OR/16/047 Key findings

Nkisi-Orji, I. 2016. Semantic information retrieval for geoscience resources: results and analysis of an online questionnaire of current web search experiences. British Geological Survey Internal Report, OR/16/047.

In this section, (Section 1: literature or data gathering searches) are referred to as type 1 search while searches that ask specific questions (Section 2: searches that ask a specific question) are referred to as type 2. The key findings from responses collected in this survey are as follows.

• Popular search engines (e.g. Google) and publication citations (e.g. Science Direct) were reported as most useful applications for search. While publication citations were slightly favoured (59%) for type 1 search, popular search engines were found most useful (79%) for type 2 search.
• There was a higher tendency for respondents to assess more search results in type 1 searches than in type 2 searches. While about 88% will assess more than 10 results in type 1, only 50% will do same in type 2. However, in both search types, most respondents will not assess more than 20 search results.
• Eighty two percent (82%) usually or sometimes performed multiple searches or constructed advanced search queries in an attempt to include narrower or equivalent terms to original search intent.
• Ninety five percent (95%) think that a search feature that expands search terms with narrower or equivalent terms will be beneficial. Ninety percent (90%) of this proportion will prefer control over when to use such feature.
• All respondents reported that their search results were at some point, dominated by irrelevant result entries. An irrelevant result can be a document which contains supplied search terms but in non-intended sense. This was not a rare occurrence for 77% of respondents.
• About 95% think that a feature which disambiguates search terms will be beneficial. Eighty one percent (81%) will like to be able to specify intended context/meaning of search terms but only when such terms are ambiguous.