Martijn van Teffelen

General discussion 157 7 antecedents that lower inhibition capacity such as increased cognitive load (Vasquez & Howard-Field, 2016) or the acute presence of physical threat (Blanchard et al., 2001). In turn, hostile interpretation tendencies may be more affected by the (emotional) ambiguity of threats (Blanchard et al., 2011). Results from other chapters in this dissertation further illustrate the importance of differentiating between different hostility dimensions. First, in chapter 3 we differentiated cognitive (i.e., perceived threat), affective (i.e., negative affect), and behavioral (i.e., physical aggressive behavior) factors and found that –generally- these do not differentially alter after experiencing social exclusion or insult. However, our findings showed that psychopathic personality traits are positively related to reduced negative affective change, and in turn to reduced aggressive responding. Results also revealed that narcissistic personality traits are positively related to perceived threat under insult conditions, and in turn to reduced initial aggressive responding. Thus, dimensions of hostility appear to have differential mediating effects on relationships between psychopathology and provoked physical aggression. Second, in chapter 4 we differentiated hostile beliefs (i.e., cognitive dimension), anger (i.e., affective dimension), and aggression (i.e., behavioral dimension). We found that compared to an active control condition (AC), traditional cognitive restructuring (CR) and I-CR were more efficacious in reducing hostile beliefs, anger, and aggression. However, I-CR was sustainably more efficacious in reducing hostile beliefs and aggression compared to AC and CR. Therefore, I-CR and CR seem to impact the majority of hostility dimensions compared to AC. Finally, in chapter 6 we differentiated hostile interpretation bias (i.e., cognitive dimension), anger and aggression. Results showed that compared to AC, CBM-I was more efficacious in reducing hostile interpretation bias and aggression. Hence, CBM-I seems to impact cognitive and behavioral dimensions, but not affective dimensions of hostility. Moreover, in both chapters 4 and 6 we found no differential impact between the active interventions and AC on a broad ‘capture-it-all’ hostility measure (i.e., the PID-5H). Thus, our findings illustrate that distinct interventions indeed have differential impact on dimensions of hostility. Altogether, our work consistently stresses the importance of differentiating hostility components. Hostility components do not only require distinct measurement; they can also help to disentangle the differential impact of personality traits, proactive and reactive aggression, and treatment. In treatment for example, it is vital to understand which ‘button’ elicits change in which domain. Treatment mechanism studies could provide insight into what intervention impacts which hostility dimension. Similarly, it is important to understand which domains are not affected by therapeutic interventions and would require additional or other forms of therapy. Leaving some hostility dimensions untreated may lead to a suboptimal reduction of hostility as a whole.