Evaluation of an informational Web site: Three variants of the think-aloud method compared

To evaluate Web sites, usability experts often use methods that were originally employed for the evaluation of software applications. In doing so, they assume that these methods will work exactly the same for both types of test objects. However, there is a major difference between transactional software applications and informational Web sites, a difference that could have an effect on the workings of various usability methods. As such, we felt that it was valuable to repeat one of our previous studies in which we compared concurrent think-aloud protocols, retrospective think-aloud protocols, and constructive interaction to evaluate a Web application, this time using a Web site. The results of our study showed that in some respects, the methods did work differently depending on the test object they were applied to. However, we conclude that the three methods are largely interchangeable and that the decision to choose one variant of the think-aloud method over the other should be based on practical considerations.