The language of violence: distinguishing terrorist from nonterrorist groups by thematic content analysis

Abstract

This study examined whether quantitative content analysis of the value references, motive imagery, and integrative complexity expressed in the documents of two terrorist groups and two nonterrorist comparison groups could distinguish the violent groups from their nonviolent counterparts. The two terrorist groups were Central al Qa'ida and al Qa'ida in the Arabian Peninsula. For each, a comparison group that operated in the same context and had a similar ideology but did not engage in terrorist violence was chosen. Statistical analyses revealed that, compared with their nonterrorist counterparts, both terrorist groups described themselves by using more positive morality, religion, and aggression value references, and described their enemies by using more negative religion value references (e.g., references to being infidels). Relative to their nonviolent comparison groups, terrorist groups also used more power, ingroup affiliation, and achievement motive imagery, and expressed lower levels of integrative complexity.When people communicate, they often convey more than the explicit content of their words. Indeed, psychological research using content analytic techniques to examine value demonstrated the importance of attending not only to what people say but how they say it. The purpose of this paper is to analyze the communications of two terrorist groups and two comparison groups along these lines. In doing so, it builds upon and expands previous research examining the relationship between group rhetoric and terrorist violence (Smith, 2004;2008).During the past several decades, research on political violence has demonstrated that several quantitative content analysis coding systems provide scientifically rigorous and valid means of exploring the complex relationship between violent words and violent actions. Psychological variables related to individuals' and groups' values, motives, and quality of information processing have been coded in documents and found to be

Analyzing the semantic content and persuasive composition of extremist media: A case study of texts produced during the Gaza conflict

“…Content analysis involves examining texts for the occurrence of a defined set of coding categories, which denote specific characteristics of a message. Coding text for occurrences of an eclectic mix of 'features' can be problematic when seeking to capture intention because it either fails to exhaust the possible variations found in the texts (and so is not generalizable) or it becomes too coarse to provide the measurement necessary to distinguish meaningful similarities and differences (see Krippendorff 1980;Smith et al 2008). Consequently, we build our analysis around existing categorization schemes (particularly the "Table of Ten"; Giebels and Noelanders 2004;Giebels and Taylor 2009) and three conceptual distinctions that are fundamental to social scientific understanding of communication behavior and influence.…”

“…Thus, while their research focuses specifically on the issue of legitimization, it begins to tease out some of the common arguments and concepts that authors use to promote their cause. A second indication of what this approach can offer is given by two studies (Smith 2004;Smith et al 2008) that scored the use of value references (e.g., goals or standards aspired to) and motive imagery (e.g., the focus of underlying concerns) in terrorist literature. These studies show that terrorist literature contains more dominance, morality, and religion-based value references than the nonterrorist literature, and that the references groups attribute to themselves differentiates whether or not the group engages in violence.…”

Guiding Principles for Countering and Displacing Extremist Narratives

“…Such counternarratives should, wherever possible, incorporate efforts to promote integrative complexity, or the ability of an individual to see the world and competing values in a nuanced, rather than a binary, manner. Prior research suggests that lowered integrative complexity substantially raises the probability of engaging in violence (Smith, et al 2008). Programs promoting increased integrative complexity, such as "Being Muslim, Being British, " appear to lead to positive anti-violence outcomes (Savage 2011;Savage, Liht, and Williams 2011;Liht and Savage 2013).…”

Content Analysis for Proactive Protective Intelligence

“…Steps 1-4 are described in [8,9]. Steps 5 and 6 are described in [10], where we discuss a specific application of Frames in Action to a dataset containing texts issued by four groups: Central al Qa'ida (AQC), al Qa'ida in the Arabian Peninsula (AQAP), Hizb ut-Tahrir (HuT), and the Movement for Islamic Reform in Arabia (MIRA) [15]. We first divide the dataset in two subsets: document from terrorist sources (AQC, AQAP) and documents from non-terrorist sources (MIRA, HuT).…”

“…Once established that Frames in Action annotations are useful in identifying texts originating from terrorist sources, we use machine learning techniques to infer a decision-tree model from the data [15] that identifies terrorist texts in terms of Frames in Action annotation features. Training on 90% of the data to learn the decision-tree model and testing the decision-tree model on remaining 10% using 10-fold cross-validation, we obtained the results shown in Figure 2.…”