Using online databases to form subject collections for informetric analyses

Abstract

The online databases of the Dialog System retrieve only 26% of documents in an exhaustively compiled collection on the subject of Bradford's Law of Scattering, with some documents being retrieved from many databases. However, when the Exhaustive Collection is more stringently defined to include only those documents more about the subject, the retrieval rate of Dialog improves to 61%, while its most productive database, LISA, alone retrieves 37%. Both of these ‘samples’ give good estimates of the size-invariant properties of the Exhaustive Collection which are typically studied in Bradford and Growth Analyses—vindicating this use of online searching. However, without additional information, online searches are of little use in determining size-related properties of subject literature collections. Whether the analysis reported here—which relies on identical interpretations of a ‘subject’—has secure foundations is briefly considered.