The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	Investigating graphical user interface usability on task sequence and display structure dependencies
Author:	Sengupta, Tirthankar
View Online:	njit-etd2004-125 (xix, 224 pages ~ 17.8 MB pdf)
Department:	Department of Industrial and Manufacturing Engineering
Degree:	Doctor of Philosophy
Program:	Industrial Engineering
Document Type:	Dissertation
Advisory Committee:	Jeng, One-Jang (Committee chair) Bladikas, Athanassios K. (Committee member) Das, Sanchoy K. (Committee member) Abdou, George Hanna (Committee member) Sodhi, R. S. (Committee member) Blanchard, Harry Edward (Committee member)
Date:	2004-08
Keywords:	Usability Graphical user interface Activity theory Eye movements Systemic-structural approach Display
Availability:	Unrestricted
Abstract:	Designing Graphical User Interfaces (GUI) requires the consideration of task sequence requirements (sequence of operations arising from task structures and application constraints) and display structure (layout of the elements of the interface) relationships. The basic purpose was to understand the usability differences of the interfaces through efficiency, motor performance, and search performance. Thirty-two subjects performed experiments in four groups. The experiments differed in display structure and compatibility of task sequences. Subject mouse actions, mouse coordinates and eye positions were recorded. The derived measures, click efficiency, mouse traversal and eye visits to different areas of interest (namely the tool, object, and goal), were analyzed in a repeated measures factorial design with compatibility and display structure as the between subjects factors and phase of learning as the within subject factor. A significant interaction between compatibility and phase of learning (p<.01) was observed. Mouse traversal per unit time increased significantly (p<. 05) across phases of learning. The phase of learning affected the number of eye visits for all groups. Compatibility had a significant ((p<.005) effect on the average processing time during search. The results establish that the compatibility of task sequence requirement with the display structure affecting the performance of subjects and hence the usability of the interface was thus obtained. However, through learning, subject performance showed considerable improvement and the effects of task sequence and display structure diminished at final stages of user learning. Based on this evidence, a systemic structural activity approach was used to develop a model of human performance on the eye movement and mouse action data. This structural model of human performance is defined as an algorithm and can be used for estimating complexity of task performance. In this study only the assumptions for development of the model and the formulation of the model are explained as an application of the results of the study. The study hence served a dual purpose in the long run: understanding the compatibility of the task sequence with the interface display structure as well as establishing eye and mouse movements as a viable tool to study task performance at human computer interfaces.