Understanding male domestic partner abusers

Foreword | Research in the past decade has found that certain ways of responding to domestic partner abusers, particularly psycho-educational approaches, can be effective in modifying abusive behaviours. The study described in this paper sought to classify male domestic partner abusers by certain identified characteristics and determine whether they responded differently to a Men's Behaviour Change Program conducted by community agencies in regional Victoria. It was found that some types of male abusers appeared to derive greater benefits from the standard 12-week program than others. Those with antisocial personality disorders tended not to respond favourably. Further research work is now required to identify interventions that will be successful with abusers who have significant antisocial personalities.

Toni MakkaiActing Director

Previous research has found that the psychological, behavioural and physiological characteristics of male domestic partner abusers may be used to assess suitability for interventions aimed at changing behaviours (Jacobson, Gottman & Shortt 1995). For example, it appears that men with antisocial personality characteristics may respond less favourably to certain psycho-educational interventions than men with other characteristics (Gottman et al. 1995; Walker 1995). A failure to match types of abusers to types of interventions can lead to less than optimal outcomes (Gondolf 1997; Gottman et al. 1995; Ornduff, Kelsey & O'Leary 1995; Walker 1995). The development of a well-defined typology of men who abuse their domestic partners is important for increasing positive intervention outcomes (National Crime Prevention 1999).

Abuser typologies

Research into the physiological reactions associated with male abusive behaviour has suggested there are two distinct types of abusers - type 1 and type 2 (Ornduff, Kelsey & O'Leary 1995; Walker 1995; National Crime Prevention 1999; Jacobson, Gottman & Shortt 1995). The typology is outlined in Table 1. However, more research is required to confirm the distinction between type 1 and type 2 and to investigate further the psychological, behavioural and emotional characteristics of these types. In particular, the roles of anger and hostility need to be clarified. Furthermore, while evidence seems to be accumulating that both types of abusers may benefit from different interventions, more research is required to clarify how each group responds to specific programs (Walker 1995).

Table 1: Typology of domestic violence abusers
Type 1 abusers Type 2 abusers
Lowered heart rates in response to conflict with their domestic partner Increased heart rates in response to conflict with their domestic partner
Exhibit increased mental vigilance and sadism associated with their apparent suppression of emotional arousal Display other signs of over-arousal in the face of escalating conflict, including increases in emotional involvement and overt anger
Exhibit antisocial personality disorder characteristics (for example, diminished conformity to social norms, consistent disregard for the rights of others, repeated deceitfulness, repeated physical fights or assaults) More likely to exhibit characteristics of borderline personality disorder - for example, pattern of unstable relationships, unstable sense of identity, emotional instability, fears of abandonment
  May further subdivide into those who have explicit psychopathology and those who do not

The present study aims to:

  1. investigate further the distinction between type 1 and type 2 abusers; and
  2. investigate whether types of abusers responded differently to participation in the Men's Behaviour Change Program conducted by community agencies in regional Victoria. The Men's Behaviour Change Program typically includes instruction around power and control issues, gender role attitude restructuring and anger management. That is, the focus is on the abuser assuming responsibility for his abusive behaviours, developing non-oppressive attitudes to women, and learning ways to manage and reduce angry and violent behaviours.

The study predicted that:

  • type 1 abusers would exhibit characteristics associated with antisocial personality disorder, and type 2 abusers' characteristics would be associated with borderline personality disorder;
  • measures of hostility, anger, sexist attitudes, tactics used in partner conflict, and measures relating to frequency and severity of violence would be different between type 1 and type 2; and
  • type 1 abusers would be less responsive to the program than type 2.


Male participants

One hundred self-referred men participated in the study. They were recruited over a period of months through counsellors at four organisations adhering to the 'No to violence' guidelines. The men were invited to participate in the research after being identified by organisations during initial assessment interviews as current partner abusers. Men deemed not to be in relationships and/or men who did not participate voluntarily were excluded (for example, court-mandated clients). Groups of participants were assessed on three occasions:

  1. prior to participation in the 12-week Men's Behaviour Change Program (test 1);
  2. at the end of the program when 30 participants were re-assessed (test 2);
  3. after a further six months when 14 participants were followed up.

Participants were paid $25 for each assessment. The results of the six-month follow-up are not reported here. The source and age of participants is described in Table 2. The overall characteristics of participants are described in Table 3.

Test 1

At the first detailed assessment, men filled out six questionnaires which took 50 minutes to complete (see Box 1). A multi-parameter pulse-oximeter was then used to assess physiological responses.

Initially, the men were asked to sit quietly with their eyes closed for two minutes while non-invasive baseline measures of heart rate, blood oxygen saturation and blood pressure (all estimates of autonomic nervous system arousal) were taken. They were then asked to listen to an audiotape featuring three hypothetical situations, recorded successively. Each scenario lasted four minutes.

There was a neutral scenario, and two scenarios describing hypothetical interactions with their partner - designed to engender jealousy and then anger. The men were encouraged to imagine vividly that they were involved in each scenario. Their physiological responses were recorded at the end of each scenario - that is, at the four, eight and 12-minute marks. The men were debriefed for 15 to 30 minutes.

Table 2: Source and age of participants
  Number Age range (years) Mean age (years)
Recruiting agency Test 1 Test 2 Test 1 Test 2 Test 1 Test 2
A 44 16 19-56 19-55 42.5 40.5
B 15 8 22-58 24-51 43.1 41.1
C 30 0 21-59 n/a 39.1 n/a
D 11 6 24-55 22-54 37.7 39.8
Total 100 30 19-68 19-58 45.2 43.8

Test 2

Thirty men were re-assessed after finishing the Men's Behaviour Change Program. In test 2, all testing undertaken in test 1 was repeated except for the MCMI-3 (which measures personality traits, stable over time).

Female participants

All 100 men were asked for permission to interview their partner (after participating in the Men's Behaviour Change Program), to source further information about the men's patterns of abusiveness. Only a small proportion agreed. Of these, the partners of the men who were re-assessed after finishing the program were invited to participate in the study.

Fourteen women were interviewed by a female research officer - a social worker with considerable experience in working with women and children who had experienced violence at home. Eight of the women were partners of type 1 abusers, and six of type 2 abusers. Eight were living with their partner when interviewed.

The invitation to participate was sent in the form of a letter at the end of the program and included a consent form and a conflicts tactics scale. Women who returned the form and scale were contacted by the interviewer to set up an appointment. The interview included exploring qualitative themes relating to abuse. Of particular interest were themes relating to changes in behaviour made by the abuser over the program and the women's views of that behaviour. The interviews lasted between 40 and 60 minutes. At the end, women were debriefed for 15 to 30 minutes and paid $25 for participating.


Characteristics of type 1 and type 2 abusers found at test 1

Men were partitioned into type 1 and type 2 groups depending on whether their heart rate in the anger scenario was higher or lower than at baseline. Forty-seven men had lower heart rates and were classified as type 1. Fifty-three had higher heart rates and were classified as type 2. However, type 1 men had higher heart rates than type 2 both before listening to the anger scenarios and during the neutral scenario. During the anger scenario their heart rates dropped. By contrast, type 2 men had lower heart rates before listening to the anger scenarios and during the neutral scenario. Their heart rates rose during the anger scenario. During the anger scenario, the actual heart rates between the two types did not differ. It appears, therefore, that the respective shifts in heart rate during the anger scenario may be a result of where their heart rates were at the beginning of the anger-inducing tape rather than a real difference between the two types.

Box 1: Questionnaires

The following six questionnaires were completed by all participants.

1. Millon clinical multiaxial inventory-3 (MCMI-3)

This questionnaire has 22 scales that measure anxiety, somatoform, bipolar-manic, dysthymic and mood disorders, and alcohol and drug abuse and dependence. It also measures personality disorders such as dependent, antisocial, aggressive-sadistic, passive-aggressive, borderline and paranoid disorders.

2. State-trait anger expression inventory-2

This questionnaire measures state and trait anger (the latter consists of angry temperament - temperamental disposition to experience and express anger without specific provocation), and angry reaction (the propensity to react with anger when treated unfairly, criticised or frustrated).

The questionnaire also measures:

  • anger-in - tendency to suppress angry feelings;
  • anger-out - tendency to express anger outwardly toward other people or objects in the environment;
  • anger control-in - tendency to constructively manage suppressed anger; and
  • anger control-out - ability to constructively manage outward expressions of anger.

3. The macho scale

This scale has 28 items measuring individual differences in sex-role stereotypical beliefs. High scores indicate more sexist beliefs about gender-appropriate behaviour.

4. Conflicts tactics scale

This standardised scale measures the frequency and severity of 19 tactics used in relationships to resolve conflict. It is a well accepted measure of inter-partner violence. Tactics are grouped into three subscales which assess reasoning, verbal abuse and physical violence.

5. Cook medley hostility scale

This questionnaire measures hostility. High scores suggest attitudes and behaviours indicative of resentment, bitterness, cynicism and mistrust of others. Five subscales have been identified in this scale: cynicism, hostile attribution, hostile affect, aggressive responding and social avoidance.

6. Buss-Durkee hostility inventory

This measures hostility. High scores suggest attitudes and behaviours indicative of suspicion, anger, cynicism and negativity with respect to others. The inventory comprises seven subscales designed to tap different aspects of hostility. The suspicion subscale represents the cognitive component. Irritability and resentment subscales measure the affective component. Four subscales assess behavioural components: assault, indirect hostility, verbal hostility and negativity.

Before listening to the tapes, the diastolic blood pressure for each type was similar. However, during the anger scenario it was significantly lower for type 1. It had decreased from baseline for type 1, but increased from baseline for type 2. Therefore the study found partial support for the type 1 and type 2 abuser distinction, but the distinction may be related to what happens to diastolic blood pressure more than heart rate when confronted with anger-inducing scenarios.

Discriminant function analyses were conducted to determine which physiological and psychological measures distinguished the two groups. Type 1 men reported higher levels of assaultiveness and verbal aggression, and stronger sex-role stereotypical attitudes than type 2. Contrary to expectations, however, type 2 men were more likely to endorse aggression as a problem-solving strategy. These results are summarised in Table 4.

Identification of groups based on physiological variables at test 1

Physiological variables for all taped scenarios - heart rate, blood oxygen saturation, and systolic and diastolic blood pressure readings - were subjected to a principal components analysis to investigate how they grouped. The study found that heart rate measures at baseline and during neutral, jealousy and anger scenarios grouped together. The systolic blood pressure measures at these times also grouped together, as did diastolic blood pressures and, separately, measures of blood oxygenation during these four periods.

Cluster analyses were then conducted to determine whether this cohort of men could be partitioned into subgroups based on these physiological factors. Three groups of men were identified, based upon their patterns of physiological response. Thirteen men in physiological group 1 (PG1) were characterised as being primarily heart rate responders - they were identified as a group by higher heart rates and higher diastolic blood pressures, but lower systolic blood pressures and blood oxygenation. The 44 men in physiological group 2 (PG2) were identified mainly by their consistently low diastolic blood pressures. The 43 men in physiological group 3 (PG3) were characterised almost exclusively by consistently higher diastolic blood pressures, but lower heart rates.

Table 3: Characteristics of participants
Characteristic Test 1 (n=100) Test 2 (n=30)
Employment Employed - 73Unemployed - 17Between jobs - 2  
Racial Caucasian - 97Koori - 3 Caucasian - 30Koori - 0
Living with partner Yes - 63All had been in contact with their partner during the past month Yes - 21 (70%)

Discriminant function analyses were conducted to determine whether psychological and psychopathological variables distinguished between the three groups. Almost none did so. The three groups could not be distinguished from each other on self-reports of the tactics they used in conflicts with their partners, nor on sexist attitudes, nor on self-reports of angry feelings and behaviours (although there was a non-significant trend for those in PG3 to report they were better at managing suppressed anger). Measures of hostility also did not discriminate between the three groups. However, men in PG2 appeared more likely to:

  • abuse drugs;
  • exhibit excitable, disinhibited behaviours; and
  • be less prone to depressed moods than those in the other groups.

Research was then undertaken to see if the types responded differently to the Men's Behaviour Change Program as assessed at test 2 - 12 weeks after initial assessment.

Specific male responses at test 2: program effects

Initially, changes in the responses of types 1 and 2, and the responses of PG1, PG2 and PG3 were compared. Change was assessed on questionnaire measures of anger, aggression, hostility, conflict tactics with partner, and sexist beliefs. The study predicted that type 1 would be less responsive to the Men's Behaviour Change Program than type 2, and that men in PG2 would be less responsive to the program than those in PG1 and PG3 (because men in PG2 were more likely to abuse drugs and to exhibit excitable, disinhibited behaviours).

Fifteen type 1 men and 15 type 2 men were assessed after the program. Multivariate analyses of variance were performed on groups of variables, followed by univariate analyses of variance when statistical significance was found.

Trait anger significantly decreased after participating in the Men's Behaviour Change Program for both groups, but there was no difference in the groups in the amount of change on this measure. There were also non-significant decreasing trends for angry temperament and angry reaction for both groups of men.

Both groups tended to show decreases on measures of assault, indirect aggression and negativity. However, they did not change on self-reports of tactics when experiencing conflict with their partner or on other measures of hostility from the Buss-Durkee hostility inventory. After participating in the program, cynicism tended to increase for type 1 abusers and decrease for type 2.

Five men from PG1, 11 from PG2 and 14 from PG3 were re-assessed after the Men's Behaviour Change Program. Again, multivariate analyses of variance were conducted followed by univariate analyses of variance when significance was found. Both state anger and trait anger significantly decreased after participation in the program for all groups. As well, there were trends for angry temperament, angry reaction, and anger control-in to decrease for both groups.

In relation to the six measures of partner conflict, self-reported use of verbal aggression in conflict with partners tended to decrease for the men as a group. After the Men's Behaviour Change Program, sexist attitudes for men in PG2 increased while the attitudes of those in PG1 and PG3 did not change.

In summary, for the 30 men assessed at time 2, participation in the Men's Behaviour Change Program was associated with decreases in levels of angry affect - probably with a reduction in angry temperament and angry reaction. Overall they experienced less anger after being on the program. Type 1 abusers, however, experienced increased cynicism. Type 2 experienced decreased cynicism.

Specific female responses

It was apparent from the female partners' responses that the men responded in diverse ways to the program. The women indicated that involvement helped some men only. Their responses re-emphasise that the Men's Behaviour Change Program is beneficial for some abusers, but not necessarily for all. The information gleaned from the partners, therefore, supports data obtained from the men.

Policy implications

Study findings support earlier research that male domestic partner abusers may be subdivided based upon their physiological responses to hypothetical tape-recorded conflict with their partner. However, this distinction may be somewhat more complex than originally thought because differences in diastolic blood pressure seemed more meaningful than differences in heart rate.

As expected, type 1 abusers scored higher on measures of assault, verbal aggression and sexist attitudes than type 2 abusers. However, somewhat surprisingly, type 2 more strongly endorsed aggression as a problem-solving technique than did type 1. Study expectations about type 1 men being more likely to exhibit antisocial personality characteristics than type 2 men were not supported. Also, contrary to expectation, the study did not find that type 2 men exhibited borderline personality disorder characteristics.

The men were also partitioned into three groups based on their heart rate, blood pressure and blood oxygenation measures of emotional arousal. Those in PG2 more closely exhibited some antisocial personality disorder characteristics than those in the other groups - they were more likely to abuse drugs, more likely to demonstrate disinhibited and impulsive behaviours, and less likely to experience dysphoric moods.

Overall, the 30 men who agreed to be assessed a second time after participating in the Men's Behaviour Change Program reported decreases in levels of angry feelings. They reported that they tended to be less hot-headed (angry temperament) and less likely to react with anger when criticised or when they perceived they were dealt with unfairly (angry reaction). It therefore appeared they were experiencing less overall anger. There was also a tendency for the men to report lower levels of assault, negativity and verbal aggression.

Table 4: Differences between type 1 and type 2 abusers
Variable/measure Significance
Diastolic blood pressure p
Hostility measures p
Sexist attitudes p
Conflict tactics not significant
Anger measures not significant
Psychopathology measures not significant

Some men, however, appeared to derive less benefit from participating in the program. Type 1 abusers tended to become more cynical over the course of the program whereas type 2 tended to become less cynical. Furthermore, the men in PG2 (those more likely to abuse drugs, and to be more excitable and disinhibited) had stronger sexist attitudes after treatment, whereas there was little change for men in the other groups. As predicted, therefore, some types of male abusers appear to derive less benefit from the standard Men's Behaviour Change Program.


In keeping with one of its main goals, the Men's Behaviour Change Program appears to teach most men how to better manage angry feelings and behaviours. As a group, the men's angry feelings in general, and reactions to provocation specifically, decreased after participating in the program. There was also a trend for verbal aggression to decrease.

Nevertheless, for some men the program was less effective in modifying undesirable attitudes towards women. These men appeared to approximate the antisocial type of abuser identified in previous research. Psycho-educational programs such as the Men's Behaviour Change Program may not therefore necessarily be the best type of intervention for this type of male abuser. Other forms of intervention may need to be developed. Some researchers have suggested that this type of abuser requires long-term psychotherapy aimed at promoting enduring cognitive and behavioural changes (Gottman et al. 1995; Ornduff, Kelsey & O'Leary 1995; National Crime Prevention 1999). On the other hand, men with pronounced antisocial personality characteristics are notoriously difficult to engage in such treatment. Clearly, further research is required on how best to manage this type of abuser. It appears, however, that the Men's Behaviour Change Program is a useful intervention for abusers who do not manifest significant antisocial characteristics.


This research was funded by the Criminology Research Council (CRC). A full report is available on the CRC web site: http://www.criminologyresearchcouncil.gov.au/.


  • Gondolf EW 1997. Batterer programs: what we know and need to know. Journal of interpersonal violence 12(1): 83-98
  • Gottman JM, Jacobson NS, Rushe RH, Shortt JW, Babcock J, La Taillade JJ & Waltz J 1995. The relationship between heart rate reactivity, emotionally aggressive behaviour, and general violence in batterers. Journal of family psychology 9(3): 227-48
  • Jacobson NS, Gottman JM & Shortt JW 1995. The distinction between type 1 and type 2 batterers further considerations: Reply to Ornduff et al. 1995, Margolin et al. 1995 and Walker 1995. Journal of family psychology 9(3): 272-79
  • National crime prevention 1999. Ending domestic violence: programs for perpetrators. Canberra: Attorney-General's Department
  • Ornduff SR, Kelsey RM & O'Leary KD 1995. What do we know about typologies of batterers? Comments on Gottman et al. 1995. Journal of family psychology 9(3): 249-52
  • Walker LEA 1995. Current perspectives on men who batter women - implications for intervention and treatment to stop violence against women. Comment on Gottman et al. 1995. Journal of family psychology 9(3): 264-27


This paper is taken from the report of research undertaken with the assistance of a grant from the Criminology Research Council.

Full report: Identification of characteristics and patterns of male domestic partner abusers

About the authors

Professor Jeffrey Richards works at Monash University. Dr Angus McLachlan, Wayne Scott and Raeleene Gregory work at the University of Ballarat.