The field of visual design effects on respondent’s behavior in online surveys is well researched; however in several cases the outcomes of different studies have provided contradictory findings. In this thesis, the focus will be on experiments dealing mainly with

(VAS) and ratings scales. A VAS is an instrument that tries to measure a characteristic or attitude that is believed to range across a continuum of values and is verbally anchored at eac end (e.g. VAS is relatively rare, in part because of operational difficulties (Couper et al. (2006)). Hence a detailed view on technical possibilities and pitfalls should be given.

Three main studies with the same experimental design (ensuring that occurring effects were reproducible) were carried out, whereby 6 different types of such scales were presented to the interviewees in order to measure the effect of varying appearance and functionality of the controls used for implementing the scales. To run these experiments, software was developed that focused on a good support system for Web survey experimenting. The results refer to the general fill out behavior, completion time, dropout, reliability and usage of extreme points.