Variation in number of hits for complex searches in Google Scholar

• Published in: Journal of the Medical Library Association Vol. 104; no. 2; pp. 143 - 145
• Main Author: Bramer, Wichor Matthijs
• Format: Journal Article
• Language: English
• Published: United States Medical Library Association 01.04.2016; University Library System, University of Pittsburgh
• Subjects: Algorithms; Bibliographic data bases; Citations; Database searching; Evaluation; Humans; Information Dissemination - methods; Information Storage and Retrieval - statistics & numerical data; Internet; Internet/Web search services; Journal Impact Factor; Library Science; Management; Medical journals; Medical Subject Headings; Online searching; Online services; Queries; Research Communications; Search Engine; Search engines; Search strategies; Netherlands; Search Engine; Reproducibility of Results; Review Literature as Topic; Information Storage and Retrieval
• ISSN: 1536-5050; 1558-9439; 1558-9439
• DOI: 10.3163/1536-5050.104.2.009

Google Scholar is often used to search for medical literature. Numbers of results reported by Google Scholar outperform the numbers reported by traditional databases. How reliable are these numbers? Why are often not all available 1,000 references shown? For several complex search strategies used in systematic review projects, the number of citations and the total number of versions were calculated. Several search strategies were followed over a two-year period, registering fluctuations in reported search results. Changes in numbers of reported search results varied enormously between search strategies and dates. Theories for calculations of the reported and shown number of hits were not proved. The number of hits reported in Google Scholar is an unreliable measure. Therefore, its repeatability is problematic, at least when equal results are needed.