Develop a Search Strategy
Definition: A systematic plan for identifying relevant literature.
Steps:
- Identify key concepts (e.g., "stress reduction," "mindfulness").
- Generate synonyms (e.g., "anxiety" OR "worry").
- Combine with Boolean operators (AND, OR, NOT).
- Translate strategies between databases using tools like the Polyglot Search Translator.
A high quality search strategy includes two kinds of search terms: keywords (any words or phrases that you choose) and subject headings (specific terms created by the database to represent particular concepts). Including multiple search terms for each of your concepts ensures that your search strategy captures all literature relevant to your question. The goal is to retrieve a comprehensive collection of literature, not just from authors who'd have used the same words, word forms, spelling and hyphenization as you.
Example Search String:
(mindfulness OR meditation) AND (stress OR anxiety) AND (college students)
If you feel stuck, there are several sources available to help you discover additional search terms:
- Abstracts/full texts of pre-identified relevant articles and results of preliminary searches on your topic
- Relevant broader and narrower subject headings in different databases
- Yale MeSH Analyzer: copy in a description of your topic (potentially your protocol) to have a list of relevant MeSH (PubMed's subject headings) suggested for you.
- Experts on the topic, including your librarian
- Textbooks on the topic
Select Databases
Key Resources:
- Multidisciplinary: Web of Science
- Sciences: Science Citation Index
- Social Sciences: PsycINFO, ERIC, Sociological Abstracts
- Grey Literature: OpenGrey, arXiv, conference proceedings, thesis repositories
AI-assisted Literature Discovery
1. Open Knowledge Maps: A Harvard supported open-source tool that generates visual knowledge maps that cluster related research papers by topics, enabling intuitive exploration of scientific literature.
2. ElicitAI: Searches across 125 million publications from the Semantic Scholar corpus, based on natural language query.
3. Research Rabbit: Facilitates literature exploration and visualization of study connections and research clusters.
4. Connected Papers: Maps related studies in a visual graph to support discovery of influential works and trends.
