yila-thematic-synthesis/README.md

Synthesising Qualitative Studies: A Practical Guide to Thematic Synthesis

Quantitative reviews have well-worn tools: effect sizes, forest plots and heterogeneity statistics. Reviews of qualitative research, such as interview studies, focus groups and ethnographies, need different methods, and many researchers are unsure how to move from a pile of papers full of quotations and themes to a synthesis that says something new. Thematic synthesis is one of the most widely used and accessible approaches. This guide explains how it works step by step, how to keep the process transparent, how to assess the studies you include, and how a source-linked assistant such as Yila can help organise the material while keeping every theme traceable to the studies that support it.

What thematic synthesis is for

Thematic synthesis, described by Thomas and Harden in 2008, adapts thematic analysis of primary data to the findings of published qualitative studies. It is well suited to questions about experiences, perceptions, barriers and facilitators: for example, how students experience online tutoring, or what prevents clinicians from adopting a new guideline. Its output is a set of descriptive themes that summarise what the studies report, followed by analytical themes that go beyond the original studies to answer the review question.

Other approaches exist, including meta-ethnography, framework synthesis and meta-aggregation. Each has strengths. Thematic synthesis is often chosen because its steps are explicit and it works well with moderate numbers of studies of varying depth.

Decide what counts as "data"

The first methodological decision is which parts of each paper you will analyse. Most thematic syntheses treat as data everything in the results or findings section: participant quotations and the authors' interpretations. Some reviewers also include relevant material from the discussion, but this risks mixing findings with speculation. Whatever you decide, apply it consistently and report it.

Import the relevant sections into a qualitative analysis tool such as NVivo, ATLAS.ti or a simpler alternative such as a spreadsheet with one row per extract. Record the study identifier and page for each extract so that every code can be traced back.

Stage one: line-by-line coding

Read the findings of each study and code them line by line, staying close to the text. Codes at this stage describe what is said, not what it means for your question: "felt embarrassed asking questions in a group", "valued flexible timing", "unsure whether tutor was qualified". Expect to create many codes, and create new ones freely as you move from study to study.

As codes accumulate, translate concepts between studies. Different studies may use different language for the same idea, and part of the work is recognising that "fear of judgement" in one paper and "worry about looking stupid" in another are the same concept. Keep a codebook with a short definition of each code and examples from more than one study.

Coding should be done by at least two people for a sample of studies, with discussion of differences. The goal is not perfect agreement but a shared understanding of what each code means.

Stage two: descriptive themes

Once all studies are coded, group related codes into descriptive themes. This is an iterative process: draft a hierarchy, check it against the data, revise. Descriptive themes remain close to the original studies. They summarise what participants and authors reported, for example "social risk of asking for help" or "practical convenience as a driver of uptake".

At this stage, write a short narrative for each descriptive theme with supporting quotations and a list of contributing studies. A simple matrix of themes by studies is useful here. It shows which themes are supported broadly and which rest on one or two papers, and it helps readers judge the strength of each theme.

Stage three: analytical themes

The step that distinguishes synthesis from summary is the development of analytical themes. Here you use the descriptive themes to answer your review question directly, often in ways the original studies did not. If the question is "what would make peer tutoring more acceptable to first-year students?", analytical themes might propose that tutoring is more acceptable when it reduces social exposure, when it fits around other commitments, and when the tutor's competence is visible.

Analytical themes require interpretation, and interpretation should be transparent. Record how each analytical theme draws on the descriptive themes, discuss alternatives within the team and be clear in the write-up about which statements are reported findings and which are your inferences.

Appraising the included studies

Qualitative studies vary in rigour, and a synthesis should take that into account. Tools such as the CASP qualitative checklist or the JBI checklist help assess whether a study's aims, methods, sampling, data collection, analysis and reflexivity are adequate. Reviewers rarely exclude studies on quality alone, but appraisal can inform how much weight a theme receives, or be used in a sensitivity check: does a theme survive if the weakest studies are removed?

Frameworks such as GRADE-CERQual go further and assess confidence in each review finding, considering methodological limitations, coherence, adequacy of data and relevance. Even if you do not apply the full framework, thinking through those four dimensions for each theme improves the discussion.

Reporting transparently

Follow a reporting guideline such as ENTREQ. Describe how studies were found and selected, which parts of each paper were treated as data, how coding was done and by whom, and how descriptive and analytical themes were developed. Include the theme-by-study matrix, either in the main text or as a supplement. Provide enough quotations for readers to see the grounding of each theme without overwhelming the narrative.

Where an assistant helps

The most time-consuming parts of thematic synthesis are locating relevant extracts and keeping track of which studies support which codes and themes. An assistant can help with both, provided its output stays traceable.

Yila is designed to turn papers and data into cited reviews, figures and slides, keeping each statement tied to the passage it came from. In a qualitative synthesis, that means a proposed theme summary can link directly to the extracts and studies behind it, which makes it easier to check that a theme is genuinely supported and to build the theme-by-study matrix. The interpretive work of developing analytical themes remains with the review team, but having the evidence trail intact makes that work easier to discuss and defend.

A step-by-step summary

1. Decide which parts of each paper count as data and import them with study and page references.
2. Code line by line, translating concepts between studies, and keep a codebook.
3. Double-code a sample and discuss differences.
4. Group codes into descriptive themes and build a theme-by-study matrix.
5. Develop analytical themes that answer the review question, recording the reasoning.
6. Appraise studies and consider confidence in each finding.
7. Report with ENTREQ, including the matrix and supporting quotations.

Conclusion

Thematic synthesis offers a clear path from many qualitative studies to findings that are more than the sum of their parts. Its strength lies in explicit stages: close coding, descriptive themes and analytical themes, each traceable to the original studies. Keeping that trail intact, whether by hand or with a source-linked assistant like Yila, is what allows readers to trust the interpretation and see where it came from.