Parties’ group appeals across space and time: an effort towards an automated, large-scale analysis of parties’ election manifestos.

Political parties play a central role in democratic systems, representing the interests and preferences of voters and competing for their support. Parties use various forms of communication to persuade voters and promote their agendas, including policy appeals and group appeals. While policy appeals refer to the specific policies that parties support or oppose, group appeals refer to the social groups that are addressed in parties' statements. While policy appeals have been extensively studied over the last decades, group appeals have only recently gained scholarly attention. Existing research has shown that political parties frequently use group appeals in their campaign communication, and that these appeals can have a significant impact on voter attitudes and behavior. However, there is a lack of empirical data to conduct cross-national and longitudinal empirical studies of group appeals. This data gap represents a challenge for our understanding of party behavior, political representation, and the party-voter linkage. In other words, how parties try to shape the political agenda and influence voter preferences by appealing to specific social groups. To address this challenge, we present a method for automating the process of extracting parties’ group appeals from election manifestos. Our method employs validated, multilingual input lexicons in a joint aspect-detection and sentiment analysis in order to identify both the social group that is being appealed to as well as the tone of the appeal (positive or negative). As a test case, we apply our approach to manifestos from major British and German parties since the 1970s. This analysis reveals important trends in party behavior which have implications for party competition, voter behavior, and political representation. Our approach offers several advantages over existing methods for collecting data on group appeals via hand-coding. First, it is scalable and efficient, allowing for the analysis of large volumes of text data without the need for extensive resources or complex automation. Second, it is multilingual, enabling the analysis of group appeals in different languages. Third, it provides a systematic and standardized method for analyzing group appeals, allowing for the comparison across time and space. Overall, our work represents a significant step forward in the study of group appeals and their role in political communication. We believe that our approach can be widely adopted and applied to a variety of contexts and research questions.