خشم در زندانیان: زنان متفاوت از مردان هستند

Anger can contribute to offending behaviour and to behavioural difficulties in prison environments. As such, training in self-management of anger has been a common strategy in an attempt to reduce such behaviours. However, the vast majority of research into anger in offenders has been conducted using male participants. This has led to a lack of knowledge specific to the treatment needs of angry female prisoners. This paper investigates the extent to which a sample of Australian female offenders differs from Australian male offenders in their expression and experience of anger. Fifty women and 121 men were given the State-Trait Anger Expression Inventory [Spielberger, C.D. (1991). State-Trait Anger Expression Inventory: STAXI Professional Manual. Florida: Psychological Assessment Resources] and the Novaco Anger Scale [Novaco, R. W. (1994). Anger as a risk factor for violence among the mentally disordered. In J. Monohan, & H. J. Steadman (Eds.), Violence and mental disorder. Chicago: University of Chicago Press]. The data collected from female participants was then contrasted with identical data collected from male inmates in the separate study [Howells, K., Day, A., Bubner, S., Jauncey, S. (2000). Anger needs and treatment responsivity in male prisoners. Unpublished manuscript: University of South Australia.]. Results indicated significant main effects for gender in a majority of the subscales of the two measures, with significant differences found in both the experience and expression of anger for male and female prisoners. The results are discussed in terms of the implications for correctional service providers with respect to the specific psychological needs of female offenders.

The vast majority of studies into anger in offenders have been conducted using male participants, (a problem that is not unique to the study of anger). Given that the ratio of male to female offenders averages 1:20 in USA, UK and Australia (Casale, 1998, Easteal, 1992, Koons et al., 1997 and Walsh, 1997), females remain a neglected population in many areas of criminological research. Consequently, correctional providers lack adequate knowledge about many facets of female offending. Despite the many differences that exist between male and female offenders, programs and services designed for males are extended to female offenders with little alteration (Koons et al., 1997). This is true of anger management programs, (Horn & Towl, 1997). The lack of knowledge related to the treatment needs of angry female prisoners is particularly significant given that anger has been viewed as both a primary cause of female imprisonment (McDonagh, 1999), and as a consequence of their imprisonment (Pennix, 1999). An understanding of how male and female offenders compare in their experience and expression of anger will serve to ameliorate this situation. In this paper we will explore such comparisons, propose possible explanations for any differences observed and suggest a course for future research. 1.1. Defining anger Anger is a universally experienced emotion. While anger does not necessarily result in antisocial behaviours, it has been demonstrated to be a major contributor to offending behaviour as well as a management concern for prison administrators (Howells, 1998). It was not until the latter half of the twentieth century that anger was subjected to rigorous scientific appraisal. This has largely been led by the work of Novaco (Meichenbaum & Novaco, 1985, Novaco, 1976, Novaco, 1978, Novaco, 1993 and Novaco, 1997) and Spielberger (Spielberger, 1991 and Spielberger et al., 1983). Novaco’s conceptualisation of anger has provided the theoretical foundation upon which to base the treatment of anger problems (Jones et al., 1999 and Novaco, 1994). The Novaco model describes anger as comprising three components which exert a reciprocal influence upon each other in response to an external trigger or environmental circumstance. The first of these are cognitions, which are viewed as central to the experience of anger and involve the angered individual’s appraisals, expectations, attitudes and beliefs. Recent research confirms that individuals high in anger have distinctive patterns of cognitive appraisal ( Hazebroek, Howells, & Day, 2001). The second component relates to physiological arousal. Where this occurs concurrently with an anger-related cognitive interpretation of the triggering circumstance, the physiological arousal is also likely to be subjectively labelled as anger. The emotion of anger can then give rise to Novaco’s third component, behavioural reactions, which can range from verbal confrontation to physical assault or destruction of property ( Daffenbacher et al., 1996). Novaco subsequently designed the Novaco Anger Scale, which assesses the experience of anger and the specific triggers that elicit an individual’s anger ( Jones et al., 1999, Novaco, 1994 and O’Neill, 1995a). Spielberger (Spielberger, 1991 and Spielberger et al., 1983), on the other hand, has concentrated on ‘dispositional’ aspects of the individual in relation to anger. Spielberger has distinguished state and trait anger, the distinction reflecting the understanding that anger is both an emotional state, varying across time, situation and intensity, as well as a stable personality trait, reflecting a person’s tendency to experience anger frequently or intensely (Spielberger & Sydeman, 1994 and Spielberger et al., 1983). Spielberger has also introduced assessment of the forms of expression of anger, in the State-Trait Anger Expression Inventory (STAXI; Spielberger, Sydeman, Owen, & Marsh, 1999). 1.2. Sex differences in anger As a universal human emotion, specific variables which might have an impact upon anger have been the subject of investigation. One such area has been the impact of gender, where research into sex differences in anger has produced mixed results. Studies which have failed to find sex differences have used both child (Brody, 1985, Brody et al., 1995, Buntaine & Costenbader, 1997 and Zenman & Shipman, 1996) and adult samples (Averil, 1983, Kopper, 1993, Kopper & Epperson, 1996 and Kopper & Epperson, 1991). In one of the few studies to report on a prisoner sample, Swaffer and Epps (1999) found that the anger of a sample of male and female adolescents held at a secure treatment centre failed to differ on either the STAXI; (Spielberger, 1991) or all but one sub-scale of the Novaco Anger Scale (NAS, indirect expression, Novaco, 1994). However, other studies do report differences. For example, while investigating the validity of the Anger Expression Scale, Spielberger et al. (1983) found that girls reported higher anger expression than boys. Other studies have reported differences in anger expression (Faber & Burns, 1996), higher frequency of female anger (Brody et al., 1995) and differences in anger management training needs of police officers (Abernethy & Cox, 1994). The reasons for these mixed results are unclear. However, it seems likely that one possible explanation lies in the specific characteristics of the sample population and how these characteristics influence the measurement used. Some insight is provided through research using clinical populations, where sex differences have been reported. For example, Funabiki, Bologna, Pepping, and FitzGerald (1980) found sex differences in the verbal hostility displayed by depressed patients, while Novaco (1994, cited in O’Neill, 1995b) found sex differences while collecting normative data for the Novaco Anger Scale, with females scoring higher than males. Sex differences have also been found in the behavioural manifestations of anger. Kelsall, Dolan, and Bailey (1995) reported that females accounted for almost half of the violent incidents reported at an adolescent forensic unit, despite constituting only a third of the population under study. While these results appear counter-intuitive, Kelsall et al. (1995) included self-harm in their measures of violent behaviour, which may be relevant to the gender imbalance of reported violent incidents. Such a finding is supported by the above-named study (Swaffer & Epps, 1999) in which females scored higher on the indirect expression of anger. These authors hypothesised a link between such scores and self-harming behaviour. One inference that may be drawn from studies such as these is that the clinical features of the sample population are of significance when interpreting the results of anger measures. This would mean that the clinical profile of women in general, and female offenders in particular, has relevance to understanding the results of gender comparisons. While there is relatively little research specifically related to psychopathology among female offenders (Hurley & Dunne, 1991, Keaveny & Zauszniewski, 1999 and Raeside, 1994), there is a consensus that mental health problems are more common among female prisoners than their male counterparts (Daniel et al., 1988, Gorsuch, 1998, Mohan et al., 1997, Morash et al., 1998 and Raeside, 1994). For example, women offenders have more extensive traumatic histories than males, with higher levels of physical and sexual abuse reported (Gorsuch, 1998 and Sheridan, 1996). Frothingham, Hobbs, Wynne, Yee, Goyal, and Wadsworth (2000) report that sexually abused children are significantly more likely to experience subsequent problems across life domains while Jehu (1991) suggests that victims of childhood sexual abuse often direct their intense anger inward toward themselves. Observations such as these may shed light on the findings of Kelsall et al. (1995), cited above. 1.3. Anger and crime Anger remains an important area of study given the rising community concern over violent crimes (Howells & Hollin, 1989 and Novaco, 1994). While anger can be an antecedent to some violent offending, it is “neither necessary nor sufficient” as an explanation of violence (Howells, 1998). Anger primarily plays a role in hostile or retaliatory violence (Brown & Howells, 1996 and Holbrook, 1997). Hostile violence refers to spontaneous violence resulting from social factors/interactions, as opposed to instrumental violence where the behaviour is premeditated and the violence simply forms a means to an end (Brown & Howells, 1996). Although the angry–instrumental distinction has been disputed by some theorists (Indermaur, 1995), anger can nonetheless be viewed as an element which sometimes influences the likelihood of violent behaviour (Howells, 1998). Anger can (and does) play a role in both the prediction of violence and the treatment of violent offenders. The prediction of violent recidivism is an important area of focus for correctional services (Byrne et al., 2001 and Rice, 1997). While some studies have failed to discriminate violent from non-violent offenders using behavioural measures (Williams, Boyd, Casardi, & Poythress, 1996), those using anger measures have been successful in distinguishing violent from non-violent offenders (Cornell et al., 1999, Granic & Butler, 1998 and Selby, 1984). However, most research has been directed at male offenders and the utility of anger measures for the prediction of female violence remains unclear. 1.4. Aims of the study This study investigated the experience and expression of anger in a sample of Australian female prisoners and compared obtained results with those of an Australian male sample recruited for a separate study (Howells, Day, Bubner, & Jauncey, 2000). Two measures of anger were used: the STAXI (Spielberger, 1991) and the NAS (Novaco, 1994). The study was exploratory in nature, with a primary focus upon whether or not adult offenders differ on measures of anger experience and expression. In view of the dearth of literature on such differences, significant findings might have implications for methods used in the treatment of anger of female offenders. It was expected, in light of previous studies (e.g. Kelsall et al., 1995), that such differences may exist and that these differences were likely to be in the direction of greater anger experience and expression among female offenders. At a general level, it was expected that high levels of anger across gender groups would be reported, consistent with Spielberger’s (1991) observation that prison inmates score higher on measures of anger than other populations. Furthermore, it was anticipated that violent offenders would score higher than non-violent offenders, consistent with the observations of Howells et al. (2001) and Mills, Kroner, and Forth (1998). To facilitate the exploration of gender differences as described above, data were analysed at the level of sub-scales within both measures. This was of particular importance with the STAXI. Differences in the expression of anger and its control may assist speculation about the reasons for any reported gender differences as well as inform review of the applicability of anger management programs across gender groups.

This study found significant differences between female and male prisoners on a wide range of anger measures. How such differences are to be explained remains unclear. While neither mental pathology nor prior abuse were measured in the present study, we have hypothesized that the differences observed between male and female prisoners in their experience and expression of anger, and the relatively (to non-offender norms) high level of anger in this female prisoner sample may be the result of a higher incidence of psychopathology in this group. We have attributed this psychopathology to experiential differences between the sexes, such as exposure to childhood and adulthood trauma. However, these interpretations are highly speculative, with our research being limited in a number of areas. First, the sample size was smaller than would be desired. This was the result of the small population of female offenders in the state of South Australia and could be overcome by replicating the study in other states, using larger samples. A larger sample would also enable comparison of violent and non-violent offenders not possible in the present study. Clearly, the inclusion of measures of psychopathology, trauma experience and anger-related behaviour (including self-harm) would enhance the veracity of our interpretations. This might be achieved through the further use of self-report measures and the design of anger-related behaviour rating scales for use by custodial staff. However, the assumption of environmental (experiential) causes, as opposed to biological factors, also needs to be tested. Inclusion of biological variables, such as hormonal factors, would strengthen the veracity of our conclusions. A third and important limitation relates to the use of self-report instruments per se. Self-report measures hold the potential to provide extremely useful clinical information assuming that the respondent has insight into their own thinking and behaviour and is motivated to respond honestly (Cohen & Swerdlik, 1999). Our assertion that the female sample represents a more pathological group would seem to challenge the assumption of motivation and, perhaps, insight. Indeed, the testing environment (a prison), is certain to influence item response, with previous research demonstrating an elevation in paranoia within such environments (Sutton, Byrne, & Byrne, 2000). One possible response to this difficulty might be the inclusion of peer or third party ratings of respondents against which self-report measures may be compared (Costa & McCrae, 1997). This research has implications for correctional service providers. It is important for those involved in the treatment and management of female prisoners to note that women differed from men in both their experience and expression of anger as well as the triggers that promulgate angry feelings and behaviours. This means that assumptions that interventions and management strategies developed using male participants will equally benefit women may be questionable. Women prisoners appear to have gender specific needs in the area of anger. Potentially, these may relate to the form of anger expression (self-harm) and the type of cognitions related to anger experience (generalisation of ‘unfair treatment’). Future research may build on this study to clarify specifically how women differ from men.