At the top of the JSTOR search results page, you’ll see a toggle that allows you to switch between JSTOR’s standard Keyword results or Semantic results.
While keyword results rely on exact matches to your search term(s), JSTOR’s semantic results use machine learning to find content that is conceptually related to your query, even if different wording is used. This may help you surface relevant items that may be missed in a traditional keyword search due to differences in terminology for the same subject.
For example, semantic results for “ecological resilience in urban planning” might also return results about “city adaptability” or “sustainable city design” as these concepts are contextually similar.
You can toggle between result types at any time and your search terms will be preserved.
| Keyword results | Semantic results | |
|---|---|---|
| How it works... |
Looks for exact matches between your search term(s) and an item’s metadata and content. Matches in an item’s Title and Author fields are weighted more heavily in terms of relevance. |
Uses machine learning to return the 25 most-relevant results based on the conceptual similarity of your search query and an item’s content. Your exact search terms may or may not appear within the item itself. |
| Results include... |
All academic and primary source content on JSTOR, including text, image, audio, and video content. |
Journal articles, book chapters, and research reports only. |
| Refine your results with... |
Use Boolean operators (AND, OR, or NOT), exact phrase matching (quotation marks), wildcards, and proximity searching to construct your query. You can also search within specific fields such as Title, Author, or Publication. In addition, you can refine your results using the facets and filters in the Refine Results sidebar. See An Introduction to Searching on JSTOR for more information. |
You can modify or adjust your search query in the search bar; however, semantic results do not support refinements in the query itself (such as Boolean operators or wildcards) or the use of facets and filters found in the Refine Results sidebar. Please note, you’ll only be shown content that you can access. |
| Example search queries |
|
|
Using the AI research tool with semantic results
In addition to the item page, individual participants in JSTOR’s AI research tool beta program will also see the option to use the tool on the semantic results page. This can help you identify items relevant to your research to focus on further.
When asking questions on the results page, the research tool will answer using only the top 25 results and will include citations to help you identify which item(s) a response is based on.