Understanding the influence of goal-structure on the disbandment of terrorist groups

Table of Contents

List of abbreviations.

List of graphs, illustrations, and tables.

Introduction

Literature Review
Roots of Terrorism
Characteristics of Terrorist Groups
Terrorist Groups’ End
Gap & Contribution

Theory
Terrorism
Goal-structure & Criteria
Argument & Causal Mechanisms

Research Design
Datasets, Case Selection & Timeframe
Operationalisation
Regression Model

Analysis
Empirics
Politicisation
Victory
Splintering
Policing
Military Force
Alternative Explanations
Reflections
On the Impact of Goal-Structure

Conclusion

References

Abstract

Very little is known about the circumstances that lead a terrorist group to cease their operations in one way rather than another. This study seeks to answer the question what impact the goal-structure of a terrorist group has on their disbandment. A new framework describes the structure of goals by differentiating operative and official goals of terrorist groups along their geographical scope, their final goal’s extent, and the respective effectiveness this is accomplished by. I argue that strategic or abstract goal-structures determine the way a terrorist organisation ends. By applying a multinomial logistic regression model with a novel dataset on 155 groups’ goal-structures, I show that a more strategic goal-structure increases the likelihood of splintering and victory, while an abstract goal-structure increases the likelihood of defeat by police or military force whilst also leading terrorist groups to join non-violent political participation.

Keywords: goal-structure; terrorism; disbandment; ideology; counterterrorism

List of abbreviations

Abbildung in dieser Leseprobe nicht enthalten

List of graphs, illustrations, and tables

Graph 1: Distribution of scope, extent, and effectiveness

Graph 2: Marginal effects of predicted probabilities for politicisation

Graph 3: Marginal effects for politicisation by different ideologies

Graph 4: Marginal effects of predicted probabilities for victory

Graph 5: Marginal effects of predicted probabilities for splintering

Graph 6: Marginal effects for splintering by different ideologies

Graph 7: Marginal effects of predicted probabilities for policing

Graph 8: Marginal effects for policing by different ideologies

Graph 9: Marginal effects of predicted probabilities for military force

Graph 10: Impact of GSS on predicted probabilities for all types of disbandment

Illustration 1: Causal mechanism: strategic goal-structure leads to politicisation

Illustration 2: Causal mechanism: strategic goal-structure leads to victory

Illustration 3: Causal mechanism: strategic goal-structure leads to splintering

Illustration 4: Causal mechanism: abstract goal-structure leads to policing or military force

Table 1: Comparing different concepts of terrorist groups’ disbandment

Table 2: Distribution of independent and dependent variable

Appendices

Appendix 1 Results of the main model with GSS as independent variable

Appendix 2 Results of the main model with scope, extent, and effectiveness as independent variables

Introduction

Why do terrorist groups dissolve? What explains the different patterns of terrorist organisations’ disbandments? How is a group that refrains from violence to join the political process different from a group that gets imprisoned through counterterrorism operations? Researchers have been improving our understanding of terrorism but so far very little remains known about the circumstances of demise of terrorist organisations.

It is striking how much research has been done both quantitatively and qualitatively on terrorist groups: why they appear, the causes of terrorism, and its effects. On the other hand, besides numerous valuable case studies, to date there are only three published quantitative works on the question how terrorist groups end (Cronin, 2009; Jones & Libicki, 2008; Weinberg, 2012). Hence, our knowledge on causes of different kinds of disbandment of terrorist groups and its general patterns is very limited. This thesis aims to advance the understanding of the disbandment of terrorist groups by connecting the well-developed body of literature on terrorist groups’ characteristics with the understudied field of terrorist groups’ disbandment. In particular, I will pick up Piazza’s (2009) concept of goal-structure of terrorist groups and apply it on Jones and Libicki’s (2008) dataset of terrorists’ demise in order to answer the research question: what impact does the goal-structure of a terrorist group have on their disbandment?

The obvious question now is: why this concept? Existing literature has focused on ideology as the major explanatory variable for different disbandments of terrorist groups. We can observe the same trend in literature on characteristics of terrorist groups: religiously motivated groups and in particular Islamic terrorist groups are often claimed to be more dangerous, more lethal, or sometimes even more irrational than other types of terrorist groups (Gunning & Jackson, 2011). This accusation is overly simplifying and has been challenged by researchers including Piazza (2008a, 2009) and his work on goal-structure. Similarly, various other theories shine a light on terrorist organisations and emphasise the importance of other explanatory factors supplementary than religion (Hoffman & McCormick, 2004; Klein, 2015; Pearson, Akbulut, & Lounsbery, 2015). In this context, goal-structure has proven to explain varying lethality of terrorist groups rather than their ideology, thereby offering an alternative concept to understand how terrorist groups work.

Since the literature on terrorists’ disbandment has focused exceedingly on ideology, I introduce goal-structure as an alternative explanation for terrorist group’s disbandment. This concept refers to the goal achievement strategy of an organisation by distinguishing between official and operative goals instead of focusing on the sole purpose a group ascribes itself (e.g. territorial sovereignty or policy change). I define goal-structure as a group of aims aligned in a priority system, which is characterised by varying scope, extent, and effectiveness, to achieve a final goal. Scope can be understood as the geographical level a group’s demands refer to, extent of aims as how much change from the initial status quo is necessary to achieve their final goal, and effectiveness as the relationship between stated goals and actual behaviour. Terrorist groups can be characterised by either a strategic or an abstract form of goal-structure. I argue that goal-structure clearly determines a group’s disbandment and therefore has important implications for our understanding of terrorist groups and how to approach them. While a strategic goal-structure leads a group onto the path of negotiations, victory, or splintering, I expect an abstract goal-structure to almost doom the group to experience defeat either by the military or police forces.

Part of my contribution is reconceptualising the goal-structure concept and establishing a solid foundation of its theory in order to operationalise its dimensions comprehensibly. Another contribution of this paper, lies in coding the dimensions of goal-structure for all 155 terrorist groups ended between 1998 and 2006. Eventually, I merge my novel data with Jones and Libicki’s (2008) dataset on terrorist groups’ disbandment and apply a multinomial logistic regression model in order to explain the five types demise: policing (i.e. judiciary actions and prosecution), military force (i.e. defeat through army), politicisation (i.e. transformation to non-violent political participation), victory, and splintering.

My results reveal significant relationships of varying strength between a terrorist group’s goal-structure and their disbandment. I find support for my hypotheses that a terrorist group characterised by a strategic goal-structure is associated with splintering and victory, just as I hypothesised. An abstract goal-structure surprisingly leads to an increased likelihood of politicisation – opposed to my argument. I only find limited support for the expected relationship between abstract goal-structure and an end by policing or military force. Scope and effectiveness provide additional significant results and allow for a more enriching interpretation and discussion.

On the following pages, I will give an overview on the current literature on terrorism studies with regards to terrorist groups’ cessation in order to locate this study’s contribution within the research field. In the next step, I will introduce the key concepts my work is relying on and extend the concept of goal-structure based on Piazza’s groundwork. After developing my hypotheses, the methodological chapter introduces the datasets I base my thesis on, as well as the operationalisation of my variables and the statistical model applied. Consequently, I present the analysis of my results, starting with some general remarks and empirics, followed by the discussion of the main results for each kind of terrorist groups’ disbandment. After considering alternative explanations and reflecting on limitations of the study, I conclude by briefly summarising the thesis’ major findings and implications for future research.

Literature Review

Terrorism Studies is, like the concept of terrorism itself, a highly politicised field and research on its elements has been carried out with varying intensity depending on public interest and funding provided by private and state donors. Since 9/11 and particularly 2004^[1] we witness an increasing amount of incidents of terrorism, which led to an enduring public interest in the phenomenon of terrorism and a growing body of literature published in the field. Terrorism Studies constitutes one of the most important fields for policy makers, which pressures researchers to permanently try to answer the question: How can we end terrorism? Seemingly the Holy Grail of Terrorism Studies.

The terrorism literature I am speaking to can be grouped in three categories with reference to how terrorism can be ended: firstly, authors who have focused on the roots of terrorist activity, assuming that by tackling the causes of terrorism we can eradicate the phenomenon as a whole. Secondly, scholars who have directed their work towards understanding terrorist groups as organisations, studying their patterns and characteristics. This branch hopes to identify what makes terrorist organisations more dangerous or lethal in order to direct counterterrorism efforts to these groups. Finally, one group of researchers looks particularly on terrorist organisations that have ended their violent campaigns in the past, hoping that by studying those cases they can inductively conclude how other terrorist groups can be dissolved as well. Following, I will present the three different groups in more detail.

Roots of Terrorism

Terrorism scholars, who focused on the causes of terrorist activity researched a diverse range of possible conditions that stimulate the emergence of terrorist violence in a country. Here, I just want to focus on the three most researched – and publicly often claimed major – causes of terrorism: regime type, poverty, and education.

Crenshaw’s (1981) argument that terrorism is related to a society’s degree of freedom that allows its members to speak up is supported by various researchers who found evidence that terrorism is more likely to occur in democracies than in autocracies. Aksoy, Carter, and Wright (2012) provided evidence that democracies offer a stage for terrorist organisations to perpetrate their attacks while also allowing them to prepare them more freely than it would be possible in more restrictive regimes. Another mechanism explaining this counterintuitive relationship rests on the capability of autocracies to use repression to counter terrorist activities (Aksoy et al., 2012; Daxecker & Hess, 2012). On the contrary, Baylouny (2004) argues that democracies should experience less terrorist violence since they are more capable of coping with terrorists by including and co-opting them. In fact, Wilson and Piazza (2013) provide evidence that a mix of coercion and co-option is the best strategy for any type of regime to end terrorist attacks.

Secondly, the prevalence of poverty as cause of terrorism is conventional wisdom but actually a highly debated claim. This hypothesis has been rejected by a number of scholars, who have tested various indicators like income, GDP, GDP per capita, GDP growth, or inflation but found no significant correlation with incidents of terrorism in a country (Abadie, 2006; Atran, 2003; Krueger & Malečková, 2003; Li & Schaub, 2004; Piazza, 2006, 2010). In fact, Krueger and Malečková (2003) find that individuals who decide to join a terrorist campaign are often having an economically stable middle-class background. Piazza (2011, 2012) advocates for a more nuanced relationship between poverty and terrorism by proving that countries are more likely to experience terrorist attacks when they accommodate economically discriminated minority groups. Li and Schaub (2004) have shown that trade, portfolio investments, and received FDIs do not have a direct effect on transnational terrorist attacks committed in a country but that those attacks are less likely to occur in economically more developed countries. Whereas for the case of failed states Piazza (2008b, 2008c) found evidence that they were indeed more likely to become hosts of groups committing transnational terrorist attacks.

Another cause of terrorism brought forward on a regular basis is the lack of education. Numerous scholars have shown that the relationship is very complex and cannot be reduced to a lack of education leading to an increase in terrorist attacks (Azam & Thelen, 2008; Berrebi, 2007; Brockhoff, Krieger, & Meierrieks, 2010, 2015; Drakos & Gofas, 2006a; Krueger & Malečková, 2003). Bravo and Dias (2006) find a generally weak significant relationship between increased literacy and a decrease in terrorist attacks, researching a subsample of 60 countries in Eurasia. Brockhoff et al. (2015) also researched this nexus more in depth and surprisingly found that an increase in literacy and primary education also leads to an increased likelihood of domestic terrorism when looking only at less developed countries. On the contrary, more developed countries face less domestic violence the more people are enrolled in university education.

Characteristics of Terrorist Groups

By today, numerous authors have engaged in answering the question why terrorist groups behave in certain ways, what explains their choice of strategies, and how their activities are influenced by external factors. One of the most important authors in the field remains Pape (2003) with his empirical study on foreign occupation and the use of suicide terrorism, providing evidence that terrorist groups decide to apply this strategy because it is successful in achieving their aims. This finding has been disputed by several scholars (Cook, 2007; Horowitz, 2010) but most importantly, on the grounds of this study, it is also often claimed that terrorists are rational and strategic actors. This assumption has been questioned for example by Abrahms who argues that “the preponderance of evidence is that people participate in terrorist organisations for the social solidarity, not for their political return.” (2008, p. 94).

A different approach was chosen by Asal and Rethemeyer (2008) who found that a group’s network as well as the controlled territory are reliable explanatory factors for its deadliness. Building on this, Brathwaite (2013) found a positive relationship of both territorial claims and group competition, with terrorist groups likelihood to participate in elections. Cook and Lounsbery (2011) emphasise the importance of increased group size and ideology – especially religious ideology – that explains terrorist organisations’ increased levels of violence, which is also supported by the findings of the previously mentioned Asal and Rethemeyer’s (2008) study. In response to the claim that religiously groups are more lethal than other terrorist groups, Piazza (2009) used the concept of goal-structure and looked at Islamist groups in particular. He demonstrates that goal-structure explains the variation of Islamist terrorist groups’ lethality and that Islamist terrorist groups with a strategic goal-structure are just as deadly as any other ideology. Goal-structure was previously developed by Piazza (2008a) as an explanatory concept for suicide terrorism in general. He understands goal-structure as what a group’s aims are oriented upon, either ideological values or discrete goals. In his terminology terrorist groups either have a universal/abstract goal-structure or a strategic goal-structure. According to Piazza “the former are distinguished by highly ambitious, abstract, complex, and nebulous goals that are driven primarily by ideology” (Piazza, 2009, p. 65) while the latter are mainly characterised by their “limited and discrete” aims. This conceptualisation is very vague, which is due to the fact that his typology is mostly driven by empirical findings and merely inductive (Piazza, 2008a, 2009). In fact, Piazza (2009) proves that Al-Qaida-affiliated terrorist groups account for most of the lethality caused by Islamist terrorist groups. In order to theoretically explain the difference between Al-Qaida-affiliated and other terrorist groups he introduces the goal-structure concept, ex post facto labelling Al-Qaida-affiliated groups ‘abstract’ and those groups not related to Al-Qaida ‘strategic’. As an explanatory concept for suicide terrorism, it has been criticised by Klein (2015) for not taking into account the importance of signalling theory. Overall, Piazza’s attempt to establish goal-structure as a novel concept lacks theoretical underpinning, but bears potential explanatory power to understand terrorist groups’ conduct.

Terrorist Groups’ End

Finally, in recent years more scholars have researched the reasons why some terrorist groups end and more specifically what determines different types of disbandment. Jones and Libicki (2008) were the first to quantitatively study how terrorist groups end by looking at 648 different organisations that were active between 1968 and 2006 deriving from the former RAND-MIPT^[2] dataset. They composed a dataset covering valuable information about the group’s size, finances, ideology, goals, and how it ended (or if it is still active).

Jones and Libicki (2008) code five options on their dependent variable how a terrorist group can possibly end if it do not remain active: policing, military force, victory, politicisation, or splintering. Another important work on how terrorist groups end published by Cronin (2009) defines six ways how terrorist groups usually end: decapitation, negotiations, success, failure, repression, and reorientation. Besides these two theories, only Weinberg (2012) published a noteworthy study on the disbandment of different terrorist groups naming defeat, success, and transformation as possible ending scenarios. Since they all study the same population, it is surprising to find them using different schemes and therefore worthwhile comparing the different concepts of ending they developed.

Abbildung in dieser Leseprobe nicht enthalten

Table 1: Comparing different concepts of terrorist groups’ disbandment

As shown in the table 1, for most types of disbandments we can identify a counterpart describing one and the same phenomenon in each study. But an interesting question raised from comparing these typologies is: How are the terrorist groups that are classified as fail in Cronin’s (2009) concept categorised in Jones and Libicki’s (2008) or Weinberg’s (2012) typology? Cronin (2009, pp. 94–114) conceptualises failure in a broad sense, including groups which have failed to hand over operations between generations, fractionalisation (aka splintering), marginalisation due to an irrelevant ideology or decreasing popular support, the loss of operational control and targeting errors as well as amnesties granted to terrorist groups or individuals by the government. Weinberg (2012) would understand parts of this ‘failure’ concept as transformation. For the Jones and Libicki (2008) typology it has to be assumed that most of those ‘failed’ terrorist groups which have not splintered are categorised as ended by policing or politicisation. This has important implications for their dataset, since they possibly overestimate the real size and effect of those two strategies.

Continuing to the datasets these three books compiled on terrorist groups’ disbandment, Jones and Libicki (2008) find that overall 268 groups out of 648 have ended between 1968 and 2006 of which 43% ended because of politicisation, 40% because of policing while only 10% achieved victory and even less were defeated by military force (7%). A total of 136 groups ended because of splintering, which is the largest number of all despite those remaining active. However, the authors do not consider those groups ended but argue that they remain active - just in a different form. Cronin (2009) accounts for the same timeframe and data source but relies on a smaller population of 457 terrorist groups – both, active and ended – overall conditioned by her exclusion of all terrorist groups “with only one attack or a single series of coordinated attacks within several days of one another” (Cronin, 2009, p. 208). In her dataset she did not code the particular types of end for every terrorist group, but focused on several criteria related to the end of the group. For instance, she found that 81 terrorist groups participated in negotiations but only 24 managed to resolve their conflicts through them. Her mostly qualitative work establishes different mechanisms using a number of case studies to illustrate her theory. Weinberg (2012) follows her example and focuses on establishing explanatory mechanisms, while relying on an individually coded dataset. In contrast to previous datasets, he chooses a timeframe ranging from 1900 to 2006 and bases his coding on different sources, including the RAND-MIPT dataset. Despite a much longer timespan, he only identifies 433 terrorist groups that were active at any point during the given timeframe – in fact the smallest sample of all three studies. Interestingly, he shows that 17.7% of all terrorist groups in this dataset have been ended by capturing their leader plus 12.7% that ended due to government repression. Another interesting figure in his data are the 7.2% terrorist groups which ended by joining non-violent political participation. Unfortunately, he merges his nuanced data on terrorists’ disbandment in broad categories like defeat instead of analysing each type individually.

Jones and Libicki (2008) find that terrorist groups with narrower goals^[3] are more likely to end by politicisation. On the contrary, Cronin (2009) shows in her quantitative analysis that groups which negotiate are almost always fostering territorial control – a rather broad goal in Jones and Libicki’s (2008) dataset. Weinberg (2012) points out that more than 40% of the terrorist groups in his dataset from 1900 to 2006 are left-wing oriented, while 16% state right-wing views, 20% are nationalistic or separatist, and only 11 % account for religiously motivated terrorist groups. This is particularly interesting when compared to Jones and Libicki (2008) who have only collected data on terrorist groups from 1968 onwards and find similar shares on left-wing groups but overwhelming 36% nationalistic terrorist groups plus 6 % which embody right-wing perspectives. Moreover, the growing number of religiously motivated groups becomes apparent with a share of 22%. Further findings by Jones and Libicki (2008) suggest that religious terrorist groups in general are more durable than any other group ideology^[4] and that they are a rather recent phenomenon compared to other groups; Large and very large groups are more likely to end by politicisation while smaller groups are more likely to end by policing. Adding to this, Cronin (2009) shows that older (longer-lived) terrorist groups are more likely to negotiate than younger organisations.

Even though all three analyses composed impressive datasets, they surprisingly did not conduct any statistical analysis of their data at all. Therefore, they mainly focus on descriptive statistics and case studies, where they use their data to underpin their arguments. Luckily, other authors picked up their rich data and analysed it more in-depth, especially the dataset composed by Jones and Libicki (2008). Blomberg, Gaibulloev, and Sandler (2011) for instance conducted a survival analysis and found that – in line with what Jones and Libicki (2008) described – religious groups as well as large groups which commit a variation of terrorist attack strategies survive significantly longer. Gaibulloev and Sandler (2014) computed discrete-time duration models to identify determinants of terrorist groups’ disbandment, which allows them to mine the explanatory power of the dataset. Their analysis of the Jones and Libicki (2008) data leads to a number of interesting findings, one of them being that terrorist groups are less likely to end by politicisation or victory when they foster empire, regime change, social revolution, or territorial change as goals. Adding to this, they find that religious terrorist groups are also less likely to end by politicisation or victory than any other type of group. However, they were not able to identify particular causes that could help explaining why some groups experience military defeat.

Blomberg, Engel, and Sawyer (2010) researched the question what influences the duration of a terrorist group, which is closely related to the quest why terrorism ends. They find that for each year a group remains active the more likely it is to survive to the next year. The likelihood to survive can be increased by a group through an increased use of violence. This leads them to the conclusion that the field of terrorist organisations is constantly ‘monopolised’ by some actors who are more violent and last longer than other groups.

Other studies that touch upon the topic of disbandment of terrorist groups often focus on durability or longevity of a terrorist group instead of investigating the variation of disbandments. However, their findings are considerable and deserve attention. Carter (2012) researched the effect of state-sponsorship and safe haven on a terrorist groups’ likelihood to end, while Phillips (2014), for the first time considers interaction between terrorist groups as a determining factor for the durability of terrorist groups. He shows the importance of cooperation between terrorists for their survival. This is particularly interesting when taking into account Kaplow’s (2015) findings that conflict groups might refuse to take part in negotiations because they fear to alienate their partners and supporters or signal weakness towards them. Phillips (2015) also found that inter-field rivalries between terrorist groups increases their longevity, while Pearson et al. (2015) researched the effect of group structure, networks, and partnerships on the durability of terrorist groups as well as their target choice. Among other findings, they confirm Phillips’s (2014) hypothesis that connectedness increases longevity.

Gap & Contribution

As shown in the literature review, comparably little attention has been given to study the different ways terrorist groups end. Despite the availability of data on this topic, only a handful of scholars have investigated the disbandment of terrorist groups so far. As I also pointed out, the rich research on roots and characteristics of terrorist groups has already provided numerous explanatory variables and causal mechanisms. I argue that combining these concepts with the understudied research question why terrorist groups end, offers great potential to expand the explanatory power of existing theories.

While Jones and Libicki (2008) consider type of goals and ideologies of terrorist organisations as explanations for the disbandment of terrorist groups, Gunning and Jackson (2011) make a strong case for not focusing on terrorists’ ideology, but rather understand the means by which they seek to achieve their goals. Previous research popularly divides religious from nationalist, left-wing from right-wing groups, although these terms actually do not tell us anything about the real motivation of a terrorist group or how they want to achieve their goals at all. In fact, there is good reason to believe that also the content of goals does not matter that much for terrorist groups, since Abrahms (2008) has shown that members of a terrorist organisation often are not even aware of the particular policy goals their group has. Therefore, there has to be another driving reason for terrorist groups to end in different ways than the purport of goals.

Piazza (2009) demonstrated that goal-structure explains better than their ideology why some groups are more lethal than others. So maybe it is again not ideologies that drive terrorist groups when they end but their underlying goal-structure. No further attempt has been made so far to gain a deeper understanding of how a group’s structure of goals impacts their likelihood of for example joining the political process or ending by policing. I consider goal-structure to be a more helpful analytical framework to understand the incentive structure of a group by analysing the scope of their goals (global or local), the goals’ extent (limited or total), and the effectiveness when achieving them. I expect these newly developed structural criteria to be more suitable to understand terrorist groups and to predict the way they will end.

There are also some more simplistic reasons, why even the few studies conducted on the end of terrorism deserve a critical assessment. Most of them have relied on the (most comprehensive) dataset provided by Jones and Libicki (2008), which only covers all types of terrorism from 1998 to 2006. From 1968 to 1997 only international terrorism was included, which barely accounts for 10% of the total terrorist incidents (Hoffman & Hoffman, 1995, p. 180). This means that every study based on this database so far has suffered a major selection bias by international terrorist groups between 1968 and 1997, that have in fact all ended, accounting for 248 out of 404 groups that have ended. Besides this selection bias, research so far cannot provide an explanation why some terrorist groups face military defeat or achieve victory while others do not.

This work’s contribution is threefold: Firstly, I develop the theory of goal-structure and its characteristics further in order to explain terrorist groups’ disbandment. Secondly, I build a Goal-Structure Score (GSS) dataset covering 155 terrorist groups that have ended between 1998 and 2006. Thirdly, I use the developed theory and its data to explain the different types of terrorists’ end, a novel nexus connecting two strands of literature in the research field. In sum, I will investigate if goal-structure can help us understand the decisions terrorist groups make or make not with perspective on how they end and why some groups are more likely to be defeated by counterterrorism measures, while others join the political process or transform into peaceful modes of protest or civic activism. The research question I’m seeking to answer on the following pages is:

What impact does the goal-structure of a terrorist group have on their disbandment?

Theory

In order to answer this question, I will first spell out the relevant theoretical concepts my study relies upon. I start out with a definition of terrorism and a short conceptualisation of terrorists’ disbandment, followed by a comprehensive introduction of the concept of goal-structure. After I presented the key concepts of my thesis, I will proceed with my main argument, articulate my hypotheses and line out the underlying causal mechanisms.

Terrorism

Despite the centuries of debate on the definition of terrorism, there is only little agreement on what terrorism is, what it involves, and who can perpetrate it (Hoffman, 2006). This work is based on the following definition of terrorism used by Jones and Libicki (2008) since I rely on their data drawn from the former RAND-MIPT database:

“ Terrorism involves the use of politically motivated violence against noncombatants to cause intimidation or fear among a target audience.” (Jones & Libicki, 2008, p. 30)

Given the dispute surrounding a common definition of terrorism, I want to point out what the presented definition means and more importantly, what it does not mean. Firstly, terrorism according to this definition is limited to politically motivated violence, thereby excluding environmentally, religiously, or economically motivated violence (Hoffman, 2006). Jones and Libicki (2008) would probably push back and argue that many of these excluded forms of terrorist motivation include some kind of political agenda and are indeed included in their definition. Secondly, terrorism is directed against non-combatants i.e. civilians. This excludes all terrorism directed against infrastructure or objects causing no fatalities but are a common appearance (START, 2015). Thirdly, Jones and Libicki (2008) later specify that terrorism “is restricted to organisations other than a national government” (p. 3) and thereby exclude state terrorism, which they argue to be a different phenomenon requiring separate consideration. It is much debated if state-terrorism and non-state-terrorism are two different things, as in both cases terrorist strategies are applied but by different actors. A criterion this definition lacks but which would help to distinguish terrorist groups from rebel groups for example, is that terrorism is committed by organisations which have “an identifiable chain of command or conspiratorial cell structure (whose members wear no uniform or identifying insignia)” ((Hoffman, 2006), p. 40). In fact, Jones and Libicki (2008) cover a lot of groups in their collection that would be labelled ‘rebel group’ in different datasets. Finally, the purpose of terrorism is defined to be causing intimidation or fear among a target audience. This part of the concept is largely uncontested, sometimes specified though as: “[terrorism is] designed to have far-reaching psychological repercussions beyond the immediate victim or target;” (Hoffman, 2006).

Summarising, Jones and Libicki (2008) use a very broad definition of terrorism, which naturally cuts both ways in a disputed academic field. Readers should be aware of the fine differences and possible implications every included or excluded criterion described has for the understanding of terrorism. As next step, I will specify what is understood as disbandment of terrorist groups.

My concept of terrorist groups’ disbandments relies to a large extent on the conceptualisation of Jones and Libicki (2008) who introduce policing, military force, victory, or politicisation as possible ways of groups to end their terrorist operations. Policing includes all kinds of judiciary actions undertaken against terrorists as well as classical prosecution by police and intelligence services. Military force belongs to the same category of counterterrorism strategy but by other means, namely a state’s armed forces. The use of military forces against terrorists obviously constitutes a different league involving armed ground troops, air force, or the use of heavy armaments. Politicisation describes all kinds of transformation to non-violent political participation, including negotiations and peace agreements, running as political party, or engaging in civic activism. When terrorist groups achieve their goals and stop operating as a terrorist organisation, their end will be defined as victory.

Opposed to Jones and Libicki (2008), I also consider splintering as a type of ending. I argue that even though the members of a splintered group might keep on executing terrorist violence, the original terrorist group ceases to exist and therefore has to be considered ended in this analysis.^[5] A terrorist group’s disbandment is understood as splintered if a group divides into two or more groups because some members either have founded a separate organisation or they join another terrorist group. Also, the original terrorist group has to cease to exist in order to be considered as a ‘splintering’ case. If some members decide to leave their group and found a new one but the original group stays active, this organisation would not be considered as splintered.

Goal-structure & Criteria

As explained earlier, goal-structure in the context of terrorism groups has been introduced by Piazza (2008a) as an explanatory variable for suicide terrorism. Later, he developed the concept further in order to demonstrate that Islamist Terrorism is not more lethal than other kinds of terrorism but that the goal-structure of Islamist terrorist groups determines their lethality (Piazza, 2009). His work points in a very interesting direction but is not sufficiently theorised, even lacking a proper definition of the concept. Therefore, I will develop my own theory of goal-structure of terrorist groups based on Piazza’s groundwork.

First of all, I define goal-structure as a group of aims aligned in a priority system to achieve a final goal. In other words, goal-structure can be thought of as the overall structure of goals. Organisations and groups always rally around a common goal. But as is known from sociology and the research on organisational behaviour, there is an important difference between official and operative goals (Perrow, 1961). If we only consider the stated goal of a terrorist organisation, for example a territorial change, we cannot gain a full understanding of this organisation’s behaviour since it is largely shaped by operative goals. Those operative goals account for “the host of decisions that must be made among alternative ways of achieving official goals and the priority of multiple goals, and the many unofficial goals pursued by groups within the organisation” (Perrow, 1961, p. 855). Goal-structure contains both dimensions of aims^[6] and following, I will describe how the concept can be used on terrorist organisations as an analytical framework.

Therefore, I present the two archetypal types of goal-structure that prevail among terrorist organisations: a strategic goal-structure and an abstract goal-structure (Piazza, 2009). Groups with a strategic goal-structure have a very particular set of prioritised operative goals that lead them progressively to achieving their official (i.e. final) goal. Typically, their official aim is rather limited and tangible and they use terrorism as a strategy to achieve this aim. Their strategicness is indicated by their ability to pinpoint their demands, which are typically focused on a local level and do not exceed national borders. Since their operative goals are negligible and their official goal not as complex, the amount of decisions that have to be made in order to succeed is smaller than in case of an abstract goal-structure. They can define a roadmap more easily, which is meant to ultimately lead them to achieving their official goal.

On the other end of the scale, groups with an abstract goal-structure follow official goals outside the ordinary political discourse. They foster extensive changes in the political landscape and demand a profound change of polity. Their aims are usually much broader and overarching than those of groups with a strategic goal-structure. This necessarily leads to an increased amount of operative goals before the official final aim can be achieved, which in turn means that more decisions need to be made along the way. Since an abstract goal-structure brings about extensive goals, a terrorist group has to have a lot of staying power in order to achieve them. Therefore, these terrorist groups will have to prioritize operative goals much differently than groups with a strategic goal-structure. Long-term survival strategies need to be part of the overall agenda as well as achieving the official goal itself. They will use terrorism not only to achieve their official goal but as a strategy to spread their message and recruit new group members, too. Subsequently, I will illustrate the differences of those archetypes in more detail based on the criteria of goal-structure.

All goal-structure systems are characterised by their scope, extent, and effectiveness. In contrast to studies before, I do not take into account the actual content of a group’s aims in terms of ideology as justified earlier, but focus on the structure of those goals. That means I understand a goal system to be structured along the scope of aims (on what level do their demands focus?), the extent of aims (how much change is necessary to achieve the goals?), and the relationship between stated goals and actual behaviour, doing and saying, labelled as effectiveness (are the operative goals designed progressively towards a final goal?). Those three categories help to systematically understand the differences between strategic and abstract goal-structure and will be introduced in more details as follows.

Scope refers to the geographical dimension of a group’s aims. While some terrorist groups want to achieve final goals within their own community or country, other groups foster final goals that are overarching a number of countries like the Middle East and North Africa region or even aspire a new global order. While goals related to a group’s country or a certain region within that country are likely to be more concrete and tangible – and therefore indicating a strategic goal-structure – goals related to a whole region or even the entire world are likely to indicate a rather abstract goal-structure.

Extent of a goal defines how far off a final goal is from the initial status quo. This criterion is borrowed from Stedman (1997) who distinguished two dimensions defining the different types of spoilers: Their goals and their commitment to achieve those. The former can vary from rebel groups with limited goals to total goals. The same holds true for terrorist groups. While some only foster limited goals targeting a few if not even only one concrete policy field e.g. the health sector, other terrorist groups demand fundamental changes not only in several policy fields but in polity itself. Jones and Libicki (2008) covered this criterion only incompletely by looking at the breadth of terrorist groups’ goals. However, it is of the utmost importance that the extent of a terrorist group’s goal is considered when defining its goal-structure. Therefore, I conceptualise extent of a final goal distinct from the goal’s content by referring to the question: How far off are a group’s final goals from the current status quo? I expect goal-structures with a limited extent – that means closer to the current status quo – to be more strategic, while a total extent – far away from the current status quo – is more likely to indicate an abstract goal-structure.

Finally, e ffectiveness focuses on how target-aimed a group’s operative goals are towards their final aim. Do terrorist attacks account for a group’s official goal? Or are terrorist attacks only hardly connectable to what they claim their final goal is? Groups with a strategic goal-structure will align their operative goals more linear and progressively towards the achievement of their official goal. Since their final goal is not as extensive, they do not need to allocate major parts of their resources to recruit new members or spread a particular message, but can focus on the most effective way to achieve their goals. For terrorist groups with an abstract goal-structure, achieving their broad goals is much more difficult and requires a vast amount of operative goals that need to be reached before they succeed. Their goal-structure cannot be as effective since they also need to make sure that their long-term struggle remains funded and new members keep on joining the group. Terrorist groups with an abstract goal-structure might appear more random in the selection of targets and actions but they are highly rational, just not as effective.

All in all, goal-structure is defined as a group of aims aligned in a priority system, which is characterised by varying scope, extent, and effectiveness, to achieve a final goal. This newly introduced definition builds upon the empirical findings of Piazza (2008a, 2009) while providing further conceptual clarification. I will adopt his terminology of strategic versus abstract to describe the different extremes of the goal-structure spectrum. After defining and conceptualising my key concepts terrorism, goal-structure, and terrorists’ disbandment, I will now move on to my main arguments in order to develop my hypotheses.

Argument & Causal Mechanisms

What impact does the goal-structure of a terrorist group have on their disbandment? I argue that terrorist organisations with strategic goal-structures will end differently from those with abstract goal-structures – but why? For each kind of ending, I will develop one hypothesis each based on logical reasoning, starting with politicisation.

Terrorist groups with strategic goal-structures are limited in scope, extent, and effectiveness of their goals. I argue that they are more likely to end by politicisation than groups with an abstract goal-structure for two straightforward reasons that I will explain subsequently in more detail. Firstly, terrorist groups with strategic goal-structures have aims that are limited in geographical scope, which allows them to identify a direct negotiation counterpart, for instance a local or country government. Secondly, their limited aims are more likely to be within the bargaining range of this counterpart, which makes them more likely to initiate negotiations and successfully conclude them.

This is mainly due to the fact that groups with strategic goal-structures are highly rational actors and have a clear political agenda, which they want to achieve through the use of targeted and effective terrorism. This reasoning might lead to the question, why they are using terrorism as a strategy in the first place? Because through terrorism a group can raise attention to their matter and gain leverage, which they would have hardly attained without applying terrorism first. Motives for this are manifold, maybe there is no other way to participate or oppose a political system (e.g. in one-party autocracies) or their demands cannot be achieved through legal democratic structures (for instance releasing a prisoner) or they are simply unlikely to be achieved through traditional politics (if the terrorists for instance only represent a small minority).

Abbildung in dieser Leseprobe nicht enthalten

Ill. 1: Causal mechanism: strategic goal-structure leads to politicisation

As a consequence of their strategic goal-structure, these terrorist groups are capable of pinpointing their limited demands opposed to groups with an abstract goal-structure that brings upon large-scale goals including numerous necessary changes from the status quo. Since a strategic goal-structure is limited both in terms of extent and scope, they will find it easier to approach a negotiation partner and eventually settle an agreement. Their tactical behaviour works as a mechanism of strategic goal-structure: their matter is vaulted into the public sphere by committing terrorist attacks related to their demands. Now that they gained attention, they can address their grievances and threaten politicians with further attacks. In turn, politicians have two choices: Either crack down on the group or give in and find a compromise in order to stop the group from further targeting the civilian population. At this point the limited scope and extent of their aims comes in again, making the latter alternative less costly for politicians. I argue that terrorist groups which end by politicisation usually have to negotiate with the authorities to some extent in order to achieve some kind of agreement allowing them to join the political process. Since they committed terrorist attacks in the past, they would have to face prosecution if no agreement with the authorities is reached. Therefore, I conclude my first hypothesis:

H1: Terrorist groups with strategic goal-structures are more likely to end by politicisation

Adding to this, terrorist groups with strategic goal-structures are also more likely to reach victory. This is due to the already mentioned limited extent of their aims as well as a local scope and the effectiveness of their attacks. Building upon the previous argumentation, groups with strategic goal-structures could also aim for victory instead of compromising on their already limited aims. Instead of waiting for an offer to negotiate, they keep on committing terrorist attacks until e.g. the government yields to the pressure and implements their demands. Their local scope allows them to direct their demands towards an addressee, while effectiveness creates accountability regarding who committed an attack. It must be obvious to the recipient of a terrorist attack’s message who the sender is and most importantly, what their demands are.

If the terrorist group is not strategic in regards to their effectiveness, the recipient cannot implement their demands and their final aim will not be achieved. If their scope is global it would also be difficult for them to address their demands. A terrorist group with an abstract goal-structure is therefore unlikely to experience victory. All criteria, their global scope, total extent of aims, and no effectiveness, hinder them to achieve this success. No government would be willing to adapt their demands as long as they are total in extent. An example to illustrate this is a terrorist group that seeks to overthrow the current government or create a new state, e.g. a socialist or Islamic state. It is practically impossible with these demands to hope for any kind of victory unless the group achieves it themselves. This however, would require large-scale warfare comparable to the LTTE’s activities in Sir Lanka.

Abbildung in dieser Leseprobe nicht enthalten

Ill. 2: Causal mechanism: strategic goal-structure leads to victory

Consequently, my second hypothesis reads as follows:

H2: Terrorist Groups with strategic goal-structures are more likely to achieve victory.

In regards to splintering, I argue that again terrorist groups with a strategic goal-structure are more likely to splinter. This builds upon the previous argumentation that groups with a strategic goal-structure are more likely to negotiate and therefore need to compromise at least some of their goals. Groups with an abstract goal-structure are less likely to find themselves in this situation. When a group with a strategic goal-structure then decides to give in on some of their demands, some member might be left disappointed and frustrated. A section of the group bound to the compromised goals is likely to splinter away from the core group. It is even imaginable that the core group splinters from a moderate wing that is willing to negotiate. In any case, splintering is likely to occur if a group compromises their goals, which is almost a prerequisite for politicisation. I therefore conclude that it is more likely for a terrorist group with a strategic goal-structure to experience splintering. My third hypothesis therefore is: Politicisation

H3: Terrorist Groups with strategic goal-structures are more likely to end by splintering.

Abbildung in dieser Leseprobe nicht enthalten

Ill. 3: Causal mechanism: strategic goal-structure leads to splintering

As shown, it is almost impossible to give in or even make concessions for groups with an abstract goal-structure. Not because they are irrational but because their aims are total and their scope global, they will find it difficult to compromise for two simple reasons: Firstly, whom should they realistically talk to? Their demands are most likely outside the bargaining range of any government or representatives they could possibly negotiate with. Also, depending on their final aim, it might not even make sense to talk with a government, for example if they want to overthrow exactly this government to build a new order or if they want to found a new state on the territory of several other sovereign states. In conclusion, a government would also never comply with a terrorist group that has an abstract goal-structure – leaving few alternatives for the terrorist group. Secondly, if a group with an abstract goal-structure would compromise on their final aims, they would signal weakness and lose credibility towards competing and partner terrorist groups, their own members, support networks, and the general public.

Abbildung in dieser Leseprobe nicht enthalten

Ill. 4: Causal mechanism: abstract goal-structure leads to policing or military force

This kind of signalling towards several audiences is at the very heart of terrorism and therefore crucial to consider. Even if the group would have limited goals, it is still worth bearing in mind the signalling effects but they are by far not as damaging as for groups characterised by an abstract goal-structure. Maybe most importantly the effect on their own group members would be disastrous. Surrender the total aims and compromising them could lead to disintegration of the group and renegade of its members, maybe splintering. Just as important as current members are potential future members. Terrorist groups depend on an ongoing supply of new recruits, which will be jeopardised if the organisation signals any deviation from their stated goals.

Given the just described scenario, a terrorist group with an abstract goal-structure faces a dead-end situation. In order to adhere to their abstract goal-structure they must continue their operations until the final aim is reached. Otherwise they risk splintering or failure of the group. This in turn means for the government that they need to react with coercion and counterterrorism efforts. The costs of those measures will always remain lower for the government than giving in on a terrorist group’s total aims. I therefore conclude the following two hypotheses:

H4a: Terrorist Groups with abstract goal-structures are more likely to end by policing,

H4b: Terrorist Groups with abstract goal-structures are more likely to end by Military force.

Since every research rests on a number of important assumptions, I briefly want to conclude my argument by acknowledging the main assumptions this thesis is built upon. Firstly, I assume that one can observe the intentions and motives of a group by looking at their actions and in this case, particularly by looking at the terrorist attacks committed. I am aware that actions are only a proxy of one’s intentions but I argue that a group’s character is rather shaped by their deeds than their internal ideological agenda. The more attacks a group commits, the more comprehensive is the picture we can paint. This implies that groups which only commit one attack are more difficult to classify in terms of their motives. Secondly, I assume that these motives and the goals that result out of them do not change radically over time. I look at terrorist groups as constant organisations, fostering the same final aim over the course of their existence. It is possible that this final aim changes slightly but if a terrorist group would decide to seek an entirely different goal, this would violate one of my basic assumptions of how terrorist groups work. Since they are defined as ‘politically motivated’, I argue that group members would revolt against an about-turn of the group. Thirdly, I assume that the use of terrorism gives power to terrorist groups they did not have before. By killing civilians, terrorists gain attention from the general public and in particular from media and politicians. Media coverage creates pressure on politicians to solve the problem and react in one way or another. This mechanism provides a channel to terrorists, which they can use to spread their ideas or force politicians to take some kind of action.

Research Design

This chapter introduces the methodology applied in this paper. In order to answer my research question, I have chosen a quantitative approach for three reasons: Firstly, it allows me to investigate not only a small number of cases but the whole population of ended terrorist groups limited only by the chosen timeframe. By looking at an entire population I can guarantee to not leave out any important cases or information and avoid selection biases connected to all models of sampling. Secondly, by looking at the bigger picture, quantitative research – if executed in a proper manner – allows us to understand the underlying patterns that are driving numerous cases at the same time but remain unnoticed if only a small number of cases are investigated. It can provide us with more generalizable findings regarding trends than other approaches (Johnson & Reynolds, 2012). Finally, in the recent literature on terrorist’s disbandment qualitative case studies prevail and we notice a profound lack of quantitative analysis of the phenomenon.

Following, I introduce the datasets used and discuss implicit case selection along with biases of the data as well as the chosen timeframe. Subsequently, I develop indicators to measure my independent and dependent variable and discuss relevant control variables and their measures. Finally, I present and explain the statistical model applied in my study, concluding with some summarising words on the research question posed and the hypotheses to be tested.

Datasets, Case Selection & Timeframe

Technically, when studying an entire population it is not necessary to mention case selection. But as the definitions of who is a terrorist and what is terrorism are not common sense as of the present day, we need to be aware that by applying one kind of definition we implicitly ‘select’ a number of cases and exclude others. My definition of terrorism is driven by the data availability of the dependent variable. The only available dataset on terrorist groups’ disbandment, my dependent variable, to the day is the one published by Jones and Libicki (2008). The previously discussed Cronin (2009) dataset does not cover the types of disbandment for each group and Weinberg (2012) did not publish it at all. In the Jones and Libicki (2008) dataset they coded 648 cases of terrorist groups based on the RAND-MIPT Terrorism Knowledge Base (TKB). Unfortunately TKB was shut down on the 31st of March 2008 but parts of the data, in particular the terrorist organisation profiles, have been transferred to START at the University of Maryland and are made available through their website, although the data has not been updated since TKB’s shutdown in March 2008 (START, 2016). It raises doubts though that Jones and Libicki (2008) only coded 648 cases while the TKB by the time the book and its dataset were published contained a total of 856 cases. One explanation might be that Jones and Libicki’s (2008) definition of terrorism is more narrow than TKB’s^[7]. Another explanation could be that Jones and Libicki (2008) did not include any groups that were result of earlier splintering because they count the groups as ended once it splintered. As the authors do not mention this gap of 208 groups, we do not know for what reason they were excluded and if there was a particular (scientific or political) motivation behind this decision. It would be advisable to investigate those missing cases and add them to the Jones and Libicki (2008) dataset. Given the limited timeframe and word count of this thesis as well as my focus on the independent variable, I unfortunately cannot provide this service myself.

Out of 648 cases available I focus my research on those terrorist groups that have ended. Unlike Jones and Libicki (2008, p. 13), I argue that splintering is a way of ending for terrorist groups, since the unit of analysis in this study is terrorist group and the group as such stops existing as soon as it splinters. Not considered in my study are groups that have remained active: In line with my research question, I try to explain the variation within terrorist groups’ disbandments. It has to be acknowledged that this might lead to a selection bias by only taking into account groups that have ended while ignoring the active ones. As Reed (2000, p. 92) points out by using the example of escalation and onset of armed conflict, it can be dangerous to ignore the non-cases, because the researcher cannot study any possible variation between cases and non-cases and focuses only on the variation within one group of one type, namely ended terrorist groups in this case. Bearing this in mind, I would still not advance my study by including the active terrorist groups in my model. The stated research question focuses exclusively on the variation within ended groups and is therefore not interested in the variation between groups that have ended and that have not.

Jones and Libicki’s (2008) dataset contains information on name of the organisation, operating country, time of operation, its peak size, their economic capabilities, regime type based on the Freedom House index, a group’s ideology, their goals, and most importantly how they ended. Out of all 648 analysed groups, 244 remained active, 136 splintered, and 268 ended (Jones & Libicki, 2008, p. 35). An important constraint of this dataset is its timeframe: They started coding from the earliest point of data availability in 1968 until the latest data in 2006 before they published. However, it was not until 1998 before the RAND-MIPT TKB started to track domestic terrorist incidents. Therefore, from 1969-1997 they only included international, that means transnational, terrorist attacks, where one terrorist group strikes in a foreign country (RAND, 2016b). Since 1998 they also include domestic terrorist incidents and groups. Obviously, this is a major bias in the data, as Hoffman and Hoffman pointed out: “Scholars estimate that international terrorism accounts for from five to ten per cent of total terrorist events world-wide.” (1995, p. 180). In order to generate valid results, it is necessary to focus on a coherent population. Therefore, I seek to reduce the risk of selection bias by only taking into account the latest period from 1998 to 2006, where both international and domestic terrorist groups are included. Excluding the data from 1969-1997 is methodologically necessary although an unfortunate limitation since older groups are more likely to have ended in the meantime and therefore, would have been a great resource. In fact this means I have to dismiss 248 cases from before 1998 of which every single one has had ended or splintered – a considerable bias that would drive my findings if included. Another advantage of dismissing the data before 1998 is Rapoport’s (2004) argument that terrorism in the past has emerged in differentiable waves. He argues that from the late 20th century on we witness ‘the fourth wave of terrorism’ (or ‘religious wave’), which I can investigate on the side in terms of these groups’ disbandments since the timeframe from 1998 to 2006 represents mainly groups of this fourth wave. The chosen timeframe leaves us with a group of 155 terrorist groups that have ended before the Jones and Libicki (2008) dataset was published.

With regards to the data used for the independent variable, I will build a new dataset based on the Terrorist Organisation Profiles (TOPs) dataset provided by the START (2016), which is a replication of parts of the former TKB until its end in 2008. The TOPs dataset provides information on 856 terrorist groups including their mother tongue name, possible aliases, bases of operation, date formed, strengths, classifications in terms of ideological motivation, financial sources, and most importantly for the coding of my new dataset the founding philosophy as well as current goals. If the information from the TOPs dataset is insufficient to surely classify and code the variables, I will consult the RAND Database of Worldwide Terrorism Incidents (RDWTI) (RAND, 2016a). It has been used as one out of several sources to initially create the TKB and is therefore perfectly applicable to the data I need to generate. It contains short descriptions on terrorism incidents, sometimes including demands, claims of responsibility, or targets, which is particularly interesting and important to know in order to classify the three criteria scope, extent, and effectiveness correctly.

Besides the fact that the TOPs data is no longer updated, its greatest downside is that most terrorist groups are not covered by it due to the common behaviour of terrorist groups to act anonymously, stay under the radar, and not communicate their particular goals or founding philosophy at all (Abrahms, 2008, pp. 89–90). We cannot blame the creators of the dataset for this flaw but need to keep in mind the general challenges connected to the study (and in particular quantitative study) of terrorism. As Drakos and Gofas (2006b) pointed out, datasets on terrorism suffer from a major underreporting bias of terrorist incidents particularly in authoritarian states. This might not affect my study as much as studies focusing on numbers or casualties of terrorist incidents, but it is still important to bear in mind that terrorist groups from more authoritarian countries are likely to be underrepresented in all datasets used.

One final word on the datasets used in this study: START as well as the former TKB or today’s TOPs and the RDWTI, are all funded by the United States Department of Homeland Security. RAND is a think tank closely affiliated with the United States government administration in various ways. Their databases on terrorism were initially started on behalf of the U.S. government’s ‘Cabinet Committee to Combat Terrorism’, it has been used as template by the federal government, and RAND analysts keep on working closely with various bodies of the U.S. government (RAND, 2016a). Jones and Libicki are RAND researchers, funding their research through RAND and “independent research and development funds provided by the Department of Defense.” (2008, ii). Being aware that terrorism is a very political concept, it has to be clear that my research too is driven by the United States of America’s perspective on terrorism, their policies, and their interests. Most databases on terrorism are hosted by U.S. universities or think tanks all funded through Homeland Security, which makes it difficult to gain a quantitative understanding of terrorism beyond the U.S.’s perspective and of course grants those institutions a prerogative of interpretation. There is little I can change about these circumstances but pointing them out and bearing them in mind when looking at the results.

Operationalisation

Following, I will outline the operationalisation of my independent and dependent variable. I focus on the independent variable as I code the data on this side myself, while on the side of the dependent variable the operationalisation is based on Jones and Libicki’s (2008) groundwork. Finally, I will discuss possible confounders that I will control for and how they are operationalised.

I previously conceptualised my independent variable goal-structure as a group of aims aligned in a priority system, which is characterised by varying scope, extent, and effectiveness, to achieve a final goal. In order to measure the goal-structure of a terrorist group, I apply the three criteria of scope, extent, and effectiveness as indicators. Each indicator will be coded into a dummy variable that has two possible manifestations (0 and 1) and eventually aggregated to an overall measurement for goal-structure. In addition, this allows me to track the causal mechanisms for each hypothesis in detail. Following, I will explain the coding of each indicator in detail.

Scope is fairly straight forward: it will be coded 0 if the group’s goals have a local focus that means everything including the country-level downwards, for example goals that touch upon national policies or local autonomy for an ethnicity. On the contrary, scope is coded 1 if the goals are more globally-oriented, meaning everything above the country-level. This includes goals that demand the merging of countries for example. Often but not exclusively transnational terrorist attacks are an indicator of a global scope.

Extent will be coded 0 if the final aims of the group are limited and focus on policy changes or adjustments inside the political process. Limited goals are close to the initial status quo and involve all kinds of aims regarding national polices but also demands for prisoner release or extortion for ransom as long as they are not related to broader goals. Extent is coded 1 if the aims of a group are total, which means they foster a profound polity change (i.e. overthrow the system) or an insurgency against the current government, which also includes terrorist groups which want to replace the current leader of a state.

[...]

---

^[1] In fact, terrorist incidents between 2004 and 2014 increased by 13-times when only taking into account the attacks where there is essentially no doubt of terrorism according to the START (2015). Reasons for that are entirely speculative but might include the invasion of Afghanistan and particularly Iraq, the Arab Spring, and the rise of ISIL.

^[2] Today RAND-MIPT has been merged into the RAND Database of Worldwide Terrorism Incidents

^[3] In terms of goals of a terrorist organisation as explanatory factor, they measured the aims of terrorists on an ordinal scale ranging from narrow aims like a status quo to broader aims, e.g. social revolution or empire.

^[4] Ideology of a terrorist organisation is coded as either religious, nationalist, left- or right-wing.

^[5] In my terminology, whenever I speak of groups that have ended, I also include those that have splintered.

^[6] The words ‚aim‘ and ‚goal‘ are used interchangeably.

^[7] Indeed TKB defines terrorism in a broad sense: „Terrorism is violence, or the threat of violence, calculated to create an atmosphere of fear and alarm. These acts are designed to coerce others into actions they would not otherwise undertake, or refrain from actions they desired to take. All terrorist acts are crimes. Many would also be violation of the rules of war if a state of war existed. This violence or threat of violence is generally directed against civilian targets. The motives of all terrorists are political, and terrorist actions are generally carried out in a way that will achieve maximum publicity. Unlike other criminal acts, terrorists often claim credit for their acts. Finally, terrorist acts are intended to produce effects beyond the immediate physical damage of the cause, having long-term psychological repercussions on a particular target audience. The fear created by terrorists may be intended to cause people to exaggerate the strengths of the terrorist and the importance of the cause, to provoke governmental overreaction, to discourage dissent, or simply to intimidate and thereby enforce compliance with their demands.“ TKB (2008)