Building a Literature Matrix
Core summary
A literature matrix (also called an evidence table or summary table) is a spreadsheet where each row is a study and each column is a piece of information you want to compare across studies. It transforms a pile of papers into an organized, at-a-glance comparison.
Detailed explanation
Detailed explanation
Why you need a literature matrix: Without a matrix, you read 20 papers and end up with scattered notes, highlighted PDFs, and a vague sense of what the literature says. With a matrix, you have a single table that reveals patterns, contradictions, and gaps at a glance. How to build one: Step 1 — Set up your spreadsheet. Use Excel, Google Sheets, or any spreadsheet tool. Step 2 — Define your columns. Standard columns include: - Author, Year - Study Design (RCT, cohort, case-control, etc.) - Population (who was studied, sample size) - Intervention/Exposure - Comparator - Primary Outcome - Key Finding (effect size, CI, p-value) - Quality Assessment (CASP/RoB rating) - Notes/Limitations Step 3 — Add one row per study as you read each paper. Step 4 — Analyze the matrix. Look for: - Consistency: Do most studies agree? If 8 of 10 show a benefit, the evidence is more convincing. - Contradictions: Do some studies disagree? Why? (Different populations? Different doses? Different follow-up periods?) - Gaps: What questions remain unanswered? What populations, interventions, or outcomes have not been studied? - Trends: Do newer studies show different results than older ones? For systematic reviews, the matrix becomes the data extraction table — a required component of the review. Tips: - Start simple. You can always add columns later. - Use color coding: green for positive findings, red for negative, yellow for mixed. - Include a 'Quality' column from the beginning. A pattern of 'all positive results from low-quality studies' is very different from 'all positive results from high-quality studies.' - Sort by study design or year to see trends.
Clinical example
A resident building a literature review on statins in chronic kidney disease creates a matrix with 15 studies. The matrix reveals that all 4 RCTs in advanced CKD (stages 4-5) showed no benefit, while all 5 cohort studies in early CKD (stages 1-3) showed benefit. This population-specific pattern would be hard to spot without the organized comparison.
Research example
Garrard's 'Health Sciences Literature Review Made Easy' recommends the matrix method as the foundation of all literature reviews, noting that researchers who use matrices produce reviews with 'greater depth, organization, and analytical clarity' than those who rely on narrative note-taking.
Knowledge check
Q1. What is the main purpose of a literature matrix?
Q2. Which of these should be included as a column in a literature matrix?
Q3. What pattern in a literature matrix would raise concern about the evidence?