Understanding Aggression: Key Concepts, Theories, and Causes, Summaries of Psychology

Download Understanding Aggression: Key Concepts, Theories, and Causes and more Summaries Psychology in PDF only on Docsity! 8 Aggression Barbara Krahé KEY CONCEPTS aggression Aggression Questionnaire aggressive cues aggressive scripts anger management training bullying catharsis cognitive neo-associationist model Conflict Tactics Scales displaced aggression excitation transfer frustration-aggression hypothesis general aggression model geographic regions approach habituation heat hypothesis hostile aggression hostile attribution bias instrumental aggression intimate partner violence media violence–aggression link modelling peer nominations post-traumatic stress disorder sexual aggression steam-boiler model time periods approach trait aggressiveness weapons effect 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 156 CHAPTER OUTLINE This chapter presents an introduction to social psychological theory and research on aggression. After a brief discussion of how to define and measure aggression, we review the main theories of aggressive behaviour. This is followed by an analysis of individual differences in aggression and the role of situational variables, such as alcohol and high temperature, in eliciting aggressive behaviour. Special attention is devoted to the impact of violent media content on viewers’ aggressive tenden- cies. In the second part of the chapter, different forms of aggression in society are examined, such as family violence, sexual aggression, and both school and workplace bullying. The chapter con- cludes with a review of strategies designed to reduce and prevent aggression. Introduction Three-year-old Karolina had a short and sad life. She was found dying in a hospital toilet in 2004, dumped there by her mother after months of torture and abuse from the mother’s new partner. The list of his atrocities brought to light during the court case included tying her to a chair for hours on end, smashing a telephone over her head, and stubbing out a burning cigarette on her arm and then putting muscle warming cream on the burns to increase the pain (Der Tagesspiegel [German news- paper], 13 April 2005). In April 2002, Germany was shocked by an unprecedented school shooting in which 17 people, including the assailant, were killed. It was soon established that the 19-year-old killer, a former pupil at the school who had been expelled some weeks prior to the attack, had not only been fascinated by firearms but had also spent much of his time playing violent electronic games (BBC News, Friday, 26 April 2002). These two examples highlight the upsetting but undeniable fact that aggression as a destructive form of social behaviour is prevalent in human interactions on a large scale. Aggression permeates close relationships (e.g., child abuse and intimate partner violence), workplace interactions (e.g., bullying), intergroup relationships (e.g., gang violence and racially motivated aggression) and con- tacts between large-scale ethnic or political groups (e.g., international warfare). Therefore, social psy- chologists’ concern with understanding the processes that trigger, intensify or suppress aggressive behaviour is by no means a purely scientific one. Instead, it is motivated by the aim to create a knowledge base from which we can develop interventions to reduce and prevent aggression. The review of social psychological aggression research offered in this chapter is guided by five key questions: 1 How do social psychologists define aggressive behaviour and what are their main methods for studying it? 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 157 CHAPTER 8 AGGRESSION160 that aggressive energy is produced continuously within the organism until it is released by an external cue, such as the appear- ance of a rival in the contest for a mating partner. If the amount of energy rises beyond a certain level without being released by an ex- ternal stimulus, it will overflow, leading to spontaneous aggression. Psychologists have challenged Lorenz’s application of his find- ings from animal studies to human aggression. An important criti- cism is directed at the assumption that once the internal reservoir of aggressive energy has been used up by an aggressive act, it is impossible to trigger another aggressive response for as long as it takes the organism to rebuild a sufficient energy level. There is ample evidence that humans can perform several aggressive behaviours in quick succession and that one aggressive act often serves to precipitate rather than suppress further aggressive acts. (2) Researchers in the field of behaviour genetics examine the extent to which individual differences in aggressive behaviour can be linked to differences in genetic make-up (Plomin, Nitz & Rowe, 1990). Specifically, behaviour geneticists have sought to demon- strate that genetically related individuals are more similar in terms of their aggressive tendencies than individuals who are not genet- ically related. A meta-analysis of twin and adoption studies by Miles and Carey (1997) concluded that shared genetic make-up accounts to a significant extent for similarities in self-ratings as well as parents’ ratings of aggressiveness, explaining up to 50 per cent of the variance. However, an important qualification comes from studies that used behavioural observation as a measure of aggres- sion. In these studies, the impact of shared environment was substantially greater than that of genetic similarity. A subsequent meta-analysis by Rhee and Waldman (2002) also found substantial effects of genetic similarity, but the effects of environmental influ- ences were found to be even stronger. Thus, the evidence from a broad range of studies suggests that aggressive behaviour is af- fected both by genetic dispositions and by socialization experiences in the course of individual development. An individual’s genetic make-up may dispose him or her towards becoming an aggressive Table 8.1 Major theories of aggression Biological approaches Ethology Behaviour genetics Hormonal explanations Psychological approaches Frustration-aggression hypothesis Cognitive neo-associationist model and excitation transfer Learning theory Social cognitive approaches Aggression conceptualized as . . . . . . internal energy released by external cues; steam-boiler model . . . transmitted as part of genetic make-up . . . influenced by male sex hormones and cholesterol . . . as a likely response to frustration, likelihood enhanced by aggressive cues . . . as a result of affect elicited by aversive stimulation that is interpreted as anger . . . as a result of reinforcement, either direct or indirect (observed) . . . as a result of social information processing, enactment of learned scripts Data base Animal studies Twin and adoption studies Developmental studies Experimental studies Experimental studies Experimental + observational studies Experimental + longitudinal studies Empirical evidence No support as a model for human aggression, but still popular in lay discourse Support for the predictive value of genetic similarity Inconclusive evidence Supported by empirical evidence Supported by empirical evidence Supported by empirical evidence Supported by empirical evidence Plate 8.1 Our genes partly determine how aggressive we are, but so too does the social environment. 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 160 THEORIES OF AGGRESSION 161 person, but environmental factors play a crucial role in determin- ing whether that disposition will be reinforced or counteracted. (3) Another line of biological research on aggression is con- cerned with the role of hormones in relation to aggressive behavi- our. The dramatic increase in the male sex hormone testosterone in boys during puberty has been linked to an increase in the pre- valence of aggressive behaviour in this developmental period, but meta-analyses found only moderate positive correlations between testosterone and aggression among adolescent boys (Book, Starzyk & Quinsey, 2001). Cortisol has been examined as another hor- monal correlate of aggression, and results were also mixed: while some studies showed that low levels of cortisol were related to aggressive behaviour and conduct problems, other studies found high cortisol levels to be predictive of aggression (cf. Ramirez, 2003, for a review). Altogether, there is as yet no conclusive evid- ence that hormones such as testosterone and cortisol play a causal role in the emergence of aggressive behaviour patterns. Psychological approaches Early psychological models also assumed aggression to be an innate response tendency. Freud’s (1920) view of aggression as an instinct in the service of the pleasure principle inspired the frustration-aggression hypoth- esis, which regards aggression as driven by a desire to over- come frustration. Subsequent psychological approaches widened the frustration–aggression link into a more general model of negative affect and highlighted the role of cognitive factors, learning experiences and decision-making processes in predicting aggressive responses. The frustration-aggression hypothesis One of the earliest empirically tested theories about the origins of aggressive beha- viour is the frustration-aggression hypothesis (Dollard, Doob, Miller, Mowrer & Sears, 1939; Miller, 1941). It states that ‘frustra- tion produces instigations to a number of different types of re- sponse, one of which is an instigation of some form of aggression’ (Miller, 1941, p. 338). In this view, aggression is not the only but a possible response to frustration. Whether or not frustration will result in an aggressive response depends on the influence of additional variables in the individual or the environment. Fear of punishment for overt aggression or unavailability of the frustrator are factors that inhibit aggression. However, frustration that cannot be expressed in the form of aggressive retaliation against the original source is often ‘displaced’, i.e., directed at an innocent target person who is more easily accessible or less threatening. In a meta-analysis including 49 studies, Marcus-Newhall, Pedersen, Carlson and Miller (2000) found consistent evidence that frustrated individuals show displacement of aggression from the source of the frustra- tion onto a less powerful or more accessible target. If aggression is one of several potential consequences of frus- tration, it is important to identify the conditions under which individuals are likely to show aggressive behaviour when frus- trated. One variable shown to enhance the probability of an aggressive response to a frustration is the presence of aggressive cues. Aggressive cues are as- pects of the situation that draw the actor’s attention to the possibility of an aggress- ive response, such as seeing pictures of people fighting or being presented with the names of famous boxing champions. In a much-cited study, Berkowitz and LePage (1967) demonstrated that participants who had previously been frustrated by receiving negative feedback administered more elec- tric shocks (as a measure of aggression) in the presence of weapons, i.e., aggressive cues, than in the presence of a badminton racket, i.e., a neutral object. Although subsequent studies have not always replicated the effect – some failing to find a weapons effect and others finding an effect in non-frustrated participants as well – overall support for the role of aggression-related cues in facilitat- ing aggressive behaviour is impressive. From their meta-analysis of 57 studies, Carlson, Marcus-Newhall and Miller (1990, p. 632) concluded that ‘aggression-related cues present in experimental settings act to increase aggressive responding’. They also found an effect, albeit weaker, of aggressive cues on participants in a neutral mood state. The finding that the impact of aggressive cues is not limited to situations where the person is already in an angry mood suggests that aggressive cues have a wide-ranging potential to activate (‘prime’) cognitive schemata related to aggression and thus increase the salience of aggressive response options. PIONEER Neal E. Miller (1909–2002) was the architect of the frustra- tion-aggression hypothesis that laid the groundwork for sub- sequent socio-cognitive and neo-associationist models of aggression. He stressed that aggression is not inevitable but a likely response to frustration, and drew attention to the need to specify conditions under which frustration is likely to lead to aggression. Berkowitz’s model of aggress- ive cues and subsequent cognitive neoasso- ciationist view built upon his ideas. His views on displaced aggression were recently revi- talized and elaborated by Norman Miller and colleagues in their ‘triggered displaced ag- gression model’ (Miller, Pedersen, Earleywine & Pollock, 2003). frustration-aggression hypothesis assumes that frustration, i.e., blockage of a goal-directed activity, increases the likelihood of aggressive behaviour displaced aggression tendency to respond to frustration with an aggressive response directed not at the original source of the frustration but at an unrelated, more easily accessible target aggressive cues situational cues with an aggressive meaning that increase the accessibility of aggressive cognitions weapons effect finding that individuals who were previously frustrated showed more aggressive behaviour in the presence of weapons than in the presence of neutral objects 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 161 CHAPTER 8 AGGRESSION162 Cognitive neo-associationism and excitation transfer In his cognitive neo-associationist model, Berkowitz (1993) extended the frustration-aggression hypothesis into a more general con- ceptualization of the link between negative affect and aggressive behaviour. He ar- gued that frustration is just one type of stimulus that eli- cits negative affective arousal, and that other aversive stim- uli, such as pain or loud noise, may trigger aggressive responses in the same way. He proposed that aversive (unpleasant) stimuli give rise to unspecific negative feelings that evoke two immediate reactions, fight and flight. In a swift and automatic appraisal process that occurs with little or no conscious awareness, the fight impulse is associated with aggression-related thoughts, memories and behavioural responses, whereas flight is associated with escape-related responses. These responses serve to channel quickly the initially undifferentiated negative affect into the more specific emotional states of (rudi- mentary) anger or (rudimentary) fear. In a subsequent, more elab- orate and controlled appraisal process, the person interprets these basic or rudimentary feelings. They are considered in relation to the situational input and the person arrives at a more specific and consolidated emotional state, i.e., anger or fear. This cognitive processing also involves the evaluation of potential outcomes, memories of similar experiences and social norms associated with the expression of different emotions. Figure 8.1 illustrates this process. For example, when a child is hit by a stone thrown by a class- mate, he will immediately experience pain associated with nega- tive affect, probably a combination of anger, inducing the urge to fight, and fear, inducing the urge to run away. Depending on the context and the child’s past experience, either the anger or the fear response is likely to dominate and guide his further analysis of the situation. Before deciding how to respond, the child will engage in a more careful appraisal process, including an assess- ment of his classmate’s motives. If he concludes that his classmate threw the stone on purpose, the immediate feeling of anger will be consolidated, and retaliation will be considered as an appropriate response. Because all the components of the emotional experi- ence are associated with each other, activating one component is assumed to trigger other components relative to the strength of their association, hence the term ‘associationism’. The weapons effect described earlier can be explained in the context of this model as a result of aggression-related associations elicited by the Plate 8.2 Do guns make us more likely to behave aggressively? Unpleasant event (e.g., frustration, pain, social stress) Negative affect Primitive associational reaction Aggression-related thoughts, memories, physiological and motor responses Escape-related thoughts, memories, physiological and motor responses Rudimentary anger Rudimentary fear More elaborate thinking (attributions, outcome expectancies, social rules about appropriate emotion in situation, etc.) Differentiated feelings Irritation, annoyance or anger Fear PIONEER Leonard Berkowitz (b. 1926) is a key figure in aggression research. He has promoted both theoretical development and empirical evidence with respect to the role of negative affect and cognitive appraisal in aggressive behaviour. His study on the ‘weapons effect’ (Berkowitz & LePage, 1967) became a classic in social psychology. In this study, it was shown that people who were previously frustrated were more likely to show aggressive behaviour if aggressive cues, such as guns, were present in the situation. Subsequently, he developed his ideas into the cognitive neo-associationist model of aggressive behaviour. He also studied the other side of human nature, helping behavior. His book Aggression: Its Causes, Consequences, and Control, published in 1993, provides a comprehensive and authoritative review of aggression research. cognitive neo-associationist model explains aggressive behaviour as the result of negative affect that is subjected to cognitive processing and activates a network of aggression-related thoughts and feelings Figure 8.1 The cognitive neo-associationist model of aggression (adapted from Berkowitz, 1993, p. 57). 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 162 PERSONAL AND SITUATIONAL VARIABLES AFFECTING AGGRESSIVE BEHAVIOUR 165 PERSONAL AND SITUATIONAL VARIABLES AFFECTING AGGRESSIVE BEHAVIOUR Do people differ in their propensity to engage in aggressive behaviour, and what are variables associated with such individual differences? What are critical factors in the situation or the social environment that make aggressive behaviour more likely? In this section, we will take a closer look at some of the factors asso- ciated with differences between persons and between situations in the likelihood of aggression. In terms of the general aggression model, these are the input variables that are crucial in eliciting cognitive, affective and physiological responses that may or may not lead to an aggressive response. The guiding questions for this section are the following: how can we distinguish between more or less aggressive individuals and groups of individuals, and what situational influences of a transient or persistent nature precipitate aggressive behaviour? Individual differences in aggressive behaviour Researchers have suggested several variables as predictors of indi- vidual differences in aggressive behaviour. In the present section, we will focus on three of them that have received intense research attention: trait aggressiveness, hostile attributional style and gen- der (cf. Krahé, 2001, Ch. 3, for a coverage of additional person vari- ables related to aggression). Trait aggressiveness The concept of trait aggressive- ness describes dispositional, i.e., temporally and cross- situationally stable, differ- ences between individuals with respect to the likelihood of showing aggressive behaviour. Whereas some individuals are easy to anger and quickly get ‘hot under the collar’, others are generally less inclined to respond with aggression. Longitudinal studies fol- lowing the same research participants over many years from child- hood into adulthood have shown that the tendency to engage in aggressive behaviour is remarkably stable over time. Drawing on findings from 16 studies exploring the temporal stability of men’s aggressive behaviour, Olweus (1979) found a stability coefficient of r = .76 over a one-year period, of r = .69 over five years and still of r = .60 over a period of 10 years. These figures are matched only by the stability of intelligence scores over time and indicate that aggression in later stages of development may be predicted on the basis of earlier aggression scores. Interestingly, the stability was highest among those individuals who had very high scores and very low scores of aggression at the beginning of the measurement period, whereas individuals with moderate aggression scores at the beginning were comparatively less stable over time. Trait aggressiveness is conceptualized as a multidimensional construct comprising four different components: physical aggres- sion, verbal aggression, anger and hostility. It is typically assessed by self-report questionnaires in which participants indicate the like- lihood of showing different forms of physical aggression, verbal aggression, anger and hostility. The most widely used instrument is the Aggression Questionnaire (Buss & Warren, 2000; see above), but there are also instruments specially designed for adolescents (e.g., Orpinas & Frankowski, 2001). Hostile attribution bias Another variable linked to stable differences in the ten- dency to show aggressive behaviour is the hostile attri- bution bias. This construct refers to the tendency to interpret ambiguous behaviour by another person as an expression of the actor’s hostile intent. For example, in deciding whether or not another person causes harm accidentally or on purpose, individuals with a hostile attributional style prefer an attribution to hostile intent rather than seeing the actor’s behaviour as unintentional or caused by carelessness. The hostile attribution bias is typically measured by presenting short films or written scenarios in which one actor causes harm to another person, but the stimulus material is unclear as to whether the harm was caused by accident or on purpose (e.g., Dodge, 1980). For example, children are shown a video in which two boys build a tower of bricks. One boy then knocks down the tower, and the film is ambiguous as to whether he did so intentionally. The participants are asked to indicate if they think the child knocked down the tower by mistake or on purpose. Respondents who con- sistently prefer explanations that attribute the damage to the actor’s intent are seen as having a hostile attribution bias. Studies with adults demonstrate that individuals with a hostile attributional style are more likely to show aggressive behaviour and that differences in trait aggressiveness are predictive of the hostile attribution bias (Dill, Anderson, Anderson & Deuser, 1997). In a longitudinal study by Burks, Laird, Dodge, Pettit and Bates (1999), children who showed hostile attributional tendencies were also more likely to develop aggressive behaviour patterns. From this perspective, individual differences in aggression may be the result of schematic, habitual ways of information processing which highlight the hostile nature of social interactions and thereby lower the threshold for aggressive responses. To explain the development of the hostile attribution bias, sev- eral studies point to the role of exposure to violent media content. Correlational studies found a relationship between attraction to media violence and hostile attribution bias (Krahé & Möller, 2004). Other studies investigated whether hostile attributional styles are transmitted from mothers to their children. MacBrayer, Milich and Hundley (2003) found that mothers of aggressive children perceived more hostile intent and were more likely to report an intention to respond aggressively than mothers of non-aggressive children. However, mothers’ and children’s hostile attributions trait aggressiveness denotes stable differences between individuals in the likelihood and intensity of aggressive behaviour hostile attribution bias tendency to attribute hostile intentions to a person who has caused damage when it is unclear whether the damage was caused accidentally or on purpose 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 165 CHAPTER 8 AGGRESSION166 and aggressive behavioural intentions were found to be signific- antly correlated only for the girls, not for the boys. The authors explain this sex-specific effect with reference to the principle of learning by modelling, which states that similar models (here: models of the same sex) are more likely to be imitated than dis- similar models. Unfortunately, no studies have yet examined the correspondence between fathers’ and sons’ hostile attribution biases to substantiate this explanation. Gender differences A final variable associated with individual differences in aggression is gender, with the underlying hypo- thesis that men are more aggressive than women. Support for this hypothesis comes from the analysis of crime statistics across a range of countries, which show that men are overrepresented as perpetrators of violent crime at a ratio of about 8 to 1 (Archer & Lloyd, 2002). Meta-analyses of the psychological literature also found significant sex differences in aggression, with men showing more physical and verbal aggression than do women (Archer, 2004; Eagly & Steffen, 1986). However, despite being significant, the size of the effects is moderate at best, and smaller for verbal than for physical aggression. Cross-cultural analyses sug- gest that this is a general pattern across different societies (Archer & McDaniel, 1995). The picture changes somewhat when rela- tional aggression is included as a form of aggressive behaviour. Relational aggression is defined as harming others through pur- poseful manipulation and damage of their peer relationships (e.g., passing lies about someone to her friend, so that their relationship is harmed), and several authors have suggested that women may be as, if not more, involved than men in this type of aggression (e.g., Österman et al., 1998). Therefore, the ‘myth of the non- aggressive woman’ should be critically examined in the context of a broader range of behavioural types and contextual conditions of aggression (White & Kowalski, 1994). Situational influences on aggressive behaviour Just as it is clear that not all individuals respond with aggression in a given situation, it is clear that not all situations elicit aggressive responses to the same extent. In this section, we examine evidence concerning the role of three situational input variables that affect the occurrence of aggressive behaviour: alcohol consumption, high temperatures and exposure to violent media content. Alcohol From the evidence available to date, it seems safe to conclude that even moderate amounts of alcohol lead to increased aggressive behaviour. Alcohol plays an important role in the perpetration of violent crime, such as homicide (Parker & Auerhahn, 1999), domestic violence, including the physical and sexual abuse of children, sexual aggression and wife battering (Wiehe, 1998), and many forms of group violence, such as sports violence, rioting and vandalism (Russell, 2004). Experimental stud- ies show that alcohol has a causal effect on aggressive behaviour. These studies compare the aggressive responses of individuals who were given alcohol to those of individuals in a control condi- tion who did not receive alcohol. Two meta-analyses examined evidence from a wide range of studies comparing alcohol vs. con- trol groups and found that alcohol was a significant predictor of aggressive behaviour (Bushman & Cooper, 1990; Ito, Miller & Pollock, 1996). It is important to note, however, that general measures of the strength of the alcohol–aggression link mask the fact that the effects of alcohol may be strong for some people, but weak for others. For example, a recent study by Giancola (2003) showed that alcohol dramatically increased the administration of (supposedly) painful electric shocks to an opponent for individuals low in dispositional empathy (the ease with which people can adopt the perspective of another person), but failed to affect the behaviour of participants high in dispositional empathy (see Chapter 9, this volume, for more detail on empathy). In terms of explaining the effects of alcohol on aggression, the attentional hypothesis suggests that alcohol has an indirect effect on aggression by reducing the attentional capacity of the individual, preventing a comprehensive appraisal of situational cues (Laplace, Chermack & Taylor, 1994). As a result, only the most salient cues Plate 8.4 Mothers’ and children’s levels of aggression are significantly correlated only for girls. Plate 8.5 Are women really less aggressive than men? 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 166 PERSONAL AND SITUATIONAL VARIABLES AFFECTING AGGRESSIVE BEHAVIOUR 167 present in a situation receive attention, and if these cues suggest aggressive rather than non-aggressive responses, aggressive beha- viour is likely to be shown. This view is supported by evidence on the impact of aggression-related cues discussed earlier. High temperature Another situational input variable affecting aggressive behaviour is high temperature (Anderson et al., 2000). The heat hypothesis predicts that aggression should increase as temperature goes up (see Everyday Social Psychology 8.1). Two paradigms were de- veloped to test this hypothe- sis under natural conditions. The first paradigm is the geo- graphic regions approach com- paring violent crime rates in hotter vs. cooler regions, find- ing support for a link between hotter climates and higher violence rates in archival data. However, the regions included in the comparison, typically the north vs. the south of the United States, differed in aspects other than temperature, such as unemployment rates or normative beliefs condoning violence, that could be relevant to aggression. This potential alternative explanation is ruled out by the second para- digm, the time periods approach, which compares changes in violent crime rates within the same region as a function of fluctuations inPlate 8.6 Even moderate amounts of alcohol lead to increased aggressive behaviour. heat hypothesis hypothesis that aggression increases with higher temperatures geographic regions approach method for testing the heat hypothesis by comparing violence rates in cooler and hotter climates time periods approach method for testing the heat hypothesis by comparing violence rates during cooler and hotter periods EVERYDAY SOCIAL PSYCHOLOGY 8.1 The heat hypothesis and effects of global warming The heat hypothesis states that high temperature increases the likelihood of violent behaviour. The implications of studies sup- porting the heat hypothesis are worrying in the face of global warming. If increases in temperature are systematically related to increases in violent crime, then the continuous rise in global temperature presents a risk factor for the rise of violent crime. Based on archival data on the link between temperature and violent crime in the United States over 48 years from 1950 to 1997, Anderson et al. (2000) estimated the magnitude of this danger, as shown in Figure 8.3. Their analysis predicts that an increase in temperature by 2 degrees Fahrenheit increases the murder and assault rate by 9 cases per 100,000 people. For a US population of 270 million, this increase translates into 24,000 additional murder/assault cases per year. For readers more familiar with Celsius than Fahrenheit, with an increase in average temperature of just 1ºC, the murder and assault rate is projected to go up by 24,000 cases in the US. As Anderson et al. (2000) point out, it is important to bear in mind that temperature is only one of many factors affecting violent crime rates. However, it remains significant when other contributory factors are controlled for. Research on the heat hypothesis alerts both policy makers and the general public to the fact that the dangers of global warming are not restricted to our natural environment but also pose a threat to the social functioning of human communities. 0 2 4 6 8 Increase in average annual temperature (Fahrenheit) 50 40 30 20 10 0 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0In cr ea se in n u m b er o f m u rd er s/ as sa u lt c as es p er 1 00 ,0 00 p eo p le In crease in m u rd er/ assau lt cases (ab so lu te n u m b ers) Figure 8.3 Estimates of global warming effect on murders and assaults per year in the United States with a population of 270 million (based on Anderson et al., 2000, p. 124). 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 167 CHAPTER 8 AGGRESSION170 SUMMARY The research reviewed in this section shows that aggressive behaviour varies both as a function of person variables and as a function of situational context. Stable individual differ- ences in the propensity to act aggressively (trait aggressive- ness) and to interpret others’ actions as an expression of hostile intent (hostile attribution bias) predict differences in the ease with which aggressive responses are triggered in a particular situation. Research has also identified consist- ent gender differences, with men showing more physical aggression than do women. Some studies suggest that for relational aggression, such as social exclusion, the gender difference may be reversed, but more research is needed to consolidate this finding. Among the situational variables affecting the likelihood of aggressive behaviour, alcohol consumption, high temperature and exposure to media violence were shown with high consistency to lower the threshold for aggressive behaviour. In the case of media violence, longitudinal studies demonstrate that negative effects can be found over extended periods of time. AGGRESSION AS A SOCIAL PROBLEM Are there gender differences in the perpetration of intimate partner violence and sexual aggression? What is bullying and what do we know about the characteristics of bullies and victims? The theoretical and empirical contributions discussed so far identified critical input variables as well as mediating processes that explain the occurrence of aggressive behaviour. In this section, we will look at specific forms of aggressive behaviour between indi- viduals and between groups and discuss how the theories and find- ings examined so far can contribute to a better understanding of these social problems. Intimate partner violence Intimate partner violence is defined as the perpetration or threat of an act of physical violence by one partner on the other in the con- text of a dating/marital relationship. It is a serious problem across the world, even though the prevalence rates vary enormously not only between but also within countries (see reviews of the international evidence by Krahé, Bieneck & Möller, 2005; Krug, Dahlberg, Mercy, Zwi & Lozano, 2002). Mirrlees-Black (1999) found that 23 per cent of women and 15 per cent of men in the UK reported that they had experienced violence from an intimate partner at some point in their lives. In a Dutch study by Römkens (1997), 21 per cent of women and 7 per cent of men reported having experienced assault by an intimate partner at least once in their lives. One of the most contentious issues in this field of research refers to the question of whether men and women perpetrate intimate partner violence to the same or a different degree. Two main data sources are available to address the scale of inti- mate partner violence and the question of men’s and women’s involvement as perpetrators: (1) official crime statistics and crime victimization surveys of representative samples, and (2) research collecting self-reports of per- petration of, or victimization by, relationship aggression, using the Conflict Tactics Scales (Straus, 1979; revised version: CTS 2, Straus, Plate 8.7 Intimate partner violence is a serious problem across the world. Research looks at whether men and women are involved as perpetrators to the same or a different degree. intimate partner violence perpetration or threat of an act of physical violence within the context of a dating/marital relationship Conflict Tactics Scales instrument for measuring intimate partner violence by collecting self-reports of perpetration and/or victimization 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 170 AGGRESSION AS A SOCIAL PROBLEM 171 Hamby, Boney-McCoy & Sugarman, 1996). Official crime victim- ization figures show that a much greater proportion of women than men are victims of partner violence and that the rate of injuries from partner violence is higher for female than for male victims (e.g., Rennison & Welchans, 2000). Studies using the Conflict Tactics Scales, however, portray a different picture. In this measure, participants are presented with a list of minor (e.g., ‘I pushed or shoved my partner’) and severe (e.g., ‘I slammed my partner against a wall’) acts of physical aggression and asked to indicate whether and how many times they have shown the behaviour in question towards an intimate partner. A large body of evidence has shown that on the CTS women feature as much or even more in the perpetration of physical aggression towards a partner than men do. In a meta-analysis of 82 studies, Archer (2000) found no evidence of the overrepresentation of men in the perpe- tration of physical aggression. Instead, he concluded that women were slightly more likely than men to show physical aggression towards a partner. Critics have argued that the picture of gender symmetry por- trayed by studies using the CTS is largely due to the fact that this instrument records acts of violence without considering their con- text. It is now widely acknowledged by researchers that progress in the understanding of the dynamics of intimate partner violence will have to pay greater attention to the specific forms and contexts in which assaults on intimate partners take place (Frieze, 2000). Sexual aggression Sexual aggression includes a range of forced sexual activities, such as sexual intercourse, oral sex, kissing and petting, using a range of coercive strategies, such as threat or use of physical force, exploitation of the victim’s inability to resist or verbal pressure. It also includes un- wanted sexual attention in the form of sexual harassment, stalking and obscene phone calls (Belknap, Fisher & Cullen, 1999; Frieze & Davis, 2002). Official crime statistics show that sexual aggression is a large-scale problem. In Germany, 8,766 cases of rape and sexual assault were reported to the police in 2003, which corresponds to a rate of 10.6 per 100,000 citizens (Polizeiliche Kriminalstatistik 2003). UK crime statistics revealed that 9,743 rapes were reported to the police in 2001, corresponding to a victimization rate of 18.7 per 100,000 members of the popula- tion (Regan & Kelly, 2003). The majority of sexual assaults are committed by a perpetrator known to the victim, either as an acquaintance or as an intimate partner. Despite the persistence of the ‘real rape stereotype’ picturing rape as a violent surprise attack in a dark alleyway, sexual assaults by strangers are the exception rather than the rule. Complementing crime statistics that only reflect cases reported to the police, large-scale studies have been conducted to record sexual victimization of women by men. A summary of this data base is presented in Table 8.2. In contrast to intimate partner violence, it is undisputed that sexual violence is gender asymmetrical, with the vast majority of sexual assaults committed by male perpetrators against female vic- tims. However, it should be noted that sexual violence is also a problem in same-sex relationships (e.g., Krahé, Schütze, Fritsche & Waizenhöfer, 2000) and that women do show sexual aggression against men (Anderson & Struckman-Johnson, 1998; Krahé, Waizenhöfer & Möller, 2003). The consequences of a sexual assault on the victim are severe. A substantial number of rape victims develop the clinical symp- tomatology of post-traumatic stress disorder (PTSD). Victims re- experience the assault in dreams, images and intrusive memories, they try to avoid cues reminding them of the assault, and experience a general emotional numbness (Foa & Rothbaum, 1998). Contrary to a widely held public belief, assaults by partners and acquaintances are equally traumatizing for the victim as stranger assaults (Culbertson & Dehle, 2001). Victims of sexual aggression not only have to come to terms with the emotional trauma of the assault itself. They also have to cope with the reactions of others who learn about their fate. There is a widespread tendency to blame the victim of a sexual assault, unparalleled in judgements of victims of other criminal offences. A large body of evidence has shown that certain victim character- istics, such as low social status, higher number of sexual partners, pre-rape behaviour that is at odds with female role expectations, are linked to higher attributions of responsibility to the victim, and often correspondingly lower responsibility attributed to the attacker (Krahé, 1991). The tendency to hold victims responsible for being sexually assaulted is seen as a major factor in the low conviction rates for rape that have plagued the legal systems of many western countries (Temkin & Krahé, 2007). Table 8.2 Prevalence of men’s sexual aggression against women (based on Spitzberg, 1999) Form of sexual Women’s Men’s Number victimization/ victimization perpetration of studies aggression reports reports (%) (%) Rapea 12.9 4.7 63 Attempted rape 18.3 10.8 35 Sexual assaultb 22.0 8.9 40 Sexual contactc 24.0 13.4 28 Sexual coerciond 24.9 24.0 39 a Completed sexual intercourse through threat or use of force. b Penetration of the body through threat or use of force. c Sexual acts without penetration of the body through continued arguments, authority, force or threat of force. d Sexual intercourse through verbal pressure or abuse of position of authority. sexual aggression forcing another person into sexual activities through a range of coercive strategies, such as threat or use of physical force, exploitation of the victim’s inability to resist or verbal pressure post-traumatic stress disorder characteristic patterns of symptoms observed in survivors of traumatic experiences such as rape 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 171 CHAPTER 8 AGGRESSION172 Bullying in school and the workplace The last 25 years have seen a growing concern about aggressive behaviour in school and work settings (Olweus, 1994; Randall, 1997). Referred to by different terms, such as bullying, mob- bing or workplace aggression, this phenomenon denotes aggressive behaviour directed at victims who cannot easily defend themselves (Smith, Ananiadou & Cowie, 2003). Bullying typically carries on over extended periods of time and involves a power differential between bully and victim based on physical strength or superior status that undermines the victims’ ability to defend themselves or retaliate. Forms of bullying include physical, verbal and relational aggression, i.e., behaviour directed at dam- aging the victim’s peer relationships. The typical victim is an anxi- ous, socially withdrawn child or adolescent, isolated from his or her peer group and likely to be physically weaker than most peers. In contrast, bullies are typically strong, dominant and assertive, showing aggressive behaviour not just towards their victims but also towards parents, teachers and other adults (cf. Griffin & Gross, 2004, for a comprehensive review). Boys feature more promi- nently than girls as victims as well as perpetrators of bullying (Olweus, 1994). They are also more likely to use physical aggres- sion than are girls, who rely more on verbal and relational forms of aggression, as shown in a cross-national comparison involving 21 countries (Smith et al., 1999). Workplace bullying has only recently become the object of sys- tematic research, and empirical evidence is still limited. Like school bullying, the core of the construct refers to behaviours intended to make another person feel miserable at work over longer periods of time, with the target persons being unable to defend themselves due to an imbalance of power between perpetrator and victim. According to Hoel, Rayner and Cooper (1999), both the prevalence and the nature of experienced bullying in the workplace are similar for men and women. However, women appear to be more negatively affected by bullying than men. A large-scale study by Smith, Singer, Hoel and Cooper (2003) explored potential links between individuals’ experience of bullying at school and at the workplace. A sample of more than 5,000 adults employed by a wide range of companies in the United Kingdom completed a measure of experience of workplace bullying and provided retro- spective reports of bullying victimization while at school. Thirty- three per cent of participants identified themselves as victims of school bullying, and 25 per cent reported that they had experienced workplace bullying in the last five years. A significant association was found between school and workplace bullying: respondents victimized at school were more likely to have been bullied at work in the last five years than respondents who had not been bullied at school. It is important to note, however, that the relationship was inferred on the basis of retrospective reports of school bullying that may have been inaccurately recalled or distorted in the light of subsequent experiences of bullying in the workplace. SUMMARY Intimate partner violence, sexual aggression and bullying are widespread forms of aggression in everyday life. They can lead to lasting negative effects on the victims’ psycho- logical functioning and well-being. In research on intimate partner violence, the issue of whether men or women fea- ture more prominently as perpetrators is controversial, but there is consistent evidence that women are more likely to be injured by an intimate partner than are men. Sexual aggression is perpetrated mostly by men against women. Bullying in school and the workplace is characterized by a power differential between perpetrator and victim. Some studies suggest that experiences of being bullied in school make victims vulnerable to subsequent workplace bullying. PSYCHOLOGICAL PREVENTION AND INTERVENTION: WHAT CAN BE DONE ABOUT AGGRESSION? Is there evidence to support the popular catharsis hypothesis, i.e., the notion that releasing aggressive tension through symbolic action reduces the likelihood of aggressive behaviour? What are viable strategies to reduce individuals’ tendencies to show aggressive behaviour? Plate 8.8 Bullying, either in schools or in the workplace, denotes aggressive behaviour directed at victims who cannot easily defend themselves. bullying denotes aggressive behaviour directed at victims who cannot easily defend themselves, typically in schools and at the workplace 9781405124003_4_008.qxd 10/31/07 3:04 PM Page 172