On-line versus postal data collection: a re-examination of the response differences using demographically matched data

Past research has identified differences between online and mail collected responses to the same survey, but differences in the demographics of respondents had also been noted making the cause of the variation unclear. In the research reported here, responses to the same questionnaire, delivered via mail and internet surveys, were demographically matched across a range of variables. This removed the impact of response differences caused by age, gender, type of product consumed and length of customer relationship. Across all the different question types and response scales, significant differences were still found between mail and online respondents, even when data were ipsatised. Notably, online respondents were far less likely to use the end-points of the scale, perhaps indicating issues with the online collection methodology. The conclusion is that the two methods of data collection can not be assumed to be directly inter-changeable, and that the method used can lead to different results if not managed carefully.