Speed and Accuracy Using Four Boolean Query Systems

Tenth Midwest Artificial intelligence and Cognitive Science Conference

Volume

Tenth Midwest Artificial intelligence and Cognitive Science Conference

Contents

Existing research has shown that naive users find it difficult to correctly express queries using the standard Boolean connectives AND, OR, and NOT. We conducted an empirical study examining the query production performance of naive users on four different systems: One of these systems was non-procedural textual (Boolean connectives); another was procedural textual (If-Then-Else), another non-procedural graphical (based on Venn diagrams), and the last was procedural graphical (Filter-Flow). We found significant differences in both the accuracy and speed of users’ query production based on the underlying Boolean forms of the queries they were required to generate. User performance in accuracy and speed was highest for the system using Boolean connectives, followed by Venn, Filter-Flow, and then If-Then-Else. No overall speed/accuracy tradeoff was observed. However, interactions were discovered amongst the factors of Boolean form, procedurality, and text vs. graphics. These results suggest lines of further work that can be used to improve Boolean interfaces to Web search engines. 6.3 User Interfaces