Human Factors and Ergonomics Society
September 27 - October 1, 2010
San Francisco, CA
Identifying Aggregate Scanning Strategies to Improve Usability Evaluations
Presenters: Joe Goldberg, Ph.D., CPE
Automated analysis methods are needed to convert eye tracking data into meaningful descriptions of high-level scanning strategies that can help usability professionals understand how designs are viewed. Prior techniques for analyzing eye tracking sequences, or scanpaths, are limited in their ability to find aggregated visual search strategies for a group of observers. We describe a pattern analysis tool that finds clusters of sequentially matching patterns among multiple scanpaths. Scanning strategies are clustered hierarchically, then represented as aggregate scanpaths. The tool can also match scanpaths against a hypothetical scanning strategy, input by mouse gesture. Use of the pattern analysis tool was demonstrated using a 114-participant eye tracking dataset in which several aggregate scanning strategies were identified. Matching against a hypothetical, counter-clockwise scanning strategy was also included in the demonstration. The analysis tool provides a valuable resource for aggregating or grouping scanning strategies, a vital step toward generalizing eye tracking results to guide design recommendations for improving usability.
Date: Thursday, September 30, 2010
Time: 8:30 - 10:00 AM