Qualitative usability studies have few users and variable protocol; numbers obtained from such studies are likely to poorly reflect the true behavior of your population due to large measurement errors.| Nielsen Norman Group