Evaluate effectiveness of web design using information-search tasks

As web pages become a mainstream mode of communication there is growing interest in evaluating their effectiveness. A common goal of web designers is to effectively represent information on a web site so that viewers can easily access information for which they are seeking. Eye-tracking technology provides a useful tool to help with this and other goals. Three straightforward examples of how the eye-tracker can be used in evaluating web pages is to look at

1) response time to reaching a targeted link (RT-to-target),
2) response time to select some information (RT-to-select)
3) accuracy in selection

RT-to-target measure informs us of the effectiveness of the (low-level) features in capturing a person's attention. An effective feature could be one that is salient on a simple dimension (such as having a unique color) relative to surrounding information. This salient feature is likely to 'grab' our attention and thus 'pull' our eyes rapidly to this information.

RT-to-select and accuracy of selecting targeted information informs us as to the effectiveness of the (higher-level) semantic content of the depicted information. If the link effectively conveys the meaning of the targeted information RT-to-select will be rapid. The more ambiguous the meaning, the slower the RT will be. Accuracy will be critical factor in confirming the RT trends. It is quite conceivable to have web information where the two types of RT information have an inverse relation. 'Flashy' but uninformative web information will trigger a fast initial eye-movements (RT-to-target). but RT-to-select the correct target is likely to be slow. Conversely, a 'dull but informative' web page might take the viewer a long initial time to find the appropriate link but could be quick in making the appropriate selection.

Two demonstration studies
Search for web information on.

------------------------------------------------------------------------------------------------------------------------------------

Different goals in web search determine what eye movement behaviors to study:
Determining which aspects of eye-movement (i.e., dependent variables) to study depends on the goals of those creating the page ( presumably determined by sponsoring organization) as well as those searching for information. When focusing on marketing-- the organizational goal is effective advertising to maximize product 'hits' and resulting sales . One such marketing concern in education settings, involves student recruitment and filling classrooms (virtual or otherwise), But presuming the mandate of 'education' as the priority. the goal now is optimizing student learning and student success. What makes web design optimal for 'selling a product' vs. 'student learning' can be dramatically different. More superficial information heaped with 'bells & whistles' can be very effective in marketing. In this case, optimal web design would strive for the fastest performance on RT-to-target and RT-to-select variables with little regard to accuracy. But if the goal of optimal design is on learning than semantic content becomes important and maximizing accuracy is top priority, RT-to-target is lowest priority, As for the RT-to-select , what would be considered to be optimal for student learning may depend on educational philosopy. Instructors interested in links that are easy-to-process and extract their meaning may strive for rapid RT-to-select performance but those instructors who believe that challenge and struggle are required for effective learning may argue for slower RT-to-select performance. Perhaps the solution is to have initial tests on experts who should achieve rapid RT-to-select performance to demonstrate that the links do effectively convey intended meaning for those with demonstrated expertise and understanding. In subsequent testing of these same pages on student novices, RT-to-select may now serve as a measure of their initial understanding and with repeated testing may serve as a measure of their progress in learning.The research literature likely will offer useful insights here but key message: goals matter when deciding which variables to study.

------------------------------------------------------------------------------------------------------------------------------------

Additional info on web design

D.J. Aks
8/29/06

 

|D.J. Aks | Eye-tracking research | System noise | Time series & fractal analysis | Visual search | Attention | Satellite Imagery | Tumor detection | Web eval & info search | Illusions