Synthesis in research is the process of combining, integrating, and interpreting findings from multiple studies to build a bigger picture than any single study provides on its own. Rather than just listing what individual researchers found, synthesis looks across studies to identify patterns, contradictions, and broader conclusions. It’s a core skill in literature reviews, systematic reviews, and any project that pulls together existing knowledge on a topic.
How Synthesis Differs From Summary
This distinction trips up a lot of people, so it’s worth getting clear. A summary describes what individual studies found, one by one. You might write a paragraph about Study A, then a paragraph about Study B, and so on. Each source gets its own moment, but they never really talk to each other.
Synthesis does something fundamentally different. It organizes information by themes, patterns, or questions rather than by individual sources. Instead of saying “Smith (2019) found X” and “Jones (2020) found Y,” a synthesis might say “Three studies found that early intervention improved outcomes, while two others found no significant effect, likely because they measured different populations.” You’re weaving sources together, making connections, and drawing conclusions that no single study made on its own. A summary functions like a neutral broker gathering facts. A synthesis is an act of analysis.
Why Synthesis Matters
Research literature grows fast. In many fields, thousands of papers are published each year on a single topic. Synthesis helps readers stay abreast of that flood by distilling it into coherent findings. It also reveals things that individual studies can’t: whether a result holds up across different populations, whether conflicting findings can be explained by differences in methodology, and where the genuine gaps in knowledge lie.
Synthesis also tests the limits of what we can generalize. If a treatment works in five studies conducted in urban hospitals but fails in two studies conducted in rural clinics, that pattern only becomes visible when someone pulls all seven studies together and looks at them side by side.
Types of Research Synthesis
Quantitative Synthesis (Meta-Analysis)
When multiple studies measure the same outcome using comparable methods, their results can be combined statistically. This is called meta-analysis. Researchers calculate a standardized effect size for each study, then pool those effects to estimate an overall result with greater precision than any single study could offer. A meta-analysis requires, at minimum, an estimate of the effect and a measure of its precision (like a confidence interval) from each included study. Simply counting how many studies had statistically significant results, sometimes called “vote counting,” is considered an unacceptable method because it ignores effect sizes and sample differences.
The Cochrane Handbook, a widely used reference for systematic reviews, specifies that only one outcome per study should contribute to any given synthesis, preventing a single study from being double-counted and inflating results.
Qualitative Synthesis
Not all research produces numbers. Studies based on interviews, focus groups, or ethnographic observation generate rich, text-based findings. Synthesizing these requires different tools. The two most common are thematic synthesis and meta-ethnography.
Thematic synthesis borrows methods from primary qualitative research and applies them across studies. It typically involves three overlapping stages: coding the findings of each study line by line, grouping those codes into descriptive themes, and then developing higher-level analytical themes that go beyond what any individual study reported. This approach is particularly good at identifying commonalities across studies.
Meta-ethnography, first proposed by Noblit and Hare in 1988, takes a different angle. It translates concepts from one study into the language and framework of another, building overarching metaphors or theories. It uses three techniques: reciprocal translational analysis (finding shared concepts), refutational synthesis (exploring contradictions between studies), and lines-of-argument synthesis (constructing a whole picture from studies of its parts). Meta-ethnography tends to be better at making visible the differences between studies, including where findings genuinely conflict.
Mixed Methods Synthesis
Some research questions require combining both quantitative and qualitative evidence. When findings from these two streams agree, integration is straightforward. When they conflict, researchers have several options: looking for sources of bias, re-examining methodological assumptions, seeking explanations from theory, or gathering additional data to resolve the discrepancy. Conflicting results aren’t necessarily a failure. They often point to important nuances that a single method would miss.
How to Actually Do It
If you’re working on a literature review or thesis, the practical side of synthesis comes down to a few key steps.
First, organize your sources by theme or variable rather than by author. A synthesis matrix is one of the most useful tools for this. It’s a simple table where each row is a study and each column is a theme, variable, or research question. Filling it in forces you to compare what each study says about the same topic, making patterns and gaps immediately visible. Many university libraries, including Johns Hopkins, recommend this approach as a starting point.
Second, assess the quality of each study before giving it equal weight in your synthesis. Not all evidence is equally reliable. Quality assessment (sometimes called critical appraisal) examines factors like whether participants were randomly assigned, whether outcome assessors were blinded to group assignments, whether dropout rates were acceptably low, and whether confounding variables were accounted for. Tools developed by organizations like the National Heart, Lung, and Blood Institute provide structured checklists for different study designs. A well-conducted synthesis doesn’t just combine findings. It weighs them by the rigor of the studies behind them.
Third, write by putting sources in conversation with each other. Use transitions that make relationships between findings explicit. Words like “similarly,” “in contrast,” “consistent with,” and “extending this finding” signal to the reader that you’re synthesizing rather than summarizing. The goal is a narrative where your voice drives the argument and individual sources serve as evidence, not where each source gets a standalone paragraph.
Reporting Standards for Formal Synthesis
If you’re conducting a systematic review, the PRISMA 2020 statement provides the most widely accepted reporting guidelines. It’s a 27-item checklist that covers everything from how you searched for studies to how you synthesized them. The 2020 update expanded its guidance on synthesis specifically, breaking it into six sub-items: which studies were eligible for each synthesis, how data were prepared, how results were displayed, what methods were used to synthesize results, how differences between studies were explored, and what sensitivity analyses were run.
Even if you’re not doing a formal systematic review, understanding these standards helps you evaluate the quality of published syntheses you encounter. A well-reported synthesis should make its methods transparent enough that someone else could replicate the process and reach similar conclusions.
Common Pitfalls
The most frequent mistake is writing a summary and calling it a synthesis. If your review reads as a series of “Study A found… Study B found…” paragraphs with no connective analysis, you haven’t synthesized anything. Another pitfall is treating all studies as equally credible without assessing their quality, which can lead you to draw conclusions heavily influenced by poorly designed research. A third is ignoring contradictory findings. Genuine synthesis doesn’t cherry-pick studies that agree with a preferred conclusion. It accounts for the full range of evidence, including results that complicate the picture, and tries to explain why studies reached different conclusions.