Framing the Taxation-Democratization Link: An Automated Content Analysis of Cross-National Newspaper Data

Abstract

Taxpaying constitutes a major opportunity for citizens to relate to their governments. Although it is true that paying taxes is a responsibility, it also entitles citizens to claim control over government spending, which may facilitate a greater democratization of a country’s political regime. Consistent with this reasoning, a growing body of scholarship has documented a positive relationship between the size of tax revenues extracted by the state and the adherence of the country’s regime to democratic values. What has been left underexplored is the role in this relationship of the media, a commonly available and relied-upon source of information about taxpaying for the public. This study offers a first contribution in this direction, by exploring the relationship between the nature of the political regime and the rhetorical construction of the concept of a taxpayer in the national press. Based on an automated content analysis of articles (N=24,969) published by ninety-two newspapers and news agencies in fifty-one countries using a set of pretrained and validated machine-learning algorithms, the study demonstrates that the less democratic a state is, the more likely it is for the national press to frame a taxpayer as a subordinate in a hierarchical relationship with the state, by discussing taxpaying in tax collection, rather than public spending, terms. The study furthers a more nuanced understanding of the place of the media in the taxation-democratization link and demonstrates the applicability of the supervised machine-learning approach to classifying frames in large cross-national samples of newspaper data.

Computer-assisted digital text analysis for journalism and communications research: introducing corpus linguistic techniques that do not require programming

“…Opperhuizen et al, 2019), automated content analysis of news articles (e.g. Kananovich, 2018), sentiment analysis of Facebook posts (e.g. Ji et al, 2019), ‘crowdcoding’ for analysis of political sentiment on Twitter (Guo et al, 2019), or a myriad of other techniques (see Boumans and Trilling, 2016), that digital methods are upon us is undeniable.…”

Who set the narrative? Assessing the influence of Chinese global media on news coverage of COVID-19 in 30 African countries

“…With computational tools, as opposed to manual content analysis, researchers can uncover general patterns, sometimes in large volumes of textual data (De Grove et al, 2020;Lewis et al, 2013), and tackle questions that might have been difficult to address with manual content analysis (Trilling & Jonkman, 2018). The approaches employed in this article, (a) using document similarity algorithms to identify text reuse (Nicholls, 2019;Vogler et al, 2020) and (b) fitting STMs to identify themes in news texts (Barberá et al, 2020;Kananovich, 2018) have been shown to produce reliable data to answer similar questions in the past. These approaches are only applicable to written text, which is the focus of this article, and do not consider non-written content.…”

Anti‐Indigenous policy formation : Settler colonialism and neoliberal political advocacy

“…In the context of emerging work on tax and ‘identity’ (Björklund Larsen 2017 ; Carrillo, 2020; Kananovich, 2018; Kiel, 2019; Sheild Johansson 2018; Simpson, 2008; Stanley, 2016; Walsh, 2018), there has been some work on taxpayer groups generally (Lo, 1990; Martin, 2008; 2013; Willmott, 2017). More broadly, political sociologists have examined the rise of professionalized citizen mobilization firms (Lee et al., 2013; Walker, 2014) that rise and fall at the behest of powerful clientele.…”