%202.png)
Systematic Literature Review
Protocol, documented search, two-stage screening, quality appraisal, structured extraction and synthesis, so your review is reproducible and survives examination.
Free. Works on Claude, ChatGPT, Gemini or any assistant that accepts a skill file.
What this skill does
The method, encoded.
A systematic review is a study whose data are published studies, and it is examined as one. That means a reader has to be able to repeat your search and arrive at your included set. Most reviews that fail do so for procedural reasons rather than intellectual ones: the protocol was written after screening, the flow diagram does not add up, the reasons for excluding thirty full texts were never recorded, or the appraisal instrument was designed for a study design the review did not include.
This skill runs the method end to end. It starts with a proportionality gate, because the most expensive mistake at masters level is starting a full systematic review and finishing a partial one, and it tells you when a scoping or structured review is the better and equally defensible choice. It then covers the protocol, the structured review question, decidable inclusion criteria, the database search and log, two-stage screening with reason codes, the reconciling flow diagram, design-appropriate appraisal, a piloted extraction form, and synthesis by claim rather than study by study.
Its distinctive step separates what the literature establishes from what it assumes and what it merely repeats, which is where a defensible research gap actually comes from. It produces protocols, tables, logs and critique. It does not write your review for you, and it will not populate a table with studies you have not read.
Best used for
- Review-based masters dissertations
- Review chapters that will be examined by a methodologist
- Defensibly establishing that a research gap exists
- Adjudicating a literature that appears to disagree
- Deciding whether a full systematic review fits the timeline
Typical inputs
What you give it.
A review question specific enough to write inclusion criteria against, Departmental expectations and marking rubric for a review at this level, Confirmed access to bibliographic databases and to full texts, A search log and retrieved record set, Full texts of candidate studies, The available hours and the submission date
Typical outputs
What you get back.
A dated review protocol with inclusion and exclusion criteria, A reproducible search appendix with full strings per database, A screening log with reason codes at full-text stage, A reconciling four-stage flow diagram, A characteristics of included studies table, A per-domain quality appraisal table, A synthesis organised by review sub-question, An establishes, assumes and repeats classification with citation traces, Separate review and evidence-base limitations, with confidence levels
Method coverage
What the skill works through.
- What a systematic review has to prove
- Is a full systematic review realistic for your timeline
- Writing the protocol before you search
- Turning a review question into decidable criteria
- Building and piloting the search
- Two-stage screening and recording reasons
- Making the flow diagram reconcile
- Choosing an appraisal instrument that fits your study designs
- Extraction that you only have to do once
- Synthesis by claim, not by study
- What the literature establishes, assumes and repeats
- Reporting limitations of the review and of the evidence
Download
Free skill. One file.
Enter your email once. Every skill you download after that takes a single click.
How to install
Add the skill file and the five kernel protocols to a Claude Project, a ChatGPT Project, a Gemini Gem, or paste them at the top of any assistant conversation. Then give it your real research material, not a description of it.
Download skillQuestions
Common questions.
What is the difference between a systematic review and a literature review?
A literature review is a reading of the field, shaped by the reviewer's judgement about what is worth including. A systematic review pre-specifies the question, the sources, the inclusion criteria and the synthesis method before searching, and documents every screening decision, so that someone else could repeat it and get the same included studies. The difference is procedure and reproducibility, not thoroughness.
Is a systematic review realistic for a masters dissertation?
Sometimes, and often not on the question as first drafted. Estimate the work from a pilot search before committing: screening runs at roughly 60 to 120 abstracts an hour, and appraisal plus extraction at 1 to 3 papers an hour. If the arithmetic does not fit, either narrow the question with a stated reason, or run a scoping or structured review and name it accurately. A well-executed scoping review marks better than a half-finished systematic one.
Do I need a second screener?
Full dual screening is best practice and often impossible for a single student. The workable alternative is to dual-screen a random subset of 10 to 20 percent with a peer, report the agreement, and declare single-reviewer screening as a limitation. That is a real quality claim and takes hours rather than weeks.
How do I choose a quality appraisal tool?
Choose it after you know which study designs your search actually returned, and pick an instrument built for those designs from the standard set used in your discipline. Applying a randomised-trial instrument to survey or interview research produces uniformly poor ratings that carry no information. Report appraisal per domain rather than as a single overall score.
Why does my flow diagram not add up?
Usually one of three reasons: duplicates were removed in two passes and counted once, studies found through citation searching never entered the identification stage, or reports and studies were conflated when one study generated several papers. Fix the underlying counts in the screening log rather than adjusting the diagram, because examiners check this arithmetic first.
Can I use AI to help with a systematic review?
For structure, protocol drafting, criteria wording, extraction form design and critique of your own draft, subject to your institution's AI policy and its declaration requirements, which vary by institution and by assessment. Not for generating search results, record counts, appraisal ratings or table rows for studies you have not retrieved and read. Where your institution prohibits AI assistance for the task, do not use it for that task.
The skill chain
Works well with.
Research where people already are.
Analyse it where you already work.
Yazi helps researchers conduct surveys, AI interviews and longitudinal research directly through WhatsApp.
