Suchergebnisse

Frames describe the way issues are presented, i.e., what aspects are made salient when communicating about these issues. Field of application/theoretical foundation: The concept of frames is directly based on the theory of "Framing". However, many studies using automated content analysis are lacking a clear theoretical definition of what constitutes a frame. As an exception, Walter and Ophir (2019) use automated content analysis to explore issue and strategy frames as defined by Cappella and Jamieson (1997). Vu and Lynn (2020) refer to Entman's (1991) understanding of frames. The datasets referred to in the table are described in the following paragraph: Van der Meer et al. (2010) use a dataset consisting of Dutch newspaper articles (1991-2015, N = 9,443) and LDA topic modeling in combination with k-means clustering to identify frames. Walter and Ophir (2019) use three different datasets and a combination of topic modeling, network analysis and community detection algorithms to analyze frames. Their datasets consist of political newspaper articles and wire service coverage (N = 8,337), newspaper articles on foreign nations (2010-2015, N = 18,216) and health-related newspaper coverage (2009-2016, N = 5,005). Lastly, Vu and Lynn (2020) analyze newspaper coverage of the Rohingya crisis (2017-2018, N = 747) concerning frames. References/combination with other methods of data collection: While most approaches only rely on automated data collection and analyses, some also combine automated and manual coding. For example, a recent study by Vu and Lynn (2020) proposes to combine semantic networks and manual coding to identify frames. Table 1. Measurement of "Frames" using automated content analysis. Author(s) Sample Procedure Formal validity check with manual coding as benchmark* Code Vu & Lynn (2020) Newspaper articles Semantic networks; manual coding Reported Not available van der Meer et al. (2019) Newspaper articles LDA topic modeling; k-means clustering Not reported Not available Walter & Ophir (2019) (a) ...

Sentiment/tone describes the way issues or specific actors are described in coverage. Many analyses differentiate between negative, neutral/balanced or positive sentiment/tone as broader categories, but analyses might also measure expressions of incivility, fear, or happiness, for example, as more granular types of sentiment/tone. Analyses can detect sentiment/tone in full texts (e.g., general sentiment in financial news) or concerning specific issues (e.g., specific sentiment towards the stock market in financial news or a specific actor). The datasets referred to in the table are described in the following paragraph: Puschmann (2019) uses four data sets to demonstrate how sentiment/tone may be analyzed by the computer. Using Sherlock Holmes stories (18th century, N = 12), tweets (2016, N = 18,826), Swiss newspaper articles (2007-2012, N = 21,280), and debate transcripts (2013-2017, N = 205,584), he illustrates how dictionaries may be applied for such a task. Rauh (2019) uses three data sets to validate his organic German language dictionary for sentiment/tone. His data consists of sentences from German parliament speeches (1991-2013, N = 1,500), German-language quasi-sentences from German, Austrian and Swiss party manifestos (1998-2013, N = 14,008) and newspaper, journal and news wire articles (2011-2012, N = 4,038). Silge and Robinson (2020) use six Jane Austen novels to demonstrate how dictionaries may be used for sentiment analysis. Van Atteveldt and Welbers (2020) use state of the Union speeches (1789-2017, N = 58) for the same purpose. The same authors (van Atteveldt & Welbers, 2019) show based on a dataset of N = 2,000 movie reviews how supervised machine learning might also do the trick. In their Quanteda tutorials, Watanabe and Müller (2019) demonstrate the use of dictionaries and supervised machine learning for sentiment analysis on UK newspaper articles (2012-2016, N = 6,000) as well as the same set of movie reviews (n = 2,000). Lastly, Wiedemann and Niekler (2017) use state of the Union speeches ...

Topics describe the main issue discussed in an article, for example: Does an article deal with politics, economics or sports? Field of application/theoretical foundation: In the context of "Agenda Setting", studies analyze which issues are on the public agenda. In the context of "News Values", studies may analyze why some topics are more prominently covered than others. References/combination with other methods of data collection: Many studies combine manual inspection of topics with their automated detection. Quinn et al. (2010) demonstrate for their analyses of legislative speeches how manual inspection may increase the validity of results. Similarly, Hase et al. (2020) use automated content analysis to find and map similar topics for which manual coding is then conducted. Such combinations may contribute to a better and more detailed understanding of topics than automated analyses by themselves. The datasets referred to in the table are described in the following paragraph: Puschmann (2019a) uses New York Times articles (1996-2006, N = 30,862) as well as articles from Die Zeit (2011-2016, N = 377) to identify topics using supervised machine learning. In another tutorial, Puschmann (2019b) uses Sherlock Holmes stories (18th century, N = 12), articles from Die Zeit (2011-2016, N = 377) and debate transcripts (1970-2017, N = 7,897) to apply LDA and structural topic modeling. In her tutorials, Silge (2018a, 2018b) also uses Sherlock Holmes stories (18th century, N = 12) and a news corpus also containing comments (2006-ongoing, N = 100,000). Silge and Robinson (2020) apply LDA topic modeling on news stories by the Associated Press (1992, N = 2,246) as well as books by Dickens, Wells, Verne and Austen (18th century, N = 4). Roberts et al. (2019) use blogposts (2008, N = 13,248) for structural topic modeling. Watanabe and Müller (2019) apply LDA topic modeling on newspaper articles from The Guardian (2016, N = 6,000). Van Atteveldt and Welbers (2019, 2020) use State of the Union speeches (1981-2017, N = 10 and ...

"Key issue" describes the main issue or perspective an article focuses on when reporting on a news topic. There might be different key issues for the same topic: When reporting on terrorism, articles can for example concentrate on the incident itself, the perpetrator behind it, victims and/or political reactions to terrorism. Field of application/theoretical foundation: Key issues share similarities with other variables such as news "frames", "issue salience" or "issue ownership" that also try to identify different perspectives for the same or different news topics. Therefore, studies based on "Framing" (Entman, 1993) work with similar variables to analyze what issues journalists focus on and many studies cited here use the concept of framing to identify key issues, for example Li (2007) or Zhang & Hellmüller (2016). References/combination with other methods of data collection: Studies for example combine content analysis and interviews with journalists to shed more light on dynamics and structures of terrorism coverage, including key issues (Larsen, 2019). Example studies: Li (2007); Matthews (2016) Information on Li, 2007 Authors: Li (2007) Research question: How did television outlets frame 9/11 during the first 24 hours of coverage and how did this framing change over time? Object of analysis: News coverage by five TV outlets (ABC, CBS, NBC, CNN, and FOX news) Time frame of analysis: 24 hours after attacks on September 9 2001 occurred Info about variables Variable name/definition: Coverage frame: "The coverage frame is defined as the aspects of a perceived reality identified through a story that makes these aspects more salient in the news coverage" (Li, 2007, p. 676). Level of analysis: News story (TV) Variables and values: Political coverage frame, economic coverage frame, criminal coverage frame, environment coverage frame, safety coverage frame, human interest coverage frame, religion coverage frame, disaster coverage frame, other coverage frame Reliability: Scott's pi: .8 Information on Matthews, ...

Labeling of groups and events describes how groups connected to religious, political or other forms of violence as well as their acts are labeled or evaluated. These labels might vary from more nominal descriptions (e.g., "gunmen") to more judgmental descriptions (e.g., "terrorist"), leading to different perceptions of these groups and acts by the public. Field of application/theoretical foundation: Labels for groups and events are of interest in journalism research, political communication, research on terrorism and violence as well as stereotyping. These measurements are often based on "Social Identity Theory" (Brown, 2000) as a theoretical foundation for why some groups and events connected to violence are described in a negative way – i.e., as an out-group –, whilst others are described in a neutral way or even positively, i.e., as an in-group. References/combination with other methods of data collection: A study by Huff and Kertzer (2017) for example combines a conjoint experiment with an "Automated Content Analysis" of media coverage to understand how the public would label different acts of violence in comparison to the media. Two studies that have been particularly influential in studying the labeling of violent acts and perpetrators will be discussed in more detail in the following sections. Example studies: Nagar (2010); Weimann (1985) Information on Nagar, 2010 Author: Nagar Research question: How did American news media cover politically violent organizations that are not linked to Al Qaeda or the events of 9/11? Object of analysis: News coverage by two American newspapers (The New York Times, The Washington Post) Time frame of analysis: 1998–2004 Info about variables Variable name/definition: Media frame: "First, the labels that describe political violence were coded separately for each segment. Second, the article frame was determined based on the most frequent label." (Nagar, 2010, p. 537) Level of analysis: Headline, lead paragraph, text Variables and values: four different label categories for ...

Sources describe the actors quoted by journalists to support or refute their argumentation or to introduce new aspects into a discussion. Sources might be used for direct or indirect quotes and can be attributed to a variety of actors, such as government officials, witnesses or PR sources. In terrorism coverage, the media tends to mostly rely on official sources such as the government or police officials. Field of application/theoretical foundation: Content analyses focus on journalistic sources beyond terrorism coverage. Such analyses are often based on "Agenda-Setting" theories (McCombs & Shaw, 1972), models conceptualizing the relationship between journalists and PR, power hierarchies, or studies on working routines of journalists. References/combination with other methods of data collection: Similar analyses in the context of "Automated Content Analysis" try to grasp news "Actors", of which news sources might be one, automatically (for example Burggraaf & Trilling, 2020). In addition, interviews with journalists can shed light on their sourcing routines (Larsen, 2018). Two studies are of particular value when analyzing sources used in terrorism coverage since they analyze a large variety of different sources and will hence be discussed in the following section. Example studies: Larsen (2019); Venger (2019) Information on Larsen, 2019 Author: Larsen (2019) Research question: How are radicalization and violent extremism framed in the news, including the sources used in these articles? Object of analysis: Online news from four Norwegian news outlets (Aftenposten, NRK, TV2, and VG) Time frame of analysis: 2014–2015 Info about variables Variable name/definition: Sources Level of analysis: News stories Variables and values: 27 different values, namely (1) no source, (2) police, (3) security services, (4) national politician in position, (5) national politician in opposition, (6) local politician, (7) bureaucracy/administration, (8) lawyer, (9) military/intelligence, (10) expert/researcher, (11) ...

Abstract
Possibilities of applying automated content analysis in journalism studies include, for example, machine learning to identify topics in journalistic coverage or measuring news diffusion via automated approaches. But how have computational methods been applied thus far? And what are consequences of the "computational turn" in communication science, especially concerning interdisciplinarity? Based on a systematic literature review, this article summarizes the use of automated content analysis in journalism studies. Results illustrate an increasing use of the method by communication scientists, as yet another indicator of methodological interdisciplinarity in communication science. However, there is little evidence of an increase in theoretical interdisciplinarity: Studies relying on computational methods do not increasingly refer to theories from other disciplines. With respect to practical interdisciplinarity, for instance collaborations, our discipline is by no means becoming more interdisciplinary. Instead, we find a shift in favor of technical disciplines. At least up to now, the "computational turn" in communication science should thus not be equated with an "interdisciplinary turn."

Die automatisierte Inhaltsanalyse wird auch in der Journalismusforschung zunehmend genutzt, um Texte (teil)-automatisiert zu analysieren. Sie hat damit zu einer Methodeninnovation im Fach beigetragen, die jedoch selten kritisch diskutiert wird. Der Beitrag gibt einen Überblick über Chancen und Grenzen der automatisierten Inhaltsanalyse in Forschung und Lehre. Er argumentiert für die Vereinbarkeit der Methode mit klassischen Journalismustheorien, weist aber auch auf Grenzen, etwa fehlende Analysetiefe oder mangelnde Validität, hin. Auch für die Vermittlung von Computational Methods werden Probleme offengelegt und Lösungsansätze auf organisatorischer, technischer und didaktischer Ebene diskutiert. Insgesamt bedeutet die automatisierte Inhaltsanalyse für die Journalismusforschung keine Neujustierung, sondern eine Erweiterung - sofern Grenzen der Methode beachtet werden. Anders sieht es bei der Lehre aus, wo eine verstärkte Vermittlung von Computational Methods gefordert wird.