Survey Administration Effects?

Abstract

The module “environmental attitudes and values” of the International Social Survey Programme (ISSP) was administered both as a paper-and-pencil questionnaire to a representative sample of the German population and as a web-based survey to an online access panel representative for German Internet users. The two samples differ significantly with regard to sociodemographic and substantive variables. An attempt to weight the data of the web-based sample on the basis of distribution characteristics of several sociodemographic variables resulted in rather inconsistent findings for the comparison of substantive items. However, if only respondents with similar education levels are compared, neither relevant differences in item means nor differences in selected interitem correlations are observable. This means that for high coverage groups (e.g., participants with a high education level), the data gathered via the web are basically identical to those obtained in a traditional self-administered mode.