Accepted version

The terrorist attacks of 9/11, and subsequent terrorist acts around the world, have alerted social psychologists to the need to examine the antecedents and consequences of terrorist threat perception. In these two studies we examined the predictive power of demographic factors (age, sex, location), individual values and normative influences on threat perception and the consequences of this perception for behavioural change and close relationships. In study 1 (N = 100) gender, benevolence values and normative influences were all correlates of threat perception, whilst sense of personal threat was correlated with increased contact with friends and family. In study 2 (N = 240) age, sex, location, and the values of Openness to Change and Hedonism, all predicted threat perception, which in turn predicted behavioural change and relationship contact. Such findings point to the important role social psychologists should play in understanding responses to these new terrorist threats. Terror threat perception and its consequences in contemporary Britain. The terrorist attacks of September 11 2001, and subsequent suicide attacks in Africa, Russia, Spain and the Middle East, have alerted Western countries to the new threat posed by large-scale, co-ordinated terror attacks. The increasing global spread of the terrorist threat means that increasing numbers of people, previously remote from conflict areas, are now faced with threats either at home or when travelling or living overseas. However, research on this topic has been largely confined to the discipline of political science or military medicine, and has rarely investigated the part played by psychological factors in predicting threat perception or its consequences (Levant, Barbanel & DeLeon, 2004; Silke, 2003, 2004). A number of social psychological theories can be employed to help us understand threat perception and its consequences (Moghaddom & Marsella, 2004). However, with the exception of researchers working on Terror Management Theory (TMT), briefly reviewed below, few social psychologists have attempted to integrate and test these in a coherent model of terrorism and its consequences. In this paper we report two studies conducted in a country (Britain) widely regarded as a major terror target but which had not, at the time of writing, suffered from a major terrorist outrage. In doing so, we aimed to bring together several complementary theoretical approaches in an attempt to gain a greater insight into how individuals perceive and cope with this increasing threat. Perceiving terror Several individual and wider, group-level factors are likely to underlie how an individual perceives terror threats (Pyszcynski, Solomon, & Greenberg, 2003). In Study One, we consider the role of values, the influence of others (normative factors) and the role of sex in terror perception. In Study Two we add the influence of location and age as additional predictors of perceived threat. Values and threat perception. The study of values has re-emerged as a major topic for social psychological investigation in the past decade, with value systems seen as important super-ordinate cognitive structures with considerable implications for individual behaviour (Rohan, 2000; Schwartz & Bilsky, 1990). Research on trauma has suggested that an understanding of core personality and values may better predict variations in trauma symptoms than the actual severity of a trauma (Durodie, 2003). At the same time, traumatic life events can challenge an individual’s “assumptive world”, undermining implicit but fundamental beliefs and values (Janoff-Bulman, 1989). For example, work using the personality dimension of authoritarianism suggests that an increase in perceptions of the world as a dangerous and threatening place can lead to a desire for security and the development of authoritarian attitudes (Duckitt & Fisher, 2003). In the present research, we considered specific values taken from a widely used model circumplex model of values in assessing the role of values in predicting terror perception (Schwartz, 1992). Building on Rokeach ’s conceptualization of values (Rokeach, 1973), Schwartz et al. describe ten, individual-level value types which satisfy biological needs, interactional requirements and institutional and social demands for group welfare and survival (Schwartz & Bilsky, 1987). These values are organized in a quasi-circumplex format, and can be divided along two dimensions (see Table 1). Dimension One compares Openness to Change (which emphasises independent thought and action and change and is represented by values of Self-direction and Stimulation) with Conservation (which stresses submission and self-restriction and includes values of Conformity, Tradition and Security). Dimension Two contrasts a more egalitarian Self-Transcendence (represented by Universalism and Benevolence values) with Self-enhancement (the values of Achievement and Power — values which emphasize the pursuit of success and dominance over others). A tenth value, Hedonism, is related to both Openness to Change and Self-Enhancement (Schwartz, 1994). Those who hold strong Security values emphasise the safety of both their intimate relationships and the wider society. Such individuals might be expected to show the greatest fear of a terror attack in general. In contrast Stimulation values are the values most directly opposed to Security in the Schwartz circumflex model. Stimulation values emphasize a daring, varied life, which might be expected to correlate with behaviors, which involve exposure to, enhanced terrorism risk (e.g. traveling overseas to potentially ‘risky’ regions). We therefore anticipate a negative correlation between Stimulation values and general threat perception. Finally, Benevolence values concern the preservation of the welfare of intimate others. Previous studies have suggested the development of relatively strong, ‘benevolent’ ties following terror incidents (Vertzberger, 1997), and we predict a significant correlation between Benevolence values and threat perception, particularly the personal sense of threat that affects the individual or their immediate family. Partial support for these hypotheses was evident from two studies using the Schwartz value scale following actual terror attacks. Frink, Rose, & Canty (2004) reported significant increases in Security values following the Oklahoma terrorist bombing in the US. In a naturally occurring, “quasi-experimental” study, Verkasalo, Goodwin & Bezmenova (in press) compared the values of matched groups of Finnish school children and University students before and after the 9/11 attacks. In their study, Security values rose significantly in the aftermath of the attacks. In contrast, levels of ‘Stimulation’ were lower following the terrorist incidents. ** Insert table 1 about here ** Normative influences. Perceptions of terror threat do not exist in a ‘vacuum’, with those around us likely to influence our risk perceptions (Kaniasty & Norris, 2004). Almost seven decades of social psychological research have underlined the importance of group norms in determining the assessment of a situation (Sherif, 1936). Hatfield & Rapson (2004) describe a process of ‘emotional contagion’, in which people ‘catch’ other’s emotions, mimicking the emotional experiences of others around them. During times of enhanced, shared threat individuals may often seek to ‘socially share’ with others their anxieties in an attempt to reduce such anxiety, or may feel reassured by significant members of their social networks about the risks posed (Dumont, Yzerbyt, Wigboldus & Gordijn, 2003). The Theory of Planned Action (Ajzen, 1991) suggests group norms are also likely to be important in forming an intention to respond. We hypothesise a significant positive correlation between an individual’s perception of threat of attack and the perception of attack probability reported by his or her friends and family. Location. Perceptions of threat are likely to be at least partly based on actual risk, with this actual risk likely to vary across locations (Huddy, Feldman, Capleos & Provost, 2002). However, risk perceptions are also likely to be influenced by the difficulties involved in changing everyday behavioural patterns, and the control individuals have over how and where they spend their time. We examined the impact of location on the risk perception of students from a range of locations, a cohort whose workweek gave them some control over their travel schedule. Using a dissonance approach, we suggested that those who live and study in ‘high risk’ areas will find themselves in a ‘dissonant state’ where their desire for safety may clash with their potential ‘high risk’ habitat (Jonas, Greenberg & Frey, 2003). As a result we predicted that this group will downplay the perceived threat. We also predicted that those living outside of London (in our study in Oxford, 65 miles from Central London) would also be less likely to see themselves at risk. In contrast, we predicted that those living in suburban locations – within the boundaries of London but not in a central location – will have neither dissonance, or distance, reasons for downplaying the terror threat. We therefore hypothesise that this group will perceive the highest risk of attack. Age. Although there is little systematic research on this topic, Thomas (2003) reports relatively high rates of anxiety amongst his middle-aged sample following the September 11 attacks compared to other sample data that has examined reactions to this event. We tentatively suggest that older respondents will exhibit greater anxiety. Sex. Girls and women have been shown to report greater threat following traumatic events than boys or men (Norris, Friedman, Watson, Bryne, Diaz & Kaniasty, 2002; Raviv, Sadey, Raviv, Silberstein & Diver, 2000), and women have generally been reported to experience higher rates of anxiety than men following a terror threat (Huddy et al., 2002). Examining responses on the Perceived Stress Scale following the 9/11 attacks, Thomas (2003) found the 35-60 year old women in her sample remained distressed, fearful and vigilant some six months after the attacks. We therefore anticipate that women will experience greater anxiety about the possibility of a terror attack than men. Managing Terror: coping with the terror threats There are likely to be considerable individual differences in the manner in which individuals cope with terror threats (Silke, 2003). In our studies, we examined a series of possible behavioural changes that might arise from anxiety about an attack, and build on recent advances in Terror Management Theory to examine the way individuals may turn to close others when faced with enhanced anxiety. Behaviour change. A number of important ‘adaptive’ behavioural changes may follow an increased terror threat. These include delaying or cancelling air travel plans, restricting destinations for holidays and limiting the use of public transportation or places visited (Huddy et al, 2002). In the weeks following the September 11 attacks, 26% of Huddy et al’s American respondents had delayed plans to travel by air, and 18% said they were travelling into Manhattan less frequently. We predict that an increased sense of threat would lead to a reduction in using public transportation, a change in daily routine to avoid places perceived to be of high risk, and a reduction in planned air travel. Relationship interactions and terror management. As attachment theorists have long since noted, one important way of coping with personal threats to one’s safety and mortality is to seek support from others (Bowlby, 1969). Primary support is likely to be derived from romantic partners and friends and family (Lazarus & Folkman, 1984). Clinical studies have demonstrated that such support may be of particular importance during times of war or following the witnessing of a terrorist incident (Applewhite & Dickins, 1997; Kalicanin & Lecic-Tosevski, 1994, Putnik & Lauri, 2004). For example, in their study of those directly affected by bombings in Yugoslavia, Putnik & Lauri (2004) found interpersonal relationships became closer, with respondents reporting the provision and receipt of heightened levels of support. In their study of responses to the WTC attacks, Huddy et al (2002) reported that 31% had made adjustments to their daily routine in order to spend greater time with their families. Derived from the theories of cultural anthropologist Ernest Becker (e.g. Becker, 1973) and his theory of human motivation and behaviour, Terror Management Theory (TMT) has primarily focused on the ways in which individuals try to boost their self-esteem and cultural worldviews when faced with their own mortality (Greenberg, Solomon & Pyszcynski, 1997). According to this theory, a fear for one’s life can lead to paralysing terror and at least a partial “cognitive shutdown”. This in turn can promote a striving to maintain or enhance self-esteem and the limiting of information processing, often resulting in an exaggeration of enemy images in order to validate worldviews and help ‘simplify the world’. In a series of laboratory experiments, TMT theorists have shown that individuals contemplating their own mortality are more likely to try to reinforce ‘mainstream’ cultural viewpoints and to reject others who they see as threatening these viewpoints (Greenberg et al, 1997). Recently the importance of interpersonal support, and in particular the support provided by romantic partners, has been added to the theory (Pyszcynski et al, 2003). For example, Florian, Mikulincer & Hirschberger (2002) have argued that romantic relationships provide a sense of security that allows people to function with relative equanimity during time of personal existential concern. During time of stress, close relationships can act as a fundamental anxiety buffer, providing a “symbolic shield against the awareness of one’s finitude” (Mikulincer et al, 2003, p.37). Mikulincer et al (2003) argue that the formation of close relationships during periods of mortality salience functions “side by side and in interaction with other mechanisms” (p. 26). In particular, death awareness leads to a desire for long-term, committed and emotionally-driven relationships with significant others, and the avoidance of conflict with these others during times of anxiety. Indeed, some TMT theorists have argued that, when made aware of the threat of death, the desire for affiliation may override the need to maintain positive world views or self-esteem (Hirschberger, Florian & Mikulincer, 2003; Mikulincer et al, 2003). We therefore anticipate that the heightened sense of one’s own mortality from a terror threat will lead to greater closeness with romantic partners, and greater contact with family members and close friends. Combined Model and summary of hypotheses We combine these hypotheses into a model examining both predictors of threat perception and the consequences of that perception for behaviour change and relationship interactions. Because we anticipate that some predictors of threat (e.g. benevolence values), and some outcomes of threat perception (e.g. behavioural changes), would be most strongly related to perception of micro-fears that concern personal threat, we examine threat perception in two ways: (1) the macro-level perception of threat of an attack in general (general perception of threat), and (2) a more micro-level perception of threat of an attack directly affecting you or your family (personal risk). Overall, this model suggests that the demographic variables of age, sex, and location, normative factors and individual values are all predictors of threat perception. This threat perception in turn will predict reported behavioural change and relationship closeness. In Study 1 we hypothesised that an individual’s sex, individual values and group norms would all predict perceived probability of attack. Female participants, those high on the values of Security and Benevolence, and those whose friends and family perceived a greater risk, were all expected to perceive a higher risk of attack. In Study 2 we anticipated that older respondents, female participants, those living in the suburbs and those high on the values of Security and Benevolence, but low on Stimulation values, were likely to perceive the highest threat. In both studies we predicted that perception of threat would influence ‘adaptive’ behaviour change, that personal threat would be most strongly correlated with Benevolence values, and that a personal sense of threat would influence the quality and frequency of relationship interactions with romantic partners and family members.