Examining the pseudo-standard web search engine results page

Nearly every web search engine presents its results in an identical format: a ranked list of web page summaries. Each summary comprises a title; some sentence fragments usually containing words used in the query; and URL information about the page. In this study we present data from our pilot experiments with eye tracking equipment to examine how users interact with this standard list of results as presented by the Australian sensis.com.au web search service. In particular, we observe: different behaviours for navigational and informational queries; that users generally scan the list top to bottom; and that eyes rarely wander from the left of the page. We also attempt to correlate the number of bold words (query words) in a summary with the amount of time spent reading the summary. Unfortunately there is no substantial correlation, and so studies relying heavily on this assumption in the literature should be treated with caution.