Faceted search interfaces for metadata-rich datasets such as product information have been around for a while. e-Bay and Amazon are two obvious examples. Faceted search for textual data is only slowly making its way into the commercial realm (see NewsSift, for example) but have been receiving increasing attention in research. Villa et al. presented an interesting paper at SIGIR09 in which they compared different interface layouts for handling aspects, and compared the effectiveness of aspectual search with a conventional interface for different tasks.
Two aspectual interfaces were constructed: a parallel view that showed aspects in columns side by side, and a tabbed view that showed some metadata side-by-side, but made the user select a facet to show the detailed search results. These were compared with a baseline non-aspectual interface that resembled conventional search. It wasn’t clear from the paper whether both aspectual interfaces were presented to users, or if one was discarded during pilot testing. The paper does allude to problems with the parallel view due to lack of space to display search results. Perhaps one way of dealing with this issue is to assign users to each aspectual interface in a balanced way, and allow them to dynamically select one interface or another. Measuring how often and under what conditions users switched views might inform some post-interview questions.
The results of the experiment showed that for difficult exploratory search tasks where users had to identify multiple aspects of an information need (rather than searching on aspects identified for them up front), aspectual interfaces provided some advantage. Users tended to use a broader vocabulary in their queries, to run more queries, view more results, and mark more documents relevant, and they engaged with the problem for the entire length of the session, rather than quitting early as some did in the baseline condition. It would have been interesting to see from their log data how saved documents were distributed over time in the session by interface condition: did people using aspectual interfaces continue to find useful information, or were they just clicking around more?
Overall, it is encouraging to see more exploration of this space in recent work presented here and at JCDL (see, for example, “What do exploratory searchers look at in a faceted search interface?“). Perhaps these findings will complement text-mining efforts such as Google Squared to provide users with useful alternatives to the single ranked list for exploring complex topics.
For other discussion of this work, see the blog post by David Karger who was actually at the conference rather than following it through Twitter!