Read Between the Lines - How Linguistics Can Unearth Parties’ True Voting Advice Application Positions

Voting Advice Applications (VAAs) offer a valuable tool for voters to shortcut the process of information gathering in a complex political landscape. However, parties have incentives to strategically employ neutral stances in VAAs while implicitly conveying policy positions in their detailed answers, thereby invalidating the final voting advice. This German case study focuses on what we call ‘hedging’ in VAAs, where parties hide their true policy preferences behind neutral self-classification. Using data from 20 German municipal VAAs from 2024, we analyse party responses to VAA statements, focusing on those categorised as ‘neutral’. We argue that three linguistic features help to predict and detect hedging: the use of neutral words, temporal adverbs, and the subjunctive II. Our findings show that the prevalence of these factors significantly affects the likelihood of a party’s ‘neutral’ statement containing an underlying policy position. Additionally, we provide evidence that parties generally claim false neutrality at a high frequency and are significantly more likely to do so when VAA statements present policy trade-offs. The findings indicate limitations of VAAs as a reliable source of party policy positions and highlight that neutral party responses should not be taken at face value by voters or scholars.