# How to Synthesize Findings from a Literature Review Matrix

## Quick Answer

- Use your completed literature review matrix as the raw material for thematic synthesis by coding each row's key findings, then grouping codes into themes that answer your research question.
- Build the synthesis section around themes instead of study-by-study summaries, supporting each theme with evidence from multiple matrix rows and noting where studies agree or conflict.
- A matrix organizes evidence but does not interpret it, so you must document your coding decisions and report limitations such as selective inclusion or inconsistent outcome measures across studies.

## What a Literature Review Matrix Does and Does Not Do

A literature review matrix is a structured table that records the essential features of each study you have read. Typical columns include author and year, study design, population or sample, intervention or exposure, outcome measures, key findings, and quality or risk-of-bias notes. The matrix turns a stack of PDFs into a comparable dataset that you can sort, filter, and scan without rereading every full text.

The matrix serves three practical functions. First, it forces you to extract the same categories of information from every study, which reduces the chance that you will remember findings from one paper more vividly than another. Second, it makes gaps visible, such as populations that no study covered or outcomes that only one research group measured. Third, it gives you a stable reference during writing so that every claim in your narrative can be traced back to a specific row and source.

What the matrix cannot do is interpret the evidence for you. A row that says "n = 120, significant improvement" does not tell you whether the study was well designed, whether the effect size matters in practice, or whether the finding applies to the population you care about. Synthesis is the intellectual work of comparing, contrasting, and combining the evidence that the matrix has organized. The matrix is a tool for that work, not a substitute for it.

The distinction matters for your writing. If you simply restate each row in paragraph form, you have produced an annotated bibliography, not a synthesis. A synthesis answers questions such as: What do these studies collectively say about the effectiveness of the intervention? Where do findings diverge and why might that be? What methodological weaknesses limit confidence in the overall picture? What remains unknown and needs further study?

## Preparing Your Matrix for Synthesis

Before you begin coding and grouping, check that your matrix is complete enough to support a defensible synthesis. A matrix built from a haphazard search will produce a haphazard narrative regardless of how carefully you analyze it.

### Confirm Your Search Was Systematic

A synthesis is only as trustworthy as the search that fed it. If you gathered studies by asking colleagues or reading only the first page of search results, your matrix will reflect those gaps. Systematic searching means documenting your databases, search terms, inclusion and exclusion criteria, and the dates you ran each search. The National Library of Medicine's bookshelf provides access to authoritative biomedical research-method references that describe how to plan and report a systematic search, including how to document your process so another researcher could reproduce it.

Your matrix should include a column for how each study was identified, such as database name and search date. This column lets you check whether your synthesis accidentally overweights studies from one database or one search strategy. If you find that most of your included studies came from a single source, you may need to run additional searches before you trust your thematic conclusions.

### Verify Each Row Contains Comparable Information

Synthesis requires comparison, and comparison requires that you extracted the same kind of information from each study. If one row records the intervention dose but the next row does not, you cannot compare doses across studies. If one row records the statistical significance but not the effect size, you cannot judge whether a significant result was practically meaningful.

Fill in missing cells before you start coding. Return to the full text of any study where your extraction is incomplete. If a study genuinely does not report a particular detail, write "not reported" in the cell instead of leaving it blank. A blank cell is ambiguous, but "not reported" is a finding in itself that you can discuss in your limitations section.

### Add a Quality or Risk-of-Bias Column

Your synthesis will need to weigh studies differently based on their methodological strength. A large randomized trial should carry more weight in your conclusions than a small uncontrolled case series, even if both appear in your matrix. Add a column where you record your assessment of each study's quality or risk of bias.

Use a consistent and transparent approach to this assessment. You might record a simple rating such as high, moderate, or low quality, or you might use a structured tool that scores specific domains such as randomization, blinding, and attrition. Whichever approach you choose, document the criteria you used so that a reader can understand why you rated one study high quality and another low. The EQUATOR Network maintains a collection of reporting guidelines that can help you select an appropriate quality-assessment approach for your study design and topic area.

## Coding Your Matrix for Thematic Synthesis

Coding is the process of labeling segments of your matrix data with short descriptive tags that capture their meaning. In thematic synthesis, you code the findings recorded in your matrix rows instead of the full text of each paper. This is what makes matrix-based synthesis efficient, because you work with your extracted summaries instead of rereading every article.

### Choose a Coding Unit

Decide what counts as a codable unit before you begin. The most practical unit is a single finding or result within a matrix row. A study that measured three different outcomes might contribute three separate codes, one for each outcome. A study that found a positive effect in one subgroup but no effect in another might contribute two codes that capture that distinction.

Record your coding unit decision in your methods notes. This decision affects how many codes you generate and how you count evidence for each theme. If you code at the level of individual findings, a study with many outcomes will contribute more codes than a study with one outcome, which can skew your sense of how much evidence supports a theme. Be aware of this dynamic and consider whether you need to weight by study instead of by finding.

### Generate Codes From the Matrix Content

Read through your matrix rows and assign a short label to each distinct finding or observation. The label should describe the content in your own words instead of quote the study's language. For example, a row that reports "participants in the intervention group showed improved glycemic control at 12 weeks" might receive the code "glycemic improvement at 12 weeks."

Keep your codes close to the data at this stage. Resist the temptation to create abstract or interpretive codes too early. The goal is to capture what the studies found, not what you think the findings mean. Interpretation comes later when you group codes into themes.

Work through the entire matrix before you review your code list. You will likely generate dozens of codes, and some will overlap or duplicate each other. That is expected. The next step is to consolidate and group them.

### Consolidate Duplicate and Overlapping Codes

After you have coded every row, review your full list of codes and merge those that describe the same finding in different words. For example, "glycemic improvement at 12 weeks" and "better glucose control after three months" probably describe the same outcome and should become one code.

This consolidation step is where you begin to see patterns. If many studies produced a code about short-term improvement but few produced a code about long-term maintenance, that imbalance is a substantive finding about the state of the evidence. Record these observations as you go, because they will feed directly into your synthesis narrative.

## Grouping Codes Into Themes

Themes are broader patterns that emerge when you group related codes together. A theme answers a question that spans multiple studies, such as "What evidence exists for the effectiveness of this intervention?" or "What factors appear to influence whether the intervention works?"

### Sort Codes by Similarity

Group codes that share a common idea or refer to the same aspect of your research question. You might sort codes into piles based on the outcome they describe, the population they studied, the intervention they tested, or the direction of their findings. The sorting criterion depends on your research question.

For a question about intervention effectiveness, you might group codes by outcome domain, such as clinical outcomes, patient-reported outcomes, and adverse events. For a question about mechanisms, you might group codes by proposed pathway or mediator. For a question about applicability, you might group codes by population subgroup or setting.

### Name Each Theme With a Descriptive Label

Once you have grouped your codes, give each group a name that captures its unifying idea. The theme name should be descriptive and specific enough that a reader can tell what evidence sits under it. A theme called "effectiveness" is too broad to be useful. A theme called "short-term glycemic improvement in adults with type 2 diabetes" tells the reader exactly what evidence the theme covers.

Write a short definition for each theme that states what kind of evidence belongs in it and what question it answers. This definition will keep you consistent as you write, and it will help you decide where to place a finding that could fit under more than one theme.

### Check That Each Theme Has Sufficient Support

A theme needs evidence from more than one study to be credible in a synthesis. If a theme contains codes from only a single study, you have a finding, not a theme. You can still report that finding in your narrative, but you should flag it as based on limited evidence instead of presenting it as a robust pattern.

Count the number of studies and the number of findings that support each theme. Record these counts in your notes. They will help you write accurately about the strength of the evidence, and they will alert you to themes that are thinner than they first appeared.

## Building the Thematic Framework

The set of themes you have identified, along with the relationships between them, forms your thematic framework. This framework is the skeleton of your synthesis section. Before you write, arrange your themes into a logical order that tells a coherent story.

### Order Themes by Your Research Question

The order of your themes should follow the logic of your research question, not the order in which you happened to code your matrix. If your question asks about effectiveness, you might start with themes about what the intervention does, then move to themes about how it works, then themes about factors that modify its effects, and finally themes about harms or limitations.

If your question asks about a body of evidence more broadly, you might start with themes about the volume and types of studies available, then move to themes about what the studies found, and end with themes about gaps and uncertainties.

### Identify Relationships Between Themes

Themes rarely stand alone. Look for relationships such as one theme explaining another, two themes that conflict with each other, or one theme that only appears under certain conditions. These relationships become the analytical content of your synthesis.

For example, you might find a theme about short-term effectiveness and a separate theme about long-term attrition. The relationship between these themes, that early gains may not persist because participants drop out, is a synthesis insight that neither theme alone captures. Make these connections explicit in your writing.

### Consider Whether a Visual Framework Helps

A simple diagram of your thematic framework can help you see gaps and relationships before you write. You might draw boxes for each theme and arrows for the relationships between them. This is a working tool for your own thinking, and you can decide later whether to include a version of it in your paper.

```mermaid
flowchart TD
    A[Completed Literature Review Matrix] --> B[Code Each Row's Findings]
    B --> C[Consolidate Duplicate Codes]
    C --> D[Group Codes Into Themes]
    D --> E[Check Theme Support Across Studies]
    E --> F[Arrange Themes by Research Question]
    F --> G[Write Synthesis Section]
    G --> H[Document Limitations and Gaps]
```

## Writing the Synthesis Section

The synthesis section translates your thematic framework into prose. The structure should follow your themes, with each theme getting its own subsection or paragraph cluster. Within each theme, you compare and contrast the evidence from the studies in your matrix.

### Open With the Overall Picture

Start your synthesis section with a paragraph that states what the body of evidence shows at a general level. This paragraph should summarize the volume and type of evidence you found, the overall direction of findings, and the main caveats that will be developed in the rest of the section.

Do not start with a study-by-study description. The reader needs the big picture first, then the detailed comparison of evidence within each theme.

### Develop Each Theme With Comparative Evidence

For each theme, present the evidence from multiple studies and show how it fits together. Use your matrix to identify which studies support the theme, which studies contradict it, and which studies are neutral or unclear. The synthesis is in the comparison, so do not simply list what each study found.

A useful structure for each theme is to state the theme's claim, present the supporting evidence from the strongest studies, note any conflicting or inconsistent findings, and explain how you interpret the balance of evidence. Where studies disagree, consider whether differences in design, population, intervention, or outcome measurement explain the disagreement.

### Use Your Quality Ratings to Weigh Evidence

Your quality or risk-of-bias ratings should influence how you describe the evidence. When high-quality studies support a theme, you can state the finding with more confidence. When only low-quality studies support a theme, you should describe the evidence as preliminary or suggestive and explain the methodological limitations that reduce confidence.

Be transparent about this weighing process. If you conclude that a theme is supported mainly by low-quality studies, say so and name the specific weaknesses, such as small sample sizes, lack of control groups, or high attrition. The EQUATOR Network's reporting guidelines can help you identify the specific reporting elements that a well-conducted study should include, which gives you a concrete basis for judging quality.

### Address Contradictions Directly

Do not hide or smooth over contradictions between studies. A synthesis that reports only agreement is incomplete and potentially misleading. When studies conflict, describe the conflict, explore possible explanations, and state what additional evidence would be needed to resolve it.

Common explanations for conflicting findings include differences in study populations, differences in intervention dose or duration, differences in outcome measurement, and differences in methodological quality. Your matrix should contain the information needed to explore these explanations, because you recorded population, intervention, outcome, and quality for every study.

### End With Gaps and Uncertainties

Close your synthesis section with a paragraph or subsection that identifies what the evidence does not tell you. These gaps should emerge naturally from your thematic analysis. If a theme had support from only one study, that is a gap. If no study measured a particular outcome, that is a gap. If all studies were conducted in one country or one population, that is a gap.

Gaps are not failures in your work. They are legitimate findings that help readers understand the limits of the evidence and that point toward future research. Reporting them accurately is part of responsible synthesis.

## At a Glance

| Synthesis Step | What You Do | Common Output | Typical Time |
| --- | --- | --- | --- |
| Prepare matrix | Complete missing cells, add quality ratings, verify comparable extraction | Complete and consistent matrix | 2 to 4 hours |
| Code findings | Label each distinct finding with a short descriptive tag | Code list linked to matrix rows | 1 to 2 hours per 20 studies |
| Group codes into themes | Sort codes by similarity, name and define each theme | Thematic framework with defined themes | 1 to 2 hours |
| Check theme support | Count studies and findings per theme, identify thin themes | Support table for each theme | 30 to 60 minutes |
| Write synthesis | Draft theme-by-theme narrative with comparisons and gaps | Draft synthesis section | 4 to 8 hours |

## Common Failure Patterns in Synthesis

Several recurring problems produce weak synthesis sections. Recognizing these patterns in your own draft is the first step to fixing them.

### The Annotated Bibliography Pattern

This failure occurs when you write one paragraph per study, describing what each study found without comparing studies to each other. The resulting text reads like a list, and the reader cannot see the overall picture. The fix is to reorganize your writing around themes instead of around studies. Each paragraph should discuss a theme and draw evidence from multiple studies.

### The Vote-Counting Pattern

This failure occurs when you count how many studies found a significant effect and how many did not, then declare the side with more studies the winner. Vote counting ignores study quality, effect size, and the precision of estimates. A single large high-quality study can provide more reliable evidence than ten small underpowered studies. The fix is to weigh studies by quality and to report the magnitude and consistency of effects, beyond their statistical significance.

### The Cherry-Picking Pattern

This failure occurs when you select only the studies that support your preferred conclusion and ignore or downplay studies that conflict with it. Cherry picking can be intentional or unintentional, but the result is the same, a synthesis that misrepresents the evidence. The fix is to report all relevant findings from your matrix, including those that complicate your narrative, and to address contradictions explicitly.

### The Overreach Pattern

This failure occurs when you draw conclusions that go beyond what the evidence supports. You might generalize from studies conducted in one population to a different population, or you might claim causality from observational studies, or you might recommend a practice based on a single small study. The fix is to match your claims to the strength and type of evidence, and to state the limits of generalizability explicitly.

### The Lost-Matrix Pattern

This failure occurs when you write a synthesis that cannot be traced back to your matrix. The reader cannot tell which studies support which claims, and you cannot defend your conclusions against challenge. The fix is to cite specific studies for each claim in your synthesis and to keep your matrix updated as you write so that every statement has a source.

## Records and Documentation for Synthesis

Good documentation of your synthesis process supports the credibility of your work and protects you if a reader or reviewer questions your conclusions. Keep the following records as you work.

### Your Coding Log

Record the codes you generated, the matrix rows each code came from, and the date you created each code. This log lets you reconstruct how you moved from matrix to themes, and it helps you notice if your coding drifted over time, such as becoming more abstract or more detailed as you worked through later studies.

### Your Theme Definitions

Write a definition for each theme that states its scope, the question it answers, and the criteria for including evidence under it. These definitions prevent you from shifting the meaning of a theme as you write, and they help a reader understand exactly what you mean by each theme name.

### Your Support Counts

Record how many studies and how many findings support each theme. This count is the factual basis for statements such as "most studies found" or "evidence is limited to two small studies." Without this record, you risk overstating or understating the strength of the evidence.

### Your Search Documentation

Keep the records of your database searches, including the databases, search terms, filters, and dates. This documentation is part of responsible research practice and may be required by your institution or by a journal if you submit your work for publication. The National Library of Medicine's research-method references describe how to document a search so that it is reproducible.

## Ethical and Professional Considerations

Synthesis is also a technical exercise. It carries ethical responsibilities to the research community and to the readers who will rely on your conclusions.

### Report Honestly and Completely

Your synthesis should represent the evidence fairly, including findings that complicate your narrative. Selective reporting of favorable results is a form of research misconduct that undermines trust in the scholarly record. The Committee on Publication Ethics describes core practices for research integrity, including expectations for honest reporting and transparent handling of data and findings.

### Acknowledge the Limits of Your Synthesis

Every synthesis has limits. Your search may have missed relevant studies, your quality assessments involve judgment, and your thematic groupings reflect your interpretive choices. Acknowledge these limits in your writing so that readers can calibrate their confidence in your conclusions.

### Follow Reporting Guidelines Where Applicable

If your synthesis is part of a systematic review or another formal research product, reporting guidelines can help you structure your work and ensure that you report all necessary elements. The EQUATOR Network provides a searchable collection of reporting guidelines for many study types, and consulting the appropriate guideline before you write can prevent omissions that would weaken your report.

### Manage Your Researcher Identity

If your synthesis will be published or shared publicly, make sure your researcher identity is accurate and up to date. An ORCID record provides a persistent identifier that links your work to you, which is useful when you share your synthesis across platforms or submit it for publication. Keeping your record current ensures that your work is correctly attributed to you.

## Limitations of Matrix-Based Synthesis

Matrix-based thematic synthesis is a practical and widely used approach, but it has limitations that you should understand and acknowledge.

### Loss of Context and Nuance

Your matrix contains extracted summaries, not the full text of each study. The extraction process necessarily loses context, such as the precise wording of outcome measures, the details of subgroup analyses, and the authors' own interpretations of their findings. If a nuance matters for your synthesis, you must return to the full text to recover it.

### Dependence on Extraction Quality

Your synthesis is only as good as your extraction. If you extracted information inaccurately or inconsistently, your codes and themes will reflect those errors. This is why checking your matrix for completeness and comparability before coding is essential.

### Interpretive Judgment in Coding and Grouping

Thematic synthesis requires judgment at every step, from choosing codes to grouping them into themes. Two researchers working from the same matrix could produce different thematic frameworks. This does not make the approach invalid, but it means you should document your decisions and be open about the interpretive nature of your analysis.

### Risk of Overweighting Well-Reported Studies

Studies that report their methods and findings clearly are easier to code and may contribute more codes than studies with sparse reporting. This can lead you to overweight well-reported studies in your themes. Your quality ratings can partially correct for this, but you should be aware of the dynamic.

## Professional Escalation Criteria

Some situations call for seeking additional help instead of proceeding on your own. Consider consulting a supervisor, librarian, methodologist, or statistician in the following circumstances.

### When Your Search Is Not Reproducible

If you cannot document your search process well enough for another person to reproduce it, stop and get help before proceeding. A librarian or information specialist can help you design and document a systematic search. The National Library of Medicine's research-method references describe the elements of a reproducible search.

### When You Cannot Assess Study Quality

If you are uncertain how to judge the quality or risk of bias of the studies in your matrix, consult a methodologist or use a structured quality-assessment tool. Making up quality ratings without a defensible basis will undermine your synthesis.

### When Your Themes Are Too Thin

If your thematic analysis produces many themes with support from only one or two studies, your synthesis will be weak. Consider whether your research question is too broad, whether your search missed relevant studies, or whether you need to consult a supervisor about reframing your question.

### When You Find Evidence of Research Misconduct

If you encounter a study that appears to involve fabricated data, duplicate publication, or other misconduct, do not ignore it. The Committee on Publication Ethics describes how journals and institutions handle allegations of misconduct, and you should report your concerns through the appropriate channel, such as the journal that published the study or your institution's research integrity office.

## Frequently Asked Questions

### How is a synthesis different from a summary?

A summary restates what individual studies found. A synthesis compares and combines findings across studies to answer a broader question. In a synthesis, you group evidence by theme, weigh studies by quality, address contradictions, and identify gaps. A summary stays within each study, while a synthesis moves across studies.

### How many studies do I need before I can synthesize?

There is no fixed minimum number. Thematic synthesis can work with as few as three or four studies if they address the same question, though the conclusions will be limited. With more studies, you can identify stronger patterns and explore reasons for conflicting findings. The key is that your synthesis must accurately reflect the amount and quality of evidence you have, not that you reach a particular count.

### Can I synthesize studies with different designs?

Yes, and this is often necessary because the evidence on a question may come from a mix of randomized trials, observational studies, and qualitative research. When you synthesize across designs, you should address how design affects the strength of conclusions. For example, you might treat findings from randomized trials as stronger evidence for causality than findings from cross-sectional surveys.

### What do I do when studies in my matrix report conflicting findings?

Report the conflict directly and explore possible explanations. Look at your matrix columns for differences in population, intervention, outcome measurement, and study quality that might explain why findings diverge. State what additional evidence would be needed to resolve the conflict. Do not simply count studies on each side and declare a winner.

### How do I decide which studies to emphasize in

## Using the Evidence

| Source | Best use in this topic | Important limitation |
|---|---|---|
| [Research Methods Resources](https://www.ncbi.nlm.nih.gov/books) | official guidance | Check the linked page for current local requirements |
| [EQUATOR Network](https://www.equator-network.org/) | official guidance | Check the linked page for current local requirements |
| [Core Practices](https://publicationethics.org/core-practices) | official guidance | Check the linked page for current local requirements |

## Related Articles

- [Protein Synthesis Step By Step](/blog/guides/protein-synthesis-step-by-step)
- [Protein Synthesis: A Step-by-Step Guide to Transcription and Translation](/blog/guides/protein-synthesis-a-step-by-step-guide-to-transcription-and-translation)
- [Scientific Literature Review: A Reproducible Search and Synthesis Workflow](/blog/careers/scientific-literature-review-a-reproducible-search-and-synthesis-workflow)
- [Scientific Literature Review Format: A Reproducible Structure for Researchers](/blog/careers/scientific-literature-review-format-a-reproducible-structure-for-researchers)
- [The Purpose of Meta-Analysis: Why Synthesize Studies Quantitatively?](/blog/guides/the-purpose-of-meta-analysis-why-synthesize-studies-quantitatively)

## References and Further Reading

- [Research Methods Resources](https://www.ncbi.nlm.nih.gov/books). National Library of Medicine.
- [EQUATOR Network](https://www.equator-network.org/). EQUATOR Network.
- [Core Practices](https://publicationethics.org/core-practices). Committee on Publication Ethics.
- [NIH Grants and Funding](https://grants.nih.gov/). National Institutes of Health.
- [ORCID for Researchers](https://info.orcid.org/researchers). ORCID.
- [Data Management and Sharing Policy](https://sharing.nih.gov/data-management-and-sharing-policy). National Institutes of Health.
- [NCBI Literature Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information.
- [PubMed](https://pubmed.ncbi.nlm.nih.gov/). National Library of Medicine.

> This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.