Individual Differences in Alcohol-Related Aggression

Acute alcohol consumption facilitates aggression for some, but not for all, persons. An important task for researchers in this area is to determine in whom and in what contexts alcohol is most likely to lead to aggression. Over the last 5 years, the PI (Dr. Giancola) has been directing an NIAAA-funded (ROI-AA-11691)research program aimed at isolating key individual difference and contextual factors responsible for alcohol'seffect on aggression. The current proposal is a competitive renewal application to fund the continuation of thisprogram of research. The overarching goal of this research program is to create a "risk profile" to identify: 1)who will, and who will not, behave aggressively when intoxicated and 2) what contexts/situations are most likely to elicit intoxicated aggression. A more distant goal of this work is to utilize its results to influence treatment and prevention interventions for alcohol-related violence. The immediate goal for the proposed funding period is to elucidate the role of 3 individual difference variables (i.e., beliefs about aggression, beliefs about alcohol, and dispositional empathy) and 2 contextual variables (i.e., induced empathy and provocation) on the alcohol-aggression relation for men and women. This work is important because it will identify risk factors for intoxicated aggression, it will improve current theories of alcohol-related aggression, and it will help clinical researchers to develop more effective empirically-based treatment and prevention interventions for alcohol-related violence. B.2. Relation Between Alcohol and Aggression: An Overview B.2.I. Survey Studies. Survey research has determined that alcohol is present in about 50% of violent crimes (Lunetta et aI., 2001; Pemanen, 1991; Wells et aI., 2000). Data from the National Crime Victimization Survey indicate that alcohol was present, during the time of the transgression, in 63% of intimate partner violence incidents, 39%-45% of murders, 32%-40% of sexual assaults, and 45%-46% of physical assaults (Greenfeld & Henneberg, 2001). Furthermore, a review of26 studies carried out in II countries corroborated these findings by demonstrating that 63% of violent criminals committed their offenses while under the influence of alcohol (Murdoch et aI., 1990). Survey studies have also determined that it is the acute effects of alcohol, rather than its chronic effects, that have the greatest impact on aggressive behavior (e.g., Fals-Stewart, 2003; Wiley & Weisner, 1995). By design, survey studies do not manipulate alcohol consumption and they typically measure violent behavior in a post-hoc fashion either by self-report methods or by accessing archival data. Such studies have contributed invaluable information to the alcohol and aggression literature (e.g., Leonard & Quigley, 1999; Murphy & O'Farrell, 1996). However, given the documented importance of the acute effects of alcohol on aggression, another approach that would provide distinct, yet complementary, data to those of survey studies would be to measure the acute effects of alcohol on in vivo physically-aggressive behavior. 16 Giancola, Peter R. B.2.2. Laboratory Studies and Measurement of Aggression. Laboratory tasks that can measure the acute effects of alcohol on "live" aggressive behavior have been used for years. One of the most widely used, and well-validated, laboratory measures of physical aggression is the Taylor Aggression Paradigm (TAP; Taylor, 1967). The TAP operationalizes aggression as the administration of mild electric shocks to a fictitious opponent during a supposed competitive interpersonal task (Taylor, 1993). This operationalization conforms very well to Baron and Richardson's (1994) well-accepted definition of aggression which states that aggression is " ... any form of behavior directed toward the goal of harming or injuring another living being who is motivated to avoid such treatment. "(pg. 7). Similar definitions have also been advanced by other leading aggression theorists (Anderson & Bushman, 2002; Berkowitz, 1993; Geen, 2001). In addition to the above survey studies, a large number of laboratory investigations also report a robust relation between acute alcohol consumption and aggressive behavior. Over 60 studies using the TAP, its modified versions, and other similar laboratory aggression paradigms, have found that persons who receive alcohol are more aggressive than those who receive a placebo or a nonalcohol control beverage (reviewed in Bushman & Cooper, 1990; Chermack & Giancola, 1997; Ito et aI., 1996; Taylor & Chermack, 1993). B.3. Theories of Alcohol-Related Aggression A number of models have been advanced to explain the mechanisms through which alcohol facilitates aggressive behavior. Three of the most well-accepted models contend that acute alcohol consumption facilitates aggression by 1) disrupting cognitive functions involved in maintaining inhibitory control over behavior (Giancola, 2000; Steele & Josephs, 1990), 2) inhibiting fear responses (lto et aI., 1996; Pihl et aI., 1993), and 3) increasing psychological and physiological arousal (Giancola & Zeichner, 1997; Graham et aI., 1997). B.3.1. What _D_oWe Know? As noted above, alcohol consumption facilitates aggression for some, but not for all, persons. Meta-analytic studies indicate that alcohol has a "medium" effect size (d = .49 to .61) on aggression (Bushman, 1993; Bushman & Cooper, 1990; Ito et aI., 1996). However, it is argued herein that this effect size masks alcohol's true effect on aggression by not taking into account key moderating factors. In other words, alcohol is purported to have no effect on aggression for some persons but a very strong effect for others. Thus, it should be made clear that the models described above only pertain to certain "at risk" individuals. Accordingly, more general models of alcohol-related aggression have argued that alcohol is most likely to facilitate aggression for persons who are already predisposed to behave in such a manner (Collins, 1988; Graham et aI., 1997; Pemanen, 1991). B.3.2. What _D_o_n_'t We Know? Very few studies have attempted to examine how various individual difference factors affect the alcohol-aggression relation. In fact, Lang and Sibrel (1989) noted that "Given that the legacy of psychology is an inclination to focus on individual differences, it is remarkable that they have been so neglected where the alcohol-aggression relation is concerned" (pg. 315). Unfortunately, almost 10 years after the publication of Lang and Sibrel, Graham and colleagues (1998) noted that it is still the case that "the study of how individual differences moderate and mediate the relationship between alcohol and aggression is an area of practical relevance that has not been studied a great deal" (pg. 663). Clearly, researchers can significantly advance this field of study by developing a "risk profile" that will identify who is most (and least) likely to behave aggressively when intoxicated. BA. Building A "Risk Profile" Based on the available research literature, this project advances 3 individual difference (i.e., beliefs about aggression, beliefs about alcohol, and dispositional empathy) and 2 contextual (i.e., induced empathy and provocation) variables that are hypothesized to be risk factors for intoxicated aggression. The individual difference risk factors for this project were chosen according to 3 criteria: 1) Empirical Support, 2) Theoretical Support, and 3) Treatment/Prevention Applicability. 17 Giancola, Peter R. 1) Empirical Support. As will be discussed below, the individual difference risk factors proposed in this project have all been repeatedly shown to be associated with aggressive behavior. 2) Theoretical Support. As noted earlier, general models of alcohol-related aggression postulate that alcohol is most likely to facilitate aggression for persons who are already predisposed to behave in such a manner (Collins, 1988; Graham et aI., 1997; Pemanen, 1991). Based on the empirical evidence establishing a relation between the 3 proposed individual difference risk factors and aggression, more specific theories of alcoholrelated aggression have predicted that alcohol is more likely tofacilitate aggression for a) persons with accepting/permissive beliefs about aggression (Giancola, 2002a), b) beliefs that drinking alcohol causes aggression (MacAndrew & Edgerton, 1969; Marlatt & Rohsenow, 1980), and c) low dispositional empathy (Giancola, 2000), than for persons without these traits. 3) Treatment and Prevention Applicability. Understanding why a person becomes aggressive when intoxicated is obviously important. However, equally important is utilizing that information to treat or prevent future alcohol-related violence. Each of the 3 proposed individual difference risk factors are clearly amenable to modification. Specifically, well-established cognitive-behavioral techniques (Beck & Freeman, 1990; Kazdin et aI., 1992) can be applied to alter erroneous beliefs that violence is an acceptable and desirable form of behavior and that drinking alcohol directly causes violence. Moreover, there is also a literature demonstrating that "empathy-training" is an effective method for increasing prosocial behaviors and decreasing aggression in children and adults (see below). Although this is not a treatment project, the findings from this investigation will have the potential to impact treatment and prevention interventions. Individual Difference Risk Factors B.4.1. Beliefs About Aggression. Huesmann (1998) recently put forth a social information processing model of aggression. One aspect of this model logically posits that beliefs about aggression will influence the expression of aggression. In this context, beliefs refer to an individual's ideas about the acceptability or unacceptability of a particular behavior. Such beliefs will then determine the individual's behavior by identifying which responses are permissible and which are not. Recent research has shown that beliefs about aggression, whether they refer to general beliefs (e.g., "it is always okay to hit others") or situation-specific beliefs (e.g., "it is okay to hit others if they hit you first"), are significantly related to the expression of aggression (Guerra et aI., 1995; Huesmann & Guerra, 1997; Zelli, et aI., 1999). Based on Huesmann's research as well as general models of alcohol-related aggression (Collins, 1988; Graham et aI., 1997), Giancola (2002a) postulated that beliefs about the acceptability/permissiveness of aggression will moderate the alcohol-aggression relation. No study has tested this hypothesis. Nevertheless, the PI has collected preliminary data showing that beliefs about the acceptability/permissiveness of aggression are positively related to aggression on the TAP in the sober state (Giancola, unpublished data). Given the promising preliminary empirical data and theoretical support, systematic study of beliefs about the acceptability/permissiveness of aggression as a moderator of the alcohol-aggression relation is warranted. B.4.2. Beliefs About Alcohol (aka, alcohol expectancies). According to alcohol expectancy theory, intoxicated aggression is engendered by one's belief that drinking alcohol causes aggression (MacAndrew & Edgerton, 1969; Marlatt & Rohsenow, 1980). However, the empirical data show that the effects of alcohol expectancies on intoxicated aggression are equivocal (reviewed in Leonard, 2002). Specifically, whereas some self-report studies have shown that the relation between alcohol and aggression is stronger among persons who expect alcohol to increase aggression than among those who do not hold this belief (Derman & George, 1989; Leonard & Senchak, 1993), other research does not support this finding (Quigley & Leonard, 1999). 18 Giancola, Peter R. Furthermore, the only study that assessed the influence of beliefs about the effects of alcohol on intoxicated aggression also produced equivocal results (Chermack & Taylor, 1995). Under conditions of high provocation (not low or medium), intoxicated subjects who believed that drinking alcohol causes aggression were more aggressive than those who did not hold this belief. However, this was not a reliable finding in that it was observed for only lout of 3 dependent variables and only under lout of 3 levels of provocation. A variety of reasons can account for this inconsistent finding such as low statistical power brought on by using to few subjects and/or dichotomizing the continuous variable of beliefs about alcohol. Another problem might have been the restricted variability of the beliefs about alcohol measure (it only had 3 items). Finally, following Reviewer 3's comments, we conducted another pilot investigation (N = 24; only men) similar to the Chermack & Taylor (1995) study described above. Results indicated that alcohol only mildly increased aggression for men who believed that alcohol causes aggression. However, this effect was observed only under conditions of high, but not low, provocation. Conversely, alcohol similarly increased aggression for men who did not believe that alcohol causes aggression under low, but not high, provocation. Given all of the inconclusive findings reviewed in this section, further research (with proper sample sizes and both genders) is clearly required to elucidate the interactive effects of acute alcohol consumption and beliefs about alcohol on aggressive behavior. BA,3. Dispositional Empathy. It is well established that the construct of empathy contains both cognitive and affective components (Davis, 1983; Miller & Eisenberg, 1988). From a cognitive perspective, empathy has been defined as the ability to take another person's psychological point of view (Davis, 1983) and the ability to comprehend their situation as well as their cognitive and affective status (Hoffman, 1984). From an affective perspective, empathy has been defined as an emotional reaction congruent with, but not identical to, another person's perceived welfare (Batson & Coke, 1981; Batson et aI., 1987). Meta-analytic and narrative reviews have concluded that persons, particularly adults, with low levels of dispositional empathy are significantly more aggressive than persons with high levels of dispositional empathy (Bjorkqvist et aI., 2000; Miller & Eisenberg, 1988). Furthermore, clinical interventions aimed at improving empathy have increased prosocial behaviors and decreased aggression in children (Chandler, 1973; Feshbach, 1989) as well as reduced the reported likelihood of committing sexual assault in men (Pithers, 1994; Schewe & O'Donohue, 1993). Relatedly, research on prosocial behavior has shown that persons with high levels of empathy exhibit more altruistic and helping behaviors than those with low levels of empathy (Batson, 1987; Eisenberg & Miller, 1987). These findings clearly highlight the empirical link between dispositional empathy and aggression as well as the treatment and prevention implications of studying dispositional empathy. Within the framework of a cognitive model of alcohol-related aggression, Giancola (2000) hypothesized that dispositional empathy is a moderator of the alcohol-aggression relation. Giancola (2003a) then conducted the first study to test this hypothesis. Results indicated that alcohol had no effect on aggression for persons with high levels of cognitive or affective empathy. However, alcohol had a powerful effect on increasing aggression for men with low levels of empathy. The results for women were similar in direction, although relatively "weak." Given that this is the only study that has examined this effect, and because findings for women were weak, its results should be considered preliminary and in need of further investigation. BAA. Theoretical Integration and Hypotheses. It was noted above that 3 of the most well-accepted models of alcohol-related aggression contend that alcohol's pharmacological effects facilitate aggression by 1) disrupting higher-order cognitive functions important in maintaining inhibitory control over behavior (Giancola, 2000; Steele & Josephs, 1990),2) inhibiting fear responses (Ito et aI., 1996; Pihl et aI., 1993), and 3) increasing psychological and physiological arousal (Giancola & Zeichner, 1997; Graham et aI., 1997). It was also noted that more general models have argued that alcohol is most likely to engender aggression for persons who are already predisposed to behave in such a manner (Collins, 1988; Graham et aI., 1997; Pemanen, 1991). 19 Giancola, Peter R. The empirical data just reviewed show that persons who: 1) have accepting/permissive beliefs about aggression, 2) believe that drinking alcohol causes aggression, and 3) have low empathy levels are indeed predisposed to behave in an aggressive manner. Hence, based on the above models, it is hypothesized that alcohol consumption will be most likely to facilitate aggression (through the mechanisms of cognitive disruption, reduced fear, and increased arousal) for persons with the appropriate risk factors (i.e., accepting/permissive beliefs about aggression, beliefs that drinking alcohol causes aggression, and low empathy). Conversely, alcohol will be least likely to increase aggression for a) persons with nonaccepting/ non-permissive beliefs about aggression, b) those who do not think that drinking alcohol causes aggression, and c) those with high levels of dispositional empathy, because by their nature, these traits will serve to inhibit aggression. This reasoning is supported by research from the PI's laboratory showing that, compared with placebo, alcohol had essentially no effect on aggression for persons with low levels of dispositional aggressivity (Giancola, 2002b), anger (Giancola, 2002c), and irritability (Giancola, 2002d). However, alcohol had strong potentiating effects on aggression for persons with high levels of these traits. Finally, in keeping with theory (Ito et aI., 1996) and previous findings (Giancola et aI., 2002; Ito et aI., 1996), it is also predicted that alcohol's strongest effect on aggression will be observed under conditions of low, rather than high, provocation in all analyses. Contextual Risk Factors B.4.5. Induced Empathy. According to Batson (1991), there are two necessary and sufficient criteria for the creation of empathy: 1) one's perception that another person is "in need" and 2) adopting that other person's perspective. Only two published laboratory studies have assessed whether an "empathy induction" can reduce aggression (Eliasz, 1980; Richardson et aI., 1994). Results from these studies were equivocal and difficult to interpret because of methodological problems. For example, the empathy manipulation in both studies was "weak" in that it involved merely asking subjects to "try to put yourself in your opponent's shoes." This manipulation is additionally problematic because it does not conform to Batson's criteria. Both of these limitations, and others, could have produced the equivocal results in both studies. Furthermore, Eliasz's measure of aggression had questionable ecological validity. As such, the PI recently completed a study that corrected these problems (Giancola & Phillips, 2003). Based on the work of leading empathy researchers (Batson et aI., 1997; Davis et aI., 1987), empathy was manipulated by having the subject read a sad note that hislher opponent wrote about him/herself (Batson Criterion #1). The subject was then asked to adopt this person's perspective (Batson Criterion #2) while competing against him/her on the TAP in the sober state. The empathy induction significantly reduced aggression for both men and women (see Preliminary Studies section for more detail). This study (Giancola & Phillips, 2003) is important because it shows that induced empathy can suppress aggression and it also suggests that previous studies on this topic may have been methodologically flawed. This study also provides the foundation and rationale to investigate the effects of induced empathy on the alcohol-aggression relation. B.4.6. Provocation. Provocation is one of the most powerful elicitors of human aggression (Bettencourt & Miller, 1996; Ito et aI., 1996; Murdoch et aI., 1990). A recent meta-analysis demonstrated that alcohol has a greater effect on aggression under conditions oflow, rather than high, provocation (Ito et aI., 1996). Specifically, because high provocation has such a strong impact on aggression, alcohol is unlikely to significantly add to this effect. However, inasmuch as low provocation has a smaller effect on aggression, alcohol has "more room" to make a large impact. This finding has been replicated repeatedly in the PI's laboratory (Giancola, 2002b, 2002c; Giancola et aI., 2002). In keeping with Baron and Richardson's (1994) definition of aggression noted earlier, provocation can be elicited by having a person physically "attack" another individual. The TAP is well suited for this purpose because subjects receive both "low" and "high" intensity shocks from their fictitious opponent. Responses on the TAP have been shown to increase in direct proportion with increases in provocation (Bailey & Taylor,1991; Giancola et aI., 2002). B.4.7. Theoretical Integration and Hypotheses. The rationale for testing the moderating effects of dispositional empathy on the alcohol-aggression relation was described above. The added component in this Contextual section is the induced-empathy variable. As noted above, the only study on the effects of dispositional empathy on intoxicated aggression showed that alcohol increased aggression only for persons with low (not high) dispositional empathy (Giancola, 2003a). Based on these findings it is predicted that alcohol's strongest aggression-potentiating effect will be observed for persons with low dispositional empathy who are in the "No Empathy-Induction" group. In contrast, intoxicated aggression will be significantly suppressed Jorpersons with low dispositional empathy who are in the "Empathy-Induction" group. It is also hypothesized that the empathy manipulation will have little effect on aggression Jor persons with high dispositional empathy given that such individuals have been shown to display very low levels oj aggression, even under alcohol (Giancola, 20ma). Finally, in keeping with theory (Ito et aI., 1996) and previous findings (Giancola et aI., 2002; Ito et aI., 1996), it is also predicted that alcohol's strongest effect on aggression will be observed under conditions oj low, rather than high, provocation in all analyses. B.4.8. Why are BelieJs About Aggression and Beliefs About Alcohol not being Manipulated? Beliefs about aggression and alcohol are not targeted for experimental manipulation because this would invalidate the TAP results. As will be seen below, subjects are intentionally not told that our study is examining the alcohol-aggression relation. Informing subjects that we are studying aggression will cause them to artificially attenuate their responses in order to appear more socially-desirable. Clearly, manipulating subjects' beliefs about aggression or their beliefs about the effects of alcohol on aggression will convey the message that this study is measuring aggression and will thus invalidate the results. B.s. Justification for, and Strengths of, this Research As noted above, acute alcohol consumption facilitates aggression for some, but not for all, persons. Hence, the primarv justification Jor this project is to address the need to develop a risk profile Jor alcoholrelated aggression in men and women. Having said this, it is important to note that there is virtually no research focusing on whether men and women differ in their risk factors for intoxicated aggression. As such, a major strength of this project is its ability to examine these gender differences. Given the rising incidence rates of aggression, and alcohol-related aggression, in women (D .S. Department of Justice, 2001), this void requires prompt attention. This project will have the largest sample of women of any laboratory study investigating the alcohol-aggression relation. Another strength of this project is its simultaneous use of both experimental and non-experimental design techniques to assess the combined moderating effects of dispositional and contextual empathy on the alcohol-aggression relation (see Study 2). Whereas demonstrating that dispositional empathy (non-experimental design; Study 1) moderates the alcohol-aggression relation is clearly important, it will not address the question of causality. However, the empathy manipulation in Study 2 (experimental design) will speak directly to this question. Furthermore, it is well known that current methods for predicting aggression, particularly intoxicated aggression, are highly limited (Monahan, 1984, 1988). Thus, another strength of this project is its ability to predict which individuals will, and will not, exhibit intoxicated aggression within controlled laboratory conditions. The results can then be validated in more naturalistic settings. In conclusion, from a conceptual perspective, the results from this project will improve current theories of alcohol-related aggression. From a practical perspective, the results will help develop a "risk profile" that will be of assistance in predicting who will, and who will not, become aggressive when under the influence of alcohol which will clearly have important treatment, preventative, forensic, public health, and social policy implications. 21