Web-based surveys present market researchers with a number of advantages in terms of speed and cost compared with other modes of research. However, online questionnaires often suffer from low response rates and technical issues. This prompted Dillman, Totora and Bowker (1998) to put forward a series of respondent-friendly design principles aimed at improving web questionnaire response rates and data quality. While these principles appear useful, empirical evidence validating them is sparse. This paper reports the results of a research project that tested a subset of Dillman et al's (1998) principles relating to: the structure of the first question; the use of graphical symbols conveying point of completion; and the use of double banking for multiple response questions. The results do not lend strong support to the principles tested, and further research is therefore required to clarify whether their adoption can significantly improve response to online questionnaires.