Selection biases in Voting Advice Application research

In: Electoral Studies, Band 36, S. 272-280

Abstract

This paper focuses on methodological aspects of Voting Advice Application (VAA) research. Self-selection into the treatment and self-selection into the sample need to be controlled for if the aim is to deduce causal effects from VAA use in observational data. This paper outlines omnipresent endogeneity issues, partly imposed through unobserved factors that affect both whether individuals use VAAs and their subsequent electoral behavior. Some research on VAAs has acknowledged the problems associated with selection biases, but practically none have taken steps towards improving these shortcomings. This paper introduces various approaches for handling selection biases and applies them to VAA data from Switzerland. These methodological considerations are of outmost importance for the relevance and validity of VAA research. [Copyright Elsevier Ltd.]