Martijn van Teffelen

General discussion 155 7 GENERAL DISCUSSION The present work set out to find a better understanding of hostility and how to influence it. Specifically, this thesis had three main goals: (1) examining the dimensions of hostility ( chapter 2 ); (2) comparing laboratory provocation methods that induce hostility and assess how these interact with psychopathic and narcissistic personality traits ( chapter 3 ); and (3) developing new ways to advance treatment options for hostility ( chapters 4 , 5 and 6 ) by testing the efficacy of imagery-enhanced cognitive restructuring (I-CR) and cognitive bias modification of hostile interpretation bias (CBM-I). In the following, the main findings are summarized and discussed and recommendations for future research will be provided. Last, implications for the clinical field will be outlined. MAIN FINDINGS Dimensions of Hostility In hostility research many scholars use the same term for different constructs, or different terms for the same construct. Some refer to anger (Ramirez & Andreu, 2006; Spielberger, 1999) or irritability (Vidal-Ribas et al., 2016) as the affective dimension of hostility, while others include both cognitive and affective aspects of hostility when referring to anger (Wilkowski & Robinson, 2008). Along the same lines, there are scholars who coin the term hostility for cognitive aspects of antagonistic phenomena (Chida & Steptoe, 2009; Smith, 1992; Spielberger et al., 1985). Others define hostility as the sum of affective, behavioral, and cognitive aspects (American Psychiatric Association, 2013; Barefoot, 1992). There are two potential reasons for the confusion in concept identity of hostility. One reason is the diversity in the number of factors that are reported in psychometric studies. For example, some report two factors aligning with the two-dimensional anger experience/anger expression model (see e.g., Martin et al., 2000), while others report factor solutions in line with the three- factor anger-hostility-aggression (AHA) or affect-behavior-cognition (ABC) model, sometimes even within the same study (Martin et al., 2000). Up to five factors have been reported in the literature (see e.g., Maier et al., 2009). A second, more fundamental reason that provides an explanation for the concept identity confusion in hostility research is that hostility components are simply not accurately measured. Despite the fact that many hostility self-reports hold excellent psychometric properties, closer inspection shows that many self-report items tap into and subsequently load on more than one dimension. That is, an item that intends to measure aggressive behavior may also assess the affective component of anger. As an example, the Buss-Perry Aggression Questionnaire (AQ) (Buss & Perry, 1992) includes items such as “I have become so mad that I have broken things” (measuring both anger and physical aggression),