An acceptability predictor for websites

User acceptance is a high priority for website design and implementation. Two significant, but largely separate, approaches to acceptability are: First, the Web Accessibility Initiative (WAI) has explored the measurement of technical features of a website to gauge its accessibility. Second, human judgments about acceptability are obtained from intended users or experts. The present work explores the important question of how best to combine these two methods. Experiment One required new users to explore automatic website evaluation systems. They found two of four systems difficult or impossible to use and system outputs difficult to understand. Experiment Two combines formal properties and user judgments, using an automatic system to predict user judgments from formal website properties. A simple system was able to predict user judgments within 91% accuracy. Clearly, user judgments about websites can be predicted reliably, a result of value to designers.

Proceedings of the 5th International Conference on Universal Access in Human-Computer Interaction. Part III: Applications and Services. San Diego, CA, July 19 - 24, 2009. Section: Universal Access to the World Wide Web.