Evaluating Visualizations in Voting Advice Applications

While the different visualisations employed by VAAs are the main source of information VAAs provide to their users, their design and function has until now rarely been evaluated. Drawing on the broader information visualization literature, this study will evaluate the three main types of visualisations in VAAs - percentage match bar plots, multi-dimensional radar charts and two-dimensional political maps - in terms of their usability and validity. This will be done in two stages. First, user data regarding the perceived usability of a certain visualization by the user will be matched to user demographics. This will allows us to establish which visualisations users prefer in different circumstances and how these preferences relate to their demographics. Second, experimental data is provided in which users were asked to use the visualizations to carry out certain tasks that are common to VAA use. This data allows us to test whether users interpret VAA visualisations in the way VAA designers expect them to and how they solve certain tasks. The study concludes with several recommendations for VAA designers and several suggestions on how to improve the visualisations current VAAs employ. This not only leads to less misunderstanding, but might also allow VAAs to even better live up to their promise of providing useful voting advice to their users.