A small‐scale test of Metals abstracts index, with ten profiles and five sections of the abstracts as data base, was carried out with a methodology which endeavoured to simulate real life use of this Index by an information scientist for current awareness searching. The performance results were 67 per cent recall, 40 per cent precision and 2 per cent fallout. An error analysis showed that 65 per cent of the errors were attibutable to human (searcher) factors. Some of the factors that must inevitably affect the assessment of printed indexes are discussed, together with the observation of two unexpected phenomena, namely an apparent relationship between precision and searching effort, and relationships between generality and both recall and fallout.
CitationDownload as .RIS
MCB UP Ltd
Copyright © 1973, MCB UP Limited