Identifying Political Topics in Social Media Messages

Jackson, Sam; Zhang, Feifei; Boichak, Olga; Bryant, Lauren; Li, Yingya; Hemsley, Jeff; Stromer‐Galley, Jennifer; Semaan, Bryan; McCracken, Nancy

doi:10.1145/3097286.3097298

Cited by 10 publications

(7 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance of the multilabel model is in Table 1. Candidates' messages were identified by topic using the lexicon-based approach developed by Jackson et al (2017). The lexicon was created based on three datasets: 1)…”

Section: Automated Classification Of Campaign Messagesmentioning

confidence: 99%

See 1 more Smart Citation

A wall of incivility? Public discourse and immigration in the 2016 U.S. Primaries

Rossini¹,

Sturm-Wikerson²,

Johnson

2020

Journal of Information Technology & Politics

View full text Add to dashboard Cite

This study uses computational methods to investigate public incivility in Facebook comments to campaign messages during the primaries of the 2016 U.S. Presidential Election, focusing on candidates' posts about immigration. Specifically, we investigate the relationship between campaign messaging strategies, such as advocacy and attacks, and the presence of incivility in public comments. We find that Trump dominated both in terms of number of posts and of public conversation and was the least likely to receive uncivil comments. Attack messages by the candidates were more likely to receive uncivil comments, suggesting that campaign negativity may indeed affect how the public engages with candidates online. These findings are important because uncivil discourse may contribute to further polarize the public around the issues, which may in turn lead to more extreme views.

show abstract

Section: Automated Classification Of Campaign Messagesmentioning

confidence: 99%

“…For immigration, the lexicon included 289 words. The performance of the lexicon for the topic of immigration had an F1 score of .92, with .86 precision and .98 recall (Jackson et al, 2017).…”

Section: Automated Classification Of Campaign Messagesmentioning

confidence: 99%

A wall of incivility? Public discourse and immigration in the 2016 U.S. Primaries

Rossini¹,

Sturm-Wikerson²,

Johnson

2020

Journal of Information Technology & Politics

View full text Add to dashboard Cite

show abstract

“…In fact, some of the results could be invalidated if the recall differs for different policy categories as we show it happens in the context of political ads. Finally, Jackson et al [28] proposed to use a lexicon-based approach to built a list of language cues for nine political topics to deal with the lack of training data. The authors evaluated the method over 500 labeled texts and they achieved an accuracy of over 85% for eight out of nine categories.…”

Section: Related Workmentioning

confidence: 99%

On Detecting Policy-Related Political Ads: An Exploratory Analysis of Meta Ads in 2022 French Election

Sosnovik,

Kessi,

Coavoux

et al. 2023

Preprint

View full text Add to dashboard Cite

To curb such risks, both online platforms and regulators (through the DSA act proposed by the European Commission) have agreed that researchers, journalists, and civil society need to be able to scrutinize the political ads running on large online platforms. Consequently, online platforms such as Meta and Google have implemented Ad Libraries that contain information about all political ads running on their platforms. This is the first step on a long path. Due to the volume of available data, it is impossible to go through these ads manually, and we now need automated methods and tools to assist in the scrutiny of political ads.In this paper, we focus on political ads that are related to policy. Understanding which policies politicians or organizations promote and to whom is essential in determining dishonest representations. This paper proposes automated methods based on pre-trained models to classify ads in 14 main policy groups identified by the Comparative Agenda Project (CAP). We discuss several inherent challenges that arise. Finally, we analyze policy-related ads featured on Meta platforms during the 2022 French presidential elections period.

show abstract

“…Recently, semi-automated lexicon-based approaches have been applied regularly by researchers. However, the topicality and coverage of such lexicons and thus the accuracy of this approach can vary greatly [20,29,30,33]. To overcome this problem, researchers have devised strategies to improve the approach.…”

Section: Issue Identification In Social Media Postsmentioning

confidence: 99%

Using Social Media Data to Analyse Issue Engagement During the 2017 German Federal Election

Meier

Bazo

Elsweiler

2021

ACM Trans. Internet Technol.

View full text Add to dashboard Cite

A fundamental tenet of democracy is that political parties present policy alternatives, such that the public can participate in the decision-making process. Parties, however, strategically control public discussion by emphasising topics that they believe will highlight their strengths in voters’ minds. Political strategy has been studied for decades, mostly by manually annotating and analysing party statements, press coverage, or TV ads. Here we build on recent work in the areas of computational social science and eDemocracy, which studied these concepts computationally with social media. We operationalize issue engagement and related political science theories to measure and quantify politicians’ communication behavior using more than 366k Tweets posted by over 1,000 prominent German politicians in the 2017 election year. To this end, we first identify issues in posted Tweets by utilising a hashtag-based approach well known in the literature. This method allows several prominent issues featuring in the political debate on Twitter that year to be identified. We show that different political parties engage to a larger or lesser extent with these issues. The findings reveal differing social media strategies by parties located at different sides of the political left-right scale, in terms of which issues they engage with, how confrontational they are and how their strategies evolve in the lead-up to the election. Whereas previous work has analysed the general public’s use of Twitter or politicians’ communication in terms of cross-party polarisation, this is the first study of political science theories, relating to issue engagement, using politicians’ social media data.

show abstract

Identifying Political Topics in Social Media Messages

Cited by 10 publications

References 7 publications

A wall of incivility? Public discourse and immigration in the 2016 U.S. Primaries

A wall of incivility? Public discourse and immigration in the 2016 U.S. Primaries

On Detecting Policy-Related Political Ads: An Exploratory Analysis of Meta Ads in 2022 French Election

Using Social Media Data to Analyse Issue Engagement During the 2017 German Federal Election

Contact Info

Product

Resources

About