Between the Lines: Analyzing the Effect of Publication Type on Sentiment in Political Scandal Coverage
Authors: Aida Bustos, Annabelle Hill, Daniel Lovett, & Liz O Neill
Date: 8 May 2023
The manner in which the media covers political scandals can significantly affect citizens' perceptions surrounding such controversies, helping to frame their beliefs and attitudes. As a result, numerous contemporary studies have explored how specific media characteristics impact both coverage and public digestion of such reporting. Leveraging texts consisting of newspaper articles from the United Kingdom from 2019 to 2023, we look to contribute to this scholarship, investigating whether we can find a link between publication type (broadsheet or tabloid) and article sentiment in media coverage of political scandals. Our results indicate that such a relationship can be substantiated with a high degree of confidence, as we observed that tabloid reporting on scandals was consistently more negative than similar coverage in broadsheets.
Despite substantial transformations in the way citizens across the world consume media coverage of politics as a result of the proliferation of social media, blogs, television, and podcasts, newspaper coverage of events still plays a prominent role in framing and shaping public attitudes and beliefs towards political developments. Within such coverage by newspaper publications, scandals involving notable elected officials have become increasingly prominent story topics.
To respond to the above phenomena, the academic community has dedicated a significant amount of attention to exploring the drivers of variation in newspaper coverage of political and economic topics, including numerous studies analyzing differences in article sentiment, as well as the outcomes of such differences in terms of electoral consequences and public trust in the government. In conjunction, numerous studies have also investigated various components of political scandals to shed light on the dynamics underpinning the determinants and effects of significant scandals in the news cycle, both historically and contemporarily.
While there has been substantial progress made in understanding the mechanisms at work within each of these areas of interest individually, a substantial scholastic gap remains in examining the overlap between media coverage and political scandals specifically. As a result, in our research we set out to inspect one aspect of this relationship, evaluating whether a statistically significant linkage can be established between the publication type (tabloid vs broadsheet) of articles covering political scandals, and their sentiment, controlling for the left-right ideological position of the newspaper.
To do so, we used the LexisNexis database to create corpora consisting of newspaper articles related to general political coverage over the last three years to establish a comparative baseline and articles related to two high-profile UK cases - the “Partygate” and “Prorogation” scandals. We chose these scandals because of their comparably succinct time frame and their high degree of notoriety. We additionally leveraged the Lexicoder and Harvard IV dictionaries of words connoting either positive or negative sentiment towards a subject to determine overall article sentiment.
With these corpora and their accompanying sentiment scores, we employed numerous quantitative text analysis methods to first explore and describe various characteristics of the data. Among these were identifying the most frequent words, examining the keyness (uniqueness) of dictionary terms by publication type, and calculating numerous summary statistics such as mean, median, mode, range, interquartile range, and standard deviation by both publication type and left-right ideological position for each corpus. Figure 1 helps to visualize the findings from this exploratory analysis, demonstrating the differences in mean sentiment scores for each dictionary between the two publication types. More specifically, it helps to highlight the consistently more negative sentiment score observed in articles from tabloid newspapers compared to those of broadsheets, as well as the greater magnitude of the difference in coverage of each political scandal when compared to the variation for general political coverage.
Following this, we finally used multivariate linear regression modeling to evaluate our central research question and determine whether a statistically significant independent relationship could be observed between publication type and article sentiment for newspaper coverage of political scandals, holding ideological leaning constant. The findings generated from this examination indicate that we can discern such a relationship with a high degree of confidence. Figure 2 further illustrates these observations, displaying the sentiment scores predicted by the regression modeling for each dictionary and coverage area, sorted by publication type. As can be observed, this figure reiterates our central finding that tabloid reporting of political scandals is independently associated with more negative sentiment than coverage in broadsheet publications. Furthermore, the substantive size of this effect is greater in the texts related to each of the scandals than in general political news articles, highlighting the fact that this relationship can be viewed as unique to scandals specifically.
While these findings help uncover the factors involved in the media’s coverage of political scandals, they are far from providing a complete answer. Extending this work into the future, potential avenues for additional research and evaluation of this central question could include making use of different forms of media, domestic contexts, or time frames in which to test these results, as well as employing new methodologies such as leveraging different operational techniques and controlling for additional explanatory variables in the analysis. Such research would thus help to further shed light on an area relatively underinvestigated but of growing importance and relevance in the increasingly volatile and hyper-polarized arena of political coverage in the media.
About the authors: Aida Bustos, Annabelle Hill, Daniel Lovett, & Liz O Neill are current MSc Politics and Data students at University College Dublin. The work presented in this blog post was conducted under the supervision of Stefan Muller, an assistant professor in the School of Politics and International Relations, and was submitted for the “Connected_Politics” module.