A comparison of three main scientific literature databases using a search in aquatic ecology

Abstract

Online searches for relevant scientific references using keywords have become common practice. Several multidisciplinary scientific online databases are available, of which Web of Science, Scopus (both payable) and Google Scholar (free of charge) are the most commonly used. We test the hypothesis that results of highly similar searches in these three databases do not necessarily give comparable results. We set out to query the three databases with a real example on “diapause in microcrustaceans” (Cladocera, Copepoda and Ostracoda), using the same time period (2012–2021), the same keywords with the same syntaxis and the same sorting criterion (“relevance”), and compared the first 100 hits provided by each database. There were several references provided which were irrelevant to the search, especially in the Web of Science, and of the remaining relevant references, only 9.84% were provided by all three databases. Our survey showed significant differences amongst the results provided by the databases, especially for “hydroperiod” and “type of environment”. These differences can be the result of different coverage of the scientific literature by the databases, but also of the different ways by which the criterion “relevance” is calculated by the three algorithms. We, therefore, recommend that literature surveys must be based on several databases; otherwise, the results might become biased.