Do Mail and Web Surveys Provide Same Results?
Katja Lozar Manfreda and Vasja Vehovar
Web surveys are a relatively new data collection mode, used only since the mid 90's. However, they already represent an important part of the contemporary survey industry (http://websm.org).
Despite their broad implementation, the validity of Web survey mode still needs to be confirmed, especially the convergent validity, e.g., the comparison of Web survey results with results of other survey modes. In our case, such validity shows whether different survey modes can be replaced by each other.
An establishment survey of educational institutions within the RIS (Research on Internet in Slovenia, http://www.ris.org) project at the Faculty of Social Sciences, University of Ljubljana, enables the comparison of a Web and a mail survey. Participants were randomly assigned to two experimental groups. The first group received a mail questionnaire (two follow-ups), while the other group was invited to answer a questionnaire on the Web (a mail advance letter with two follow-ups). The advance letters and questionnaires were made as similar as possible. Automated skips were implemented in the Web questionnaire while arrows and text were used for skips in the paper questionnaire. The substantive results, data quality and respondents' satisfaction from both questionnaires were compared. Results show that there are no major differences in substantive responses, however there do exist differences in item non-response with the Web mode having much higher item non-response rates.