Systematic Literature Review: Data Extraction and Synthesis (S4.2)

Systematic literature reviews hinge on a disciplined two-step workflow after studies are selected: extract the right data in a consistent format, then synthesize results across studies to answer the review’s research questions. Data extraction turns raw study details into an organized evidence set—making it possible to compare findings, assess patterns, and ultimately draw conclusions that are traceable back to individual papers.

Extraction starts with a predefined data extraction form or spreadsheet to keep information consistent across included studies. Common fields include author, publication year, and journal, along with study objectives or research questions. Methodological details typically captured include study design, sample size, and data collection techniques. Reviews also extract key variables and how they were measured, plus each study’s main findings and conclusions. Limitations noted by the original authors are recorded as well, since they later affect how strong the evidence is considered. Depending on the review’s goals, additional columns may capture the theoretical framework used to explain relationships between variables, the study context, and how that context informs understanding of the review topic.

To support this work, extraction can be managed in Excel or Google Sheets, or through systematic review software such as Covidence and EPI Reviewer. Once extraction is complete, the next move is to create a structured summary of the extracted data. This often takes the form of an evidence table or data matrix that organizes studies by themes, methodologies, outcomes, or theoretical frameworks. Visual summaries—descriptive characteristic tables, frequency charts, or concept maps—help surface patterns, trends, and outliers. This structured overview becomes the foundation for synthesis and for identifying where evidence is consistent, mixed, or missing.

Synthesis integrates findings from multiple studies to produce broader conclusions. For example, in a review of corporate social responsibility (CSR) and organizational performance, extracted results might show some studies reporting no relationship, others reporting positive effects, and still others reporting negative effects. When results are mixed, the synthesis can conclude that evidence is inconclusive and use that inconsistency to highlight research gaps.

Synthesis can be qualitative or quantitative. Qualitative synthesis includes thematic analysis or narrative synthesis, particularly suited to conceptual or exploratory reviews, while quantitative synthesis uses meta-analysis when studies report comparable numerical data. Regardless of approach, synthesis typically involves identifying common themes or categories, comparing and contrasting outcomes to find consistencies and discrepancies, and assessing the strength of evidence while accounting for study quality. Evidence strength can depend on factors such as whether studies use secondary data with measurable impact (e.g., CSR investment effects) or whether proposed measurement tools have established psychometric properties like reliability and validity.

Finally, synthesis decisions must be transparent and systematic. That means documenting each step, reporting potential biases, heterogeneity, and limitations in the synthesis process, and being explicit about how implications connect back to the review’s research questions. The extracted spreadsheet can also directly feed tables and figures for reporting—such as tables of study characteristics, measures used, item counts, questionnaire administration methods, and analysis types—so future researchers can select appropriate tools and understand how conclusions were built.