ارزیابی خشم و خصومت: نقد

While the emotion of anger has become an increasingly important part of clinical assessment, the theoretical and psychometric adequacy of the instruments used to assess anger and hostility have long been questioned. In the present review, we first provide definitions of anger and hostility in order to provide a theoretical context from which to evaluate the scope of current measures of these constructs. Second, we review the major self-report scales used to assess anger and hostility in light of these definitions and provide a detailed evaluation of psychometric evidence concerning their reliability and validity. Finally, we offer specific recommendations concerning how anger and hostility assessment instruments can be improved and expanded. In particular, we note the need for (a) an expansion of anger assessment methods beyond traditional endorsement approaches, (b) scales to assess specific domains of anger experience, (c) scales that assess unique content domains of anger experience and expressions, such as spouse-specific or driving-related anger scales, and (d) scales that assess the clinical aspects of the anger construct.

It is well accepted that the first step in constructing any psychological assessment device is a careful theoretical consideration of how the construct should be defined. Once the theoretical starting point is established, operational definitions of each component of that construct can be forged and scale construction can begin, with the quality of the final version depending in part on how closely the supporting data match the scale's original theoretical basis (Cronbach & Meehl, 1955). Unfortunately, as suggested by numerous researchers (e.g., Berkowitz, 1993, Eckhardt & Deffenbacher, 1995 and Novaco, 1994), lack of theoretical progress in achieving a basic understanding of anger as a socially and clinically relevant emotion has been a major hindrance to the scale development process described above. As indicated by Berkowitz (1994 p. 35), “Any really close and thorough examination of the psychological research into the origins of anger and emotional aggression must leave the thoughtful reader somewhat dissatisfied. The literature presents us with occasional inconsistencies and unexpected findings that most of the investigators seem not to have noticed....” Given this theoretical confusion, it seems unlikely that methods developed to assess anger share sufficient conceptual variance and measure a similar construct. However, researchers investigating the causes, consequences, and correlates of anger must erect their conclusions on the foundation of assessment; to the extent that anger assessment instruments cannot adequately capture the construct, the quality of anger research will continue to be suspect. Given the complexity of the anger construct (for reviews, see Berkowitz, 1993 and Kassinove, 1995), anger assessment must involve careful attention toward a wealth of internal and external variables. Thus, theoretical contributions to a measure's development must be sound, and the methods used during the development of the measure must be empirically anchored. Across the last 20 years, several reviewers have expressed concern over the status of anger assessment. For example, Biaggio et al. Biaggio, 1980 and Biaggio et al., 1981 reviewed several widely used self-report scales and concluded that, as a whole, these instruments possessed weak psychometric characteristics, were only marginally related to each other, and were generally lacking in validity. In addition, few relationships were obtained between paper-and-pencil measures of anger and subjects' diary records of anger experiences. Spielberger et al. Spielberger et al., 1983 and Spielberger et al., 1985 approached the problematic status of anger research by noting that available scales were greatly lacking in conceptual clarity and construct validity. Scales appeared to be constructed without explicit definitions of anger, hostility, and aggression in mind, and there was little differentiation between the experience and expression of anger. This point was echoed by Novaco (1994), who noted that neglect of theory and lack of operational definitions were disturbing communalities in the development of existing anger scales. In addition, most instruments assess anger as a basic personality dimension as opposed to a clinically relevant affective domain, making it difficult to shift the field's perspective from anger as a basic personality dimension, to anger as a variable of potential clinical importance (Eckhardt & Deffenbacher, 1995). Together, these reviews indicate that despite the importance of anger as a theoretical and socially relevant construct, researchers had made little progress in developing construct valid anger assessment methods. With this in mind, the purpose of the present paper is to critically review a wide range of anger assessment methods in terms of their psychometric quality and conceptual clarity. We will begin by discussing conceptual distinctions between anger and hostility, review specific assessment instruments to measure hostility and anger, and provide concluding remarks.

In the present paper, we attempted to review the major assessment instruments used for the measurement of anger and hostility. The scales discussed varied widely in their theoretical underpinnings, psychometric quality, and construct validity. While some scales were developed with clear, empirically based definitions and possess excellent psychometric quality (e.g., NAS, STAXI-2, AQ), others did not clearly distinguish between the constructs of anger and hostility or had an otherwise suspect empirical basis. The points we would like to emphasize in this conclusion are less concerned with issues regarding the various psychometric details of the instruments reviewed, for these are themselves dependent upon what we see as the critical aspects of anger assessment—the theoretical model underlying the instrument and the researcher's or clinician's purpose in using a particular scale. These factors have been identified as central components of modern definitions of construct validity (e.g., Messick, 1995) and suggest that statements about the quality of instruments to assess anger and hostility should be largely based upon whether data from a particular scale are consistent with the theory underlying the measure as well as the particular purpose for which the measure was used. In terms of scale development, it does not appear that most of the instruments reviewed emerged from a clear and consistent theoretical basis. While it would be tempting to suggest that scale developers have neglected to adequately review the empirical and theoretical literature on anger and hostility before constructing their particular scales, recall the quote from Berkowitz (1994) in the first paragraph of this paper, wherein he makes the case that the anger literature itself is replete with inconsistent and unexpected findings that resist clear theoretical integration. For example, what are the boundaries between the constructs of anger and hostility? To what extent do these labels describe the same construct (e.g., the cognitive underpinnings of anger versus the negative, pessimistic cognitions thought to underlie hostility)? To what extent does the evidence support their distinction (e.g., the multiple dimensions of the anger construct versus the narrower definitions of hostility)? Should anger be measured as a basic personality trait or a dimension of diagnostic and clinical relevance? The available literature offers few if any solid answers to these vexing questions. The point here is that one's theoretical model of a construct defines the nature and range of that construct, and in turn dictates the manner in which the construct can be measured Loevinger, 1957 and Smith & McCarthy, 1995. Thus, the theoretical problems that exist within the scientific literature on anger and hostility are likely to carry over into similar inconsistencies in attempts to measure these constructs.