Ali Bahari; Homayoun Sadeghi-Bazargani; Fahimeh Bakhtiari
Abstract
Introduction: Use of databases for accessing electronic documents has increased the necessity of reviewing the performance and coverage of these databases. The aim of this study was to assess the recall and precision of Persian databases for retrieval of scholarly information in the field of occupational ...
Read More
Introduction: Use of databases for accessing electronic documents has increased the necessity of reviewing the performance and coverage of these databases. The aim of this study was to assess the recall and precision of Persian databases for retrieval of scholarly information in the field of occupational injuries.Methods: This descriptive study was carried out using observation and comparison approach. The study population included national databases of SafeLir, Scientific Information Database (SID), Magiran, and Irandoc, and the sample size equaled the number of documents retrieved from these databases in the last two decades. Keywords were determined based on the opinions of the members of the research team and review of the relevant documents. Finally, forty key words were chosen for searching the sites were searched, and the number of recovered documents was recorded. Indices and graphs related to recall and precision were then prepared.Results: The number of recovered documents from Magiran, SID, SafeLir, and Irandoc were 411, 153, 523, and 149, respectively. SafeLir database had a higher relative recall (83.7%) and SID had a higher precision (89.54%) in retrieving relevant documents compared to other databases. The greatest overlap in results was observed between Magiran and SafeLir (269 documents).Conclusion: The findings of this study will help researchers in the field of injuries and related fields to identify the appropriate database (with an appropriate recall and precision) for retrieval of documents in their specialized fields. If recall of the search or precision of the search or both of them are important. It is recommended to use SafeLir, SID, and SafeLir databases, respectively, in cases where precision, recall and both are important.
