it.
| Source | Design | Sample and materials | Key finding | Limitation or note | Theme |
|---|---|---|---|---|---|
| Roediger & Karpicke (2006) | Two experiments: one or three free-recall tests without feedback vs. restudying | Students studying prose passages; final test after 5 min, 2 days, or 1 week | Restudying gave better recall at 5 minutes. Testing gave substantially greater retention on the delayed tests. | Longest delay was one week | Effect depends on conditions |
| Karpicke & Blunt (2011) | Experiments comparing retrieval practice with concept mapping as an elaborative study method | Students studying science texts | Retrieval practice produced greater gains in meaningful learning, including on inference questions and concept-map tests. | Compares one elaborative method, not every study strategy | Beats other study methods |
| Adesope et al. (2017) | Meta-analysis of practice tests vs. non-testing conditions | 118 included studies | Practice tests were more beneficial than restudying and all other comparison conditions. Effect sizes varied with test features, participants, outcomes, and methods. | Averages hide variation; read the moderators | Effect depends on conditions |
| Dunlosky et al. (2013) | Review of 10 learning techniques | Published research on each technique | Practice testing rated high utility, along with distributed practice. It benefits learners of different ages and abilities across many tasks. | A review of existing evidence, not new data | Beats other study methods |
Reading the example
Agreement. All four rows favor practice testing over restudying, rereading, or concept mapping, at least under some conditions. That is a claim you can state with confidence, and each row backs it differently: two experiments, one meta-analysis, and one review.
A conflict that is really a condition. Row 1 shows restudying winning at five minutes and testing winning later, and row 3 says effect sizes vary with test features and study characteristics. So “testing always works” is too strong. The Theme column groups these two rows, which tells you the paragraph to write is about conditions.
A sample pattern. Both primary studies in this sample used students working with text. If your study looks at a different group or material, the Sample column shows that nothing here covers it. That empty space is worth explaining in your review, not ignoring.
From matrix to paragraph
Here is the same material written two ways. The first follows the rows. The second follows the theme.
Roediger and Karpicke (2006) studied testing. Karpicke and Blunt (2011) compared retrieval practice with concept mapping. Adesope et al. (2017) ran a meta-analysis of practice tests. Dunlosky et al. (2013) reviewed ten learning techniques.
Practice testing tends to beat restudying, but the advantage depends on conditions. In two experiments, restudying produced better recall at a five-minute test, while testing produced substantially greater retention after two days and one week (Roediger & Karpicke, 2006). A meta-analysis found practice tests more beneficial than restudying and other comparison conditions, with effect sizes varying by test features, participant and study characteristics, and outcome measures (Adesope et al., 2017). The conditions are therefore part of the finding. What remains to be shown is whether the pattern holds for [your population or material].
The strong version makes a claim, uses rows as evidence, qualifies the claim, and ends where your study begins. The weak version could be written without the matrix.
Where can you get a literature review matrix template?
Copy the template below into Excel or Google Sheets, or download it as a CSV. It has the nine columns from Step 2. Nothing is saved or sent anywhere.
| Source (author, year) | Research question or purpose | Theory or framework | Design and method | Sample and context | Key findings | Limitations | Theme | Relevance to my study |
|---|---|---|---|---|---|---|---|---|
Paste the header row into cell A1. Freeze the header row and the Source column, then use filters to sort by Theme and by year.
A spreadsheet wins because you can sort by theme and year. A Word table works if your program wants a document, but nine columns is wide, so use landscape orientation or trim the columns.
If you want a ready-made version built around themes and gaps, the free Literature Review Tracker sorts sources by theme and method. The Pro workbook adds a Source Matrix, Gap Finder, Thesis Builder, and Outline Handoff, which take you from sources to an argument in four steps.
What mistakes make a literature review matrix useless?
A matrix fails quietly. The table fills up, and the review still reads like a list. Six mistakes cause most of it.
| Mistake | What it looks like | Fix |
|---|---|---|
| Copying the abstract | Cells repeat the abstract’s wording | Read the full text and write each cell in your own words |
| Different fields per row | Some rows record the sample and others skip it | Fix the columns first, and use “NR” for anything not reported |
| Too many columns | Fifteen columns, half of them never read | Cut to the columns that answer your question |
| No theme column | The table cannot be sorted into paragraphs | Add theme labels once several rows are filled, and word them as claims |
| Stopping at the table | The review follows the row order | Write one paragraph per theme |
| Never updating | New sources are in the reference list but not in the matrix | Add a row the day a source enters your reading list |
If a committee comment like “needs more depth” lands on your literature review, how to read vague advisor feedback helps turn it into an edit list, and the theme column is the first place to look. A finished matrix also pays off at the defense. When a member asks how your work connects to a paper you did not mention, which is one of the dissertation defense questions, the matrix tells you quickly whether that paper is already a row.
FAQ
Is a literature review matrix the same as a synthesis matrix?
Often, but not always. A literature review matrix usually has one row per source, and a synthesis matrix is organized around the points you plan to discuss. Many programs use the names interchangeably, so ask your advisor which one is expected. A single matrix with a theme column covers both jobs.
Should I use Excel, Google Sheets, or Word for a literature review matrix?
A spreadsheet is easier because you can sort by theme or year. Use Word only if your program asks for a document, and switch to landscape orientation to fit the columns. Walden’s Writing Center offers sample matrices in both formats.
How many columns should a literature review matrix have?
There is no fixed number. Keep the first few columns the same in every matrix, then add the ones your question needs. Six to nine columns is workable for most projects, and any column you never read down can go.
How many sources should a literature review matrix include?
It depends on the scope of the review, your program, and whether the review sits in a proposal or a full dissertation. Ask your advisor what scope is expected. Add a source only when you can fill the “Relevance to my study” column in one line.
Can I use a literature review matrix for a systematic review?
As a starting point, yes, but formal systematic reviews follow stricter methods. The Cochrane Handbook stresses structured, easy-to-use data collection forms that capture accurate and complete data, and Cochrane’s standards call for piloting the form. Check your review protocol before you rely on a plain matrix.
Can AI fill in my literature review matrix?
AI tools can draft entries, but they can misread a paper or invent a detail, so check every cell against the full text. Also check your university’s policy on AI use before you rely on one for coursework or a dissertation.
Final thoughts
A full matrix is a stored reading list. It becomes a literature review when you read down the columns, name the themes, and write one paragraph for each. Choose columns that fit your question, fill each row from the full text, add a theme column, and look for agreement, conflict, and empty cells before you write a sentence. Ready to build yours? Copy the template in this guide and fill in the first row.
Literature Review Tracker
The matrix stores your sources. The tracker helps you sort them by theme and method, which is the step before you build your argument.
Get the tracker →