Hour 1 · SLR · Step 3 of 1520%
1.1
Question & search strategy
Frame the question. Draft the Boolean string.
~10 min
Use any free AI chat tool (ChatGPT, Claude, Gemini) to frame the question with PICO or SPIDER, and to draft a Boolean string with the synonyms you'd otherwise miss. Remember: a generated string is only a starting point — it hasn't searched anything yet. Run it in the real databases, refine, and that run becomes your record.
AI prompts (3)
Prompt
PICO / SPIDER question framer
When: You have a rough topic and need a structured, reviewable research question.
You are a research methodologist helping me frame a systematic literature review question. My rough topic: <PASTE TOPIC HERE> My field: <e.g. education technology, public health, management> Please: 1. Reframe the topic as a PICO question (Population, Intervention, Comparison, Outcome). If qualitative, use SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type) instead, and say which you chose and why. 2. List 3 alternative framings at narrower/broader scope so I can pick the right grain. 3. Flag any concept that is ambiguous and would need an operational definition before searching. Be concise. Use bullet points.
Prompt
Boolean string builder (Scopus / Web of Science)
When: You have a framed question and want a first-draft search string with synonyms and field tags.
Act as a research librarian. Build a Boolean search string for Scopus AND a separate one for Web of Science. Research question: <PASTE PICO/SPIDER QUESTION> Time window: <e.g. 2015–2025> Document types: <e.g. articles and reviews only> Language: <e.g. English> For each concept block: - List synonyms, acronyms, plurals, British/American spellings. - Use truncation (*) and phrase quoting where appropriate. - Combine with AND between concepts, OR within a concept. - Apply field tags: TITLE-ABS-KEY( ... ) for Scopus, TS=( ... ) for Web of Science. Return: 1. Final Scopus string (one code block) 2. Final WoS string (one code block) 3. A short table of the concept blocks and their synonyms so I can audit your choices. Do not invent terminology — if you're unsure about a synonym, mark it [verify].
Always run the string in the actual database and refine. The documented database run is your auditable record, not the AI conversation.
Prompt
Inclusion / exclusion criteria drafter
When: Before screening — to lock down what counts as relevant.
Help me draft inclusion and exclusion criteria for a systematic literature review. Research question: <PASTE> Stage where each criterion will be applied: title/abstract OR full-text (state for each). Produce a single table with columns: | Criterion | Type (inclusion/exclusion) | Applied at | Rationale | Cover at minimum: population, study design, intervention/phenomenon, outcome, time window, language, document type, peer-review status. Add anything else my question implies. Keep each criterion atomic — one decision per row.