Widget manipulation revisited: a case study in modeling interactions between experimental conditions

Abstract

Widgets are often used to perform control tasks in three-dimensional (3D) virtual environments (VEs). Spatial interactions through widgets require precise 3D manipulations, and several design aspects of VEs contribute to the ease, accuracy, and speed with which users can perform these interactions. Throughout the years, VE researchers have studied relevant design aspects; for example, the location and size of the widgets, monoscopic versus stereoscopic viewing, the presence or absence of co-location, or the inclusion of (passive) tactile feedback, are all design aspects that have been studied. However, researchers have mostly studied design aspects in isolation and have paid little attention to possible interactions between conditions. In this paper, we introduce a method for modeling interaction effects between experimental conditions and illus- trate it using data from a specific case study, i.e., widget manipulation tasks. More specifically, we model how the effect of passive tactile feedback interacts with stereoscopic viewing for three widget manipulation tasks. We also model how these effects vary between two tasks, i.e., button and menu item selection. Models that include inter- action effects between experimental conditions can be used to get a deeper understanding in the system design trade-offs of a virtual environment.