Political issues, in general, focus on the content of political actors’ communication and most often describe either the main issue or several issues that are in the focus of a political actor’s statement or any other relevant text (e.g., press release, news article, tweet, etc.). The basic premise of analyzing political issues in the self-presentation of political actors is that one major goal of political actors’ communication is to place specific issues on the political agenda (Strömbäck & Esser, 2017). Political issues are most often coded based on a list of pre-defined issues that refer to different policies and sometimes also to polity or politics. The scope and detail of the individual issues depend on the purpose and the focus of the analysis.
Field of application/theoretical foundation:
Apart from being a common descriptive and control variable, the coding of issues in political actors’ communication can serve as the basis for more complex variables or concepts such as agenda building or issue ownership.
Agenda building, at large, refers to the process of how media content is shaped by societal forces (Lang & Lang, 1981). With regard to analyses of politicians’ self-presentation, most work focuses on the processes of communication by which political actors aim to obtain media coverage for their issues (Norris et al., 1999; Seethaler & Melischek, 2019). Analyses on agenda building usually compare issue agendas between at least two different forms of communication, e.g., between channels where political actors have high control (such as press releases, party manifestos, social media messages) and journalistic outlets where political actors have less control (e.g., Harder et al., 2017; Kiousis et al., 2006; Seethaler & Melischek, 2019).
Content analyses on agenda building usually start by, first, identifying relevant issue fields and categories (inductively or deductively). Second, the dominant political issues in political actors’ communication and/or other forms of communication (e.g., news articles) are coded according to predefined lists. Third, the occurrence of specific issues or issue agendas are compared between the different forms of communication, often over time (see, e.g., Seethaler & Melischek, 2019).
Issue ownership, in broad terms, means that some parties are considered by the public in general as being more adept to deal with, or more attentive to, certain issues (Lachat, 2014; Petrocik, 1996; Walgrave et al., 2015). Traditionally, issue ownership has been analyzed from a demand-side perspective, based on surveys, as the connection between issues and parties in voters’ minds. Definitions of issue ownership usually comprise at least two dimensions: competence issue ownership (parties’ perceived capacity to competently handle or solve a certain issue) and associative issue ownership (the spontaneous link between some parties and some issues) (Walgrave et al., 2015). Content analyses build on these definitions to investigate to what extent political actors focus on issues that they (respectively their parties) own and what factors may explain the (non-)reliance on owned issues (e.g., Dalmus et al., 2017; Peeters et al., 2019). Other content analyses use issue ownership as an independent variable, for example, to explain user reactions to parties’ social media messages (e.g., Staender et al., 2019).
Content analyses on issue ownership usually start by, first, identifying relevant issue fields and categories (inductively or deductively). Second, the dominant political issues in political actors’ communication are coded according to predefined lists. Third, political actors are assigned issue ownership for specific issues based on theoretical considerations, existing literature, and/or survey data. Fourth, an index for owned issues is calculated at the statement or text level based on the coded issues and the predefined ownership for specific issues.
References/combination with other methods of data collection:
Political issues can be analyzed using both manual and automated content analysis (e.g. topic modeling or dictionary approach). Analyses use both inductive or deductive approaches and/or a combination of both to identify issue categories and extend or amend previous lists of political issues.
Example studies:
Dalmus et al. (2019), Peeters et al. (2019); Seethaler & Melischek (2019)
Table 1: Summary of a selection of studies on agenda building and/or issue ownership
Author(s)
Sample
Unit of Analysis
Values
Reliability
Seethaler & Melischek (2019)
Content type: parties’ news releases and tweets, media reports
Country: Austria
Political actors: all parliamentary parties (ÖVP, SPÖ, FPÖ, Grüne, NEOS, Liste Pilz)
Outlets: all party news releases, parties’ and top candidates’ twitter accounts, five legacy media outlets
Sampling period: 6 weeks before the national election day in 2017 (4 September 2017–14 October 2017)
Sample size: 1,009 news releases, 9,088 tweets, 2,422 news stories
Unit of analysis: individual news releases, tweets, and news stories
Level of analysis: issue agendas
Dominant issue: 13 issue areas based on the Comparative Agendas Project: civil rights, government operations, law and crime, international affairs and defence, European integration, macroeconomics, domestic commerce, transportation and technology, environment and agriculture, education, labour, social welfare and housing, health
Cohen’s Kappa between .91 and .95
Harder, Sevenans, & Van Aelst (2017)
Content type: newspaper, television, radio, news website, and Twitter items featuring a political topic, a domestic political actor, or an election-specific term
Country: Belgium
(Political) actors: tweets by 678 professional journalists, 44 accounts affiliated with legacy media organizations, 467 politicians, 19 civil society organizations, 109 “influentials”
Outlets: 5 print newspapers, 3 news websites, 2 daily television newscasts, 6 daily radio newscasts, current affairs tv programs, and election-specific tv shows
Sampling period: Belgian 2014 election campaign (1 May to 24 May 2014)
Sample size: n = 9,935
Unit of analysis: news items and tweets
Level of analysis: news items (n = 5,260) / news stories (n = 414)
Issues (up to three issues per item): list of 28 broad issues based on the Comparative Agendas Project
Categorization of news stories: inductive coding of individual time- and place-specific events based on news items from traditional news outlets. Non-news items and tweets were then assigned to the already-identified news stories
Krippendorff’s alpha = .70
Krippendorff’s alpha = .76 (for assigning news story to tweet)
Dalmus, Hänggli, Bernhard (2019)
Content type: party manifestos, party press releases, and newspaper coverage
Countries: CH, DE, FR, UK
Political actors: parties
Outlets: 1 quality newspaper and 1 tabloid per country, all party press releases and manifestos
Sampling period: election campaigns between 2010 and 2013 (8 weeks prior to the respective election days)
Sample size: 4,191
Unit of analysis: Actor statements on issues concerning national politics and containing either an explicitly mentioned position or interpretation/ elaboration on the issue
Level of analysis: text level
Main issue: Economy, Welfare, Budget, Freedom and Rights, Europe/ Globalization, Education, Immigration, Army, Security, Ecology, Institutional Reforms, Infrastructure, Elections and Events (each of these top-issue categories is made up of several more detailed sub-issues leading to a total of 127 issue options)
Issue emphasis: percentage of statements devoted to a certain issue
Issue ownership: issue fully belongs to one party (1), issue belongs to center-left / center-right parties (0.5), issue is unowned (0) (based on Seeberg, 2016; Tresch et al., 2017, for more details see the paper)
Cohen’s Kappa ?.3 for sub-issues; Cohen’s Kappa ?.5 for top-issues
Peeters, Van Aelst, & Praet (2019)
Content type: politicians’ tweets, online media coverage, and parliamentary documents
Country: Belgium (Flemish part)
Political actors: 144 MPs from the 6 parties represented in the Flemish and federal parliament
Outlets: 13 Flemish news outlets
Sampling period: 1 January to 1 September, 2018
Sample size: n = 51,691 tweets, n = 8,857 articles, n = 12,638 parliamentary documents
Unit of analysis: text level
Level of analysis: issue agendas
Index for issue concentration: Herfindahl index (to assess how diverse/ concentrated the individual issue agendas are across platforms)
Issues: automated coding of 20 issue topics using the Dutch dictionary based on the Comparative Agendas Project
Issue ownership: operationalization based on survey data; relative party ownership scores for each politician were assigned based on the percentage of respondents that linked a certain party with the topic
NA
(A manual check on 200 randomly selected documents shows that a little over 70% of the automated non-codings were in fact non-classifiable documents. For the other 30%, the dictionary was not able to properly classify the documents.)