Performance, preference, and visual scan patterns on a menu-based system: Implications for interface design

Abstract

This study was conducted to provide evidence for the nature of visual search processes with menus, and to derive design principles for menu-based natural language (MBNL) interfaces to databases. The effects of window size, window activity, and query length were investigated. It was found that longer queries were performed faster with single active windows, but multiple active windows were rated as more 'natural'. Query times increased with query length, as did eye fixation frequencies, fixation durations, and dwell times. Errors were most likely to occur on the longest query. Fixation durations also varied with window size. However, visual behavior depended on the area being viewed and on the interaction between window activity and query length. In contrast with previous studies, it was also found that menus were not scanned randomly. However, scanpaths were less deterministic with multiple active windows and became even more unconstrained as query length increased. User interface design recommendations were derived from the findings. © 1989 ACM.