Synthesis Matrix Generator

A synthesis matrix is the working document that turns a pile of included studies into a structured narrative. In every major guidance for systematic and scoping reviews, the synthesis matrix sits between data extraction and the written synthesis section as the place where cross-study patterns become visible. Popay and colleagues (2006) describe the matrix as the central organising tool for narrative synthesis, while Thomas and Harden (2008) use a similar grid in thematic synthesis to map descriptive themes against the analytical themes that follow. The Joanna Briggs Institute (JBI) Manual for Evidence Synthesis recommends matrices for both qualitative and mixed-methods reviews, and the Cochrane Handbook recommends descriptive matrices for narrative summaries when meta-analysis is not appropriate.

The literature synthesis matrix differs from a data extraction form in purpose. The extraction form is filled in once per study and captures the structured fields (sample, intervention, comparator, outcome, effect size, risk of bias). The matrix is the second-pass document: it takes the extracted data and reorganises it so themes run across the top and studies run down the side, putting cross-study comparison in the foreground. Each cell typically contains one or two sentences, an effect estimate with confidence interval, a verbatim quote (for qualitative work), or a short judgment. The matrix is iterative; columns may merge or split as the synthesis matures and new themes emerge from the evidence. For implementation reviews, the matrix is often pre-structured around the Consolidated Framework for Implementation Research (CFIR) or RE-AIM. For diagnostic accuracy reviews, columns map to QUADAS-2 domains plus sensitivity, specificity, and prevalence. For prognosis reviews, QUIPS domains and outcome trajectories form the column set.

A defensible narrative synthesis tool for systematic reviews makes three things easy: building the matrix, exporting it for an appendix, and showing your synthesis decisions in a traceable way. Reviewers and supervisors will ask why a particular conclusion follows from the evidence, and the matrix is your audit trail. PRISMA 2020 expects the synthesis approach to be described in the methods, and the matrix is the cleanest artefact to attach as a supplementary file. For thesis literature reviews, the matrix doubles as a chapter-writing scaffold: each themed column becomes a section, the cells become the citations and arguments inside that section, and the row-level study attributes (sample, setting, design) become the contextualising sentences that open each evidence statement. Pair this tool with the data extraction form builder upstream to feed the matrix, the risk of bias assessment tool to add quality columns, and the GRADE evidence tool to translate the matrix into a summary-of-findings table when the synthesis is complete.

For mixed-methods and convergent designs, the matrix is the structural hinge between quantitative and qualitative evidence. The columns split into a quantitative side (effect estimates, sample sizes, statistical significance) and a qualitative side (themes, quotes, theoretical insights), with an integration column that records where the two converge, diverge, or extend each other. JBI and the Pluye and Hong (2014) framework both describe this convergent integration pattern explicitly. For framework synthesis, the columns are pre-specified from theory (for example, the COM-B model for behaviour change, or the Theoretical Domains Framework) and the matrix tests whether the included evidence fills, modifies, or contradicts each framework element. For meta-ethnography, the matrix supports reciprocal and refutational translation across studies, recording where one study's concept maps onto, contradicts, or extends a concept from another. In every case, the matrix is not the final synthesis; it is the structured scaffold from which the synthesis prose is written.