A carregar...
Publicação
Using Candidates’ Tweets to Predict an election outcome. The United States 2022 Midterm elections study
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 842.75 KB | Adobe PDF |
Resumo(s)
Understanding social media’s role in political communication is crucial in the evolving media landscape.
Motivated by the transformative impact of social media on political engagement and discourse, this
research fills an under-explored academic gap, studying the effects of geographic focus—local versus
national—in candidates’ tweets on U.S. Senate election outcomes. It reveals a modest but significant
correlation between the nature of political discourse and election competitiveness. Interestingly, strict
adherence to party-centric topics did not significantly influence electoral success. The study assessed
the performance of regression and classification models in forecasting election outcomes, with
classification models demonstrating superior results. Both models provide a new benchmark for future
studies in political communication on social media. These findings bear considerable implications for
political practitioners, indicating that election success is not merely guaranteed by echoing party centric issues or predominantly adopting a national communication scope.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Data Driven Marketing, specialization in Marketing Intelligence
This thesis has originated the article «Using Candidates’ Tweets to Predict an Election Outcome» in Political Research Quarterly: https://doi.org/10.1177/10659129241286827
This thesis has originated the article «Using Candidates’ Tweets to Predict an Election Outcome» in Political Research Quarterly: https://doi.org/10.1177/10659129241286827
Palavras-chave
Machine Learning Predictive Modelling Topic Analysis Political Communication Social Media Political Marketing Twitter