Score reliability in Web- or Internet-based surveys: Unnumbered graphic rating scales versus likert-type scales

Unnumbered graphic rating scales theoretically can lead to greater score va riance and thus greater score reliability. The use of Web-based graphic sli ders overcomes the practical problems of administering graphic scales on pa per measures: the excessive costs of scoring and excessive scoring errors a rising from the tediousness of scoring done by hand. The study investigated whether graphic scales realize theoretically expected improved score relia bility and how coarseness in slider scoring affects reliability. A large sa mple size (n = 4,407) and diverse respondent groups were employed so that m ore confidence could be invested in results.