An Empirical Foundation for
Automated Web Interface Evaluation,
Melody Yvette Ivory,
PhD Dissertation, UC Berkeley
Computer Science Division, 2001.
This dissertation presents the first research in which a large collection of Web sites rated by Internet experts was analyzed to develop statistical models to support automated evaluation of new sites. It makes several key contributions to the advancement of automated Web interface evaluation, including the following.
- It summarizes the current state of automated evaluation methods for graphical and Web interfaces and proposes ways to expand existing support.
- It describes how methods used for evaluating the performance of computer systems can be applied towards the problem of automated interface evaluation.
- It presents a methodology and tools for automated evaluation of Web interfaces that is a synthesis of approaches used in the usability and performance evaluation domains and based on empirical data.
- It describes an extensive set of quantitative Web interface measures for assessing many aspects discussed in the Web design literature.
- It describes the development of highly-accurate statistical models for assessing Web interface quality.
- It documents the efficacy of the statistical models for improving Web interface quality and for validating established Web design guidelines.
HTML, PDF (76 KB)
PDF (5 MB)
Individual Chapters (PDF)