ABSTRACT

This study investigates the political topics that triggered user engagement on candidates’ Facebook pages during the 2018 Hungarian general election campaign on a full dataset of politicians’ communication by using automated content analysis approaches. We have limited knowledge on what topics spread well on Facebook. This study aims to fill this gap by investigating the engagement-triggering effects of prominent topics of the 2018 Hungarian election campaign. It hypothesizes that salient issues (immigration and corruption) and campaign-related posts will obtain more user reaction, while policy topics and mobilization-related content will be less engaging. The effects of issue ownership on user engagement are also tested. For this, the study draws upon an extensive dataset that includes all Facebook posts of each candidate (511 candidates 38030 posts) running for MP positions, posted between the official starting day of the campaign and Election Day. Topics are identified by dictionary-based and supervised learning methods that allow us to categorize a larger set of data than earlier works. The results show that corruption, development policy, and campaign are the most popular topics with users. The most interesting finding, however, is related to the reversed issue ownership effect on user engagement.

摘要

本研究通过使用自动内容分析方法，在政客交流的完整数据集上调查了在 2018 年匈牙利大选期间引发用户参与候选人 Facebook 页面的政治话题。我们对 Facebook 上哪些话题传播得很好的知识有限。本研究旨在通过调查 2018 年匈牙利竞选活动突出主题的参与触发效应来填补这一空白。它假设突出问题（移民和腐败）和与竞选相关的帖子将获得更多用户反应，而政策主题和动员相关内容的吸引力将降低。还测试了问题所有权对用户参与度的影响。为了这，该研究利用了一个广泛的数据集，其中包括竞选 MP 职位的每位候选人的所有 Facebook 帖子（511 名候选人 38030 个帖子），在竞选正式开始之日和选举日之间发布。主题由基于字典的监督学习方法识别，这使我们能够对比早期作品更大的数据集进行分类。结果表明，腐败、发展政策和运动是最受用户欢迎的话题。然而，最有趣的发现与反向问题所有权对用户参与度的影响有关。结果表明，腐败、发展政策和运动是最受用户欢迎的话题。然而，最有趣的发现与反向问题所有权对用户参与度的影响有关。结果表明，腐败、发展政策和运动是最受用户欢迎的话题。然而，最有趣的发现与反向问题所有权对用户参与度的影响有关。