When left to themselves without an experimenter present, users actually tried out fewer system functions than in the previous evaluation. There were twent-four sessions altogether --- five users had two sessions with the system. Seventeen participants used the titles list to choose which documents to look at. Eleven chose to do a new search at some point in the session; no one initiated an expanded search. Five participants added new terms at some point and three deleted terms from their query. Only two selected thesaurus terms from the suggested words list and no one navigated the thesaurus.
Users' lack of exploration of the thesaurus and expanded search functions was disappointing to the experimenters. Various possible reasons can be put forward:
One way to investigate why users were not taking advantage of the expansion facilities was to look at usage of the ``New Search'' facility If it was being used to repeat a previous query, it would indicate that users were dissatisfied with their existing results but did not understand how to exploit the system to find more books. However, out of the eleven users who carried out a new search in one session, only one repeated his original query exactly, although two other users almost did. One typed in ``strategic cost management'' the first time and then tried ``strategic_cost_management'' because he thought this would give him an adjacency search --- a facility which was not available in the system. The other typed ``relationship marketing'' as a first query and then entered the complete title ``Genuine Interpersonal Service and Relationship Marketing'' in an attempt to retrieve details of a specific book.