The Synthesis Matrix: Turning a Pile of Papers Into an Argument

You have read forty papers. You have careful notes on each one. And when you sit down to write the literature review, what comes out is a list: this paper found X, that paper found Y, a third paper found Z. It is accurate, it is thorough, and it is not a literature review. It is a stack of summaries wearing one.

The missing step is synthesis: saying what the papers mean together, where they agree, where they clash, and what the pattern across them adds up to. There is a simple tool that makes synthesis almost mechanical, and it is called a synthesis matrix.

Summarizing is not synthesizing

A summary answers the question "what does this paper say?" Synthesis answers a harder one: "what do these papers, together, say about my question?" That difference is the whole game. Jane Webster and Richard Watson made the same point in a 2002 editorial on writing literature reviews⁠[1]: a review that simply presents a summary of the relevant articles "fails to synthesize the literature." A reviewer or an advisor can tell within a paragraph whether you are marching through sources one at a time or actually reasoning across them, and only the second reads like scholarship.

The trouble is that synthesis is hard to do in your head. Hold forty papers and a dozen themes in mind at once and the connections blur. That is not a failing on your part: working memory, the system that holds information while you reason with it, has a central store of only about three to five meaningful items⁠[2] in young adults. You need a way to lay everything out so the patterns become visible instead of imagined. That is exactly what the matrix does.

What a synthesis matrix is

A synthesis matrix is a grid. As Williams College Libraries puts it⁠[3], it organizes your sources by theme: each row is a source, each column is a theme or question that runs across your reading, and each cell holds what that source says about that theme. That is the entire structure. A spreadsheet works perfectly, and so does a plain table in a document.

The power is in the layout. Once your reading sits in a grid, the two directions come to mean different things. A row shows you everything one paper touches. A column shows you what every paper says about a single theme, lined up side by side, and that column is the raw material of a synthesized paragraph.

How to build and read one

Building it is mechanical, which is the point. Start with your themes as the columns. These come from your own reading: the recurring questions, methods, or debates you kept meeting across papers. Then go through your notes, source by source, and drop each paper's relevant point into the right cell, in your own words and kept short. You are not rewriting the paper. You are recording what it contributes to each theme, and leaving a cell empty where the paper is silent.

Then read down the columns, because that is where synthesis happens. Reading a single column, you see at a glance that four papers agree on a point, two disagree, and one never addresses it. Agreement becomes a claim you can state with confidence. Disagreement becomes a tension worth discussing. An empty stretch in a column is a gap⁠[4], which is often the most interesting thing on the page, because a gap is where your own contribution can go. This kind of comparison across papers is what Researcher+'s compare mode is built for, surfacing what several papers say on the same question with citations to the source pages. However you assemble it, the matrix is what makes the pattern visible.

From grid to argument

The matrix is scaffolding, not the finished review. To turn it into prose, write from the columns, not the rows. It is the same move Webster and Watson recommend: build a concept matrix as you read each article, then write the review by discussing each concept. Each theme, each column, becomes a passage that states what the literature collectively says about it, names who agrees and who dissents, and points to the gap if there is one. String those theme passages together in an order that builds, and you have the body of a literature review that argues rather than lists⁠[5].

None of this replaces judgment. Choosing the themes, weighing whose evidence is stronger, deciding what the gaps mean: that is the thinking only you can do, and the matrix exists to give that thinking something to stand on. Summarizing forty papers is data entry. Laying them in a grid and reading down the columns is how the pile finally becomes an argument.