| Popis: |
Objectives: To assess the quality of data downloaded from a variety of bibliographic databases commonly searched in systematic reviews, in order to provide guidance on the order in which database records should be uploaded into a citation manager and which record should be used as the “primary record” during the duplicate removal process. Methods: We downloaded the bibliographic records of a random sampling of two hundred journal articles from the following databases: MEDLINE (via PubMed and Ovid), Embase, Web of Science, Scopus, Cochrane, CINAHL (via Ebsco), International Pharmaceutical Abstracts (via Ebsco), Sociological Abstracts (via Proquest), and Google Scholar. We required that each article be available in at least three databases and that each database contain at least fifty of these articles. We then assessed the completeness and correctness of the downloads using the full-text article as our gold standard. Records were scored on presence, completeness, and correctness of the following fields: title, authors, digital object identifier (DOI), volume, issue, pages, publication year, accession number, abstract, and URL. Using these data, we calculated an overall score for each database and a head-to-head score for each database combination using the article overlap for each database pair. |
