Using Voting Advice Application Data in Empirical Research: Practical, Ethical, and Methodological Issues

The massive popularity of Voting Advice Applications (VAAs) has led to a proliferation of VAA user response data. Despite this, the use of VAA data in scientific publications has never been put under critical scrutiny. This paper looks at 57 peer-reviewed articles published between 2007 and 2019, which analyzed more than 65 datasets from 40 different countries, accounting for more than 14 million respondents. The paper discusses practical issues regarding preprocessing and the accuracy of the reported numbers, ethical issues on informed consent, anonymity, data sharing and replication, and methodological issues on the self-selection effects arising from listwise deletion and the reliance on opt-in and follow-up questionnaires. The paper concludes with recommendations for VAA developers, journal editors and ethics committees, as well as third party users of VAA data.