جانبداری انتساب خصمانه در خشونت تکانشی و از پیش برنامه ریزی شده

Previous research has demonstrated an association between trait aggression and hostile attribution bias, or the tendency to interpret others’ actions as hostile, yet little research has been devoted to exploring its role in subtypes of aggression. We used hypothetical vignettes to explore hostile attribution bias in impulsive aggressors, premeditated aggressors, and non-aggressive controls. Contrary to our prediction that impulsive aggressors would be more prone to hostile attribution bias, we did not observe direct evidence of this; however, a bias was indirectly evident in the behavior of both types of aggressors. Although they did not specifically articulate feeling as though ambiguous acts were committed with hostile intent, their verbal and physical reactions indicated otherwise. Future research should focus on delineating the full sequence of social cognitions that occur during aggressive encounters in order to determine whether these reactions were produced by common or divergent motivations.

Unlike anger, which has been identified as an affective feature of aggressive behavior, hostility has been characterized as a more cognitive component of aggression (Epps & Kendall, 1995) and as the attitude that motivates aggression (Spielberger, 1988). Hostile attribution bias (HAB) can be defined as a tendency to interpret the intent of others as hostile, despite the fact that environmental cues fail to indicate clear intent (Milich & Dodge, 1984)⁎. The role of biased interpretation of social cues in aggressive behavior has been explained using the social information processing model (Crick & Dodge, 1994). According to this model, the way that children ultimately behave in a given social situation involves a complex sequence of events (encoding of cues, interpretation of cues, clarification of goals, response access or construction, response decision, and behavioral enactment) and is affected by the child’s past experiences. Aggressive children frequently misinterpret incoming social cues, tending to automatically assume that the actions of others are negative and, as a result, respond maladaptively in ambiguous interpersonal situations. Research conducted with aggressive children and adolescents has consistently demonstrated a robust relationship between aggression and HAB, both in hypothetical and actual situations (for a review, see Crick & Dodge, 1994). Other research comparing subtypes of aggressive children has revealed greater HAB in reactive versus proactive aggression (Dodge & Coie, 1987). Although the research on HAB began with children, the concept has since been extended to adults. The most common method of assessing HAB in adults involves presenting ambiguous stimuli that can be interpreted as either hostile/threatening or benign. Participants are shown hypothetical scenarios ranging in nature from hostile to ambiguous to benign and asked to rate how angry they would be if the events happened to them and how intentional and hostile the incidents seemed. Using this design, Epps and Kendall (1995) found that college students high in trait anger and aggression attributed hostility in all three conditions (hostile, ambiguous, and benign), although their HAB was less pronounced in the benign condition. In contrast, those low in trait anger and aggression were less likely to attribute hostile intent in any condition. In a community-based study of drivers, researchers measured participants’ reactions to everyday road scenarios and found that drivers high on measures of trait aggressiveness interpreted ambiguous driving incidents as more hostile and intentional than drivers low in aggressiveness, who tended to be reluctant to attribute hostility and generally assumed road incidents to be accidental (Matthews & Norris, 2002). This bias did not extend to scenarios that were blatantly benign or malign, suggesting that individuals’ own level of aggressiveness may color the way they interpret interpersonal situations when objective evidence is not available. Hostile attribution bias has also been observed in men with a history of marital violence. In their study of maritally violent husbands, maritally distressed but nonviolent husbands, and nonviolent/nondistressed husbands, Holtzworth-Munroe and Hutchinson (1993) presented men with vignettes depicting problematic marital situations and asked them to rate the wife’s behavior in each scenario. Violent husbands were more likely than both other groups to attribute negative intentions to the wife, especially in situations depicting jealousy, rejection from the wife, or potential public embarrassment. Other research involving maritally violent men has shown similar results (Eckhardt et al., 1998 and Jin et al., 2008). Although a substantial literature exists documenting differences in neurochemistry, neuropsychology, psychophysiology, and treatment efficacy between impulsive and premeditated subtypes of aggression (Houston et al., 2004 and Scarpa and Raine, 2000), social-cognitive factors such as HAB have been largely overlooked. Impulsive (or reactive) aggression involves frequent outbursts in which the person becomes emotionally agitated and loses control of his or her behavior; this agitation is typically not warranted by the situation and often occurs in response to perceived provocation (Stanford, Greve, & Dickens, 1995). In contrast, premeditated (or proactive, instrumental) aggression is marked by planned or goal-directed aggressive acts that are carried out in a controlled, unemotional manner, often for the purposes of social gain or dominance (Barratt et al., 1999 and Stanford et al., 2003a). IAs but not PMs tend to exhibit verbal and executive function deficits (Stanford et al., 1997 and Villemarette-Pittman et al., 2003), abnormally low prefrontal functioning (Raine & Venables, 1988), p300 abnormalities (Barratt et al., 1997, Houston et al., 2004, Mathias and Stanford, 1999 and Stanford et al., 2003b) and low serotonin (Linnoila et al., 1983). Hostile attribution bias can serve to maintain chronic aggressive behavior by increasing the likelihood of an aggressive response; when hostile intent is perceived in the behavior of others, an aggressive individual may feel that violent behavior is justified because he or she may view it as retaliation rather than instigation (Holtzworth-Munroe, 1991). This is consistent with research showing that the perception of aggressive intent in others is a powerful cause of anger and aggressive behavior (Epstein & Taylor, 1967). Based on previous research suggesting personality differences between impulsive and premeditated aggressors (Helfritz and Stanford, 2006 and Houston et al., 2004), it is likely that IAs are more prone to HAB. In fact, Giancola (1995) has proposed that the impairments in executive function seen in hostile (impulsive) aggression may lead to a cognitive bias that could increase the likelihood of aggressive behavior in stressful or provocative situations. He explains that, “…deficient self/social monitoring, abstract reasoning, and attention skills may compromise one’s ability to read and correctly interpret potentially ambiguous social cues which can conceivably lead to misunderstandings and possibly aggression in conflict situations” (pp. 444,445). Premeditated aggressors, on the other hand, do not have executive function deficits and often use aggression as a tool to get what they want. Since they are in control of the situation, it stands to reason that they would be more immune to emotional overreactions that could affect their reading of a given social situation. Although this is an understudied area of research, Bailey and Ostrov (2008) investigated the relationship between HAB and aggressive subtypes in a non-selected sample of emerging adults and their results, albeit tentative, showed that reactive physical aggression was associated with a HAB for ambiguous situations related to physical provocation whereas proactive physical aggression was not. Seager (2005) used a binocular rivalry task related to weapons perception in combination with hypothetical vignettes depicting social situation to assess the self-schemas of incarcerated men and found that hostile self-schemas, in conjunction with trait impulsivity, were related to both the frequency of past violent offenses and the degree of psychopathy. Other researchers have also observed a link between HAB and psychopathy (Serin, 1991 and Vitale et al., 2005); however, despite the fact that aggressive men often possess psychopathic tendencies (e.g. disregarding the rights of others), one need not be physically aggressive to be considered a psychopath, as there are many nonviolent ways to violate the rights of others (e.g. “conning” someone out of his or her life savings; Hare, 1991) so it is unclear how these results fit into the aggression literature. To our knowledge, ours is the first study to explore HAB in impulsive and premeditated aggressors with a lifetime history of chronic physical aggression problems. Additionally, previous research on adult HAB has indicated a need to investigate the role of gender of the provocateur (Tremblay & Belchevski, 2004). Due to social norms, we expected an overall trend of decreased aggression toward women as compared to men across all participants; however, given the uncontrolled nature of impulsive aggression, we predicted that this trend would be less pronounced in IAs as compared to PMs and controls. The purpose of the current research was to explore how HAB may differ across subtypes of aggression. Due to differences in personality suggested by our previous work (Helfritz and Stanford, 2006 and Houston et al., 2004), we predicted that IAs would show greater evidence of HAB as compared to both PMs and controls, as indicated by their responses to the Conflict Situation Vignettes (Tremblay & Belchevski, 2004).

Results were analyzed using repeated measures ANOVA, with group as the between-subjects factor and gender of instigator and situation (intentional, ambiguous, or unintentional) as within-subjects factors†. For all univariate ANOVAs and post hoc t-tests, the Bonferroni inequality was used to control for Type I error rate inflation. For all repeated measures analyses, the Geisser–Greenhouse (1958) conservative F test was used as a correction to guard against violations of the sphericity assumption. Responses to the question, “How angry would you be in this situation?” resulted in a significant gender × situation interaction (F(1.55,81.34) = 5.54, p = 0.01, partial η2 = 0.092, power = 0.77). Because follow-up contrasts revealed an ordinal interaction, the main effects of gender (F(1,55) = 101.31, p < 0.001, partial η2 = 0.648, power = 1.00) and situation (F(1.48,81.34) = 235.88, p < 0.001, partial η2 = 0.811, power = 1.00) were interpreted. Results showed that for female instigators (M = 4.24, SE = 0.208), participants reported significantly less anger than for male instigators (M = 5.39, SE = 0.208). Additionally, follow-up contrasts of the main effect of situation showed significant differences between all conditions (p < 0.001), indicating that participants were angriest in situations in which the offending act was intentional (M = 6.41, SE = 0.248), less angry when intentionality was ambiguous (M = 4.91, SE = 0.213), and least angry when it was unintentional (M = 3.12, SE = 0.189). Responses to the question, “How likely is it that you would express to him/her that you are angry?” resulted in significant main effects for gender (F(1,55) = 108.23, p < 0.001, partial η2 = 0.663, power = 1.00) and situation (F(1.59,87.64) = 186.53, p < 0.001, partial η2 = 0.772, power = 1.00). Follow-up contrasts showed that for female instigators (M = 3.48, SE = 0.230), participants reported significantly less anger expression than for male instigators (M = 4.94, SE = 0.246). Follow-up contrasts of situation showed significant differences between all three conditions (p < .001), with anger expression systematically decreasing from intentional situations (M = 6.01, SE = 0.302) to ambiguous situations (M = 4.21, SE = 0.242) to unintentional situations (M = 2.42, SE = 0.199). Responses to the question, “How likely is it that you would be rude to him/her?” resulted in significant gender × situation (F(1.82,100.13) = 7.83, p = 0.001, partial η2 = 0.125, power = 0.932) and group × situation (F(3.12,86.85) = 3.08, p = 0.03, partial η2 = 0.101, power = 0.714) interactions. The gender × situation interaction was ordinal, so the main effects of gender (F(1,55) = 71.64, p < 0.001, partial η2 = 0.566, power = 1.00) and situation (F(1.56,85.85) = 231.85, p < 0.001, partial η2 = 0.808, power = 1.00) were interpreted. All follow-up contrasts were significant (p < 0.001) and results followed the same pattern as anger and anger expression, with participants reporting greater rudeness to men (M = 4.54, SE = 0.253) than to women (M = 3.29, SE = 0.242), and decreasing rudeness from intentional (M = 6.02, SE = 0.331) to ambiguous (M = 4.16, SE = 0.267) to unintentional (M = 1.56, SE = 0.175) situations. Follow-up contrasts for the group × situation interaction showed that PMs differed from controls in ambiguous situations (F(2,55) = 4.38, p = 0.017, partial η2 = 0.137, power = 1.00), whereas IAs did not significantly differ from either group. No significant differences were found for the intentional or unintentional situations (see Fig. 1). Full-size image (17 K) Fig. 1. Conflict Situation Vignettes: Rude by Situation. ∗denotes significant difference from controls (p < 0.05). Vertical lines represent standard error. Figure options Responses to the question, “How likely is it that you would yell or swear at him/her?” resulted in a gender × situation interaction (F(1.83,100.42) = 33.86, p < 0.001, partial η2 = 0.381, power = 1.00), and follow-up contrasts revealed that it was an ordinal interaction so the main effects of gender (F(1,55) = 161.22, p < 0.001, partial η2 = 0.746, power = 1.00) and situation (F(1.26,69.19) = 110.95, p < 0.001, partial η2 = 0.669, power = 1.00) were interpreted. All follow-up contrasts were significant (p < 0.001) and results followed the same pattern found above, with participants reporting a greater likelihood of yelling or swearing at men (M = 3.55, SE = 0.259) than at women (M = 1.36, SE = 0.180), and a decreasing likelihood of yelling/swearing from intentional (M = 3.96, SE = 0.334) to ambiguous (M = 2.32, SE = 0.186) to unintentional (M = 1.09, SE = 0.137) situations. Responses to the question, “How likely is it that you would threaten him/her if the situation was not resolved?” resulted in a gender × situation × group interaction (F(3.53,97.09) = 2.66, p = 0.043, partial η2 = 0.088, power = 0.684), and follow-up contrasts revealed several significant results. For intentional situations, IAs were significantly more likely to threaten than controls were (F(2,55) = 4.58, p = 0.014, partial η2 = 0.143, power = 0.754), and for ambiguous situations, both groups of aggressors were more likely to threaten than controls were (F(2,55) = 4.89, p = 0.011, partial η2 = 0.151, power = 0.783); however, for unintentional situations, no group differences were found (see Fig. 2). For situations in which the instigator was a man, both groups of aggressors were significantly more likely to threaten than controls were (F(2,55) = 4.90, p = 0.011, partial η2 = 0.151, power = 0.783); however, no group differences were found when the instigator was a woman (see Fig. 3). Responses to the question, “How likely is it that you would use physical force (e.g. push or grab) him/her if the situation was not resolved?” resulted in a gender × situation × group interaction (F(3.53,97.06) = 5.29, p = 0.001, partial η2 = 0.161, power = 0.949), and follow-up contrasts showed a pattern of significant results that was similar to those found above. For intentional situations, both groups of aggressors were significantly more likely to report that they would use physical force than controls were (F(2,55) = 5.20, p = 0.009, partial η2 = 0.159, power = 0.809), and for ambiguous situations, PMs were more likely to report that they would use physical force than controls were (F(2,55) = 6.34, p = 0.003, partial η2 = 0.187, power = 0.883); however, for unintentional situations, no group differences were found (see Fig. 4). Full-size image (16 K) Fig. 2. Conflict Situation Vignettes: Threaten by Situation. ∗denotes significant difference from controls (p < 0.05). Vertical lines represent standard error. Figure options Full-size image (13 K) Fig. 3. Conflict Situation Vignettes: Threaten by Gender of Instigator. ∗denotes significant difference from controls (p < 0.05). Vertical lines represent standard error. Figure options Full-size image (15 K) Fig. 4. Conflict Situation Vignettes: Physical Force by Situation. ∗denotes significant difference from controls (p < 0.05); ∗∗denotes significant difference from controls (p < 0.01). Vertical lines represent standard error. Figure options For situations in which the instigator was a man, PMs were significantly more likely to report that they would use physical force than controls were (F(2,55) = 5.20, p = 0.009, partial η2 = 0.159, power = 0.809); however, no group differences were found when the instigator was a woman (see Fig. 5). Full-size image (12 K) Fig. 5. Conflict Situation Vignettes: Physical Force by Gender of Instigator. ∗denotes significant difference from controls (p < 0.01). Vertical lines represent standard error. Figure options Responses to the question, “How likely is it that he/she did it on purpose?” resulted in no significant interactions, but significant main effects for gender (F(1,55) = 9.14, p = 0.004, partial η2 = 0.143, power = 0.844) and situation (F(1.78,97.76) = 318.45, p < 0.001, partial η2 = 0.853, power = 1.00) were found. Follow-up contrasts revealed the same pattern of results that had been obtained for anger, anger expression, rudeness, and yelling/swearing, with participants reporting a lower likelihood that female instigators (M = 3.88, SE = 0.231) “did it on purpose” as compared to male instigators (M = 4.34, SE = 0.182). All follow-up contrasts for situation were significant (p < 0.001) and showed a decreasing pattern of scores from intentional (M = 6.70, SE = 0.290) to ambiguous (M = 4.70, SE = 0.251) to unintentional situations (M = 0.94, SE = 0.137). Responses to the question, “How likely is it that he/she did it to be mean or hostile?” resulted in a gender × situation interaction (F(1.98,109.00) = 11.84, p < 0.001, partial η2 = 0.177, power = 0.993), and follow-up contrasts revealed an ordinal interaction so the main effects of gender (F(1,55) = 45.96, p < 0.001, partial η2 = 0.455, power = 1.00) and situation (F(1.43,78.71) = 239.73, p < 0.001, partial η2 = 0.813, power = 1.00) were interpreted. All follow-up contrasts were significant (p < 0.001) and results followed the same pattern as above, with participants inferring a lower level of hostility from female instigators (M = 2.50, SE = 0.199) than from male instigators (M = 3.36, SE = 0.182), and decreasing levels of inferred hostility from intentional (M = 5.44, SE = 0.312) to ambiguous (M = 2.86, SE = 0.200) to unintentional situations (M = 0.50, SE = 0.100).