The Maudsley Violence Questionnaire: initial validation and reliability
Introduction
Research into beliefs and thinking specifically associated with violent offending is limited in comparison to similar research into sex offending (see Abel, Gore, Holland, & Camp, 1989; Blumenthal, Gudjonsson, & Burns, 1999; Hayashino, Wurtele, & Klebe, 1995; McGrath, Cann, & Konopasky, 1998; Vanhouche & Vertommen, 1999; Ward, Hudson, Johnston, & Marshall, 1997). The majority of research into violence and aggression has focussed on risk factors (Bjørkly, 1997; Monahan, 1984; Monahan & Steadman, 1994) and anger (Novaco, 1975, Novaco, 1994; Novaco & Welsh, 1989). Although Novaco’s work has considered triggering thoughts for anger, anger itself is neither necessary nor sufficient for violence and therefore is only one of the mediating psychological variables for aggression.
Slaby and Guerra (1988) showed the significant predictive validity of beliefs supporting the use of aggression in 15–18 year old. Their 18 item belief measure included five subscales: the legitimacy of aggression; aggression increases self esteem; aggression helps avoid a negative self image; victims deserve aggression; and, victims don’t suffer. The victim related subscales were least predictive of violence and the most predictive subscales were avoiding a negative self image, increasing self esteem and legitimising aggression. The items were not factor analysed, but had been generated in line with social learning theory.
Other studies have considered offending by assessing criminogenic or criminal factors, e.g. the Psychological Inventory of Criminal Thinking styles (Walters, 1995, Walters, 1996) and the Psychopathy Checklist––Revised (Cornell et al., 1996; Hare, 1991; Hare et al., 1990). There may be a number of problems with these approaches. First, uncertain conceptual basis, e.g. the PICTS did not factor analyse into the clinical scales used in the measure (Walters, 1995) and psychopathy has never been recognised as a diagnosis by either of the classification systems, ICD-10 (1992) or DSM-IV (1994). Second, both the PICTS and the PCL-R are generic measures of offending rather than specific to violence. Third, the complexity of violence (Monahan & Steadman, 1994)––in a violent act there are likely to be multiple causal and exacerbating factors, ameliorating or protective factors, individual differences, and the influences of other factors such as the environment and victim behaviour and characteristics. This problem has lead McGuire and Priestly (1995) to suggest that the whole focus of risk assessment needs to shift towards understanding what makes this particular person commit this particular violent act at this particular time. Such an individualised approach requires the assessment of a variety of individual factors, psychological mechanisms and thoughts.
Numerous measures exist for anger, hostility, paranoia, impulsivity and empathy. A search of the literature (by the author for this paper) revealed over 40 in current use and previous reviews cite a similar or greater number (Gothelf, Apter, & Van Praag, 1997; Miller, Smith, Turner, Guijarro, & Hallet, 1996) which may all relate to violent and aggressive behaviour, but these tend to be personality measures, risk assessments or behavioural measures. Those that purport to measure thoughts do not comprehensively measure the specific thoughts, beliefs and attitudes which may principally relate to violence.
For the purpose of this study a questionnaire was developed to measure these thoughts, with the potential for use with various groups of individuals. The intention was to evaluate individuals’ thoughts and beliefs about violence, violent acts and beliefs about what is acceptable, justifiable and reasonable in various situations. The Maudsley Violence Questionnaire (MVQ) was designed with reference to violence literature and to clinical experience with violent offenders, but with a view to evaluating the statistical performance of the measure, through factor analysis, on a normal population. The author has run a Violence Clinic for 6 years which has resulted in the treatment of a number of violent individuals, some of whom have completed the MVQ. The MVQ was also given to other professionals and several lay people to give their reflections and to make improvements. The feedback was generally positive from all groups and minor changes were made before it was used in the current research. Furthermore, the MVQ was also reviewed by Dr. L. Firestone (author of the FAVT, Firestone Assessment of Violent Thoughts; Doucette-Gates, Firestone, & Firestone, 1999) who has used the MVQ in a similar (unpublished) study in California, USA.
The FAVT is the only other violence specific measure which has undergone adequate factor analysis and assesses thoughts and feelings from the perspective of an internal ‘voice’––‘an integrated system of negative thoughts and attitudes, antithetical to self and cynical and hostile thoughts toward others’ (Doucette-Gates et al., 1999, p. 116) e.g. ‘you’re stupid’, ‘you’re a failure’. The FAVT factor analysed into four robust factors: social mistrust; perceived disrespect/disregard; negative critical thoughts; and, expression of overt anger. There are other attitude scales which have been developed for use with children, but one is specific to gun violence (AGVQ, Shapiro, Dorman, Burkey, Welker, & Clough, 1997); and the ‘Attitudes towards violence scale’ is brief (15 items on two factors) and has not been compared to self-reported violence (Funk, Elliott, Urman, Flores, & Mock, 1999). From the perspective of a generic cognitive model (e.g. Beck & Freeman, 1990), the FAVT appears to assess automatic thoughts (e.g. ‘they don’t give a damn about you’; ‘you’ll show them who’s the boss’) while the MVQ looks at rules or dysfunctional assumptions (e.g. ‘it is ok to hit someone if they make you look stupid’; ‘if you are not willing to fight it means you are weak and pathetic’), core beliefs (e.g. ‘I see myself as a violent person’, ‘violence is second nature to me’) and related cognitions. Given the lack of therapeutically driven research into scales assessing violent thoughts, the development of the MVQ was seen as justified from a research perspective and necessary from a clinical perspective. Furthermore, compared to measures such as the PCL-R and the PICTS which can only be used on offender populations, the MVQ was designed for use with individuals violent or otherwise, regardless of previous convictions.
It is clear from past research and work by Novaco and others that there are patterns of dysfunctional thinking which may be associated with anger, aggression and violence (reviewed by Novaco & Welsh, 1989). Furthermore, in sex offending literature, many scales exist which measure cognitions, beliefs and distortions that have been found to be important in sex offending (e.g. Abel et al., 1989; Burt, 1980; Bumby, 1996), but few comparable measures exist for violence. From theoretical perspectives whether a cognitive model (e.g. Beck & Freeman, 1990), Novaco’s specific model of anger (1994), social learning theory (see Slaby & Guerra, 1988) or from attitude behaviour research (Kraus, 1995) it is likely that certain cognitions predispose towards violence, and that certain cognitions are reinforced by violent and aggressive behaviour, thus maintaining the cycle of violent behaviour, as in other types of offending.
The items in the MVQ were designed to assess attitudes and beliefs which predispose towards or legitimise violence and to reflect the existing underlying theories of violent/aggressive behaviour from a cognitive perspective. Items were developed in particular which related to the masking of low self esteem with aggression, or the promotion of violence in response to threatened self esteem (see Salmivalli, 2001). Items were generated by applying a cognitive behavioural formulation to violence (Walker & Bright, in preparation) and it appeared that many dysfunctional assumptions (rules), attitudes and beliefs were extensions of normal beliefs which justify an aggressive response to a variety of perceived provocations. The central theme of perceived provocation concerned seeing violence as a justifiable and perhaps the only response to embarrassment or threatened self esteem (Gilligan, 1996). That is not to say that violence was used to promote self esteem in an instrumental way, but that it is used when seen as necessary and justified by the aggressor. Gilligan’s (1996) formulation for violence suggested that several factors in combination are necessary for violence. These factors include low self esteem and anger, but also the crucial trigger of humiliation as perceived by the aggressor, and in addition the lack of adequate alternative responses to ‘save face’. These key concepts of respect and physical bravery in humiliating situations were the central themes underlying the development of the items for the MVQ. The items were generated in the form of rules (or dysfunctional assumptions) for various potentially humiliating situations, or beliefs which may support or legitimise violence. Furthermore, the frequency of violence in society lead to the development of a questionnaire based on normal thinking processes to explore whether a more extreme thinking pattern could be found that related to increased violence. To the author’s knowledge, this is the first time such a measure of this specific set of violent thoughts has been adequately developed with reference to thinking structures within the cognitive model (which has become a treatment model of choice for aggression and violence, both in the community with various populations, and in prison; Beck & Fernandez, 1998; Robinson & Porporino, 2001); appropriately factor analysed with a large sample; and, compared against a normal distribution of violent behaviour.
In the present study, the MVQ was validated initially with a normal adolescent population for several reasons. First, the beliefs used, although partly generated from an offending perspective, represent extensions of normal beliefs. Second, it remains difficult to acquire a large enough sample of reliably responding offenders for an initial analysis of such a questionnaire. Third, adolescents are old enough to have developed a belief system which is important in predicting behaviour (Kraus, 1995). Fourth, there is considerable evidence of the continuity of antisocial behaviour through adolescence into adulthood (Farrington & West, 1993), and in particular aggressive behaviour (Eley, Lichtenstein, & Moffitt, 2003). The developmental understanding of violence has also been emphasised by other authors (e.g. Fonagy, 2003; Gilligan, 1996; Slaby & Guerra, 1988).
This paper presents the results of an exploratory factor analysis of the MVQ completed by 785 school students (age 16–18) attending schools in London. Concurrent validity with self report delinquency is also presented. It was predicted that those showing more evidence of violence related cognitions would show elevated rates of self reported violent behaviour.
Section snippets
Participants
The participants were 785 students from nine schools within the London area. Schools approached were from a range of London Education Authorities, a range of academic achievement levels and different areas of London. The majority of schools were mixed comprehensive; however, single sex, voluntary aided and foundation schools were also represented. Schools were chosen to reflect the broad socio-economic and educational background of school students aged 16–18 in the London area. Of the potential
Results
The first analysis was to assess the adequacy of the sample for factor analysis. The Kaiser–Meyer–Olkin Measure of Sampling Adequacy was 0.849 which is above the recommended cut off of 0.6 (Kaiser, 1974); Field (2000) suggests that KMO’s above 0.8 are very good; and Bartlett’s test of Sphericity was significant (p<0.001), so factor analysis was seen as appropriate for this data set. Kass and Tinsley (1979) recommended a minimum of five participants per variable (97 variables in this study) up
Discussion
This study demonstrated that attitudes, beliefs and rules that theoretically relate to violence and aggression can be measured, and correlate with self-reported violent behaviour. The factor analysis of the MVQ showed that these beliefs are associated with violence for both men and women and comprise two robust and reliable factors: ‘machismo’ and ‘acceptance’ of violence. The Attitudes Towards Violence Scale (Funk et al., 1999) also factor analysed into two factors, ‘culture of violence’ which
Acknowledgements
Prof. Gisli Gudjonsson for supervision, support and encouragement, Dr. Paul Barrett for patient and comprehensive statistical help, Dr. Lisa Firestone for enthusiasm about the MVQ and reviewing the project; the South London and Maudsely NHS Trust for allowing me time to undertake this research; and, the schools, teachers and pupils who took part in the research.
References (51)
Clinical assessment of dangerousness in psychotic patients: some risk indicators and pitfalls
Aggression and Violent Behavior
(1997)- et al.
Cognitive distortions and blame attribution in sex offenders against adults and children
Child Abuse & Neglect
(1999) - et al.
Measurement of aggression in psychiatric patients
Psychiatry Research
(1997) Feeling good about oneself, being bad to others. Remarks on self-esteem, hostility, and aggressive behaviour
Aggression and Violent Behaviour
(2001)- et al.
Anger in prisoners: women are different from men
Personality and Individual Differences
(2002) - et al.
Cognitive distortions in sex offenders: An integrative review
Clinical Psychology Review
(1997) - et al.
The measurement of the cognitive distortions of child molesters
Annals of Sex Research
(1989) - et al.
Cognitive therapy of personality disorders
(1990) - et al.
Cognitive-behavioural therapy in the treatment of anger: a meta-analysis
Cognitive Therapy and Research
(1998) Assessing the cognitive distortions of child molesters and rapists: development and validation of the MOLEST and RAPE Scales
Sexual Abuse: Journal of Research & Treatment
(1996)
Cultural myths and supports for rape
Journal of Personality and Social Psychology
An adaptation of Mak’s self reported delinquency scale for Western Australian adolescents
Australian Journal of Psychology
A first course in factor analysis
Psychopathy in instrumental and reactive violent offenders
Journal of Consulting and Clinical Psychology
Assessing violent thoughts: the relationship between thought processes and violent behaviour
Psychologica Belgica
A longitudinal behavioural genetic analysis of the etiology of aggressive and nonaggressive antisocial behaviour
Development and Psychopathology
Criminal, penal and life histories of chronic offenders; Risk and protective factors and early identification
Criminal Behaviour and Mental Health
Discovering statistics: using SPSS for windows
Towards a developmental understanding of violence
British Journal of Psychiatry
The attitudes towards violence scale: a measure for adolescents
Journal of Interpersonal Violence
Violence: reflections on our deadliest epidemic
Manual for the hare psychopathy checklist-revised
The revised psychopathy checklist: descriptive statistics, reliability, and factor structure
Psychological Assessment: A Journal of Consulting and Clinical Psychology
Cited by (58)
Effectiveness of psychological and psychosocial interventions for forensic mental health inpatients: A meta-analysis
2021, Aggression and Violent BehaviorThe Role of Social Support in Machismo and Acceptance of Violence Among Adolescents in Europe: Lights4Violence Baseline Results
2021, Journal of Adolescent HealthCitation Excerpt :These gender differences were also confirmed in terms of other attitudes, such as sexism [30,31] and aggressiveness [32]. Thus, it can be said that cognitions relating to violent offenses may operate differently depending on sex [4]. Violence and aggression are not exclusively male characteristics.
A systematic review of group work interventions in UK high secure hospitals
2018, Aggression and Violent BehaviorCitation Excerpt :Young et al. (2010) compared pre- post-treatment assessments for patients who completed the programme (n = 22) to that of a waiting list control group (n = 12; 8% medium secure, 92% high secure). Results demonstrated that the treatment group made a significant improvement in violent attitudes and behaviour as measured by the Maudsley Violence Questionnaire (Walker, 2005) and Disruptive Behaviour and Social Problems Scale (Young, Gudjonsson, Ball, & Lam, 2003). Although Young et al. (2010) report positive findings for R&R2M the small and unevenly matched control group prevents any firm conclusions being made regarding the effectiveness of the programme as no direct statistical comparisons could be made between the two groups.
A Systematic Review of Autistic People and the Criminal Justice System: An Update of King and Murphy (2014)
2023, Journal of Autism and Developmental Disorders