Skip to Main Content

Evidence Synthesis: Searching

Systematic Searching

Systematic searches are the foundation for your review. A systematic search is one that can be reproduced accurately by another person. 

There are several aspects of the search that must be documented in order to make it transparent and reproducible:

  • Databases searched 
    • Date range and/or version for each database
    • Platform used to access each database 
    • Date each database was searched
  • Keywords
    • Plain text keywords
    • Controlled vocabulary (i.e. subject headings)
    • Field tags and limiters
  • Search filters (i.e. language, publication type, etc.)
  • Steps in the search process
  • Number of results from each database
  • Number of results after deduplicating 

 

Search Methodology Guidelines

There are several trustworthy sources you can reference for guidance on performing systematic searches:

Why are multiple databases required for an evidence synthesis?

Evidence syntheses are meant to be exhaustive analyses of all the literature on a given topic. No single database is going to have every piece of literature on a topic, and this is why you need to search multiple databases.

How many databases are required?

You will need to conduct searches in multiple databases. The exact databases you search, and the number of databases you choose, will depend on your research topic. This is especially relevant for social science questions that are highly multidisciplinary; social science researchers may find that they need to search more databases than health science researchers when doing an evidence synthesis. 

Identifying Search Terms

Starting with the main concepts identified in your research question framework, identify words and phrases that can be used to describe each concept. Evidence syntheses are intended to be comprehensive scans of the literature, so you'll want to identify as many relevant synonyms as you can. Synonyms and concepts are joined together using Boolean operators to create a search query for an academic database.

Plain Text Variations

In order to be comprehensive, you'll want to think of every possible way that each concept of your research question might be described. Think about:

  • How the concept is described by professionals in the discipline
  • How the concept is described in plain language
  • Slang words describing the concept
  • Outdated terms that were previously used to describe the concept
  • Name brands used in lieu of more general descriptors (e.g., "Facebook" when talking about "social media")
  • Alternate spellings (e.g., behaviour, colour)  

Subject Headings

Some databases make use of subject headings. These subject headings act as keyword tags placed on materials ingested into and indexed in the database. The subject headings used for any given database are listed in its thesaurus. 

Using subject headings in evidence synthesis searches ensures that you cast a wide net and collect as many potentially relevant articles as possible. You will want to be careful that you don't overuse subject headings, however, because they could bring in too many unrelated results that will add time to your screening process.

Search Fields

Always indicate the field being searched for each keyword listed in your search strategy. It is common to search for each keyword in the title and abstract fields, although parts of the search may be narrowed to title-only or broadened to all-text.

Field formatting varies significantly between databases. When in doubt, use a database's advanced search page to generate the correct field tags. 

Translating Searches

After you create a search query for one database, you'll need to adjust the formatting to be readable in your other databases. This is called translating your search. A translated search does not need to be an exact duplication of your original search. You may opt to add in or leave out certain keywords due to the database's content scope. The same keyword can produce vastly different results in different databases.

Note that if you are using databases with subject headings, you'll need to manually look for equivalent subject headings when translating the search for another database. For databases that don't have a comprehensive subject heading thesaurus, remove all subject heading terms from those search queries.  

Supplemental Search Methods

Structured database searches should serve as the foundation of your search. You may choose to supplement them with less structured methods in order to ensure comprehensiveness.

Grey Literature

Grey literature (or gray literature) is a broad term referring to any materials produced outside of the traditional academic publishing environment and distribution channels. Grey literature can include conference proceedings, dissertations, white papers, government reports, pre-print articles, and much more.

Benefits of including grey literature in an evidence synthesis:

  • Addresses bias against the null 
  • Provides greater comprehensiveness 

Disadvantages of including grey literature in an evidence synthesis:

  • Very difficult to search - both in terms of documentation and reproducibility 
  • Not peer-reviewed 

 Where to find grey literature:

 

Hand Searching

If you know of any specific journals that are likely to have articles relevant to your research question, you can manually hand search by reviewing the tables of contents. You can also do this for conference proceedings. To document this type of search, identify the journals/conferences and volumes, issues, and/or years you will be reviewing.  

 

Forward and Backward Citation Chaining

Citation chaining (also called citation chasing or citation snowballing) is a specific type of hand searching. It can be implemented after the full text screening stage to make sure you have uncovered every possible relevant source. 

Backward chaining is when you look at the reference list of a relevant article to see which sources it cited. Forward chaining is when you look for documents that were published after that relevant article and cited it. Google Scholar, Scopus, and the Lens allow for forward citation chaining.

The tool citationchaser can also be used to facilitate citation chaining. 

Tips for Searching

Tips for Searching

  • Always run each database search separately, even if multiple databases are available on the same vendor platform (e.g. EBSCO or ProQuest).
    • This will allow you to report specific database results in your PRISMA diagram, which will increase the transparency of your search process and your review.
    • Some vendors automatically deduplicate citations when multiple databases are searched at the same time. This results in a search that is not transparent or replicable.
  • Include the vendor platform (e.g. EBSCO or ProQuest) used to access the database, if relevant, as well as the date range of each database's coverage. Some databases are available on multiple platforms and are offered with varying date range coverages. Individuals can retrieve different results when using the same database if they are using a different vendor platform or a different date range.
  • Unsure about which databases to search or which keywords to use? Look at methodology sections in reviews that have already been published on your topic to see what other experts in the field have used.
  • Start your project with "benchmarking articles" (also called seed articles or gold-standard articles) that fit every piece of inclusion criteria. These are articles that you expect to see in your search results. If you aren't seeing the articles in your search results, it may be an indication that you need to edit your search.

 

Sources

Anderson, N. K., & Jayaratne, Y. S. N. (2015). Methodological challenges when performing a systematic review. European Journal of Orthodontics, 37(3), 248-250. 10.1093/ejo/cjv022 Find@UNC

JBI Manual for Evidence Synthesis. (2024). (E. Aromataris, C. Lockwood, K. Porritt, B. Pilla, & Z. Jordan, Eds.). JBI. 10.46658/JBIMES-24-01. Link to full text

Haddaway, N. R., Grainger, M. J., & Gray, C. T. (2022). Citationchaser: A tool for transparent and efficient forward and backward citation chasing in systematic searching. Research Synthesis Methods, 13(4), 533-545. 10.1002/jrsm.1563 Link to full text