Computational models of the perceptual, cognitive, and motor processes involved in the visual search of pull -down menus and computer screens.

Abstract

A major challenge in making software easy for people to use is to design screen layouts that people can search efficiently. This dissertation demonstrates that for this very important practical problem in the context of human-computer interaction, a detailed computational model of the mental and physical processes embodied in the human user provides detailed theoretical, empirical, and practical insights with which to understand the user and make fundamental improvements to software and hardware design. Specifically, these insights are relevant to addressing basic issues associated with graphical user interface design that relate to computational cognitive modeling, visual search, menu search, and manual motor control. This dissertation presents the first empirically validated computational models of the perceptual, cognitive, and motor processes involved in the visual search of pull-down menus on a computer, models that can be generalized to explain the processes involved in more general computer screen layout visual search tasks. The cognitive models are built using Kieras and Meyer's EPIC cognitive architecture. Major findings of the models include: For randomly-ordered menus, people consider more than one menu item at a time and adopt search strategies that are somewhat random and somewhat systematic. For numerically-ordered menus, people develop considerable knowledge of the locations of the menu items, but their ability to make predictive aimed movements to those known locations degrades for items further down in the menu.