How to Synthesize Sources in a Literature Review

What it means to synthesize sources

A literature review can list accurate accounts of many papers and still read as a set of separate notes. The missing step is usually synthesis: discussing sources together around a question or an idea, rather than giving each source its own turn.

To synthesize sources is to examine several of them together and show how they relate. Those relationships may be agreement, disagreement, different explanations, different populations, different methods, different contexts, a theme that keeps returning, a change over time, a repeated limitation, or a question the studies leave open.

The relationship has to be one the sources actually support. A writer should not join two findings just because they sit in the same paragraph.

If you are still deciding how the whole review should be organised, start with How to Structure a Literature Review. This page stays with the smaller job: combining sources inside that structure.

Summary and source-by-source writing

A summary reports one source at a time. The shape is familiar: one author found this, the next found that, a third found something else. Each sentence may be accurate. The reader still has to guess how the studies speak to each other.

A synthesis organises several sources around an idea. A generic shape, not a sentence to copy, is: several studies associate one thing with another, but they explain that association differently. Studies using one method emphasise one account. Studies in another setting report a different one.

The names Study A, Study B and Study C below are placeholders. They are not real papers, authors, journals, years, or statistics.

When one paragraph per study is weak

A common draft gives Study A its own paragraph, then Study B, then Study C. That pattern can hide themes, repeat the same kind of sentence, and leave the review closer to an annotated bibliography. The reader has to build the comparison alone.

Source-by-source order is not always wrong. A short account of one important study, or a chronological discussion of how a debate moved, can be the right choice. The problem is using that order by default, when the point of the section is a relationship among studies.

Start from a question

Synthesis is easier when you know what you are trying to understand. These are prompts for reading, not a research question this page can choose for you:

• What do the studies generally agree on?
• Where do the explanations differ?
• Which methods produce different kinds of evidence?
• Do findings differ across contexts or populations?
• Which limitations appear more than once?
• What is still unresolved?

Do not bend the literature so that it justifies a question you had already decided to ask. If the sources do not support that question, the honest result is to change the question or to say the literature does not settle it.

Read for relationships and group sources

Notes that only answer “what did this source say?” are hard to synthesize. While you read, also ask what question the study addressed, what context it used, what method it used, what it found, what explanation it offered, which limitation matters, and which other sources it agrees or disagrees with. There is no single correct note-taking system.

Sources can then be grouped by theme, concept, finding, argument, method, population, context, period, theoretical perspective, or a disagreement. A source may belong in more than one group. Do not force every paper into exactly one theme so the outline looks tidy.

A matrix is optional

A synthesis matrix is a working table for your own notes. Columns might include the source, its research focus, the method or context, the key finding or argument, the theme it speaks to, how it relates to other sources, and a limitation or caution.

It is a way of laying notes beside each other. It is not a table most universities require, and it does not need to appear in the review. This page does not include an interactive matrix.

Name the pattern

After the notes are grouped, ask what pattern actually appears. This is not a meta-analysis and it does not produce a pooled statistic.

• Agreement. Several sources support a similar conclusion.
• Disagreement. Sources reach different conclusions.
• Qualification. A pattern appears, but only under some conditions.
• Method difference. Results differ partly because the studies were done differently.
• Context difference. Findings vary across settings or populations.
• Development over time. Later work changes or challenges an earlier explanation.

Name only the pattern you can point to in the sources. A difference in method does not automatically explain a conflict in findings. Prefer “may help explain”, “could reflect”, or “the studies differ in” until the sources themselves make the reason clear.

Build the paragraph around an idea

A synthesis paragraph is usually organised by an idea, not by an author. A useful shape, not a five-sentence rule, is: the point of the paragraph, evidence from more than one source where that evidence exists, the relationship among those sources, what that relationship means for the review, and a link to the next issue when one is needed.

Each of those moves has to stay inside what the sources support. The interpretation is yours. The findings are not.

A flexible framework

1. Point. State the relationship or theme this paragraph is about.
2. Evidence. Bring in the sources that speak to it.
3. Relationship. Show agreement, difference, a condition, a method, a context, or a change over time.
4. Interpretation. Say why that relationship matters to this review.
5. Transition. Move to the next issue only when the paragraph has finished its point.

A fictional example

Illustrative example using fictional source notes. These are not real studies. Do not cite them. Do not add findings that are not in the notes.

The topic is remote teamwork and communication. The only notes are these:

• Study A. Participants reported faster information sharing in remote teams.
• Study B. Participants reported coordination problems across time zones.
• Study C. Communication quality varied according to team routines.

The first version tells a reader what each note said. The second uses only those three notes: faster sharing, coordination trouble, and a difference tied to routines. It does not add a sample size, a country, a tool, or a cause the notes do not contain.

Agreement, disagreement, method and context

Agreement

“Many studies agree” is weak unless you say what they agree about, and unless the sources you have actually cluster that way. Where method or context matters, name it. Do not stretch a few similar findings into a field-wide consensus.

Disagreement

Disagreement is useful when it is real. Ask whether the studies differ in population, definition, method, setting, period, or in how authors read similar evidence. Do not assume the first difference you notice is the explanation. If the literature does point one way, say that. Do not invent a split so the paragraph looks balanced.

Method

Method can change what a finding is able to show. Interview studies may describe experiences or processes. Survey or other quantitative designs may estimate patterns or associations. That depends on the actual design. One approach is not automatically better than the other. A separate guide on qualitative and quantitative research is not part of this page.

Context and population

Studies may involve different ages, countries, industries, institutions, periods, or settings. A difference can matter. A finding from one context is not, by itself, a finding about every context.

Themes, and other ways to organise

Themes

A theme is not a broad label such as “technology” unless the sources are actually about technology as a whole. A more useful theme names a pattern, for example technology access as a condition that affects participation, and only when the sources support that pattern. A heading does not create the synthesis. Inside the theme, the sources still need to be compared.

Chronological synthesis

Order by time can still be synthesis when it shows what changed, what was challenged, and what stayed in place. A list of studies by year is not that.

Methodological grouping

Group by method when the method is relevant to the question you are asking of the literature. Interview-based work, surveys, experiments, and document-based work are examples, not a list every review must use.

Theoretical or conceptual grouping

Sources may use different theories, models, concepts, or explanatory frameworks. A review can compare what each framework helps explain and where its limit shows up. Do not invent a theoretical relationship the sources do not make.

What synthesis is not yet

Careful comparison may show repeated limitations, contexts with little study, conflicting findings, open questions, or areas that are weakly connected. Not every difference is a research gap. A gap has to follow from the literature, not from a topic you hope is empty. How to identify a research gap explains that reasoning.

The same comparison may show which questions still matter. Do not force the sources to defend a question you chose in advance. Research questions versus objectives explains how a question and an objective differ.

Language, overclaiming and thin evidence

Relationship words should match the relationship. Words such as “similarly”, “consistent with”, and “both studies indicate” fit agreement. “However”, “in contrast”, and “whereas” fit a difference. “Although”, “in some contexts”, and “this pattern was less clear when” fit a qualification. “Additionally”, “later work extends”, and “another perspective emphasises” fit an extension. Choose the word from the sources, not from a list.

These claims are often too wide: “all studies show”, “the literature proves”, “there is no research on”, and “researchers agree that”. Use them only when the review you have actually done supports them. Safer sentences keep the scope, the disagreement, and the uncertainty visible.

If only one source supports a point, do not invent or misquote other sources to complete a theme. Reconsider the paragraph, look for further real literature through your usual search, or say that the evidence you have is limited. This page does not teach database searching.

Sources must be real, findings must be reported as the source reported them, and a citation must point at the source you used. Disagreement and gaps are not props. The academic integrity commitment states the same boundary for ClarityScholar’s services. This article is not that policy.

Citations and how many sources

Each claim that depends on a source should be attributed to that source. How the citation looks is a formatting question. Citation and formatting reviews presentation against a style you supply. This article does not teach every style, and it does not set a number of citations per paragraph.

There is also no required number of sources in a synthesis paragraph. Two sources, several sources, or occasionally one source beside a wider analytical point can each be right. The test is whether the paragraph shows the relationship clearly.

One source may appear in more than one theme when it genuinely contributes to more than one issue. Do not give every source equal space for the sake of fairness. Space should follow relevance to the point.

If the literature sits inside a proposal and the problem is the proposal as a whole, research proposal editing is the closer service. It reviews a proposal you are writing. It does not supply the literature.

Common mistakes

• Giving every source its own paragraph when the section is about a relationship.
• Using a theme heading and then summarising the papers underneath it one by one.
• Stacking citations without saying how the sources relate.
• Treating every difference as a contradiction.
• Ignoring method or context when they change what the finding means.
• Writing a consensus the sources do not support.
• Inventing a gap, a disagreement, or an extra study.
• Giving every source the same amount of space.
• Reporting results without saying why they matter to this review.

A short checklist and one possible workflow

This checklist stays on this page. It is not a separate resource, and it is not a required template.

• Does the paragraph have an analytical point?
• Are sources brought together where more than one speaks to that point?
• Is the relationship among them stated?
• Have context or method been considered where they matter?
• Is the evidence separate from your interpretation?
• Is any claim wider than the literature you actually discuss?
• Does the section serve the purpose of the review?

One possible sequence, not the only one, is: clarify what the review is trying to understand, pull comparable information from the sources, group them by relationships that are real, name patterns and differences, plan paragraphs around ideas, draft with the sources those ideas need, then revise for overclaiming and for paragraphs that still read as isolated summaries.