Keywords

1 Introduction

Terrorism research has seen an increase in scholarly debate and empirical studies since the 1970s, with an abrupt rise after the September 11, 2001 attacks (Jones 2007; Kocks et al. 2011; Miller and Mills 2009). Terrorism studies is an interdisciplinary research field, with most studies stemming from disciplines such as political studies, psychology, and sociology and only a few from journalism or media studies (Jones 2007; Silke 2004). The relationship between journalism and terrorism has been described as symbiotic (Elter 2008; Miller 1982; Surette et al. 2009; Wilkinson 1987), meaning that the two are mutually dependent. Terrorists rely on the media to make their acts and atrocities public and reach a wider audience; journalists cannot refuse to report about a terrorist act because of their function as news providers. Moreover, terrorism is a newsworthy topic, so coverage will lead to higher economic revenue. However, some researchers have described the relationship as parasitic rather than symbiotic (Schaffert 1992; Schultz 2017) because terrorists “force” the media to report. This can be regarded as an act of exploitation in which the parasite (the terrorists) (mis)uses the functions of the media that are obliged to report important events to the public.

Common research questions in content analyses include a focus on which terrorist attacks make it into the headlines (Sui et al. 2017; Weimann and Brosius 1991), coverage, labeling, and framing of violent groups (Dunn et al. 2005; Hellmueller et al. 2022); Zhang and Hellmüller 2016), and—to a lesser extent—coverage of individual terrorists (Knight 2007). Studies examining news values, for instance, find that attacks in Western countries with a high number of fatalities and attacks by Muslim perpetrators are considered more newsworthy by the Western media (Hase 2021; Kearns et al. 2019). Studies also analyze radicalization narratives in the media (Auer et al. 2019; Nacos 2005) and radicalizing narratives in terrorist propaganda (Ingram 2017; Mahood and Rane 2017) with the latter strand of research illustrating that terrorists focus on identity building via in- and outgroup dichotomy, crisis reinforcement, and depiction of violence and military success (Colas 2017; Lorenzo-Dus et al. 2018; Rothenberger et al. 2018). Research also includes audience reactions to attacks on social media (Fischer-Preßler et al. 2019). Single case studies prevail, but we find comparative angles across these lines of research (e.g., through comparisons by group, ideology, or motivation of the attack) (Decker and Rainey 1982) or by media from various countries (Peralta Ruiz 2000). Content analyses across media channels (e.g., TV and print coverage), however, are not common.

2 Common Research Designs and Combinations of Methods

Frequent study designs include quantitative and qualitative content analyses of legacy and non-legacy media and of political statements and speeches, with the formal research object being texts and, to a lesser extent, images or videos (Gerstenfeld 2003; Beuthner et al. 2003). We also find surveys and experiments on the understanding and effects of terrorism or terrorism coverage from the recipients’ perspective (Huff and Kertzer 2018; Shoshani and Slone 2008), while surveys concerning the communicators (be it journalists, politicians, or terrorists) are missing. Guided interviews reconstructing working routines have sometimes been used (Konow-Lund and Olsson 2016), but observations of journalists’ working routines in times of terror are largely missing.

The main method of scientific data collection and analysis used in research on terrorism and media is content analysis, mostly of articles in Western print media. A review of the state of research revealed a strong focus on content analyses of print and television coverage of specific terrorist acts (event centered) and few analyses of background information during attacks (context centered) (Rothenberger 2022). Discourse analyses and longitudinal comparative studies are less frequent. Traditional manual content analyses are the most common, but computational methods, such as automated content analysis or network analysis, are increasingly used, such as to uncover journalistic sourcing (Rauchfleisch et al. 2017) or automatically analyze topics of social media discussions on terrorism (Fischer-Preßler et al. 2019). Some studies also combine media data with extra-media data such as databases recording terrorist attacks and their characteristics (Hase 2021; Hellmueller et al. 2022). Knelangen (2009) examined the pros and cons of qualitative versus quantitative terrorism research. He concluded that it is ultimately the research question that decides which method or combination of methods is appropriate.

Concerning the research object, there is a lack of content analyses of radio programs and of ethnographic studies of terrorists’ daily lives in the respective areas, including observations of their media use (Ross 2007). Although Gerstenfeld’s (2003) content analysis of 157 extremist websites concluded that the internet is a particularly powerful and effective instrument for extremists to reach an international audience, recruit members, network with other groups, and enable a high level of image control, content analyses of traditional mass media are still most common.

If we follow the order determined by the course of communication from the attack to the audience, we have studies dealing with terrorists’ utterances on the internet (An et al. 2018), media coverage (with word choice playing an important role during the journalistic production phase) (Badr 2017; Glück 2007), and the effects on recipients, including emotional reception (Haußecker 2013).

One always has to take into account that normative prescriptions of how to cover terrorism may change over time according to editorial guidelines (e.g., insofar as a TV channel speaks of “martyrs” or not or whether perpetrators are named and visualized). Beermann (2004), Cohen-Almagor (2005) and Schmid (1992) notice many differences in editorial policies and coherence of using the label “terrorism”. According to Beermann (2004), the Reuters news agency, for example, does not refer to specific events as “terrorism” since the 1960s.

3 Main Constructs

While studies deal with different research objects (e.g., terrorist attacks, terrorist groups, single terrorists, or radicalization) and use different methods of analysis, spanning from computational to discourse analysis, six variables are measured repeatedly:

  1. 1.

    Key issue: Many studies analyze the key issue in an article (i.e., what aspect is emphasized when it comes to terrorism). Studies thereby differentiate between descriptions of a terrorist attack and its course, the severity and consequences of an attack, political reactions in the form of official statements by politicians, military reactions in the form of the “war against terror,” and how to counter terrorism as key issues (Guzek 2019; Mogensen et al. 2002; Zhang and Hellmüller 2016).

  2. 2.

    Sources: Many studies analyze frequent sources used in terrorism coverage. They find that government officials/politicians are predominantly sourced, but so are other media outlets/news agencies, security/police, witnesses/ordinary people, or terrorists themselves (Matthews 2013; Venger 2019; Zhang and Hellmüller 2016). As social media allows for quick updates, the use of Twitter, for example, normalized over time (Bennett 2016).

  3. 3.

    Labeling of groups and events: Content analyses also examine labels of violent events and groups (i.e., what attacks and perpetrators are called in news coverage). Labeling is an important aspect of coverage, as different labels for groups (e.g., “terrorists” vs. “freedom fighters”) and incidents (e.g., “terrorist attack” vs. “crime”) attach dissimilar normative associations and motives (Bhatia 2005). Military action against a perpetrator, for example, might be more easily supported if that person is explicitly labeled a “terrorist” (Baele et al. 2019), as the term implies “that a given action is illegitimate” (Steuter 1990, p. 261) and thus legitimizes military reactions. This might also be why news media generally tend to be cautious about attaching that label (Nagar 2010).

    Studies operationalizing such labeling differentiate between labels for events (e.g., “attack,” “explosion,” “seizure”) and groups (e.g., “rebel,” “terrorist,” “murderer”) and find that these vary with incident and perpetrator characteristics and by who is providing the description (Elmasry and el-Nawawy 2020; Lavie-Dinur et al. 2018; Hase 2021; Simmons and Lowry 1990). With the rise of religious terrorism, many studies find that outlets have created new labels connecting Islam and terrorism (e.g., “radical Islam” or “Islamic terrorists”) (Mahony 2010).

  4. 4.

    Radicalization narratives: News media try to make sense of the radicalization of especially young people by constructing narratives around their turn toward terrorism and attributing certain reasons for that choice. Common media explanations proposed for the radicalization of (mostly male) adolescents are religious fanaticism, being the victim of brainwashing by extremists, or naively joining in search for an adventure (Berbers et al. 2016). Interestingly, studies find these radicalization narratives to be gendered: when the media describes suspected female terrorists, it concentrates more on their physical appearance, sexuality, supposed naivety, and troubling social life (Conway and McInerney 2012; Martini 2018; Nacos 2005).

  5. 5.

    Emotionalization: Emotionalization is another aspect of coverage and can be operationalized as a the inclusion of persons who express or are ascribed emotions in text or images (Gerhards et al. 2011). Unsurprisingly, studies find around half the coverage to be emotionalized (Wolf 2010) and that the dominant emotions are mostly implicit and negative, such as sorrow, fear, and devastation (Gerhards et al. 2011).

  6. 6.

    Visualization: In addition, research shows that visualization (operationalized, for example, as the number of visuals accompanying textual coverage or what is shown on these) plays an important role for memorial purposes, personification, and calls for solidarity in the wake of an attack. Studies on visualization thereby differentiate between the visuals shown (e.g., if buildings, memorial gatherings, victims, perpetrators, or flags are displayed in news or social media content) (Berkowitz 2017; Beuthner et al. 2003; Kim 2012). Interestingly, violence—an inherent part of terrorism—is more likely to be illustrated in the form of destroyed buildings than mutilations/wounded people. Legacy media mostly show dead people as covered up (Jirschitzka et al. 2010; Linder 2011)—in contrast to propaganda content, where violent images often are part of the terrorists’ message (Winkler et al. 2019).

4 Research Desiderata

This overview clarified that the operationalization of terrorism-related variables remains a challenge. Terrorism as a label that signifies an asymmetric relation always includes a component of “power” that manifests in media content. As only few international studies outside of the US and “the West” exist, it is usually the US/Western perspective on terrorism that we find in empirical studies (this is to some extent also due to language barriers and thus the international visibility of research). Further, analyses often deal with spectacular, singular attacks in Western countries instead of including a broader view (e.g., by including longitudinal coverage of certain groups or radicalization narratives). The selection of studies in this article accordingly reflects this bias.

Content analyses of web content face the difficulty of co** with a plethora of texts and visuals, from politicians, journalists, citizen journalists, extremists, terrorists, etc. Computer-assisted methods of data collection (e.g., scra**) and data analysis (e.g., automated content analyses) are required to deal with such “big data” but are—as of yet—rarely applied. Additionally, it is not only the commonly known web that is of interest to terrorism researchers: “terrorists’/extremists’ internet usage is still under-researched because of the lack of systematic Dark Web content collection and analysis methodologies” (Quin et al., 2006, p. 4). Another problem of content analyses in the field of terrorism studies is that many miss a close link to a coherent theoretical framework guiding the research. Beck and Quandt (2011) complain about many of the content analyses after the September 11 attacks that were published under great pressure of topicality and that leave it unclear to what extent journalism and communication studies can provide a theoretical understanding of the phenomenon.

Relevant Variables in DOCA—Database of Variables for Content Analysis

Key issue: https://doi.org/10.34778/2u

Sources: https://doi.org/10.34778/2w

Labeling of groups and events: https://doi.org/10.34778/2v