Adaptive Questionnaires for Voting Advice Applications in Practice: A Study on User Experience, Trust and Recommendation Accuracy

Since the rise of Machine Learning techniques, data collection processes have become more and more effective. Also, for Voting Advice Applications (VAA), adaptive questionnaires with active learning components have been proposed to speed up the recommendation process. However, it remains unclear whether such methods will also be helpful in practice. This study implements an adaptive questionnaire for VAA and quantifies its benefits regarding user experience and self-reported propensity to vote through a series of user experiments.