The effects of design parameters of rating scales on the perceived quality of interaction with web sites were investigated, using four scales (Disorientation, Perceived ease of use, Perceived usefulness and Flow). Overall, the scales exhibited good psychometric properties. In Experiment 1, psychometric results generally converged between two response formats (visual analogue scale and Likert scale). However, in Experiment 2, presentation of one questionnaire item per page was better than all items presented on a single page and direct interaction (using radio buttons) was better than indirect interaction (using a drop-down box). Practical implications and a framework for measurement are presented.