I Need a CAVAA: How Conversational Agent Voting Advice Applications (CAVAAs) Affect Users' Political Knowledge and Tool Experience.

In election times, millions of voters consult Voting Advice Applications (VAAs) to learn more about political parties and their standpoints. While VAAs have been shown to enhance political knowledge and increase electoral turnout, research also demonstrates that voters frequently experience comprehension problems when responding to the political attitude statements in a VAA. We describe two studies in which we test a new type of VAA, called Conversational Agent VAA (CAVAA), in which users can easily access relevant information about the political issues in the VAA statements by asking questions to a chatbot. Study 1 reports about an online experiment (N = 229) with a 2 (Type: traditional VAA/CAVAA) x 2 (Political sophistication: low/high) design. Results show that CAVAA users report higher perceived political knowledge scores and also answer more factual knowledge questions correctly than users of a regular VAA. Also, participants' CAVAA experience was evaluated better. In Study 2 (N = 180), we compared three CAVAA designs (a structured design with buttons, a non-structured design with an open text field, and a semi-structured design with both buttons and an open text field), again for higher and lower politically sophisticated users. While the three designs score equally high on factual and perceived knowledge indicators, the experience of the structured CAVAA was evaluated more positively than the non-structured version. To explore the possible cause for these results, we conducted an additional qualitative content analysis on 90 chatbot-conversations (30 per chatbot version). This analysis shows that users more frequently access additional information in a structured design than in a non-structured design, whereas the number of break-offs is the same. This suggests that the structured design delivers the best experience, because it provides the best trigger to ask questions to the chatbot.

The affect of negativity: testing the Foreign Language Effect in three types of valence framing and a moral dilemma.

In decision-making people react differently to positive wordings than to negatives, which may be caused by negativity bias: a difference in emotional force of these wordings. Because emotions are assumed to be activated more strongly in one's mother tongue, we predict a Foreign Language Effect, being that such framing effects are larger in a native language than in a foreign one. In two experimental studies (N = 475 and N = 503) we tested this prediction for balanced and unbalanced second language users of Spanish and English and for three types of valence framing effects. In Study 1 we observed risky-choice framing effects and attribute framing effects, but these were always equally large for native and foreign-language speakers. In our second study, we added a footbridge dilemma to the framing materials. Only for this task we did observe a Foreign Language Effect, indicating more utilitarian choices when the dilemma is presented in L2. Hence, across two studies, we find no Foreign Language Effect for three types of valence framing but we do find evidence for such an effect in a moral decision task. We discuss several alternative explanations for these results.

A good tennis player does not lose matches. The effects of valence congruency in processing stance-argument pairs.

According to the principle of Argumentative Orientation (AO), speakers and writers adjust their frame choice to the opinion they want to convey and hearers and readers are sensitive to this profile choice. In three reaction time studies (N = 68; N = 97; N = 60) we investigated whether, in line with AO, stance-argument pairs congruent in valence are easier to process and to verify than incongruent pairs. Second, we tested whether, in line with predictions from the Markedness Principle (MP), positive congruent pairs are easier to process than negative congruent pairs. In line with AO, participants made faster and more accurate judgments of congruent pairs than of incongruent pairs. This effect was observed when controlling for word length and word frequency, and occurred irrespective of the distance between the evaluative word in the stance and argument. No unambiguous effect of Markedness was found.

Issue framing in online voting advice applications: The effect of left-wing and right-wing headers on reported attitudes.

Voting Advice Applications (VAAs) provide voting recommendations to millions of people. As these voting recommendations are based on users' answers to attitude questions, the framing of these questions can have far-reaching consequences. The current study reports on a field experiment in which the framing of the header above VAA statements (N = 17) was manipulated (condition 1: no header; condition 2: a right-wing header, e.g., finance; condition 3: a left-wing header, e.g., nature and environment). Visitors of a VAA developed for Utrecht, the fourth largest municipality in the Netherlands, were randomly guided to one of the versions of the tool in which the header type was varied. Results (based on Nrespondents = 27,404) show that providing a header (left-wing or right-wing) leads to more left-wing answers as compared a condition where there is no header above the attitude statement. This effect, however, is only observed for respondents with lower levels of political sophistication.

Positive vs. Negative: The Impact of Question Polarity in Voting Advice Applications.

Online Voting Advice Applications (VAAs) are survey-like instruments that help citizens to shape their political preferences and compare them with those of political parties. Especially in multi-party democracies, their increasing popularity indicates that VAAs play an important role in opinion formation for citizens, as well as in the public debate prior to elections. Hence, the objectivity and transparency of VAAs are crucial. In the design of VAAs, many choices have to be made. Extant research in survey methodology shows that the seemingly arbitrary choice to word questions positively (e.g., 'The city council should allow cars into the city centre') or negatively ('The city council should ban cars from the city centre') systematically affects the answers. This asymmetry in answers is in line with work on negativity bias in other areas of linguistics and psychology. Building on these findings, this study investigated whether question polarity also affects the answers to VAA statements. In a field experiment (N = 31,112) during the Dutch municipal elections we analysed the effects of polarity for 16 out of 30 VAA statements with a large variety of linguistic contrasts. Analyses show a significant effect of question wording for questions containing a wide range of implicit negations (such as 'forbid' vs. 'allow'), as well as for questions with explicit negations (e.g., 'not'). These effects of question polarity are found especially for VAA users with lower levels of political sophistication. As these citizens are an important target group for Voting Advice Applications, this stresses the need for VAA builders to be sensitive to wording choices when designing VAAs. This study is the first to show such consistent wording effects not only for political attitude questions with implicit negations in VAAs, but also for political questions containing explicit negations.