| Popis: |
Conversational Agent Voting Advice Applications (CAVAAs) are chatbot-based information retrieval systems for citizens who aim to inform themselves about the political issues at stake in times of political elections. Previous studies investigating these relatively young tools primarily focused on the effects of CAVAAs that include a text-based chatbot. In order to further optimize their design, current research compared the effects of CAVAAs with a text, voice, and combined chatbot. In an experimental lab study among young voters (N = 60) these three modalities have been compared on usage measures (the amount of information retrieved from the chatbot, and miscommunication), evaluation measures (ease of use, usefulness, and enjoyment), and political measures (perceived and factual political knowledge). Results show that the three CAVAA modalities score equally high on political measures and the perception of enjoyment. At the same time, the textual and combined CAVAA outperform the voice CAVAA on several aspects: the voice CAVAA received lower ease of use and usefulness scores, respondents requested less additional information, and they experienced more miscommunication when interacting with the voice chatbot. Analyses of the usage data also indicate that in the combined condition users hardly use the voiceoption and instead almost exclusively rely on text-functionalities like clicking on suggestion buttons. This seems to suggest that using voice is too much of an effort for CAVAA users; we therefore recommend the usage of text-bots in this specific usage context. |
