From Gates to Grids: Future of Terrorism Research
This paper discusses the evolving landscape of terrorism research in response to emerging technologies and shifting geopolitical dynamics. It discusses the gated access to information and how techniques of online data collection, advanced analytical tools and artificial intelligence have complemented researchers. Nonetheless, the study highlights the indispensability of collecting quality data, theoretical frameworks and contextual (area studies) knowledge in understanding the field.
Introduction
The future of terrorism studies depends on a healthy population of established scholars – that is to say, core researchers who achieve a certain level of acclaim through relevant and scientifically rigorous contributions to the field. Yet, in considering research production, the field is faced with a significant hurdle: data scarcity. Traditionally, primary data collection within terrorism studies has been dependent on research methods that feature high entry barriers, largely due to its reliance on access to detainees and confidential reports; such access is oftentimes restricted to a small circle of scholars.
This in turn can give rise to concerns regarding the accuracy and bias of the sources in question, especially given that they are gated from the wider community of researchers.[1] Yet, the primary data drawn from such research methods remains the gold standard within the field, for good reason. From a traditional standpoint, valuable contributions to the field are not possible without first attaining access to such primary data.
This is, of course, a false catch-22, common within not just terrorism studies, but many academic fields in general. Groundbreaking research has been penned by many emerging researchers, most of whom have impacted the field without the benefit of privileged access. Yet, we should not dismiss the hurdle as presented – the issue at its core remains relevant as without access to primary data, emerging researchers face significant challenges in establishing themselves, leading to a field characterised by a “miniscule” core of established researchers[2] and an inability to “attract new researchers and then hold onto them”.[3] Furthermore, emerging researchers also find themselves unable to substantiate or expand on the “invisible college” of orthodox scholarship, an issue that has only in recent years been mitigated.[4]
What is worth examining, then, in the context of pondering the future of terrorism studies, is how some emerging researchers have navigated this hurdle and produced worthwhile contributions without privileged access, but instead with sophisticated and innovative techniques. Perhaps more importantly, it is then crucial that we consider how future generations of scholars might be empowered with such capabilities, via the opportunities of augmentation presented by cutting-edge developments such as generative artificial intelligence (AI).
At the same time, we caution against sensationalist approaches which, in the pursuit of augmented data collection and analysis capabilities, lose sight of the importance of primary data and knowledge in producing quality research. As such, we reject the false dichotomy which pits ‘traditional’ research methods against ‘new’ research methods – rather, we recognise that the two build upon one another to generate valuable contributions to the field.
This paper explores the dimensions of terrorism studies research, beginning with an overview of the current landscape and how it shapes the barriers faced by new entrants. Then, it examines the techniques used by emerging scholars to circumvent these obstacles, focusing on data collation and analysis to carve out new niches within the field. The next section explores recent developments in generative AI and its potential to augment the conduct of large-N data analysis, further lowering the barrier of entry to research. Finally, it considers how, despite these technological advancements, the authenticity and accuracy of terrorism research remains inherently reliant on the knowledge drawn from primary sources. In doing so, the study hopes to equip emerging researchers not only with vital tools and techniques drawn from both innovators within the field and cutting-edge technological developments, but also with the wisdom of applying them judiciously.
Terrorism Studies: The Lay of the Land
The study of terrorism and counter terrorism has come a long way since the September 11, 2001 attacks. Academic evaluation of research methodology in the field, for instance, during this period was marked by criticism from key figures who would later become influential voices within the community, including Schmid and Jongman (1988),[5] Hoffman (1992)[6] and Crenshaw (2000).[7] Ranstorp’s (2009) survey of the field, which this article owes much of its insights to, noted that much of the pre-9/11 critique of the scholarship centred on an “over-reliance on recycled secondary sources and… academics being ensconced in ivory towers, instead of field research and talking to actual terrorists”.[8] At the same time, Silke’s (1990-1999) examination of the scholarship revealed that the majority of articles were written by ‘one-time authors’, revealing the field’s inability to hold on to emerging researchers.[9]
To view such critique as gatekeeping by key figures in the field would be a vast oversimplification, ignoring that the essentials of social science research – verifiable primary data and scientific rigour – are in fact necessary for quality research. Without these essentials, the field is left vulnerable to unverifiable claims and political agendas. This critique was reinforced in the post-9/11 era, in which the volume of contributions to the field rose significantly. Ranstorp, writing amidst this period, called to attention how the field had been affected by “the dominance of an ‘invisible college of scholars’… and the frequency of one-time visitors to the field”.[10]
Since then, perceptions of the field have shifted towards cautious optimism, highlighted by Morrison’s (2022) work drawn from the analysis of interviews with guests from the Talking Terror podcast.[11] In response to Sageman’s (2014) claim that terrorism research had stagnated,[12] Morrison instead argued that collective review has revealed a “surge of highly trained interdisciplinary researchers, coupled with greater access to data”,[13] as well as the development of a “more consistent community of researchers”.[14] He also highlighted Schuurman’s (2018) work, which revealed that “the use of primary data has increased and appears to be continuing to do so”.[15] Similarly, Schmid’s (2021) survey-driven review of terrorism research revealed that among 45 respondents, 14.9 percent viewed “primary data usage” as the key factor characterising “progress/achievements in the field” – the answer which received the highest consensus in response to the question.[16] While both Morrison and Schuurman were careful in providing caveats to the abovementioned enduring problems, their evaluations nevertheless present a field that in recent years has moved towards mitigating these problems.
Barriers to Access Within Terrorism Studies and Their Impact
Yet, even in this era of cautious optimism, many emerging scholars have failed to overcome the hurdle – that is, inaccessible primary data. Access to such data is fraught with several challenges, which include but are not limited to the following. First, it is difficult to access privileged data and terrorist prisoners for interviews.[17] Such access continues to be gatekept by government institutions[18] and, in today’s terrorism landscape, increasingly by social media platforms.[19] Second, primary data that is accessible by some is rarely shared with all. Hegghammer (2014) characterised this issue as a situation in which “each person [is] collecting his own data, for different purposes, different lengths of time, and in different ways. And worst of all, we’re not sharing”.[20]
While the development of online repositories such as the Global Terrorist Database (GTD) has come some way in dealing with this issue, it remains an issue of reasonable self-interest that – especially in environments where funding and practitioner demand decreases – researchers would be hesitant in pooling their primary sources and data. Furthermore, even if we set aside self-interest, it is common that much primary data access granted to established researchers is itself predicated upon confidentiality clauses and agreements. Third, as Berger (2019) noted, safety issues during field research remain a significant concern for researchers, in which they might be exposed to threats not only from extremists, but also from “governments in areas where extremist movements operate”.[21] Mitigating such threats requires substantial funding, in the form of private security and threat assessments, again presenting an obstacle for emerging researchers, who likely face greater difficulties obtaining such funding than established researchers.
Apart from issues of data access, the Eurocentric focus of terrorism studies represents a different form of barrier faced by some emerging researchers. Facing difficulties due to English being the lingua franca of the field, as well as issues of familiarity regarding European- or United States (US)-based publishing platforms, emerging scholars in the Global South may at times find it difficult to establish themselves in the community. As such, scholars writing within critical terrorism studies have pointed out that the field “continues to perpetuate the reproduction of Eurocentric research and the exclusion of non-Western voices”.[22] The loss of these indigenous voices and contributions from the Global South then hampers the addition of further nuance and rigour into existing discourses on terrorism research.[23]
The potential consequences of these challenges are significant, and in some cases have been demonstrated through precedent within the field. While Morrison noted that his interviewees have observed an increase in “career terrorism researchers”,[24] quantitative research into this claim still needs to be conducted before it can be confirmed – a fact that Morrison himself alluded to.[25] As such, complacency remains a danger given that the abovementioned challenges may still stymie the growth of new researchers in the field, motivating them to find greener pastures elsewhere. Alongside the issue of retaining researchers in the field, there remain the issues of the “invisible college” and groupthink, in which the lack of valuable contributions by emerging researchers then leads to the ossification of knowledge. Ranstorp (2009) noted that the field has been subjected to a situation in which “mutually reinforced camaraderie is often valued over scientific scrutiny of methods, theory, and data”.[26]
Emerging Scholars and Innovative Techniques
To circumvent the lack of access to primary sources, some emerging researchers have adopted sophisticated techniques. These innovative approaches, which can be categorised into online data collection and data analysis innovations, are compensatory measures which enable these researchers to carve out new niches and contribute valuable insights within the field. An example of emerging researchers who subsequently established themselves through such innovative techniques would be Jacob Shapiro, whose 2012 work alongside David Siegel[27] used complex game theory principles to explain why terrorist organisations continue to rely on “security-reducing” bureaucratic tools to manage operatives – a course of inquiry that eventually led to his book The Terrorist’s Dilemma (2013).[28] Another example would be Amy Pedahzur, who with Arie Perliger used social network analysis methods in 2006 to highlight the role of local struggles and family groups in deciding terrorist networks’ use of suicide bombings in Palestine.[29]
Online Data Collection
Emerging scholars are increasingly relying on online data gathering techniques to supplement or replace traditional methods of data gathering. While online data gathering in the current landscape necessitates that researchers sift through a vast expanse of rough data, much of which is simply noise in the context of most terrorism research questions, it does present the advantage of accessibility, allowing scholars to uncover and analyse primary data from publicly available sources, social media platforms and online forums.[30] Furthermore, such data specifically deepens research into the sub-fields of online terrorist activity and radicalisation, a field that is relevant amidst an increasingly online global population.
For example, one of the key pieces of emerging scholarship examining online far right extremism was conducted by Rieger, Kumpel, Wich, Kiening and Groh in 2021.[31] In their study, they employed advanced data trawling and scraping technologies, utilising various scrapers and publicly available online datasets, such as the Pushshift Reddit dataset, which “makes available all the submissions and comments posted on Reddit between June 2005 and April 2019”.[32] While the Pushshift Reddit dataset was not specifically built for terrorism-related research, scholars have built datasets based on social media platforms to study terrorism-related trends. For instance, a 2019 dataset was compiled by Kayode-Adedeji, Oyero and Stella, comprising “150 mass media YouTube videos on Al-Shahab, Boko Haram and IS terrorist groups from 2014 to 2016”, in which the attached discussions were categorised into 13 sub-topics.[33] This dataset has proven invaluable in understanding terrorism-related rhetoric and narratives on mainstream online platforms, and assisting researchers in tackling underlying theoretical questions regarding ideology and outreach.
Other advanced data mining techniques have been deployed alongside the development of technology. For example, a 2019 article by Garcia-Retuerta, Bartoleme, Chamoso and Corchado presents an innovative method for identifying new terrorist propaganda videos built from fragments of older terrorist media, via “a web-scraping method for retrieving relevant videos and a Hash-based algorithm which identifies the original content of a video.”[34]
Data Analysis Techniques
Beyond online data collection, many emerging researchers have also employed sophisticated data analysis techniques to dissect either publicly available data, or privileged data presented by other established researchers. Such analytical methods can unveil patterns that are not immediately apparent, offering a deeper understanding of the mechanisms of radicalisation, the spread of extremist ideologies and the operational tactics of terrorist groups.
For example, a 2018 study by Frissen, Toguslu, Ostaeyen and d’Haenens undertook an analysis of Quranic references in the Islamic State (IS)’s Dabiq magazine, providing a taxonomy of how religious texts are manipulated to fuel online violent radicalisation. This required the adoption of a funnel approach to accommodate a “broad contextual exploration of the surahs (chapters)” and a “more detailed textual examination of the ayat (verses) quoted in Dabiq” – as well as the use of various software such as NVivo and IBM SPSS Statistics.[35] Similarly, a 2019 study by Kling, Stock, Ilhan and Henkel provided an informetric analysis of strategic communications from IS. It adopted a mixed-method approach of content analysis via “coding categories of incitement, condemnation and rewards”, as well as the text-word method, “a knowledge representation method specifically suited for application on these non-scientific texts”.[36]
The use of sophisticated data analysis techniques for terrorism studies research has also undergone evaluation from emerging researchers possessing technical knowledge regarding the techniques. For instance, a 2019 paper by Kumar, Mazzara, Messina and Lee analysed “the performance of classifiers such as Lazy Tree, Multilayer Perceptron, Multiclass and Naïve Bayes classifiers for observing the trends for terrorist attacks around the world”.[37] Such an endeavour allowed them to not only evaluate the accuracy of various commonly used classification models in the context of terrorism studies research, but also identify the situations in which each might be best deployed. Furthermore, their study presented further avenues for improvement and research with regard to data analysis techniques, noting that a central objective would be to increase the “sub-classification layers and attributes both in order to find more useful trends”.[38]
The purpose of highlighting recent innovations in these techniques is twofold. First, it is to emphasise the fact that valuable terrorism studies research can in fact be conducted via open-source data, particularly in specific areas such as online terrorist activity and radicalisation. Second, most such studies rely on advanced technical skills, given that they comprise as much of the ‘hard’ sciences as they do of social science knowledge. This gives rise to a further quandary – not all emerging researchers have the desire or the capacity to develop such advanced technical skills while undertaking terrorism studies research. This then leads us to a potential alternative: the augmentation of one’s capabilities via generative AI.
Augmentation from Generative AI
Generative AI refers to a subset of AI technologies that can generate new content, such as text, images and code, which is similar to human-generated content. This is achieved through the use of algorithms and models that have been trained on large datasets of existing content. One of the key components of generative AI that can contribute to terrorism research is low- or no-code programming. Such developments significantly simplify the process of data analysis by minimising the amount of coding required, especially for analysts with rudimentary programming knowledge.
Low- or no-code programming empowers terrorism scholars by enhancing data analysis capabilities. Terrorism scholars could leverage the intuitive low- or no-code interface to clean, analyse and visualise data – making it easier to uncover trends without the need for complex coding. Beyond basic regression analysis, off-the-shelf generative AI tools often include functions like advanced data analysis packages, machine learning models, network analysis and natural language processing.[39] Scholars can leverage these advanced tools for more complex data analysis, enabling them to conduct deeper research on the intricacies of terrorist networks and to understand the underlying narratives within large volumes of textual data. This democratisation of data analysis tools allows for a broader base of researchers to contribute meaningful insights to terrorism research, fostering innovation and collaboration across disciplinary boundaries.
By simplifying the technical aspects of data manipulation and analysis, low- or no-code platforms ensure that scholars can focus more on the substantive aspects of their research. This shifts the focus of quantitative research from troubleshooting the syntax of programming languages to the comprehension of the conceptual and theoretical basis of each analytical method.[40] Generative AI can automate routine data processing tasks, allowing scholars to concentrate on interpreting results, drawing insights and making informed assessments about the direction of their studies. Consequently, quantitative analysis becomes less about the ability to code and more about the capacity to think critically about data and to apply the appropriate analytical frameworks that align with the objectives of the research.
This democratisation of sophisticated analytical tools through low- or no-code platforms significantly reduces the barrier of entry for terrorism research. Traditionally, the field has been somewhat inaccessible to those without a strong background in computer science or data analysis, limiting participation to a relatively small group of highly specialised researchers. However, by making powerful data analysis tools more user-friendly and accessible, a wider range of scholars – including those with expertise in political science, sociology, psychology and other related disciplines – can now contribute to terrorism research. This inclusivity not only enriches the field with diverse perspectives, but also encourages a multidisciplinary approach to understanding and combating terrorism.[41]
The Primacy of Quality Data
While AI and low- or no-code tools have made data analysis more accessible, their effectiveness is inherently tied to the quality of the data they process. Essentially, garbage in, garbage out remains a fundamental principle. Even the most sophisticated AI model cannot produce accurate, useful insights from flawed or biased data. This highlights the primacy of quality data: AI tools, no matter how sophisticated, cannot produce relevant and accurate data for analysis.[42] Turning back to the critique levelled at the terrorism studies field in the post-9/11 era, we find that – despite technological developments – the central concept of prioritising quality data remains the same.[43] Quality data in research is paramount, particularly for scholarship in terrorism. Research emerging from terrorism studies has real policy implications. Therefore, the accuracy of data collected from studying terrorist behaviours has far-reaching consequences, influencing not only academic discourse, but also national security strategies, law enforcement practice and international relations. Inaccurate or misleading data may distort analysis and misguide policies – this may be costly to lives and personal liberties.[44]
Collecting and organising high-quality data in terrorism research continue to be a challenge. Data in terrorism research usually comes from three sources: field research, official sources and online open-source. Each of these sources comes with its own set of challenges and limitations. Field research data, while invaluable for its depth and first-hand insights, can be difficult to obtain due to access restrictions and security risks. Official sources, though authoritative, may be limited by classification, redaction or even political biases that could affect reliability. Online open-source data, despite its timeliness, accessibility and volume, presents challenges due to its overwhelming mass and needs to be verified to extract meaningful information.
To leverage AI tools for research effectively, data management and organisation in terrorism research are crucial. Hence, to maximise the capabilities of AI in terrorism research, the fundamentals of collecting relevant and accurate data remain crucial.
The Indispensability of Theoretical Knowledge and Regional Context
Besides quality data, researchers must still familiarise themselves with fundamental theoretical principles of terrorism and counter terrorism and their respective region-specific contexts (area studies). While AI tools can automate data processing and uncover patterns within large datasets, they cannot replace the nuanced comprehension of terrorism theories and the complex sociopolitical dynamics that influence terrorism.
These theories are crucial to comprehend because terrorism is a manifestation of deeper societal, political and ideological undercurrents. Theories within terrorism studies, such as the radicalisation process,[45] intra- and inter-group dynamics,[46] and the strategies of asymmetric warfare,[47] are fundamental frameworks for comprehending data analysis. Additionally, the sociopolitical context where terrorism occurs varies significantly from one region to another. The researcher’s familiarity with the unique national and regional dynamics that may influence terrorism trends cannot be replaced by AI tools. Moreover, while the drivers of terrorism may share common characteristics, there are regional variations that warrant a more customised approach in evaluating the threat.
Conclusion
Terrorism research has evolved significantly, moving from an area dominated by gatekeeping and high entry barriers, to one that is increasingly accessible due to technological advancements. Emerging scholars now harness technology to collect data online and apply complex data analysis techniques, thus democratising access to information and enabling a broader range of perspectives within the field. The advent of AI’s low- or no-code programming platforms represents a further step in this evolution, offering researchers powerful tools to employ advanced statistical methods with greater ease, thereby lowering the barriers to entry even further.
However, while these technological advancements have transformed the landscape of terrorism research, they do not negate the need for thorough data collection, nor do they replace fundamental theoretical and region-specific knowledge. The essence of effective terrorism research lies not just in the ability to process large datasets, but in the interpretation of those datasets within the appropriate theoretical and contextual frameworks.
About the Authors
Kenneth Yeo and Benjamin Mok are Senior Analysts at the International Centre for Political Violence and Terrorism Research (ICPVTR), a constituent unit of the S. Rajaratnam School of International Studies (RSIS), Nanyang Technological University (NTU), Singapore. They can be reached at [email protected] and [email protected], respectively.
Citations
[1] It should be noted that this also gives rise to legal issues from a policy standpoint and ethical issues from an academic standpoint, which – while outside the scope of this paper – remain key concerns within the field of terrorism research.
[2] Thomas Hegghammer, “The Future of Terrorism Studies,” (presentation, EMC Chair Conference, Newport, 2013).
[3] Andrew Silke, “Contemporary Terrorism Studies: Issues in Research,” in Critical Terrorism Studies: A New Research Agenda, eds. Richard Jackson, Marie Smyth and Jeroen Gunning (New York: Routledge, 2009), pp. 34-48.
[4] Magnus Ranstorp, “Mapping Terrorism Studies After 9/11: An Academic Field of Old Problems and New Prospects,” in Critical Terrorism Studies: A New Research Agenda, eds. Richard Jackson, Marie Smyth and Jeroen Gunning (New York: Routledge, 2009), pp.13-33; John F. Morrison, “Talking Stagnation: Thematic Analysis of Terrorism Experts’ Perception of the Health of Terrorism Research,” Terrorism and Political Violence, Vol. 34, No. 8 (2022), pp. 1509-1529, https://doi.org/10.1080/09546553.2020.1804879.
[5] Alex P. Schmid and Albert J. Jongman, Political Terrorism: A New Guide to Actors, Authors, Concepts, Data Bases, Theories, And Literature (New Jersey: Transaction Publishers, 1988).
[6] Bruce Hoffman, “Current Research on Terrorism and Low-Intensity Conflict,” Studies in Conflict & Terrorism, Vol. 15, No. 1 (1992), pp. 25-37, https://doi.org/10.1080/10576109208435889.
[7] Martha Crenshaw, “The Psychology of Terrorism: An Agenda for the 21st Century,” Political Psychology, Vol. 21, No. 2 (2000), pp. 405-20, https://doi.org/10.1111/0162-895X.00195.
[8] Ranstorp, “Mapping Terrorism Studies after 9/11,” p. 17.
[9] Andrew Silke, “The Road Less Travelled: Recent Trends in Terrorism Research,” in Research on Terrorism, ed. Andrew Silke (London: Routledge, 2004), pp. 186-213, https://doi.org/10.4324/9780203500972.
[10] Ranstorp, “Mapping Terrorism Studies,” p. 17.
[11] Morrison, “Talking Stagnation.”
[12] Marc Sageman, “The Stagnation in Terrorism Research,” Terrorism and Political Violence, Vol. 26, No. 4 (2014), pp. 565-580, https://doi.org/10.1080/09546553.2014.895649.
[13] Morrison, “Talking Stagnation,” p. 1524.
[14] Ibid., p. 1518.
[15] Bart Schuurman, “Research on Terrorism, 2007–2016: A Review of Data, Methods, and Authorship,” Terrorism and Political Violence, Vol. 32, No. 5 (J2020), pp. 1011-1026, https://doi.org/10.1080/09546553.2018.1439023.
[16] Alex P. Schmid, James J. F. Forest and Timothy Lowe, “Counter-Terrorism Studies: A Glimpse at the Current State of Research (2020/2021): Results from a Questionnaire Sent to Scholars and (Former) CT Practitioners,” Perspectives on Terrorism, Vol. 15, no. 4 (2021), pp. 155-183.
[17] Morrison, “Talking Stagnation,” p. 1521.
[18] John F. Morrison, “Gary Lafree: The Global Terrorism Database,” 2017, in Talking Terror, Season 1, Episode 20, podcast, https://soundcloud.com/user-366747443/gary-lafree, quoted in Morrison, “Talking Stagnation,” p. 1521.
[19] J.M. Berger, Researching Violent Extremism: The State of Play (Washington, D.C.: RESOLVE Network, 2019), https://doi.org/10.37805/rve2019.3, p. 9.
[20] Hegghammer, “The Future of Terrorism Studies,” p. 3.
[21] Berger, “Researching Violent Extremism,” p. 9.
[22] Shirley Achieng’, Samwel Oando and Richard Jackson, “Critical Terrorism Studies,” in A Research Agenda for Terrorism Studies, eds. Lara Frumkin, John Morrison and Andrew Silke (Cheltenham: Edward Elgar Publishing, 2023), pp. 63–76, https://doi.org/10.4337/9781789909104.00009.
[23] There exist also wider issues regarding how the silencing of indigenous and Global South voices fuels the political bias of knowledge production, often in concert with certain political agendas. This is a concern that affects not only emerging researchers, but also established researchers, and is thus not explored in this paper. See Ilyas Mohammed, “Decolonialisation and the Terrorism Industry,” Critical Studies on Terrorism, Vol. 15, No. 2 (2022), pp. 417-440, https://doi.org/10.1080/17539153.2022.2047440.
[24] Morrison, “Talking Stagnation,” p. 1518.
[25] Morrison, “Talking Stagnation,” p. 1525.
[26] Ranstorp, “Mapping Terrorism Studies,” p. 30.
[27] Jacob N. Shapiro and David A. Siegel, “Moral Hazard, Discipline, and the Management of Terrorist Organizations,” World Politics, Vol. 64, No. 1 (2012), pp. 39-78, https://doi.org/10.1017/S0043887111000293.
[28] Jacob N. Shapiro, The Terrorist’s Dilemma: Managing Violent Covert Organizations (Princeton: Princeton University Press, 2013).
[29] Ami Pedahzur and Arie Perliger, “The Changing Nature of Suicide Attacks: A Social Network Perspective,” Social Forces, Vol. 84, No. 4 (2006), pp. 1987-2008.
[30] Megha Chaudhary and Divya Bansal, “Open Source Intelligence Extraction for Terrorism‐related Information: A Review,” WIREs Data Mining and Knowledge Discovery, Vol. 12, no. 5 (2022), p. 1473, https://doi.org/10.1002/widm.1473.
[31] Diana Rieger et al., “Assessing the Extent and Types of Hate Speech in Fringe Communities: A Case Study of Alt-Right Communities on 8chan, 4chan, and Reddit,” Social Media + Society, Vol. 7, No. 4 (2021).
[32] “Papers With Code – The Pushshift Reddit Dataset,” https://paperswithcode.com/dataset/pushshift-reddit.
[33] Tolulope Kayode-Adedeji, Olusola Oyero and Stella Aririguzoh, “Dataset on Online Mass Media Engagements on YouTube for Terrorism Related Discussions,” Data in Brief, Vol. 23 (2019), https://doi.org/10.1016/j.dib.2018.12.020.
[34] David García-Retuerta et al., “Counter-Terrorism Video Analysis Using Hash-Based Algorithms,” Algorithms, Vol. 12, No. 5 (2019), p. 110, https://doi.org/10.3390/a12050110.
[35] Thomas Frissen et al., “Capitalizing on the Koran to Fuel Online Violent Radicalization: A Taxonomy of Koranic References in ISIS’s Dabiq,” Telematics and Informatics, Vol. 35, No. 2 (2018), pp. 491-503, https://doi.org/10.1016/j.tele.2018.01.008.
[36] Frauke Kling et al., “The Islamic State’s Strategic Communication: An Informetric Topic Analysis,” Proceedings of the Association for Information Science and Technology, Vol. 55, No. 1 (2018), pp. 264-273, https://doi.org/10.1002/pra2.2018.14505501029.
[37] Vivek Kumar et al., “A Conjoint Application of Data Mining Techniques for Analysis of Global Terrorist Attacks — Prevention and Prediction for Combating Terrorism” arXiv, February 21, 2019, http://arxiv.org/abs/1901.06483.
[38] Ibid.
[39] Iqbal H. Sarker, “AI-Based Modeling: Techniques, Applications and Research Issues Towards Automation, Intelligent and Smart Systems,” SN Computer Science, Vol. 3, No. 2 (2022), p. 158, https://doi.org/10.1007/s42979-022-01043-x.
[40] Daniel Pinho, Ademar Aguiar and Vasco Amaral, “What About the Usability in Low-Code Platforms? A Systematic Literature Review,” Journal of Computer Languages, Vol. 74 (2023), https://doi.org/10.1016/j.cola.2022.101185.
[41] Harmonie Toros, “Interdisciplinarity, Globality and Downsizing: Aspirations for the Future of Terrorism Studies,” in A Research Agenda for Terrorism Studies, eds. Lara Frumkin, John Morrison and Andrew Silke (Cheltenham: Edward Elgar Publishing, 2023), pp. 263-274, https://doi.org/10.4337/9781789909104.00025.
[42] Tyler H. McCormick et al., “Using Twitter for Demographic and Social Science Research: Tools for Data Collection and Processing,” Sociological Methods & Research, Vol. 46, No. 3 (2017), pp. 390-421, https://doi.org/10.1177/0049124115605339.
[43] Ranstorp, “Mapping Terrorism Studies,” p. 17.
[44] Mary DeRosa, Data Mining and Data Analysis for Counterterrorism (Washington, D.C.: The CSIS Press, 2004).
[45] Fathali M. Moghaddam, “The Staircase to Terrorism: A Psychological Exploration,” American Psychologist, Vol. 60, No. 2 (2005), p. 161; Fathali M. Moghaddam, From the Terrorists’ Point of View: What They Experience and Why They Come to Destroy (Westport: Praeger, 2006); Randy Borum, “Radicalization into Violent Extremism I: A Review of Social Science Theories,” Journal of Strategic Security, Vol. 4, No. 4 (2011), pp. 7-36.
[46] Mia M. Bloom, “Palestinian Suicide Bombing: Public Support, Market Share, and Outbidding,” Political Science Quarterly, Vol. 119, No. 1 (2004), pp. 61-88; Bruce Hoffman, Inside Terrorism (New York: Columbia University Press, 2006); Joshua Kilberg, “A Basic Model Explaining Terrorist Group Organizational Structure,” Studies in Conflict & Terrorism, Vol. 35, No. 11 (2012), pp. 810-830; Shapiro, The Terrorist’s Dilemma.
[47] David C. Rapoport, “The Four Waves of Modern Terrorism,” in Attacking Terrorism: Elements of a Grand Strategy, eds. Audrey Kurth Cronin and James M. Ludes (Washington, D.C.: Georgetown University Press, 2004), pp. 46-73; Andrew H. Kydd and Barbara F. Walter, “The Strategies of Terrorism,” International Security, Vol. 31, No. 1 (2006), pp. 49-80, https://doi.org/10.1162/isec.2006.31.1.49; Ethan Bueno de Mesquita and Eric S. Dickson, “The Propaganda of the Deed: Terrorism, Counterterrorism, and Mobilization,” American Journal of Political Science, Vol. 51, No. 2 (2007), pp. 364-381; Jeffrey W. Lewis, “Precision Terror: Suicide Bombing as Control Technology,” Terrorism and Political Violence, Vol. 19, No. 2 (2007), pp. 223-245.
