Necesidades, familia y discapacidad intelectual

Treating the terrorist as a rational actor is not a new approach. Although the goals of a terrorist may be irrational, their actions will be guided by rationality. The rational choice perspective has been useful in understanding political violence including terrorism (Pape, 2005; Clarke and Newman 2006) and literature

consistently supports the presumption that terrorists are rational actors (Sandler et al, 1983; Enders et al., 1990; Enders and Sandler, 1999; Crenshaw, 2000;

Silke, 2001; Pape, 2005; Taylor and Horgan, 2006; Caplan, 2006). Committing an act of terrorism, whether under the guidance of a wider network or as a lone attacker, is a purposeful behaviour that is guided by rationality. Although the ‘best’

choice may not be taken, a deliberative process of thinking will have been engaged with.

Terrorists make carefully calculated choices that are value-maximising (Asal et al., 2009) with the intention of increasing their probability of success (Hoffman, 2006). There is evidence to suggest that there is a calculation of perceived risks in the selection of targets at the group and individual level (Mickolus, 1980;

Sandler et al., 1983; Sandler and Lapan, 1988). Airline hijackings and chemical, biological, radiological and nuclear (CBRN) incidents, which are associated with high risk, are among the lowest level occurrences amongst terrorist activities, whilst low risk incidents such as bombings and shootings are amongst the highest (Sandler et al., 1983; National Consortium for the Study of Terrorism and Responses to Terrorism, 2016). This implies that there is some consideration of risk by the offenders. Dugan et al. (2005) used a rational choice model to analyse the rewards, risks and costs associated with airline hijackings. They found that hijacking frequencies decreased after the installation of metal detectors and an increased presence of law enforcement in airport checkpoints, due to the increased risk of detection. When applying the rational choice perspective to terrorist incidents in Israel, Brophy-Baermann and Conybeare (1994) found that initial Israeli retaliations against terrorism led the terrorists to change their strategies as they expected further retaliations.

Considerations of issues like security, avoiding detection, and ease of access and escape are regularly engaged upon. Jenkins (1985) observed that terrorists spent a lot of time deliberating over targets and deciding which were the most vulnerable. This cost-benefit consideration was demonstrated in a recent analysis of terrorist autobiographies providing further support for the rationality of terrorists (Gill et al., 2018). One finding particularly relevant to this thesis was that “terrorists

often keep several potential targets in mind and choose the one with relatively fewest risks” (p. 5). The factors considered encompassed both subjective and objective factors and, in many ways, mirrored criminological findings related to criminal cost–benefit decision making. There were many depictions of how fear and nerves negatively impacted the decision-making processes in planning and carrying out an attack, which were reflective of findings from studies of urban criminals including street robbers, shoplifters, and burglars. These appeared to be most intense during the commission of an attack. Another prominent finding was the consideration of situational security features. They found the offender’s evaluation of security features at the target necessitates hostile reconnaissance.

This awareness of security factors often led to doubts and irregular behaviour at the target, increasing the risk of the terrorist being detected. Consistent with findings from studies of urban crime (Taylor and Nee, 1988; Butler, 2005; Nee and Meenagham, 2006; Bernasco and Jacques, 2015), perceptions of how effectively deployed security was important in this process.

Spatial analyses in the field of terrorism research do not fully reflect the advances made in the study of general crime. This thesis attempts to bridge that gap. Like urban crimes, terrorist attacks do not occur randomly in time and space. Both criminals and terrorists are subject to geographical constraints and other limitations associated with access to resources. Although the ideological underpinnings of their actions may be irrational, a terrorist’s decision-making process will follow some form of logic, and the locations they choose to attack will not be arbitrary. Therefore, analyses of this kind can inform strategies by identifying areas that may benefit from disruptive and preventive measures, such as SCP.

The first applications of spatial analyses to examine the distribution of urban crimes were mostly concerned with identifying patterns at the macro level, e.g. at state or nation level. This has been true too for the study of terrorism. For example, studies have analysed whether there is a contagious diffusion like element of terrorism (Midlarsky et al., 1980; LaFree, 2018), the transnational displacement of terrorism following 9/11 (Enders and Sandler, 2006), and the

clustering of attacks cross-nationally (Midlarsky et al., 1980; Braithwaite & Li, 2007; LaFree et al., 2012). The main limitation of studies at the macro level is that they assume that all space within each delimited geographic area is equally likely to experience the same amount of terrorism risk. This means the impacts of these attacks at a finer aggregation cannot be estimated. Whilst the results of macro-level studies have limited practical use, their findings of spatial clustering at the country level provide a great starting point for further spatial analyses at finer spatial resolutions within these countries (Li, 2005; Piazza, 2008).

Most research on the spatial patterns of terrorism has focused on meso-level analyses. Meso-level analyses are concerned with examining the space in between macro-level (national or international) and micro-level (individual) factors, and typically focus on sub-national regions and communities. In an analysis of group attacks in Israel, Berrebi & Lakdawalla (2007) found four key determinants of risk variation according to space, the most useful being that of proximity of terrorist operational bases. Similarly, Rossmo and Harries (2011) found that terrorist cell sites were clustered and found evidence for distance decay in a study of organisations in Turkey.

Most geospatial research is guided by the least effort principle (Zipf, 1965) which expresses that when considering a “number of identical alternatives for action, an offender selects the one closest to him in order to minimize the effort involved”

(Lundrigan and Czarnomski, 2006, p.220). When considering urban crimes an offender’s journey to their crime location typically demonstrates a distance decay function, where the frequency of offences decreases as the distance from the home increases. Proximity to a terrorist’s home location has shown potential to be a useful predictor of where an attack may take place for group-based terrorism. Cothren et al., (2008) found that just under half of group-based attacks occurred within 30 miles of the offender’s home location, while Clarke and Newman (2006) argue that “proximity to the target is the most important target

characteristic to terrorists” (p.154).⁸ However, within these studies, there have been very few attempts to disaggregate the data, meaning that all types of attacks are treated as a homogenous construct.

There is also a crucial temporal element to criminal activity. Whilst it has been long noted that victimisation increases the probability of future victimisation (Farrell, 1995; Pease, 1998), spatial and temporal analysis has provided the means to model the subsequent variations in risk. With regard to residential burglary, research shows that following one offense there is a temporary elevation in risk of further offenses at the same home and those nearby (Townsley et al. 2003; Johnson and Bowers, 2004a). In other words, the risk of burglary displays a contagious quality in terms of its space-time distribution.

Further research shows such variations in risk are largely ubiquitous, with similar patterns observed across different areas and within different countries (Johnson et al., 2007). For example, the risk of victimisation is similarly contagious in relation to motor vehicle theft (Lockwood, 2012), shootings (Ratcliffe and Rengert, 2008), assaults and robberies (Grubesic and Mack, 2008), and maritime piracy (Marchione & Johnson, 2013). In each case, an elevation in risk extends beyond the location of the original incident and then decays over time.

Like urban crime, these temporal variations are also evident in attack patterns in sustained conflicts. Hotspots of violence during violent campaigns have been identified and spatio-temporal trends of terrorism decay in similar manner to traditional crimes. Townsley et al. (2008) used the Knox (1964) test to analyse IED attacks by insurgents in Iraq. Attacks were non-random and were clustered in space and time. After an initial attack, a further attack was likely within 1km and within two days. Braithwaite and Johnson (2012) found similar results in their analysis of insurgent attacks alongside counter-insurgency operations. Insurgent attacks clustered, and there was an immediate increase in risk in the immediate vicinity of the attack, which sharply decreased after. Berrebi and Lakdawalla

8 A more detailed discussion of the distance decay function and analyses will be presented in chapter 3.

(2007) found that the risk of subsequent related incidents rose after an initial attack in Israel before returning to the baseline after approximately eight weeks.

Similarly, Marchione and Johnson (2013) found that following an initial incident of maritime piracy, the risk of a subsequent incident increased temporarily.

Behlendorf et al. (2012) found spatio-temporal clustering in attacks by Euskadi Ta Askatasuna (ETA) in Spain and the Farabundo Marti National Liberation Front (FMLN) in El Salvador. Terrorist attacks are not randomly distributed and factors such as the location of the attack and the time passed since the previous incident have been shown to help determine the location of future attacks (LaFree et al., 2012). Braithwaite and Johnson (2015) conclude that risk heterogeneity is an especially important factor when understanding spatial-temporal patterns of IED attacks.

Local infrastructure is another important element to consider as variations offer different opportunities, risks and rewards. However, a consideration of how the environmental backcloth of a city shapes the behaviour of terrorists has largely been neglected. Zhukov (2012) demonstrated the importance of road networks in a study of insurgent activity in North Caucasus and concluded that they were the most important determining factor for the location of attacks. Johnson and Braithwaite (2009) postulate that attacks by violent actors such as insurgents are concentrated in certain areas for tactical reasons, in an attempt to exhaust the resources of the opposition. The identifications of patterns such as these have implications for predicting where group attacks are likely to occur in future.

However, these studies neglect to explore how targeted locations differ from one another, i.e. why one location is chosen from a number of very similar discrete alternatives. Another weakness of these studies is that they assume underlying processes that determine the locations of the attacks are homogenous and there is no consideration of ideological factors that may shape targeting behaviours.