1With theoretical concepts pertaining to inter-group relations, motivation, and self, and a research domain that comprises ideology, fanaticism, violence and aggression, social psychology seems uniquely equipped to shed light on some of the most pressing issues of our time: Why do people choose the path of terrorism? And, what makes people prefer the path of violence over more peaceful means to realize personal and collective goals? As a phenomenon, however, terrorism (not to be read as response to terrorism) appears to have received little attention in mainstream social psychological studies, except perhaps as a way to introduce more focused research topics, or to discuss the relevance of particular findings. This article deals with the experimental social psychological study of terrorism. The connection between struggle and violence will be designated as the key topic in the experimental psychology of terrorism. An exploratory study on this link will be reported.

Social Psychological Models of Terrorism

2In recent years, the emerging field of terrorism studies has generated a number of models that treat terrorism as a social psychological phenomenon and try to explain it as such. The ziggurat of zealotry, the staircase of terrorism by Moghaddam, and the pyramid model of McCauley are particularly influential.

3The ziggurat of zealotry has not been published yet. However, it influences the field, mainly because it is infused with years of experience of a senior CIA Analyst, Cindy Storer (e.g., Shainin, 2006). A “ziggurat” is a Mesopotamian pyramid, which is characterized by several clearly identifiable levels. This type of pyramid is therefore particularly suited to depict the radicalization process that plays a role in terrorism. There are five distinct layers. The base layer comprises that part of the population that to some extent identifies with the struggle, but does not actively engage, and is not ready to use violence for the cause. At the second level, one finds people who are actively committed to the struggle of the community, for example through community service, but who are not willing to use violence. The third level includes the group of individuals with more violent intentions. At this level, citizens become militants and aim to change the prevailing societal order by means of violence. On the fourth level, the collective struggle and its political background become subordinate to this use of violence. The fifth, top level of the ziggurat includes those individuals for whom the struggle has become equated with the objective of mass destruction and disruption.

4In the staircase to terrorism model (Moghaddam, 2005), the ground floor also comprises individuals who feel disadvantaged as a result of their group membership, and from that background, identify with a particular political ideology that conveys a message of collective struggle. The first level of the staircase entails a consideration of the options that effectively help to give shape to the struggle. Should mainstream political participation turn out to be ineffective or impossible, the model suggests it will increase displacement of aggression (floor 2) whereby feelings of anger and frustration are expressed through the use of force for the sake of the struggle. This aggression obtains moral legitimacy through the rhetoric of a (terrorist) organization, thus contributing to the transition to a third floor of the staircase model, where moral engagement is shaped and strengthened. On the fourth level, the conviction that one is engaged in a righteous struggle is further amplified particularly through enhanced categorical “us versus them” thinking. A final step removes all potentially inhibiting factors regarding the use of violence against fellow humans, enabling the unreflective use of violence as the prime means to advance the struggle.
In his general depiction of the “Jiu Jitsu” dynamics of terrorism, McCauley also emphasizes the pyramid structure of terrorism (e.g., McCauley & Moskalenko, 2008). The first assumption regarding the radicalization process is that the individual group members who are actively involved are only the top of a broader social movement. The basis of his terrorism pyramid is formed by a group of people who identify with a collective struggle or at least perceive the legitimacy and pertinence of the struggle. McCauley and colleagues’ depiction of terrorism essentially entails a pyramid with a larger base of support on the ground floor relative to a smaller number of people that emerge out of this base and who carry out violent attacks.

The Struggle-Violence Link

5These model-based representations of terrorism assume a lengthy, continuous developmental path that occurs within a social context and within a complex configuration of economic and political circumstances. Nevertheless, the models can well serve as basis for an experimental analysis of terrorism.

6The assumptions that are shared by all models seem particularly important in this regard. First, all assume that on all levels, there is a perceived value of a collective struggle. At the most basic levels of McCauley’s pyramid model, the ziggurat, and the ground floor of the staircase to terrorism, this awareness originates from identification with the group, and awareness that the group is unfairly deprived of opportunities or resources. A second common characteristic concerns the increasing extent to which this struggle is linked to violence. In all models, the ground floor has a low degree of linkage between the struggle and use of violence, while the highest echelons of the models assume a full equation of struggle and violence.

7Regardless of the model, the transition from floor to top may thus be depicted as a gradual process in which a struggle is increasingly associated with violence (cf. McCauley & Moskalenko, 2008). The models contribute separately to an understanding of the factors affecting the strength of this link. The models distinguish: 1. determinants of the use of violence as an option to give shape to the struggle (e.g., moral commitment, Martyrdom cults); 2. Factors that undermine the availability of alternative options to give shape to the struggle (e.g., lack of opportunity); 3. Factors that remove inhibition of the association between violence and struggle (e.g., training, infrahumanization). Accordingly, the foregoing integration enables a less historical and more systematic psychological analysis of terrorism.

8In the integrative depiction, the most important variable is the association between the struggle and violence. As people identify more strongly with the struggle, and the struggle is more strongly associated with violence, terrorism is more likely to occur (or in “pyramid terminology”, there is greater upward mobility within the pyramid structure). The strength of the association between struggle and violence is not only caused by factors that directly strengthen the association, but also by alternative associations with struggle (for example, violent protests) that may dilute the struggle-violent link, and the presence or absence of inhibitory factors.

9In many respects, this view has similarities with the “Tool” approach to terrorism provided by Kruglanski and Fishman (2006). The tool view explains terrorism as a means to achieve a political goal, and assigns a central role to the association between the political objective and terrorism as a means among alternative means. Furthermore, the presently advanced integrative model shares interests with a recent study by Bushman, Ridge, Das, Key, and Busath (2007), which demonstrated that if a struggle is of religious nature, participants are willing to use more violence than when the struggle is not of religious nature. The connotation of the struggle thus influences the extent to which people are willing to use violence.
In sum, there seems to be sufficient empirical and theoretical basis to designate the link between struggle and violence as a focal area in the study of terrorism. A laboratory setup was construed that aimed to assess the strength of association and aimed to identify the predictors of associative strength.

The Current Study

10A primed decision reaction time task was used for that purpose. An example is an earlier experiment by Payne (2001). To show that people tend to recognize weaponry faster when dealing with people with darker skin than people with white skin, he asked his participants to indicate as quickly and accurately as possible whether a weapon or tool appeared on a computer screen. Just before the presentation of the weapon or the tool, participants were very briefly exposed to a black or white face. Consistent with the reasoning that black faces are more strongly associated with weapons than white faces, it was found that subjects responded faster after they had been exposed to a black face than after a white face. The black face activates concepts associated with violence through the association between dark-skin and violence, thus making the participant more “ready” to recognize the weapon.

11In the current discussion, it is assumed that people vary in the extent to which engagement in a struggle implies engagement in violence. Here, it is not about white or black faces, but words associated with the struggle that should impact speed of responses to weapons or more generally, words related to violence. A more rapid response to violence related words after exposure to words associated with the struggle, compared to words that are not related to the struggle, would serve as an indication of the extent to which a participant associates struggle with violence.

12A first step towards validation of the instrument is an examination of the effects on this measure of the psychological states and characteristics that also have a predictive value in relation to terrorism. Although there is no such thing as a terrorist personality or a terrorist psychopathology, several scholars have emphasized that it requires a particular mindset to choose the path of terrorism. Jerrold Post, one of the world’s leading experts on terrorism, for example, has stressed that there is often a psychological basis for radicalization (e.g., Post, 2007). Post speaks in this context of splitting, whereby narcissistic wounds, for example as a result of childhood experiences, lead to anger that may incite militant ideological commitment. This analysis is consistent with several studies that have shown that the injuries inflicted on the self, such as through exclusion (DeWall, Twenge, Gitter, & Baumeister, 2009), through thoughts of their own death (see Motyl & Pyszczynski, 2009, this issue; Pyszczynski, Solomon, & Greenberg, 2003), or a fragile self-esteem (Bushman & Baumeister, 1998), often promote hostility and aggression. Instability of the self-concept, and narcissism may therefore be predictors of stronger association between struggle and violence.
Accordingly, the idea came about to conduct a study whereby participants filled out various personality questionnaires, answered questions about their childhood, high school, and college experiences, and completed a word recognition task in which they were instructed to respond as quickly and accurately as possible whether they saw a word or not, whereby several of the words were associated with violence. The personality scales were indicators of narcissism and self-stability. The questions on the youth, high school, and college year experiences served to identify people with positive and adverse experiences. The word recognition task served to determine the strength of the association between struggle and violence. Prior to the words, participants were briefly exposed to words associated with struggle (i.e., injustice, power, struggle). Among those who strongly associate struggle with violence, faster responses to violence-related words will be observed as a result of this exposure. The analysis laid out above assumes that youth experience, stability of the self, and narcissism, should predict such faster responding.

Results

Data preparation

18Before conducting our main analysis, participants’ input was first prepared. Six coders were asked to code all descriptions of childhood, high school, and college experience for valence (positive versus negative) and social nature (social versus non-social). An example of a positive social experience is a description of a relation with parents. An example of a positive non-social experience is the acquisition of money or a toy. A negative social experience is for example the divorce of parents or the experience of loneliness. A negative non-social experience is for example the experience of intense physical pain or suffering from a disease.

19Cases were assigned to a particular category on the basis of the coders’ majority decision. In 23 percent of the cases, there was perfect agreement between coders; in 34 percent of the cases, 5 out of 6 coders were in agreement; in 20 percent of the cases, 4 out 6 were in agreement; In 14 percent of the case, 3 of 6 coders were in agreement (none of the case involved a 3-3 split); In 9 percent of the cases, 2 coders agreed on one category, while two other coders agreed on another categories. These cases were discussed among coders until agreement was reached. Table 1 summarizes the number of assigned cases per category.

Table 1

Number of cases per condition

Social Non-Social Positive 60 38 Negative 19 11

Number of cases per condition

20Three steps were taken with regard to the reaction time data. First, all errors were eliminated from further analysis. Second, response latencies exceeding 2000ms were set to 2000ms. An index of the extent to which individuals associate struggle with violence was computed by dividing the response latencies to violence related words when preceded by struggle related primes by the response latencies to violence related words when preceded by non-struggle related primes. Higher scores indicate greater propensity to associate struggle with violence, although the absolute scores should be treated with caution, because a score exceeding 1 not necessarily implies a greater tendency to associate struggle with violence.

Primary Analysis

21The main interest was to examine the effect of youth, high school, and college experiences on the strength of association between struggle and violence related words. To that effect, a 2(Valence of Experience: Positive vs negative) × 2(Experience Type: social vs. non-social) between-subjects × 3(Experience Phase: Youth vs. High school vs. College) within-subjects Multivariate Analysis of Variance (MANOVA) was conducted with the struggle-violence index as dependent variable. This analysis yielded a significant 3-way interaction, F(2, 116)=4.04, p<.04. Tests of the between-subjects effects tests showed that Experience Type and Experience Valence only affected the response latencies in the Youth phase, F(1,124)=8.61, p<.01 for the Valence main effect, and F(1,124)=3.92, p <.07 for the Type x Valence interaction, but not in the High School, and College year phases, F’s(1,124)<1.23, ns. The relevant means of the three way-interaction are displayed in Figure 1.

Figure 1

Response index as a function of condition. Higher scores indicate greater propensity to associate struggle with violence

Response index as a function of condition. Higher scores indicate greater propensity to associate struggle with violence

22Follow-up simple effect tests were used to specify the nature of the two-way interaction found in the Youth Phase. Negative experiences were found to be more strongly associated with a propensity to associate struggle with violence than positive experiences. F(1,124)=5.88, p<.02. As the observed marginal two-way interaction indicates, this effect was most strongly pronounced among participants reporting negative social experiences rather than negative non-social experiences, F(1,124)=2.81, p<.10. Overall, then, consistent with Post’s wounded-self hypothesis, negative-social-youth experiences were shown to be the most strongly associated with a cognitive linkage of struggle with violence.

Auxiliary measures

23To further examine predictors of the struggle-violence link, composite scores were formed of the self-concept stability and narcissism measures, and correlated with the outcome of the response latency task. A significant negative correlation was found between self-concept stability and the struggle-violence index, r=-.19, p<.02. Here, lesser self-concept stability is associated with a stronger tendency to associate struggle with violence. No significant correlation was found between narcissism and struggle-violence association, r=.08, ns.

Conclusion and Discussion

24The aim of the present study was to identify factors that influence the strength of the association between struggle and violence. To that end, an instrument was developed that assessed the association between violence and struggle on the basis of response latencies. Students completed this task, and were also asked to answer questions about their experience during youth, high school and college. They also completed self-concept stability and narcissism scales. Participants who reported negative social youth experiences exhibited greater strength of association between struggle and violence. Also, greater self-concept instability was found to be marginally associated with stronger association between struggle and violence.

25Although obtained in laboratory situation that is by all accounts far removed from the real situation in which terrorism is enacted, the findings nonetheless appear to converge with some insights into the psychological dynamics at play in the world of terrorism. Although quite a number of terrorism scholars have emphasized the absence of a defining personality structure or developmental trajectory that is specific for terrorists, case descriptions frequently reveal some personal or vicarious trauma to underlie militant ideological ardor (e.g., Kruglanski, Chen, Dechesne, Fishman, & Orehek, in press; Speckhard & Akhmedova, 2006). The present findings clearly converge with these descriptions.

26Terrorism is a complex phenomenon that is difficult to study as a whole in an artificial environment such as a psychological laboratory. Nevertheless, the central link between struggle and violence is a defining component of many analyses of terrorism, and can be subjected to systematic study outside of the natural context in which it occurs. The classic survey of terrorism definitions by Schmid and Jongman of 1988 shows that “Violence” and “Politics” are the two most frequently mentioned elements. Furthermore, analyses of terrorism increasingly focus on the differences between violent and non-violent terrorist organizations (Karagiannis & McCauley, 2006; Smith, 2008; Dechesne, 2009). The “Pyramid” models, in their various forms, all imply that terrorism is best described as a process starting with identification with the collective cause and culminating in a firm belief in a linkage between the collective cause and the use of violence. Also, careful analysis of the history of terrorist organizations reveals that many came to existence after a split with more moderate forces over the use of violence to advance the cause. Importantly, the political decision-making processes involved in the choice for violence is not specific to a particular constellation of circumstances. While the informational content of the choice varies from case to case, the underlying structure of the process can be regarded as the same. In all cases, violence is increasingly seen as a relevant mean to advance the collective cause. In other words, the rule “IF Struggle then Violence” is increasingly becoming feasible during the radicalization process.

27Indeed, the present discussion could very well be framed in terms of the very general Unimodel of Social Judgment that assigns a central role to such IF-THEN production rules (e.g., Kruglanski, Dechesne, & Orehek, in press). The Unimodel assumes that judgment formation can be understood as a syllogistic reasoning process whereby convergence of major premise and minor premise affords a judgment regarding a particular topic. In the present discussion, the belief in a linkage between struggle and violence leads to the decision to engage in violence when confronted with an actual struggle. The unimodel further assumes that the degree to which the convergence of struggle and the belief in the rule that “struggle implies violence” depends on a host of factors that can be plotted on various parameters, including stimulus characteristics, processing capacity and capability, and directional and non-directional motivation. For example, the awareness of the struggle may not have sunk in sufficiently because of a lack of communication (i.e., a stimulus characteristics), or strategic considerations (i.e., directional motivation) may prevent full violent engagement.

28The key idea is here that independent of the case specifics, terrorism can be studied as a generic phenomenon when the link between struggle and violence is taken as study object. If one follows the suggestion proposed by many to place the link between struggle and violence at the heart of the analysis of terrorism, one can use a situation independent analysis to account for a key component in terrorism. In turn, once the analysis of terrorism has become independent of specific cases or instances, social psychological, experimental methodology may be of considerable use to elucidate the nature of the phenomenon.
In many respects, the present type of research converges with work by Asal, Lemieux, and Wilkenfeld (2008). They also investigated the extent to which people emphasize with the decision to use violence to address grievance felt by an (non-existing) ethnic group. They specifically examined the extent to which grievances and risks affected the extent to which participants were willing to use explosives to address injustices inflicted on the group. Among other findings, they showed that social dominance orientation (Pratto, Sidanius, Stallworth, & Malle, 1994) predicted the extent to which one is willing to use violence once decided to engage in social protest. They also showed that stronger grievances were associated with a stronger inclination to justify the use of violence, but not with an inclination to use violence to redress the grievance.
The current study and the study by Asal, et al. (2008) contribute to a field of the experimental psychology of terrorism. Such an approach may not only provide powerful illustrations of the willingness to use terror, but may be especially useful to identify psychological factors that put people at risk of engaging in terrorism. If, as I just argued, terrorism can be studied in the lab, it is also possible to examine how various roles, challenges, threats, personality, motivations, and emotions, affect the propensity to engage in political violence, behaviorally, psychologically, and physiologically.