Title: Modus Operandi and Personality Disorder in Incarcerated Spousal Killers

Authors: Donald G. Dutton *, Greg Kerry†

*Professor of Psychology, University of British Columbia, Vancouver, BC, Canada

†Psychologist, Correctional Services of Canada, Warkworth Penitentiary, Campbellford, Ontario, Canada

Information on the corresponding author: Prof. Donald G. Dutton, Department of Psychology, University of British Columbia, Vancouver, Canada, V6H 1Z4. Tel: 604-822-2151; Fax: 604-822-6923

Previous studies of uxoricide (wife murder) have tended to focus on the ratio of uxoricide to femicide (killing of women), prior histories of physical abuse, use of excessive force ("overkill"), and the role of estrangement as risk factors. However, little attention has been directed toward personality disorder in perpetrators of uxoricide and whether personality disorder relates to the modus operandi pattern of the murder.

Campbell (1992) found that 28 of 65 femicides in Ohio were uxoricides. Of these, 18 of 28 (68%) had recorded histories of previous physical abuse, and 17 of 28 (61%) used "excessive force" during the killing. Wolfgang (1956) had found that 44 of 53 husbands who killed their wives did so with excessive force. His definition of excessive force was two or more separate actions of stabbing, cutting, or shooting, involved in the process of slaying the victim. Cooper (1994) also reported excessive force as frequent (51% of cases of uxoricide).

In a study of 896 woman killings (femicides) with identified perpetrators in

Ontario from 1974 - 1990, Crawford & Gartner (1992) found that 551 (62%) were killed by intimate partners. Of all femicides where a motive for the killing could be established from police records, 32% were "estrangement killings", another 11% were

based on beliefs that the female partner was sexually unfaithful, a variation of the abandonment fear without actual physical estrangement. (Dutton, 1995).

Crawford & Gartner (1992) also reported excessive force as frequent (60% of uxoricides).Crawford & Gartner (1992) found that 551 of 896 (62%) of femicides in Ontario were uxoricides. Of the uxoricides, 297 had prior recorded histories of domestic violence, 166 had records of threats to the woman, and 130 had records of prior police intervention. Both Stout (1993) and Cooper (1994) reported that use of a knife was the most common modus operandi of uxoricide. Both studies found that 41% of uxoricides were the result of stabbings. The use of firearms declines from non-family murders

(60% ) to family murders (40%) (Dawson & Langan 1994).

Biro, Vucovic & Djuric (1992) examined MMPI profiles of 112 convicted homicide perpetrators in Eastern Europe. They found that 49% of their sample showed a "hypersensitive-aggressive" profile characterized by scale elevations on the Pa and D scales, in other words, exhibiting signs of paranoia and depression. Factor analysis of MMPI scores revealed six factors. One of these (Factor 5) depicted "overcontrolled" murderers (Megargee 1966), however, the authors did not analyze personality factors by spousal versus stranger homicides.

Kalichman (1988) collected MMPI data on 16 women and twenty men convicted of murdering spouses and 19 other men convicted of murdering strangers during the course of a crime. Although there was a high incidence of psychopathology in each group, the patterns varied. Sixty nine percent of the female perpetrators had no significant scale elevations on the MMPI, whereas only 15% of the male perpetrators had "normal" profiles (i.e. no scale elevations). The most frequently elevated scale in the male uxoricide group was Pd. (35% had elevations over T scores of 70). Men who had killed strangers had more scale elevations than the uxoricidal group, with 37% showing elevations on the Ma scale as well as the Pd scale. We could find no previous studies reporting MCMI profiles of murderers, although MCMI profiles of non-lethal spouse assaulters have been reported (Dutton & Starzomski 1993).

Wilson & Daly (1993) point out that equality in the sex ratio of killing does not connote equal motives for spousal homicide. One motive that appears disproportionately by sex in the homicide data is what they call "estrangement" , a misnomer that appears to mean recent, or immanent abandonment. Abandonment means that the eventual perpetrator was left or expects to be left by the eventual victim; whereas estrangement means simply that the perpetrator and victim are separated. Clearly, if the male left, he would have a reduced motive to kill.

Recent abandonment serves as a risk factor for uxoricide but not mariticide (husband killing). Wilson & Daly (1993) examined police records in Canada, Australia (New South Wales), and the U.S. (Chicago) and found that estrangement was associated with an elevated risk of uxoricide. (A risk factor of 3 (Chicago) , 6 (Australia) or 7.5 (Canada)). This effect did not replicate for husband killings. Comparable increases in mariticide for separated men are as follows: Chicago and Canada the risk factor was 2, in New South Wales it was zero. Hence, abandonment appears to disproportionately affect uxoricide.

The duration of estrangement is also of note. Wallace (1986) reported that 47% (15\38) slain wives in Australia were killed within the first two months and 91% (29\38) within the first year of separation. Comparable data are reported for Chicago (50% within 2 months, 85% within a year (Wilson & Daly 1993). Stout (1993) echoed the risk involved in early stages of estrangement. In a study of 23 men incarcerated for uxoricide in Missouri, she found 52% of all uxoricides occurred during estrangement. The modal period of separation was less than one month (25%). Cooper & Eaves (1996) found that 46 of 147 "family homicides" were precipitated by separation from the partner.

Wilson & Daly (1993) point out that the temporal association between separation and uxoricide does not necessarily prove that separation caused uxoricide. For example, if a woman left a relationship where violence was escalating to a dangerous level, then the

immediate post-separation period might be a time of elevated uxoricide risk regardless of whether men respond violently to separation per se…However, the case descriptions often make it clear that the link between separation and murder is more than incidental: Homicidal husbands are often noted to have threatened to do exactly what they did, should their wives ever leave them, and they often explain their homicides as responses to the intolerable stimulus of the wife’s departure. (Wilson & Daly, 1993, p.10).

At present, there are no available published data to suggest that escalation of violence necessarily precedes uxoricide, although this belief is found in "lethality checklists"(e.g. Campbell 1995, Sonkin, 1997). There are also no published data on cases of estranged women who were abuse victims of their husband, but who were killed by someone else.

A perspective on the motivational basis for abandonment-precipitated uxoricide comes from a study of non-lethal males who had been convicted of wife assault (Dutton & Browning 1988). After viewing videotaped conflict scenarios with a theme of a women "abandoning" her husband, maritally violent men reported significantly more anger and anxiety than non-violent controls. Dutton (1995) suggested that abandonment rage had its’ origins in early development, including attachment and object relations.

Rage-motivated killings should demonstrate greater violence than other murders. Wolfgang (1958) defined violent homicides (overkill) as "… involving two or more acts of stabbing, cutting, or shooting or a severe beating" (cited in Straus 1976, p.55). Crawford & Gartner (1992) noted overkill as a frequent feature of intimate femicides (uxoricides).

..in almost 60% of the intimate femicides offenders stabbed, bludgeoned, beat, strangled, and\or slashed the throats of their victims. Often these fatal assaults involved lengthy and excessive violence. In many of these killings the offenders inflicted harm far beyond what would have been necessary to kill their victims. For example, more than half of the stabbings involved multiple stab wounds—in many cases, dozens of wounds over the entire body. Similarly, beatings and bludgeonings were typically prolonged attacks with numerous injuries. In addition, almost one-fifth of the cases involved multiple methods. Case histories indicate that it was not unusual for the killer to continue to attack his victim after it was clear she was dead." (Crawford & Gartner, 1992 p.45-46).

Eighty percent of the uxoricides occurred either in the victims’ (30%) or offenders’ (50%) home. Cazanave & Zahn (1992), in a study of 83 uxoricides, found that 46% of uxoricides but only 12% of mariticides were cases of overkill.

Goetting (1995) compared 123 femicides in Detroit (1982-1983) to other homicides in the city at that time. She found femicides somewhat less likely to be committed by gun (66% vs. 54%), and somewhat more likely to be stabbed or beaten with a weapon (29% vs. 35%). Another 17% of femicide victims were killed by the perpetrator’s bare hands. Less than a third of the offenders had been drinking at the time of the murder. Police records indicated that the most typical scenario (96%) was a verbal\physical confrontation over sexual indiscretions and\or the threat of terminating the relationship. Goetting (1995) describes the "homicidal marriages and other romantic relationships appear to have been strongly ambivalent in nature, and the deadly act seems to have dissipated hateful sentiments on the part of the offender, leaving a sense of despair at the loss of the loved one" (p.26). Similar findings were obtained by Campbell (1992), in a study of 73 femicides in Dayton, Ohio (1975 - 1979). 52% were killed by sexual intimates (half from whom they were estranged), 64% were known to have been previously physically abused by the perpetrator. In 61% of the cases overkill or excessive violence was used. Police reports of male jealousy indicate that it was the most frequent motive. Campbell’s investigation indicates that conjugal paranoia was frequent. No studies of the modus operandi of intimate killings report comparable data for non-intimate femicides.

Marzuk, Tardiff & Hirsch (1992) report 1000-1500 homicide-suicides in the U.S. annually. Coid (1983), in a review of homicide-suicide in 17 studies and 10 nations, found that incidence rates were remarkably constant (averaging .2 - .3\100,000) despite greater variation in national suicide or homicide incidence rates. Marzuk et al. (1992) describe the most common form of homicide-suicide as "amorous jealousy", which involves one half to three quarters of all murder-suicides in the U.S. Amorous jealousy involves beliefs, real or delusional, of the partners’ infidelity. It is usually committed by males aged 18-60. The beliefs range from ruminative to obsessional; the latter being termed the "Othello Syndrome", and accompanied by irritability, despondency and murder. As the authors put it "more often there has been a chronically chaotic relationship fraught with jealous suspicions, verbal abuse and sublethal physical violence. The triggering event is often the female’s rejection of her lover and her immediate threat of withdrawal and estrangement" (op. cit., p. 3180) Cooper & Eaves (1996) reported 20 of 65 intimate partner homicides were followed by suicide or by a suicide attempt. Rosenbaum (1990) compared 12 homicide-suicide perpetrators with 24 intimate killers who were not suicidal. While both groups had relationships that were "chaotic" with histories of physical abuse, and all perpetrators were extremely jealous, murder-suicide perpetrators were exclusively male and were depressed at the time of the murder ( the murder-only group was 50% female perpetrators).

The present study was designed to explore the behavioral contingencies of spousal homicide in order to establish whether patterns of modi operandi exist. A behavioral contingency means, given that action A occurred, is the probability of action B affected? Inter alia, we were interested in evidence for estrangement, overkill, suicidal impulses and the relationship of each of these to the general m.o. of the murder.

Secondly, we wanted to know whether certain forms of personality disorder are related to the modus operandi of spousal homicide. Dependent men, for example, should be more likely to enact an estrangement murder with abandonment rage and overkill. Antisocial personality disorder, on the other hand, may kill more instrumentally. To ascertain this, The Millon Clinical Multiaxial Inventory (version 2) was given to all respondents who had not already had this assessment made by the correctional system. MCMI profiles of spousal killers are compared to those from a non-lethal group of men in treatment for spousal assault.

Method

Interviews and inspection of institutional records were conducted with 90 federally incarcerated prisoners in Canadian prisons. All men had been convicted of spousal homicide. Institutional records included detailed reports of the actual killing based on police and court records. In addition,for 50 men , psychiatric reports, including the Millon Clinical Multiaxial Inventory (version 2) and statements made by the perpetrators at the time of the homicide were collected. Comparisons were made with 50 non-lethal spouse abusers drawn from a court-mandated treatment programme based in Vancouver, B.C. Age and race demographics for the two groups were similar, both had average ages of 32 and were predominantly Caucasian.

From the institutional records, a coding sheet was derived that included information on whether 1) evidence existed for the homicide being planned 2) whether a weapon was used 3) the type of weapon 4) the number of blows, stabs etc. 5) whether another person was killed 6)whether the perpetrator confessed 7) whether a suicide attempt was made 8) whether the perpetrator attempted to flee 9) whether the perpetrator called the police 10) whether the perpetrator tried to hide the body 11) whether evidence existed for estrangement 12) whether there was a recorded history of prior violence and 13) the location of the homicide.

The Millon Clinical Multiaxial Inventory (version 2, Millon, 1987), a 175-item self-report inventory, was designed to assist psychologists, psychiatrists and others making assessment and treatment decisions for persons with interpersonal and emotional problems. The scales on this test measure constructs closely related to those in the DSM-IV.

Millon (1987) suggests that scores over 75 on the clinical syndromes (scales A through T) and severe syndromes (scales SS through PP) are suggestive of a disorder, and scores over 85 provide strong evidence for a disorder. For the severe personality pathologies, scales S through P, scores over 75 suggest moderately severe levels of chronic personality functioning, and scores over 85 suggest a more decompensated personality pattern. Millon also suggests examination of the highest two or three clinical personality scale scores, scales 1 through 8B.

Results Insert table around here

As can be seen in Table 1, most (78 of 90) of the uxoricides in this sample were reactive, unplanned events. Surprisingly, in almost a third of the cases, the victim was killed with the perpetrators bare hands. Where a weapon was used, it was usually a knife, the most prevalent and accessible weapon. Evidence for "overkill" was present in the modus operandi of these murders, with 5 to 25 blows or stabs occurring according to coroners’ reports. It was rare for another person to be killed (9 of 90) but when this occurred, it was generally the victims’ current lover.

Also somewhat surprisingly, 20 of 90 killers called the police themselves. About a third attempted suicide but another third tried to hide the body. Evidence for estrangement killings was strong. About two thirds of all uxoricides were estrangement motivated, these typically had histories of domestic violence (56\60), and prior separations 54\60. ( See table 2).

Presence of personality disorder was determined using the MCMI criterion cutoff of 85^th percentile above baseline. According to this criterion, the most prevalent type of personality disorder associated with uxoricide were the Passive-Aggressive (M =88.6), Avoidant (M = 80.1), Self Defeating ( 81.0) and Dependent (M = 78.8) diagnoses. These "overcontrolled" personality types are generally believed to suppress rage. (See Table 3).

In comparison to non-lethal spouse assaulters, these "overcontrolled" personalities were more prevalent than the more expressive types (Antisocial, Aggressive-Sadistic).

Estrangement killings were exclusively committed by men with overcontrolled "dependent" type personality disorders. By contrast , Antisocial Personality Disorder. engaged in instrumental killings that were more likely to be planned. (See Table 4). Antisocial Personality Disorders were more likely to hide the body ( t = 3.4, p <.01) and less likely to report post homicide suicidal feelings (t= 4.1, p<.01).

Men who kill their wives and who complete valid MCMI reports are almost invariably personality disordered. This finding is consistent with the findings of Hart, Dutton & Newlove (1994) that, as severity of non-lethal physical abuse of spouses increased, so did the likelihood of personality disorder. The surprising finding in the current research however, is that the so called "overcontrolled’ personality disorders are overrepresented amongst spousal killers. Passive aggressive and Dependent personality disorders are most frequent, and Antisocial personality disorder less frequent compared to non lethal spouse abusers.

Also, clear patterns of modi operandi occur with various personality disorders. Overcontrolled men are more likely to commit abandonment murders, to attempt suicide after the murder, and to kill "reactively". Many of the murders occurred during a failed reconciliation, or when a woman first announced her intentions to leave.

By contrast, Antisocial men killed for instrumental reasons ( such as to collect insurance policies) and their killings were planned in a "cold-blooded" way. The abandonment killing may best be explained by the concept of "catathymic crisis".

Impending or recent abandonment can precipitate what Wertham (1937), Revitch & Schlesinger (1981), and Meloy (1992) call a "catathymic crisis". Catathymic crisis is characterized by a seemingly unsolvable psychic state of chronic and aversive emotional tension (called an incubation period), projection of responsibility for the internal tension

state onto the external situation (the marriage or the spouse), with the perception arising that violence is the only way out of the situation. After this prolonged period of internal conflict and rumination, the violent act is carried out. The internal tension dissipates after the violent act (with no insight), "superficial normalcy" is apparent for several months and internal psychic equilibrium is eventually established with development of insight. Meloy cites Wertham (1937) as first applying this concept to criminal psychology. Wertham viewed the violent act as an "expression of the fight on the part of the patient for safeguarding of his personality". As such, the act derives from the "psychological deep structure", that is the "inborn (sic) modes of organizing experience that provides the templates for representations of self and others. The actual representations, and their respective affects, are the product of early learning and subsequent development."(Wertham, cited in Meloy, 1992, p. 41).

Meloy cites a study by Weiss et al. (1960) of "sudden murderers" who commit a "single, isolated, unexpected episode of violent, impulsive acting out behavior - behavior which is never well thought out and has no obvious purpose or hope for personal advantage" (Weiss, Cited in Meloy, 1992, p.43). Such sudden murderers came from large, mother dominated families with hostile or rejecting fathers. Attachment between mother and child was ambivalent, characterized by "an underlying hostility that had to be repressed or split off during affectionate moments by both parties" (Weiss, Cited in Meloy, 1992, p.44). The sudden murderers differed from habitual criminals in two important ways; they showed the highest dependency needs on assessment, and they did not use projection completely successfully, in that they were intermittently aware of their own failures. This latter aspect of their personality makes them similar to suicidal people as described by Baumeister (1990). That is, both sudden murderers and suicides appear to have a pre-violence acute notion of themselves as failing. It is this notion, in the case of suicides, that creates "aversive self awareness" from which the suicide is an attempt at escape. In the case of some homicides, the aversive feelings are seen as being caused by the eventual object of murder (see also Dutton & Yamini 1995).

Meloy (1992) describes catathymic violence as "delineated by intense autonomic arousal, overwhelming anger during the violence, the perception of the victim as an imminent threat to the ego structure of the perpetrator, a time limited behavioral sequence, and a goal of threat reduction and a return to homeostasis" (Meloy, 1992 p. 47). Revitch & Schlesinger (1981) described catathymic violence as having three stages; incubation, violent act, and relief. The incubation stage includes cognitions that are marked by an "obsessive preoccupation with the future victim" and affect states including depression, frustration, and helplessness. The victim is typically known to the perpetrator, and an emotional bond or attachment exists between them.

Meloy (1992) describes the lack of an integrated identity in perpetrators of catathymic violence. Self and object representations vacillate between good or bad, and are not held constant. Meloy cites the "homicidal-suicidal wish" that emerges during the incubation stage "Either he or I must die, something has to give" (Meloy, 1992, p. 58). During the incubation stage, "there is no consciousness of the good self or good object, instead, what is focused on, is the aggressive aspect of the self-object dyad". This polarization of self and object representation as either all good or all bad is comparable to that demonstrated by children aged 2 - 7, after the development of the notion of a separate self (Mahler, Pine & Bergman 1975), but before the development of "conservation of the self" (Kagan 1981). The fixation at this level may indicate profound disturbances in the interpersonal relationship with the parent occurring at that age in the subsequent perpetrator.

Meloy (1992) describes another feature of catathymia which is essential in fully understanding this form of aggression. In catathymia what occurs is "oscillation of the destructive wish", also referred to as projective-introjective cycling. That is, the origins of aggression, pain and dysphoria are alternately projected onto the "object or introjected into the self. The suicidal or homicidal nature of the aggressive wish depends on the perceived origin of the emotional distress, which may rapidly oscillate" (p.59). Although such thinking still assumes a distinction between self and other, the blurring of the origin of the pain or the "attributional dilemma" suggest an imperfect separation. Confusion is apparent about the origin of psychological content: Who is doing what to whom? Am I feeling a certain way or is she?

Meloy (1992) also points out that typically such thinking is extreme; more moderate choices (such as leaving the relationship) are never consciously considered. "Splitting, a characteristic of borderline personality organization, may occur during the experience of intense, polarized and alternating affect states. Rage may be felt toward the self and then toward the object. The intensity of the rage at the moment of violence also implicates splitting, since there is no blocking or neutralization of discharge that is seen in higher level repressive operation" (p. 60). Aggressive drives remain raw, unmodulated and "split off" from conscious expression during the incubation phase. For this reason, the actual violence is often experienced as "out of character" and "overwhelming " by the perpetrator. Typically, the perpetrator experiences some tension release during the violent act and subsequent relief.

The clinical description of catathymic homicides is consistent with abandonment-precipitated rage killings. The pent up rage, for example, generates "overkill" as described by Crawford & Gartner (1992). The abandonment probably initiates the catathymic crisis: rumination on the thought that "if I can’t have her, no one else will" (Campbell 1992), or on the notion that the woman’s act of leaving and the image of her being with another man is the cause of the perpetrator’s psychic pain. Dutton & Browning (1988) found that, even in non-lethal abusive men, videotapes of a woman expressing her desire for more independence produced greatly elevated levels of both anger and anxiety, over and above those reported by non-abusive men. The notion of catathymic crisis raises the issue of personality disorder and homicide. All the "overcontrolled" types of personality disorder found to be overrepresented in this study (compared to non-lethal wife assaulters) are men who have self-disturbances as part of their clinical makeup. However, borderlines also have self disturbances. Borderlines were represented amongst both lethal (34%) and non-lethal (38%) groups in this study.

Some risk assessment scales have been published (Sonkin, 1997, Campbell, 1995), however, these scales appear to better predict the kind of abusive personality who would terrorize a woman (eg. Psychopathic, aggressive-sadistic). For example, a scale published by Campbell (1995) was based on a woman’s belief that her partner might kill her. Certain terrorist tactics, such as choking to the point of the woman losing consciousness contribute to this belief. However, this belief notwithstanding, the personality disorder most likely to actually kill his spouse is dependent and passive-aggressive, not the profile predicted by these scales. A study by Showalter, Bonnie & Roddy (1980) concluded that, "especially significant was the fact that most of these men (spousal killers) lacked recorded histories of assaultive or other socially disturbing behavior…in only two cases had the offender had any previous contact with the mental health system and in both the intervention was precipitated by spousal conflict." (Showalter, Bonnie, & Roddy, 1980, p. 125).

Hence, there is some evidence that spousal homicide may not necessarily be predictable on the basis of non-lethal violence. The "risk prediction" notion that escalating violence combines with other factors to make homicide likely has no empirical support. Suppressed rage rather that expressed violence may be more indicative of subsequent spousal homicide. However, this conclusion must remain tentative. One retrospective study found that 90% of spousal homicide cases had prior interventions by police for "domestic disturbance" (Wilt & Breedlove, 1977), although there was no analysis of whether the violence was escalating. Overcontrolled-Dependent men are still at risk for prior wife assault, and a more comprehensive study, involving interviews with intimate friends or families of the victim would probably reveal a higher incidence of prior domestic violence than institutional records. Nevertheless, if we consider personality disorder as a risk factor for spousal homicide, Overcontrolled-Dependent men appear to have been overlooked by prior risk assessments.

Some limitations of the present study should be noted. The data were reconstructed from institutional files. As mentioned above, some types of information, especially histories of prior abusiveness may be underreported by this method. Subsequent studies should include an interview with an intimate friend of the victim. Also, a question must always be raised about whether an incarcerated sample is representative of all perpetrators, although in the case of spousal homicide, the arrest-conviction-incarceration rates are quite high.

References

Baumeister, R.F. (1990). Suicide as escape from self. Psychological Review, 97, 90-113.

Biro, M., Vucovic, N., & Djuric, V. (1992). Towards a typology of homicides on the basis of personality. British Journal of Criminology, 32, (3) , 361 – 371.

Campbell, J.C. (1992). "If I can’t have you, no one can": Power and control in homicide of female partners. In J. Radford & D.E.H. Russell (Eds.) Femicide: The Politics of Woman Killing. New York: Twayne Publishers.

Campbell, J.C. (1995). Assessing Dangerousness: Violence by sexual offenders, batterers and child abusers. Newbury Park: Sage.

Cazanave, N.A., & Zahn, M.A. (1992). Women, murder and male domination: Police reports of domestic violence in Chicago and Philadelphia. In E.C. Viano (Ed.) Intimate Violence:Interdisciplinary Perspectives Washington, D.C.: Hemisphere Publishing Corp.

Coid, J. (1983). The epidemiology of abnormal homicide and murder followed by suicide. Psychological Medicine, 13, 855 – 860.

Cooper, M., & Eaves, D. (1996). Suicide followed by homicide in the family. Violence and Victims, 11(2), 99-112.

Cooper, M. (1994). Wasted Lives: The Tragedy of Homicide in the Family.

Vancouver: BCIFV.

Cote, R. & Hodgins, S. (1992). The prevalence of major mental disorders among homicide offenders. International Journal of Law and Psychiatry, 15, 89 – 99.

Crawford, M. & Gartner, R. (1992). Woman Killing: Intimate Femicide in Ontario 1974 – 1990. Ontario: Woman’s Directorate: Ministry of Community and Social Services.

Dawson, J. & Langan, P. (1994). Murder in Families. (Special Report, Bureau of Justice Statistics). Washington, D.C.: US Dept of Justice.

Dutton, D.G. (1995). Male abusiveness in intimate relationships. Clinical Psychology Review, 15(6), 207-224.

Dutton, D.G., & Browning, J.J. (1988). Power struggles and intimacy anxiety as causative factors in wife assault. In G. Russell (Ed.), Violence in Intimate Relationships. Great Neck, NY: PMA publishing. (22pages).

Dutton, D.G., & Starzomski, A. (1993). Borderline personality in perpetrators of psychological and physical abuse. Violence and Victims, 8(4), 327-338.

Dutton, D.G. & Yamini, S. (1995). Cognitive deconstruction and projective-introjective cycling in cases of adolescent parricide. American Journal of Orthopsychiatry, 65(1), 39-47.

Dutton, D.G., Starzomski, A. & Ryan, L. (1996). Antecedents of borderline personality organization in wife assaulters. Journal of family Violence, 11(2), 7(4), 113-132.

Goetting, A. (1995). Homicide in Families and Other Special Populations. New York: Springer.

Hart, S.D., Dutton, D.G. & Newlove, T. (1993). The prevalence of personality disorder among wife assaulters. Journal of Personality Disorders, 7(4), 329-341.

Kalichan, S. (1988). MMPI profiles of women and men convicted of domestic homicide. Journal of Clinical Psychology, 44, (6), 847 – 852.

Kagan, J. (1981). The Second Year: The Emergence of Self-Awareness. Cambridge, MA.: Harvard University Press.

Mahler, M. Pine, F., & Bergman, A. (1975). The Psychological Birth of the Human Infant. New York: Basic.

Marzuk, P.M., Tardif, K. & Hirsch, C.S. (1992). The epidemiology of murder-suicide. Journal of American Medical Association, 267(23), 3179-3183.

Megargee, E.I. (1966). Undercontrolled and overcontrolled personality types in extreme anti-social aggression. Psychological Monographs, 80, 611 –617.

Meloy, J.R. (1992). Violent Attachments. Jason Aronson: Northvale, New Jersey.

Millon, T. (1987). Manual for the MCMI (2^nd Ed.). Minneapolis, Minnesota: National Computer Systems.

Revitch, E. & Schlesinger, L.B. (1981). The psychopathology of Homicide. Springfield, Illinois: Charles Thomas.

Rosenbaum, M. (1990). The role of depression in couples involved in murder-suicide and homicide. American Journal of Psychiatry, 147, 1036-1039.

Showalter, C.R., Bonnie, R.J. & Roddy, V. (1980). The spousal homicide syndrome. International Journal of Law and Psychiatry, 3, 117 – 141.

Sonkin, D. (1997). The perpetrator assessment handbook. Sausalito,CA.

Stout, K. (1993). Intimate femicide: A study of men who have killed their mates. Journal of Offender Therapy, 19, ¾, 81 - 94.

Straus, M.A. (1976). Sexual Inequality, cultural norms, and wife beating. Victimology, 1, 54-76.

Wallace, A. (1986). Homicide: The Social Reality. Sydney: New South Wales Bureau of Crime Statistics and Research.

Weiss, J., Lambati, J. & Blackman, N. (1960). The sudden murderer: A comparative analysis. Archives of General Psychiatry, 2, 669-678.

Wertham, F. (1937). The catathymic crisis. Archives of Neurology and Psychiatry, 37, 974 – 978.

Wilson, M. & Daly, M. (1993). Spousal homicide risk and estrangement. Violence & Victims, 8, 1, 3 –16.

Wilt, M. & Breedlove, R. (1977) Domestic violence and the police:Studies in Detroit and Kansas City. Police Foundation:Washington, D.C.

Wolfgang, M.E. (1958). Patterns of Criminal Homicide. Philadelphia: Pennsylvania Press.

Wolfgang, M. (1956) Husband-wife homicides. Journal of Corrective Psychiatry and Social Therapy, 2, 263 -271.

Evidence for planning? Yes 12 No 78.

Weapon used? Yes 62 No 28

Number of blows or stabs (Range) 5 - 25

Another person killed? Yes 9 No 81

Perpetrator calls police? Yes 20 No 70

Suicide attempt? Yes 30 No 60

Try to hide body? Yes 35 No 55

Evidence for

estrangement? Yes 60 No 30

History of

domestic violence? Yes 60 No 30

Prior separations? Yes 70 No 20

X (s.d.) %>85 X (s.d.) % >85 t . p.

Instrumental Estrangement

Antisocial (n=15) 15 0

Dependent\

Avoidant\

Self Defeating (n = 26) 0 26