OpenNeedle Ask your own

Question explored with the scientific record

Answered on this page ↓

How do you decide what studies to retrieve for a question?

Sep 3, 2026 · 6 sources used · OpenNeedle synthesis
The short version: you are asking about my method for selecting studies, and the evidence block is mostly about technical search algorithms, not about how I decide what to retrieve.

The evidence block contains 21 records. Most are about information retrieval systems: ant colony optimization for web graphs [1], query expansion in MEDLINE [3], relevance ranking in search engines [4], and multi-agent distributed retrieval [5]. One record argues that using LLMs to make relevance judgments makes the generating model the ceiling for measurable performance [2]. Another evaluates AI tools for systematic reviews and finds they improve search efficiency but raise ethical concerns about bias and the need for human oversight [23].

None of these records describe how this specific system decides what studies to retrieve for a question. The retrieval system matched keywords in your question and returned what it found. What it found was mostly about how other systems retrieve information, not about this system's own method.

My call: the evidence here does not answer your question about how I decide what studies to retrieve. The retrieval was a miss on the exact question. Confidence: high that the evidence does not address your question, not clear on the question itself.

Keep digging

Sources used 6

  1. Biological-inspired algorithms for dynamic search in web graphs 2011 IEEE Symposium on Computers and Communications (ISCC) (2011) Thin

    This study presents a novel Ant Colony Optimization algorithm that dynamically constructs a search graph to improve the precision of web searches by mimicking the foraging behavior of ants, achieving a satisfactory effectiveness rate of approximately 84.4% in retrieving relevant…

    DOI: 10.1109/ISCC.2011.5983954
  2. Don't Use LLMs to Make Relevance Judgments Information Retrieval Research (2025) commentary Mixed

    A position paper arguing that using LLMs to generate relevance judgments makes the generating model the ceiling for measurable IR performance, so such judgments cannot reliably evaluate systems that improve on the model.

    DOI: 10.54195/irrj.19625
  3. Query expansion and MEDLINE Information Processing & Management (1996) Thin

    This study evaluates automatic MEDLINE query expansion strategies to add MeSH terms to user queries using a statistical thesaurus and retrieval feedback within the SMART system, showing significant retrieval gains over a baseline and identifying retrieval feedback as the most ef…

    DOI: 10.1016/0306-4573(95)00076-3
  4. An evaluation of relevancy ranking techniques used by internet search engines Library and Information Research (2013) Thin

    Subjective evaluation of relevancy ranking across five Internet search engines using 39 student participants shows Excite and Infoseek outperform Lycos and other engines, with around 40% relevance for WebCrawler, Excite, Infoseek, and AltaVista, and results limited to January 19…

    DOI: 10.29173/lirg306
  5. Personalised distributed information retrieval-based agents International Journal of Intelligent Systems Technologies and Applications (2010) primary study Strong

    PDIRBA uses a multi-agent, personalised approach to distributed information retrieval to improve relevance, reduce response times, and increase extensibility.

    DOI: 10.1504/ijista.2010.033896
  6. Evaluating the Efficacy of AI Tools in Systematic Literature Reviews: A Comprehensive Analysis Journal of Information Systems and Informatics (2025) systematic review Strong

    This systematic review evaluates the efficacy of AI tools in conducting systematic literature reviews, finding that they improve search, screening, and data extraction efficiency but raise ethical concerns such as bias and the need for human oversight.

    DOI: 10.51519/journalisi.v7i1.1035

Your question next

What do you want to know?

No question is too uncomfortable for the evidence. Bring yours.

Ask your question