Cost-Justifying Usability, 2nd Edition

to publicly observable events and can be captured on videotape. Surely metrics

such as these offer the best hopes for establishing the scientific credibility of

usability evaluation? Unfortunately, these metrics have not been empirically

researched in terms of their metric properties as thoroughly as the Satisfaction

metrics have been. Not only do we not know how reliable such performance

metrics are, but we also have no information about their population parameters

(although they do have a lot of what is known as face or apparent validity). As

we shall see later, this has serious consequences on the sample sizes needed to

demonstrate an effect. Satisfaction ...