Comparing Internet and phone survey mode effects across countries and research contexts*

Abstract

We compare internet and telephone survey responses across 27 European Union nations and two research contexts, one a choice experiment of willingness to pay to avoid power outages, and the other a social acceptance of energy infrastructure question. The various forms of survey mode effects and the challenges of survey mode choice are documented and developed in the context of statistical theory and an application to an economics survey. We find evidence that survey-mode effects vary across nations and research contexts. We interpret our findings to suggest that the degree of measurement bias can vary strongly between research contexts, for example based on the availability of a perceived socially correct response within a given context. Future survey-based research should evaluate the choice of survey mode in a context and region-specific manner.