Title page for ETD etd-04192004-155229 ( Browse

Type of Document

Dissertation

Author

Zeng, Xiaoming

Author's Email Address

xizst9@pitt.edu

URN

etd-04192004-155229

Title

Evaluation and Enhancement of Web Content Accessibility for Persons with Disabilities

Degree

Doctor of Philosophy

Program

Rehabilitation Science

School

School of Health and Rehabilitation Sciences

Advisory Committee

Advisor Name	Title
Bambang Parmanto	Committee Chair
Miriam F. Hertz	Committee Member
Ronald E. LaPorte	Committee Member
Valerie J.M. Watzlaf	Committee Member

Keywords

Web content accessibility
Persons with disabilties
usability
usability testing
blind
website design
disability study
web accessibility measurement
human computer interaction

Date of Defense

2004-04-16

Availability

unrestricted

Abstract

Web content accessibility is to which degree a site is accessible to the largest possible range of people, especially for persons with disabilities. The current study comprises four interrelated parts revolving around evaluating and enhancing Web content accessibility.
First, a novel measurement metrics called Web Accessibility Barriers (WAB) score is constructed based on published Web accessibility design guidelines. The performance of the measurement metrics is assessed using a Web accessibility gold standard. The Receiver Operating Characteristic (ROC) curve shows that WAB score separate inaccessible Web pages well from accessible ones with Area Under the Curve (AUC) value of 0.917.
Second, using the WAB score, I evaluate the degree of Web accessibility of consumer health information websites. Mean of WAB score of the 108 selected websites is 9.31 with standard deviation of 6.29. Among those websites, government and education websites have lower WAB score – 0.39 and 1.16 respectively. ANalysis Of VAriance (ANOVA) test shows statistical significances across different functional categories of the websites (F=9.705, p < 0.001). In addition, this study shows that WAB score correlated with traffic rank of the websites with (r = 0.32, p < 0.01).
Third, a usability study is conducted to examine the performance of a Web Transcoder Gateway (WTG) server for blind Web users to access online information. The WTG server removes Web accessibility barriers in real time based on Web accessibility design guidelines. The study design is a within-subject cross-over design. A mixed model analysis is employed to examine the effect of the WTG server on time, success, errors, and subjective measurement. Sixteen blind Web users took part in the study. Results show that participants spent less time, make few errors, and succeed more on the tasks via WTG server. Participants also feel more satisfactory, less frustrated, and more confident when access online information via WTG server. Observational and anecdotal findings imply that only removing accessibility barriers may not be sufficient to achieve the best usability for blind Web users.
Finally, a usability study is conducted to examine the performance of a Web Transcoder Gateway (WTG) server for PDA users to access Web sites. The study design is also a within-subject cross-over design. A mixed model analysis is employed to examine the effect of WTG server on time, success, and subjective measurements. Twenty subjects took part in the study. Results show that participants spend less time on each task via WTG server. Each participant can finish the tasks successfully via either WTG server or non WTG server. This implies that the WTG server can serve as “Electronic Curb Cut” for Web users under constraints. However, the participants show no statistically significant preference to WTG server.

Files

Filename		Size	Approximate Download Time (Hours:Minutes:Seconds)
Filename		Size	28.8 Modem	56K Modem	ISDN (64 Kb)	ISDN (128 Kb)	Higher-speed Access
	XiaomingZeng_April2004.pdf	3.83 Mb	00:17:42	00:09:06	00:07:58	00:03:59	00:00:20