%202.png)
Academic Literature Search Strategy
Build a documented, repeatable literature search: concept blocks, real synonyms, Boolean strings, database choice, citation chasing and a search log.
Free. Works on Claude, ChatGPT, Gemini or any assistant that accepts a skill file.
What this skill does
The method, encoded.
Most students search the way they search for anything else: type the topic into the nearest box, read the first page, stop. The result is either nothing or forty thousand items, and neither number tells you anything about the literature.
This skill replaces that with a method. It decomposes a research question into two to four concept blocks, then builds a synonym set for each one from real database records rather than from imagination, including the terms the field used before the current ones. That single step is usually the difference between a student who concludes their topic is unresearched and a student who finds the literature that has been there all along.
It then covers Boolean construction, phrases, truncation, proximity and field limits generically, so the principles transfer between databases; how to choose databases by discipline and why an indexed database is a different instrument from a general search engine; inclusion criteria written and dated before searching rather than fitted afterwards; two-pass screening with recorded reasons; backward and forward citation chasing; when grey literature belongs; and how to recognise genuine saturation rather than a string that keeps returning the same items.
The output is a search log: the artefact that makes the search reproducible, that is frequently marked directly, and that turns "I could not find much" into a defensible statement.
This is a rigorous search for a normal literature review. A formal protocol-driven systematic review is a different method.
Best used for
- Building the search behind a dissertation literature review
- Rescuing a search returning nothing or returning thousands of items
- Finding the vocabulary a field actually uses, including older terms
- Choosing databases for a cross-disciplinary question
- Producing the documented search log a marking rubric requires
- Defensibly establishing that little has been published on a question
- Updating a search before submission without redoing it
Typical inputs
What you give it.
Committed research question with population, phenomenon and context, Discipline, and any second discipline the question crosses into, Access to at least one indexed academic database, Optional two or three known-relevant items to test the string against, Optional institutional database list or subject guide, Optional date or language restrictions required by the discipline, Rubric wording on what the search must document
Typical outputs
What you get back.
Concept table with synonyms per block and the source of each term, Inclusion and exclusion criteria, written and dated before searching, Search log with database, exact string, limits, date, results and items taken forward, Database rationale, including databases not searched and why, Screening funnel counts with standard exclusion reasons, Citation chasing record, backward and forward, with what each round added, Explicit grey literature decision, Included source list with grey literature marked, Method paragraph written from the log, plus a limitations statement
Method coverage
What the skill works through.
- Why ad hoc searching fails
- Breaking a research question into concept blocks
- Building synonym sets from real records, not from memory
- Boolean operators, phrases, truncation and proximity
- Field limits and why title and abstract beats full text
- Choosing databases, and how they differ from a search engine
- Writing inclusion criteria before you search
- Piloting a search string against items you already have
- Screening in two passes with recorded reasons
- Backward and forward citation chasing
- When grey literature belongs in a review
- Recognising saturation and knowing when to stop
- Keeping a search log a marker can follow
- Documented search versus systematic review
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.
Why does my literature search return no results?
Almost always vocabulary or over-constraint. Either your field uses a different word for your construct, in which case harvest synonyms from the keyword fields of articles you already have and from the database's own thesaurus, or your string contains too many concept blocks joined by AND. Three blocks is usually the ceiling; extra requirements belong at screening, not in the string.
How do I build a Boolean search string?
Group synonyms for the same idea with OR, and join the different ideas with AND. Enclose multi-word concepts as phrases, truncate word stems carefully so variants are captured without pulling in unrelated words, and limit matching to title, abstract and subject headings rather than full text. Avoid NOT, which silently removes relevant records that mention the excluded term in passing.
How many databases should I search?
More than one, always. At minimum a subject-specific index for your discipline plus a broad multidisciplinary index, and one from the adjacent discipline if your question crosses fields. No database indexes everything, and a single-database review usually shows a visible bias toward one publisher or journal family.
Is a general web search engine enough for a literature review?
General search tools, including the scholarly ones, are useful for citation chasing, for grey literature and for finding a known item. They are a poor basis for the searching you report, because they have no controlled vocabulary, no stable result set and no inspectable ranking, so the search cannot be documented in a way anyone can repeat. Use indexed academic databases for the search you describe in your method.
What is a search log and why do I need one?
A record of every search you ran: database, exact string, limits, date, number of results, number taken forward. It is frequently a marking criterion, it makes your search reproducible, it lets you update the search before submission in an hour, and it is the only thing that turns "I could not find much" into a defensible statement rather than an admission. Record the searches that returned nothing too.
What is citation chasing?
Following the literature's own connections rather than the index's. Backwards means mining the reference lists of your best sources, which is the fastest route to foundational work and older terminology. Forwards means finding what has cited those sources since, which reaches recent work and critiques. Both routinely retrieve material no search string reaches.
Should I include grey literature?
Include it where the question is applied, policy-related, recent or geographically specific, and where practitioner work leads academic work. Grey literature is unrefereed rather than inferior, so the appraisal burden shifts to you. Whichever way you decide, state the decision and the reason rather than excluding by silence.
How do I know when to stop searching?
When three things are true together: new searches using new vocabulary return only items you have seen, a round of citation chasing adds nothing, and each concept has been searched in at least two databases and two vocabularies. Getting the same results from the same string is a fact about the string, not about the literature.
What is the difference between a documented search and a systematic review?
A documented search is transparent, logged and repeatable, and it supports a normal narrative or thematic review. A systematic review additionally requires a protocol written before searching, exhaustive rather than sufficient searching, pre-specified eligibility, dual independent screening, formal risk-of-bias assessment and reporting to a recognised standard. Do not describe a documented search as systematic; the claim is checkable against your log.
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.
